- Email: [email protected]
Effect of primary recrystallization microstructure on abnormal growth of Goss grains in a twin-roll cast grain-oriented electrical steel
Materials & Design 131 (2017) 167–176
Contents lists available at ScienceDirect
Materials & Design journal homepage: www.elsevier.com/locate/matdes
Effect of primary recrystallization microstructure on abnormal growth of Goss grains in a twin-roll cast grain-oriented electrical steel
MARK
Hong-Yu Songa, Hai-Tao Liua,⁎, John J. Jonasb, Guo-Dong Wanga a b
State Key Laboratory of Rolling and Automation, Northeastern University, Shenyang 110819, PR China Materials Engineering Department, McGill University, 3610 University Street, Montreal H3A 0C5, Canada
A R T I C L E I N F O
A B S T R A C T
Keywords: Goss grains Twin-roll casting Grain-oriented electrical steel Cold rolling Abnormal grain growth
Various primary recrystallization microstructures and textures were produced in a twin-roll cast grain-oriented electrical steel by employing different routes. The relationship between the primary recrystallization microstructure and texture and the abnormal growth of secondary Goss grains was investigated. The results show that the cold rolling reductions have a significant influence on secondary recrystallization by changing the primary recrystallization microstructure. Sheet processed by single-stage cold rolling with 88.3% reduction displayed the poorest secondary recrystallization microstructure as it contained many small equiaxed grains. By contrast, the employment of two-stage cold rolling markedly improved the secondary recrystallization microstructure. In the case of two moderate reductions of 65.2% and 66.3%, dense deformation substructures formed during both the first and second cold rolling, leading to a homogeneous primary recrystallization microstructure together with a strong γ-fiber texture. In this way, a suitable secondary recrystallization microstructure consisting of large Goss grains was produced. In the case of the inappropriate reductions, many large λ- and α-grains in the primary recrystallization matrix blocked the growth of secondary Goss grains along the transverse direction, resulting in a poor secondary recrystallization microstructure.
1. Introduction Twin-roll casting is a novel ‘near-net-shape’ forming process by means of which thin strips can be produced directly from the melt [1,2]. Recent progress in twin-roll casting has made it possible to produce grain-oriented electrical steels using this technique [3–5]. For example, Liu et al. [3] and Song et al. [4,5] have shown that 0.23–0.27 mm thick grain-oriented electrical steel sheets can be successfully fabricated using twin-roll casting. In conventional processing, the {110}〈001〉 Goss texture is introduced near the surface during hot rolling as a result of the intense shear associated with roll friction [6,7]. This component survives through subsequent processing and provides the Goss nuclei necessary for successful secondary recrystallization [8,9]. The importance of the surface layer of the hot rolled sheet was demonstrated in experiments in which the Goss-containing layer was removed, resulting in incomplete secondary recrystallization [10]. Thus, hot rolling is generally considered to be indispensible during the conventional processing of grain-oriented electrical steel. More recently, Song et al. [11] and Fang et al. [12] have shown that grain-oriented electrical steel sheet can be successfully produced without hot rolling by employing a particular twin-roll casting route.
⁎
Corresponding author. E-mail address: [email protected] (H.-T. Liu).
http://dx.doi.org/10.1016/j.matdes.2017.06.016 Received 10 January 2017; Received in revised form 6 June 2017; Accepted 6 June 2017 Available online 07 June 2017 0264-1275/ © 2017 Elsevier Ltd. All rights reserved.
According to this method, the origin and development of the Goss orientation differs from those in the conventional process. In the conventional process, the effects of primary recrystallization on the secondary recrystallization behavior have been investigated in detail. Various models have been developed to explain the selective growth of the Goss grains, i.e. the size advantage model [13,14], the coincidence site lattice (CSL) model [15–17], the high energy (HE) model [18–20] and the solid-state wetting model [21–23]. When the twin-roll casting route is employed, the microstructures and textures produced by primary recrystallization differ considerably from those resulting from the conventional process. In the latter, the Goss texture is the major component of the primary recrystallization texture when the two-stage route is employed. By contrast, in the strip casting process, a strong γfiber and a weak Goss component are produced, leading to distinct secondary recrystallization behaviors. The relation between the characteristics of the primary recrystallization microstructure and the abnormal growth of secondary Goss grains is still not clearly understood. In the present paper, various primary recrystallization microstructures and textures were produced in a twin-roll cast grain-oriented electrical steel by employing different processing routes. The interaction between the primary recrystallization behavior and the abnormal
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
ND
ND RD
100 m
(c)
RD
100 m
(d)
ND
ND RD
100 m RD
100 m
Fig. 1. Microstructures of the (a) RI-PRA, (b) RII-PRA, (c) RIII-PRA and (d) RIV-PRA sheets.
3. Results
growth of secondary Goss grains was investigated. This work has led to an increased understanding of the development of sharp Goss textures in twin-roll cast grain-oriented electrical steels, as will be seen below.
3.1. Microstructure and texture of the as-cast strip The microstructure of the as-cast strip was composed of large columnar ferrite grains, as reported previously [11]. The associated texture consisted of the λ-fiber texture (〈001〉//ND) and did not vary significantly through the thickness [11]. These characteristics have been attributed to the high melt superheat [28]. It is important to note that the as-cast strip was not subjected to any deformation before annealing. Thus, the as-cast microstructure and texture were not changed because the as-cast strip did not undergo any recrystallization or phase transformation.
2. Experimental procedures The chemical composition (wt%) of the tested steel was 3.32 Si, 0.0026 C, 0.064 Mn, 0.024 S, 0.035 Al, 0.0093 N and balance Fe. A 2.3 mm-thick as-cast strip was produced using a vertical type twin-roll caster. During casting, a relatively high melt superheat was employed so as to produce a coarse solidification microstructure composed of large columnar grains. The experimental details of this process have been reported previously [24,25]. The as-cast strips were annealed at 950 °C for 4 min and water quenched. Here, four different routes were employed to produce the cold rolled sheets: (RI) the annealed sheets were directly cold rolled to the final thickness of 0.27 mm using a reduction of 88.3%; (RII) the annealed sheets were first subjected to a cold rolling using a reduction of 56.5%, intermediate annealed at 830 °C for 3 min, and then cold rolled to 0.27 mm using a reduction of 73.0%; (RIII) the annealed sheets were first cold rolled to 0.8 mm using a reduction of 65.2%, intermediate annealed at 830 °C for 3 min, and then cold rolled to 0.27 mm using a reduction of 66.3%; (RIV) the annealed sheets were first cold rolled down to 0.6 mm using a reduction of 73.9%, intermediate annealed at 830 °C for 3 min, and cold rolled to a final thickness of 0.27 mm using a reduction of 55.0%. In all four cases, primary recrystallization annealing was carried out at 830 °C for 3 min. Finally, the sheets were heated up to 1200 °C at a heating rate of 20 °C/h and held at 1200 °C for 10 h. For convenience, the samples are denoted here as first cold rolled (FCR) sheet, primary recrystallization annealed (PRA) sheet and secondary recrystallization annealed (SRA) sheet. The primary recrystallization microstructures were etched with 4% nital and examined using a Leica optical microscope. The secondary recrystallization microstructures were revealed by etching suitably ground samples with 10% hydrochloric acid. For characterization of the texture, the orientation distribution functions (ODFs) were calculated from three incomplete {110}, {200} and {211} pole figures by the series expansion method (lmax = 22) developed by Bunge [26,27]. For the electron backscatter diffraction (EBSD) measurements, the samples were wet ground using silicon carbide papers, mechanically polished and electropolished with a 14% perchloric acid/alcohol solution. The EBSD system was installed on a Zeiss Ultra 55 scanning electron microscope (SEM). The size distribution of the inhibitors was determined using SEM. The chemical composition of the inhibitors was verified by means of energy dispersive X-ray spectroscopy.
3.2. Primary recrystallization microstructure and texture The microstructures and textures of the primary recrystallization annealed sheets subjected to the four different routes are shown in Figs. 1 and 2, respectively. As can be seen, the RIII-PRA sheet had the most uniform microstructure consisting entirely of equiaxed ferrite grains with a size range of 10–23 μm, Fig. 1c. The texture was characterized by a strong γ-fiber texture, a medium α-fiber texture, and a weak Goss component. It is of particular note that many Goss grains were distributed throughout the thickness (Fig. 3b) and that these were characterized by grain boundaries of intermediate misorientations (20–45°). By contrast, the boundaries of the matrix as a whole had fewer misorientations in this range (Fig. 4a), in consistency with earlier observations [18–20,29]. Such boundaries have higher energies and thus facilitate coarsening of the inhibitors as well as the reduction of their pinning effects, leading to the selective growth of the Goss grains. This kind of grain boundary characteristic distribution (GBCD) is considered to facilitate the abnormal growth of Goss grains during secondary recrystallization annealing [18–20]. By contrast, the PRA sheets produced by the three other routes exhibited similar inhomogeneous microstructures with large elongated grains at the surface and relatively fine grains in the center, Fig. 1a, b and d. On the RD-TD section, the large elongated grains can be seen to be distributed inhomogeneously and separated by fine grained regions. The RI-PRA sheet exhibited a strong α-fiber texture at the surface, a strong {111}〈112〉 component and a medium λ-fiber texture through the thickness. The texture of the RII-PRA sheet was characterized by a strong partial α-fiber running from {001}〈110〉 to {111}〈110〉, a distinct λ-fiber and a weak γ-fiber from the surface to the 1/4 layer, these were associated with a medium α-fiber and a {111}〈112〉 component in the center. It is of interest that the RIV-PRA sheet displays a texture that is similar to that of the RII-PRA sheet. One difference is 168
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
(c)
(d)
Fig. 2. Textures of the (a) RI-PRA, (b) RII-PRA, (c) RIII-PRA and (d) RIV-PRA sheets.
that the strong α-fiber and λ-fiber texture are only present at the surface layer in the former sheet. Thus, although the schedules of the cold rolling reductions were totally different in routes RI, RII and RIV, strong α- and γ-fiber textures evolved at the surface layers in all three of these PRA sheets. A typical EBSD map of the RD-TD section of the RII-PRA sheet is illustrated in Fig. 5. As can be seen, the large elongated grains exhibit a strong partial α-fiber running from {001}〈110〉 to {223}〈110〉 and a partial λ-fiber texture running from {001}〈210〉 to {001}〈110〉. By contrast, the fine grained regions are mainly characterized by a strong γ-fiber texture and a weak {001}〈100〉 component (cube texture). In addition, many fine Goss grains are distributed throughout these regions; these again had more boundaries with misorientations of 20–45° than the fine grains as a whole, Fig. 5b. Thus, the microstructure, texture and GBCD around the Goss grains in the fine grained
regions were similar to those in the RIII-PRA sheet. In this work, the precipitation of dense AlN, MnS and co-precipitates of AlN and MnS was largely completed during annealing [11], resulting in essentially the same inhibitor characteristics in all the PRA sheets. As shown in Fig. 6, most of the precipitates ranged in size from 20 to 90 nm. 3.3. Secondary recrystallization microstructure and texture Some typical microstructures produced by secondary recrystallization annealing are illustrated in Fig. 7. As can be seen, only the RIII-SRA sheet displayed a complete secondary recrystallization microstructure. This route also produced the most homogeneous primary recrystallization microstructure. By contrast, the RI-SRA sheet had the poorest microstructure as it was composed largely of equiaxed grains. 169
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
4. Discussion 4.1. Origin and evolution of the Goss orientation prior to primary recrystallization Although hot rolling was eliminated in this work, successful secondary recrystallization was still achieved (Fig. 7c), even though the origin of the Goss oriented grains differed significantly from that in the conventional case. After the first cold rolling, Goss orientations evolved within the shear bands in the {111}〈112〉 and {111}〈110〉 deformed grains (Fig. 9a and b), as described by previous researchers [30–32]. It is important to note that more Goss components formed within the shear bands inclined at ~29° than in those inclined at ~17°. In addition, the Goss components within the bands with the ~29° inclination exhibited less displacement along the 〈110〉 axis, Fig. 9c. After 10 s of intermediate annealing, some Goss grains nucleated along the shear bands while the matrix remained in the deformed state, Fig. 9d. The preferential nucleation of Goss grains is considered to be related to its low stored energy [32–34]. Still more Goss grains formed after 20 more seconds of annealing, resulting in an increase in the intensity of this component, Fig. 10. After 3 min of annealing, an inhomogeneous microstructure consisting of fine grains and large elongated grains was produced. The large grains were characterized by a λand α-fiber texture, while the fine grains were associated with a γ-fiber texture together with a strong Goss component, Fig. 10. By comparing Fig. 10b and d, it can be seen that the Goss grains grew at the expense of the surrounding deformed matrix, which consisted largely of γ-grains. Continued rapid growth resulted in the formation of a strong Goss texture, Fig. 10. After the second cold rolling and primary annealing, many dispersed Goss grains were distributed throughout the thickness of the sheet, Fig. 3b. As shown in Fig. 3a, few Goss grains formed in the RI-PRA sheet processed according to the single-stage cold rolling route. This orientation originates within the shear bands formed during cold rolling. However, the heavy rolling reduction (88.3%) that was applied facilitated the formation of γ oriented grains instead and suppressed the development of Goss grains during primary recrystallization, apparently because of the lack of dense shear bands in the cold rolled sheet.
ND RD
200µm
{110}<001> (Deviation≤15°)
RD
200µm
{110}<001> (Deviation≤15°)
(b)
ND
Fig. 3. Grains within 15° of the exact Goss orientation identified in green in the (a) RIPRA and (b) RIII-PRA sheets. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The other two sheets produced by the two-stage cold rolling route were also characterized by poor secondary recrystallization microstructures; these contained narrow and elongated Goss grains as well as many fine equiaxed grains. A typical EBSD map illustrating the RD-TD microstructure of the RII-SRA sheet is presented in Fig. 8. The large elongated λ- and α-grains present in the primary recrystallization state have disappeared. As can be seen, growth of the secondary Goss grain along the TD is being blocked by many equiaxed grains, most of which have diameters of 100–300 μm. Some extremely large grains with the diameters exceeding 1000 μm were also observed, in agreement with the microstructure of Fig. 7d. The orientations of the equiaxed grains were characterized by a strong γ-fiber texture and a Goss component, Fig. 8. Once again, the Goss grains within the equiaxed grains exhibited a narrower and more intense intermediate misorientation angle distribution (20–45°) than their neighbors, Fig. 8. Thus, the equiaxed grains around the Goss grains in the RII-SRA sheet had similar textures and GBCDs as the fine-grained regions in the corresponding RII-PRA sheet.
4.2. Development of the primary recrystallization microstructure and texture The deformation microstructure has considerable influence on the final recrystallization microstructure and texture as it affects the nucleation and growth of new grains during annealing [35,36]. The frequency of nucleation of new grains is closely correlated with the degree of inhomogeneity of the deformation microstructure, that is, with the
(a)
(b)
0.24 0.21
0.21
Goss/neighbours Matrix
0.18
0.15
Frequency
Frequency
0.18
0.24
Goss/neighbours Matrix
0.12 0.09
0.15 0.12 0.09
0.06
0.06
0.03
0.03 0.00
0.00 0 5 10 15 20 25 30 35 40 45 50 55 60 65
0 5 10 15 20 25 30 35 40 45 50 55 60 65
Misorientation angle (degree)
Misorientation angle (degree)
Fig. 4. Grain boundary characteristic distribution in the (a) RIII-PRA and (b) RII-PRA sheets.
170
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(c)
(b)
(d)
Fig. 5. (a) Inverse pole figure map (normal direction) of the RII-PRA sheet, (b) image quality map with Goss grains identified in green (15° deviation), (c) ODF of the large elongated grains, and (d) ODF of the recrystallized grains within the fine grained regions. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(a)
(b) Fe
Fe
Al N
Si S
Mn
Fe
100nm Fig. 6. (a) SEM image showing the inhibitor size distribution and (b) a typical EDX spectrum in the RIII-PRA sheet.
to the elimination of the large elongated grains, Fig. 1c. This was probably related to the presence of the shear bands. The coarse initial microstructure, to which a moderate reduction of 65.2% was applied, favored the formation of shear bands during the first cold rolling [11]. These shear bands provided the additional nucleation sites for recrystallization, leading to the improved microstructure produced by intermediate annealing. As a result, the retained large grains underwent more constraints from their neighbors during second cold rolling,
presence of shear bands and deformation bands [36]. In the present work, the cold rolled sheet processed by route RI was characterized by two types of deformed grains, i.e. grains with uniform microstructures and grains containing dense deformation bands. Primary recrystallization annealing resulted in the formation of large elongated λ- and αgrains, indicating that it is difficult to eliminate such large grains by single-stage cold rolling, even at large reductions (88.3%). By contrast, the employment of two-stage cold rolling route RIII led 171
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
TD
TD RD
10mm
(c)
RD
10mm
(d)
TD
TD RD
10mm
RD
10mm
Fig. 7. Macrostructures on the RD-TD sections of the (a) RI-SRA, (b) RII-SRA, (c) RIII-SRA and (d) RIV-SRA sheets.
(a)
(b)
(c)
(d)
Fig. 8. (a) Inverse pole figure map (normal direction) of the RII-SRA sheet, (b) grain size distribution of the equiaxed grains, (c) ODF of the equiaxed grains and (d) misorientation angle distribution around the Goss grains compared to that of all the equiaxed grains.
4.3. Relation between the primary recrystallization microstructure and incomplete secondary recrystallization
introducing more deformation bands due to the interaction between adjacent grains. In addition, dense shear bands were also formed under the moderate rolling reduction of 66.3%. Thus, lots of nucleation sites for recrystallization were created, giving rise to the formation of homogeneous primary recrystallization microstructure. It should be noted that the other two PRA sheets processed by the two-stage cold rolling route contained similar inhomogeneous microstructures, Fig. 1b and d. These observations indicate that the differences in the primary recrystallization microstructures are probably related to the choice of the cold rolling reductions. Under the present conditions, only route RIII, with balanced first and second cold rolling reductions of 65.2% and 66.3%, produced dense shear bands in both first and second cold rolling steps.
In this work, the inhibitor conditions were similar in all the PRA sheets. Thus, it appears that it was the primary recrystallization microstructure that determined the secondary recrystallization microstructure. The poorest microstructure was that of the RI-SRA sheet, Fig. 7a. Here there were many elongated α-grains near the surface and only a few Goss grains present in the RI-PRA sheet, Fig. 3a. Thus, the conditions for successful secondary recrystallization were not satisfied. In the case of the two-stage cold rolling route, the unsuitable selection of reductions (RII and RIV) led to incomplete secondary recrystallization microstructures, Fig. 7b and d. The behavior of a large elongated grain during secondary recrystallization annealing by route RII is depicted in Fig. 11. As can be seen, the left side has retained its
172
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
(e) (c)
(d)
Fig. 9. (a) Inverse pole figure map (normal direction) of the RIII-FCR sheet, (b) image quality map; here grains that are within 15° of the ideal Goss orientation are identified in green, {100} pole figures of the Goss component within shear bands inclined at (c) ~ 29°and (d) ~ 17° with respect to the rolling direction, (e) image quality map of the RIII-FCR sheet annealed at 830 °C for 10 s. Here grains that are within 15° of the ideal Goss orientation are identified in green. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(a)
(b)
(c)
(d)
Fig. 10. (a) Image quality map of the RIIIFCR sheet annealed at 830 °C for 30 s. Here grains that are within 15° of the ideal Goss orientation are identified in green, (b) ODF of the recrystallized grains. (c) Image quality map of the RIII-FCR sheet annealed at 830 °C for 3 min, Goss grains shown in green, (d) ODF of the fine recrystallized grains. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4.4. Formation of successful secondary recrystallization microstructures
initial shape and orientation, while the rest has been consumed by large equiaxed λ-, α- and γ-grains. Thus it appears that most of the elongated grains present in the PRA sheets are eventually replaced by large equiaxed grains prior to the onset temperature of secondary recrystallization. The microstructural evolution of samples extracted at different temperatures during secondary recrystallization annealing along routes RII and RIV was investigated by EBSD. A schematic diagram illustrating this evolution is illustrated in Fig. 12, where it can be seen that growth along the TD was blocked by the large equiaxed grains because of their large sizes and unsuitable orientations. These only grew normally, leading to a poor secondary recrystallization microstructure consisting of Goss grains of limited size.
In this work, the RIII-PRA sheet exhibited the most homogeneous microstructure together with a strong γ-fiber texture and well-developed Goss components. The formation of this microstructure is illustrated in Fig. 13 on the RD-TD section. Here, some Goss grains start to grow at the onset temperature of secondary recrystallization. On increasing the temperature, the Goss grains grow rapidly along both the RD and TD directions by consuming the matrix grains. Finally, a complete secondary recrystallization microstructure composed of large Goss grains is formed. This can be attributed to the presence of a homogeneous primary recrystallization microstructure containing many Goss nuclei and a strong γ-fiber texture. As shown above, in the cast of initial coarse solidification microstructure, it is important to eliminate the large primary recrystallization 173
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
Fig. 11. (a) Inverse pole figure map (normal direction) of the RII-PRA sheet, (b) ODF of the grains within the rectangular frame. Fig. 12. Schematic diagram illustrating microstructural evolution during secondary recrystallization of a sheet with an undesirable primary recrystallization microstructure and texture (Note: the equiaxed matrix grains are generally < 100 μm in diameters during secondary recrystallization, while the matrix grains and lengths of the narrow Goss grains after secondary recrystallization annealing are about several millimeters and several centimeters, respectively.).
174
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
Fig. 13. Schematic diagram illustrating microstructural evolution during secondary recrystallization of a sheet with a desirable primary recrystallization microstructure and texture (Note: the matrix grains are generally < 100 μm in diameter during secondary recrystallization, while the lengths of the coarse Goss grains after secondary recrystallization annealing are about several centimeters.).
large equiaxed grains in the primary recrystallization material impeded secondary recrystallization by blocking the abnormal growth of secondary Goss grains along the transverse direction. The balanced twostage cold rolling route led to the best secondary recrystallization microstructure. In this case, the presence of a homogeneous primary recrystallization microstructure together with a relatively strong γ-fiber and a Goss component promoted successful secondary recrystallization.
grains by employing two-stage cold rolling with balanced amounts of cold rolling reduction. 5. Conclusions Various processing routes were employed to produce different primary recrystallization microstructures and textures in a twin-roll cast grain-oriented electrical steel. The relations between the primary recrystallization microstructures and the abnormal growth of secondary Goss grains were investigated. The primary recrystallization annealed sheet processed by single-stage cold rolling exhibited a strong α-fiber at the surface and a medium λ-fiber throughout the thickness. This type of sheet displayed the poorest secondary recrystallization microstructure as it was mainly composed of fine equiaxed grains. The presence of
Acknowledgments Hongyu Song is grateful to the support from Chinese Scholarship Council (No. 201506080039). The authors also acknowledge the financial support received from the National Natural Science Foundation of China (Grant nos. 51004035, 51374002, 51574078), the China 175
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
[17] P. Lin, G. Palumbo, J. Harase, K.T. Aust, Coincidence site lattice (CSL) grain boundaries and Goss texture development in Fe–3% Si alloy, Acta Mater. 44 (12) (1996) 4677–4683. [18] Y. Hayakawa, J.A. Szpunar, The role of grain boundary character distribution in Goss texture development in electrical steels, J. Magn. Magn. Mater. 160 (1996) 143–144. [19] Y. Hayakawa, J.A. Szpunar, A new model of Goss texture development during secondary recrystallization of electrical steel, Acta Mater. 45 (11) (1997) 4713–4720. [20] N. Rajmohan, J.A. Szpunar, Y. Hayakawa, A role of fractions of mobile grain boundaries in secondary recrystallization of Fe–Si steels, Acta Mater. 47 (10) (1999) 2999–3008. [21] K.J. Ko, A.D. Rollett, N.M. Hwang, Abnormal grain growth of Goss grains in Fe–3% Si steel driven by sub-boundary-enhanced solid-state wetting: analysis by Monte Carlo simulation, Acta Mater. 58 (2010) 4414–4423. [22] K.J. Ko, P.R. Cha, D. Srolovitz, N.M. Hwang, Abnormal grain growth induced by sub-boundary-enhanced solid-state wetting: analysis by phase-field model simulations, Acta Mater. 57 (2009) 838–845. [23] K.J. Ko, J.T. Park, J.K. Kim, N.M. Hwang, Morphological evidence that Goss abnormally growing grains grow by triple junction wetting during secondary recrystallization of Fe–3% Si steel, Scr. Mater. 59 (2008) 764–767. [24] H. Liu, Z. Liu, G. Cao, C. Li, G. Wang, Microstructure and texture evolution of strip casting 3 wt% Si non-oriented silicon steel with columnar structure, J. Magn. Magn. Mater. 323 (2011) 2648–2651. [25] H.T. Liu, Z.Y. Liu, Y. Sun, F. Gao, G.D. Wang, Development of λ-fiber recrystallization texture and magnetic property in Fe–6.5 wt% Si thin sheet produced by strip casting and warm rolling method, Mater. Lett. 91 (2013) 150–153. [26] H.J. Bunge, Texture Analysis in Materials Science: Mathematical Methods, Butterworths, London, 1982. [27] O. Engler, V. Randle, Introduction to Texture Analysis, second ed., CRC Press, London, 2009. [28] H. Liu, Z. Liu, C. Li, G. Cao, G. Wang, Solidification structure and crystallographic texture of strip casting 3 wt% Si non-oriented silicon steel, Mater. Charact. 62 (2011) 463–468. [29] T. Imamura, Y. Shingaki, Y. Hayakawa, Effect of cold rolling reduction rate on secondary recrystallized texture in 3 Pct Si–Fe steel, Metall. Trans. A. 44 (4) (2013) 1785–1792. [30] R. Haratani, B. Hutchinson, I.L. Dillamore, P. Bate, Contribution of shear banding to origin of Goss texture in silicon steel, Met. Sci. 18 (1984) 57–65. [31] K. Ushioda, B. Hutchinson, Role of shear bands in annealing texture formation in 3%Si–Fe(111)[112] single crystals, ISIJ Int. 29 (10) (1989) 862–867. [32] J.T. Park, J.A. Szpunar, Evolution of recrystallization texture in nonoriented electrical steels, Acta Mater. 51 (11) (2003) 3037–3051. [33] M.R. Barnett, J.J. Jonas, Influence of ferrite rolling temperature on microstructure and texture in deformed low C and IF steels, ISIJ Int. 37 (7) (1997) 697–705. [34] L. Kestens, J.J. Jonas, Modelling texture change during the static recrystallization of a cold rolled and annealed ultra low carbon steel previously warm rolled in the ferrite region, ISIJ Int. 37 (8) (1997) 807–814. [35] R.D. Doherty, D.A. Hughes, F.J. Humphreys, J.J. Jonas, D. Juul Jensen, M.E. Kassner, W.E. King, T.R. McNelly, H.J. McQueen, A.D. Rollett, Current issues in recrystallization: a review, Mater. Sci. Eng. A 238 (1997) 219–274. [36] J.J. Jonas, Effects of shear band formation on texture development in warm-rolled IF steels, J. Mater. Process. Technol. 117 (2001) 293–299.
Postdoctoral Science Foundation (Grant nos. 2014M560218, 2016T90228), the National Key Research and Development Program of China (2016YFB0300305) and the Fundamental Research Funds for the Central Universities (Grant nos. N140705001, N160705001). JJJ is indebted to the Natural Sciences and Engineering Research Council of Canada for financial support. References [1] K. Shibuya, M. Ozawa, Strip casting techniques for steel, ISIJ Int. 31 (7) (1991) 661–668. [2] S. Ge, M. Isac, R.L. Guthrie, Progress in strip casting technologies for steel: technical developments, ISIJ Int. 53 (5) (2013) 729–742. [3] H.T. Liu, S.J. Yao, Y. Sun, F. Gao, H.Y. Song, G.H. Liu, L. Li, D.Q. Geng, Z.Y. Liu, G.D. Wang, Evolution of microstructure, texture and inhibitor along the processing route for grain-oriented electrical steels using strip casting, Mater. Charact. 106 (2015) 273–282. [4] H.Y. Song, H.H. Lu, H.T. Liu, H.Z. Li, D.Q. Geng, R.D.K. Misra, Z.Y. Liu, G.D. Wang, Microstructure and texture of strip cast grain-oriented silicon steel after symmetrical and asymmetrical hot rolling, Steel Res. Int. 85 (10) (2014) 1477–1482. [5] H.Y. Song, H.T. Liu, H.H. Lu, L.Z. An, B.G. Zhang, W.Q. Liu, G.M. Cao, C.G. Li, Z.Y. Liu, G.D. Wang, Fabrication of grain-oriented silicon steel by a novel way strip casting process, Mater. Lett. 137 (2014) 475–478. [6] S. Mishra, C. Darmann, K. Lucke, On the development of the Goss texture in iron-3% silicon, Acta Metall. 32 (12) (1984) 2185–2201. [7] Y. Shimizu, Y. Ito, Y. Iida, Formation of the Goss orientation near the surface of 3 Pct silicon steel during hot rolling, Metall. Trans. A. 17 (1986) 1323–1334. [8] M. Matsuo, Texture control in the production of grain oriented silicon steels, ISIJ Int. 29 (10) (1989) 809–827. [9] D. Dorner, S. Zaefferer, D. Raabe, Retention of the Goss orientation between microbands during cold rolling of an Fe3%Si single crystal, Acta Mater. 55 (2007) 2519–2530. [10] A. Böttcher, K. Lücke, Acta Metall. Mater. 41 (1993) 2503–2514. [11] H.Y. Song, H.T. Liu, W.Q. Liu, Y.P. Wang, Z.Y. Liu, G.D. Wang, Effects of two-stage cold rolling schedule on microstructure and texture evolution of strip casting grainoriented silicon steel with extra-low carbon, Metall. Trans. A. 47 (2016) 1770–1781. [12] F. Fang, Y. Zhang, X. Lu, Y. Wang, G. Cao, G. Yuan, Y. Xu, G. Wang, R.D.K. Misra, Inhibitor induced secondary recrystallization in thin-gauge grain oriented silicon steel with high permeability, Mater. Des. 105 (2016) 398–403. [13] J.E. May, D. Turnbull, Secondary recrystallization in silicon steel, Trans. Metall. Soc. AIME 212 (1958) 769. [14] Y. Inokuti, C. Maeda, Transmission Kossel study of the formation of Goss grains after an intermediate annealing in grain oriented silicon steel, Trans. Iron Steel Inst. Jpn. 24 (8) (1984) 655–662. [15] R. Shimizu, J. Harase, Coincidence grain boundary and texture evolution in Fe–3% Si, Acta Metall. 37 (4) (1989) 1241–1249. [16] Y. Yoshitomi, Y. Ushigami, J. Harase, T. Nakayama, H. Masui, N. Takahashi, Coincidence grain boundary and role of primary recrystallized grain growth on secondary recrystallization texture evolution in Fe–3%Si alloy, Acta Metall. Mater. 42 (8) (1994) 2593–2602.
176
Contents lists available at ScienceDirect
Materials & Design journal homepage: www.elsevier.com/locate/matdes
Effect of primary recrystallization microstructure on abnormal growth of Goss grains in a twin-roll cast grain-oriented electrical steel
MARK
Hong-Yu Songa, Hai-Tao Liua,⁎, John J. Jonasb, Guo-Dong Wanga a b
State Key Laboratory of Rolling and Automation, Northeastern University, Shenyang 110819, PR China Materials Engineering Department, McGill University, 3610 University Street, Montreal H3A 0C5, Canada
A R T I C L E I N F O
A B S T R A C T
Keywords: Goss grains Twin-roll casting Grain-oriented electrical steel Cold rolling Abnormal grain growth
Various primary recrystallization microstructures and textures were produced in a twin-roll cast grain-oriented electrical steel by employing different routes. The relationship between the primary recrystallization microstructure and texture and the abnormal growth of secondary Goss grains was investigated. The results show that the cold rolling reductions have a significant influence on secondary recrystallization by changing the primary recrystallization microstructure. Sheet processed by single-stage cold rolling with 88.3% reduction displayed the poorest secondary recrystallization microstructure as it contained many small equiaxed grains. By contrast, the employment of two-stage cold rolling markedly improved the secondary recrystallization microstructure. In the case of two moderate reductions of 65.2% and 66.3%, dense deformation substructures formed during both the first and second cold rolling, leading to a homogeneous primary recrystallization microstructure together with a strong γ-fiber texture. In this way, a suitable secondary recrystallization microstructure consisting of large Goss grains was produced. In the case of the inappropriate reductions, many large λ- and α-grains in the primary recrystallization matrix blocked the growth of secondary Goss grains along the transverse direction, resulting in a poor secondary recrystallization microstructure.
1. Introduction Twin-roll casting is a novel ‘near-net-shape’ forming process by means of which thin strips can be produced directly from the melt [1,2]. Recent progress in twin-roll casting has made it possible to produce grain-oriented electrical steels using this technique [3–5]. For example, Liu et al. [3] and Song et al. [4,5] have shown that 0.23–0.27 mm thick grain-oriented electrical steel sheets can be successfully fabricated using twin-roll casting. In conventional processing, the {110}〈001〉 Goss texture is introduced near the surface during hot rolling as a result of the intense shear associated with roll friction [6,7]. This component survives through subsequent processing and provides the Goss nuclei necessary for successful secondary recrystallization [8,9]. The importance of the surface layer of the hot rolled sheet was demonstrated in experiments in which the Goss-containing layer was removed, resulting in incomplete secondary recrystallization [10]. Thus, hot rolling is generally considered to be indispensible during the conventional processing of grain-oriented electrical steel. More recently, Song et al. [11] and Fang et al. [12] have shown that grain-oriented electrical steel sheet can be successfully produced without hot rolling by employing a particular twin-roll casting route.
⁎
Corresponding author. E-mail address: [email protected] (H.-T. Liu).
http://dx.doi.org/10.1016/j.matdes.2017.06.016 Received 10 January 2017; Received in revised form 6 June 2017; Accepted 6 June 2017 Available online 07 June 2017 0264-1275/ © 2017 Elsevier Ltd. All rights reserved.
According to this method, the origin and development of the Goss orientation differs from those in the conventional process. In the conventional process, the effects of primary recrystallization on the secondary recrystallization behavior have been investigated in detail. Various models have been developed to explain the selective growth of the Goss grains, i.e. the size advantage model [13,14], the coincidence site lattice (CSL) model [15–17], the high energy (HE) model [18–20] and the solid-state wetting model [21–23]. When the twin-roll casting route is employed, the microstructures and textures produced by primary recrystallization differ considerably from those resulting from the conventional process. In the latter, the Goss texture is the major component of the primary recrystallization texture when the two-stage route is employed. By contrast, in the strip casting process, a strong γfiber and a weak Goss component are produced, leading to distinct secondary recrystallization behaviors. The relation between the characteristics of the primary recrystallization microstructure and the abnormal growth of secondary Goss grains is still not clearly understood. In the present paper, various primary recrystallization microstructures and textures were produced in a twin-roll cast grain-oriented electrical steel by employing different processing routes. The interaction between the primary recrystallization behavior and the abnormal
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
ND
ND RD
100 m
(c)
RD
100 m
(d)
ND
ND RD
100 m RD
100 m
Fig. 1. Microstructures of the (a) RI-PRA, (b) RII-PRA, (c) RIII-PRA and (d) RIV-PRA sheets.
3. Results
growth of secondary Goss grains was investigated. This work has led to an increased understanding of the development of sharp Goss textures in twin-roll cast grain-oriented electrical steels, as will be seen below.
3.1. Microstructure and texture of the as-cast strip The microstructure of the as-cast strip was composed of large columnar ferrite grains, as reported previously [11]. The associated texture consisted of the λ-fiber texture (〈001〉//ND) and did not vary significantly through the thickness [11]. These characteristics have been attributed to the high melt superheat [28]. It is important to note that the as-cast strip was not subjected to any deformation before annealing. Thus, the as-cast microstructure and texture were not changed because the as-cast strip did not undergo any recrystallization or phase transformation.
2. Experimental procedures The chemical composition (wt%) of the tested steel was 3.32 Si, 0.0026 C, 0.064 Mn, 0.024 S, 0.035 Al, 0.0093 N and balance Fe. A 2.3 mm-thick as-cast strip was produced using a vertical type twin-roll caster. During casting, a relatively high melt superheat was employed so as to produce a coarse solidification microstructure composed of large columnar grains. The experimental details of this process have been reported previously [24,25]. The as-cast strips were annealed at 950 °C for 4 min and water quenched. Here, four different routes were employed to produce the cold rolled sheets: (RI) the annealed sheets were directly cold rolled to the final thickness of 0.27 mm using a reduction of 88.3%; (RII) the annealed sheets were first subjected to a cold rolling using a reduction of 56.5%, intermediate annealed at 830 °C for 3 min, and then cold rolled to 0.27 mm using a reduction of 73.0%; (RIII) the annealed sheets were first cold rolled to 0.8 mm using a reduction of 65.2%, intermediate annealed at 830 °C for 3 min, and then cold rolled to 0.27 mm using a reduction of 66.3%; (RIV) the annealed sheets were first cold rolled down to 0.6 mm using a reduction of 73.9%, intermediate annealed at 830 °C for 3 min, and cold rolled to a final thickness of 0.27 mm using a reduction of 55.0%. In all four cases, primary recrystallization annealing was carried out at 830 °C for 3 min. Finally, the sheets were heated up to 1200 °C at a heating rate of 20 °C/h and held at 1200 °C for 10 h. For convenience, the samples are denoted here as first cold rolled (FCR) sheet, primary recrystallization annealed (PRA) sheet and secondary recrystallization annealed (SRA) sheet. The primary recrystallization microstructures were etched with 4% nital and examined using a Leica optical microscope. The secondary recrystallization microstructures were revealed by etching suitably ground samples with 10% hydrochloric acid. For characterization of the texture, the orientation distribution functions (ODFs) were calculated from three incomplete {110}, {200} and {211} pole figures by the series expansion method (lmax = 22) developed by Bunge [26,27]. For the electron backscatter diffraction (EBSD) measurements, the samples were wet ground using silicon carbide papers, mechanically polished and electropolished with a 14% perchloric acid/alcohol solution. The EBSD system was installed on a Zeiss Ultra 55 scanning electron microscope (SEM). The size distribution of the inhibitors was determined using SEM. The chemical composition of the inhibitors was verified by means of energy dispersive X-ray spectroscopy.
3.2. Primary recrystallization microstructure and texture The microstructures and textures of the primary recrystallization annealed sheets subjected to the four different routes are shown in Figs. 1 and 2, respectively. As can be seen, the RIII-PRA sheet had the most uniform microstructure consisting entirely of equiaxed ferrite grains with a size range of 10–23 μm, Fig. 1c. The texture was characterized by a strong γ-fiber texture, a medium α-fiber texture, and a weak Goss component. It is of particular note that many Goss grains were distributed throughout the thickness (Fig. 3b) and that these were characterized by grain boundaries of intermediate misorientations (20–45°). By contrast, the boundaries of the matrix as a whole had fewer misorientations in this range (Fig. 4a), in consistency with earlier observations [18–20,29]. Such boundaries have higher energies and thus facilitate coarsening of the inhibitors as well as the reduction of their pinning effects, leading to the selective growth of the Goss grains. This kind of grain boundary characteristic distribution (GBCD) is considered to facilitate the abnormal growth of Goss grains during secondary recrystallization annealing [18–20]. By contrast, the PRA sheets produced by the three other routes exhibited similar inhomogeneous microstructures with large elongated grains at the surface and relatively fine grains in the center, Fig. 1a, b and d. On the RD-TD section, the large elongated grains can be seen to be distributed inhomogeneously and separated by fine grained regions. The RI-PRA sheet exhibited a strong α-fiber texture at the surface, a strong {111}〈112〉 component and a medium λ-fiber texture through the thickness. The texture of the RII-PRA sheet was characterized by a strong partial α-fiber running from {001}〈110〉 to {111}〈110〉, a distinct λ-fiber and a weak γ-fiber from the surface to the 1/4 layer, these were associated with a medium α-fiber and a {111}〈112〉 component in the center. It is of interest that the RIV-PRA sheet displays a texture that is similar to that of the RII-PRA sheet. One difference is 168
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
(c)
(d)
Fig. 2. Textures of the (a) RI-PRA, (b) RII-PRA, (c) RIII-PRA and (d) RIV-PRA sheets.
that the strong α-fiber and λ-fiber texture are only present at the surface layer in the former sheet. Thus, although the schedules of the cold rolling reductions were totally different in routes RI, RII and RIV, strong α- and γ-fiber textures evolved at the surface layers in all three of these PRA sheets. A typical EBSD map of the RD-TD section of the RII-PRA sheet is illustrated in Fig. 5. As can be seen, the large elongated grains exhibit a strong partial α-fiber running from {001}〈110〉 to {223}〈110〉 and a partial λ-fiber texture running from {001}〈210〉 to {001}〈110〉. By contrast, the fine grained regions are mainly characterized by a strong γ-fiber texture and a weak {001}〈100〉 component (cube texture). In addition, many fine Goss grains are distributed throughout these regions; these again had more boundaries with misorientations of 20–45° than the fine grains as a whole, Fig. 5b. Thus, the microstructure, texture and GBCD around the Goss grains in the fine grained
regions were similar to those in the RIII-PRA sheet. In this work, the precipitation of dense AlN, MnS and co-precipitates of AlN and MnS was largely completed during annealing [11], resulting in essentially the same inhibitor characteristics in all the PRA sheets. As shown in Fig. 6, most of the precipitates ranged in size from 20 to 90 nm. 3.3. Secondary recrystallization microstructure and texture Some typical microstructures produced by secondary recrystallization annealing are illustrated in Fig. 7. As can be seen, only the RIII-SRA sheet displayed a complete secondary recrystallization microstructure. This route also produced the most homogeneous primary recrystallization microstructure. By contrast, the RI-SRA sheet had the poorest microstructure as it was composed largely of equiaxed grains. 169
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
4. Discussion 4.1. Origin and evolution of the Goss orientation prior to primary recrystallization Although hot rolling was eliminated in this work, successful secondary recrystallization was still achieved (Fig. 7c), even though the origin of the Goss oriented grains differed significantly from that in the conventional case. After the first cold rolling, Goss orientations evolved within the shear bands in the {111}〈112〉 and {111}〈110〉 deformed grains (Fig. 9a and b), as described by previous researchers [30–32]. It is important to note that more Goss components formed within the shear bands inclined at ~29° than in those inclined at ~17°. In addition, the Goss components within the bands with the ~29° inclination exhibited less displacement along the 〈110〉 axis, Fig. 9c. After 10 s of intermediate annealing, some Goss grains nucleated along the shear bands while the matrix remained in the deformed state, Fig. 9d. The preferential nucleation of Goss grains is considered to be related to its low stored energy [32–34]. Still more Goss grains formed after 20 more seconds of annealing, resulting in an increase in the intensity of this component, Fig. 10. After 3 min of annealing, an inhomogeneous microstructure consisting of fine grains and large elongated grains was produced. The large grains were characterized by a λand α-fiber texture, while the fine grains were associated with a γ-fiber texture together with a strong Goss component, Fig. 10. By comparing Fig. 10b and d, it can be seen that the Goss grains grew at the expense of the surrounding deformed matrix, which consisted largely of γ-grains. Continued rapid growth resulted in the formation of a strong Goss texture, Fig. 10. After the second cold rolling and primary annealing, many dispersed Goss grains were distributed throughout the thickness of the sheet, Fig. 3b. As shown in Fig. 3a, few Goss grains formed in the RI-PRA sheet processed according to the single-stage cold rolling route. This orientation originates within the shear bands formed during cold rolling. However, the heavy rolling reduction (88.3%) that was applied facilitated the formation of γ oriented grains instead and suppressed the development of Goss grains during primary recrystallization, apparently because of the lack of dense shear bands in the cold rolled sheet.
ND RD
200µm
{110}<001> (Deviation≤15°)
RD
200µm
{110}<001> (Deviation≤15°)
(b)
ND
Fig. 3. Grains within 15° of the exact Goss orientation identified in green in the (a) RIPRA and (b) RIII-PRA sheets. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
The other two sheets produced by the two-stage cold rolling route were also characterized by poor secondary recrystallization microstructures; these contained narrow and elongated Goss grains as well as many fine equiaxed grains. A typical EBSD map illustrating the RD-TD microstructure of the RII-SRA sheet is presented in Fig. 8. The large elongated λ- and α-grains present in the primary recrystallization state have disappeared. As can be seen, growth of the secondary Goss grain along the TD is being blocked by many equiaxed grains, most of which have diameters of 100–300 μm. Some extremely large grains with the diameters exceeding 1000 μm were also observed, in agreement with the microstructure of Fig. 7d. The orientations of the equiaxed grains were characterized by a strong γ-fiber texture and a Goss component, Fig. 8. Once again, the Goss grains within the equiaxed grains exhibited a narrower and more intense intermediate misorientation angle distribution (20–45°) than their neighbors, Fig. 8. Thus, the equiaxed grains around the Goss grains in the RII-SRA sheet had similar textures and GBCDs as the fine-grained regions in the corresponding RII-PRA sheet.
4.2. Development of the primary recrystallization microstructure and texture The deformation microstructure has considerable influence on the final recrystallization microstructure and texture as it affects the nucleation and growth of new grains during annealing [35,36]. The frequency of nucleation of new grains is closely correlated with the degree of inhomogeneity of the deformation microstructure, that is, with the
(a)
(b)
0.24 0.21
0.21
Goss/neighbours Matrix
0.18
0.15
Frequency
Frequency
0.18
0.24
Goss/neighbours Matrix
0.12 0.09
0.15 0.12 0.09
0.06
0.06
0.03
0.03 0.00
0.00 0 5 10 15 20 25 30 35 40 45 50 55 60 65
0 5 10 15 20 25 30 35 40 45 50 55 60 65
Misorientation angle (degree)
Misorientation angle (degree)
Fig. 4. Grain boundary characteristic distribution in the (a) RIII-PRA and (b) RII-PRA sheets.
170
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(c)
(b)
(d)
Fig. 5. (a) Inverse pole figure map (normal direction) of the RII-PRA sheet, (b) image quality map with Goss grains identified in green (15° deviation), (c) ODF of the large elongated grains, and (d) ODF of the recrystallized grains within the fine grained regions. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(a)
(b) Fe
Fe
Al N
Si S
Mn
Fe
100nm Fig. 6. (a) SEM image showing the inhibitor size distribution and (b) a typical EDX spectrum in the RIII-PRA sheet.
to the elimination of the large elongated grains, Fig. 1c. This was probably related to the presence of the shear bands. The coarse initial microstructure, to which a moderate reduction of 65.2% was applied, favored the formation of shear bands during the first cold rolling [11]. These shear bands provided the additional nucleation sites for recrystallization, leading to the improved microstructure produced by intermediate annealing. As a result, the retained large grains underwent more constraints from their neighbors during second cold rolling,
presence of shear bands and deformation bands [36]. In the present work, the cold rolled sheet processed by route RI was characterized by two types of deformed grains, i.e. grains with uniform microstructures and grains containing dense deformation bands. Primary recrystallization annealing resulted in the formation of large elongated λ- and αgrains, indicating that it is difficult to eliminate such large grains by single-stage cold rolling, even at large reductions (88.3%). By contrast, the employment of two-stage cold rolling route RIII led 171
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
TD
TD RD
10mm
(c)
RD
10mm
(d)
TD
TD RD
10mm
RD
10mm
Fig. 7. Macrostructures on the RD-TD sections of the (a) RI-SRA, (b) RII-SRA, (c) RIII-SRA and (d) RIV-SRA sheets.
(a)
(b)
(c)
(d)
Fig. 8. (a) Inverse pole figure map (normal direction) of the RII-SRA sheet, (b) grain size distribution of the equiaxed grains, (c) ODF of the equiaxed grains and (d) misorientation angle distribution around the Goss grains compared to that of all the equiaxed grains.
4.3. Relation between the primary recrystallization microstructure and incomplete secondary recrystallization
introducing more deformation bands due to the interaction between adjacent grains. In addition, dense shear bands were also formed under the moderate rolling reduction of 66.3%. Thus, lots of nucleation sites for recrystallization were created, giving rise to the formation of homogeneous primary recrystallization microstructure. It should be noted that the other two PRA sheets processed by the two-stage cold rolling route contained similar inhomogeneous microstructures, Fig. 1b and d. These observations indicate that the differences in the primary recrystallization microstructures are probably related to the choice of the cold rolling reductions. Under the present conditions, only route RIII, with balanced first and second cold rolling reductions of 65.2% and 66.3%, produced dense shear bands in both first and second cold rolling steps.
In this work, the inhibitor conditions were similar in all the PRA sheets. Thus, it appears that it was the primary recrystallization microstructure that determined the secondary recrystallization microstructure. The poorest microstructure was that of the RI-SRA sheet, Fig. 7a. Here there were many elongated α-grains near the surface and only a few Goss grains present in the RI-PRA sheet, Fig. 3a. Thus, the conditions for successful secondary recrystallization were not satisfied. In the case of the two-stage cold rolling route, the unsuitable selection of reductions (RII and RIV) led to incomplete secondary recrystallization microstructures, Fig. 7b and d. The behavior of a large elongated grain during secondary recrystallization annealing by route RII is depicted in Fig. 11. As can be seen, the left side has retained its
172
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
(e) (c)
(d)
Fig. 9. (a) Inverse pole figure map (normal direction) of the RIII-FCR sheet, (b) image quality map; here grains that are within 15° of the ideal Goss orientation are identified in green, {100} pole figures of the Goss component within shear bands inclined at (c) ~ 29°and (d) ~ 17° with respect to the rolling direction, (e) image quality map of the RIII-FCR sheet annealed at 830 °C for 10 s. Here grains that are within 15° of the ideal Goss orientation are identified in green. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(a)
(b)
(c)
(d)
Fig. 10. (a) Image quality map of the RIIIFCR sheet annealed at 830 °C for 30 s. Here grains that are within 15° of the ideal Goss orientation are identified in green, (b) ODF of the recrystallized grains. (c) Image quality map of the RIII-FCR sheet annealed at 830 °C for 3 min, Goss grains shown in green, (d) ODF of the fine recrystallized grains. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
4.4. Formation of successful secondary recrystallization microstructures
initial shape and orientation, while the rest has been consumed by large equiaxed λ-, α- and γ-grains. Thus it appears that most of the elongated grains present in the PRA sheets are eventually replaced by large equiaxed grains prior to the onset temperature of secondary recrystallization. The microstructural evolution of samples extracted at different temperatures during secondary recrystallization annealing along routes RII and RIV was investigated by EBSD. A schematic diagram illustrating this evolution is illustrated in Fig. 12, where it can be seen that growth along the TD was blocked by the large equiaxed grains because of their large sizes and unsuitable orientations. These only grew normally, leading to a poor secondary recrystallization microstructure consisting of Goss grains of limited size.
In this work, the RIII-PRA sheet exhibited the most homogeneous microstructure together with a strong γ-fiber texture and well-developed Goss components. The formation of this microstructure is illustrated in Fig. 13 on the RD-TD section. Here, some Goss grains start to grow at the onset temperature of secondary recrystallization. On increasing the temperature, the Goss grains grow rapidly along both the RD and TD directions by consuming the matrix grains. Finally, a complete secondary recrystallization microstructure composed of large Goss grains is formed. This can be attributed to the presence of a homogeneous primary recrystallization microstructure containing many Goss nuclei and a strong γ-fiber texture. As shown above, in the cast of initial coarse solidification microstructure, it is important to eliminate the large primary recrystallization 173
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
(a)
(b)
Fig. 11. (a) Inverse pole figure map (normal direction) of the RII-PRA sheet, (b) ODF of the grains within the rectangular frame. Fig. 12. Schematic diagram illustrating microstructural evolution during secondary recrystallization of a sheet with an undesirable primary recrystallization microstructure and texture (Note: the equiaxed matrix grains are generally < 100 μm in diameters during secondary recrystallization, while the matrix grains and lengths of the narrow Goss grains after secondary recrystallization annealing are about several millimeters and several centimeters, respectively.).
174
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
Fig. 13. Schematic diagram illustrating microstructural evolution during secondary recrystallization of a sheet with a desirable primary recrystallization microstructure and texture (Note: the matrix grains are generally < 100 μm in diameter during secondary recrystallization, while the lengths of the coarse Goss grains after secondary recrystallization annealing are about several centimeters.).
large equiaxed grains in the primary recrystallization material impeded secondary recrystallization by blocking the abnormal growth of secondary Goss grains along the transverse direction. The balanced twostage cold rolling route led to the best secondary recrystallization microstructure. In this case, the presence of a homogeneous primary recrystallization microstructure together with a relatively strong γ-fiber and a Goss component promoted successful secondary recrystallization.
grains by employing two-stage cold rolling with balanced amounts of cold rolling reduction. 5. Conclusions Various processing routes were employed to produce different primary recrystallization microstructures and textures in a twin-roll cast grain-oriented electrical steel. The relations between the primary recrystallization microstructures and the abnormal growth of secondary Goss grains were investigated. The primary recrystallization annealed sheet processed by single-stage cold rolling exhibited a strong α-fiber at the surface and a medium λ-fiber throughout the thickness. This type of sheet displayed the poorest secondary recrystallization microstructure as it was mainly composed of fine equiaxed grains. The presence of
Acknowledgments Hongyu Song is grateful to the support from Chinese Scholarship Council (No. 201506080039). The authors also acknowledge the financial support received from the National Natural Science Foundation of China (Grant nos. 51004035, 51374002, 51574078), the China 175
Materials & Design 131 (2017) 167–176
H.-Y. Song et al.
[17] P. Lin, G. Palumbo, J. Harase, K.T. Aust, Coincidence site lattice (CSL) grain boundaries and Goss texture development in Fe–3% Si alloy, Acta Mater. 44 (12) (1996) 4677–4683. [18] Y. Hayakawa, J.A. Szpunar, The role of grain boundary character distribution in Goss texture development in electrical steels, J. Magn. Magn. Mater. 160 (1996) 143–144. [19] Y. Hayakawa, J.A. Szpunar, A new model of Goss texture development during secondary recrystallization of electrical steel, Acta Mater. 45 (11) (1997) 4713–4720. [20] N. Rajmohan, J.A. Szpunar, Y. Hayakawa, A role of fractions of mobile grain boundaries in secondary recrystallization of Fe–Si steels, Acta Mater. 47 (10) (1999) 2999–3008. [21] K.J. Ko, A.D. Rollett, N.M. Hwang, Abnormal grain growth of Goss grains in Fe–3% Si steel driven by sub-boundary-enhanced solid-state wetting: analysis by Monte Carlo simulation, Acta Mater. 58 (2010) 4414–4423. [22] K.J. Ko, P.R. Cha, D. Srolovitz, N.M. Hwang, Abnormal grain growth induced by sub-boundary-enhanced solid-state wetting: analysis by phase-field model simulations, Acta Mater. 57 (2009) 838–845. [23] K.J. Ko, J.T. Park, J.K. Kim, N.M. Hwang, Morphological evidence that Goss abnormally growing grains grow by triple junction wetting during secondary recrystallization of Fe–3% Si steel, Scr. Mater. 59 (2008) 764–767. [24] H. Liu, Z. Liu, G. Cao, C. Li, G. Wang, Microstructure and texture evolution of strip casting 3 wt% Si non-oriented silicon steel with columnar structure, J. Magn. Magn. Mater. 323 (2011) 2648–2651. [25] H.T. Liu, Z.Y. Liu, Y. Sun, F. Gao, G.D. Wang, Development of λ-fiber recrystallization texture and magnetic property in Fe–6.5 wt% Si thin sheet produced by strip casting and warm rolling method, Mater. Lett. 91 (2013) 150–153. [26] H.J. Bunge, Texture Analysis in Materials Science: Mathematical Methods, Butterworths, London, 1982. [27] O. Engler, V. Randle, Introduction to Texture Analysis, second ed., CRC Press, London, 2009. [28] H. Liu, Z. Liu, C. Li, G. Cao, G. Wang, Solidification structure and crystallographic texture of strip casting 3 wt% Si non-oriented silicon steel, Mater. Charact. 62 (2011) 463–468. [29] T. Imamura, Y. Shingaki, Y. Hayakawa, Effect of cold rolling reduction rate on secondary recrystallized texture in 3 Pct Si–Fe steel, Metall. Trans. A. 44 (4) (2013) 1785–1792. [30] R. Haratani, B. Hutchinson, I.L. Dillamore, P. Bate, Contribution of shear banding to origin of Goss texture in silicon steel, Met. Sci. 18 (1984) 57–65. [31] K. Ushioda, B. Hutchinson, Role of shear bands in annealing texture formation in 3%Si–Fe(111)[112] single crystals, ISIJ Int. 29 (10) (1989) 862–867. [32] J.T. Park, J.A. Szpunar, Evolution of recrystallization texture in nonoriented electrical steels, Acta Mater. 51 (11) (2003) 3037–3051. [33] M.R. Barnett, J.J. Jonas, Influence of ferrite rolling temperature on microstructure and texture in deformed low C and IF steels, ISIJ Int. 37 (7) (1997) 697–705. [34] L. Kestens, J.J. Jonas, Modelling texture change during the static recrystallization of a cold rolled and annealed ultra low carbon steel previously warm rolled in the ferrite region, ISIJ Int. 37 (8) (1997) 807–814. [35] R.D. Doherty, D.A. Hughes, F.J. Humphreys, J.J. Jonas, D. Juul Jensen, M.E. Kassner, W.E. King, T.R. McNelly, H.J. McQueen, A.D. Rollett, Current issues in recrystallization: a review, Mater. Sci. Eng. A 238 (1997) 219–274. [36] J.J. Jonas, Effects of shear band formation on texture development in warm-rolled IF steels, J. Mater. Process. Technol. 117 (2001) 293–299.
Postdoctoral Science Foundation (Grant nos. 2014M560218, 2016T90228), the National Key Research and Development Program of China (2016YFB0300305) and the Fundamental Research Funds for the Central Universities (Grant nos. N140705001, N160705001). JJJ is indebted to the Natural Sciences and Engineering Research Council of Canada for financial support. References [1] K. Shibuya, M. Ozawa, Strip casting techniques for steel, ISIJ Int. 31 (7) (1991) 661–668. [2] S. Ge, M. Isac, R.L. Guthrie, Progress in strip casting technologies for steel: technical developments, ISIJ Int. 53 (5) (2013) 729–742. [3] H.T. Liu, S.J. Yao, Y. Sun, F. Gao, H.Y. Song, G.H. Liu, L. Li, D.Q. Geng, Z.Y. Liu, G.D. Wang, Evolution of microstructure, texture and inhibitor along the processing route for grain-oriented electrical steels using strip casting, Mater. Charact. 106 (2015) 273–282. [4] H.Y. Song, H.H. Lu, H.T. Liu, H.Z. Li, D.Q. Geng, R.D.K. Misra, Z.Y. Liu, G.D. Wang, Microstructure and texture of strip cast grain-oriented silicon steel after symmetrical and asymmetrical hot rolling, Steel Res. Int. 85 (10) (2014) 1477–1482. [5] H.Y. Song, H.T. Liu, H.H. Lu, L.Z. An, B.G. Zhang, W.Q. Liu, G.M. Cao, C.G. Li, Z.Y. Liu, G.D. Wang, Fabrication of grain-oriented silicon steel by a novel way strip casting process, Mater. Lett. 137 (2014) 475–478. [6] S. Mishra, C. Darmann, K. Lucke, On the development of the Goss texture in iron-3% silicon, Acta Metall. 32 (12) (1984) 2185–2201. [7] Y. Shimizu, Y. Ito, Y. Iida, Formation of the Goss orientation near the surface of 3 Pct silicon steel during hot rolling, Metall. Trans. A. 17 (1986) 1323–1334. [8] M. Matsuo, Texture control in the production of grain oriented silicon steels, ISIJ Int. 29 (10) (1989) 809–827. [9] D. Dorner, S. Zaefferer, D. Raabe, Retention of the Goss orientation between microbands during cold rolling of an Fe3%Si single crystal, Acta Mater. 55 (2007) 2519–2530. [10] A. Böttcher, K. Lücke, Acta Metall. Mater. 41 (1993) 2503–2514. [11] H.Y. Song, H.T. Liu, W.Q. Liu, Y.P. Wang, Z.Y. Liu, G.D. Wang, Effects of two-stage cold rolling schedule on microstructure and texture evolution of strip casting grainoriented silicon steel with extra-low carbon, Metall. Trans. A. 47 (2016) 1770–1781. [12] F. Fang, Y. Zhang, X. Lu, Y. Wang, G. Cao, G. Yuan, Y. Xu, G. Wang, R.D.K. Misra, Inhibitor induced secondary recrystallization in thin-gauge grain oriented silicon steel with high permeability, Mater. Des. 105 (2016) 398–403. [13] J.E. May, D. Turnbull, Secondary recrystallization in silicon steel, Trans. Metall. Soc. AIME 212 (1958) 769. [14] Y. Inokuti, C. Maeda, Transmission Kossel study of the formation of Goss grains after an intermediate annealing in grain oriented silicon steel, Trans. Iron Steel Inst. Jpn. 24 (8) (1984) 655–662. [15] R. Shimizu, J. Harase, Coincidence grain boundary and texture evolution in Fe–3% Si, Acta Metall. 37 (4) (1989) 1241–1249. [16] Y. Yoshitomi, Y. Ushigami, J. Harase, T. Nakayama, H. Masui, N. Takahashi, Coincidence grain boundary and role of primary recrystallized grain growth on secondary recrystallization texture evolution in Fe–3%Si alloy, Acta Metall. Mater. 42 (8) (1994) 2593–2602.
176