Improved magnetic properties of self-composite SrFe₁₂O₁₉ powder prepared by Fe₃O₄ nanoparticles

2.1 Synthesis of Fe₃O₄ magnetic nanoparticles

Fe₃O₄ magnetic nanoparticles (NPs) with a diameter of 12 nm were synthesized by thermal decomposition in organic solutions (Wu et al., 2014). Fe (III) acetylacetonate (3 mmol, 99%, Alfa), sodium oleate (2 mmol, 90%, Sigma), oleic acid (15 mmol, 90%, Aldrich) and 20 mL benzyl ether (99%, TongGuang) were mixed and magnetically stirred under a gentle flow of Ar. Then the mixture was heated to 120 °C for 30 min. Under Ar blanket, the solution was heated up to 295 °C at a heating rate of 5 °C/min and kept at this temperature for 2 h. The solution was then cooled down to room temperature after removing the heat source. After that, 20 mL of hexane (97%, TongGuang) and 20 mL of ethanol (99.7%, TongGuang) were added to the mixture to precipitate the product. The Fe₃O₄ NPs were collected by the magnetic separation method, and the final product was dispersed in hexane for further use.

2.2 Synthesis of SrM powders

The synthesized Fe₃O₄ NPs and reagent grade powder SrCO₃ (99.9%, Innochem) were mixed with different Fe/Sr atomic ratios of 12:1, 11:1, and 10:1. Then, the 12:1 mixture was sintered at different temperatures of 500–1100 °C at intervals of 100 °C and 1250 °C with a heating rate of 5  °C/min, while the others were sintered at 1100 °C. As a comparison, purchased reagent grade Fe₃O₄ (99%, Aladdin) and SrCO₃ were mixed with a Fe/Sr atomic ratio of 12:1 and sintered under the same conditions. Finally, the structure and magnetic properties of sintered SrM powders were characterized. To distinguish the two kinds of iron sources and the different Fe/Sr atomic ratios, the four SrM specimens were labeled as NP12, NP11, NP10, and RG12.

2.3 Characterization and measurements

The phase composition of sintered specimens at different temperatures were identified by an X-ray diffractometer (XRD, Rigaku D/Max-2400) with Cu Kα radiation. The DSC and TG curves of precursor powders were measured in the N₂ atmosphere at a heating rate of 5 °C/min by thermogravimetric and differential scanning calorimeter (TG/DSC, Netzsch STA-449). The morphology of Fe₃O₄ NPs and SrM powders were detected by a transmission electron microscope (TEM, Hitachi H-7650) and a scanning electron microscope (SEM, Phenom Pro), respectively. The particle size distribution of SrM powders were examined by a laser particle size analyzer (LDPSA, Helos-Oasis). The room-temperature magnetic hysteresis loops of specimens were measured on a vibrating sample magnetometer (SQUID-VSM, Quantum Design) with the maximum external field of 7 T.

3.1 Formation mechanism of SrM

The XRD patterns of Fe₃O₄ NPs synthesized by the thermal decomposition method indicated that all the peaks had a good agreement with the standard PDF card NO.19-0629. This confirmed the existence of a single cubic phase structure, as shown in Fig. 1(a). The inset of Fig. 1(a) shows the TEM image with the histogram of the size distribution of Fe₃O₄ NPs with a relatively uniform size of 12 nm, stable morphology, and good mono-dispersity. And the Since Fe/Sr atom ratio affects the magnetic properties of SrM, specimens with different Fe/Sr ratios of 10, 11, and 12 were sintered from Fe₃O₄ NPs at 1100 °C for comparison. Fig. 1(b) shows the XRD patterns of NP12, NP11, and NP10. Only NP10 had no α-Fe₂O₃ at 1100 °C, indicating that the iron-deficient formulation was more beneficial to the formation of the pure phase. This is consistent with the conclusions of previous studies (Fu and Lin, 2005; Jean et al., 2010; Masoudpanah and Ebrahimi, 2011).

Close

Fig. 1

XRD patterns of (a) Fe₃O₄ NPs with the inset of their TEM image and particle size distribution, (b) SrM powders prepared by Fe₃O₄ NPs with different Fe/Sr atom ratio of 10, 11 and 12 sintered at 1100 °C, (c) reagent grade Fe₃O₄ and SrCO₃ sintered at 500 to 1100 °C, (d) Fe₃O₄ NPs and SrCO₃ sintered at 500 to 1250 °C.

Export to PPT

Fig. 1(c) and (d) show the phase evolution of the mixtures of SrCO₃ with reagent grade Fe₃O₄ and Fe₃O₄ NPs respectively with a Fe/Sr atom ratio of 12. The formation mechanism of SrM generated by the two kinds of Fe sources through the solid-state method was well reflected with an increase in sintering temperature. More specifically, the reagent grade Fe₃O₄ was completely oxidized to α-Fe₂O₃ (600–700 °C) at first, and then subsequently followed the general reaction process. Immediately, some of α-Fe₂O₃ reacted with SrCO₃ to form perovskite phase SrFeO_3−x (800 °C). Thereafter, SrFeO_3−x reacted with the remaining α-Fe₂O₃ to form hexagonal SrFe₁₂O₁₉ starting from 900 °C. The entire reaction process is described by the following formula:

However, when Fe₃O₄ NPs were used as an iron source, SrFeO_3−x and SrFe₁₂O₁₉ appeared simultaneously at 700 °C as shown in Fig. 1(d). The SrFeO_3−x phase still existed at 1000 °C and was not exhausted until the reaction was completed at 1100 °C. This phenomenon indicated that not only did the formation of SrFeO_3−x was advanced to the condition that Fe₃O₄ nanoparticles had not been completely oxidized, but also Fe₃O₄ nanoparticles directly reacted to generate SrFe₁₂O₁₉ as described in reaction formula (4). In other words, reaction formula (4) may occur with reaction formulas (1) to (3) at the same time. This is attributed to the small size of Fe₃O₄ NPs with easy reactivity, which is key to optimizing the structure, morphology, and magnetic properties of hexagonal ferrites. A similar formation mechanism has been reported that the reaction of Fe₂O₃ and SrCO₃ after mechanical ball milling became Sr-Fe₃O₄, and then directly reacted with oxygen to form SrFe₁₂O₁₉ (Kaczmarek et al., 1998; Xu et al., 2017). The difference in the reaction process between Fe₃O₄ NPs and reagent grade Fe₃O₄ may lead to a variety of magnetic properties, whereas the intermediate phase transformation was often overlooked.

It can be seen from Fig. 1(d) that there is still unreacted α-Fe₂O₃ at 1100 °C which was eliminated at 1250 °C, the normal temperature of the solid-state method (Shen et al., 2021; Sozeri et al., 2012).

To further understand the formation mechanism of SrM, the DSC and TG curves of precursor powders were measured from room temperature to 1300 °C for comprehensive consideration with XRD results, as shown in black and blue solid lines in Fig. 2. The TG curves were differentiated to help understand the sintering reaction process of the mixture, as shown by the green dot lines in the DTG curves. As the AB segment of the TG curve in Fig. 2(a) shows, the mass of the mixture of Fe₃O₄ NPs and SrCO₃ decreased slowly with increasing temperature between room temperature and 400 °C; while the mass had no change in temperature ranging from 400 °C to 600 °C, as shown in the BC segment. The former was due to the release of gas and vapor absorbed on the surface of the mixture, and the latter implied there was no chemical reaction between Fe oxide and SrCO₃. The little hump in the BC segment represents the oxidation of Fe₃O₄ as described in the formula (1). However, when the temperature exceeded 600 °C, the TG curve dropped sharply, as shown in the CD segment. Meanwhile, there are two obvious peaks in the DTG curve between 600 °C and 750 °C, corresponding to the small endothermic peaks in the DSC curve around 667.8 °C and 708.6 °C. It means there are chemical reactions occurred. As shown in Fig. 1(c), SrFeO_3−x and SrFe₁₂O₁₉ were observed at 700 °C, indicating that the reaction formulas (2) to (4) took place in the CD segment. A nearly 4 % reduction of mass fraction also indicated the release of CO₂. Since the TG curve did not change as shown in the DEF segment in Fig. 2(a), no additional reaction occurred above 800 °C. That meant the main reactions were completed at 800 °C, a relatively low temperature in the solid-state method, and SrM particles mainly grew up rather than nucleated at this stage. Except for point E on the TG curve, corresponding to the endothermic peak at 1063.7 °C on the DSC curve, the reaction from the remaining SrFeO_3−x to SrFe₁₂O₁₉ was completed. Therefore, the pure SrM phase can be generated at 1100 °C or above, after which the particles grew up with increasing temperature.

Close

Fig. 2

DSC (blue) and TG (black) curves of (a) the mixture of Fe₃O₄ NPs and SrCO₃ and (b) the mixture of reagent grade Fe₃O₄ and SrCO₃. The green dot lines represent differential lines of TG curves, i.e., DTG curves.

Export to PPT

The DSC and TG curves for the reagent agent Fe₃O₄ and SrCO₃ mixture were measured from room temperature to 1300 °C, as shown in Fig. 2(b). Dissimilar to the TG curve in Fig. 2(a), the AB segment of the TG curve in Fig. 2(b) which ranged from room temperature to 700 °C represented the evaporation of absorbed water and the oxidation of Fe₃O₄. The reactions did not occur until the mass fraction decreased by almost 4 % from 700 °C to 900 °C as shown in the BC segment. Corresponding to Fig. 1(d), SrFeO_3−x and SrFe₁₂O₁₉ were found in the sequence above 700 to 900 °C, as shown in formulas (2) and (3). Besides, a broad endothermic peak without humps occurred in the DSC curve. This meant no additional reactions took place in the reagent agent Fe₃O₄ mixture. With the increase in temperature, the α-Fe₂O₃ content gradually, while the content of the SrFe₁₂O₁₉ turned to be much more compact such that the TG curve decreased slightly from 900 °C up to 1300 °C, as shown in the CD segment in Fig. 2(b). A similar TG curve of Sr-Fe₃O₄ mixture after 120 h-milled was observed by Xu et al., (2017). Based on the above discussion, it can be inferred that using Fe₃O₄ NPs as an iron resource can complete the reaction at a lower temperature and react more quickly, such that the SrFe₁₂O₁₉ phase can be obtained at 1100 °C. Due to the difference in particle size, reagent agent Fe₃O₄ reacts slowly at a higher temperature, and presents different reaction processes with Fe₃O₄ NPs.

3.2 Microstructure of SrM

Fig. 3(a) to (c) show the SEM images of SrM powders prepared by Fe₃O₄ NPs with a Fe/Sr atom ratio of 12, i.e., NP12, sintered at different temperatures, from which the changes of particle morphology and size during the formation process of SrM can be seen. As Fig. 3(a) shows, NP12 had an irregular shape, indicating that the SrM phase was not formed fully at 1000 °C. However, the hexagonal shape of the particles in Fig. 3(b) indicated that the SrM phase had been formed at 1100 °C. Moreover, NP12 at 1100 °C revealed a very nonuniform particle size distribution, with large particles on the micrometer scale but small on the nanometer. This phenomenon has special implications for magnetic properties, which are discussed below. Fig. 3(g) shows the energy dispersive spectrometer (EDS) measurement of NP11 at 1100 °C, which proves the homogeneity of our sample. And the atomic ratio shows a slight excess of Fe, which should be contributed to unreacted α-Fe₂O₃. When the temperature was 1250 °C, the particle size of SrM powders increased to ∼3 μm as shown in Fig. 3(c), which was to the disadvantage of H_cJ. In conclusion, 1100 °C is the most suitable temperature for SrM preparation from Fe₃O₄ NPs by the solid-state method.

Close

Fig. 3

SEM images (a) NP12 sintered at 1000 °C, (b) NP12 at 1100 °C, (c) NP12 at 1250 °C, (d) RG12 at 1100 °C, (e) NP10 at 1100 °C, and (f) NP11 at 1100 °C. And (g) EDS of NP12 at 1100 °C with elements identification.

Export to PPT

In contrast, the particle size of RG12 at 1100 °C was much more uniform and they were mostly greater than 1 μm as shown in Fig. 3(d). The difference between the microstructure of NP12 and RG12 is related to the formation mechanism described in Section 3.1, which is vividly displayed in Fig. 4. Since some of Fe₃O₄ NPs reacted directly to form SrM at low temperature (700 °C), this part of the particles underwent the process of nucleation and growth with temperature increasing, resulting in large particles. While the rest of Fe₃O₄ NPs were oxidized to α-Fe₂O₃, then formed SrFeO_3−x with SrCO₃, and finally SrM was generated. This process kept corresponding SrM particles small and led to a disparity in the particle size distribution.

Close

Fig. 4

The formation mechanism diagram of SrM ferrite prepared by solid-state reaction using Fe₃O₄ nanoparticles as raw material. The SrM(L) and SrM(S) represent large and small particles of SrFe₁₂O₁₉, respectively.

Export to PPT

Fig. 3(e) and (f) show the particle sizes of NP10 and NP11 at 1100 °C, which are both unevenly distributed. In terms of shape, the grains of NP10 were plate-like but NP11 had no regular shape. It can be concluded from Fig. 3(b), (e), and (f) that the Fe/Sr atom ratio had a significant effect on the microstructure of SrM. As the Fe atom ratio increased (i.e., more Fe₃O₄ NPs participated in the reaction), more nanoscale SrM particles were formed. Therefore, the final particle size distribution was more disparate. Besides, the smaller the Fe/Sr ratio was, the more favorable for the formation of plate-like particles.

After the laser particle size analyzer measurement, it can be intuitively seen in Fig. 5 that the SrM powder prepared by Fe₃O₄ NPs and reagent grade Fe₃O₄ at 1100 °C had a significant difference in particle distribution. It can be seen from the The NP12, NP11 and NP10 all had “double peaks”. This meant that there were two kinds of grain composition with a large difference in particle size. Meanwhile RG12 only had a single peak distribution. Fig. 5(a) shows the particle size distribution of NP12 with smaller particle size of 0.387 μm and larger particle size of 6.158 μm. These were consistent with the results observed by SEM in Fig. 3(b). It has been reported that the SrM powders in nanocrystalline and micron-size can be sintered together, which is called self-composite SrM, to achieve enhanced magnetic properties (Shashanka et al., 2019). The microstructure of NP12 shown in Fig. 3(b) is just like self-composite SrM as described, and a rough proportion can be obtained from Fig. 5(a) to be 70 % of nano and 30 % of micro.

Close

Fig. 5

Particle size distribution histograms of SrM powder: (a) NP12 sintered at 1100 °C; (b) RG12 sintered at 1100 °C; (c) NP10 sintered at 1100 °C; (d) NP11 sintered at 1100 °C. The red and blue lines are Gauss fit for smaller and larger particles.

Export to PPT

Fig. 5(b) shows the particle size distribution of RG12 at 1100 °C. The average particle size is 1.705 μm, larger than the single domain (∼1 μm) of SrM (Wohlfarth, 1983), which is not conducive to the high H_cJ. According to Fig. 5(c) and (d), changing the Fe/Sr atom ratio resulted in different proportions of small and large particles in SrM powders. For NP10, 43 % of small particles with size of 1.434 μm and 57 % of large particles with size of 16.431 μm were found. While for NP11, small particles with size of 1.302 μm accounted for 52 %, and large particles with size of 12.531 μm accounted for 48 %. Due to the special formation mechanism, the SrM powders prepared by Fe₃O₄ NPs present the self-composite of nanocrystalline and micron-sized particles, which may realize the regulation of the microstructure and enhancement of magnetic properties.

3.3 Magnetic properties of SrM

The effect of sintering temperature, types of Fe resource and Fe/Sr atom ratio on the magnetic properties for the obtained SrM are illustrated in Fig. 6(a) to (d). As can be seen from Fig. 6(a), the coercivity H_cJ and saturated magnetization M_S of NP12 sintered at 1100 °C and 1250 °C were found to be 348.9 kA/m, 97.8 kA/m and 65.8 emu/g, 69.5 emu/g, respectively. Due to the presence of α-Fe₂O₃, the M_S of NP12 sintered at 1100 °C was slightly lower than 1250 °C. The SrM ferrite has high uniaxial magnetocrystalline anisotropy, while the remanence M_r in randomly oriented SrM samples is usually ∼50% of the M_S, resulting a ratio of M_r/M_S (squareness ratio, SQR) of 0.5 (Liu et al., 2019). The M_r and SQR of NP12 sintered at 1100 °C and 1250 °C was 0.48 and 0.37, respectively. The SQR value of NP12 at 1100 °C is close to randomly oriented samples in literature (Pullar et al., 2012), and it is obvious that high temperature is harmful to magnetic properties. Combined with the differences in microstructure of the two samples, the lower sintering temperature had a significant effect on grain size for anisotropy. In this preparation method, the high temperature of 1250 °C caused the grain size of NP12 to be too large and this decreased the H_cJ. It has been reported that higher H_cJ can be obtained if SrM is sintered from nanocrystalline powders compared to the material derived from bulk powders (Guzman-Minguez et al., 2021; Holscher et al., 2020). The increased H_cJ of NP12 sintered at 1100 °C was mainly due to the presence of relatively smaller grains, which accounted for 70 %, in the compact sintered from the nano-sized Fe₃O₄ particles. Such high H_cJ (348.9 kA/m) of randomly oriented SrM is an improvement compared with other iron resources in solid-state method, as Table 1 shows.

Close

Fig. 6

The room temperature M−H hysteresis loops of: (a) NP12 sintered at 1100 °C and 1250 °C; (b) NP12 and RG12 sintered at 1100 °C; (c) NP10, NP11 and NP12 sintered at 1100 °C. And (d) temperature dependence of hysteresis loops, with the inset of enlargement of the second quadrant.

Export to PPT

The room temperature hysteresis loops of NP12 and RG12 sintered at 1100 °C can be seen from Fig. 6(b). RG12 with H_cJ being 259 kA/m and M_S being 52.3 emu/g had worse magnetic properties compared to NP12. This clearly shows that the different morphologies of Fe₃O₄ have a sensitive influence on the structure and properties of SrM. Moreover, the maximum magnetic product (BH)_max is also an important parameter for permanent magnets (Saura-Muzquiz et al., 2018), and the calculated (BH)_max of NP12 and RG12 are 1.521 and 1.018 MGOe, respectively. Obviously, using Fe₃O₄ NPs has advantages for higher (BH)_max.

Fig. 6(c) reveals the influence of the Fe/Sr atom ratio on the hysteresis loops of NP12, NP11 and NP10 sintered at 1100 °C. As mentioned in Section 3.2, the nano/micron ratio of the self-composition of these three specimens showed that H_cJ enhanced with an increase in Fe/Sr ratio, due to contribution from the single domain nano-sized grains (Gu et al., 2013). But the M_S of NP12, NP11 and NP10 showed a different trend, because in NP10 there was no unreacted α-Fe₂O₃ at 1100 °C and good plate-like grains as shown in Fig. 3(e), which is contributed to M_S (Kubo et al., 1985). While NP11 had α-Fe₂O₃ and the particles were mostly spherical, which seems detrimental to M_S. It can be seen from Fig. 1(b) that NP12 had the best crystallinity, resulting in a higher M_S. The different tendency of H_cJ and M_S indicates the ratio of nano- and micron-particles is a crucial factor affecting magnetic properties. The self-composite material with 50 w.t.% nano- and 50 w.t.% micron-particles have been observed to make the H_cJ decrease instead (Shashanka et al., 2019). The number of our samples is too small to see the tendency of magnetic properties with Fe/Sr ratio, but from above results, the Fe/Sr ratio of 12 seems to be the most suitable in this preparation method. Moreover, the magnetocrystalline anisotropy K₁ of our samples can be determined from magnetic hysteresis loops by applying the law of approach to saturation (LATS) (Fabian, 2006). The obtained K₁ of NP12, NP11, NP10 and RG12 are 1.45, 1.38, 1.14 and 0.97 × 10⁵ J/m³, respectively, and the value of NP12 is comparable with reported in literature (Dagar et al., 2020). It can also be seen that the high H_cJ is originated from the high K₁.

The magnetic properties of the studied SrM were compared in detail with those reported recently using different Fe sources by solid-state method (Table 1). The comprehensive magnetic properties reported in other works are not as good as this work, especially the enhancement of H_cJ without losing M_S which is beneficial to practical applications as permanent magnets. High-quality iron sources can improve the purity of the product and enhance the M_S, but the high H_cJ is difficult to obtain. In this work, the high H_cJ of SrM was attributed to the realization of self-composition with 70 % nanocrystalline and 30 % micron-sized particles by using Fe₃O₄ NPs as iron resource.

It is worth noting that the hysteresis loop of NP12 at 1100 °C shows significant “kinks” in the second and fourth quadrans, generally occurring when the hard and soft magnetic phases are uncoupled or weakly coupled (Viet Nga and Lan, 2020). So, we wonder whether there is still unreacted soft magnetic Fe₃O₄ or oxidized γ-Fe₂O₃ in this case, although these two phases are almost impossible to exist at a high temperature like 1100 °C (Feitknecht and Gallagher, 1970). In order to distinguish whether the “kinks” in the hysteresis loop are cause by the exchange coupling between the two phases in nanocomposite or by the single-phase magnet, the hysteresis loops of NP12 at low temperate were measured. Generally, the increased anisotropy of hard phase and reduced length of the exchange coupling result in decoupling at low temperature, causing two magnetization reversals in the hysteresis loops (Kneller and Hawig, 1991). However, it can be seen from Fig. 6(d) that the hysteresis loops at low temperature are smooth, indicating that there were no soft magnetic phases in NP12 at 1100 °C.

For further verify the single hard magnetic phase of NP12, we sintered Fe₃O₄ NPs individually to 1100 °C with the same heating procedure. Because the X-ray diffraction peaks of Fe₃O₄ and γ-Fe₂O₃ are the same, but almost overlap with that of SrFe₁₂O₁₉ and cannot be easily distinguished. We analyzed the phase and magnetic properties of the obtained product, displayed in Fig. 7. As shown in Fig. 7(a), only α-Fe₂O₃ was detected in the XRD measurement, indicating that Fe₃O₄ NPs were completely oxidized at 1100 °C. Then ZFC-FC curves were measured as shown in Fig. 7(b) and the ZFC was differentiated (see in the inset of Fig. 7(b)) to obtain the Morin phase transition temperature. The peak of the differential curve corresponds to the Morin phase transition temperature of 260.1 K, which was slightly lower than the reported 263 K (Morin, 1950). It has been reported that the Morin phase transition temperature will decrease with decreasing particle size (Amin and Arajs, 1987). To verify the weak ferromagnetism above the Morin phase transition temperature and the antiferromagnetism below it, the hysteresis loops of the samples at 200 and 300 K were measured. As shown in Fig. 7(c), the M−H curve at 200 K was almost a straight line, and there was almost no hysteresis at low field, indicating that the antiferromagnetism of the sample below the Morin phase transition; while the M−H curve at 300 K exhibits obvious ferromagnetism in the low field with a magnetic moment of 0.18 emu/g and a coercivity of 1 kOe. All the results show that Fe₃O₄ NPs are completely oxidized to α-Fe₂O₃ without Fe₃O₄ or γ-Fe₂O₃, further indicating that the “kinks” are not caused by the composite phase. It is because the weak ferromagnetism of α-Fe₂O₃ at 300 K, the “kinks” on the hysteresis loop of NP12 at room temperature are likely to be interfered with by residual α-Fe₂O₃, while α-Fe₂O₃ shows antiferromagnetism at low temperate, so the hysteresis loops of NP12 become smooth at low temperature.

Close

Fig. 7

The phase and magnetic properties of the product after sintering at 1100 °C from individual Fe₃O₄ NPs: (a) the XRD pattern; (b) ZFC and FC curves under an external field of 100 Oe, with the inset of differential of the ZFC curve; (c) M−H curves at 200 and 300 K with the inset of an enlarge of the curve at 300 K at low field.

Export to PPT

The results discussed above show, firstly that the SrM ferrite prepared by the simple solid-state reaction using Fe₃O₄ NPs as iron source exhibit increased coercivity of 348.9 kA/m due to a unique formation mechanism, which is also the mechanism of improved magnetic properties by mechanical ball-milling. Since the process of using Fe₃O₄ NPs to form nano- and micron- composites is shorter than that of self-composite method, we commend the method in this work as the idea of high-performance magnets, and the related microstructure of SrM ferrite deserves further research, such as combing with the molten salt method. Although “kinks” appear in the hysteresis loop at room temperature due to the weak ferromagnetism of unreacted α-Fe₂O₃, the (BH)_max exhibits to be 1.521 MGOe, which doesn’t affect it at all as a possible material for magnetic recording, data storage, etc.