www.steel-research.

de

FULL PAPER
Effect of Hot Rolling on Texture, Precipitation,
and Magnetic Properties of Strip-Cast
Grain-Oriented Silicon Steel
Yang Wang,
Yuan-Xiang Zhang, Xiang Lu, Feng Fang, Guang-Ming Cao, Cheng-Gang Li,
Yun-Bo Xu, R. Devesh Kumar Misra, and Guo-Dong Wang

An ultra-low carbon grain-oriented silicon steel as-cast strip is produced by twin-roll strip
casting process, and subsequently subjected to one-pass hot rolling, one-stage cold rolling,
primary annealing, and secondary annealing. The effect of hot rolling process on
microstructure, texture, and inhibitor evolution is studied. It is shown that the precipitation
behavior of AlN particles is significantly affected by the hot rolling temperature and the
optimal precipitation temperature is 1100 8C. On the other hand, the intensity of Goss texture
in hot-rolled sheets is determined by hot rolling reduction, which is limited in the strip casting
process and the hot-rolled Goss texture is relatively weak compared to the conventional
process. The primary annealed Goss texture can originate from the cold rolling process and
this texture is homogeneous through the thickness, besides the inherited hot-rolled Goss
texture mainly presented in the subsurface layer. Thus, relatively strong Goss texture in
primary annealed sheets is obtained. The results indicate that the hot rolling process is an
efficient way to optimize the texture and precipitation in strip-cast silicon steel, and relatively
good magnetic properties can be obtained by the strip casting process.

1. Introduction multi-pass hot rolling and hot band annealing.[4] Later on,
two different methods were developed to decrease the
Grain-oriented silicon steel is the only one product reheating temperature.[5–9] In the first method, Cu was
manufactured in the steel industry that utilizes the added to the steel melt and formed a new type of
secondary recrystallization phenomenon.[1] The Goss Cu2S inhibitor with lower solid solution temperature, and
grains grow abnormally during high-temperature anneal- the slab reheating temperature could be decreased to
ing and lead to the easy magnetization <001> axis parallel 1200–1280 8C. In the second, the inhibitor control was
to the rolling direction. This texture characteristic enables abandoned in the hot rolling and annealing processes and
grain-oriented silicon steels to be used as core materials the main inhibitors were produced by nitriding the cold-
for transformers.[2] It is well-known that grain growth rolled sheet during decarburization annealing. The slab
inhibitors are key factors that determine whether the reheating temperature was decreased to 1150–1200 8C, and
secondary recrystallization occurs or not.[3] In conven- is the lowest reheating temperature applied in the steel
tional slab process, the slab should be reheated at high industry.
temperature greater than 1350 8C, in order to dissolve the Recently, a twin-roll strip casting process was proposed
coarse precipitates formed during continuous casting, and to manufacture grain-oriented silicon steels.[10] This
then the fine and dispersed inhibitors are formed during new advanced and energy-efficient technology can pro-
vide a thin steel strip directly from molten steel and
the multi-pass hot rolling with heavy reduction can
[
] Y. Wang, Y.-X. Zhang, X. Lu, F. Fang, G.-M. Cao, C.-G. Li, Y.-B. Xu,
be eliminated.[10–13] More importantly, due to rapid
G.-D. Wang solidification of the as-cast strip, the inhibitors are
State Key Laboratory of Rolling and Automation, Northeastern precipitated finely or in supersaturated solid solution. It
University, Shenyang 110819, China
Email: [email protected]
is not necessary to apply the high-temperature reheating
R. D. K. Misra process. The previous results reported that the micro-
Laboratory of Excellence in Advanced Steel Research, Department of structure, texture, and precipitate distribution in the
Metallurgical, Materials and Biomedical Engineering, University of
Texas at El Paso, Texas 79968, USA as-cast strip is totally different from the conventional
hot-rolled band.[14–15] In order to guarantee the surface
DOI: 10.1002/srin.201600009 quality of the as-cast strip and the final magnetic

ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 87 (2016) No. 12 1601
 www.steel-research.de
FULL PAPER

properties, a single hot rolling followed by strip casting was backscattered diffraction (EBSD). The micrographs and
needed. Song[16,17] studied the effect of hot rolling kinds of precipitates were studied using electron probe
reduction on microstructure, texture, and ductility of microanalyzer (EPMA), and precipitates were statistically
strip-cast silicon steel, and reported that the asymmetrical counted in a random selected area of 150 mm2 using
hot rolling promoted the formation of Goss texture. image analyzer (Image J). The magnetic induction at
Similarly, Li[18] studied the effect of rolling temperature 800 A m1 (B8) and iron losses at 1.7 T, 50 Hz (P1.7/50)
on texture and formability of the as-cast strip of Fe–6.5%Si were measured using a single-sheet tester in the rolling
steel. It can be seen that the previous research mainly direction of the secondary annealed specimens with
focused on the effect of hot rolling process on the 100 mm length and 30 mm width.
microstructure and texture, while the precipitation of
inhibitors and final magnetic properties were not studied.
In this study, an ultra-low carbon grain-oriented silicon
steel strip was produced by twin roll strip casting process 3. Results and Discussion
and then subjected to different hot rolling process. Next,
the hot-rolled sheets were cold rolled and annealed. The 3.1. The Effect of Hot Rolling Temperature on the
effect of hot rolling process on inhibitor precipitation, Precipitation Behavior of Second Phase Particles
Goss texture formation, and final magnetic properties was
studied in detail. Figure 1 shows micrographs of precipitates in hot-rolled
sheets with different hot rolling temperature. When the hot
rolling temperature was 900 8C, the particle density was
very low and the corresponding EDS elemental maps by
2. Experimental Section EPMA indicated that the main particles were MnS with size
of 100 nm, while the grain growth inhibitor AlN was not
A 2.0 mm-thick as-cast strip containing 0.005% C, 3.0% Si, observed. When the rolling temperature was increased to
0.027% Al, 0.2% Mn, 0.0083% N, and 0.02% S was produced 1000 and 1100 8C, two different types of particles were
using a laboratory twin roll strip caster, as described in found. The first were the coarse particles of MnS and
detail earlier by Wang et al.[14] The superheat in the melt AlN that were present in combination of size 100 nm.
pool was controlled to 30 8C in order to produce the fine- The other were new fine AlN particles of size 30–60 nm.
grained as-cast strips. Subsequent processing routes are Furthermore, the density of fine AlN particles was
presented in Table 1. The as-cast strips were reheated to significantly greater than coarse compounded particles.
900, 1000, and 1100 8C and then hot rolled in one pass with On the other hand, the size and distribution of precipitates
the identical reduction of 35%, (referred as process 1, 4, was significantly influenced by the hot rolling tempera-
and 5). Additionally, to study the Goss texture evolution, ture. As shown in Figure 2, the particle density at 1100 and
the as-cast strip reheated to 1000 8C was hot rolled with a 1000 8C in hot-rolled sheets were 21.9 and 15.4 mm2,
reduction of 15 and 25%, respectively (referred as and the corresponding average size of particle was 55.1
processes 2 and 3). All the hot-rolled strips were cold and 39.9 nm, respectively. This result indicates that higher
rolled to 0.27 mm via one-stage cold rolling process, and hot rolling temperature promotes the precipitation of AlN
then annealed at 800 8C for 5 min. The primary annealed particles.
sheet was finally heated to 1200 8C at a rate of 20 8C h1 for In order to further investigate the precipitation behavior
secondary recrystallization. of AlN particles in the tested silicon steel, a relative PTT
The microstructure and texture of hot-rolled strips and (precipitation–temperature–time) curve was calculated
primary annealed sheets were studied using electron based on the traditional nucleation growth kinetics

Hot rolling Hot rolling Cold rolling Thickness of

Process temperature reduction reduction cold-rolled Primary Secondary
number [8C] [%] [%] sheet annealing annealing

1 900 35 80 0.27 mm 800 8C for Heated from 800 to 1200 8C

5 min at a rate of 20 8C h1
2 1000 15 85
3 1000 25 82
4 1000 35 80
5 1100 35 80

Table 1. The processing parameters of strip-cast grain-oriented silicon steel.

1602 steel research int. 87 (2016) No. 12 ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.steel-research.de

FULL PAPER
Figure 1. Micrographs and EDS elemental mapping of precipitates in hot-rolled sheets with different hot rolling temperatures. a) 900 8C,
b) 1000 8C, and c) 1100 8C.

theory.[19–21] The black and red lines represented 5 and atomic diffusion played a dominant role in the precipita-
95% of the relative phase fraction of AlN precipitated at tion of AlN. Therefore, the density of AlN particles in
dislocations, respectively. (t0 indicates the product of hot-rolled sheets was increased with increase in hot rolling
temperature-independent constant, which cannot be temperature. On the other hand, the previous work
accurately calculated. However, the relative precipitation reported that MnS particles were precipitated in the
time (lg(t/t0)) that changes with temperature can be as-cast strip.[15] In view of the spontaneous decrease of
calculated and plotted.) As shown in Figure 3, the PTT Gibbs-Energy, the following precipitation of AlN during
curve exhibited a typical “C” shaped and the nose the hot rolling stage will preferentially occur at the
temperature was 1090 8C. It is known that the precipita- interface to reduce interface energy, which then gives rise
tion kinetics of AlN particles is controlled by chemical free- to the formation of the composite precipitate composed of
energy and atomic diffusion. The chemical free-energy was MnS and AlN compounds, as shown in Figure 1b and c.
decreased and activation energy for diffusion was in-
creased with increase in the hot rolling temperature. Thus,
there was an optimal precipitation temperature (nose 3.2. The Effect of Hot Rolling Reduction on the
temperature) for AlN particles. In our study, the hot-rolled Development of Goss Texture
sheets at 1100 8C had highest particle density, which is in
good agreement with the calculated results, i.e., the nose Figure 4 presents the microstructure and texture of hot-
temperature was 1090 8C. In addition, the hot rolling rolled sheets with different hot rolling reduction. The 15%
temperature was in the range of 900–1100 8C and the hot-rolled sheet was composed of coarse ferrite grains with

Figure 2. Size distribution of particles in hot-rolled sheets for Figure 3. Relative PTT curves of AlN in experimental silicon steels
different hot rolling temperatures. with dislocation nucleation mechanism.

ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 87 (2016) No. 12 1603
 www.steel-research.de
FULL PAPER

Figure 4. Orientation image maps of all orientations and Goss texture with a deviation of 158, as well as constant w2 ¼ 458 section of the
ODFs of hot-rolled sheets with different hot rolling reductions. a) 15%, b) 25%, and c) 35%.

a slight deviation from the normal direction, which is 95%, and hence an amount of Goss orientation is formed
similar to the initial solidification microstructure reported due to the severe shear deformation. However, in our
earlier.[16] When the hot rolling reduction increased from study, the thickness of the as-cast strip was 2 mm. The
25 to 35%, the microstructure was obviously refined by hot-rolled process involved only one pass and the total hot
severe deformation and recrystallization. On the other rolling reduction was limited. The intensity of Goss texture
hand, the texture of hot-rolled sheets exhibited a random was still very weak even though the hot-rolled reduction
texture, for instance, l-fiber inherited from the as-cast increased to 35%, which is unfavorable for the abnormal
strip and the deformed a-fiber with peaks at {112}<110> growth of Goss texture compared to the conventional hot
and {111}<110>. Besides, the Goss texture was formed in rolling process. However, surprisingly, a relative strong
the subsurface layer of hot-rolled sheets, and the number Goss texture was obtained after cold rolling and primary
of Goss grain was increased with increase in hot rolling annealing, as shown in Figure 5. The amount of Goss
reduction. texture was similar to the conventional primary annealed
It is well accepted that the Goss nuclei for second sheets.
recrystallization was originated in the hot rolling stage due Usually, the hot-rolled Goss texture is almost fully
to the shear deformation caused by high friction between rotated to the g-fiber during the one-stage cold rolling
the rolls and strip surface.[22] In the conventional process, process and very small Goss nuclei are survived in the
the thick slabs of 240 mm thickness are continuously hot interior of {111}<112> deformed bands. After subsequent
rolled via many passes with a total reduction exceeding primary annealing, these Goss nuclei are preferred grown

1604 steel research int. 87 (2016) No. 12 ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.steel-research.de

FULL PAPER
Figure 5. Orientation image maps of all orientations and Goss texture with a deviation of 158, as well as constant w2 ¼ 458 section of the
ODFs of primary annealed sheets. a) Process 1, b) process 2, c) process 3, d) process 4, and e) process 5.

due to the special grain boundaries.[22] Nevertheless, the grain size prior to deformation.[23,24] The microstructure of
intensity of the inherited Goss texture becomes weak to conventional hot-rolled sheets was fully refined by severe
some extent compared to the hot-rolled sheet and the deformation and dynamic recrystallization. The shear
primary annealed sheets were dominated by g-fiber with a bands were hardly formed during cold rolling and thus the
peak at {111}<112>. Furthermore, this primary annealed Goss texture mainly originated during the hot rolling
Goss texture inherited from the hot-rolled sheet was also process. However, in this study, the hot rolling reduction
distributed at the subsurface layer. However, in our study, was very limited. The relatively coarse hot-rolled grains
the cold rolling reduction was smaller than the conven- may promote the formation of shear bands though the
tional process, leading to more hot-rolled Goss texture was thickness, resulting in cold-rolled Goss texture appeared
survived at cold rolling and primary annealing. On the in the center layer of primary annealed sheets.
other hand, besides the inheritance of surface layer Goss
grains from the hot-rolled sheets, the center layer of
primary annealed sheets appeared some Goss grains as 3.3. The Effect of Hot Rolling Processes on the Secondary
well (shown in Figure 5 and circled by black-dotted line), Recrystallization and Magnetic Properties
which implied that these center layer Goss grains may have
formed during subsequent cold rolling and primary The precipitation distribution in hot-rolled sheets and the
annealing processes. It is known that Goss texture can intensity of Goss texture in primary annealed sheets had
be nucleated at the shear bands of the deformed grains and a significant effect on secondary recrystallization and
the formation of the shear bands is related to the initial magnetic properties. The effect of hot rolling process on

ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 87 (2016) No. 12 1605
 www.steel-research.de
FULL PAPER

Precipitation in Goss texture in Goss texture in primary

Process number hot-rolled sheets hot-rolled sheets annealed sheets

1 Few coarse MnS – Goss texture fraction: 1.75%

No fine AlN Distributed surface and
center layers
2 – Weak Goss texture Goss texture fraction: 1.04%
Distributed subsurface Distributed surface and
layer center layers
3 – Weak Goss texture Goss texture fraction: 1.43%
Distributed subsurface Distributed surface and
layer center layers
4 High-density fine AlN Weak Goss texture Goss texture fraction: 1.73%
Mean size: 39.9 nm Distributed subsurface Distributed surface and
layer center layers
Density: 15.4 mm2
5 High-density fine AlN – Goss texture fraction: 2.11%
Mean size: 55.1 nm Distributed surface and
center layers
Density: 21.9 mm2
Conventional Process[22,25] High-density fine MnS Strong Goss texture Goss texture fraction: 1.8%
and high-density fine
AlN after hot band Distributed subsurface Mainly distributed
annealing layer subsurface layers

Table 2. Precipitation and Goss texture distribution in hot-rolled or primary annealed sheets.

precipitation and Goss texture is summarized in Table 2. secondary annealing, leading to the absence of secondary
Figure 6 shows the macrostructure of secondary recrystal- recrystallization (shown in Figure 6a). On the other hand,
lized sheets with different hot rolling process. When the the hot rolling reduction significantly affected the second-
hot rolling temperature was 900 8C (process 1), the AlN ary recrystallization as well. When the hot rolling reduction
inhibitor was not precipitated (shown in Figure 1a), was 15% (process 2), the refinement of hot-rolled grains
the primary recrystallized grains grew normally during was inadequate. The relatively coarse grains were inclined

Figure 6. Macrostructure of secondary recrystallization annealed sheet; a) process 1, b) process 2, c) process 3, d) process 4, and
e) process 5.

1606 steel research int. 87 (2016) No. 12 ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.steel-research.de

FULL PAPER
2. As the hot rolling reduction was increased, the
microstructure was refined and the Goss texture was
increased. However, in the strip casting process, the hot
rolling reduction was limited and the hot-rolled Goss
texture was relatively weak compared to the conven-
tional process.
3. The primary annealed Goss texture was composed of
two parts: the first part originated during hot rolling and
was mainly distributed in the subsurface layer. The
other was formed during subsequent cold rolling and
was distributed homogeneously through the thickness.
4. The secondary recrystallization and magnetic proper-
ties were significantly affected by the hot rolling
process, and the optimal magnetic induction B8 of
Figure 7. Magnetic properties of secondary recrystallization strip-cast grain-oriented silicon steel was 1.85 T.
annealed sheet with different processes.

to form inhomogeneous primary annealed microstructure Acknowledgements

(shown in Figure 5b circled by black solid line) and the This work was financially supported by the National
relatively coarse primary grains grew easily and preferen- Natural Science Foundation of China (Nos. 51174059,
tially during secondary annealing due to its size advantage. 51374002, 51404155, and U1260204), the Fundamental
The secondary Goss grains cannot consume these grains, Research Funds for the Central Universities (N130407003),
resulting in an incomplete secondary recrystallization the Program for New Century Excellent Talents in
(shown in Figure 6b). By contrast, when the hot rolling University (NCET-13-0111).
reduction was 35% and the hot rolling temperature was
1000 and 1100 8C (processes 4 and 5), the inhibitor Received: January 10, 2016; Revised: March 15, 2016;
precipitation during hot rolling was suitable, and the Published online: April 4, 2016
microstructure and texture of primary annealed sheets was
favorable to abnormal growth for Goss grains. Thus, the Keywords: grain-oriented silicon steels; strip casting; hot
secondary recrystallization was completed and the size of rolling; Goss texture; magnetic properties
secondary grains up to 10–50 mm.
Figure 7 shows the magnetic properties of secondary
annealed sheets. The magnetic properties were related to References
the macrostructure of secondary recrystallization sheets.
The secondary recrystallization in processes 1, 2, and 3 was [1] T. Kumano, T. Haratani, Y. Ushigamo, ISIJ Int. 2002,
absent or incomplete, and the magnetic induction B8 was 42, 440.
1.52, 1.55, and 1.62 T, respectively. The corresponding core [2] S. M. Shin, S. K. Chang, B. C. Decooman, ISIJ Int.
loss P1.7/50 was 2.42, 2.66, and 2.41 W kg1, respectively. 2008, 48, 1788.
While, for processes 4 and 5, the secondary recrystalliza- [3] K. Jenkins, M. Lindenmo, J. Magn. Magn. Mater.
tion was completed and the magnetic induction B8 was 2008, 320, 2423.
obvious increased to 1.73 and 1.85 T. The corresponding [4] K. Iwayama, T. Haratani, J. Magn. Magn. Mater. 1980,
core loss P1.7/50 was decreased to 2.09 and 1.68 W kg1. The 19, 15.
optimal B8 was comparable to the conventional process [5] Z. S. Xia, Y. L. Kang, Q. L. Wang, J. Magn. Magn.
reported by Guo[26] (process number 0 in Figure 7). In Mater. 2008, 320, 3229.
addition, the core loss was relatively high, even though an [6] Z. S. Xia, Y. L. Kang, Q. L. Wang, J. Univ. Sci. Technol.
optimized hot rolling process was applied, which may be Beijing 2009, 31, 439.
related to the elimination of high-temperature pure [7] K. Günther, G. Abbruzzese, S. Fortunati, G. Ligi, Steel
annealing in this study. Res. Int. 2005, 76, 413.
[8] S. Mishra, V. Kumar, Mater. Sci. Eng. B 1995, 32, 177.
[9] T. Kumano, T. Haratani, N. Fujii, ISIJ Int. 2005, 45, 95.
[10] A. Girgensohn, A. R. Büchner, K. H. Tacke, Ironmak.
4. Conclusions Steelmak. 2000, 27, 317.
[11] A. Hunter, M. Ferry, Scr. Mater. 2002, 47, 349.
1. The nose temperature of AlN precipitation was [12] J. Y. Park, K. H. Oh, H. Y. Ra, Scr. Mater. 1999, 40, 881.
1100 8C in the ultra-low-carbon silicon steel, and [13] D. Raabe, Acta Mater. 1997, 45, 1137.
the particle density was increased with increase in hot [14] Y. Wang, Y. B. Xu, Y. X. Zhang, J. Magn. Magn. Mater.
rolling temperature from 900 to 1100 8C. 2015, 379, 161.

ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 87 (2016) No. 12 1607
 www.steel-research.de
FULL PAPER

[15] Y. Wang, Y. B. Xu, Y. X. Zhang, Mater. Charact. 2015, [20] H. A. Wriedt, Metall. Trans. A 1980, 11, 1731.
107, 79. [21] H. J. Wang, B. Fu, L. Xiang, J. Iron Steel Res. 2015,
[16] H. Y. Song, H. T. Liu, H. H. Lu, Mater. Sci. Eng. A 2014, 27, 40.
605, 260. [22] M. Matsuo, ISIJ Int. 1989, 29, 809.
[17] H. Y. Song, H. H. Lu, H. T. Liu, Steel Res. Int. 2014, 85, [23] J. T. Park, J. A. Szpunar, Acta Mater. 2003, 51, 3037.
1477. [24] D. Dorner, S. Zaefferer, D. Raabe, Acta Mater. 2007,
[18] H. Z. Li, H. T. Liu, Y. Liu, J. Magn. Magn. Mater. 2014, 55, 2519.
370, 6. [25] N. J. Park, H. D. Joo, J. T. Park, ISIJ 2013, 55, 125.
[19] R. Radis, S. Schwarz, S. Zamberger, Steel Res. Int. [26] W. Guo, W. M. Mao, Y. Li, Z. G. An, Mater. Sci. Eng. A
2011, 82, 905. 2011, 528, 931.

1608 steel research int. 87 (2016) No. 12 ß 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim