Abstract

    A ferritic stainless steel, with a chromium/nickel equivalent ratio of 7.4, was deformed in torsion over a temperature range of 900°C to 1200°C and a strain rate of 1 s^-1. The flow curves attained showed two different shapes: at higher temperatures, the materials typically softened only by dynamic recovery while, at lower temperatures, the flow curves were characterized by a rapid rise to a peak stress, followed by extensive flow softening. There was a very significant reduction of grain size during deformation under all the conditions studied, suggesting that dynamic recrystallization occurred. These observations are incongruent with the literature, which reports that the restoration mechanisms for ferrite change from dynamic recrystallization to dynamic recovery when the deformation temperature decreases, at a constant strain rate. Observations of the microstructural evolution after interrupting the tests at different levels of strain and analysis of the dependence of the work-hardening rate on the applied stress lead to a discussion about the competition between recovery and recrystallization during deformation of ferritic stainless steel at high temperatures and upon extensive strain.

Key Words: dynamic recovery, dynamic recrystallization, continuous dynamic recrystallization

Resumo

    Um aço inoxidável austenítico, com razão cromo/níquel equivalente de 7,4, foi deformado na faixa de temperaturas de 900°C a 1200°C com taxa de deformação de 1 s^-1. As curvas tensão x deformação obtidas mostraram duas formas diferentes: em altas temperaturas a forma típica de materiais que amaciam apenas por recuperação dinâmica enquanto que, em baixas temperaturas, as curvas tensão x deformação foram caracterizadas por um crescimento rápido até uma tensão de pico, seguido por um amaciamento extensivo. Foi observada uma redução significativa no tamanho de grão durante a deformação em todas as condições estudadas, sugerindo que a recristalização dinâmica tenha ocorrido. Estas observações são incongruentes com a literatura, que reporta que os mecanismos de restauração para a ferrita variam de recristalização dinâmica para recuperação dinâmica quando a temperatura de deformação diminui a uma taxa constante. Observações da evolução microestrutural em ensaios interrompidos em diferentes níveis de deformação e a análise da dependência da taxa de encruamento com a tensão aplicada possibilitam uma discussão sobre a competição entre recuperação e recristalização durante a deformação de aços inoxidáveis ferríticos em altas temperaturas e submetidos a grandes deformações.
Palavras Chaves: recuperação dinâmica, recristalização dinâmica, recristalização dinâmica contínua

1. Introduction

    During the cold deformation of metals, a small fraction of the expended work remains in the form of stored energy in the crystal lattice, work-hardening the material. When processed at high temperatures, the work-hardening rate decreases due to the operation of restoration mechanisms such as recovery and recrystallization. It has been demonstrated that the cold-worked state can be described by a single internal parameter, total dislocation density, to which other internal parameters can be related [1]. As a consequence, the measured flow stress can be used as a probe of the strained structure and the work-hardening rate as a measure of the structure’s strain-induced change.
Dynamic recovery and recrystallization is activated by different amounts of stored energy; consequently, these processes depend upon the density, distribution and arrangement of dislocations. Materials whose dislocations are able to cross-slip and climb, rearranging into a polygonal subgrain substructure, tend to show a high degree of dynamic recovery, while materials with low stacking fault energy have a much lower level of dynamic recovery. In this case, the dislocation density rises until a critical condition is reached, at which point new grains are nucleated and grow during straining [2]. The rapid elimination of dislocations during recrystallization by the growth of new grains leads to a flow curve peak followed by rapid work softening, when the material is strained at a constant temperature and strain rate.
Ferritic steels provide a good example of high levels of dynamic recovery [3,4]. As a result of developing constant dislocation density after a certain amount of work hardening, the flow curve rises monotonically to a steady state plateau, when strained at a constant temperature and strain rate. It has been suggested that dynamic recovery is the only mechanism required in the hot deformation of this type of material [5,6]. However, new grains have been observed to form during the deformation of a-iron [7] and ferritic stainless steel [8]. The purpose of this work was to investigate the active softening mechanism correlating the microstructural evolution with the shape of the flow curve in a ferritic stainless steel subjected to extensive straining by hot torsion.

2. Materials and Experimental Procedures

    The material used in this work was a ferritic stainless steel with a chromium/nickel equivalent ratio of 7.4, whose chemical composition is given in Table I. This steel was prepared by vacuum induction melting and the resulting ingot was divided into two parts. Some of the samples were machined from the ingot and tested in the as cast condition while others were tested in the as wrought condition after hot forging with a reduction of area of 80%. In both cases, the samples were subjected to hot torsion testing at a strain rate of 1 s^-1 over a temperature range of 900°C to 1200°C. The as cast material was tested on heating; the samples were heated to the test temperatures, kept at these temperatures for 1 minute and strained until the final fracture occurred. In the as wrought condition, the samples were tested on cooling, i.e., they were heated to 1250°C and held for 10 minutes at this temperature, then cooled to the test temperatures, kept at these temperatures for 1 minute and strained until the specimen fractured.
The mechanical tests were carried out using a computerized hot torsion machine [9]. The samples, which were 10 mm in length and 10 mm in diameter in the reduced central gage section, were heated in an induction furnace assembled on the testing machine. Chromel-alumel thermocouples set on the surface of the samples were used to measure and control the temperature. To prevent oxidation, the sample was enclosed in a 2% hydrogen argon atmosphere surrounded by a quartz tube. In order to retain the high-temperature microstructures for further observations, water was injected into the quartz tube, either immediately upon reaching the fracture strain or, in some cases, after attaining the required strain. A metallographic examination was carried out on tangential sections at a depth of less than 1mm below the surface, using optical microscopy (OM) and scanning electron microscopy (SEM).

3. Results

    Figure 1 shows the microstructures of the ferritic stainless steel in the as cast and as wrought conditions before the beginning of mechanical tests. As can be seen, the microstructure of both samples consists of a matrix (average ferrite grain size in the as cast condition was 410 µmm and 187 µmm in the as wrought condition) with a certain number of second-phase particles. The high Cr/Ni ratio indicates that this steel solidified as single-phase ferrite; the small particles within the grains and at the grain boundaries are austenite formed from ferrite in the solid state by the Widmanstätten mechanism [10]. Although the volume fraction of austenite decreased as the material was processed, a certain amount of particles can be observed even in Figure 1b.

Table I – Chemical Composition of the tested Steel (wt%)

 

 

Fig. 1 . OM starting microstructures for mechanical tests carried out under the as cast (a) and as wrought (b) conditions

    Figure 2 illustrates the flow curves determined for samples tested on cooling in the as wrought condition. At higher temperatures, the flow curves display the characteristic shape expected for materials that soften by dynamic recovery, with rapid work hardening to a low steady state level. As the temperature decreases, the shape of the flow curves changes, with rapid work hardening reaching a peak stress followed by extensive flow softening. This difference in the flow behavior is better depicted in Figure 3, which illustrates the dependence of peak strain and strain to fracture on deformation temperature.
The peak strain usually increases as the deformation temperature decreases. Figure 3a displays an unusual behavior: when deformation temperature increased, the peak strains decreased to a minimum, rising again when the temperature rose. In this case, the transition temperature, which is characterized by a minimum in the peak strain, was around 1050°C. This trend was also observed in the behavior of the strain to fracture. Figure 3b displays two different behaviors in the dependence of ductility on the test temperatures for samples tested in the as cast and as wrought conditions. In both conditions, there is significant ductility at high temperatures; however, it decreases as the deformation temperature drops to below 1100°C.
Figure 4 shows the effect of deformation on the final microstructure. Figure 4a displays the microstructure of a sample strained in the low ductility region. This specimen was tested on heating to 1000°C in the as cast condition. In this figure, it is worth noting the presence of some particles of austenite precipitated in deformation bands and aligned in the direction of deformation within the ferrite matrix. Figure 4b shows the microstructure of a sample strained in the high ductility region. This sample was tested on cooling to 1200°C in the as wrought condition. Both experiments reveal that the grain size decreased significantly during deformation, as can be seen by a comparison of Figures 1 and 4.

 

Fig. 2 – Flow curves for tests carried out on cooling for samples with the as wrought structure.

 

 

 

Fig. 3. Dependence of the peak strain in the as wrought condition (A) and strain to fracture for both conditions (B) on deformation temperature.

 

 

Fig. 4 – OM microstructure observed after straining to fracture in the low ductility region (A) and in the high ductility region (B) indicated in Figure 3A.

4. Discussion

    There was a dramatic change in grain size during deformation in all the conditions studied; coarse equiaxed ferrite grains that were present at the beginning of straining were transformed into small grains after extensive straining, which is an evidence that dynamic recrystallization occurred. An observation of the shape of the flow curves in Figure 2 indicates that, at lower deformation temperatures, the stress increased as deformation proceeded, reaching a peak, after which the stress decreased to a steady state. This is a typical behavior of the materials that soften by dynamic recrystallization. However, at temperatures in excess of 1100°C, the shape of the curves is characteristic of materials that softened only by dynamic recovery. These findings are in disagreement with the literature. It has been reported that the restoration mechanisms for ferrite change from dynamic recovery to dynamic recrystallization when the value of the Zener-Hollomon parameter     is decreased [7,8,11].
To analyze this behavior, samples of the ferritic stainless steel were subjected to hot torsion testing at 1150°C and the microstructural evolution was observed after interrupting the tests at different levels of straining and rapid quenching of the samples by water spray.

    Figure 5 displays the microstructure observed at two different levels of straining. In the first, the straining was interrupted at e=0.28, which corresponds to the peak stress, see Figure 3a. Figure 5b displays the microstructure after straining to e=0.8, which represents an intermediate value between peak strain and steady state strain. At the beginning of straining, the original grains were elongated and the grain boundaries became irregular. Figure 5a shows how the original grain boundaries bulged and new fine grains were formed. This microstructure corresponds to the early stage of dynamic recrystallization [12]. Increasing the straining toward the steady state revealed a distinct substructure (Figure 5b). The substructure and the irregular shape of the grain boundaries render this phenomenon easily distinguishable from grains formed by discontinuous dynamic recrystallization.

Fig. 5 – Evolution of the microstructure (SEM) during testing at 1150°C after straining to e=0,28 (a) and e=0.8 (b)

    Bearing in mind that dynamic recovery and dynamic recrystallization are competitive mechanisms, the onset of dynamic recrystallization takes place after a certain amount of recovery, when the work-hardening rate decreases drastically.

    Figure 6 illustrates the dependence of the work-hardening rate on the applied stress in the work-hardening regime, calculated from flow curves attained in the as wrought conditions. The work-hardening rate clearly decreases continuously as the applied stress is increased to peak stress, indicating that no single significant change in the active softening mechanism occurs throughout the work-hardening region. Thus, in all the experiments carried out here, only dynamic recovery is believed to have occurred at this stage of straining.
Together with the absence of an inflexion point in the work-hardening rate curves, the presence of well formed and homogenously distributed substructures within original grains at straining greater than that of the peak stress is a strong evidence that the material softened by intense dynamic recovery. The reduction in grain size during straining leads us to conclude that a continuous new grain formation process occurred. Thus, continuous dynamic recrystallization took place in this material, with new grains forming through the growth of subgrains and a gradual increase in sub-boundary misorientation [13,14].
    The literature reports that the interface between precipitated austenite particles and the ferrite matrix is coherent [15,16]. Thus, when the austenite phase is formed by a ferrite solid-state transformation, the austenite is coherent with the ferrite

 

Fig. 6 – Dependence of the work-hardening rate on the applied stress calculated from flow curves displayed in Figure 2.

    This coherence increases the interaction between precipitates and dislocations, increasing the material’s resistance. During the initial stage of deformation, the austenite particles inhibit the deformation of the ferrite matrix, which is softer, increasing flow stress and work hardening. This effect continues until some dynamic softening process diminishes the coherence between the particles and the matrix. A recent paper [17] suggests that the coherence decreases during the formation of new grains due to the growth of subgrains and the gradual increase in sub-boundary misorientation, with the continuous dynamic recrystallization of the ferrite matrix. Thus, after a certain amount of straining, when the particles are no longer coherent with the matrix, the flow stress no longer changes as the deformation reaches the steady state.
    This interpretation is in agreement with the data presented herein. At lower deformation temperatures, particles of austenite were observed and the flow curves displayed a “rump”-like shape: these curves were characterized by a rapid rise to a peak stress followed by extensive flow softening. At higher temperatures, austenite particles were dissolved during heating or had not yet precipitated; the flow curves show the characteristic shape of materials softened by dynamic recovery. Moreover, the material was high ductile under these conditions while, at lower deformation temperatures, the austenite particles hindered the movement of new ferrite grains and the strain to fracture was markedly decreased [18].

5. Conclusions

    Although bulging of grain boundaries was observed after slight straining and the shape of flow curves suggests that a softening mechanism such as dynamic recrystallization occurred, the presence of a well formed and homogeneously distributed substructure within the original grains and the absence of an inflexion point in the work-hardening rate curves are strong evidences that the ferritic stainless steel softened by intense dynamic recovery.

    The reduction in grain size during straining leads to the conclusion that continuous dynamic recrystallization occurred during deformation, with new grains forming through the growth of subgrains and a gradual increase in sub-boundary misorientation. The difference in the shape of the flow curves is correlated with the presence/absence of austenite particles formed by the Widmanstätten mechanism. The flow stress was heightened when these particles were coherent, and flow softening was observed when the coherence disappeared with continuous dynamic recrystallization.

Acknowledgements

    The authors wish to thank the Brazilian research funding agencies FAPESP and CNPq for their support of this work.

References

1. J. Gil Sevilano: Mater. Sci. Forum, vols. 113-115, pp. 19-28, (1993).         [ Links ]
2. E. I. Poliak and J. J. Jonas: Acta Mater., vol. 44, pp. 127-136, (1996).         [ Links ]
3. G. H. Akbari, C. M. Sellars and J. A. Whiteman: Acta Mater., vol.45, pp. 5047-5058, (1997).         [ Links ]
4. P. Cizek and B. P. Wynne: Mater. Sci. Eng., vol. A230, pp. 88-94, (1997).         [ Links ]
5. H. J. McQueen and W. Blum: Mater. Sci. Eng., vol. A290, pp. 95-107, (2000).
6. E. Evangelista, P. Mengucci, J. Bowles and H. J. McQueen: High Temp. Mater. Proc., vol. 12, pp. 57-65, (1993).         [ Links ]
7. G. Glover and C.M. Sellars: Metall. Trans., vol. 4A, pp. 765-775, (1973).         [ Links ]
8. T. Maki, S. Okaguchi and I. Tamura: in ICSM 6, R. C. Gifkins, ed., Pergamon Press, vol. 1, pp. 529-534, (1982).         [ Links ]
9. A. M. Jorge Jr. and O. Balancin: Rev. Eng. Ciênc. Aplic., vol. 2, pp. 133-148, (1994/1995).         [ Links ]
10. N. Suutala, T.Takalo and T. Moisio: Metall. Trans., vol. 10A, pp. 1183-1190, (1979).         [ Links ]
11. P. R. Cetlin, S. Yue and J. J. Jonas: ISIJ Intern., vol. 33, pp. 488-497, (1993).         [ Links ]
12. W. Roberts, H. Bodén and B. Ahlblon: Met. Sci., vol. 13, pp. 195-205, (1979).         [ Links ]
13. A. Belyakov, R. Kaibyshev and R. Zaripova: Mater. Sci. Forum, vol. 113-115, pp. 385-390, (1993).         [ Links ]
14. A. Belyakov, H. Miura and T. Sakai: in Thermec’97, T. Chandra and T. Sakai, eds., The mineral, Metal & Material Society, vol. 1, pp. 257-263, (1997).         [ Links ]
15. J.B. Arboledas, J. L. M. Tirado and R. S. Rodriguez: in Proc. Int. Congr. Stainless Steels 96, Dusseldorf/Neuss, June 1996, VDEh (Verein Deutsche Eisenhütten-leute), Dusseldorf, pp. 116- 121, (1996).         [ Links ]
16. A. Iza-Mendia, A. Piñol-Juez, J. J. Urcola and I. Gutiérrez: Metall. Mat. Trans., Vol. 29A, pp. 2975- 2986, (1998).         [ Links ]
17. O. Balancin, W. A. M. Hoffmann and J. J. Jonas: Metall. Mat. Trans., vol 31A, p. 1353-1364, (2000).         [ Links ]
18. G. S. Reis, A. M. Jorge Jr. and O. Balancin: Materials Research, vol. 3, pp. 31-35, (2000).         [ Links ]