MATERIALS TRANSACTIONS
Online ISSN : 1347-5320
Print ISSN : 1345-9678
ISSN-L : 1345-9678
Materials Processing
Suppression of Solidification Defects in Partial Non-Magnetization Improvement for Silicon Steel
Norihiko Hamada, Takashi Horikawa, Hironari Mitarai, Katsunari Oikawa, Satoshi Sugimoto
Author information
Keywords: interior permanent magnet (IPM) motor, rotor core, leakage flux, partial non-magnetic improvement, boron
JOURNAL FREE ACCESS FULL-TEXT HTML

2023 Volume 64 Issue 10 Pages 2508-2514

Details
Download PDF (4226K)
Download citation RIS

(compatible with EndNote, Reference Manager, ProCite, RefWorks)

BIB TEX

(compatible with BibDesk, LaTeX)

Text
How to download citation
Contact us
Abstract

Leakage flux in rotor core bridges is a problem specific to interior permanent-magnet (IPM) motors. It is widely known that the partial non-magnetization of bridges reduces the magnetic flux leakage. In a previous study, a process was proposed whereby a part of the silicon steel sheet that bridges after pressing was non-magnetized by melting and mixing Ni–Cr alloy powder with a silicon steel sheet using a laser, and the rotor core was produced by laminating them. However, because the final solidification part had solidification defects, such as cracks and shrinkage cavity, the process was proposed to leave a homogenous part free of solidification defects. Therefore, the area of the improved portion increased. We focused on developing a new alloy for non-magnetic improvement to suppress solidification defects. The improved portion was melted and mixed using a laser with various B contents to obtain a composition of Fe–(15–20) mass%Ni–(15–20) mass%Cr–(2–3) mass%Si–(0–1.6) mass%B. Large cracks and large shrinkage cavity were observed in the boron-free alloy. The cracks and shrinkage cavity decreased with an increase in the B content. The minimization of the area of non-magnetic improvement is possible by suppressing solidification defects. Consequently, the laser processing speed per piece and the amount of expensive nickel were reduced. These new alloys show promise for practical applications in the partial non-magnetization process.

To reduce the number of solidification defects formed during the nonmagnetic improvement of silicon steel, the development of a new alloy for non-magnetic improvement was carried out. The improved portion was melted and mixed by a laser with Ni–Cr powder with various B contents and silicon steel to obtain a chemical composition of Fe–(15–20) mass%Ni–(15–20) mass%Cr–(2–3) mass%Si–(0–1.6) mass%B. Figure shows the effect of B content (x mass%) in Fe–Cr–Ni–Si on the appearance. The upper photos are the topside view of each sample, and the bottom photos are the enlarged view of the final solidification area. x in Fe–Cr–Ni–Si are (a) 0, (b) 0.2, (c) 0.9 and (d) 1.6. Large cracks and large shrinkage cavity were observed in the boron-free alloy. The cracks and shrinkage cavity became smaller with an increase in the boron content.

1. Introduction

To realize carbon neutrality, vehicles with internal combustion engines are rapidly being replaced by electric vehicles. Consequently, the number of motors used in electric vehicles is increased. Leakage flux in rotor core bridges is a problem specific to interior permanent magnet (IPM) motors and remains unresolved. It is widely known that the partial non-magnetization of bridges will reduce the magnetic flux leakage, and the rotor will have a higher magnetic flux,1) as schematically shown in Fig. 1.

Fig. 1

Schematic image of a cross section of a motor showing magnetic flux flow. (a) Current bridge and (b) partial non-magnetization improvement bridge.

Several studies related to partial non-magnetization treatment have been reported.1–9) In the case of silicon steels, the bridges of the rotor are work-hardened to obtain low magnetic permeability.8) However, the magnetic polarization of the bridges is not reduced. Previous research focused on the partial non-magnetization method, in which a slit was processed in silicon steel with a laser, and a Cu or Cu–Si alloy wire was melted into the slit.9) However, these methods have not yet been put to practical use.

In our previous study,10) we proposed a new partial non-magnetization process for silicon steel. The portion of the silicon steel sheets that become the bridge after pressing can be non-magnetized by melting and mixing the Ni–Cr alloy powder with the silicon steel sheets using a laser to fabricate the rotor core by laminating them, as shown in Fig. 2. The obtained rotor core exhibits a magnetic flux 35% higher than that of a conventional rotor core. Furthermore, the rotor core with the optimized geometry of the non-magnetic improvement part achieved a 40.1% higher magnetic flux than the conventional rotor core.11) However, because the final solidification part had solidification defects such as cracks and shrinkage cavity, the process was proposed to leave a homogenous part free of solidification defects. Therefore, the area of the improved portion increases. The large improvement area shrinks significantly during cooling after laser processing, leaving residual stress in the silicon steel sheet. This causes warpage of the silicon steel sheet and raises concerns about the die life in the press process and other issues in the subsequent pressing process. In this study, we developed a new alloy for the non-magnetic improvement of silicon steel to suppress the solidification defects.

Fig. 2

Schematic image of new partial non-magnetization process for silicon steel.

2. Experimental Procedure

2.1 Sample preparation

Figure 3 illustrates the sample preparation process. The laser process consists of three steps: 1) powder scraping, 2) melting and mixing (melt mixing), and 3) residual powder removal. In the powder scraping step, the alloy powder was placed on a silicon steel sheet (50HXT780T, Nippon Steel Corporation, Tokyo, Japan) with a thickness ranging from 0.4 to 0.8 mm. Two types of powders were prepared. One was Ni–50 mass% Cr (Nippon Welding Rod Co., Ltd.), with an average particle diameter of 35.5 µm and another was Fe–16 mass%B (Koujundo Chemical Laboratory Co., Ltd.) with sieving less than 106 µm. The amount of powder scraping and the mixture of Ni–Cr and Fe–B powders were adjusted so that the chemical composition of the improved portion after melt mixing with silicon steel by a laser was Fe–(15–20) mass%Ni–(15–20) mass%Cr–(2–3) mass%Si–(0–1.6) mass%B. In the melt-mixing step, the powder and silicon steel sheet were melt-mixed using a laser at a set value of 3 mm × 10 mm. A single-type fiber laser was used with an output of 2 kW and a laser diameter of 36 µm. The laser had an output of 300–700 W, and a scan speed of 100 mm·s−1. The scanning pattern is shown in Fig. 4.

Fig. 3

Process of sample preparation.

Fig. 4

Scanning pattern of laser process.

2.2 Evaluation

Their appearance was observed using a stereoscopic microscope and an optical microscope. The cross-section of the specimen prepared by policing and etching with aqua regia was observed using a metallurgical microscope. Cross-sectional observations and chemical compositions were performed using scanning electron microscopy (SEM) and energy-dispersive spectrometry (EDS). Only boron content measurements were map-analyzed using an electron probe microanalyzer (EPMA). The magnetic polarization of a sample in a 2 mm × 2 mm cut without cracks or shrinkage cavities from the non-magnetic improvement samples was measured using vibration sample magnetometry (VSM) under a magnetic field of up to 1.6 MA·m−1. With the same specimen, the melting point was measured by a differential thermal analysis (DTA) at a heating rate of 10°C·min−1 under Ar flow. A specified composition sample was prepared by button arc melting and was cut into 4 mm × 4 mm × 12 mm in shape for the linear thermal expansion coefficient measurement by thermomechanical analysis (TMA) on the condition of the heating rate of 5°C min−1 at a load of 10 g under Ar flow.

The solidification defects were defined by two values: the crack length (Lc) and shrinkage cavity diameter (Ds). Lc and Ds were measured using a wide-area 3D measurement system (VR-5000, Keyence Corporation). Lc was the maximum linear distance from one end of the crack to the other. If multiple cracks were present, this was the sum of the cracks. The cavity with a depth of 0.8 mm or more was defined as the shrinkage cavity. Twelve specimens of the same composition were measured, and the average values were used as the crack length (Lc) and shrinkage cavity diameter (Dc).

The microstructures were quantified using cross-sectional images obtained using a metallurgical microscope. The measurement method was as follows: Microstructures with secondary dendrite arms were defined by two values: the dendrite arm width (Sw) and the gap between the secondary dendrite arms (Sg), as shown in Fig. 5. Sw and Sg were individually measured from the image by drawing an auxiliary line at the center of the secondary dendrites. Sw was measured using the cross-sectional method, and Sg was set as the gap between each cell when the structure had a cellular structure. Fifteen measurements were taken per image, and the averages of the two images were taken as Sw and Sg.

Fig. 5

Definitions of dendrite arm width (Sw) and the gap in the secondary dendrite arm (Sg).

3. Results

3.1 Effect of B content on the solidification defects

The improved portion was laser-processed with various B contents to obtain the chemical composition of Fe–(15–20) mass%Ni–(15–20) mass%Cr–(2–3) mass%Si–(0–1.6) mass% B. The effect of the B content on the appearance of the improved portion was observed, as shown in Fig. 6. The thickness of the improved portions is 1.1 to 1.4 mm, which is almost the same thickness, and it was considered possible to compare each sample. In the boron-free alloy, large cracks and large shrinkage cavity were observed, and the shrinkage cavity penetrated to the back surface. The cracks and shrinkage cavity decreased with an increase in the B content. The crack length (Lc) and shrinkage cavity diameter (Dc) for various B contents are shown in Fig. 7. The crack length (Lc) decreased with an increase in the B content, and the sample with B at 1.6% exhibited only minor cracks. The shrinkage cavity diameter (Dc) decreased with an increase in the B content, and no shrinkage cavity was observed above a B content of 1%. The improved portions with various B contents were cut to 2 mm × 2 mm in shape without defects, and the hysteresis loops were measured, as shown in Fig. 8, and compared to silicon steel. The boron-free sample shows a magnetic polarization of approximately 0 T owing to the precipitation of the austenite phase according to the Schaeffler diagram.12–14) Up to a B of 1.6%, magnetic polarization is much lower than that of silicon steel, and the addition of B in Fe–Cr–Ni–Si does not affect magnetism in this measurement range.

Fig. 6

Effect of B content (x mass%) in Fe–Cr–Ni–Si on the appearance. The upper photos are the topside view of each sample, and the bottom photos are the enlarged view of the final solidification area. x in Fe–Cr–Ni–Si are (a) 0, (b) 0.2, (c) 0.9 and (d) 1.6.

Fig. 7

Effect of B content in Fe–Cr–Ni–Si on (a) crack length (Lc) and (b) shrinkage cavity diameter (Ds).

Fig. 8

Hysteresis curves with various B contents compared to silicon steel.

3.2 Effect of B on the structure

The cross-sectional microstructures were observed using a metallurgical microscope with various B contents, as shown in Fig. 9. In the boron-free sample, a cell-like structure, which is the primary crystal indicated by the white area in Fig. 9(a), was observed. The structure shows finer and well-developed secondary arms (complex structure), indicated by the white area in Fig. 9(b)–(d), and an increase in the amount of the eutectic structure, indicated by the black area in Fig. 9(b)–(d), with increasing B content. The dendrite arm spacing is often used to quantify dendrites.15–24) In this study, because it is necessary to quantify the fine and complex structure of the dendrite, the dendrite arm width (Sw) and the gap between the secondary dendrite arms (Sg) were defined, as shown in Fig. 5. The dendrite arm width (Sw) and gap between the secondary dendrite arms (Sg) were measured for various B contents from the cross-sectional images, as shown in Fig. 10. Since the sample in Fig. 9(a) had a cell structure, Sw and Sg were measured, as described in the experimental procedure. The dendrite arm width (Sw) decreased sharply and then slowly as the B content increased. The gap between the secondary dendrite arms (Sg) increased sharply and then slowly as the B content increased. It can be observed that the number of eutectic structures increased, which agrees with Fig. 9. The elemental mapping of the cross-section of the sample was analyzed using EPMA, as shown in Fig. 11. Large amounts of B, Cr, and Si are observed in the eutectic structure. The addition of B increases the eutectic structure.

Fig. 9

Microstructures by a metallurgical microscope with various B content (x mass%) in Fe–Cr–Ni–Si. x in Fe–Cr–Ni–Si are (a) 0, (b) 0.2, (c) 0.9 and (d) 1.6.

Fig. 10

Effect of B content in Fe–Cr–Ni–Si on (a) dendrite arm size and (b) the gap in the secondary dendrite arms.

Fig. 11

SEM and EPMA analyses with various B content (x mass%) in Fe–Cr–Ni–Si. x in Fe–Cr–Ni–Si are (a) 0, (b) 0.2, (c) 0.9 and (d) 1.6.

The effects of the dendrite arm width (Sw) and gap between the secondary dendrite arms (Sg) on the crack length (Lc) were evaluated, as shown in Fig. 12. When the dendrite arm width (Sw) decreased and the gap between the secondary dendrite arms (Sg) increased, the crack length (Lc) decreased. Because the amount of eutectic structure increases with a decrease in the dendrite arm width (Sw) and an increase in the gap between the secondary dendrite arms (Sg), the amount of eutectic structure is also an important factor for crack suppression.

Fig. 12

Effect of (a) the dendrite arm width (Sw) and (b) the gap in the secondary dendrite arms (Sg) on the crack length (Lc).

3.3 Characteristics of cracks

To evaluate the characteristics of a crack, the cross-section of the crack in boron-free and 0.9 mass%B alloys was observed, as shown in Fig. 13. The crack surfaces of each specimen were on the dendrite surface and not on the dendrites. These cracks are classified as solidification cracks.25–32) Solidification cracking occurs near the solidus line and also grow during cooling. The occurrence of solidification cracking and the growth of solidification cracks are discussed in the next chapter.

Fig. 13

Cross-section of the crack with (a) B of 0 mass% and (b) B of 0.9 mass%.

4. Discussion

First, the occurrence of solidification cracking was discussed. The relationship between the microstructures and cracks in TIG-welded austenitic stainless steel was investigated by Senda et al.33) Large cracks were observed when the microstructure exhibited relatively thick primary dendrite arms and insufficiently grown secondary arms. In contrast, small cracks were observed when the microstructure showed thin primary dendritic arms and fine complex secondary arms. The microstructure with a simple cell structure, as shown in Fig. 9(a), was similar to that observed in large cracks. As shown in Figs. 9(b), (c), (d), the microstructures with fine and complex structures, which have a low dendrite arm width and a wide gap between the secondary dendrite arms, are similar to those observed in the small cracks reported by Senda et al.33)

Kou28) and Homma et al.34) reported that the liquid phase can be fed into the initial cracks immediately and that cracks can be healed during the start of the solidification stage. However, in the final stage, the liquid feed was insufficient, and cracks occurred at the liquid film interface. The specimen without B has a small eutectic region, as shown in Fig. 9(a); therefore, the formation mechanism of cracks in the present process is similar to that reported by S. Kou28) and Homma et al.34) The boron-added specimens have a large eutectic region, as shown in Figs. 9(b), (c), and (d). The addition of B decreases and separates the liquidus and solidus temperature. In the final solidification stage, liquid feeding is sufficient among between the liquidus and the solidus temperature, and cracks must be suppressed. The addition of B resulted in a finer and more complex structure with a large number of eutectics, which suppressed the occurrence of solidification cracks.

Next, the growth of solidification cracks is discussed. The melting points and linear thermal expansion coefficients were measured, as shown in Fig. 14. Melting point represents the solidus temperature because these data were obtained during the heating process. The melting point decreased rapidly and saturated at 1300°C as the B content increased. The linear thermal expansion coefficient decreased with an increase in the B content. The suppression of the solidification crack growth was considered because the absolute amount of shrinkage of the improved portion was reduced owing to the lower melting point and lower coefficient of linear expansion with an increase in the B content. As described above, investigations into crack suppression have progressed; however, shrinkage cavity suppression requires further investigation.

Fig. 14

Effect of B content in Fe–Cr–Ni–Si on (a) melting point (solidus temperature), Tm, and (b) linear thermal expansion coefficient, a.

The suppression of solidification defects made it possible to minimize the area of the non-magnetic improved portion. Consequently, the laser processing speed per piece and the amount of expensive nickel can be reduced. The boron-added alloy promises practical applications in the partial non-magnetization process.

5. Conclusion

We focused on the non-magnetic improvement composition to suppress solidification defects such as cracks and shrinkage cavities and obtained the following conclusions:

  1. (1)    Large cracks and a large shrinkage cavity were observed in the boron free alloy. The cracks and a shrinkage cavity become small by the addition of B.
  2. (2)    Microstructures with various B contents were observed. A cell-like structure was observed in the boron-free alloy, and the microstructure showed a finer, well-developed secondary arm (complex structure) and an increase in the eutectic structure with increasing B content. Large amounts of B, Cr, and Si in the eutectic structure were analyzed using EPMA, and it was considered that the addition of B increased the eutectic structure.
  3. (3)    The cracks were classified as solidification cracks based on microstructure observations. The addition of B resulted in a finer and more complex primary dendritic structure with a large eutectic structure, which suppressed the occurrence of solidification cracks. The suppression of the solidification crack growth was considered because the absolute amount of shrinkage with the improvement portion was reduced owing to the lower melting point and lower linear thermal expansion coefficient with an increase in the B content.

Acknowledgments

This research was a result of commissioned work (JPNP20001891-0) and was supported by the Green Innovation Fund subsidy project (JPNP22100313-0) of the New Energy and Industrial Technology Development Organization (NEDO), a national research and development agency.

REFERENCES
 
© 2023 The Japan Institute of Metals and Materials
feedback
 Top