1 Introduction

The displacement of magnetization loops along the positive or negative magnetic field axis is generally attributed to exchange bias (EB). The EB was reported way back in 1956 by Meiklejohn and Bean in ferromagnetic fine Co particles coated with antiferromagnetic CoO [1]. Since then, the EB is studied in several oxidized FM particles and in bilayer or multilayer hetero-structure involving ferromagnet (FM), antiferromagnet (AFM), ferrimagnet (FiM), spin glass, etc [2–5]. The recent and renewed interest in the study of EB is due to its importance in several applications such as magnetic recording media, spintronic devices, etc [4]. Moreover, it helps in tuning and improving the coercivity and energy product of permanent media.

Recently, EB is reported in bulk materials such as orthochromites [6–11], spinel oxides [12–15], manganites [16–18], and orthoferrites [19–23]. EB in nanoparticles of orthochromite is explained in terms of core–shell structure [7] and in bulk material; it is attributed to the antiparallel coupling between the weak ferromagnetic component and the moment due to magnetic rare-earth ions or other doped magnetic ions [8, 9]. Interaction between ferrimagnetic components of two magnetic sublattices is proposed to play a role in the observed EB in spinel compounds [12, 15]. Several rare-earth orthoferrites such as LaFeO3, La(Cr,Fe)O3, Nd(Cr,Fe)O3, DyFeO3, and SmFeO3 are reported to exhibit an EB behavior [19–23]. Here, DyFeO3 and single-crystalline SmFeO3 exhibit a room temperature exchange bias field of 100 and 10 Oe, respectively. Thus by further tailoring the orthoferrites, one can achieve a large room temperature EB field.

YFeO3 is a well-known orthoferrite having a distorted perovskite structure with Pnma space group, where each Fe 3+ ion is octahedrally co-ordinated to six nearest oxygen ions [24, 25]. The superexchange interaction in Fe 3+–O 2−–Fe 3+ networks gives rise to an antiferromagnetic ordering with a Néel temperature (T N ) around 650 K [26]. In addition to that, it is reported to show a weak ferromagnetic behavior for T < T N, due to antisymmetric Dzyaloshinskii–Moriya (DM) interaction [27]. Maiti et al. [28] have reported a large EB field of 18 kOe at 10 K in nanocrystalline YFeO3 along with a large coercive field of 23 kOe. However, for application towards magnetic media, a large EB field is required at room temperature and above [29]. In order to tailor the material for EB around room temperature, we have taken up the detailed study of magnetic properties of Mn-substituted YFeO3 compounds. We have found that these materials exhibit an EB behavior at a higher temperature, i.e., above room temperature.

2 Experimental Details

YFe1−x Mn x O3 compounds for x = 0 to 0.3 were prepared by solid-state reaction route. Stoichiometric ratio of Y2O3, Mn(CH3COO)2.4H2O, and Fe2O3 of purity better than 99 % were weighed and grinded using agate mortar and pestle under acetone (99 %) medium. It was presintered at 900 ∘C for 12 h. The presintered powder was pressed into cylindrical pellets by adding polyvinyl alcohol (PVA) as a binder and by using a hydraulic press. The final sintering in pellet form was carried out at 1450 ∘C for 24 h with intermediate grindings. Crystal structure was studied by recording X-ray diffraction (XRD) patterns for Cu- K α radiation using a Rigaku make TTRAX III diffractometer. Magnetization measurements were carried out using a vibrating sample magnetometer (Lakeshore model 7410) in the temperature range 300–700 K.

3 Results and Discussions

Figure 1 shows typical XRD patterns along with the Rietveld refinement for (a) x = 0.05 and (b) x = 0.3 samples. All the samples are found to be in a single-phase form, and they crystallize in orthorhombic unit cell with Pnma space group. The reliability factors of the refinement are R p = 10.6 %, R Bragg = 7.0 %, R f = 7.0 %, and χ 2 = 3.6 for x = 0.3 sample. Typical values of lattice parameters are found to be a = 5.6069(4) Å, b = 7.6020(3) Å, and c = 5.2859(4) Å and a = 5.6443(2) Å, b = 7.5516(3) Å, and c = 5.2753(3) Å for x = 0.05 and x = 0.3 samples, respectively. The increase in Mn concentration gives rise to the increase in lattice constant a and decrease in lattice parameters b and c. Substitution of smaller Mn 3+ (0.580 Å) ions at Fe 3+ (0.645 Å) site is expected to reduce the lattice constants (b, c). Such a compression along the bc plane may lead to expansion along the a direction of the (Fe, Mn)O6 octahedra. The lattice distortion induced by Jahn–Teller active Mn 3+ ions is also expected to play a role in the increase of the a value.

Fig. 1
figure 1

XRD patterns of YFe1−x Mn x O3 compounds for a x = 0.05 and b x = 0.3. Y obs and Y cal represent experimental and Rietveld refined data, respectively

Full size image

Typical plots of temperature variation of magnetization (M– T) under zero-field-cooled (ZFC) and field-cooled (FC) conditions for an applied field of H = 100 Oe are shown in Fig. 2 for x = 0.05 and 0.3 samples. They exhibit an antiferromagnetic transition followed by a considerable increase in magnetization with a decrease in temperature. This behavior is similar to that of ferromagnetic transition but with a small magnetic moment, and it highlights the presence of a weak FM component for T < T N. Such a behavior is reported in the parent compound (YFeO3) due to spin canting [26]. Moreover, the magnitude of magnetization at a particular temperature below T N is found to decrease with an increase in Mn concentration, i.e., the weak ferromagnetic component decreases with an increase in Mn concentration. A large irreversibility is observed between magnetization measured under ZFC and FC conditions, and such behavior is generally observed in magnetic systems having a spin canted moment. The relative irreversible magnetization \(\frac {\Delta M}{M_{\text {ZFC}} }=\frac {M_{\text {FC}} -M_{\text {ZFC}} }{M_{\text {ZFC}} }\) is found to be as high as 25 for x = 0.3 sample and is comparable to other spin canted materials such as NiCr2O4 [12]. T N value was determined from the peak of dM/dT versus T plot, and it decreases from 646 K for x = 0 to 456 K for x = 0.3. The observed decrease in T N value can be attributed to the weaker superexchange interaction in Fe 3+–O 2−–Mn 3+ and Mn 3+–O 2−–Mn 3+ networks. The magnetization of x = 0.3 sample tends to fall rapidly for T < 310 K, and it can be attributed to spin reorientation transition. Such transition from Γ4(A x F y G z ) to Γ1(G x C y A z ) configuration, i.e., the rotation of antiferromagnetic spin axis from crystallographic c-axis to a-axis has been reported in literature [26, 30, 31]. YFeO3 is known to exhibit spin reorientation transition (T SR) at 70 K [26]. So, Mn doping at the Fe site not only decreases the T N value but also simultaneously increases the T SR value.

Fig. 2
figure 2

Temperature variation of magnetization measured under ZFC and FC conditions (H = 100 Oe) for a x = 0.05 and b x = 0.3 samples of YFe1−x Mn x O3

Full size image

Figure 3 shows typical M–H loops recorded under ZFC and FC conditions for x = 0.05 and 0.15 samples at 608 and 573 K, respectively. Here, the cooling field was H FC = 2 kOe. Both samples show a typical weak ferromagnetic behavior with a considerable remanent magnetization and coercive field. At the same time, the observed linear behavior at higher fields without any tendency towards saturation highlights the existing long-range AFM interaction. We can also see from Fig. 3 that M–H loops recorded under FC condition show smaller saturation and remanent magnetization values compared to that of ZFC condition. The coercivity values are also found to decrease upon FC. This can be attributed to the growth and collinear rotation of antiferromagnetic domains or due to the growth of the Γ1(G x C y A z) phase at the expense of the Γ4(A x F y G z ) phase. The Γ1(G x C y A z) magnetic phase is known to exhibit a relatively strong collinear antiferromagnetism compared to that of the Γ4(A x F y G z ) phase [32].The coexistence of Γ1 and Γ4 phases has been reported in literature from Mössbauer spectroscopy studies [33].

Fig. 3
figure 3

M–H loops recorded under ZFC and FC (H FC = 2 kOe) conditions for a x = 0.05 and b x = 0.15 samples of YFe1−x Mn x O3

Full size image

In order to study the exchange bias behavior, we have recorded M–H loops at different temperatures after field cooling through T N by applying a magnetic field of H FC = 2 kOe. Typical plots of M–H loops recorded at different temperatures for x = 0.25 and 0.3 samples are shown in Fig. 4a,b in an expanded scale. These loops are found to shift along the negative field axis, and it highlights the presence of a negative exchange bias field, H EB. The exchange bias field, H EB, and the coercive field, H C, are determined using the relations H EB = (H C1 + H C2)/2 and H C = (H C1 − H C2)/2 where \(H_{C_{\mathrm {1}}}\) and \(H_{C_{\mathrm {2}}}\) correspond to coercive field at ascending and descending branches of the M–H loop, respectively [14, 18].

Fig. 4
figure 4

M–H loops measured at different temperatures for a x = 0.25 and b x = 0.3 samples of YFe1−x Mn x O3 after field (H FC = 2 kOe) cooling through T N. c Temperature variation of H EB for x = 0.3 sample under positive field cooling (circles) and negative field cooling (triangles) conditions. d d 2 M/d H 2 versus H plots for x = 0.25 and 0.30 samples at 453 and 413 K, respectively

Full size image

In order to further confirm the EB and rule out the possibility of any experimental artifact, M–H loops at respective temperatures were recorded after reversing the cooling field, i.e., H FC = −2 kOe. In this case, we have observed the shifting of loops towards the positive field axis. Typical plots of temperature variation of H EB for H FC = + 2 and −2 kOe are shown in Fig. 4c for the x = 0.3 sample. We can see an exponential rise of negative H EB value with a decrease in temperature for H FC = + 2 kOe (positive cooling field). Similar behavior with comparable magnitude is observed for H FC = −2 kOe but with an opposite sign (positive) of H EB. This clearly demonstrates the intrinsic exchange bias behavior in the present system.

Further, in order to rule out the possibility of a minor loop effect on the measured H EB value, we have carried out the analysis of M − H loops in terms of singular point detection method as demonstrated by Harres et al. [34]. According to them, a major or saturated M–H loop exhibits a singularity at |H| = H A, the anisotropy field. In other words, dM/dH or its higher order derivatives for ascending and descending branches of saturated M–H curves overlap for |H| ≥ H A. Typical plots of d 2 M/d H 2 versus H for ascending and descending curves are shown in Fig. 4d for x = 0.25 and 0.30 samples. We can see that such plots for ascending and descending branches overlap for |H| ≥ 5 kOe, and it confirms that measured M–H loops are major in nature.

The maximum value of H EB is found to be in the order of a few kilo-oersted and is comparable to that obtained in the nanoparticles of LaFeO3 at 5 K [19] and Zn-doped LaFeO3 nanoparticles [35]. Typical plots of temperature variation of H EB for x = 0.15 and 0.25 samples are shown in Fig. 5. It is found that H EB exhibits an exponential rise in magnitude with a decrease in temperature below T N and the data were fitted to the relation H EB(T) = H EB(0)exp(−T/T 1), where T 1 and H EB(0) are constant. The fitted data are shown as solid line and they closely follow the experimental data. Same data in a semi-logarithmic scale are shown in the insets of Fig. 5, where they exhibit a linear behavior. So the present set of samples has a potential to yield large room temperature EB for various applications. The maximum H C value determined from the above measurements is 3.5 kOe at 413 K, and they mostly increase linearly with a decrease in temperature.

Fig. 5
figure 5

Temperature variation of H EB for a x = 0.15 and b x = 0.25 samples. Insets show the same data in the semi-log scale. The fitted data are shown as solid line

Full size image

Generally, EB is explained in terms of exchange interaction at the interface of two magnetic phases such as FM/AFM and FiM/AFM. EB in rare-earth orthoferrites such as nanoparticles of LaFeO3 [19], Zn-doped LaFeO3 [35], and DyFeO3 [22] is attributed to conventional core–shell model. As we are dealing with bulk materials, the core–shell model approach is not likely to play a role in the present set of samples. The contribution of magnetic impurity phases such as Y3Fe5O12, YMnO3, and Fe2O3 can be ruled out because the samples are found to be in a single-phase form as per XRD analysis. Moreover, we have not observed any trace of magnetic transition corresponding to above impurity phases. EB in spinel compounds (AB2O4) is explained by considering the anisotropic exchange interaction between the longitudinal ferrimagnetic and the transverse antiferromagnetic components due to spin canted ferrimagnetic interaction across the magnetic sublattices in tetrahedral and octahedral environments [12, 15]. Here, no such long-range ferrimagnetic interaction is known to exhibit. It is known that structurally and chemically single-phase YFeO3 undergoes spin reorientation from Γ4(A x F y G z) to Γ1(G x C y A z) phase at around 70 K as the temperature is decreased. We have also shown that the spin reorientation transition temperature increases upon Mn substitution, i.e., close to room temperature. Here, the Γ4 phase contains a considerable weak ferromagnetic component while the Γ1 phase promotes the collinear antiferromagnetism. For T < T N, i.e., much above room temperature, the presence of Γ1 phase under the matrix of Γ4 phase cannot be ruled out [33]. There is also a possibility that the concentration of Γ1 phase increases under the FC condition. The anisotropic exchange interaction at the interface of the Γ1and Γ4 phases is expected to contribute towards the observed EB field. The above argument can be further supported from the observed reduction in the magnitude of saturation and remanent magnetization upon field cooling as shown in Fig. 3. Such combination of Γ4(A x F y G z) and Γ1(G x C y A z) magnetic phases upon field cooling is reported in Y0.9Pr0.1CrO3 system [32]. So, the observed large EB field above room temperature in YFe1−x Mn x O3 compounds suggests that these materials can serve as one of the group of potential candidates for application towards magnetic recording, etc.

4 Conclusions

Single-phase samples of YFe1−x Mn x O3 for x = 0 to 0.3 were prepared by solid-state route. Temperature variation of magnetization measurements show antiferromagnetic transition with weak ferromagnetic component due to spin canting along with a large irreversibility between the zero-field-cooled and field-cooled magnetizations. The Néel temperature is found to decrease from 646 K for x = 0 to 456 K for x = 0.3 due to the dilution of Fe 3+–O 2−–Fe 3+ networks by the random substitution of Mn 3+ ions. These samples exhibit an exchange bias behavior with a maximum exchange bias field of 1.2 kOe around 413 K. The observed EB is explained by considering the anisotropic exchange interaction at the interface of two magnetic phases with one of them having a dominant collinear antiferromagnetic behavior while the other one having a considerable weak ferromagnetic component.