Structural and Magnetic Properties Study of Fe 2 O 3 /NiO/Ni 2 FeO 4 Nanocomposites

: In the current work, the nanocomposites that consist chieﬂy of three components— α -Fe 2 O 3 , NiO and Ni 2 FeO 4 , in two different ratios 2:2:1 (FNN-221) and 2:1:1 (FNN-211), respectively—were produced. The synthesis was done in two steps by following the chemical co-precipitation and mechanical ball-milling route. The presence of individual phase was identiﬁed from the XRD data without the detection of any additional impurities. The phase fraction of each component estimated from the proﬁle ﬁtting of XRD patterns were found to be 41.2%, 39.7%, 19.1% in FNN-221 sample and 49.5%, 26.4%, 24.1% for FNN-211 sample, respectively, which were consistent with the experimental values. The total magnetization at 300 K was obtained to be 13.41 emu/g and 10.95 emu/g for FNN-221 and FNN-211 samples, respectively. In FNN-211 compound the zero ﬁeld coercivity ( H C ) expanded towards the higher ﬁeld values thereby signifying the exchange bias behavior. Furthermore, the exchange bias ﬁeld ( H ex ) for FNN-211 was obtained as 35.1 Oe. Z.A.; Z.A.; visualization, C.A.; writing—original draft, Z.A., C.A., and F.A.; writing—review and editing, C.A., F.A.


Introduction
Recently, the multi-component magnetic nanoparticles captivated an enormous deal of research and technological interests among scientific communities. An important advantage of such materials over single phased magnetic nanostructures is that the magnetic properties of the former are deeply correlated to the interfacial interactions and contact between the components. Thus, their physical properties can be drastically tailored/induced simply by adjusting the chemical compositions, geometrical shapes and ratios of different components [1][2][3][4][5]. Because of these characteristics, the magnetic nanocomposites can be directly exploited in several new technological fields like spintronic devices, hyperthermia treatment, permanent magnets, photocatalysts and high-density magnetic storage media [6][7][8][9]. Nanocomposites involving the ferrites (oxides of iron) are the most extensively investigated materials in this context. Lately, the methods for the production of ferrite nanocomposites have been centered around the fabrication of metal doped ferrites into a non-magnetic and/or matrix with weak magnetic polarity (SiO 2 , ZnO, titanates) [2,[10][11][12]. The contrast in the magnetic polarity between the host matrix and the foreign particle facilitates a variation in the degree interactions among the charge carriers. Thereby the magnetic transition and anisotropic nature can be affected. The ferrites of Cobalt (Co) and Nickel (Ni) have shown synergistic improvement in their chemical and thermal stability upon dispersion in a porous silica matrix [13][14][15]. Moreover, the low temperatures anomalies in the magnetic hysteresis and modification in the internal coercive field in the case of binary ferrites-ZnO nanocomposites, have also been mentioned in the literature [10,16,17]. Not long ago Galizia et al. investigated a "wasp waist" hysteresis loop in spinel ferrites-composites. It has been noticed that the wasp-waist shape of hysteresis in the spinel ferrites composites is sensitive toward the phase fraction/ratios and particle size of the forming components [18]. Similarly, an exchange spring like physical phenomenon in the ferrites-graphene oxide nanocomposites has been unfolded by El-Khawas et al. [19]. The principal origin of exchange bias in such systems is the staggered surface spins at the interface between the two components [20]. In addition, the influence of particle size and synthesis routes on the structural and magnetic properties in ferrites has been frequently studied by numerous researchers [21][22][23]. Nevertheless, the mechanism of interaction between dual and triple magnetic phases in the ferrites, especially at the low temperatures is still not well understood from the available literature.
Thus, in the present work, we have fabricated triple phase nanocomposites chiefly including three components: spinel Ni-ferrite (Ni 2 FeO 4 ); hematite (α-Fe 2 O 3 ) and nickel oxide (NiO). In order to investigate the influence of each phase on the structural and magnetic properties, the different ratios of each component have been selected and synthesized using chemical Co-precipitation route. In general, both NiO and Ni 2 FeO 4 are antiferromagnetic below their Neel temperatures (T N ) [24]. On the other hand, depending on the synthesis conditions and particle geometries, α-Fe 2 O 3 can exhibit anti-ferromagnetism, weak ferromagnetism and even super-paramagnetism [2,6,25]. The magnetic response of the nanocomposites had been compared with the parent α-Fe 2 O 3 component in order to highlight the involvement of particle interactions. The results obtained here, signifies that due to the involvement of spin functionalities the total magnetization of the nanocomposites has been enhanced by several folds compared to the of parent α-Fe 2 O 3 compound. Meanwhile, the current work is also important in the sense that the interactions among multiple magnetic polarities have been investigated under the influence of several magnetic fields and temperature conditions in the single study. Based on these measurements the appearance of exchange bias effect in the case of composite structures has been identified. Such type of study is important to the literature in order to understand the magnetic behavior in three component systems.

Experimental Details
Initially, the nanoparticles of all the three components namely: Ni 2 FeO 4 , α-Fe 2 O 3 and NiO were individually synthesized through the chemical co-precipitation route. The highly pure (≥99.5%; Sigma Aldrich, MI, USA) analytical grade regents of FeCl 3 .6H 2 O, Ni(NO 3 ) 2 .6H 2 O and NiCl 2 .6H 2 O were used as starting precursors during the synthesis procedure. NaOH and aqueous NH 4 OH solutions were employed as precipitator for the synthesis of Ni 2 FeO 4 /NiO and α-Fe 2 O 3 , respectively. The molar amounts of each salt have been taken as such displayed in Table 1. First, the solutions of individual precursors were prepared separately by dissolving the respective molar amounts of salts in the mentioned amount of deionized (D.I) water. The solutions were subjected to continuous stirring (500 rpm) for 40 min at 60 • C to ensure the homogenous mixing of each salts in the D.I water. Once the salts were dissolved into the D.I water the reaction kinetics and the particle size were controlled through the dropwise (1 mL/min) addition of precipitator into the solutions. The reaction has been stopped upon the formation of precipitates and the reaction mixture was stirred for 1 h under ambient temperature and pressure. The obtained slurry was then washed several times with acetone and the dried in the oven at 70 • C overnight. In the second step the nanocomposites with two different ratios were prepared and labelled as FNN-221 for 2:2:1 ratio and FNN-211 for 2:1:1 ratio of α-Fe 2 O 3 , NiO and Ni 2 FeO 4 powders, respectively (see Table 1). The powders of as synthesized α-Fe 2 O 3 , NiO and Ni 2 FeO 4 compounds were weighted in the corresponding amounts and added in ball-milling vials with balls (diameter~7 mm) to powder ratio of 1:20 for each sample. A small amount of ethanol was added into the vials as a buffer media along with a few drops of polyvinyl-alcohol (PVA; 2 wt.%). The samples were ball-milled for 6 h with a milling speed of 180 rpm. The obtained mixtures were dried into the oven at 70 • C. The obtained powders were heat treated at 300 • C to dissolve PVA and fuse the particles. The crystalline structure, phase fraction and crystallite sizes of the samples were characterized through the X-ray diffraction (XRD) patterns. The XRD patterns were obtained from Bruker D8 advance diffractometer (Cu-Kα radiation). For Rietveld refinement of XRD profiles, the pseudo-Voigt function has been used. The background has been corrected through the linear interpolation between a set of background points. Scanning electron microscope (FE-SEM; Merlin compact) was employed for analyzing the particle shapes and sizes. Elemental mapping was done with Energy Dispersive X-Ray Analysis (EDX; Oxford). The magnetic measurements were carried out using vibrating sample magnetometer (VSM) equipped in physical properties measurement system (PPMS EverCool-II, Quantum Design, San Diego, CA, USA). , NiO (space group; Fm-3m) and Ni 2 FeO 4 (space group; Fd-3m), respectively. The absence of any additional peak apart from the desirable three phases suggest that the all three phases were unreacted during the heat treatment, rather expresses their individual crystalline nature. The phase fractions of each component in the XRD profiles were calculated by performing the Rietveld analysis in HighScore Plus (Malvern Panalytical) software and the derived parameters are displayed in Table 2. For FNN-221 sample, the phase fractions of α-Fe 2 O 3 , NiO and Ni 2 FeO 4 were found to be 41.2%, 39.7% and 19.1%, respectively. Meanwhile, the three phases contribute 49.5%, 26.4% and 24.1% for FNN-211 samples, respectively. It should be noted that the phase fraction of each compound in both the samples are in perfect agreement with the powder ratios weighted during the synthesis of these composites. Additionally, we observed that the lattice parameters were unaffected towards the change in the ratio of either compounds (shown in Table 2), which is obvious for composites structures. We calculated the crystallite size of the components by employing the Scherrer's equation and Caglioti function in the HighScore plus. However, for this task the reflections of individual phases were difficult to resolve from the XRD patterns of both samples. Thus, the principal XRD peaks at 2θ = 33.20 • , 43.3 • , 35.67 • corresponding to α-Fe 2 O 3 , NiO and Ni 2 FeO 4 compounds, respectively, were accounted. The obtained crystallite sizes are supplied in Table 2. From both methods the crystallite sizes were identified as~58 nm,~285 nm and~49 nm for α-Fe 2 O 3 , NiO and Ni 2 FeO 4 components, respectively. Table 2. The phase fraction (in molar) of components and concentration of each element present in the sample (obtained from XRD and EDX analysis, respectively); comparison of different figure of merits acquired from the magnetization measurements. For each component the thermal parameters were kept constant while performing the calculations.

FNN-221 FNN-211
Unit cell parameters (Å)   Although, the possibility that these particles might have agglomerated during the formation of composites cannot be neglected. Therefore, for further insights into the particle shapes and arrangements we analyzed the SEM images. The SEM micrographs of  Although, the possibility that these particles might have agglomerated during the formation of composites cannot be neglected. Therefore, for further insights into the particle shapes and arrangements we analyzed the SEM images. The SEM micrographs of  Although, the possibility that these particles might have agglomerated during the formation of composites cannot be neglected. Therefore, for further insights into the particle shapes and arrangements we analyzed the SEM images. The SEM micrographs of  Although, the possibility that these particles might have agglomerated during the formation of composites cannot be neglected. Therefore, for further insights into the particle shapes and arrangements we analyzed the SEM images. The SEM micrographs of both the composite samples are displayed in Figure 2a,b. For both samples the particles assemble in a granular structure with irregular morphology. The granular structure is due to the post synthesis heat treatment of the samples which causes the agglomeration among the particles. These attributions are in excellent consent with the nanocomposites synthesized through chemical route by various authors [26][27][28]. Another finding from the SEM images is the irregular shapes of grains, the irregularity can be described by considering the variation in particles sizes of all the three components. Because of the variable crystallite sizes (see Table 2) of individual phase the degree of agglomeration, also differ for each component, thereby exhibiting the irregularity in granular shape. Additionally, the elemental analysis could be of importance for establishing relation between ratios of each component present in the samples, for this purpose we conducted the EDX analysis. The concentration (wt.%) of each element is presented in Table 2   both the composite samples are displayed in Figure 2a,b. For both samples the particles assemble in a granular structure with irregular morphology. The granular structure is due to the post synthesis heat treatment of the samples which causes the agglomeration among the particles. These attributions are in excellent consent with the nanocomposites synthesized through chemical route by various authors [26][27][28]. Another finding from the SEM images is the irregular shapes of grains, the irregularity can be described by considering the variation in particles sizes of all the three components. Because of the variable crystallite sizes (see Table 2) of individual phase the degree of agglomeration, also differ for each component, thereby exhibiting the irregularity in granular shape. Additionally, the elemental analysis could be of importance for establishing relation between ratios of each component present in the samples, for this purpose we conducted the EDX analysis. The concentration (wt.%) of each element is presented in Table 2   The phase composition of the synthesized nanocomposites has also been verified from the Raman spectra and the deconvoluted Raman peaks of both the samples are presented in Figure 3a,b. As expected, the phonon modes corresponding to the three components α-Fe2O3, NiO and Ni2FeO4 are present in both the samples. The strong Raman signals located at 405 cm −1 , 494 cm −1 , 612 cm −1 indicates the existence of α-Fe2O3 phase in both the samples [29]. Similarly, the phonons situated at 571, ~920 cm −1 and 1034 cm −1 allied to the NiO phase [30]. Finally, the Ni2FeO4 phase has been identified from 675 cm −1 , 457 cm −1 , The phase composition of the synthesized nanocomposites has also been verified from the Raman spectra and the deconvoluted Raman peaks of both the samples are presented in Figure 3a [29]. Similarly, the phonons situated at 571,~920 cm −1 and 1034 cm −1 allied to the NiO phase [30]. Finally, the Ni 2 FeO 4 phase has been identified from 675 cm −1 , 457 cm −1 ,~567 cm −1 Raman peaks of these samples [31]. Additionally, it is fascinating to note from the Raman peaks is that in the case of FNN-221 sample the Raman signals representing NiO constitute slightly high signal density compared to the NiO phonons of FNN-211 sample. The relatively higher concentration of NiO Raman signals in FNN-221 samples are in compliance with the experiment and XRD analysis. Therefore, we rigorously analyzed the magnetic properties in the following text. In order to identify the interaction between particles we performed magnetic measurements under different sets of temperature and magnetic field conditions. First, the variation in magnetization with respect to the temperature (M-T) was analyzed in field cooled (FC) and zero field cool (ZFC) condition. The FC-ZFC curves from 50 K to 400 K under an applied field of 500 Oe for both the nanocomposites are presented in Figure 4a,b.
Crystals 2021, 11, x FOR PEER REVIEW 6 of 13 ~567 cm −1 Raman peaks of these samples [31]. Additionally, it is fascinating to note from the Raman peaks is that in the case of FNN-221 sample the Raman signals representing NiO constitute slightly high signal density compared to the NiO phonons of FNN-211 sample. The relatively higher concentration of NiO Raman signals in FNN-221 samples are in compliance with the experiment and XRD analysis. Therefore, we rigorously analyzed the magnetic properties in the following text. In order to identify the interaction between particles we performed magnetic measurements under different sets of temperature and magnetic field conditions. First, the variation in magnetization with respect to the temperature (M-T) was analyzed in field cooled (FC) and zero field cool (ZFC) condition. The FC-ZFC curves from 50 K to 400 K under an applied field of 500 Oe for both the nanocomposites are presented in Figure 4a,b.  It has been observed that the ZFC magnetization attain a maximum value at 229 K, 240 K and 269 K for FNN-221 and FNN-211 and α-Fe 2 O 3 compounds, respectively. At the downturn of the onset of this blocking temperature (T B ), the ZFC magnetization abruptly declined with decreasing the temperature values and separates from the FC magnetization. This distinction amplified at the low temperatures thereby indicating the thermomagnetic irreversibility between FC-ZFC magnetization. In general, the irreversible ZFC-FC magnetization signifies some sort of anisotropy and/or spin glass state in a magnetic material [32]. In the present work, the appearance of glassy states is obvious because of the co-existence of multiple magnetic components having contrasting magnetic polarities. Further, we see that at 50 K the height of magnetic moment corresponding to Fe 2 O 3 particles is far lower than that of composite structures. This was expected because the α-Fe 2 O 3 particles are weakly ferromagnetic only above 270 K. Below this temperature the antiferromagnetic coupling becomes dominant [6,25]. Additionally, both FC and ZFC magnetization of parent Fe 2 O 3 particles dramatically rises with increasing the temperature and reaches a maximum at 270 K. Thereby, signifying some sort of magnetic coupling below this temperature, we connect this behavior with the Morin transition [25,33]. Moreover, such transition has been completely vanished upon the fabrication of composite structures. In the literature it has been suggested that nature of Morin transition depends on the particle size and/or interparticle interactions. For particles smaller than 50 nm the Morin transition shifts towards the lower values and vanishes for particles smaller than 20 nm [34]. In the present case, we note that the particle sizes of each components hardly vary upon the formation of composite structures (see Table 2) thereby negating any involvement of particle diameters on the Morin transition. Hence, the absence of Morin transition in both the composite samples has been assigned primarily to the interactions between the particles due to their variable spin functionalities. These attributions can be supported by the fact that the temperature dependency of magnetization for the composite structures is significantly higher than parent Fe 2 O 3 compound. Now, our interest is to collect more details on the inter-particle interactions among all the three components, therefore we carried out further magnetic measurements by applying external stimulus i.e., in the presence of applied magnetic field (M-H) at various temperatures (50 K, 100 K, 200 K, 300 K). In Figure 5a,b we compared the M-H magnetization curves of FNN-221, FNN-211 composites with the parent Fe 2 O 3 sample, under a maximum applied field of 3T. In both samples regardless of the isotherms the magnetization abruptly rises with increasing the magnetic field strength, approached a maximum and the magnetization did not saturate up to 3T. It indicates that the magnetic states co-exist in all samples and as a function of applied field strength the effective magnetic moment of each components reverses their polarity towards the applied field thereby inhibiting the total magnetization to reach at the saturation. Moreover, the total magnetization (at 3T) decreases with increasing the temperature (see Figure 5(a1,b1)) which can be expected for a polycrystalline compound. The rise in the temperature promotes thermal disorders among the ordered spins and intensifies the spin degree of motion, thereby causing the reduction in the total magnetization. The effect of temperature on the magnetization and magnetic hysteresis loop can be clearly seen from Figure 5(a2,b2). The parameters obtained from the hysteresis loops of the samples are presented in Table 2. It can be seen that for all the samples the total magnetization (M t ), coercive field (H C ) and remanent magnetization (M r ) collectively increases with decreasing the measurement temperature. Here, the increase in coercive field can be described such that at far low temperatures the anisotropy energy increases and more significant than the magnetostatic parameter thereby producing large values of the coercivity [2,35]. Similarly, form Table 2 it should be noted that at all temperatures the M t values for FNN-221 sample is larger than the FNN-211 sample, the high magnetization value of FNN-221 sample has been ascribed to increase in interparticle interaction due to incorporation of higher number of NiO particles in this sample. On the other hand, the coercivity of FNN-211 sample slightly higher than the FNN-221 sample which has been attributed to competing coupling interactions between the three components. To support this explanation, we have presented the magnetization curve of parent α-Fe 2 O 3 compound separately in Figure 5c. It can be seen that the maximum value of isothermal magnetization for this sample is roughly 10-fold lower and the coercive field is barely significant compared to the FNN-221 and FNN-211 composites. The low magnetization value has been allied to the weak ferromagnetic exchange coupling in α-Fe 2 O 3 compound. So, in FNN-211 sample because of the lower number of total antiferromagnetic domains (NiO:Ni 2 FeO 4 = 1:1) than the FNN-221 (NiO:Ni 2 FeO 4 = 2:1), the weakly ferromagnetic α-Fe 2 O 3 particles compete with the antiferromagnetic domains. Thus, the ferromagnetic coercivity for this sample is larger than the FNN-221 samples. The schematic diagram representing the possible interactions among these particles has been given in Figure 6. In conclusion, it is deemed important to consider that the local competition between the magnetic states exists because of their contrasting ratios in both of the samples. Therefore, to gain more insight on the exchange coupling at the interface of these magnetic states we measured the MH magnetization with ZFC-FC conditions.  The field dependent ZFC-FC magnetization curves of both the samples are displayed in Figure 7a and the local magnification of these curves covering the coercive field and total magnetization are displayed separately in Figure 7b,c, respectively. We observed that compared to the FNN-221 nanocomposite the hysteresis loop of FNN-211 sample slightly shift towards the high field regions. Indeed, the total ZFC magnetization for FNN-211 sample is slightly lower than the FC magnetization, similarly, the coercivity expanded towards high field values for the same sample. Meanwhile, in distinction between ZFC and FC magnetization in case of the FNN-221 sample is relatively insignificant. The loop shift phenomenon in nanocomposites has been noticed by several authors and attributed to the interfacial exchange interactions (exchange bias) between the two or more magnetic states having contrasting polarities [36][37][38]. The relative shift in the magnetic hysteresis due to the exchange bias can be estimated as /2; where is the exchange bias field. The value for FNN-211 compound is found to be 35.1 Oe, this value however is relatively smaller than those reported in the literature for α-Fe2O3 and α-Fe2O3: NiO composites [39,40]. The weak exchange bias in our sample is probably due to the existence of an additional magnetic state. Even so, the exchange bias in the second sample (FNN-221) is negligible, which suggests that the exchange bias effect is mainly due to the competing interactions of weak ferromagnetic and antiferromagnetic interfaces. In FNN-221 the antiferromagnetic content is relatively higher than the FNN- The field dependent ZFC-FC magnetization curves of both the samples are displayed in Figure 7a and the local magnification of these curves covering the coercive field and total magnetization are displayed separately in Figure 7b,c, respectively. We observed that compared to the FNN-221 nanocomposite the hysteresis loop of FNN-211 sample slightly shift towards the high field regions. Indeed, the total ZFC magnetization for FNN-211 sample is slightly lower than the FC magnetization, similarly, the coercivity expanded towards high field values for the same sample. Meanwhile, in distinction between ZFC and FC magnetization in case of the FNN-221 sample is relatively insignificant. The loop shift phenomenon in nanocomposites has been noticed by several authors and attributed to the interfacial exchange interactions (exchange bias) between the two or more magnetic states having contrasting polarities [36][37][38]. The relative shift in the magnetic hysteresis due to the exchange bias can be estimated as H ex = − H right C + H le f t C /2; where H ex is the exchange bias field. The H EX value for FNN-211 compound is found to be 35.1 Oe, this value however is relatively smaller than those reported in the literature for α-Fe 2 O 3 and α-Fe 2 O 3 : NiO composites [39,40]. The weak exchange bias in our sample is probably due to the existence of an additional magnetic state. Even so, the exchange bias in the second sample (FNN-221) is negligible, which suggests that the exchange bias effect is mainly due to the competing interactions of weak ferromagnetic and antiferromagnetic interfaces. In FNN-221 the antiferromagnetic content is relatively higher than the FNN-211 sample thereby the competition with the weakly ferromagnetic states has been prevailed. This attribution satisfies our previous assumptions and it could be the main reason for comparatively small values of zero field exchange bias in the present work.

Conclusions
In summary, we have synthesized ferrite nanocomposites and confirmed the phase fractions of each three α-Fe 2 O 3 , NiO and Ni 2 FeO 4 components from XRD and SEM analysis. The temperature dependent and field dependent magnetization of FNN-221 nanocomposite are relatively higher than the FNN-211 sample. The high values of magnetization are because of the increase in interparticle interaction due to incorporation of higher number of NiO particles in this sample. Further in FNN-211 compound the zero field coercivity expands towards the higher field region (hysteresis shift) which points to the existence of exchange bias. The hysteresis loop shift in FNN-221 was completely negligible, which relates that the weak exchange bias effect in FNN-211 sample is certainly due to the local competition between the weak ferromagnetic and antiferromagnetic interfaces.