1 Introduction

Multiferroic materials have been attracting researchers recently for their interesting fundamentals as well as for the possibility of application of these materials in different spintronic devices [1,2,3,4]. In magnetoelectric multiferroic materials, ferroic (or anti-ferroic) magnetic and electric ordering coexist in a single phase giving rise to the possibility of controlling the magnetization (intrinsic polarization) with the application of electric field (magnetic field) [1]. Due to these coupling between the two properties, magnetoelectric multiferroic materials have become one of the most important materials of today [1]. The reason for the limited number of multiferroic material is the mutual exclusive origin of the two ordering (empty d shell for ferroelectricity and partially filled d shell for magnetic ordering) [1,2,3,4].

BiFeO3 is one of the most interesting and well-studied multiferroic as it is the only multiferroic material to show both the ordering (magnetic and ferroelectric) above the room temperature (ferroelectric Curie temperature TN ~ 1103 K and Neel temperature TN ~ 643 K) [5,6,7,8]. It exhibits G-type canted antiferromagnetic ordering with a cycloid frequency of ~ 62 nm [8]. BiFeO3 shows large spontaneous polarization of order 10–100 µc/cm2 because of polar displacement of cations and anions relative to each other pointing along one of the eight pseudo-cubic [111] axes [5,6,7]. The lone pair electron at the 6 s shell of Bi is considered to be the main reason behind the observed ferroelectricity on BiFeO3 while the partially filled 3d shell of Fe is responsible for the canted antiferromagnetic ordering [5,6,7]. BiFeO3 has been found to be useful in different modern-day technologies including microwave synthesis [9, 10]. In spite of having these features, BiFeO3 is not considered suitable for many applications due to many reasons. Structural instability, difficulty in synthesizing single phase, high leakage current, and low resistivity of the material due to the presence of Fe3+ and oxygen vacancies are some of them [11,12,13,14]. To overcome the shortcomings, there have been numerous attempts involving doping of different transition metal at Fe site and rare earth metals at Bi sites, inducing chemical pressure or strain in the system [15,16,17,18,19,20,21,22]. An alternate option is to prepare composite of different multiferroic materials with a similar structure involving BiFeO3. There are reports on composite structures and superlattice structures of BiFeO3–BaTiO3, BiFeO3–PbTiO3, and BiFeO3–BiMnO3, showing improvement in multiferroic properties [23,24,25]. Yu et al. [26] have reported exchange bias and other enhanced magnetic properties in BiFeO3–La0.7Sr0.3MnO3 heterostructure due to charge transfer-assisted band reconstruction near the interface which in turn creates additional ferromagnetic and antiferromagnetic exchange interactions.

In this context, we have prepared a composite of BiFeO3 and TbMnO3, belonging to a different type of multiferroic. TbMnO3 is type II multiferroic material in which the ferroelectric ordering (TC ~ 28 K) arises as a result of magnetic ordering (TN ~ 42 K), and the material is known to possess strong magnetoelectric coupling [27, 28]. It crystallizes in orthorhombic perovskite (Pbnm) close to that of BiFeO3 which has rhombohedral perovskite (R3c) structure [27, 29]. The lattice mismatch between them is expected to trigger band reconstruction near the interface [26]. Band reconstruction can influence many physical properties of the composite including transport and magnetic properties. In one of our earlier reports, it was reported that the composite of BiFeO3 and TbMnO3 (in 7:3 ratio) 0.7BFO–0.3TMO shows interesting properties like spin–phonon coupling, magnetodielectric coupling, high dielectric constant, etc. [30]. It was also seen that charge accumulation and charge transfer at the interface of two materials of the composite in determining different properties of the composite. Thus, it would be significant to study the band structure of the composite and band reconstruction if any due to the charge transfer. It is also interesting to study the changes in transport and magnetic properties due to band reconstruction. In this report, we have studied the band structure through valence band X-ray photoemission spectroscopy (XPS) and ultraviolet photoemission spectroscopy (UPS) and UV–visible absorption spectroscopy. The effect of band reconstruction on the transport and magnetic properties has also been studied.

2 Materials and methods

The compounds BiFeO3, TbMnO3, and the composite of BiFeO3 and TbMnO3 (in 7:3 ratio), i.e., 0.7BFO–0.3TMO, were prepared following a conventional solid-state reaction method by taking the precursors for all the constituent elements i.e., Bi2O3, Fe2O3, Mn2O3, and Tb4O7 (Alfa Aesar, USA) in proper stoichiometric ratio. At first, powders of TbMnO3 were prepared by solid-state reaction of Tb4O7 and Mn2O3 in proper stoichiometric ratio. The mixture was calcined 1200 °C for 12 h. The calcined powders were pelletized and sintered at 1300 and 1400 °C for 24 h. The powder was ground thoroughly in between. Then the TbMnO3 powder was mixed with Bi2O3 and Fe2O3 in stoichiometric ratio, and the mixture was ground with mortar pestle for 4 h. The mixture was calcined at 900 °C for 6 h after which it was pelletized and sintered for 10 h at 1000 °C to get the final sample. The surface morphology and grain growth are studied from field emission scanning electron microscope (FESEM, Nova Nano SEM 450). The composition of the samples was also studied from energy-dispersive X-ray analysis (TEAM EDS SYSTEM with Octane Plus SDD Detector) integrated with the SEM. X-ray photoemission spectroscopy (XPS) experiments were performed using Omicron Multi-probe® Surface Science System, GmBH, equipped with a dual-anode non-monochromatic Mg/Al X-ray source (DAR400), a monochromatic source (XM 1000), and a hemispherical electron energy analyzer (EA 125). All the XPS measurements were performed inside the analysis chamber under base vacuum of ~ 1.8 × 10−10 Torr using monochoromatized AlKα with a power of 300 W. The total energy resolution for monochromatic AlKα line with photon energy 1486.70 eV, estimated from the width of the Fermi edge, was about 0.25 eV. The pass energy for Survey Scan Spectra and core level spectra was kept at 50 eV and 30 eV, respectively. Ultraviolet photoemission spectroscopy (UPS) and UV–visible absorption spectroscopy were employed to study valence band structure and electronic properties. To investigate the fine changes near Fermi level, Fermi edge ultraviolet photoemission spectra were collected using the non-monochromatic He I (21.2 eV) line at an average base pressure of 2.8 × 10−8 Torr. The energy resolution of the analyzer and the step size were set at 0.03 eV and 0.05 eV, respectively. The magnetic measurements were carried out in Magnetic Property Measurement System (SQUID-MPMS, Quantum Design, USA). The temperature-dependent transport properties were measured in Close Cycle He cryostat (Advanced Research Systems, Inc.) using a Keithley 6517B Electrometer.

3 Result and discussion

In order to visualize surface morphology and microstructure of the composite images of sintered 0.7BFO–0.3TMO composite along with the pure BFO and TMO were recorded in the field emission scanning electron microscope (FESEM) (shown in Fig. 1a–c). It can be observed that pure BFO pellet exhibits rectangular and non-homogenously distributed grains with smooth surfaces showing crystal growth of BFO. The average grain size was calculated using ImageJ software with an average size of ~ 2.16 μm. TMO microstructure consisted of non-homogenously distributed grains of an irregular shape with average grain size of 1.19 μm. The composite also shows similar morphology with irregular-shaped grains, but the grain size varies in broad range as compared to that of BFO and TMO grains. The average grain size also lies in between that of the two pure compounds (1.59 μm) [31]. In order to identify all the elements present in 0.7BFO–0.3TMO composite, energy-dispersive X-ray analysis (EDXA) analysis was carried out, which is shown in Fig. 1d. The obtained EDXA result confirmed the homogenous mixing and phase formation of the composite. The EDXA result also indicated the presence of oxygen vacancy in the system. Manganite systems sintered at a higher temperature tend to have oxygen vacancies which in turn can influence the Mn ions present in the system to have multiple valence states in order to maintain the charge neutrality [32].

Fig. 1
figure 1

a–c FESEM images from fractured surfaces of the single-phase BFO, TMO, and composite 0.7BFO–0.3TMO, respectively, and insets show the corresponding calculation of grains size bye image j software. d shows EDXA graph of the composite 0.7BFO–0.3TMO with its chemical composition in the insets

Full size image

Fourier transform infrared (FTIR) spectroscopy has been one of the most used techniques these days to analyze different functional groups and corresponding vibrational bands. The FTIR spectrum of the composite 0.7BFO–0.3TMO was recorded in the wavenumber range of 400–2000 cm−1 which is presented in Fig. 2. The vibrational bands of metal–oxygen bonds generally lie below 1000 cm−1. In the case of pure BiFeO3, the bending and stretching vibrations of the Fe–O in FeO6 octahedra and the vibrational band of BiO6 octahedra lie in the range from 400 to 560 cm−1, which has been clearly observed in the spectrum (peaks denoted as A and B) [33, 34]. The presence of TbMnO3 perovskite phase in the composite is confirmed by identifying the stretching vibration of Mn–O–Mn bond ~ 573 cm−1 arising from MnO6 octahedra of the compound (denoted as C in the spectrum). Apart from these bands few other bands can be identified in the spectrum such as 1720 cm−1 assigned to the bending vibrations of H2O which arises due to the presence of moistures in the powder [33, 35]. The bands ~ 1070 cm−1 and 1440 cm−1 are ascribed to the stretching vibrations of C=O and C–C, respectively, which appears due to the absorption of carbon during the exposure of the sample to air [36].

Fig. 2
figure 2

FTIR spectrum of the composite 0.7BFO–0.3TMO recorded at 300 K

Full size image

X-ray photoemission spectroscopy (XPS) is a prominent technique followed by many researchers to study the chemical composition and valence state of constituent elements of material. We have studied the survey scan and detail scan of Fe2p, Mn2p, O1s, Bi4f, and Tb3d spectra to gain detail knowledge of their electronic structure and composition in the composite BiFeO3–TbMnO3. All the peak positions were matched with the National Institute of Standards and Technology (NIST) XPS database. From the survey, it was confirmed that all the constituent elements (i.e., Bi. Fe, Tb, Mn, and O) were present in the composite (Fig. 3a). The absence of any foreign element in the compound was also confirmed from the spectrum. Although a C 1s peak can be seen from the survey scan spectrum, the presence of carbon does not influence any physical property of the composite. In many systems, such surface absorbed/adsorbed carbon has been found which does not have any effect on the physical properties. The detail spectra were deconvoluted using the XPSPEAK software. In Fig. 3b, the deconvoluted spectrum of the Fe2p core level is shown, which is split into two broad peaks due to spin–orbit coupling. The characteristic peaks of Fe2p3/2 and Fe2p1/2 are observed around the binding energies 710.6 and 724.7 eV, respectively, which correspond to Fe3+ state in BFO [37–39]. In addition to the characteristic peaks, there is a satellite peak, which is observed at the binding energy position 719 eV. The presence of the satellite peak ~ 719 eV confirms that trivalent oxidation state is the dominant state of iron in BFO which is expected for BiFeO3 [32, 33]. The XPS analysis indicates that only pure BiFeO3 phase is present without any type of impurities.

Fig. 3
figure 3

a XPS survey scan of 0.7BFO–0.3TMO at 300 K. Deconvoluted detail spectra of b Fe 2p, c O 1s, d Mn 2p of composite 0.7BFO–0.3TMO. Detail spectra of e Tb 3d and f Bi 4f

Full size image

Figure 3c shows the O1s XPS spectrum of the composite which is split into two parts revealing two kinds of chemical state of oxygen present in the composite. The peaks at binding energy position ~ 529.3 eV correspond to the oxygen situated at lattice denoted as OL and the peak positioned at ~ 531.3 eV corresponds to the surface chemisorbed oxygen denoted by OV. The surface chemisorptions of the oxygen ions occur due to the presence of oxygen vacancy in the system. In BiFeO3-based compounds, it is common to find the oxygen vacancies which are created at the surface due to lattice defects [34, 40]. Moreover, our XPS results conform with EDX analysis which indicated the presence of oxygen vacancy in the system.

In Fig. 3d, the deconvoluted Mn2p spectrum is presented which is also found to be split into two characteristic spin–orbit peaks for Mn2p3/2 and Mn2p1/2 ~ 642.6 and 654.0 eV, respectively. Interestingly, the peaks were asymmetric in nature indicating multiple valence states of Mn ions in the composite. The Mn2p3/2 and peak were deconvoluted into three characteristic peaks at binding energies 640.4, 641.4, and 642.6 eV, which correspond to Mn2+ Mn3+ and Mn4+, respectively. The deconvoluted peak positions for the Mn2+, Mn3+, and Mn4+ species were found at 651.8, 652.8, and 654.0 eV, respectively. From the deconvoluted characteristic peaks, Mn3+ was found to be more intense, indicating dominant presence of Mn3+ which is obvious in case of TbMnO3. It is possible to estimate the fraction from the area under each curve representing a particular ion. The fractional composition is estimated using the relation:

$$\% {\text{fraction}} = \frac{{{\raise0.7ex\hbox{${{\text{Area}}_{j} }$} \!\mathord{\left/ {\vphantom {{{\text{Area}}_{j} } {S_{j} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${S_{j} }$}}}}{{\sum {\raise0.7ex\hbox{${{\text{Area}}_{j} }$} \!\mathord{\left/ {\vphantom {{{\text{Area}}_{j} } {S_{j} }}}\right.\kern-0pt} \!\lower0.7ex\hbox{${S_{j} }$}} }},$$

where Sj is the relative sensitivity factor. The relative fractions have been found to be 22.85%, 50.46%, and 26.68% for the Mn2+, Mn3+, and Mn4+ respectively. The mixed phase of the Mn ions evolves in the system to maintain the charge neutrality which is affected by the oxygen vacancy observed from the O1 s detail spectrum. It has been observed that, in Mn-based systems, charge transfer effect between the Mn3d orbitals and O2p ligand orbitals can give rise to the mixed phase of Mn [41 and references therein]. The creation of Mn4+ and Mn2+ species can also be due to the valence instability of Mn3+ ions. The mix valence state of Mn ions is known to modify the bulk magnetic and electrical properties of different manganite systems [42].

In Fig. 3e, core level XPS spectra of Tb3d element is presented. The spectra show spin-orbit coupling peaks of Tb 3d positioned at ~ 1276.7 eV and ~1241.8 eV for 3d3/2 and 3d5/2, respectively. The peak positions and the doublet separation clearly suggest that Tb is in 3+ oxidation state [NIST]. The core-level spectra of Bi4f are also studied and are shown in Fig. 3f. The spin–orbit split peaks of Bi4f5/2 and Bi4f7/2 are observed to be present at ~ 163.7 eV and ~ 158.3 eV, which corroborate well with earlier reports of trivalent valence states of Bi [43, 44].

The presence of different charge states of TM ions at the interface due to the charge transfer between Mn–Mn ions may lead to the band reconstruction at the interface. As a result of the reconstruction, band gap of the composite is expected to decrease [45, 46]. Thus to confirm the band gap reduction, we have measured the absorption spectrum in the UV and visible range. For studying the absorption characteristics of the composite, absorbance at different wavelengths (k) (range of 200–800 nm) was recorded and the absorption coefficients (a) were calculated at corresponding wavelengths. As can be seen from the spectrum, the absorption band edge lies beyond the range of measurement. We have followed the Tauc’s method to estimate the band gap of the composite from the absorption spectra [26]. The photon energy (hν) and the band gap energy for a particular transition are related by the equation:

$$\left( {\alpha h\nu } \right) = K\left( {h\nu - E_{\text{g}} } \right)^{n} ,$$
(3)

where α is the absorption coefficient given by α = 2.303 (Ab/t) (here, Ab is absorbance and t is the thickness of the cuvette which is 1 cm) and K is the edge width parameter. The value of n depends on the type of transition, i.e., allowed direct, allowed indirect, forbidden direct, and forbidden indirect for which it can have values 1/2, 2, 3/2, and 3, respectively. BiFeO3 is known to have a band gap of 2.1 eV to 2.7 eV in different forms of the material such as bulk, nanomaterials, or single crystalline material [47,44,49]. Since BiFeO3 and TbMnO3 both are known to be direct band gap material, the band gap of the composite was determined from the linear fitting of the straight-line part of the (αhν)2 versus photon energy (hν) plot on the hν axis [50]. From the Tauc’s plot (Fig. 4), it is evident that the band gap of the material lies ~ 01.45 eV, which is low compared to other reported values for BiFeO3 [47,44,49]. Therefore the reduction in the band gap of the composite system confirms the band reconstruction phenomenon due to the charge transfer between the TM ions in the composite.

Fig. 4
figure 4

Tauc plot for the determination of the optical band gap of the composite 0.7BFO–0.3TMO at room temperature. The black dashed arrow is to guide the eye, showing the extracted bad gap. The inset shows the absorption spectrum of the composite from which the Tauc plot has been estimated

Full size image

The XPS valence band (VB) spectra of the 0.7BFO–0.3TMO system has been recorded at room temperature (300 K) to gain better insight into its detailed electronic structure (Fig. 5b). It is evident from the figure that near Fermi level (EF = 0 eV), the spectral weight of the electronic states is extremely weak or merely absent which essentially suggests an insulating ground state of the system. Moreover, it is relevant here to mention that the UV–visible spectroscopy study yielded a high band gap of ~ 1.45 eV and the resistivity data exhibited an insulating nature with high resistance (discussed later). Hence, our VB spectra corroborate well with our earlier results. The energy position of the valence band maximum has been estimated to be ~ 1 eV by making a linear extrapolation of the sharply rising feature immediately below the Fermi level (as shown in inset of Fig. 5a) [45, 51]. Apart from this, knowing the band gap of the system (~ 1.45 eV) and the valence band maximum energy position (~ 1 eV), the position of the conduction band (CB) minimum can be readily determined to be ~ − 0.45 eV. Hence, a schematic diagram of its possible density of states (DOS) (band structure) is depicted in Fig. 5a.

Fig. 5
figure 5

a Schematic diagram of the band structure. Inset shows the XPAS valence spectra of the composite 0.7BFO–0.3TMO. The red line is to guide the eye showing the position of the edge of the valence band. b Valence band XPS spectrum of the composite at 300 K. c UPS spectrum of the composite at room temperature

Full size image

Furthermore, three main features of XPS VB spectra below the EF can be observed at ~ 2.5 eV, 7 eV, and 10.5 eV, which are denoted as VB1, VB2, and VB3 respectively. The complete occupied VB spectra below EF (ranging from 0 eV to ~ 14 eV) is mainly composed of hybridized states of Tb4f, extended Mn3d, Fe3d, and O2p orbitals. Since, for the present system, the Mn and Fe are in octahedral co-ordination with ligand oxygen ions, the crystal field effect causes the Mn/Fe3d states to split into eg and t2g levels. Moreover, for such kind of co-ordination, the t2g states always lie below the eg states. Therefore, the spectral weight (in the range of 0 eV to ~ 1 eV) immediately below EF can be mainly attributed to the contributions from partially occupied Mn3d_eg and Fe3d_eg orbitals [46]. However, the first shoulder like feature VB1 positioned at ~2.5 eV has appeared mainly due to the hybridization of Tb4f, extended Mn3d_t2g, and Fe3d_t2g with O 2p orbitals. However, a significant contribution to this feature VB1 is expected to be from Tb4f states as the other states (Mn3d_t2g, Fe3d_t2g) are extended over a long range [52]. The most intense feature of the VB spectra at ~ 7 eV (VB2) can be attributed mainly to the hybridization of the Mn3d_t2g and Fe3d_t2g orbitals with the O2p states. On the other hand, the feature VB3 at the lowest energy ~10.5 eV emanates from the hybridized states of O2p - Mn3d_t2g/Fe3d_t2g and other oxygen bonding states O2p–Mn/Fe4sp and O2p–Tb5sd, etc., with significant contribution being from O2p states [52].

Moreover, to probe how the VB spectral features get modified with higher resolution and also to augment with our previous results of a electronic structure near Fermi level, we have carried out the ultraviolet photoemission spectroscopy (UPS) measurement on the same system (Fig. 5c). It is interesting to note here that the spectral features in UPS VB spectra immediately below the EF are of relatively low intensity as compared to those observed in XPS VB spectra. This in turn predicts the fact that the region immediately below EF has significant contribution from the Mn/Fe3d states since the photo-ionization cross-section for the Mn/Fe3d photoemission is considerably higher than that for the O2p photoemission in the XPS process. As a consequence, the low-energy UPS is less sensitive to the heavier atoms and highly sensitive to the lighter atoms such as oxygen. Hence, in the whole UPS spectra, the predominant contribution comes from the O2p states, while the contributions from the Tb4f, Mn3d, and Fe3d states get suppressed. Irrespective of the above facts, the UPS VB spectra agreed mostly with the XPS VB spectra. The absence of the electronic states near the Fermi level in the UPS VB spectra supports our previous results, thus suggesting its insulating nature. Moreover, a feature (VB1) near ~ 2.5 eV (associated with the hybridized states of O2p–Mn/Fe3d and O2p–Tb4f orbitals) can be visible in the UPS VB spectra which corroborate with the XPS VB spectra. Similarly, the second feature (VB2) which is the intense broad peak near ~ 7 eV matches well with the most intense peak of the XPS VB spectra. Another weak feature (VB3) near ~ 11.5 eV can also be visible which is seemingly associated with the O2p–Mn/Fe3d hybridized states with significant contribution coming from O2p states. The position of the last feature differs by ~ 1 eV with that of the VB3 feature observed in XPS VB spectra. This can presumably be attributed to the possible charging effect due to the high resistivity of the system.

In our earlier study and in other works, it has been found that the BiFeO3-based materials show high dielectric constant and magnetodielectric coupling [21, 22, 30]. Hence, it is significant to study the transport property of the composite. Figure 6 shows the resistivity vs. temperature plot of the composite measured in the temperature range 10–300 K. The resistivity of the sample increases continuously with decreasing temperature indicating semiconducting nature of the composite. This also indicates the involvement of activation process in the transport mechanism. Below 42 K, the resistivity of the sample increases sharply. The inset (a) of Fig. 6 shows lnρ versus 1000/T plot signifying Arrhenius fitting which shows that our data poorly fit the Arrhenius equation. Hence, the Mott’s variable-range hopping (VRH) mechanism was considered to govern the transport process. The resistivity in case of VRH model can be written as: [53]

$$\rho = \rho_{0} \exp \left[ {\left( {\frac{{T_{\text{M}} }}{T}} \right)^{{\frac{1}{4}}} } \right] ,$$

where TM is known as Mott’s characteristic temperature and is expressed as:

Fig. 6
figure 6

Temperature-dependent resistivity of the composite 0.7BFO–0.3TMO. Inset a lnρ versus 1/T plot showing Arrhenius plot. Inset b lnρ versus (1/T)1/4 showing the plot for variable-range hopping

Full size image
$$T_{\text{M}} = 21.2 \alpha^{3} /k_{\text{B}} N\left( {E_{\text{F}} } \right).$$

The most probable hopping distance (R) and the hopping energy (W) can be presented as:

$$R = \left[ {{\raise0.7ex\hbox{${8\pi \alpha N\left( {E_{\text{F}} } \right)k_{\text{B}} T}$} \!\mathord{\left/ {\vphantom {{8\pi \alpha N\left( {E_{\text{F}} } \right)k_{\text{B}} T} 9}}\right.\kern-0pt} \!\lower0.7ex\hbox{$9$}} } \right]^{{ - \frac{1}{4}}} ,$$
$$W = {\raise0.7ex\hbox{$3$} \!\mathord{\left/ {\vphantom {3 {4\pi R^{3} N\left( {E_{\text{F}} } \right)}}}\right.\kern-0pt} \!\lower0.7ex\hbox{${4\pi R^{3} N\left( {E_{\text{F}} } \right)}$}},$$

where 1/α is the localization length. The average ionic radii of the two transition metal ions (Fe3+ and Mn3+) which are considered responsible for the conduction mechanism are taken as the localization length. Inset (b) of Fig. 6 shows the lnρ versus (1/T)1/4 plot, which displays a good linear fitting thus supports the transport mechanism to be VRH. It is noteworthy to mention here that in disordered systems with random distribution of defects, VRH mechanism can generally be found to control the transport process [54]. In our sample, there are site defects due to the presence of mixed valence of Mn ions. The large strain produced at the interface of the two materials due to lattice mismatch would definitely create lattice defects in the composite system. Moreover, the high resistivity of the sample is consistent with band gap ascribed from the UV–visible spectrum and the band structure calculated from the VB XPS and UPS spectra.

To gain knowledge about the effect of band reconstruction and creation of different valence state of Mn ions on the magnetic property of the composite, we have measured the room-temperature magnetization (M) versus magnetic field (H) hysteresis loop of the composite and compared with that of pure BFO. Enhancing the magnetic property of BiFeO3 is one of the most significant challenges in multiferroic material study as it is the only multiferroic whose ordering temperatures (both magnetic and electric) are above the room temperature. Figure 7 presents the measured M–H loops for the magnetic field \(\pm\) 7 T from which it can be seen that the magnetization increases linearly with the applied field. This unsaturated behavior is consistent with the antiferromagnetic nature of BiFeO3 at room temperature. A closer look around the zero field reveals that at low fields there is deviation from the linearity and the loop is wide open. This along with the loop opening even at high fields signifies ferromagnetic contribution in the magnetic property of the sample. There are reports on the room-temperature weak ferromagnetism in the nanoparticles of BiFeO3 [8]. Also BiFeO3 is known to possess canted antiferromagnetic ordering with a cycloid frequency ~ 62 nm which is governed by the Dzyaloshinskii–Moriya interaction [5, 6, 8]. The weak ferromagnetism can also arise from the canted antiferromagnetic ordering. Inset in Fig. 7 shows the M–H loop for the pure BFO sample. It can be seen that the magnetization of BFO also does not saturate showing its antiferromagnetic nature. Interestingly, it can be noticed that the magnetization value increases by an order of two. The coercivity of the material is found to increase from that of pure BFO. The increase and the low field weak ferromagnetic behavior can be understood to be due to the possible ferromagnetic superexchange interaction between different Mn ions (Mn2+, Mn3+, and Mn4+ as seen from XPS spectra) at the interface. According to Goodenough–Kanamori rules the superexchange interactions between Fe3+–O–Mn2+/Mn4+ are also expected to be ferromagnetic [26, 55, 56]. It is difficult to estimate the contribution of different interactions which dominate the magnetic behavior. However, from the XPS study, it is observed that the relative fractions are found to be 22.85%, 50.46%, and 26.68% for the Mn2+, Mn3+, and Mn4+ respectively. Mn3+ dominates the fractions as expected, whereas Mn4+ and Mn2+ have also significant fractions. From the analysis, it can be concluded that Fe3+ ions are present in the sample in the largest amount among the transition metal Mn3+, Mn4+, and Mn2+ in decreasing order. Clearly, the most dominating interactions would be Fe3+–O–Fe3+. Among the ferromagnetic interactions, Fe3+–O–Mn3+ would be the strongest, and similarly, we can say that the superexchange interaction of Mn2+–O–Mn4+ would be the weakest. Thus, the composite shows higher coercivity in comparison with pure BiFeO3.

Fig. 7
figure 7

Room-temperature M–H loop of the composite 0.7BFO–0.3TMO. Inset shows the room temperature M–H loop of pure BiFeO3

Full size image

4 Conclusion

To summarize, we have prepared the BiFeO3–TbMnO3 composite (7:3) via conventional solid-state reaction and studied the interface through different characterizations. Morphological detail and the grain growth were studied from the SEM images. Chemical state and the composition of the material were studied from XPS and EDXA which revealed that oxygen vacancy is there which in turn creates mix valence state of Mn ions to maintain the charge neutrality. A remarkable decrease in band gap was observed from the UV–visible absorption spectra from that of BiFeO3. Based on the band gap (~ 1.45 eV) and the results from XPS valence band spectra and UPS spectra, the band structure of the material was drawn in which the conduction band edge was found ~ 0.45 eV. In the valence band, three main features were observed at binding energy positions ~2.5 eV, 7 eV, and 10.5 eV (VB1, VB2, VB3, and VB4) which were composed of hybridized states of Tb4f, extended Mn3d, Fe3d, and O2p orbitals. The most intense feature (VB2) was attributed to the hybridization of the Mn3d_t2g and Fe3d_t2g orbitals with the O2p states, while the weak shoulder-like feature VB1 which close to the Fermi energy is attributed to the Tb4f states. Such band diagram and the reduction in the band graph are due to the reconstruction of the bands due to interfacial strain. Moreover, the transport property was found to be dominated by variable-range hopping mechanism, and the high resistivity of the material was also found to be consistent with the band diagram. Antiferromagnetic like non-saturating M–H loop was observed with weak ferromagnetic features at low fields which was attributed to the superexchange interaction between different Mn ions.