Microstructural Evolution during Industrial Rolling of a Duplex Stainless Steel

Transcription

1 , pp Microstructural Evolution during Industrial Rolling of a Duplex Stainless Steel G. FARGAS, 1) N. AKDUT, 2) M. ANGLADA 1) and A. MATEO 1) 1) Dept. Ciència dels Materials i Enginyeria Metal.lúrgica, Avda. Diagonal 647, Universitat Politècnica de Catalunya, 08028, Barcelona, Spain. antonio.manuel.mateo@upc.edu 2) OCAS N.V./Arcelor Research Industry Ghent, ArcelorMittal Group, John Kennedylaan 3, B-9060, Zelzate, Belgium. (Received on April 16, 2008; accepted on August 4, 2008) The microstructural evolution of a duplex stainless steel EN during the industrial rolling process has been examined. The ferritic and austenitic phases were analyzed separately after each rolling and annealing step. Optical microscopy and transmission electron microscopy, together with texture measurements, have been used to characterize the morphology and the preferred crystallographic orientations of both phases. Additionally, the hardness evolution in ferrite and austenite after each industrial step was measured by instrumented indentation. The results show that microstructural features observed after hot rolling tend to remain during further processing and they are essentially present in the final cold rolled and annealed product. KEY WORDS: duplex stainless steel; rolling; annealing; texture; anisotropy. 1. Introduction Duplex stainless steels (DSS) have a microstructure where ferrite and austenite are present in approximately equal amounts. The industrial interest on this stainless steel grade arises from its attractive combination of mechanical properties and corrosion resistance, which are better than predicted by the law of mixture. As a consequence, DSS exhibit an overall performance unattainable neither for austenitic nor for ferritic stainless steels. 1 5) Due to their properties, DSS today are being mainly used in the off-shore, chemical and petrochemical industries, e.g. for marine constructions, ships outer skins, pipelines and power plants. 6 9) DSS are mainly used in form of flat products but cast structures for e.g. off-shore applications exist as well. The processing route to manufacture DSS sheets generally consists in alternative steps of rolling and annealing, as presented in Fig. 1. After continuous casting, the slab is hot rolled in several passes in order to attain large thickness reduction. Then, the material is annealed to recover the ductility for the subsequent cold rolling up to the required final thickness. The process ends with a final annealing which gives the steel product the appropriate characteristics for future applications. During these industrial process steps the steel undergoes microstructural changes, which affect the mechanical properties of the final product. Since during rolling both phases align in the rolling direction (RD), the microstructure becomes morphologically anisotropic and even marked crystallographic textures can appear. As a result, their mechanical properties are strongly direction dependent 10 12) and this often gives raise to problems for sheet forming operations like deep drawing. Therefore, the study of the microstructural changes and texture evolution during the different stages of the rolling process is important in order to understand their influence on the properties of the final industrial product. Both, the hot and the cold working behaviour of the steel have to be analysed, together with the investigation of the restoration mechanisms active during the annealing treatments. The hot forming behaviour of austenitic and ferritic stainless steels has been widely characterized mainly by means of hot torsion and compression tests. 13,14) From these studies, it is well known that ferrite and austenite behave differently during plastic deformation. During hot working the two major mechanisms of restoration of metals are dynamic recovery (DRV) and dynamic recrystalization (DRX), whereby the stacking fault energy (SFE) dictates which mechanism will take place. While single-phase austenitic stainless steels can undergo DRX due to their low SFE, in ferritic stainless steels, which have a much higher Fig. 1. Schematic sketch of the industrial processing route for (duplex) stainless steels. 1596

2 SFE than austenite, DRV is more likely to occur. However, in the latter case, conditions that hinder DRV and favour the onset of DRX are also possible. The simultaneous presence of both phases in DSS makes the situation more complex. There are only a few studies in literature dealing with high temperature deformation of DSS in which controversial results about the austenite behaviour are presented. Some authors indicate that DRX takes place 15,16) but others suggest that this mechanism is inhibited in these two-phase alloys ) Concerning ferrite, Cizek and Wynne 20) described its softening as an extended dynamic recovery, while other authors called it continuous recrystalization. 21) The difference between these two mechanisms is defined in terms of the relative amounts of low and high angle boundaries in the microstructure. 22) Another factor to consider is that during deformation, both at high temperature and at room one, the strain partitioning between the two phases is not homogenous. Mainly it is influenced by many phenomena such yield point and strain hardening of each constitutive phase, volume fraction and phase morphology. A recent study concerning hot workability demonstrates that ferrite accommodates three to five times more strain than austenite. 23) This result was obtained by measuring the length, width and height of the austenite grains after hot rolled quenched samples. It was considered that strain partitioning between the two phases results in translations of the austenite grains in the continuous ferrite matrix. At room temperature, the microstructure of cold worked DSS is generally characterized by formation of dislocation cells in ferrite, whereas a much more uniform dislocation distribution with high incidence of stacking faults is found in austenite. 24) In this condition, both hardness and dislocation density are higher in austenite, which is the phase that exhibits more strain hardening and stored energy ) The dynamic and static mechanisms that develop during annealing are even more complex because they are not arising to the same extend as for monophasic ferritic and austenitic stainless steels. Among others, factors such as the volume fraction of each phase and the nature of the interphase will determine the softening processes during annealing in such duplex structures. Generally, extensive recovery takes place in ferrite on annealing and also recrystalization nuclei are formed. The preferred areas for recrystallization nucleation are regions with high density of shear bands. Recrystallization in austenite occurs in a more discontinuous mode, preferentially at ferrite/austenite interfaces, deformation twins and shear bands. 24,27 29) Most of the works available in the literature concern the behaviour of DSS under specific thermomechanical conditions, i.e. hot deformation, cold deformation or static annealing. The aim of this paper is to give an overview of the microstructural evolution of DSS during the whole industrial rolling process, where the material is submitted to different states and to study the influence on the mechanical properties and on formability of the final industrial product. 2. Experimental Procedure The material studied was a duplex stainless steel EN produced by UGINE & ALZ (ArcelorMittal Group). Table 1. Chemical composition (in weight-%) of the studied duplex stainless steel EN It was supplied after each of the four different industrial process steps indicated in Fig. 1. The chemical composition of the alloy is given in Table 1. The first condition studied was the hot rolled steel (at the end of the finishing mill), from now on designated as HR, which was supplied in the form of plates of mm and a thickness of 6 mm. The industrial hot rolling schedule followed to obtain these plates consists basically in two steps. First, cast slabs of 250 mm of thickness are reheated to C and subsequently pre-rolled to a thickness reduction of 25% (150 mm). Before second step, the slabs are cooled down to room temperature and their surface is ground. The final rolling operations starts with a reheating at C and then slabs are rough rolled with a total thickness reduction of 80%, i.e. from 150 to 30 mm thickness. Subsequently, the rough rolled sheets are hot rolled with the final 6 mm thickness which corresponds to a total reduction of 96%. Typical finishing temperatures are between 975 C and C. Next, the sheets are rapidly cooled to 650 C by means of water sprays in the run out table and coiled. The second supplied condition was after hot rolling and annealing (HRA) at a temperature of approximately C for around 15 min and the third condition was after cold rolling (CR). In this case the sheet ( mm) had a thickness of 2 mm. The last one was the final product (FP) of the industrial process obtained after the last solution annealing treatment at C for app. 10 min. The sample dimension was equal to the one after CR. The microstructures were examined by both optical and transmission electron microscopy (TEM). For optical microscopy, electrolytic etching in KOH solution was used. This etching leaves ferrite grains in brown colour and austenitic ones in white. A detailed crystallographic characterization was carried out using a SIEMENS D5000 texture goniometer. Incomplete pole figures were measured in the back reflection mode at the centre layer of the rolling sheet. Thereby, for the ferritic BCC phase the {110}, {200}, {211} and {310} pole figures were obtained, while for the austenitic FCC phase the {111}, {200}, {220}, and {311} pole figures were recorded. Subsequently, the orientation distribution functions (ODF) for both phases were computed. Traditionally, the ODF is represented in sections j 2 constant for the FCC and j 1 constant for the BCC phase in steps of 5. Hardness was measured in both phases after each process step at the central layer of the rolling sheet. Because of the small thickness of the austenite and ferrite layers, hardness tests were made with a MTS Nanoindenter XP instrument with a Berkovich indenter at constant deformation rate of 0.05 s 1 up to 500 nm of indentation depth. The hardness values were obtained by the method developed by Oliver and Pharr. 30) Tensile testing of the final industrial product was carried out at room temperature using an INSTRON 5585 computerized universal testing machine equipped with two exten- 1597

3 someters, i.e. in the direction of the tensile stress and perpendicular to it, to measure plastic-strain ratios (r-values) according to ASTM E ) The tests were performed on samples machined at 0, 45 and 90 to the rolling direction (RD), i.e. longitudinal (L), diagonal (D) and transversal (T), in order to determine mean anisotropy (r ) and planar anisotropy (Dr) by using the following two equations: r r r r ( )...(1) 4 r r r r ( )...(2) 2 To prove the influence of planar anisotropy on the forming properties of the sheet, earing tests according to ISO ) were carried out with a punch diameter of 20 mm and using circular samples with a diameter of 40 mm. Fig. 2. Three-dimensional microstructure of the hot rolled duplex stainless steel. Fig. 3. Fig. 4. TEM micrograph of the hot rolled duplex stainless steel. Microstructure of the hot rolled and annealed duplex stainless steel. 3. Results and Discussion 3.1. Microstructural Evolution As a result of the rolling process, the steel in the HR condition shows already a strongly oriented microstructure which is aligned in the rolling direction RD (Fig. 2). A characteristic pancake-like microstructure was observed consisting of flattened, widened and elongated austenite and ferrite bands, whose thicknesses (dimension in the normal direction ND) were between 0.5 and 11 mm for both phases. In the rolling plane, ferrite appears with very irregular morphology and wavy boundaries as a consequence of being the phase which carries most of the deformation, as demonstrated by Duprez 23) since the ferritic phase is not only the matrix phase but also the softer phase. For the HR condition, TEM observations revealed a structure of subgrains in ferrite (Fig. 3) due to the extended dynamic recovery, 20) which is similar to the continuous recrystallization that takes place in ferritic steels during deformation at high temperatures. 21) Through the recovery process part of the dislocations are annihilated. The remaining dislocations are stored in low-angle dislocation walls resulting in the formation of subgrains. The merging of these dislocations walls leads to a gradual build up of misorientation between neighbouring subgrains. Austenite appeared with a higher grain size than ferrite as a result of dynamic recrystallization and/or metadynamic recrystalization, which can take place during the cooling of the plate from hot rolling. In austenite, which has a low SFE, dislocations are less mobile and dynamic recovery is limited. Consequently, once a critical strain is exceeded, the driving force becomes large enough for dynamic recrystallization when the dislocations are eliminated through the replacement of deformed grains by new dislocation-poor grains. After the subsequent annealing at C, the microstructure of the HRA material is shown in Fig. 4. For this condition, the longitudinal section presents layers of austenite and ferrite with thickness between 3 and 11 mm. In the rolling plane, the wavy morphology of ferrite has almost disappeared and the microstructure is, compared to the HR material, characterized by thicker bands of austenite and ferrite. TEM observations of austenite did not differ considerably from the HR step because the softening mechanisms had started to act during hot rolling. During the annealing process extensive recovery took place in ferrite with an increase of the average size of the subgrains (Fig. 5). As it can be observed, the recovery is affected by the orientation present after deformation and, as a result, subgrains develop following the rolling direction. There is a marked reduction in the thickness of the layers at the end of the cold rolling step, leading to values between 3 and 5 mm. In almost all observations of the material in the CR condition, ferrite and austenite layers are essentially one grain thick, so the layer spacing may be considered as a good approximation to the grain size in this section. As a consequence of (shear) deformation, flattened broken bands of austenite and ferrite are visible at the rolling plane (Fig. 6). After the last annealing step at C, softening mechanisms lead to a final industrial product, which recovers the aligned phases along the rolling direction originated in the 1598

4 Fig. 8. TEM micrograph of the final industrial product: a) Ferrite and b) Austenite. Fig. 5. TEM micrograph after industrial hot rolling and annealing. Fig. 6. Three-dimensional microstructure of the industrially cold rolled duplex stainless steel. Fig. 9. Variation of austenitic phase volume fraction after each industrial process step. Fig. 7. Three-dimensional microstructure of the final industrial product. first stage of the industrial process (Fig. 7). For the FP a subgrain structure of ferrite is not observed anymore by TEM (Fig. 8). The main part of the new grains formed during annealing is equiaxial and they have a low dislocation density. This feature is also visible for austenite, where the presence of annealing twins is the main feature. The variation of the austenitic phase volume fraction during the industrial process is shown in Fig. 9. As it can be seen, HR steel presents the lowest volume fracture of austenite percentage compared with the rest of the studied conditions. For hot rolling, the steel first becomes re-heated at high temperatures (1 250 C) before HR. In this high temperature range the ferritic phase is more stable and less austenite is present. Industrial annealing conditions, after hot and cold rolling, proceed at lower temperatures and therefore phase transformation of ferrite to austenite is enhanced. Note that the little drop in austenite volume fraction after CR might be related to the deformation induced austenite to a -martensite transformation, which retransforms to austenite again after the final annealing process Texture Evolution In Figs. 10 and 11, the texture development of the ferritic and austenitic phase during the different industrial process steps, i.e. HR, HRA, CR and FP, is shown. Textures appear with a marked intensity in ferrite, mainly consisting of sharp a-fiber ( 110 //RD) spread from {001} 110 to {112} 110 (Fig. 12(a)). The a-fiber exhibits a similar statistic relevance in the final industrial product and in the hot rolled plate. Orientation density values along a-fiber show a maximum on the {001} 110 orientation for all the studied conditions (Fig. 10). An interesting feature is the weakness of the {111} 110 orientation, just like for the rest of the g-fiber components (Fig. 12(a)) which usually are the dominating textures for BCC metals after rolling and annealing. This result is in agreement with the previous studies carried out with duplex stainless steels ) In the austenite, texture of HR steel is well defined (Fig. 11(a)). The main part of the deformation rolling textures characteristic of FCC metals are observed (Fig. 12(b)), i.e. Brass ({011} 211 ), Copper ({112} 111 ) and S ({123} 634 ) orientations, lying on the b-fiber, which is close to the g-fiber, and the Goss ({011} 100 ) component. Moreover, the presence of cube texture reveals that during the hot rolling step, austenite recrystallizes dynamically and/or meta-dynamically, i.e. when rolling deformation is stopped but temperature is still elevated; nuclei grow into the heterogeneous ferritic matrix, which is partly dynamically recrystallized. That a dynamic recristallization and deformation occur simultaneously is noticeable by the RD-rotation of the cube orientation via the Goss to the Brass orientation, which is the ultimate deformation component in 1599

5 Fig. 10. Texture development of the ferritic phase after each industrial processing step: a) HR, b) HRA, c) CR and d) FP. Fig. 11. Texture development of the austenitic phase after each industrial processing step: a) HR, b) HRA, c) CR and d) FP. ISIJ International, Vol. 48 (2008), No

6 Fig. 13. Austenite and ferrite hardness measured after each industrial process step. Table 2. Tensile test results of the final industrial product. Fig. 12. Selected ODF sections with some important fibers and orientations: a) BCC metals and b) FCC metals. Table 3. r-values obtained from the tensile test of the final industrial product. FCC materials with a low stacking fault energy. 34) This RDrotation of the cube orientation manifests itself in different intensities in all processing steps, see Fig. 11. In addition, in the course of processing the Copper orientation shifts from its ideal position (j 1 90 ) to lower j 1 -values, which is an indication for deformation inhomogeneities like shear deformation and the formation of shear bands. It is also worth mentioning that the texture intensity decreases after each process step. This is due to the fact that after each annealing the new recrystallized grains adopt random orientations. The texture intensity is the highest after HR because of the elevated temperatures, where orientation rotations in austenite and ferrite are easier than during cold rolling. The similarity between the ferritic texture of the FP and the HR conditions results from the fact that, at high temperatures, ferrite is the softer matrix phase, and therefore it carries out most of the deformation. As a consequence, ferrite develops the same type of texture and intensity as after HR. For the same reason the ferritic matrix has higher texture intensities than the harder austenitic second phase Hardness Evolution The hardness evolution of both phases during the different steps of the process is shown in Fig. 13. At the end of the hot rolling process ferrite is harder than austenite. As demonstrated in other studies, 23) when duplex stainless steel is deformed at high temperatures most of the strain is taken by ferritic matrix which flows around austenite grains. In that way the authors also considered that this difference in strain partitioning lead to a considerable incompatibility in strain between the a/g at the phase boundaries which was supposed to be assume mostly by the ferrite. After the subsequent annealing step, an extended recovery process leads to a considerable decrease of ferrite hardness. In contrast, hardness of austenite remains unchanged as a consequence of the softening initiated in the preceding hot rolling step which reduces the driving force for recrystallization. During cold rolling both phases experienced strain hardening in a similar degree, being the hardness of the austenite higher than that of ferrite due to the nitrogen addition and the low stacking fault energy. Analogous results were reported by other authors 24 26) by measuring the hardness with increasing plastic deformation. They observed that at room temperature austenite displays more strain hardening than ferrite with a considerable increase of dislocation density. At the end of the industrial process, the difference in hardness between austenite and ferrite remains similar to that observed after cold rolling, despite both phases have softened significantly as a consequence of the recrystallization process that takes place during the last annealing step Anisotropy Evaluation of the Final Industrial Product The results obtained from the tensile tests are summarized in Table 2. The anisotropy is clearly reflected on mechanical properties considering the three orientations of the rolled sheet, i.e. L (0 ), D (45 ) and T (90 ). It can be seen that the highest values of yield stress (s y ) and ultimate tensile strength (s uts ) corresponded to the transversal direction, while diagonally orientated samples exhibited the lowest strength but the highest ductility. This marked anisotropy may have a negative effect for applications because usually the direction of the applied stresses in industrial uses corresponds with the rolling direction as a result of the limited width of the flat product. This strong directionality is also noticeable with the r coefficient, which has the maximum value at 45 and decreases drastically for samples tested at 0 and 90 with respect to the rolling direction (Table 3). The calculated mean anisotropy (r ) is inferior to one, indicating that the strength in the thickness direction is lower than the average strength in the directions in the plane of the sheet and, as a consequence, this material will exhibit limited formability by deep drawing. 1601

7 The authors owe their gratitude to ALZ (today UGINE & ALZ, ArcelorMittal Group), for providing the material via OCAS (today Arcelor Research Industry Ghent, Arcelor- Mittal Group), to OCAS for carrying out the texture measurements and to the Technologic Center of Manresa for its technical support in forming tests. REFERENCES Fig. 14. Earing after cup drawing test. The high value of planar anisotropy (Dr) obtained means that there will be no-uniform strength distribution in the plane of the sheet. Accordingly, four ears at 45 were observed after the earing test (Fig. 14). In practice, for drawing operations, it will be necessary to leave enough extra metal on the drawn cup in order to trim the ears formed. 4. Conclusions Rolled duplex stainless steel EN has been investigated in order to study the microstructural evolution during industrial processing and its effects on the anisotropy of the final industrial product. After hot rolling, the microstructure is morphologically anisotropic with austenite and ferrite bands elongated in rolling direction. This feature remains in the subsequent steps of the industrial process until the final product. The a-fiber is the predominant texture in ferrite with no development of g-fiber components. Typical rolling textures of FCC metals are observed in austenite together with the cube texture, which is an evidence of recrystallization processes. Only for cold rolled steel, the distribution of austenite textures is different and reveals characteristics of deformed FCC metals with low SFE. Hardness measurements reveal that the strain partitioning between austenite and ferrite is not equal in hot and cold rolling. While ferrite posses the highest hardness values after deformation at high temperature, austenite hardens more than ferrite at room temperature. Due to the anisotropic microstructure, the mechanical properties are strongly direction dependent, with a maximum resistance corresponding to samples oriented perpendicular to the rolling direction. Moreover, the effect of textures is reflected by a low value of normal anisotropy and considerable planar anisotropy, which enhance the formation of ears during cup drawing tests. Therefore, the industrial final product will exhibited limited formability, especially for manufacturing of components by deep drawing. Acknowledgements The authors wish to thank the Spanish MCYT (Grant MAT ) for the financial support and to Dr. J. M. Manero (UPC) for his assistance in TEM studies. 1) H. D. Solomon and T. M. Devine: Proc. of Conf. on Duplex Stainless Steels, ed. by R. A. Lula, ASM, Metal Park, Ohio, (1983), ) P. Guha and C.A. Clark: Proc. of Conf. on Duplex Stainless Steels, ed. by R. A. Lula, ASM, Metals Park, Ohio, (1983), ) J. Charles: Proc. Duplex Stainless Steels 91, Vol. 1, Les éditions de physique Beaune, France, (1991), 1. 4) J. O. Nilsson: Mater. Sci. Technol., 8 (1992), ) P. Lovland: Stainless Steel Eur., (1993), 28. 6) J. Oredsson and S. Bernhardsson: Mater Perform., 22 (1983), 35. 7) N. Sridhar, J. Kolts and L. H. Flashe: J. Met., 37 (1985), 31. 8) M. A. Streicher: Met. Prog., 128 (1985), 29. 9) R. D. Kane: Adv. Mater. Proc., 144 (1993), ) W. B. Hutchinson, K. Ushiola and G. Runnsjö: Mater. Sci. Technol., 1 (1985), ) J. L. Song and P. S. Bate: Acta Mater., 45 (1997), ) A. Mateo, L. Llanes, N. Akdut, J. Stolarz and M. Anglada: Int. J. Fatigue, 25 (2003), ) B. Ahlblom and R. Sandstrom: Int. Met. Rev., 27 (1982), 1. 14) H. J. McQueen and J. J. Jonas: J. Appl. Metalwork., 3 ( ), ) J. Botella, J. L. Martos and R. Sánchez: Proc. Int. Cong., Stainless Steels 96, ed. by VDEH, Dusseldorf, (1996), ) J. A. Jiménez, M. Carsi and O. A. Ruano: Proc. VI Cong. Nac. Propiedades Mecánicas de Sólidos 98, Badajoz, España, (1998), ) F. E. Al-Jouni and M. Sellars: Proc. Risf Symp., ed. by J. B. Blide Sfrensen et al., Roskilde, Denmark, (1983), ) A. Iza-Mendia, A. Piñol, J. J. Urcola and I. Gutiérrez: Metall. Mater. Trans. A., 29A (1998), ) A. Iza-Mendia: Ph.D. Thesis, University of Navarra, San Sebastián, (1999). 20) P. Cizek and B. P. Wyne: Mater. Sci. Eng. A, A230 (1997), ) C. G. Schmidt, C. M. Young, B. Walser, R. H. Klundt and O. D. Sherby: Mater. Trans. A, 13A (1982), ) F. J. Humphreys and M. Hatherly: Recrystallization and Related Annealing Phenomena, Pergamon Press, Elmsford, New York, (1996). 23) L. Duprez: Ph.D. Thesis, Universiteit Gent, Gent, (2003). 24) W. Reick, M. Pohl and A. F. Padilha: ISIJ Int., 38 (1998), ) W. H. M. Welman, V. J. Gadgil and B. H. Kolster: Proc. Duplex Stainless Steels 91, Vol. 1, Les éditions de physique, Beaune, (1991), ) A. L. Etter, T. Baudin, M. H. Mathin, W. Swiatnicki and R. Penelle: Scr. Mater., 54 (2006), ) B. A. Cooke, A. R. Jones and B. Ralph: Met. Sci., 13 (1979), ) M. Blicharski: Met. Sci., 18 (1984), ) J. Keichel, G. Gottstein and J. Foct: Proc. High Nitrogen Steel 98, Stockholm, (1999), ) W. C. Oliver and M. G. Pharr: J. Mater. Res., 7 (1992), ) ASTM E517-81, Annual Book of ASTM Standards, ASTM, Philadelphia, (USA), ) ISO 11531, International Standards, (1994). 33) W. B. Hutchinson, U. V. Schlippenbach and J. Jonsso: Proc. Duplex Stainless Steels 86, TWI, The Hague, The Netherlands, (1986), ) J. Hirsch, K. Lücke and M. Hatherley: Acta Metall., 36 (1998),