Introduction

Fe-Si alloys are widely used as soft magnetic materials in motors, transformers and power generators. When the Si content reaches 6.5 wt.%, Fe-Si alloy has the best magnetic properties due to its high electrical resistivity, high permeability, low magnetocrystalline anisotropy and near zero magnetostriction, which result in low core losses and low noises [1, 2].

When the Si content increases, the material becomes too brittle to be rolled without cracking [3]. The main cause for brittleness is two kinds of ordered phases: B2 and DO3 [4], because antiphase boundaries prevent the cross slips of dislocations. Many researchers have attempted to suppress the ordered phases and fabricate Fe-6.5 wt.%Si thin sheet by various methods. Liang et al. [5] reported that hot, warm and cold rolling was successfully done to final thickness of 0.03-0.05 mm by proper rolling and heat treatment process. The sheet was not fractured and exhibited good ductility between hot and cold rolling process. Researchers also investigated surface diffusion techniques in order to produce high Si steel sheet. The techniques are alternative routes to obtain high Si steel sheets without rolling problem. Yamaji et al. [6] produced Fe-6.5 wt.%Si sheets by CVD technique in the NKK Corporation. This technique used a chemical reaction, SiCl₄ + 5Fe = Fe₃Si + 2FeCl₂, on the basic material of cold rolled 3 wt.% Si steel.

Powder metallurgy is also applicable to fabricate high Si steel sheets. Li et al. [7] reported that Fe-6.5 wt.%Si alloy strip was fabricated by the direct powder rolling of the admixture containing 93.5 wt.%Fe powder and 6.5 wt.%Si powder. The strip exhibited excellent magnetic properties in high frequencies of 1-10 kHz. However the Fe-6.5 wt.%Si sheet fabricated by Fe-6.5 wt.%Si powder rolling have been little studied.

In this work hot-rolled Fe-6.5 wt.%Si sheets were fabricated from SPSed preforms which were consolidated by spark plasma sintering process after gas atomizing Fe- 6.5 wt.%Si alloy into powders. Their microstructure and soft magnetic properties were investigated to study the effect of powder hot-rolling reduction and heat treatment on magnetic properties.

Experimental

Fe-6.5 wt.%Si powder was produced by a high pressure Ar gas atomization in a clean closed system. The atomized powders with the size range from 90 to 150 mm were consolidated to disc shaped preforms with a diameter of 50 mm and height of 10 mm by spark plasma sintering (SPS).

The SPSed preforms were canned by copper container with thickness of 3 mm and were hot rolled at 1173 K with total reduction ratio of 70, 80 and 90% for 4, 5 and 8 passes, respectively. The hot rolled sheets were annealed at 1273 K for 1h in flowing Ar atmosphere and were subsequently quenched in water or liquid nitrogen to suppress the formation of ordered phases.

Magnetic properties were examined by vibration sample magnetometer (VSM) and AC B-H analyzer. The specimens with a dimension of 1×2×4 mm were prepared to measure saturation magnetization in the field strength of 10 kA/m using VSM. The toroidal specimens with a dimension of Φ 18 mm outer diameter, Φ 8 mm inner diameter and 1 mm thickness were used to measure the core losses in a frequency range of 50, 100, 400 and 1000 Hz and magnetic induction of 0.5 T by AC BH analyzer. Microstructure observations were carried out by using scanning electron microscopy (SEM). Orientations and boundaries of grains and particles were analyzed by electron back-scattered diffraction (EBSD). The elements mappings of Fe, Si and O were carried out by an electron probe micro-analyzer (EPMA).

Results & discussion

3.1. Saturation magnetic flux density and core loss of Fe-6.5 wt.%Si sheets

Fig. 1 shows the variation of the saturation magnetic flux density with the total reduction ratio of hot rolled sheets. Saturation magnetization is the intrinsic property of the material. Therefore all of the samples had almost the same saturation magnetization values of 188-191 emu/g, whereas saturation magnetic flux density gradually increased with total reduction ratio of hot rolling as shown in Fig. 1. This result was caused by increase of relative density in Fe-6.5 wt.%Si sheets, which increased from 95.5% to 99.7% with total reduction range of 70-90%.

Fig. 1

Variation of saturation magnetic flux density in Fe- 6.5 wt.%Si sheets with total reduction ratio during hot rolling process.

Fig. 2 shows the variation of core loss in magnetic induction of 0.5 T with frequency range of 50-1000 Hz in hot rolled Fe-6.5 wt.%Si sheets. When the total reduction ratio increased from 70 to 80%, core loss (W_5/1000) decreased about 10%, whereas when hot rolling reduction increased from 80 to 90% core loss (W_5/1000) increased by about 1.5 times. In order to investigate these changes, we needed to see more detailed changes in the core loss by dividing the total core loss to the hysteresis loss and the eddy current loss. The hysteresis loss is given by [8]

Fig. 2

Variation of core loss of Fe-6.5 wt.%Si sheets in magnetic induction of 0.5 T with total reduction during hot rolling process ratio in the frequency range of 50, 100, 400 and 1000 Hz.

(1)

Ph=f×Wh[W/m3]

Where, W_h is a B-H hysteresis loop area and f is the frequency. It is known that the hysteresis loss increases with existence of defects, impurities, grain boundaries, etc. because they prevent the magnetic domain wall movement [2].

As shown in Fig. 2, the hysteresis losses had almost same values at a specific frequency regardless of total reduction ratio of hot rolling. Therefore the differences of the total core losses were determined by the differences of the eddy current losses.

The eddy current loss is given by [8]

(2)

Pe=π2f2B2t2cρ[W/m3]

Where, B is amplitude of magnetic flux density, t is a sample thickness, ρ is the resistivity of the material and c is a constant. In this case f, B and t have fixed, so ρ is dominant factor to determine the eddy current loss. The electrical resistance is caused by the scattering of electrons due to disturbance in the crystal structure such as solute elements, impurities, grain boundaries, dislocations, etc [9]. Fig. 3 shows the variation of electrical resistivity with total reduction ratio in Fe-6.5 wt.%Si sheets. The changes of electrical resistivity exhibited inverse tendency with the changes of eddy current loss, whereas they were similar to the changes of hysteresis loss. Because the factors increased the resistivity resulting from giving rise to electron scattering and also hindered movement of magnetic domain wall. However, especially in the high frequency range, the contribution of the eddy current loss to total core loss was higher than that of the hysteresis loss. It was because the latter was proportional to the frequency while the former was proportional to the square of frequency. Therefore the increase of electrical resistivity leading to the decrease of the eddy current loss is dominant factor to obtain low core loss at high frequency and we needed to know the causes that increased the electrical resistivity.

Fig. 3

Variation of (a) electrical resistivity and (b) eddy current loss of Fe-6.5 wt.%Si sheets with total reduction ratio during hot rolling process.

3.2. Microstructure of Fe-6.5 wt.%Si sheets

Fig. 4 shows EBSD grain boundary maps and IPF maps of hot rolled Fe-6.5 wt.%Si sheets at the total reduction ratio of 70, 80 and 90%, respectively. In Fig. 4(a)-(c), the black and red lines indicated high angle boundaries (HABs) that having misorientation angle θ > 15° and low angle boundaries (LABs) that having misorientation angle 2 ≤ θ ≤ 15°, respectively.

Fig. 4

EBSD grain boundary maps of Fe-6.5 wt.%Si sheets from the direction of ND and EBSD IPF maps of Fe-6.5 wt.%Si sheets from the direction of ND((d), (e), (f)) and RD((g), (h), (f)): sample A, B, C were rolled with total reduction ratio of 70%, 80% and 90%, respectively.

The increase of HAB density in the reduction range of 70-80% occurred due to the thickness reductions of not only particles but also particle boundaries during hot rolling process. In Fig. 4(a), each of primary particles having single orientation represented the fact that the grains in the particle had grown as the size of one particle but it could not grow beyond the particle boundary. It was due to the particle boundaries were still imperfect and too thick to easily diffuse the atoms from the one particle to the other particle at the total hot rolling reduction of 70%. Fig. 4(b) shows the particle boundaries were compacted and LABs were created at the total reduction ratio of 80%. Additional hot rolling to 80% made the particle boundaries compact and crystallographic orientation similar. These tendencies were definitely displayed at total reduction ratio of 90%. In Fig. 4(c), the particle boundaries disappeared due to the merging of the particles and as a result the HAB density decreased. The particle boundary disappearing was caused by high strain and development of rolling texture. The true strains with reductions of 70, 80 and 90% were 1.2, 1.6 and 2.3. When the reduction ratio increased from 80 to 90%, the increase of the true strain had much higher value than that of from 70 to 80%. So particle boundary disappearing was able to be clearly observed at total reduction ratio of 90%. The creations of LABs could be understood as traces of the disappeared particle boundaries.

It was known that the crystallographic orientations in ferritic steels were assembled along <110>//RD orientations after rolling process. Therefore, with increasing hot rolling reduction, the <110>//RD rolling texture was developed and each grains had similar orientations. It could be observed in EBSD IPF maps of Fig. 4(d)-(i). Each of the colors in the maps representεd the crystallographic direction in the view of specific sample axis as ND, RD or TD. Fig. 4(d)-(f) were IPF ND maps and Fig. 4(g)-(i) were IPF RD maps. With increasing the total reduction ratio of hot rolling, the orientations were concentrated around {112}<110> or {113}<110>. These orientations were typical orientations in the <110>//RD rolling texture and the volume fractions of these orientations gradually increased with hot rolling reduction. The rolling texture development increased inter-particle coherency, it made merging of particles into one grain. Moreover, inner atoms of the primary powder were pushed out to the particle boundary region because the area of the particle surface increased with particle aspect ratio during hot rolling. The calculated aspect ratios in reduction of 70, 80 and 90% were 11, 25 and 100, respectively. But the measured aspect ratio was less than the calculated value since the particle merging. Therefore, by increase of hot rolling reduction to 90%, many particles were merged and particle boundaries tended to disappear. It was the main cause of electrical resistivity decreasing in the hot rolling reduction range of 80-90%.

Oxide layers on the particle boundaries also affect electrical resistivity because they can perform as the insulation layer [10]. Fig. 5 shows SEM micrographs and EDS dot mapping images of Fe, Si and O in the hot rolled Fe- 6.5 wt.%Si sheets with the total reduction ratio of 70, 80 and 90%, respectively. The initial atomized powder had thin oxide layer on the particle surface. As shown in Fig. 5(a)-(c), the oxide layers had broken and widely dispersed with increasing hot rolling reduction in hot rolled Fe-6.5 wt.%Si sheets. The oxides distributed along particle boundaries until the total reduction ratio of 80%. So they could perform as insulating layer and contributed to the increase of electrical resistivity. But when total reduction increased to 90%, the oxides were homogeneously dispersed all over the material. It was related to disappearing of particle boundaries and less helpful to increase electrical resistivity than oxide distribution of 80% reduction.

Fig. 5

SEM micrographs and EDS dot mapping images of Fe, Si and O in the hot rolled Fe-6.5 wt.%Si sheets with the total reduction ratio of (a) 70, (b) 80 and (c) 90%, respectively.

Conclusions

Soft magnetic properties of Fe-6.5 wt.%Si alloy sheets fabricated by powder hot rolling process were investigated to study the effect of powder hot-rolling reduction and heat treatment on magnetic properties. To identify the differences of magnetic properties in Fe-6.5 wt.%Si sheets, electrical resistivity, microstructure, crystallographic orientation and elements distribution were measured and the following conclusions can be drawn.

• (1) The saturation magnetic flux density increased with total reduction ratio of hot rolling due to the increase of density of the material. The particle boundaries in Fe-6.5 wt.%Si sheets were greatly compacted with increasing of hot rolling reduction.

• (2) Core loss decreased with increasing of hot rolling reduction except the case of hot rolling reduction range from 80 to 90%. The decrease of core loss were caused by the increase of electrical resistivity that made the eddy current loss lower. The increase of HAB density with hot rolling reduction was main cause of the increase of the electrical resistivity.

• (3) When total reduction ratio of hot rolling increased from 80 to 90%, core loss increased with showing the inverse tendency with others. It was the result of particle boundary disappearing. The development of rolling texture and high strain made the particle boundary disappeared and particles merged. Therefore electrical resistivity decreased with HAB density that made core loss higher.

• (4) The oxides made some contribution to increase electrical resistivity. The oxide layer on the initial powder surface was broken and dispersed during hot rolling process.

Acknowledgements

Acknowledgements

This work was supported by the Energy Technology Development Program (Project No. 20145020300900) funded from the Ministry of Trade, Industry and Energy (MOTIE) through the Korea Institute of Energy Technology Evaluation and Planning (KETEP).

References

• 1. R.M. Bozorth, Ferromagnetism., (1951) New York Van Nostrand.
• 2. B.D. Cullity and C.D. Graham, Introduction to Magnetic Materials., (2009) 2nd edNew Jersey John Wiley & Sons, Inc..
• 3. Y. Sato, T. Sato and Y. Okazaki: Mater. Sci. Eng. A., (1988) 99 73.
• 4. B. Viala, J. Degauque, M. Baricco, E. Ferrara, M. Pasquale and F. Fiorillo: J. Magn. Magn. Mater.., (1996) 160 315.
• 5. Y.F. Liang, F. Ye, J.P. Lin, Y.L. Wang and G.L. Chen: J. Alloys Compd.., (2010) 491 268.
• 6. T. Yamaji, M. Abe, Y. Takada, K. Okada and T. Hiratani: J. Magn. Magn. Mater.., (1994) 133 187.
• 7. R. Li, Q. Shen, L. Zhang and T. Zhang: J. Magn. Magn. Mater.., (2004) 281 135.
• 8. Das Jiten, K. Chandra, P. S. Misra and B. Sarma: J. Magn. Magn. Mater.., (2008) 320 906.
• 9. S. Nagarjuna, K. Balasubramanian and D.S. Sarma: Mater. Sci. Eng. A., (1997) 225 118.
• 10. P. Jang and B. Lee: IEEE Trans. Magn.., (2009) 45 2781.

Figure & Data

References

Citations

Citations to this article as recorded by