- Email: [email protected]
Effects of laser irradiation on iron loss reduction for Fe–3%Si grain-oriented silicon steel
Acta Materialia 53 (2005) 939–945 www.actamat-journals.com
Effects of laser irradiation on iron loss reduction for Fe–3%Si grain-oriented silicon steel Muneyuki Imafuku a,*,1, Hiroshi Suzuki b, Koichi Akita c, Keiji Iwata a, Masahiro Fujikura d a Advanced Technology Research Laboratories, Nippon Steel Corporation, Futtsu, Chiba 293-8511, Japan Neutron Science Research Center, Japan Atomic Energy Research Institute, Naka-gun, Ibaraki 319-1195, Japan Department of Mechanical Systems Engineering, Musashi Institute of Technology, Setagaya, Tokyo 158-8557, Japan d Steel Research Laboratories, Nippon Steel Corporation, Futtsu, Chiba 293-8511, Japan b
c
Received 31 August 2004; accepted 27 October 2004 Available online 23 November 2004
Abstract The effects of laser irradiation on iron loss reduction for Fe–3%Si grain-oriented silicon steel sheet were investigated. The local tensile residual stress states near the laser irradiated cavity lines were observed by using the new X-ray stress measurement method for a single crystal. Although the higher laser power induced the larger tensile residual stresses, the minimum iron loss was obtained at the medium tensile residual stress conditions of about 100–200 MPa. The increase of Vickers hardness was observed with increasing laser power, which was the mark of the plastic deformations induced by the laser irradiation. The tensile residual stress reduces eddy current loss and the plastic deformation increases hysteresis loss of the material. The total iron loss is determined by the balance of these two effects of laser irradiation. 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Grain-oriented silicon steel; Laser treatment; Residual stresses; Magnetic domain; Hardness
1. Introduction Fe–3%Si grain-oriented silicon steel, consisting of {1 1 0}Æ0 0 1æ oriented large grains, has been widely used for transformer cores. The reduction of iron loss is one of the most important industrial issues and lower iron loss materials have been developed by improving {1 1 0}Æ0 0 1æ alignment, making thinner-gauge sheet and refining magnetic domain wall spacing [1]. In particular, the magnetic domain-refining techniques, such as laser-irradiation [2], groove-forming [3] and etching [4] *
Corresponding author. Tel.: +81 439 80 2691; fax: +81 439 80 2767. E-mail address: [email protected] (M. Imafuku). 1 Present address: Materials Characterization Center, Nippon Steel Technoresearch Corporation, Futtsu, Chiba 293-0011, Japan.
techniques have been developed over the past 20 years and are known to be very effective for refining magnetic domain wall spacing and hence reducing iron loss. It has been speculated that the induced tensile stresses [5] or recrystallized micro grains [6] might be the origin of the magnetic domain refining. Very recently, the present authors have developed a new X-ray measurement method of a plane stress state for a single crystal [7,8] and showed for the first time that local residual tensile stresses are induced in laser-irradiated Fe–3%Si grainoriented silicon steels [9]. It was supposed that the local residual tensile stresses change the magnetic anisotropy of the material and destabilize the magnetic domains along the rolling direction so as to refine the domains. However, a precise study of the effects of laser irradiation on the reduction of iron loss has not been carried out hitherto.
1359-6454/$30.00 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actamat.2004.10.040
940
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
In this paper, the residual stress distributions of a plane stress state in laser-irradiated Fe–3%Si grain-oriented silicon steel sheet were measured with various laser irradiation conditions by the newly developed X-ray stress measurement method for a single crystal. We investigated the relationship between the iron core loss and the residual stress distributions and the deformations in order to clarify the mechanism of iron loss reduction by laser irradiation on this material. 2. Experimental 2.1. Sample preparation A single crystal grain of 30 · 15 mm2 was prepared by cutting from a Fe–3%Si grain oriented silicon steel sheet of 0.23 mm thick for the X-ray stress measurements. The oxide coating films on the sheet were removed beforehand. The length direction of the specimen (30 mm) was set to be parallel with the rolling direction (RD) of the sheet. Then, the specimen was annealed at 1027 K for 2 h in pure hydrogen in order to remove the effect of cutting. Four dotted cavity lines were formed in air along the transverse direction (TD) by using Nd:YAG laser with the energy levels of 1.6, 3.3, 4.9 and 6.6 mJ/pulse. The pitch of the lines was 6.0 mm. Fig. 1 shows the schematic illustration of the specimen used in this study. Although the grainoriented silicon steel is manufactured to be highly aligned in {1 1 0}Æ0 0 1æ direction, each grain has its individual orientation. The orientation of the specimen determined by Laue method was {50 66 3}Æ1 0 127æ, which is very close to the ideal orientation, {1 1 0}Æ0 0 1æ in this study. Fe–3%Si grain oriented silicon steel sheets of 60 · 300 · 0.23 mm3 were used for the specimens of magnetic measurements. These sheets were also annealed at 1027 K for 2 h in pure hydrogen before the laser-irradiation. Nd:YAG pulsed laser was irradiated on the surface of each sheet with the energy levels of 0.3–9.0 mJ/pulse in 6.0 mm pitch along TD.
Fig. 1. Schematic illustration of laser-irradiated grain-oriented silicon steel used in this study.
2.2. X-ray stress measurement The sin2w method [10] has been commonly used to investigate the stress states in various materials. This method cannot be applied, however, for the stress measurements in coarse grains or single crystals because this method is based on the assumption of randomly oriented elastic polycrystalline material. In principle, we can obtain the stress state in a single crystal, when the absolute lattice displacements of the several directions can be measured precisely. It is difficult, however, to determine the reliable lattice spacing in the stress-free condition. Several methods, such as the pseudo-Ko¨ssel pattern [11] or Debye ring measurements [12], have been examined for this purpose, but have not come into wide use. Recently, Yoshioka et al. [13] proposed a unique and practical X-ray stress measurement method for a single crystal. Very recently, Suzuki et al. [7–9,14–16] have refined this method so as to improve the accuracy of the measurement and data analysis and succeeded in determining the stress states in single crystal Si and Fe– 3%Si grain-oriented silicon steel. The following is the essential points of the stress measurement method for a cubic single crystal applied in this study. Fig. 2 shows a crystal coordinate system Xi, a laboratory coordinate system Li and a specimen coordinate system Pi. / and w are the rotation angle between P1 and L1 and P3 and L3, respectively. A lattice strain of nth plane, eL33ðnÞ in the L3 direction on the laboratory coordinate system in the plane stress condition is expressed with the stress components rS11 ; rS12 and rS22 on the specimen coordinate system by the following equation [14]: eL33ðnÞ ¼ ðhn h0 Þ cot h0 ¼ S 11 S 12 12S 44 ðc231 p211 þ c232 p212 þ c233 p213 ÞrS11 þ 2ðc231 p11 p12 þ c232 p12 p22 þ c233 p13 p23 ÞrS12 þ ðc231 p221 þ c232 p222 þ c233 p223 ÞrS22 þ S 12 ðrS11 þ rS22 Þ þ 12S44 ðrS11 sin2 / rS12 sin 2/ þ rS22 cos2 /Þsin2 w ¼ An rS11 þ Bn rS12 þ C n rS22 ;
ð1Þ
Fig. 2. Relationship between a crystal coordinate system Xi, a laboratory coordinate system Li and a specimen coordinate system Pi.
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
where Sij is the elastic compliance of the cubic crystal. The transformation matrices, p and c are between the specimen and crystal coordinate systems, and crystal and laboratory coordinate systems, respectively. In this equation, h0 is half of the diffraction angle in the stress-free condition, and is usually unknown. An, Bn and Cn are the variables which can be determined by the Miller indices of the measured diffraction planes. hn is expressed as the following equation: S 2r11 2rS12 2rS22 2hn ¼ An þ Bn þ Cn cot h0 cot h0 cot h0 þ 2h0 :
ð2Þ
By choosing the equivalent diffraction planes, rS11 ; rS12 ; rS22 and h0 can be calculated by the multiple regression analysis method. In this analysis, at least four equivalent diffraction planes should be measured. Fig. 3 shows the X-ray stress measurement apparatus for a single crystal used in this study. Newly developed vw-oscillation stage was installed in MSF system (Rigaku Corporation) with one dimensional position sensitive proportional counter (PSPC). The axes of the v- and w-oscillations correspond to the direction of collimated incident X-ray beam and the vertical direction of the v-oscillation axis, respectively. In order to obtain a perfect X-ray diffraction profile of the target diffraction plane with PSPC, coupled two-axis oscillation of the sample is necessary. The measured position of the specimen was adjusted at the rotation center of the stage within ±20 lm error by using a laser displacement meter. A shield-tube X-ray source of Cr Ka radiation with a power of 30 KV and 8 mA was used in this study. The collimated diameter of X-ray beam was about 0.4 mm. Six equivalent 2 1 1 diffraction planes of a-Fe, 2 1 1, 1 1 2, 1 2 1, 1 2 1, 1 1 2, and 2 1 1, satisfied the reflection
941
condition and were chosen for the measurement. The stresses at four positions, 0, 0.25, 0.50 and 1.00 mm from the laser-irradiated cavity line (i.e. along RD direction) were measured. 2.3. Magnetic properties The energy loss of the iron sheet, W (i.e. the iron loss), was measured by the H-coil method. The maximum magnetic-flux density, Bm was 1.7 T and the frequency of AC field, f was 50 Hz in this study. The hysteresis loss, Wh (see in the following section), which is related to the pinning force of the magnetic domain movement was measured by using a Cioffy-type DC magneto meter. Magnetic domain structures at the surface of the sheets were observed by using a colloid magnetic fluid type domain viewer, in which the magnetite particles are uniformly dispersed in a solvent. The colloid particles tend to condense at the magnetic domain boundaries by their leak magnetic field near the surface of the magnetic material sheet, and so we can observe the magnetic domain structures of the sheet.
3. Results and discussion Fig. 4 shows the changes in the iron loss with the energy level of Nd:YAG laser, E from 0.3 to 9.0 mJ/pulse. We can see that the iron loss drastically decreases by laser-irradiation and its minimum value was obtained at around 3.0 mJ/pulse. When the laser energy was increased to more than 3.0 mJ/pulse, the iron loss slightly increased. That is to say, an optimum laser energy condition for the improvement of the iron loss in Fe–3% grain-oriented silicon steel sheet exists at around 3.0 mJ/pulse. A similar result has been already reported by Fujikura et al. [17].
0.9
Iron loss, W/ W/ kg
Bm = 1.7T f = 50Hz 0.8
0.7
0.6
0
1
2
3
4
5
6
7
8
9
10
Laser energy, E/ mJ/ pulse Fig. 3. The X-ray stress measurement apparatus for a single crystal.
Fig. 4. Change in the iron loss, W, of Fe–3%Si grain-oriented silicon steel with the laser energy, E, at Bm = 1.7 T and f = 50 Hz.
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
Average ratio of magenetic domain refinement
942
0.6 0.5 0.4 0.3 0.2 0.1 0.0
1
2
3
4
5
6
7
Laser energy, E/ mJ/ pulse Fig. 6. Relationship between the laser energy and average ratio of magnetic domain refinement, 2L(after laser-irradiation)/2L(before laser-irradiation), where 2L is the magnetic domain wall spacing.
Fig. 5. Magnetic domain structures in Fe–3%Si grain-oriented silicon steel sheets before and after laser irradiation with the energy levels of (a) 1.6, (b) 3.3, (c) 4.9 and (d) 6.6 mJ/pulse.
It is well known that the total energy loss of the magnetic sheet, W, consists of two terms, as shown in the following equation: W ¼ W h þ W e;
ð3Þ
where Wh and We denote the hysteresis loss and the eddy current loss, respectively. The first term Wh is caused by the pinning effect against the movement of magnetic domain walls for AC magnetization process, and can be decreased by purifying the material. The second term We is caused by the Joule heat due to the eddy currents, which can be decreased by the addition of a second element, (for example, adding silicon in iron sheet) or using thinner-gauge material, and the energy loss by the movement of magnetic domain walls. Based on the analysis of Pry and Bean [18], when the magnetic domain wall spacing, 2L, is sufficiently larger than the thickness of the magnetic sheet, d (i.e. under the condition of 2L d), We can be approximately expressed as follows: 2L W e ¼ 1:628W c ; ð4Þ d where Wc denotes the classical loss value. Thus, We decreases when 2L becomes small (i.e. the magnetic domains are refined). That is the essential theory of
magnetic domain refining process for a magnetic material sheet. Fig. 5 shows the observed magnetic domain structures in the specimens before and after the laser irradiation. Four dotted cavity lines were formed in each sample. Slab type 180 domains were observed along RD of Æ0 0 1æ, which is the easy magnetization axis of bcc iron. It is obvious that the magnetic domains between the lines were remarkably refined when the laser energy increased. The relationship between the laser energy and the average ratio of magnetic domain wall spacing before and after the laser-irradiation is summarized in Fig. 6. The magnetic domain spacing becomes drastically narrower by increasing the laser energy up to E = 4.9 mJ/pulse and then saturates with higher laser energy. According to Eq. (4), we would expect that the total iron loss W decreases up to E = 4.9 mJ/pulse by decreasing We. Instead, the lowest iron loss value was actually obtained for the condition of E = 3.3 mJ/pulse, at which the magnetic domains were not fully refined. This means that the improvement of iron loss cannot be simply explained by the magnetic domain refining effect described above. In our previous papers [9,15,16], we showed that the local tensile residual stresses are induced at the surface of Fe–3%Si grain-oriented silicon steel sheet by the laser-irradiation and this state causes the refinement of 180 magnetic domains. In this study, we investigated the relationship between the residual stress distribution and the laser energy. Fig. 7 shows the laser energy dependence of residual stress distributions in Fe–3%Si grain-oriented silicon steel sheet as a function of distance from the laser cavity lines along RD, at (a) X = 0.0 mm, (b) X = 0.25 mm, (c) X = 0.5 mm and (d) X = 1.0 mm. In this figure, rx, ry and rxy represent the plane stresses in the rolling and transverse directions and the shear stress between these directions. At the po-
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
nomena in stainless steel by some researchers [20–23]. Mukai et al. [20] have reported that the tensile residual stress state was formed when the sample was peened in air, whereas the compressive residual stress state was formed when the sample was peened in water. They have mentioned that the dominant factors for the creation of the residual stress are the thermal effect in the former case and the shock wave effect in the latter. Since the tensile residual stress states were confirmed in all our laser irradiation conditions, the thermal history effect of heating and cooling should be the origin of the local tensile stresses and consequently the magnetic domain refinement in the laser-irradiated Fe–3% grain-oriented silicon steel. Considering the results of laser energy dependences of magnetic domain-refining (see Fig. 6) and the residual stresses (see Fig. 7(a)), it is clear that the larger the tensile stress, the finer the magnetic domain wall spacing. This means that the laser-induced tensile residual stresses cause the magnetic domain refining of this material sheet. This is also confirmed by the stress relief annealing experiment [9,15,16], in which the domain-refining effect vanished when the residual tensile stresses were released by annealing. However, the minimum value of W was not obtained under the largest tensile stresses and
sition of X = 0 mm, the residual stresses of rx and ry were proportional to the laser energy from 1.6 to 4.9 mJ/pulse and slightly decreased at 6.6 mJ/pulse. Since the yield stress of this material is about 330–360 MPa [19], the specimen might yield at the laser energy of 6.6 mJ/pulse and thus, the tensile residual stresses were released. Besides, the values of ry were larger than those of rx in all conditions (for example, ry/rx = 1.7 at E = 4.9 mJ/pulse), suggesting the neighboring cavities along TD tensed each other so as to increase the tensile stress of ry between the cavities. rx and ry are considered to be the principle stresses in the surface plane at X = 0.0 mm because the shear stress, rxy, is negligible at X = 0.0 mm. At the position of X = 0.25 mm, the gradient of tensile stresses was so sharp in the measured area that the reliable limit of the residual stresses becomes relatively large. The reason why the shear stresses were also observed at X = 0.25 mm is not yet clear. The stress-induced area was limited to less than 1.0 mm as is seen in this figure. At the position of X = 0.5 mm, a weak tensile stress state was observed for E = 6.6 mJ/ pulse, while stress-free state was observed from E = 1.6–4.9 mJ/pulse. Recently, the laser peening process has been studied for the improvement of stress corrosion cracking phe-
(c) 300
X = 0 mm σy σx σxy
200
Residual stress, σ /MPa
Residual stress, σ /MPa
(a) 300
100
0
–100
X = 0 .5 mm σy σx σxy
200
100
0
–100 1
2
3 4 5 6 Laser energy, E/ mJ/ pulse
1
7
2
3 4 5 Laser energy, E/ mJ/ pulse
6
7
(d) 300
(b) 300 X = 0.25 mm
X = 1.0 mm
σy σx σxy
200
Residual stress, σ /MPa
Residual stress, σ /MPa
943
100
0
–100
σy σx σxy
200
100
0
–100 1
2
3 4 5 Laser energy, E/ mJ/ pulse
6
7
1
2
3 4 5 Laser energy, E/ mJ/ pulse
6
7
Fig. 7. Laser energy dependence of residual stress distributions in Fe–3%Si grain-oriented silicon steel at the distance from the laser cavity lines along the rolling direction: (a) X = 0.0 mm; (b) X = 0.25 mm; (c) X = 0.5 mm; and (d) X = 1.0 mm. rx, ry and rxy represent the plane stresses in rolling direction and transverse direction and the shear stress between X and Y directions.
944
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
served at above E = 4 mJ/pulse, as is shown in Fig. 9. The increase of laser energy causes the increase of some kinds of plastic deformation, which should act as the pinning sites against the movement of magnetic domain walls, and the increase in Wh. This is why the lowest iron loss could not be obtained when the magnetic domains were fully refined above E = 4.9 mJ/pulse.
240 6.6mJ/pulse 4.9mJ/pulse 3.3mJ/pulse 1.6mJ/pulse
Hardness / Hv
220
200
4. Conclusions 180 0.0
0.2
0.4
0.6
0.8
1.0
Position, X /mm Fig. 8. Vickers hardness distributions around laser-irradiated cavities with the laser energy levels of 1.6, 3.3, 4.9 and 6.6 mJ/pulse.
with at the same time the finest magnetic domain wall spacing conditions. Another factor should be considered in order to explain the reduction of the iron loss in Fe– 3%Si silicon steel by the laser irradiation. Next, we investigated the plastic deformation effect by laser irradiation. It is known, in general, that work hardening occurs due to induced plastic strains for a metal. The amount of plastic deformation might be larger with increasing laser energy since the cavity traces become larger with increasing the laser energy. In order to estimate this effect by laser irradiation, we performed micro Vickers hardness testing experiments. Fig. 8 shows the results of Vickers hardness distributions around laser-irradiated cavities with the laser energies of 1.6, 3.3, 4.9 and 6.6 mJ/pulse. We can see from this figure that the hardness increases with increasing laser energy around the cavities up to about X = 0.3–0.6 mm. In particular, a marked increase of the hardness was observed at E = 6.6 mJ/pulse. This means that the amount of plastic deformation increases with increasing laser energy. The notable increase of Wh was also ob-
Acknowledgements The authors would like to thank M. Hashimoto and T. Kubota in Nippon Steel Corporation for supporting this work and stimulating discussions.
0.275
Hysteresis loss, Wh / W/kg
The effects of laser-irradiation on the iron loss reduction were investigated with respect to magnetic domain refining in Fe–3%Si grain-oriented silicon steel. The residual stress distributions near the laser irradiated cavity line were measured by the new X-ray stress measurement method for a single crystal. It was confirmed that the magnetic domain-refining phenomenon occurs by laser-induced tensile residual stresses. The larger residual tensile stresses were observed with higher laser energy conditions and then the magnetic domains were refined. Simultaneously, the increase of hardness, which is closely related to the plastic deformation of the material, becomes also distinguished with increasing laser energy. The laser irradiation has two effects on Fe–3%Si grain-oriented silicon steel in relation to the reduction of iron loss: one is the tensile residual stress and the other is plastic deformation at the surface of the material. These effects act in the opposite way in terms of iron loss. The former decreases We and the latter increases Wh in Eq. (3). The total iron loss W is determined by the balance of these two effects of laser irradiation, of which E = 3.3 mJ/pulse is the optimum condition in Fe–3%Si grain-oriented silicon steel.
References 0.250
0.225
0
2
4 6 8 Laser energy, E/ mJ/ pulse
10
Fig. 9. Change in the hysteresis loss, Wh, of Fe–3%Si grain-oriented silicon steel with the laser energy, E.
[1] Kubota T, Fujikura M, Ushigami Y. J Magn Magn Mater 2000;215–216:69. [2] Iuchi T, Yamaguchi S, Ichiyama T, Nakamura M, Ishimoto T, Kuroki K. J Appl Phys 1982;53:2410. [3] Kobayashi H, Kuroki K, Sasaki E, Iwasaki M, Takahashi N. Phys Scripta 1988;T24:36. [4] Sato K, Honda A, Nakano K, Ishida M, Fukada B, Kan T. J Appl Phys 1993;73:6609. [5] Nakamura M, Hirose K, Nozawa T, Matsuo M. IEEE Trans Magn 1987;MAG-23:3074. [6] Nozawa T, Matsuo Y, Kobayashi H, Iwayama K, Takahashi N. J Appl Phys 1988;63:2966.
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945 [7] Suzuki H, Akita K, Misawa H. Mater Sci Res Int 2000;6:255. [8] Akita K, Suzuki H, Yoshioka Y. Mater Sci Res Int 2001;STP-1:382. [9] Suzuki H, Akita K, Misawa H, Imafuku M. Mater Sci Res Int 2002;8:207. [10] see for example Noyan IC, Cohen JB. Residual stress – measurement by diffraction and interpretation. New York: Springer-Verlag; 1987. p. 117. [11] Imura T, Weismann S, Slade Jr JJ. J Acta Cryst 1962;15:68. [12] Onaga T, Honda K. J Jpn Soc Mech Eng 1970;73:972. [in Japanese]. [13] Yoshioka Y, Ohya S, Suyama Y. In: Proceedings of the 5th international conference on residual stresses, Sweden; 1998. p. 528. [14] Suzuki H, Akita K, Misawa H. Jpn J Appl Phys 2003;42:2876. [15] Suzuki H, Akita K, Misawa H, Imafuku M. In: Proceedings of the international conference on Advance Technology in Experimental Mechanics 2003 (ATEMÕ03), Nagoya, OS04W0201; 2003 [in CD-ROM].
945
[16] Imafuku M, Suzuki T, Suzuki H, Akita K. Adv X-ray Anal 2004;47 [accepted for publication]. [17] Fujikura M, Arai S, Mogi H, Yabumoto M, Kubota T. In: Proceedings of the 7th international workshop on 1- and 2dimensional magnetic measurement and testing, PTB-bericht, PTB-E-81; 2003. p. 23. [18] Pry RH, Bean CP. J Appl Phys 1958;29:532. [19] Private data. [20] Mukai N, Aoki N, Obata M, Ito A, Sano Y, Konagai C. In: Proceedings of the 3rd international conference on Nuclear Engineering (ICONE-3); 1995. p. 1489. [21] Sano Y, Mukai N, Okazaki K, Obata M. Nucl Inst Met Phys Res (Japan) 1997;B212:432. [22] Peyre P, Berthe L, Scherpereel X, Fabbro R. J Mater Sci 1998;33:1421. [23] Peyre P, Scherpereel X, Berthe L, Carboni C, Fabbro R, Beranger G, Lemaitre C. Mater Sci Eng A 2000;280:294.
Effects of laser irradiation on iron loss reduction for Fe–3%Si grain-oriented silicon steel Muneyuki Imafuku a,*,1, Hiroshi Suzuki b, Koichi Akita c, Keiji Iwata a, Masahiro Fujikura d a Advanced Technology Research Laboratories, Nippon Steel Corporation, Futtsu, Chiba 293-8511, Japan Neutron Science Research Center, Japan Atomic Energy Research Institute, Naka-gun, Ibaraki 319-1195, Japan Department of Mechanical Systems Engineering, Musashi Institute of Technology, Setagaya, Tokyo 158-8557, Japan d Steel Research Laboratories, Nippon Steel Corporation, Futtsu, Chiba 293-8511, Japan b
c
Received 31 August 2004; accepted 27 October 2004 Available online 23 November 2004
Abstract The effects of laser irradiation on iron loss reduction for Fe–3%Si grain-oriented silicon steel sheet were investigated. The local tensile residual stress states near the laser irradiated cavity lines were observed by using the new X-ray stress measurement method for a single crystal. Although the higher laser power induced the larger tensile residual stresses, the minimum iron loss was obtained at the medium tensile residual stress conditions of about 100–200 MPa. The increase of Vickers hardness was observed with increasing laser power, which was the mark of the plastic deformations induced by the laser irradiation. The tensile residual stress reduces eddy current loss and the plastic deformation increases hysteresis loss of the material. The total iron loss is determined by the balance of these two effects of laser irradiation. 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. Keywords: Grain-oriented silicon steel; Laser treatment; Residual stresses; Magnetic domain; Hardness
1. Introduction Fe–3%Si grain-oriented silicon steel, consisting of {1 1 0}Æ0 0 1æ oriented large grains, has been widely used for transformer cores. The reduction of iron loss is one of the most important industrial issues and lower iron loss materials have been developed by improving {1 1 0}Æ0 0 1æ alignment, making thinner-gauge sheet and refining magnetic domain wall spacing [1]. In particular, the magnetic domain-refining techniques, such as laser-irradiation [2], groove-forming [3] and etching [4] *
Corresponding author. Tel.: +81 439 80 2691; fax: +81 439 80 2767. E-mail address: [email protected] (M. Imafuku). 1 Present address: Materials Characterization Center, Nippon Steel Technoresearch Corporation, Futtsu, Chiba 293-0011, Japan.
techniques have been developed over the past 20 years and are known to be very effective for refining magnetic domain wall spacing and hence reducing iron loss. It has been speculated that the induced tensile stresses [5] or recrystallized micro grains [6] might be the origin of the magnetic domain refining. Very recently, the present authors have developed a new X-ray measurement method of a plane stress state for a single crystal [7,8] and showed for the first time that local residual tensile stresses are induced in laser-irradiated Fe–3%Si grainoriented silicon steels [9]. It was supposed that the local residual tensile stresses change the magnetic anisotropy of the material and destabilize the magnetic domains along the rolling direction so as to refine the domains. However, a precise study of the effects of laser irradiation on the reduction of iron loss has not been carried out hitherto.
1359-6454/$30.00 2004 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.actamat.2004.10.040
940
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
In this paper, the residual stress distributions of a plane stress state in laser-irradiated Fe–3%Si grain-oriented silicon steel sheet were measured with various laser irradiation conditions by the newly developed X-ray stress measurement method for a single crystal. We investigated the relationship between the iron core loss and the residual stress distributions and the deformations in order to clarify the mechanism of iron loss reduction by laser irradiation on this material. 2. Experimental 2.1. Sample preparation A single crystal grain of 30 · 15 mm2 was prepared by cutting from a Fe–3%Si grain oriented silicon steel sheet of 0.23 mm thick for the X-ray stress measurements. The oxide coating films on the sheet were removed beforehand. The length direction of the specimen (30 mm) was set to be parallel with the rolling direction (RD) of the sheet. Then, the specimen was annealed at 1027 K for 2 h in pure hydrogen in order to remove the effect of cutting. Four dotted cavity lines were formed in air along the transverse direction (TD) by using Nd:YAG laser with the energy levels of 1.6, 3.3, 4.9 and 6.6 mJ/pulse. The pitch of the lines was 6.0 mm. Fig. 1 shows the schematic illustration of the specimen used in this study. Although the grainoriented silicon steel is manufactured to be highly aligned in {1 1 0}Æ0 0 1æ direction, each grain has its individual orientation. The orientation of the specimen determined by Laue method was {50 66 3}Æ1 0 127æ, which is very close to the ideal orientation, {1 1 0}Æ0 0 1æ in this study. Fe–3%Si grain oriented silicon steel sheets of 60 · 300 · 0.23 mm3 were used for the specimens of magnetic measurements. These sheets were also annealed at 1027 K for 2 h in pure hydrogen before the laser-irradiation. Nd:YAG pulsed laser was irradiated on the surface of each sheet with the energy levels of 0.3–9.0 mJ/pulse in 6.0 mm pitch along TD.
Fig. 1. Schematic illustration of laser-irradiated grain-oriented silicon steel used in this study.
2.2. X-ray stress measurement The sin2w method [10] has been commonly used to investigate the stress states in various materials. This method cannot be applied, however, for the stress measurements in coarse grains or single crystals because this method is based on the assumption of randomly oriented elastic polycrystalline material. In principle, we can obtain the stress state in a single crystal, when the absolute lattice displacements of the several directions can be measured precisely. It is difficult, however, to determine the reliable lattice spacing in the stress-free condition. Several methods, such as the pseudo-Ko¨ssel pattern [11] or Debye ring measurements [12], have been examined for this purpose, but have not come into wide use. Recently, Yoshioka et al. [13] proposed a unique and practical X-ray stress measurement method for a single crystal. Very recently, Suzuki et al. [7–9,14–16] have refined this method so as to improve the accuracy of the measurement and data analysis and succeeded in determining the stress states in single crystal Si and Fe– 3%Si grain-oriented silicon steel. The following is the essential points of the stress measurement method for a cubic single crystal applied in this study. Fig. 2 shows a crystal coordinate system Xi, a laboratory coordinate system Li and a specimen coordinate system Pi. / and w are the rotation angle between P1 and L1 and P3 and L3, respectively. A lattice strain of nth plane, eL33ðnÞ in the L3 direction on the laboratory coordinate system in the plane stress condition is expressed with the stress components rS11 ; rS12 and rS22 on the specimen coordinate system by the following equation [14]: eL33ðnÞ ¼ ðhn h0 Þ cot h0 ¼ S 11 S 12 12S 44 ðc231 p211 þ c232 p212 þ c233 p213 ÞrS11 þ 2ðc231 p11 p12 þ c232 p12 p22 þ c233 p13 p23 ÞrS12 þ ðc231 p221 þ c232 p222 þ c233 p223 ÞrS22 þ S 12 ðrS11 þ rS22 Þ þ 12S44 ðrS11 sin2 / rS12 sin 2/ þ rS22 cos2 /Þsin2 w ¼ An rS11 þ Bn rS12 þ C n rS22 ;
ð1Þ
Fig. 2. Relationship between a crystal coordinate system Xi, a laboratory coordinate system Li and a specimen coordinate system Pi.
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
where Sij is the elastic compliance of the cubic crystal. The transformation matrices, p and c are between the specimen and crystal coordinate systems, and crystal and laboratory coordinate systems, respectively. In this equation, h0 is half of the diffraction angle in the stress-free condition, and is usually unknown. An, Bn and Cn are the variables which can be determined by the Miller indices of the measured diffraction planes. hn is expressed as the following equation: S 2r11 2rS12 2rS22 2hn ¼ An þ Bn þ Cn cot h0 cot h0 cot h0 þ 2h0 :
ð2Þ
By choosing the equivalent diffraction planes, rS11 ; rS12 ; rS22 and h0 can be calculated by the multiple regression analysis method. In this analysis, at least four equivalent diffraction planes should be measured. Fig. 3 shows the X-ray stress measurement apparatus for a single crystal used in this study. Newly developed vw-oscillation stage was installed in MSF system (Rigaku Corporation) with one dimensional position sensitive proportional counter (PSPC). The axes of the v- and w-oscillations correspond to the direction of collimated incident X-ray beam and the vertical direction of the v-oscillation axis, respectively. In order to obtain a perfect X-ray diffraction profile of the target diffraction plane with PSPC, coupled two-axis oscillation of the sample is necessary. The measured position of the specimen was adjusted at the rotation center of the stage within ±20 lm error by using a laser displacement meter. A shield-tube X-ray source of Cr Ka radiation with a power of 30 KV and 8 mA was used in this study. The collimated diameter of X-ray beam was about 0.4 mm. Six equivalent 2 1 1 diffraction planes of a-Fe, 2 1 1, 1 1 2, 1 2 1, 1 2 1, 1 1 2, and 2 1 1, satisfied the reflection
941
condition and were chosen for the measurement. The stresses at four positions, 0, 0.25, 0.50 and 1.00 mm from the laser-irradiated cavity line (i.e. along RD direction) were measured. 2.3. Magnetic properties The energy loss of the iron sheet, W (i.e. the iron loss), was measured by the H-coil method. The maximum magnetic-flux density, Bm was 1.7 T and the frequency of AC field, f was 50 Hz in this study. The hysteresis loss, Wh (see in the following section), which is related to the pinning force of the magnetic domain movement was measured by using a Cioffy-type DC magneto meter. Magnetic domain structures at the surface of the sheets were observed by using a colloid magnetic fluid type domain viewer, in which the magnetite particles are uniformly dispersed in a solvent. The colloid particles tend to condense at the magnetic domain boundaries by their leak magnetic field near the surface of the magnetic material sheet, and so we can observe the magnetic domain structures of the sheet.
3. Results and discussion Fig. 4 shows the changes in the iron loss with the energy level of Nd:YAG laser, E from 0.3 to 9.0 mJ/pulse. We can see that the iron loss drastically decreases by laser-irradiation and its minimum value was obtained at around 3.0 mJ/pulse. When the laser energy was increased to more than 3.0 mJ/pulse, the iron loss slightly increased. That is to say, an optimum laser energy condition for the improvement of the iron loss in Fe–3% grain-oriented silicon steel sheet exists at around 3.0 mJ/pulse. A similar result has been already reported by Fujikura et al. [17].
0.9
Iron loss, W/ W/ kg
Bm = 1.7T f = 50Hz 0.8
0.7
0.6
0
1
2
3
4
5
6
7
8
9
10
Laser energy, E/ mJ/ pulse Fig. 3. The X-ray stress measurement apparatus for a single crystal.
Fig. 4. Change in the iron loss, W, of Fe–3%Si grain-oriented silicon steel with the laser energy, E, at Bm = 1.7 T and f = 50 Hz.
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
Average ratio of magenetic domain refinement
942
0.6 0.5 0.4 0.3 0.2 0.1 0.0
1
2
3
4
5
6
7
Laser energy, E/ mJ/ pulse Fig. 6. Relationship between the laser energy and average ratio of magnetic domain refinement, 2L(after laser-irradiation)/2L(before laser-irradiation), where 2L is the magnetic domain wall spacing.
Fig. 5. Magnetic domain structures in Fe–3%Si grain-oriented silicon steel sheets before and after laser irradiation with the energy levels of (a) 1.6, (b) 3.3, (c) 4.9 and (d) 6.6 mJ/pulse.
It is well known that the total energy loss of the magnetic sheet, W, consists of two terms, as shown in the following equation: W ¼ W h þ W e;
ð3Þ
where Wh and We denote the hysteresis loss and the eddy current loss, respectively. The first term Wh is caused by the pinning effect against the movement of magnetic domain walls for AC magnetization process, and can be decreased by purifying the material. The second term We is caused by the Joule heat due to the eddy currents, which can be decreased by the addition of a second element, (for example, adding silicon in iron sheet) or using thinner-gauge material, and the energy loss by the movement of magnetic domain walls. Based on the analysis of Pry and Bean [18], when the magnetic domain wall spacing, 2L, is sufficiently larger than the thickness of the magnetic sheet, d (i.e. under the condition of 2L d), We can be approximately expressed as follows: 2L W e ¼ 1:628W c ; ð4Þ d where Wc denotes the classical loss value. Thus, We decreases when 2L becomes small (i.e. the magnetic domains are refined). That is the essential theory of
magnetic domain refining process for a magnetic material sheet. Fig. 5 shows the observed magnetic domain structures in the specimens before and after the laser irradiation. Four dotted cavity lines were formed in each sample. Slab type 180 domains were observed along RD of Æ0 0 1æ, which is the easy magnetization axis of bcc iron. It is obvious that the magnetic domains between the lines were remarkably refined when the laser energy increased. The relationship between the laser energy and the average ratio of magnetic domain wall spacing before and after the laser-irradiation is summarized in Fig. 6. The magnetic domain spacing becomes drastically narrower by increasing the laser energy up to E = 4.9 mJ/pulse and then saturates with higher laser energy. According to Eq. (4), we would expect that the total iron loss W decreases up to E = 4.9 mJ/pulse by decreasing We. Instead, the lowest iron loss value was actually obtained for the condition of E = 3.3 mJ/pulse, at which the magnetic domains were not fully refined. This means that the improvement of iron loss cannot be simply explained by the magnetic domain refining effect described above. In our previous papers [9,15,16], we showed that the local tensile residual stresses are induced at the surface of Fe–3%Si grain-oriented silicon steel sheet by the laser-irradiation and this state causes the refinement of 180 magnetic domains. In this study, we investigated the relationship between the residual stress distribution and the laser energy. Fig. 7 shows the laser energy dependence of residual stress distributions in Fe–3%Si grain-oriented silicon steel sheet as a function of distance from the laser cavity lines along RD, at (a) X = 0.0 mm, (b) X = 0.25 mm, (c) X = 0.5 mm and (d) X = 1.0 mm. In this figure, rx, ry and rxy represent the plane stresses in the rolling and transverse directions and the shear stress between these directions. At the po-
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
nomena in stainless steel by some researchers [20–23]. Mukai et al. [20] have reported that the tensile residual stress state was formed when the sample was peened in air, whereas the compressive residual stress state was formed when the sample was peened in water. They have mentioned that the dominant factors for the creation of the residual stress are the thermal effect in the former case and the shock wave effect in the latter. Since the tensile residual stress states were confirmed in all our laser irradiation conditions, the thermal history effect of heating and cooling should be the origin of the local tensile stresses and consequently the magnetic domain refinement in the laser-irradiated Fe–3% grain-oriented silicon steel. Considering the results of laser energy dependences of magnetic domain-refining (see Fig. 6) and the residual stresses (see Fig. 7(a)), it is clear that the larger the tensile stress, the finer the magnetic domain wall spacing. This means that the laser-induced tensile residual stresses cause the magnetic domain refining of this material sheet. This is also confirmed by the stress relief annealing experiment [9,15,16], in which the domain-refining effect vanished when the residual tensile stresses were released by annealing. However, the minimum value of W was not obtained under the largest tensile stresses and
sition of X = 0 mm, the residual stresses of rx and ry were proportional to the laser energy from 1.6 to 4.9 mJ/pulse and slightly decreased at 6.6 mJ/pulse. Since the yield stress of this material is about 330–360 MPa [19], the specimen might yield at the laser energy of 6.6 mJ/pulse and thus, the tensile residual stresses were released. Besides, the values of ry were larger than those of rx in all conditions (for example, ry/rx = 1.7 at E = 4.9 mJ/pulse), suggesting the neighboring cavities along TD tensed each other so as to increase the tensile stress of ry between the cavities. rx and ry are considered to be the principle stresses in the surface plane at X = 0.0 mm because the shear stress, rxy, is negligible at X = 0.0 mm. At the position of X = 0.25 mm, the gradient of tensile stresses was so sharp in the measured area that the reliable limit of the residual stresses becomes relatively large. The reason why the shear stresses were also observed at X = 0.25 mm is not yet clear. The stress-induced area was limited to less than 1.0 mm as is seen in this figure. At the position of X = 0.5 mm, a weak tensile stress state was observed for E = 6.6 mJ/ pulse, while stress-free state was observed from E = 1.6–4.9 mJ/pulse. Recently, the laser peening process has been studied for the improvement of stress corrosion cracking phe-
(c) 300
X = 0 mm σy σx σxy
200
Residual stress, σ /MPa
Residual stress, σ /MPa
(a) 300
100
0
–100
X = 0 .5 mm σy σx σxy
200
100
0
–100 1
2
3 4 5 6 Laser energy, E/ mJ/ pulse
1
7
2
3 4 5 Laser energy, E/ mJ/ pulse
6
7
(d) 300
(b) 300 X = 0.25 mm
X = 1.0 mm
σy σx σxy
200
Residual stress, σ /MPa
Residual stress, σ /MPa
943
100
0
–100
σy σx σxy
200
100
0
–100 1
2
3 4 5 Laser energy, E/ mJ/ pulse
6
7
1
2
3 4 5 Laser energy, E/ mJ/ pulse
6
7
Fig. 7. Laser energy dependence of residual stress distributions in Fe–3%Si grain-oriented silicon steel at the distance from the laser cavity lines along the rolling direction: (a) X = 0.0 mm; (b) X = 0.25 mm; (c) X = 0.5 mm; and (d) X = 1.0 mm. rx, ry and rxy represent the plane stresses in rolling direction and transverse direction and the shear stress between X and Y directions.
944
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945
served at above E = 4 mJ/pulse, as is shown in Fig. 9. The increase of laser energy causes the increase of some kinds of plastic deformation, which should act as the pinning sites against the movement of magnetic domain walls, and the increase in Wh. This is why the lowest iron loss could not be obtained when the magnetic domains were fully refined above E = 4.9 mJ/pulse.
240 6.6mJ/pulse 4.9mJ/pulse 3.3mJ/pulse 1.6mJ/pulse
Hardness / Hv
220
200
4. Conclusions 180 0.0
0.2
0.4
0.6
0.8
1.0
Position, X /mm Fig. 8. Vickers hardness distributions around laser-irradiated cavities with the laser energy levels of 1.6, 3.3, 4.9 and 6.6 mJ/pulse.
with at the same time the finest magnetic domain wall spacing conditions. Another factor should be considered in order to explain the reduction of the iron loss in Fe– 3%Si silicon steel by the laser irradiation. Next, we investigated the plastic deformation effect by laser irradiation. It is known, in general, that work hardening occurs due to induced plastic strains for a metal. The amount of plastic deformation might be larger with increasing laser energy since the cavity traces become larger with increasing the laser energy. In order to estimate this effect by laser irradiation, we performed micro Vickers hardness testing experiments. Fig. 8 shows the results of Vickers hardness distributions around laser-irradiated cavities with the laser energies of 1.6, 3.3, 4.9 and 6.6 mJ/pulse. We can see from this figure that the hardness increases with increasing laser energy around the cavities up to about X = 0.3–0.6 mm. In particular, a marked increase of the hardness was observed at E = 6.6 mJ/pulse. This means that the amount of plastic deformation increases with increasing laser energy. The notable increase of Wh was also ob-
Acknowledgements The authors would like to thank M. Hashimoto and T. Kubota in Nippon Steel Corporation for supporting this work and stimulating discussions.
0.275
Hysteresis loss, Wh / W/kg
The effects of laser-irradiation on the iron loss reduction were investigated with respect to magnetic domain refining in Fe–3%Si grain-oriented silicon steel. The residual stress distributions near the laser irradiated cavity line were measured by the new X-ray stress measurement method for a single crystal. It was confirmed that the magnetic domain-refining phenomenon occurs by laser-induced tensile residual stresses. The larger residual tensile stresses were observed with higher laser energy conditions and then the magnetic domains were refined. Simultaneously, the increase of hardness, which is closely related to the plastic deformation of the material, becomes also distinguished with increasing laser energy. The laser irradiation has two effects on Fe–3%Si grain-oriented silicon steel in relation to the reduction of iron loss: one is the tensile residual stress and the other is plastic deformation at the surface of the material. These effects act in the opposite way in terms of iron loss. The former decreases We and the latter increases Wh in Eq. (3). The total iron loss W is determined by the balance of these two effects of laser irradiation, of which E = 3.3 mJ/pulse is the optimum condition in Fe–3%Si grain-oriented silicon steel.
References 0.250
0.225
0
2
4 6 8 Laser energy, E/ mJ/ pulse
10
Fig. 9. Change in the hysteresis loss, Wh, of Fe–3%Si grain-oriented silicon steel with the laser energy, E.
[1] Kubota T, Fujikura M, Ushigami Y. J Magn Magn Mater 2000;215–216:69. [2] Iuchi T, Yamaguchi S, Ichiyama T, Nakamura M, Ishimoto T, Kuroki K. J Appl Phys 1982;53:2410. [3] Kobayashi H, Kuroki K, Sasaki E, Iwasaki M, Takahashi N. Phys Scripta 1988;T24:36. [4] Sato K, Honda A, Nakano K, Ishida M, Fukada B, Kan T. J Appl Phys 1993;73:6609. [5] Nakamura M, Hirose K, Nozawa T, Matsuo M. IEEE Trans Magn 1987;MAG-23:3074. [6] Nozawa T, Matsuo Y, Kobayashi H, Iwayama K, Takahashi N. J Appl Phys 1988;63:2966.
M. Imafuku et al. / Acta Materialia 53 (2005) 939–945 [7] Suzuki H, Akita K, Misawa H. Mater Sci Res Int 2000;6:255. [8] Akita K, Suzuki H, Yoshioka Y. Mater Sci Res Int 2001;STP-1:382. [9] Suzuki H, Akita K, Misawa H, Imafuku M. Mater Sci Res Int 2002;8:207. [10] see for example Noyan IC, Cohen JB. Residual stress – measurement by diffraction and interpretation. New York: Springer-Verlag; 1987. p. 117. [11] Imura T, Weismann S, Slade Jr JJ. J Acta Cryst 1962;15:68. [12] Onaga T, Honda K. J Jpn Soc Mech Eng 1970;73:972. [in Japanese]. [13] Yoshioka Y, Ohya S, Suyama Y. In: Proceedings of the 5th international conference on residual stresses, Sweden; 1998. p. 528. [14] Suzuki H, Akita K, Misawa H. Jpn J Appl Phys 2003;42:2876. [15] Suzuki H, Akita K, Misawa H, Imafuku M. In: Proceedings of the international conference on Advance Technology in Experimental Mechanics 2003 (ATEMÕ03), Nagoya, OS04W0201; 2003 [in CD-ROM].
945
[16] Imafuku M, Suzuki T, Suzuki H, Akita K. Adv X-ray Anal 2004;47 [accepted for publication]. [17] Fujikura M, Arai S, Mogi H, Yabumoto M, Kubota T. In: Proceedings of the 7th international workshop on 1- and 2dimensional magnetic measurement and testing, PTB-bericht, PTB-E-81; 2003. p. 23. [18] Pry RH, Bean CP. J Appl Phys 1958;29:532. [19] Private data. [20] Mukai N, Aoki N, Obata M, Ito A, Sano Y, Konagai C. In: Proceedings of the 3rd international conference on Nuclear Engineering (ICONE-3); 1995. p. 1489. [21] Sano Y, Mukai N, Okazaki K, Obata M. Nucl Inst Met Phys Res (Japan) 1997;B212:432. [22] Peyre P, Berthe L, Scherpereel X, Fabbro R. J Mater Sci 1998;33:1421. [23] Peyre P, Scherpereel X, Berthe L, Carboni C, Fabbro R, Beranger G, Lemaitre C. Mater Sci Eng A 2000;280:294.