ON THE MAJOR FACTORS AFFECTING GOSS TEXTURE DEVELOPMENT IN GRAIN ORIENTED SILICON STEEL

Ever since Goss discovered the method of making ’grain-oriented’ silicon steel,l, a high number of explanations have been proposed. May and Turnbull2 showed that the second phase particles, particlarly MnS influence the mobility of the boundaries. Later Misra, Dfirman and Lficke3, as well as Inokuti4 found a strong correlation between the strength of the Gosscomponent measured in the subsurface of the hot band and that of the final product. Evidence was also given to indicate that the final Goss texture formation may also occur when the hot band subsurface contained no Goss at all5. Haratani and Hutchinson6, on the other hand, based on a series of sectioning experiments of silicon steel taken at various stages between primary recrystallization and subsequent abnormal grain growth, found that shear bands of {111}<112> orientation serve as nuclei for the Goss texture formation in the primary recrystallized specimen. Another series ofresearch by Harase and coworkers7 followed the idea ofAust and Rutter8, that boundaries with a high number of coincidence site lattice points migrate with a rate higher than others. First Harase et al.6 then Rouag et al.9 have observed that the formation of a well aligned Goss texture with little scatter, observed after abnormal grain growth, can be linked to the P.(9) type CSL boundaries in the primary recrystallized structure. For some time it was evident that it is not possible to accept several differring and contradictory explanations for the same final texture development, unless these theories are rather complementary then contradictory. We shall, in this paper, show that linking the hypotheses stated above can contribute to the understanding of the final texture development in GO silicon steel.


INTRODUCTION
Ever since Goss discovered the method of making 'grain-oriented' silicon steel,l, a high number of explanations have been proposed. May and Turnbull2 showed that the second phase particles, particlarly MnS influence the mobility of the boundaries. Later Misra, Dfirman and Lficke3, as well as Inokuti4 found a strong correlation between the strength of the Gosscomponent measured in the subsurface of the hot band and that of the final product. Evidence was also given to indicate that the final Goss texture formation may also occur when the hot band subsurface contained no Goss at all5. Haratani and Hutchinson6, on the other hand, based on a series of sectioning experiments of silicon steel taken at various stages between primary recrystallization and subsequent abnormal grain growth, found that shear bands of {111}<112> orientation serve as nuclei for the Goss texture formation in the primary recrystallized specimen. Another series of research by Harase and coworkers7 followed the idea of Aust and Rutter8, that boundaries with a high number of coincidence site lattice points migrate with a rate higher than others. First Harase et al.6 then Rouag et al.9 have observed that the formation of a well aligned Goss texture with little scatter, observed after abnormal grain growth, can be linked to the P.(9) type CSL boundaries in the primary recrystallized structure.
For some time it was evident that it is not possible to accept several differring and contradictory explanations for the same final texture development, unless these theories are rather complementary then contradictory. We shall, in this paper, show that linking the hypotheses stated above can contribute to the understanding of the final texture development in GO silicon steel. 819 EXPERIMENTAL A complete process-flow investigation was carried out using an industrial source material of CGO quality. The processing steps were made according to standards accepted in the silicon steel industry, applicable for 0.22 mm materials. This involved a hot rolling, a two stage cold rolling with intermediate anneal in between and after the second cold rolling a decarburization and recrystallization anneal at 800C for 2 minutes in wet hydrogen. This was followed by a 24 hour high temperature anneal at 1150C in dry hydrogen and thermal flattening. Texture measurements were made by the X-ray diffraction technique. A Co-anode, source was used to measure (110), (200) (112) pole figures by reflection (from a=0 to 80 ).
ODFs were calculated by means of series expansion using Bunge's method0.
The ODFs were then ghost corrected using the method of Liicke et al. by means of texture modeling and the corrected 'true' ODFs were used for further analysis. Texture measurements were also coupled with microstructural determination, and scanning electron microscopy (SEM) to obtain information on the second phase MnS morphology.
The present paper deals with the hot band stage, the primary recrystallized stage and with the final product. A more detalied analysis of the complete process flow will be published separately2. Material qualities, characterized as 'good' (Series 'A') and 'average' (Series 'B') are separately analyzed.
Characteristic ODFs for the hot rolled material, measured at the subsurface (at relative through thickness height S-0.75) are given in Figs la and b corresponding to 'good' (Fig. la) and 'average' (Fig lb). 'Good' and 'average' refer to qualities in the final specific power loss as measured by the standard 25 cm Epstein probes. We note a very strong Goss texture both materials.
The textures of the final products are shown in Figs. 2a and b. We note again the sharp Goss texture in both materials, although for 'A' the Goss peaks are twice as strong as for 'B'.              Table 2. Materials 5A, 5B and 9A, 9B are from two separate series of experiments.

DISSCUSSION
We have followed different approaches to explain the development of the Goss texture after abnormal grain growth. Our results indicate that both the P.(9) CSL boundaries in the primary recrystallized matrix, as well as the Goss component in the hot band subsurface influence the final texture development. Altough the attempt to link these two theories seems to be in its preliminary stage, our experiments suggest that such texture development routes are to be found where we can establish a link between the texture of the hot band subsurface and that of the primary recrystallized material. Although a coherent model accounting for intermediate, and final stages of texture development would be more appealing, it is indeed doubtful, how we can link a texture component in the hot band subsurface to the primary matrix. We also remember that the material goes through two cold rollings and two anneals in between. Nevertheless, it cannot be excluded that the inheritance of the Goss from the hot band leads to such components in cold rolling that helps to increase the CSL boundaries to the Goss orientation. Current deformation and recrystallization models cannot follow this with sufficient acccuracy, however, a series of suitable experiments, similar to those made to support3,4 and refute5 the structure memory phenomenon could also be useful.

CONCLUSIONS Io
Both hot band subsurface and primary recrystallized matrix structure seem to play a role in influencing the quality of the final abnormal grain growth texture in grain oriented silicon steel. The linkage between the hot band subsurface and the final texture, as well as that of the r.(9) CSL boundaries on the final abnormal grain growth texture development in Fe-3%Si has been shown. That these factors alone can also lead to the development of sharp Goss orientation in the final texture does not exclude the occurrance of either one of these phenomenae. The conditions that finally lead to the dominance of one or the other influences should be further investigated.