Transformations and Microstructures
Transformation Delay and Texture Memory Effect of Columnar Grained Cast Slab in Low Grades Non-oriented Electrical Steels
2021 Volume 61 Issue 5 Pages 1669-1678
Details
2021 Volume 61 Issue 5 Pages 1669-1678
The cast slab of low grades non-oriented electrical steels experiences twice diffusional solid phase transformation and demonstrates the features of strong morphological memory referring to columnar structure and texture memory referring to the preferred <100> texture at room temperature. In this paper, electron backscatter diffraction (EBSD), quasi-insitu observation of heating samples, and dilatometry are used to study and analyze the two kinds of memory phenomena. The results show that the cast slab consists of about 70% coarse columnar grains and 30% small equiaxed grains. Many small equiaxed grains show Σ3 misorientations with columnar grains indicating K-S orientation relationship obeyed during phase transformation. Coarse columnar grains show a typical <100>||growing axis orientation which is the same as solidified columnar grains. The quasi-insitu observation shows that transformation of columnar grained ferrite to austenite is very sluggish and columnar grained ferrite can still be seen even at a superheating degree of 176°C for 1 hour. Dilatometry measurement indicates that the starting transformation temperatures for a columnar-grain-dominant sample and a small-equiaxed-grain-dominant sample are similar, whereas their transformation extents are quite different with columnar grained sample showing a low dilatational amount due to insufficient transformation. It is most likely that the coarse columnar grains in cast slab are retained and untransformed high temperature δ-ferrite are not subjected to twice complete transformations. These retaining columnar grains in low grades of electrical steels can be used to improve magnetic properties through optimizing processing parameters.
Non-oriented electrical steel is mainly used for iron core materials in electrical machines that transform electrical energy into mechanical energy. The most ideal texture for this soft magnetic functional material is {100} texture to reach the best performance. Non-oriented electrical steel can be classified into low grades non-oriented electrical steels involving phase transformation (referring to the products of 1300 and 800 grades) and medium and high grades non-oriented electrical steels without phase transformation (referring to the products of 600 grades and above), according to whether there is solid phase transformation during the solidification. The 1300 and 800 grades are the most widely used in industry. Although there are many favorable {100} columnar grains in continuous cast slabs, the products of low grade electrical steels are often composed of strong deleterious {111}<112> texture and of a fine grain size, thus leading to a poor magnetic properties. The columnar grain and the {100} orientation always can be found after the δ(δ-ferrite)→γ(austenite)→α(α-ferrite) transformation, and it indicates that the morphological memory and textural memory phenomena appear. However, the favorable initial structures and textures disappear completely after industrial hot rolling, cold rolling and annealing. From this we can raise the following two academic and applicational questions. First, why do not the columnar grain and {100} texture disappear during the twice phase transformations without external stress during continuous cast slab solidification? In other words, what is the mechanism of morphological and textural memory without external stress after δ→γ→α phase transformation, in which there is a 300°C austenite phase zone. The second is that since the hot and cold deformation will destroy the columnar grain and the favorable {100} texture, so is there any way to offset the unfavorable effect of the deformation? That is, how can we effectively retain the favorable {100} texture, so as to improve the final magnetic properties?
There have been many reports on the phase transformation treatment during final annealing to obtain the favorable {100} columnar grain structure in non-oriented electrical steel.1,2,3,4,5,6,7,8,9,10,11) Moreover, the textural memory phenomena of phase transformation, the K-S orientation relationship and the variant selection phenomena have also been widely observed in various steels,12,13,14,15,16,17,18,19,20) and the mechanism of their formation is considered to be based on grain boundary nucleation and double K-S (Kurdjumov-Sachs) relationship.20,21,22,23) However, it is rarely reported how to retain {100} textures in hot rolled sheets. We have observed in previous studies1) that it is difficult to obtain a strong {100} texture in commercial low grades electrical steel after hot rolling, cold rolling and phase transformation annealing, though the cast slabs exhibit {100} columnar grain. Therefore, this article mainly attempts to solve the two problems mentioned above by studying the textural memory behavior of the continuous cast slab after the δ→γ→α phase transformation and the reheating α→γ→α phase transformation, and provides the theoretical basis for the subsequent optimization of the microstructure and texture after hot rolling, cold rolling and annealing.
The continuous cast slab of industrial low grade non-oriented electrical steel was selected as the starting material. The chemical composition (wt%) was: 0.002% C, 0.35% Si, 0.13% Mn, 0.14% P, 0.18% Al, and Fe balance.
The critical transformation temperatures during phase transformation were measured with the DIL805A dilatometer, and the size for the test samples is φ 4 mm × 10 mm cut from the continuous cast slab. During the measurement, the sample is heated up at a rate of 200°C/h up to 1150°C and then cooled down to room temperature at a rate of 300°C/h. The critical temperatures were Ac1=980°C, Ar1=938°C, Ac3=1090°C, Ar3=973°C. The heating and cooling rates were the same as above unless otherwise specified for other measurements. The corresponding phase diagram and theoretical phase transformation temperatures were calculated using the Thermo-Calc software according to the alloy composition. The equilibrium phase transformation temperatures are A1=927°C and A3=1004°C. Three kinds of dilatation sample were measured. The columnar-grain-dominant sample was cut from the continuous cast slab, while the equiaxed-grain-dominant sample was cut from the cast slab which was subjected to reheating annealing at 1180°C for 30 min, and the third rough-rolled sample was cut from a 40 mm thick rough-rolled slab (the grain size is 63 μm).
In order to study the phase transformation behavior of the columnar grained samples, quasi-insitu method was used to characterize the microstructure of cast slab during thermal cycling. The three cuboid initial samples of 20 mm (rolling direction, RD) × 10 mm (normal direction, ND) × 8 mm (transverse direction, TD) cut from continuous cast slabs were polished and etched with 4% nital for 40 s, and then subjected to EBSD detection. Subsequently, samples were heated in the tube furnace at 1100°C for 7 min, 1100°C for 30 min or 1180°C for 60 min. The annealing temperatures are higher than the A3 phase transformation temperature by 96°C or 176°C, respectively. Finally, the samples were cooled to room temperature at the cooling rate of 300°C/h. During annealing process, hydrogen atmosphere with a flow rate of 8 L/min was supplied to prevent surface oxidation. Three cuboid annealed samples were polished and etched with 4% nital for 40 s, and then subjected to EBSD detection. The different heating temperatures and holding times are used to investigate the extent of transformation of columnar grained samples.
The orientation data were obtained by a ZEISS-ULTRA55 scanning electron microscope (SEM) equipped with an electron back-scatter diffraction (EBSD) Channel 5 orientation analysis system. Local misorientation extracted from the EBSD maps shows the average local misorientation for misorientations below the subgrain angle and can be used to quantify strain-induced changes. Here, subgrain angle was set as 5° and filter size was 3×3. It is well known that the columnar grains are very coarse, and the measured EBSD area of about 4 mm2 is still smaller than the size columnar grains. Therefore, in order to increase the measured region, a large area is obtained by combination of several EBSD maps. In this work, the maximum allowable deviation angle to a typical texture is within 15°.
Figure 1 shows the theoretically calculated phase diagram. According to the alloy composition, the corresponding phase diagram and equilibrium phase transformation temperatures are calculated using the Thermo-Calc software, and the characteristics of the non-equilibrium solidification and solid phase transformation is analyzed. The dashed line in Fig. 1 marks the theoretical equilibrium phase transformation temperature corresponding to Fe-0.35%Si. The A1 temperature is determined as 927°C and A3 is 1004°C. Comparatively, the measured transformation temperature by the dilatometer are A1=959°C, A3=1031°C. Thus, the difference values between the measured A1 and A3 and the theoretical values are 32°C and 27°C, respectively. Considering the heating and cooling speeds used, the difference between theoretical and measured values is understandable. The phase diagram in Fig. 1 shows that the alloy experiences δ→γ→α transformation after cooling from the liquid, and the stable austenite phase region is about 300°C. Moreover, the twice phase transformations are both diffusional phase transformation. Besides the diffusion of interstitial atoms C, the diffusion of substitutional alloy elements Si and Mn also occurs.
The theoretically calculated phase diagram of low grade electrical steel. (Online version in color.)
The continuous cast slab with 230 mm thickness is mainly composed of symmetrical coarse columnar grains, and the small black dots inside the columnar grains are equiaxed grains formed by phase transformation.1) This phenomenon implies the occurrence of morphological memory, i.e. the columnar grain is retained from high temperature δ-ferrite after δ→γ→α phase transformation during cooling. Figure 2 shows the EBSD data of the continuous cast slab in area 1 composed of 80% columnar grains and 20% equiaxed grains. The columnar grains are nearly <100>//ND, which is the same as the columnar grains formed by solidification. The inside of the columnar grains contains small non-{100} (<100>//ND) equiaxed grains, as shown in Fig. 2(a). The small equiaxed grains are mainly nucleated inside the columnar grains rather than at the grain boundaries, indicating that the tilting-type columnar grain boundaries have no significant nucleation advantage. The local texture in Figs. 2(b) and 2(c) is mainly near-cubic texture, and there are many Σ3 misorientation between the columnar grains and the small equiaxed grains as shown in Fig. 2(d) with the red line (the red line refers to Σ3 grain boundary). This suggests that the phase transformation during cooling obeys K-S orientation relationships, because there is no Σ3 misorientation between variants in N-W (Nishiyama-Wassermann) or other orientation relationships. The small equiaxed grains inside the columnar grains are mainly the rotated cube (<100>//ND) orientation, the {111} (<111>//ND) orientation and the nearly Goss orientation as shown in Figs. 2(e) and 2(f).
EBSD data of the cast slab (Area 1): (a) IPF-X map; (b–c) the {100} pole figure and φ2 = 45° section ODF of columnar grains; (d) the band contrast (the red line refers to Σ3 grain boundary); (e–f) the {100} pole figure and φ2 = 45° section ODF of equiaxed grains. (Online version in color.)
Figure 3 shows the variant orientations which are transformed from cube oriented austenite according to the K-S orientation relationship. According to the Figs. 3(a) and 3(b), one group of 8 variants are close to the rotated cube orientation, another group of 8 variants are nearly Goss orientation, and the third group of 8 variants are close to rotated Goss orientation. Theoretical calculations24) for martensite transformation show that the phase transformation strain induced by the formation of two variants with Σ3 relationships is only 33% of that induced by single variant formation, and the two variants with Σ3 misorientation are most likely to appear.
The possible orientation distribution schematics of 24 ferrite variants transformed from cubic austenite according to the K-S relationship: (a) schematic representation of the parent cube orientation, and the 24 corresponding K-S variants in {100} pole figure ; (b) corresponding φ2 =45° section ODF.
Figure 4 shows the EBSD data of the continuous cast slab in area 2 composed of 70% columnar grains and 30% equiaxed grains. It can be seen from Figs. 4(a)–4(f) that the structures and textures of area 2, similar to area 1, are columnar grains with <100>//ND and some non-{100} equiaxed grains inside columnar grains. Besides, the newly formed small equiaxed grains and columnar grains mostly conform to the Σ3 relationship. The misorientations between the small equiaxed grains and the coarse columnar grains in Fig. 4(a) are expressed as axis-angle pairs as listed in Table 1, and 70% of the small equiaxed grains and the columnar grains are consistent with the strict Σ3 (<111>60°) misorientation relationship. It should be noted that there are many subgrain boundaries inside {100} columnar grains as shown with arrows in Fig. 4(d).
EBSD data of the cast slab (Area 2): (a) IPF-X map; (b–c) the {100} pole figure and φ2 = 45° section ODF of columnar grains; (d) the band contrast (the red line refers to Σ3 grain boundary); (e–f) the {100} pole figure and φ2 = 45° section ODF of equiaxed grains. (Online version in color.)
| Misorientation expressed in axis-angle pairs | ||||
|---|---|---|---|---|
| <-1-11>60° | <-11-1>59.29° | <-3-34>59.12° | <2-4-1>50.6° | <142>40.19° |
| <111>59.98° | <1-1-1>59.25° | <-3-34>59.02° | <-32-2>50.17° | <-42-3>39.08° |
| <11-1>59.98° | <-1-1-1>59.13° | <43-3>58.61° | <-2-3-3>48.84° | <133>38.73° |
| <-1-1-1>59.91° | <-1-11>59.12° | <-3-34>58.45° | <-3-3-2>48.45° | <41-2>35.63° |
| <-1-11>59.87° | <11-1>59.07° | <-343>57.13° | <-1-1-1>50.05° | <334>35.6° |
| <111>59.87° | <-1-1-1>59.05° | <344>56.53° | <-1-1-1>46.15° | <101>35.44° |
| <111>59.86° | <1-1-1>59.01° | <-34-3>56.41° | <-3-4-4>45.69° | <-3-4-1>33.94° |
| <11-1>59.76° | <-1-11>58.54° | <1-11>56.31° | <-443>45.32° | <1-4-4>33.49° |
| <1-11>59.7° | <11-1>58.27° | <1-1-1>56.1° | <-1-1-1>45.14° | <101>32.94° |
| <-1-11>59.66° | <-1-11>58.2° | <1-3-3>55.41° | <2-12>49° | <0-13>32.54° |
| <-11-1>59.63° | <1-1-1>58.17° | <2-3-3>55.41° | <-122>48.21° | <-1-3-1>32.15° |
| <-11-1>59.63° | <11-1>58.08° | <-1-44>55.02° | <-1-2-2>47.2° | <-2-43>30.39° |
| <-11-1>59.6° | <-11-1>58.06° | <313>58.95° | <-33-1>46.56° | <141>29.62° |
| <11-1>59.6° | <11-1>57.39° | <-1-3-3>59.31° | <-123>44.93° | <132>28.34° |
| <-1-11>59.58° | <33-4>60.33° | <-1-4-4>58.71° | <041>20.28° | <132>27.38° |
| <-1-1-1>59.53° | <34-4>60.31° | <31-3>57.43° | <-103>19.04° | <132>26.01° |
| <111>59.51° | <334>60.12° | <-1-2-2>59.36° | <-101>18.69° | <132>25.13° |
| <-11-1>59.46° | <44-3>60.08° | <1-1-2>50.33° | <1-11>55° | <1-4-3>22.93° |
| <-1-11>59.41° | <-4-3-4>60.08° | <-134>55.99° | <-10-1>54.23° | <4-31>20.28° |
| <-1-1-1>59.4° | <-4-43>59.9° | <-14-3>53.39° | <03-4>54.1° | <-4-3-2>20.15° |
| <-1-11>59.39° | <4-33>59.44° | <41-3>52.84° | <-4-33>52.72° | <3-1-3>10.79° |
| <-1-1-1>59.36° | <-3-34>59.25° | <-1-34>52.35° | <043>44.42° | <421>4.86° |
| <-1-1-1>59.35° | <334>59.18° | <133>51.91° | <4-1-1>42.76° | |
According to the two groups of measured results (the third group is given in the reference,1) involving 10 coarse columnar grains), it is preliminarily considered that the coarse columnar grains in industrial cast slab are mainly untransformed high temperature δ-ferrite. That is to say, the transformation is very insufficient during δ→γ→α phase transformation. In addition, the characteristics of clustered {100} subgrains in columnar grains are slightly different from columnar grains of medium and high grades without phase transformation.
3.3. Microstructure Characteristics of α→γ→α Phase Transformation during ReheatingIn order to investigate whether there is a textural memory phenomena during the α→γ→α phase transformation, the cast slab was annealed above the measured Ac3 phase transformation temperature and the transformation behavior was analyzed by quasi-insitu method.
Figure 5 shows the quasi-insitu EBSD data of the cast slab before and after reheating annealing at 1100°C for 7 minutes. The observed area contains 6 columnar grains in Figs. 5(a) and 5(b), each of which has some fine equiaxed grains. Moreover, it can be seen from Fig. 5(f) that about 50% of the small equiaxed grains have the Σ3 misorientation with the surrounding columnar grains. Actually, the sample has been re-polished after phase transformation annealing in Figs. 5(e) and 5(f), so the grain sizes are not completely consistent with initial ones and new polishing grains may also appear. This suggests that the new grains are not all produced by phase transformation. The initial columnar grains and orientations in Fig. 5(a) are retained in Fig. 5(e), and many equiaxed grain are newly formed inside the columnar grains, while the newly formed small equiaxed grains mostly conform to the Σ3 relationship. This shows that many low energy Σ3 relationships can be formed during α→γ→α phase transformation only when the variants selection following the K-S relationship and the significant phase transformation strain affects the variant selection. The new small equiaxed grains are formed inside the columnar grains, rather than at the grain boundary. In other words, the new small equiaxed grains are easy to nucleate but difficult to grow up during phase transformation. Furthermore, the newly formed small equiaxed grains are mostly {111} and {110} orientation, among them the {111} <112> oriented grains are dominant, as shown in Figs. 5(i) and 5(j). There are few {100} oriented grains in the newly formed grains, which is different from that in initial cast slab, but the orientations of the initial equiaxed grains are basically retained after phase transformation as shown in Figs. 5(c), 5(g) and 5(i). Theoretically, the initial equiaxed grains have been subjected to α→γ→α phase transformation, but the orientations are essentially unchanged as shown in Figs. 5(c)–5(d) and 5(g)–5(h), which indicates that the textural memory occurs.
EBSD data of the cast slab before and after reheating annealing at 1100°C for 7 minutes: (a–d) IPF-Z map, grain boundary distribution map (the red line refers to Σ3 grain boundary), {100} pole figure and φ2 = 45° section ODF before annealing, respectively; (e–h) IPF-Z map, grain boundary distribution map (the red line refers to Σ3 grain boundary), {100} pole figure and φ2 = 45° section ODF after annealing, respectively; (i–j) the {100} pole figure and φ2 = 45° section ODF of newly formed equiaxed grains. (Online version in color.)
Figure 6 shows the quasi-insitu EBSD data of the cast slab before and after reheating annealing at 1100°C for 30 minutes. After prolonging the holding time, the columnar grains of the initial cast slab can be also retained in Figs. 6(e) and 6(f), and the main orientation did not change before and after reheating annealing as shown in Figs. 6(c)–6(d) and 6(g)–6(h), while more equiaxed grains are formed inside the columnar grains compared with Fig. 5(e). However, even though the small equiaxed grains can successfully nucleate and grow up, the proportion of small grains is less than 40%. Furthermore, it can be seen from Fig. 6(g) that the small equiaxed grains are also mainly composed of {110} and {111} textures, and the nucleation sites of equiaxed grains are mainly located inside the columnar grains, which suggest that the columnar grain boundaries cannot effectively promote nucleation.
EBSD data of the cast slab before and after reheating annealing at 1100°C for 30 minutes: (a–d) IPF-Z map, grain boundary distribution map (the red line refers to Σ3 grain boundary), {100} pole figure and φ2 = 45° section ODF before annealing, respectively; (e–h) IPF-Z map, grain boundary distribution map (the red line refers to Σ3 grain boundary), {100} pole figure and φ2 = 45° section ODF after annealing, respectively. (Online version in color.)
Figure 7 shows the average local misorientation distribution of the cast slab before and after reheating annealing at 1100°C for 30 minutes, which is merged by 18 EBSD maps. It can be seen from Fig. 7(a) that there are many subgrain boundaries inside the columnar grains. The austenite grains are prefer to nucleate on these subgrain boundaries during phase transformation, thus there are many austenite grains inside columnar grains. Compared with Figs. 7(a) and 7(b), the stress distribution near the subgrain boundary before reheating is not uniform, but the stress distribution near the subgrain boundary after reheating is relatively uniform. It suggests that the nucleation on subgrain boundaries will alleviate the phase transformation strain, while the compressive stress near subgrain boundary is large which is unbeneficial to nucleation of austenite and phase transformation.
The average local misorientation distribution of the cast slab before and after reheating annealing at 1100°C for 30 minutes: (a) before reheating; (b) after reheating. (Online version in color.)
Figure 8 shows the quasi-insitu EBSD data of the cast slab before and after reheating annealing at 1180°C for 60 minutes, and the annealing temperature is higher than the A3 equilibrium phase transformation temperature by 176°C. Comparing Figs. 8(c) and 8(g), the equiaxed grains with {111} and {110} texture are formed inside columnar grains after annealing. From Figs. 8(b) and 8(f), the newly formed small equiaxed grains mostly conform to the Σ3 relationship. The newly formed small equiaxed grains almost replace the initial columnar grains, but the initial {100} columnar grains is still partially retained (indicated by the black arrow in Figs. 8(a) and 8(e)). It implies that the small 8 mm-thickness initial cast slab still contains original columnar grains after phase transformation annealing at 1180°C for 60 minutes. Moreover, the large initial non-columnar grains have also been refined after phase transformation, and the main texture deviates slightly after reheating annealing as shown in Figs. 8(d) and 8(h).
EBSD data of the cast slab before and after reheating annealing at 1180°C for 60 minutes: (a–d) IPF-Z map, grain boundary distribution map (the red line refers to Σ3 grain boundary), {100} pole figure and φ2 = 45° section ODF before annealing, respectively; (e–h) IPF-Z map, grain boundary distribution map (the red line refers to Σ3 grain boundary), {100} pole figure and φ2 = 45° section ODF after annealing, respectively. (Online version in color.)
From the quasi-insitu measurement of three different temperatures and holding times in the section 3.3, it can be summarized that the phase transformation of the large columnar grains is very sluggish, which makes the initial structure easy to retain. Figure 9 shows the phase transformation temperatures determined by dilatometry. The axis direction of the cylinder sample is parallel to normal direction of the columnar grain. Two kinds of sample are measured. One is the original columnar grain dominant samples taken directly from cast slab, and the other is the sample with equiaxed grains as shown in Fig. 8(e). From the Fig. 9(a), it can be seen that the phase transformation range of the columnar grain samples during heating is only 7°C, which is smaller than the equiaxed grain samples (51°C). In terms of the expanding amount, the total expansion of the equiaxed grain sample is about 25 μm, while that of the columnar grain sample is about 16 μm, which is only 64% of that for the equiaxed grain. For the equiaxed grain samples, the grain size of initial structure is 238 μm, while the grain size is only 183 μm after dilatometry test, and the refinement is due to the phase transformation. Figure 9(b) shows the phase transformation temperature of the rough-rolled sheet at a heating rate of 10°C/min and a cooling rate of 180°C/min (the corresponding structure is of uniform equiaxed grains). The total expanding length is 35 μm. Therefore, the expanding length of the previous equiaxed grain sample is 74% of that for the rough-rolled sample. It can be concluded that the difference of dilatation is due to the different microstructure.
The phase transformation temperature measured by dilatometry: (a) columnar grains sample and equiaxed grains sample; (b) rough-rolled sample.
Figure 10 shows the cross-section microstructure and EBSD data corresponding to columnar-grains-dominant sample and equiaxed-grains-dominant sample in Fig. 9(a). It should be noted that the direction perpendicular to the cross section is the normal direction of the initial columnar grains, and the EBSD map has a mirroring relationship with the optical micrograph due to the scanning strategy in EBSD mapping. The Figs. 10(a)–10(e) demonstrates that there are still residual columnar grains after α→γ→α phase transformation. Moreover, the area without phase transformation contains many small dot-like particles resembling inclusions or defects, which indicates that those area have poor resistance to etching. The small equiaxed grains region in equiaxed grain sample (Fig. 10(f)) is larger than that in columnar grain sample (Fig. 10(a)), but there still exist the regions without phase transformation (those areas corresponding to poor surface quality). In a word, the difference of the microstructure, whether it experiences phase transformation or not, is the grain size and the defects within the grains. Therefore, it is speculated that the short holding time and fast cooling rate make it difficult for the austenite nucleated at the subgrain boundary to swallow up the rest of columnar grains and grow up. Because of the large grain size, the columnar grains reach the conditions for phase transformation in thermodynamics, but not in kinetics, so only a part of columnar grains experiences α→γ→α phase transformation. Finally, the morphological structure and texture of the columnar grains are retained after phase transformation. Note that Ac3 and Ar3 used in Fig. 9(a) are not the real definition of the phase transformation point, because most of the large ferrite columnar grains have not transformed into austenite. The definition of Ac3 is the temperature at which austenite transformation is completed, and Ar3 is the temperature at which ferrite transformation starts in 100% austenite structures.
The cross-section microstructure and EBSD data corresponding to columnar grain sample and equiaxed grain sample: (a–e) cross-section microstructure, IPF-Z map, the band contrast (the red line refers to Σ3 grain boundary), and the {100} pole figure of columnar grain sample, respectively; (f–j) cross-section microstructure, IPF-Z map, the band contrast (the red line refers to Σ3 grain boundary), and the {100} pole figure of equiaxed grain sample, respectively. (Online version in color.)
Firstly, more than 70% of the columnar grain microstructure and {100} texture are retained in industrial continuous cast slab. The quasi-insitu measurement and the difference of phase transformation extents between columnar grain sample and equiaxed grain sample measured by dilatometry have confirmed that phase transformation of the coarse columnar grain is very sluggish and it is difficult for the columnar grains to be swallowed up by small equiaxed austenite grains. Therefore, one possible mechanism is proposed that the retained columnar grains in cast slab are the high temperature δ-ferrite. The remaining small equiaxed grains (about 30%) experience twice phase transformations, and many of them contain Σ3 variant selection relationships, which indicates that the effect induced by phase transformation strain is significant. It should be noted that the grain boundary is generally the main nucleation site for diffusional phase transformation, while the columnar grain boundary in continuous cast slab is a tilting-type low-energy grain boundary, so it is not an effective nucleation site. Actually, the effective nucleation site is the subgrain boundary or inclusion interface. The small austenitic grains formed inside the coarse columnar grains are greatly compressed by the surrounding columnar grains due to the volume expansion of austenite and coarse ferrite grains during heating and holding, so it is difficult to effectively swallow up the columnar grains. It should be noted that there are many subgrain boundaries inside {100} columnar grains as shown with arrows in Fig. 4(d) and the observed subgrain structure looks equiaxed. The existence of equiaxed subgrains indicates that these interfaces of subgrains are very stable during heating and are not easy to disappear, or it is difficult for ferrite grains to grow at high temperatures. If the equiaxed subgrains are the products of phase transformation, the traces of phase transformation can be seen after heating for different time at high temperature which is much higher than the phase transformation temperature, but it is not. The reason for their stable existence is that the low angle grain boundaries between {100} oriented grains may be a low-energy interfaces, which are not easy to migrate and do not promote grain boundary transformation and nucleation. The subgrains may be produced by liquid erosion during high temperature continuous casting (as reported in reference25,26,27,28,29)) or by small reduction of casting roll during continuous casting.
As the observed Σ3 relationships are between the initial orientation and the orientation after twice phase transformation. Therefore, the possible mechanism to explain the relationship between the retained coarse {100} columnar grains without phase transformation and the small equiaxed grains after twice transformation is as follows: the twice phase transformation follow the K-S relationship. If there is no K-S relationship between δ-ferrite and austenite or between austenite and α-ferrite, it cannot be explained how can the Σ3 relationship be held between α-ferrite and δ-ferrite. Therefore, austenite has to have K-S relationship with δ-ferrite and transform to α-ferrite with K-S relationship again to cause the Σ3 boundaries, which can cancel out the transformation strain by a nucleation of equiaxed ferrite in the vicinity of growing δ-ferrite.
The columnar grains in the room temperature cast slab are similar to the solidified columnar grains. Since the composition is close to pure iron and the diffusion ability of elements is significant (mainly the self-diffusion of Fe), and the range of austenite phase region in equilibrium is about 300°C, a remaining question is why twice complete solid phase transformations have not occurred. It is easy to associate it with that the columnar grains in room temperature cast slabs are the high temperature δ-ferrite. Otherwise, it needs to be explained how the high temperature δ-ferrite columnar grains transform into austenite, what shape and orientation austenite has and how these austenitic grains transform into α-ferrite columnar grains. It can be seen in quasi-insitu reheating experiments that the phase transformation of the initial columnar grains is very sluggish. Therefore, it is reasonable to infer that there is no phase transformation in most areas of the initial industrial continuous cast slab during nearly unidirectional cooling condition. Furthermore, the widely observed columnar grain boundary is difficult to promote nucleation, which also indicates that the columnar grains are not easy to phase transformation. A reason may be the expanding stress. The another problem is that the small equiaxed grains transformed from initial columnar grains have shown the textural memory phenomena during reheating and cooling, whether did they transform into {100} oriented columnar grains? It is suggested that a part of equiaxed grains can transform into {100} orientation. It can be seen in Fig. 4(f) that the small grains have transformed back into {100} oriented grains. Therefore, the double textural memory phenomenon with twice K-S relationship still exists.
4.2. The Industrial Application Value of the Retention Phenomenon in Low-grade Electrical SteelThe columnar grains in continuous cast slabs are generally considered to be deleterious, because columnar grains can cause significant inhomogeneity of grain size and surface quality problems such as corrugated defect or black lines.30) In electrical steel without phase transformation, the electromagnetic stirring is usually used to reduce the proportion of columnar grains region, so as to eliminate their harmful effects. Although these methods have eliminated the harmful effects, the favorable effects of columnar grain cannot be utilized. After elucidating the evolution of columnar grains and textures, the favorable texture should be controlled to be retained to the maximum extent, and the appearance of surface quality defects should be overcome, which requires lots of detailed study. In low-grade non-oriented steel with phase transformation, cast slabs are rolled from about 230 mm to 2.5 mm with a reduction of nearly 99%, which experience twice phase transformations. Therefore, the columnar grains seem to have no effect at all because of the heavy hot rolling and the subsequent cold rolling, but in fact, the favorable texture has been retained to some extent.1) However, even though the hot-rolled sheet mostly contains favorable {100} and {110} grains, a strong {111} texture is formed after cold rolling and annealing due to the fine grain size in the hot rolled sheets, and the proportion of {111} grains is more than 40%. Therefore, the main challenge for the 1300 and 800 low grades non-oriented steels of largest production is to retain the coarse columnar grains after hot rolling. Because of the corrugated defects caused by the columnar grain, long time heat preservation and heavy reduction are often used in industries to make the microstructure uniform, so that the favorable factors in the initial cast slab cannot be effectively utilized.
According to the above results, it is expected to effectively retain the advantages of columnar grains in hot-rolled sheet, so as to ultimately improve the final magnetic properties, by means of coarsening the columnar grains during solidification, reducing the cast slab thickness, and heating the cast slab at short time and low temperature. Therefore, it can be envisaged that the industries can use a method of rapid and short time heating at 1100°C or lower temperature to hot rolling (the industrial hot rolling is generally around 1150°C with long holding time) and then use rapid cooling, so that the {100} ferrite may be retained, and the final magnetic properties are optimized. This work will be reported in future study.
The microstructure in the Fe-0.35Si low-grade electrical steel has been investigated by quasi-insitu EBSD observation in the samples annealled at different temperature, dilatometry measurement and thermodynamic calculation. The main results of the current study are summarized as follows:
(1) The Fe-0.35%Si low-grade electrical steel cast slab is composed of approximately 70% coarse <100> columnar grains and 30% fine equiaxed grains. The morphological memory and textural memory phenomena can be due to the retention of the initial high temperature δ-ferritie. There are many small subgrains inside the columnar grains, which also belong to the textural memory. Furthermore, there is a difference between the microstructures that have experienced phase transformation and those without phase transformation.
(2) Due to the large size and stress hindrance, the high temperature δ-ferrite columnar grains were not swallowed up by the new small austenite grains during phase transformation. Many small equiaxed ferrite grains show Σ3 misorientations relationship with columnar grains, which implies that the phase transformation obeys the K-S relationship and influence of phase transformation strain.
(3) The results of reheating α →γ→ α diffusional phase transformation show that the interior of the columnar grains is the main nucleation site rather than the low energy columnar grain boundaries. Moreover, the grain size significantly affects the phase transformation kinetics resulting in the insufficient transformation.
This work was financially supported by the National Natural Science Foundation of China (Grant No. 51771024).
