Fe-Co alloys are promising materials for high power density motors, since they have higher Bs than Fe-Si steels which are most commonly used as the motor cores. It is known, however, that the Fe-Co alloys generally have poor workability due to an existence of the brittle phase, ordered B2. In this study, effects of Co content and Si/Al addition with the Fe-Co alloys on the workability and magnetic properties were investigated. The neutron diffraction measurements of the Fe-Co alloys with different Co contents of 5, 18, 27, and 49 mass% were first conducted to know the existence of the ordered B2 phase. The result showed that the B2 phase was observed in the Fe-Co alloys with Co contents of 27 and 49 mass%, while the alloys with 5 and 18 mass% did not contain it, which suggests the latter two alloys have the better workability. Then, we checked the effect of Si/Al addition on the magnetic properties of the 82Fe-18Co alloy, and found that the co-addition of Si/Al reduce the coercivity, Hc to about 60 A/m and iron loss to less than 170 W/kg while maintaining a high B value at 30 000 A/m above 2.2 T.
I. INTRODUCTION
It is estimated that around 45% of global electric power is consumed by electric motors, resulting in over 6 Gt/year of CO2 emissions.1 Accordingly, it is important to improve the motor efficiency in order to realize the more sustainable society. Soft magnetic material used as the motor core is one of the key materials for the development higher efficiency motors.
It is known that the addition of Si or Al to Fe generally improve the soft magnetic properties; the electrical resistivity increases, the magnetic crystalline anisotropy decreases, and the magnetostriction constant approaches to zero, resulting in reducing the iron loss.2 This is the reason why the electrical steels, Fe-Si are most commonly used as the core materials for various electric motors.
However, it seems that the Bs of the conventional Fe-Si material is too low to meet the specification for the higher power-density motors installed into the electric vehicles, the electric aircrafts, drones, and the future mobility machines such as flying cars. Accordingly, development of the high Bs material with a low core-loss is thought to be strongly desired. It is well known that Fe-Co alloys have the highest Bs among all the bulk materials. The Permendur, 49Fe-49Co-2V (mass%) is already used in some high-performance motors, since it has good soft magnetic properties. However, drawbacks of the Permendur are poor workability and machinability due to the existence of a brittle phase, ordered B2 one. One of the ideas to suppress the formation of the B2 phase is to reduce the amount of Co in the Fe-Co alloy. However, there are few reports on the relationship between the B2 phase formation and Co content in the Fe-Co alloy, experimentally determined. Also, we cannot adopt the idea of reducing the Co content immediately, because the soft magnetic properties of Permendur deteriorate by reducing the Co content.3
The motivation of this study is to develop a high performance soft magnetic material with high Bs, low iron loss, and good machinability for the core of the high efficiency, high power-density motors. We first conducted neutron diffraction measurements on the Fe-Co alloys with different Co contents to check the existence of the B2 phase in these alloys. And then, the effect of Si/Al addition on microstructure and magnetic properties of 82Fe-18Co alloy were investigated in order to improve the soft magnetic properties.
II. EXPERIMENTAL PROCEDURE
Four ingots with different chemical compositions shown in Table I were prepared by induction-melting and casting using the vacuum melt furnace. The ingots were annealed at 1373 K for 7.2 ks in air, and then forged into a cylinder shape with an outer diameter of 20 mm.
Alloy . | Composition (mass%) . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
C . | Si . | Mn . | S . | Ni . | Cr . | V . | Co . | Al . | Fe . | |
49Fe-49Co-2V | 0.009 | 0.05 | 0.06 | 0.005 | 0.05 | 0.03 | 1.97 | 48.70 | 0.03 | Bal. |
73Fe-27Co | 0.007 | 0.24 | 0.25 | 0.005 | 0.62 | 0.61 | 0.27 | 27.16 | 0.04 | Bal. |
82Fe-18Co | <0.001 | <0.01 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 18.15 | 0.005 | Bal. |
95Fe-5Co | <0.001 | <0.01 | <0.01 | 0.0007 | 0.01 | <0.01 | <0.01 | 5.00 | 0.002 | Bal. |
Alloy . | Composition (mass%) . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
C . | Si . | Mn . | S . | Ni . | Cr . | V . | Co . | Al . | Fe . | |
49Fe-49Co-2V | 0.009 | 0.05 | 0.06 | 0.005 | 0.05 | 0.03 | 1.97 | 48.70 | 0.03 | Bal. |
73Fe-27Co | 0.007 | 0.24 | 0.25 | 0.005 | 0.62 | 0.61 | 0.27 | 27.16 | 0.04 | Bal. |
82Fe-18Co | <0.001 | <0.01 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 18.15 | 0.005 | Bal. |
95Fe-5Co | <0.001 | <0.01 | <0.01 | 0.0007 | 0.01 | <0.01 | <0.01 | 5.00 | 0.002 | Bal. |
The samples for neutron diffraction measurement were prepared by cutting into small bar-shaped pieces of 3 mm × 3 mm × 35 mm from the forged samples. The neutron diffraction measurement was conducted with iMATERIA spectrometer at the Materials and Life Sciences Facility at the J-PARC (Japan Proton Accelerator Research Complex). Ingots shown in Table II for a study of their magnetic properties and microstructures were also prepared by induction-melting and casting using the vacuum melt furnace, followed by the annealing at 1373 K for 7.2 ks in air.
Alloy . | Composition (mass%) . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
C . | Si . | Mn . | S . | Ni . | Cr . | V . | Co . | Al . | Fe . | |
82Fe-18Co | <0.001 | <0.01 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 18.15 | 0.005 | Bal. |
82Fe-18Co-1.5Si | <0.001 | 1.48 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 18.12 | 0.01 | Bal. |
82Fe-18Co-1.0Si-0.5Al | <0.001 | 0.99 | <0.01 | 0.0005 | 0.01 | <0.01 | <0.01 | 18.1 | 0.47 | Bal. |
82Fe-18Co-0.5Si-1.0Al | <0.001 | 0.50 | <0.01 | 0.0006 | 0.01 | <0.01 | <0.01 | 18.02 | 1.02 | Bal. |
82Fe-18Co-1.5Al | <0.001 | <0.01 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 17.9 | 1.51 | Bal. |
Alloy . | Composition (mass%) . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|
C . | Si . | Mn . | S . | Ni . | Cr . | V . | Co . | Al . | Fe . | |
82Fe-18Co | <0.001 | <0.01 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 18.15 | 0.005 | Bal. |
82Fe-18Co-1.5Si | <0.001 | 1.48 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 18.12 | 0.01 | Bal. |
82Fe-18Co-1.0Si-0.5Al | <0.001 | 0.99 | <0.01 | 0.0005 | 0.01 | <0.01 | <0.01 | 18.1 | 0.47 | Bal. |
82Fe-18Co-0.5Si-1.0Al | <0.001 | 0.50 | <0.01 | 0.0006 | 0.01 | <0.01 | <0.01 | 18.02 | 1.02 | Bal. |
82Fe-18Co-1.5Al | <0.001 | <0.01 | <0.01 | 0.0004 | 0.01 | <0.01 | <0.01 | 17.9 | 1.51 | Bal. |
Then, the ingots were forged into rectangle shape with dimensions of thickness 40 mm and width 100 mm The forged samples were hot-rolled into thickness of 4 mm after being annealed at 1173 K for 7.2 ks, and then post-annealed at 1023 K for 21.6 ks in an Ar atmosphere. The post-annealed samples were cut into a thickness of 0.67 mm, and then rolled into 0.2 mm, i.e., with a reduction of 70%, at 473 K where the samples are ductile. (Hereafter, these rolled samples will be abbreviated as “rolled samples.”) The evaluation procedure of ductile-to-brittle transition temperature (DBTT) will be described later. The rolled samples were heated into 1223 K with a heating rate of 287 K/s in a mixed-gas of nitrogen and hydrogen, then held at 1223 K for 240 s followed by air-cooling. (Hereafter, these samples will be abbreviated as the “annealed samples.”)
The specimens for an evaluation of the workability were prepared by cutting from the annealed samples. The impact values were measured at the temperature range from 293 K to 423 K using a Charpy impact tester. The ductile-to-brittle transition temperature (DBTT) was determined from the impact values measured at each test temperature. The DBTT was evaluated as the energy transition temperature (TrE). The microstructure observation was conducted using an optical microscope (OM). The hardness was measured by a micro-hardness tester with a load of 0.1 kg. Here, the Vickers hardness was the average of values at seven measuring points. Texture analysis of the rolled samples and annealed samples were performed by observing the polished surface with a scanning electron microscope-electron back scatter diffraction pattern (SEM-EBSD).
In order to evaluate the texture quantitatively, the A-parameters were calculated from the Orientation Distribution Function (ODF).4 The texture parameter Aθ is defined as the orientation averaged value of A(g), taking into account the volume fraction of each texture component, for an arbitrary magnetic field direction θ. This orientation weighted average value can be expressed by the convolution integral of Aθ(g) and the frequency f(g) of any crystal orientation obtained by ODF. (1) In rotating electrical machine applications, the directions of the magnetizing field will be equally distributed in plane of the rolled sheet. Accordingly, a direction averaged A-parameter must be considered as (2) in which the integral expands over all possible planar directions of the rolled sheet.
In this study, the A-parameters were calculated on the presupposition that the magnetic field is applied along rolling plane the crystallographic orientation is ⟨100⟩ direction, the easy axis of magnetization. In this case, if the calculated angle as the A-parameter is small, the ⟨100⟩ oriented grains in the rolling plane is highly concentrated, in other words, the sample has a good soft magnetic property.
The toroidal core for the measurement of the soft magnetic properties was made by the following procedure. Five ring samples with an outer diameter of 28 mm × inner diameter of 20 mm made by the etching the annealed samples were stacked and set into the holder. Here, the insulation paper was inserted into the interlayer of each ring. The holder was wound with copper wire and made into a toroidal core used for the measurement of soft magnetic properties. DC magnetization characteristics were measured using a BH tracer up to a maximum applied magnetic field of 30 kA/m. Iron loss measurements were performed using an AC BH analyzer with a constant maximum excitation flux density of 1.5 T and varying frequency with the range of 0.05 to 10 kHz.
The hysteresis loss Ph and eddy current loss Pe in the iron loss Pc were calculated from Eqs (3) and (4). We extrapolated the Pc/f to 0 Hz to obtain the hysteresis loss coefficient per period (Kh = Ph/f), from which the hysteresis loss Ph was calculated at each frequency.5
III. RESULTS AND DISCUSSION
Figure 1 shows neutron diffraction patterns of the four samples shown in Table I. The 49Fe-49Co-2V alloy and the 73Fe-27Co alloy clearly had the ordered B2 phase, while the other alloys, the 82Fe-18Co and the 95Fe-5Co did not contain B2. This result agrees with the Fe-Co binary phase diagram past reported, and it is suggested that the latter two alloys have a good workability. We chose the 82Fe-18Co alloy for the further investigation in this study.
Figure 2 shows the DBTT of the samples shown in Table II. The DBTT of the 82Fe-18Co and 82Fe-18Co-1.5Al was room temperature. On the other hand, the 82Fe-18Co-1.5Si alloy showed a higher DBTT of 423 K. The DBTT of the Si/Al-co-added 82Fe-18Co alloy was intermediate value in these samples. It is generally thought that a deformation twin, which is generated during the plastic deformation process by a concentration of the stress resulted from the accumulation of the dislocation, causes the brittle fracture in the Fe-Si alloy with the BCC structure.6,7 Also, Griffiths and Riley reported that the solid solution of Si atom into the Fe BCC crystal lowers the stacking fault energy of the Fe crystal and suppresses the cross slip of dislocations.8 The embrittlement of the Fe-Co alloy by the Si-addition shown in Fig. 2 may be caused by same reason as reported in these previous literatures.
Figures 3(a) and 3(b) show the microstructures of the cross section along the transverse direction with the as rolled samples and the annealed samples of 82Fe-18Co and Si/Al-added 82Fe-18Co observed by OM. Vickers hardness and average grain size calculated from image analysis for the annealed samples are also shown in Fig. 3. Elongated crystals along the rolling direction were observed in all as-rolled samples. After the heat-treatment, the elongated crystals changed into the larger, equiaxed shape through their recrystallization process at annealing temperature. The hardness decreased after annealing. The hardness of the Si-added samples is higher than that of the sample without Si-addition, probably due to the same reason as described for Fig. 2. The grain size of the Si/Al-added heat-treated samples was larger than that of the samples without Si/Al-addition.
The as rolled textures of the rolled samples are given in Fig. 4(a). The rolled texture of 82Fe-18Co consisted of the α-fiber (RD//⟨110⟩) and the γ-fiber (ND//⟨111⟩), and the main orientation was {111}⟨110⟩ which is classified into the γ-fiber.
Barnett and Kestens reported that the rolled texture of the Interstitial Free (IF) steel with the reduction of 65% reduction consisted of the α-fiber and the γ-fiber, and its main orientation was {111}⟨112⟩ which is classified into the γ-fiber.9 It is thought that the 82Fe-18Co in this study exhibited a similar behaviour for the rolling texture formation as that for the Fe. The rolled texture of the Si/Al-added 82Fe-18Co alloys also showed the α-fiber and the γ-fiber. The only-Al-added 82Fe-18Co showed a similar rolled texture similar as the 82Fe-18Co. When the Si content in the Si/Al-added 82Fe-18Co alloys was higher, the texture changed from the γ-fiber to the α-fiber, and the main orientation of the only-Si-added 82Fe-18Co became to {100}⟨110⟩.
Tomita et al. reported that the rolled texture of the pure iron changed from a mixed state of the α-fiber and the γ-fiber to the α-fiber when the reduction ratio increased from 90% to 99.8%.10 This change is a similar behaviour as that of the 82Fe-18Co with the increase of the Si content in this study. This suggests that the solid solution of Si atom to the Fe-Co lattice suppresses cross slip and the dislocations are more likely to accumulate as already mentioned above, even though the rolling reduction rate is relatively low of 70%. Shingaki et al.11 and Ruzakov et al.12 reported that the {100}⟨110⟩ is stable as the main orientation direction for the cold-rolled Fe-3.2 mass% Si sample with the reduction of 66% accompanied with the twin deformation. The fact that the main orientation of the 82Fe-18Co-1.5Si after rolling was {100}⟨110⟩ in this study can be thought to support the above presumption.
The textures of the annealed samples are shown in Fig. 4(b). After annealing, the texture of the recrystallized 82Fe-18Co alloy changed to the random orientation. This behavior is different from the reported IF steels, where the {111}⟨110⟩ recrystallized grains, classified into the γ-fiber, generally develop when the reduction ratio is less than 70%.9,13
Harumitsu et al. reported that when the inhomogeneity between the deformation near the grain boundary and the one of the inter-grain becomes significantly large for the pure copper, the recrystallized textures become random.14 The partial replacement of Co to Fe, or the Si/Al addition to the FeCo alloy in this study may cause the same behavior as that reported by the Harumitsu et al.
To quantitatively evaluate the texture of each annealed samples, the A-parameters calculated from the ODF’s are shown in Fig. 5. The 82Fe-18Co without the Si/Al-addition showed the smallest A-parameter. The A-parameters of annealed samples did not change significantly when the Si/Al addition ratio was changed. Kestens and Jacobs investigated the A-parameter of the (100)⟨001⟩ orientation value was be 22.5° for the material with the best soft magnetic properties.4 Since the values of the A-parameters of all samples in this study are higher than 22.5°, the differences in the A-parameters among the alloys in this study seems to be very little substantially.
This result that the random recrystallized textures in this study is unsuitable for obtaining more excellent soft magnetic properties.
The maximum relative permeability μrm and coercivity Hc of the annealed samples are shown in Figs. 6 and 7. It can be seen that the addition of Si/Al to the 82Fe-18Co improved the soft magnetic properties.
Figure 8 shows the magnetic flux density (B30 000) at an applied magnetic field of 30 000 A/m for the annealed samples. The B30 000 for each alloy showed high values ranging from 2.3 T to 2.2 T. The slight decrease in B30 000 of Si/Al-added sample compared with the no-addition one is thought to be due to the partial substitution of non-magnetic elements, Si/Al to the iron site in the BCC crystal.
Figure 9 shows the iron loss Pc with 1.5 T-1 kHz of the annealed samples. The hysteresis loss Ph and the eddy current loss Pe are also shown in Fig. 9. The Pc, Ph, and Pe decreased by the Si/Al addition. The addition of Si seems to be more effective than the Al-addition to reduce the iron loss.
Next, we will discuss on the reason why the Si/Al-addition improved the soft magnetic properties. Two major factors to influence the soft magnetic properties are the microstructure and the intrinsic properties of the material. First, we will see the results for the microstructure again.
Figure 10 shows the relationship between A-parameters and Ph for the heat-treated samples. Here, if the A-parameter is smaller, more crystal grains with an easy axis, ⟨100⟩ orientation along the rolling direction will be included, resulting in the improvement of the soft magnetic properties. However, Fig. 10 shows the opposite behaviour; the Si/Al-added samples with more excellent soft magnetic properties have the larger A-parameter than the sample without Si/Al addition. Therefore, it does not seem that the microstructure of the samples in this study is a dominant reason.
Regarding the other factor, the intrinsic properties, it is known that the Si or Al addition to Fe generally decreases the crystal magnetic anisotropy anisotropy Ku, and reduces the magnetostriction constant λ. It is thought to be reasonable that such change in the coercivity after the Si/Al-addition influenced the soft magnetic properties shown in Figs. 6 and 7. It was reported that the addition of Si to Fe decreases Ku more than the Al-addition.2 The result that the addition of Si to Fe-Co decreased the Hc more than the Al-addition as can be seen from Fig. 6 supports the presumption. On the other hand, the μrm shown in Fig. 7 had a little complicated behaviour. The only Si-added sample did not have the maximum μrm value, but the co-addition of Si/Al showed the maximum μrm. We do not have an enough data to explain this, although we think that the difference of the magnetic domain structure may be influences. We will investigate further to understand these phenomena in the future.
Figure 11 shows the relationship between electrical resistance and Pe of the annealed samples. It is obvious that the addition of Si/Al to the 82Fe-18Co alloy increased the electrical resistance resulting in the decreasing of the Pe.
IV. CONCLUSIONS
The effect of Co content in the Fe-Co alloys on the formation of ordered B2 phase was first studied by a neutron diffraction analysis. The 49Fe-49Co-2V alloy and the 73Fe-27Co alloy clearly had the ordered B2 phase, while the other alloys, the 82Fe-18Co and the 95Fe-5Co did not contain B2, which suggested that the latter two alloys have a good workability.
Then, the effects of the Si/Al addition to the 82Fe-18Co alloy on the workability and the soft magnetic properties were investigated. The DBTT of 82Fe-18Co alloy increased by the Si-addition, and also the Vickers hardness of the annealed samples increased with the Si-added alloys. On the other hand, the soft magnetic properties such as the μrm, the Ph, and the Pe were significantly improved by the addition of Si/Al, while maintaining a high B value at 30 000 A/m above 2.2T.
From these results in this study, the Si/Al-added 82Fe-18Co alloy is a promising soft magnetic material for the high-power density, high-efficiency motors.
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Takamasa. Sato: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal). Hiroyuki. Takabayashi: Conceptualization (equal).
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request.