Microstructural investigation on strengthening mechanisms of AISI ...

Microstructural investigation on strengthening mechanisms of AISI 304L austenitic stainless steel during cryogenic deformation P. Behjati*, A. Najafizadeh and A. Kermanpur In the present paper, microscopy techniques and mechanical tests were used to investigate in detail the strengthening mechanisms of an AISI 304L austenitic stainless steel during cryogenic deformation. The strain hardening rate–strain response of the alloy indicated three distinct regimes of hardening, being very similar to that previously reported for other low stacking fault energy alloys. The hardening rate initially decreased up to a strain of y6% (stage I). Then, a second stage of increasing hardening rate began (stage II). At strains larger than y25%, stage III with decreasing hardening rate followed. It was suggested that the formation of e-martensite and a9-martensite together is responsible for the appearance of stage II. The high strain hardening values of the alloy in stage II were related to the increased fraction of a9-martensite and dislocation pile-ups behind the Lomer–Cottrell locks. The appearance of stage III was attributed to the difficulty of a9-martensite nucleation and ease of dislocation cross-slip at higher strains. Keywords: Austenitic stainless steel, Cryogenic deformation, Martensite, Lomer–Cottrell locks

Introduction Austenitic stainless steels are widely used in important industries, such as chemical, petrochemical, machinery, automobile, nuclear and shipyard, due to their excellent corrosion resistance, good weldability, excellent formability and high strength.1 These alloys receive their strength mainly through solid solution strengthening, strain hardening or grain refining technique. Their microstructure is composed of a metastable austenite c phase. Plastic deformation of the c phase slightly above the Ms temperature (martensite start temperature) can give birth to the formation of e and a9 phases with hcp and bcc crystal structures respectively. It is well established that the deformation induced martensitic transformations in the 300 series austenitic stainless steels occur most probably through the cReRa9 steps. Earlier studies show that the character and intensity of the martensitic transformations depend on the type,2,3 rate,4 amount5 and temperature of deformation6 and also on the chemical composition of the alloy.7–9 The results of recent investigations reveal that deformation induced martensitic transformations can be employed to develop a new generation of austenitic stainless steels with exceptional mechanical properties.10–12 It has been found that with accurate control of such phase transformations, it is possible to enhance the strength of commercial stainless steels to the levels of y2 GPa.13,14 It

Department of Materials Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran *Corresponding author, email [email protected]

1828

ß 2011 Institute of Materials, Minerals and Mining Published by Maney on behalf of the Institute Received 16 October 2010; accepted 4 December 2010 DOI 10.1179/1743284710Y.0000000040

has been shown that the level of success in the manufacture of such superior steels depends on the morphology of the deformation induced martensite.15 These new findings have regenerated interest in the microstructural study of martensitic transformations and also in their influence on the strength and strain hardening behaviour of austenitic stainless steels. Several studies have been performed to investigate the strain hardening behaviour of austenitic stainless steels.13,16–18 Their results show that in these steels, the formation of deformation twins and deformation induced martensite can increase the strain hardening values considerably. However, few investigations have been performed on the correlation between the microstructure and mechanical properties of austenitic stainless steels at cryogenic temperature (77 K). Furthermore, the contribution of e-martensite and Lomer–Cottrell (LC) locks on the strain hardening behaviour of these steels has not received enough attention so far. In the present paper, microstructural evolutions occurring during the cryogenic deformation of an AISI 304L austenitic stainless steel are investigated in detail. Based on the observations, the influence of martensite phase and LC locks on the flow stress and strain hardening of this alloy is discussed.

Experimental The chemical composition of the AISI 304L stainless steel used in the present work is Fe–10Ni–18Cr–2Mn– 1Si–0?03C–0?04P–0?03S. The annealed raw material was provided in the form of cold rolled sheets with 0?5 mm thickness. The tensile specimens (30 mm gage length,

Materials Science and Technology

2011

VOL

27

NO

12

Behjati et al.

Strengthening mechanisms of AISI 304L austenitic stainless steel

1 a true stress–true strain and b strain hardening rate–true strain response of alloy used in present work

4 mm width and 0?5 mm thickness) cut from the cold rolled sheets were tested isothermally on a TTDM Instron machine at a strain rate of 1023 s21, temperature of 77 K and various strains. Three samples were tested for each strain level. A CX51 Olympus optical microscope (OM) was used to study the microstructure of the samples. Specimens for OM study were first electrolytically polished using a solution consisting of 350 mL H2PO4, 550 mL H2SO4 and 150 mL water at room temperature and operating voltage of 10 V. Then, they were electroetched using a solution of saturated oxalic acid in water at room temperature and operating voltage of 15 V. Nomarski differential interference contrast mode was used to study the deformation markings. Specimens of transmission electron microscopy (TEM) were prepared through thinning of 0?1 mm thickness slices obtained by mechanical grinding. Thinning was carried out by dual jet electropolishing technique at y50–60 V in a medium (with a constant temperature of 280 K) containing one part perchloric acid and four parts acetic acid. Observations were made on a Philips EM 300 TEM operating at 100 keV voltage. The formation of e phase was followed by X-ray diffraction using Co Ka1 radiation. The c and e phases are paramagnetic in contrast to the a9phase, which is strongly ferromagnetic and the only magnetic phase in the austenitic stainless steels. Therefore, the volume fraction of the a9 phase can be measured using a Feritscope (model MP30 in the present

work). To measure the a9-martensite content of the samples, the Feritscope readings were converted to actual martensite contents using the following equation16 a’-martensite content~1:71|Feritscope reading

(1)

Results Mechanical response Figure 1 shows the true stress s–true strain e and strain hardening behaviour of the alloy used in the present work respectively. The strain hardening rate h–strain response was numerically calculated from the stress–strain curve (Fig. 1a). Figure 1b clearly shows three distinct regimes of hardening, being very similar to that previously reported for other low stacking fault energy (SFE) alloys.17,18 The hardening rate initially decreases up to a strain of y6% (stage I). Then, a second stage of increasing hardening rate begins (stage II). At strains larger than y25%, stage III with a decreasing hardening rate follows.

Martensite formation In the present work, it was observed that the a9 phase grows at the expense of the e phase. In other words, with the increase in strain intensity of the e phase, X-ray diffraction peaks increase first, reach a maximum and then decrease with further strain. Figure 2a shows an Xray diffraction pattern of a sample deformed to a strain of y5%, evidencing considerable formation of the e

2 a X-ray diffraction pattern of sample deformed to strain of y5% and b formation of e and a9 phases as function of true strain for samples deformed up to 20% strain


2011

VOL

27

NO

12

1829

Behjati et al.


3 a formation of a9-martensite and b rate of a9-martensite transformation as function of true strain

phase. Figure 2b shows the formation of e and a9 phases as a function of true strain for the samples deformed up to 20% strain. It can be seen that a fraction of emartensite reaches to its maximum at a strain of y6%. Figure 3a shows the formation of a9-martensite as a function of true strain for the samples deformed up to 34% strain (rupture strain). It is observed that martensite formation follows a sigmoidal trend, reaching a saturation value before fracture. Figure 3b shows the rate of a9-martensite transformation as a function of true strain. The curve was derived by differentiating the transformation curve shown in Fig. 3a. Figure 3b shows that the rate of a9-martensite transformation has a distinctive peak around the strain of 17%. A comparison of Figs. 1b and 3b indicates that the rate of a9-martensite formation and strain hardening has a similar trend; however, it is clearly observed that the peak value of martensite transformation rate is reached at smaller strain than the peak value of the strain hardening rate (17% strain in comparison with 25% strain respectively), as reported by Talonen et al.16

shows the optical microstructure of the samples deformed to different levels of strain. After y3% deformation, the formation of long thin parallel sided deformation markings, consistent with (111) type habit, was observed in few grains. Figure 4a typically shows such markings in a sample deformed to 5% strain. These markings have been described variously as e-martensite,19 deformation twins or dislocation pile-ups,20 and their quantity increases with the decrease in temperature. Optical microscopy does not allow us to distinguish between these features precisely. Generally, these markings extended throughout the grain, but occasionally, some appeared to terminate within the grain. Moreover, few intersections of these markings could be seen at this level of strain (Fig. 4a). As shown in Fig. 4b, it was found that with the increase in deformation, some faulted sheets form on the previously described markings. Figure 4c shows the promotion of these faulted sheets with the increase in deformation so that at rupture strain the microstructure is almost completely covered by these sheets. Transmission electron microscopy images

Microstructure Optical microscopy images

Study of the as received material using OM showed extensive formation of annealing twins, which is a characteristic feature of low SFE alloys. The grain size of the raw material was determined to be y45 mm. Figure 4

Study of the samples using revealed planar arrangement of dislocations at different levels of deformation. Figure 5 denotes the TEM image of a sample deformed to 3?4% strain showing explicitly planar arrangement of dislocations. This planarity was found even at higher strains. Activation of more slip systems and extensive

4 Optical micrograph of sample deformed to a 5%, b 10% and c 34% strain

1830


2011

VOL

27

NO

12

Behjati et al.

5 Image (TEM) of sample deformed to 3?4% strain

formation of stacking faults were observed in austenite with the increase in deformation. In the present work, TEM analysis revealed that irregular overlapping process of stacking faults leads to the formation of emartensite. Figure 6 shows overlapping of stacking faults in a sample deformed to 5% strain. Furthermore, it was found that the deformation markings observed in Fig. 4a are bands of e-martensite. A remarkable feature observed in the microstructure of the samples was the extensive presence of LC locks. The number of these locks increased directly with strain. Figure 7 shows a typical TEM image of dislocation pileup behind the LC locks in a sample deformed to 10% strain. In the present study, in no case direct formation of a9 phase from the parent austenite was observed. It was found that the a9-martensite nucleation is essentially heterogeneous and occurs in the preferential sites, i.e. inside and particularly at the intersections of previously formed e bands (the faulted sheets observed in OM images). In other words, the deformation induced martensitic transformations occur through the cReRa9 steps. Figure 8 shows the TEM image of a sample deformed to 17% strain. The formation of a9-martensite is readily observed in the intersection of e phase bands.

Discussion Earlier studies have attributed the occurrence of a second stage of increasing hardening rate in the strain hardening rate–strain plot of austenitic stainless steels deformed at cryogenic temperature to the formation of

6 Overlapping of stacking faults in sample deformed to 5% strain


7 Image (TEM) of sample deformed to 10% strain

a9-martensite.16 Spencer et al.13 suggested that a9martensite in austenitic stainless steels acts as a reinforcing phase in two ways: (i) it supports a higher stress than taustenite under external loading (ii) although the martensite co-deforms with the austenite, it is expected that if the deformation temperature is reduced, the strength of the martensite will increase relative to that of the austenite, and the martensite will act as a more effective reinforcing phase. Considering the mixture of austenite, e-martensite and a9-martensite as a composite material, its strength can be expressed as follows (2) smix ~sc fc zse fe zsa’ fa’ or smix ~sc (1{fe {fa’ )zse fe zsa’ fa’

(3)

where smix is the strength of the mixture; sc, se and sa9 are the strengths of austenite, e-martensite and a9martensite respectively; and fc, fe and fa9 are the fraction of austenite, e-martensite and a9-martensite respectively. As shown in Fig. 2, a9-martensite has a considerably small fraction (maximum of 4%) at strains lower than 6% (the end of stage I). Therefore, the product of sa9 and fa9 (equation (2)) in this stage is small, and a9-martensite cannot be the sole reason of the stop in falling of strain hardening at the end of stage I. On the other hand, a fraction of e-martensite reaches its maximum (12?5%) at this level of deformation (Fig. 2). A higher fraction of emartensite suggests that the product of se and fe (equation (2)) in this stage is large enough to increase the strength of the alloy smix, i.e. in this stage, emartensite can act as a new strengthening source. With the increase in deformation, the a9 phase grows at the expense of the e phase, leading to the formation of initial softening trough in the true stress–true strain plot of the alloy (Fig. 1a).21 A decrease in deformation temperature increases the fraction of e phase and intensifies the occurrence of initial softening trough. A single stacking fault can be considered as an elementary e-martensite embryo. e bands of finite thickness can be formed from such embryos, either by a regular creation of stacking faults on alternate (111)c planes or, more probably, by an irregular overlapping process (Fig. 5b).22 In the latter case, faults are created randomly at first on parallel planes, and then the process changes gradually to the regular hcp stacking sequence. Hedstro¨m et al.23 investigated the formation of


2011

VOL

27

NO

12

1831

Behjati et al.


a bright field; b dark field image. 8 Image (TEM) of sample deformed to 17% strain

e-martensite in individual austenite grains embedded in the bulk of a metastable austenitic stainless steel. They found that the formation of e-martensite is highly localised, and only one out of 47 austenite grains formed e-martensite, which is consistent with the authors’ observations in the present work. Intersections of shear bands (faults, twins and e-martensite) are suitable sites for the nucleation of a9-martensite.24 An increase in deformation activates more slip system and enhances the number of stacking faults and also their width; therefore, the increase in deformation promotes the number of potential sites (such as intersections of previously formed e bands) for the nucleation of a9-martensite (Fig. 8). The formation of a9-martensite reinforces austenite and increases the strain hardening of the alloy, which can explain the reason of strain hardening increase in stage II of Fig. 1b. Although the formation of a9-martensite is the major reason for the high strain hardening values of the alloy in stage II, it is not the sole reason. This is due to the fact that the peak value of martensite transformation rate is reached at considerably smaller strain than the peak value of the strain hardening rate. Recently, it has been suggested that LC locks may stop the strain hardening to fall to lower values in fcc alloys with planar arrangement of dislocations.20 Lomer–Cottrell locks can act as strong barriers to the glide of active dislocations on the intersecting planes that form the lock (Fig 7). The effectiveness of these slip barriers in fcc polycrystals is expected to increase with the tendency of the material to slip planarity. In the present work, the significant planar arrangement of dislocations (Fig. 5) can be due to the considerable decrease in SFE value and increase in Peierls stress of slip systems at cryogenic temperature. With the decrease in SFE, the equilibrium distance between Shockley partials, 1/6 S112- T, increases, which makes cross-slip more difficult, confining dislocations to their original {111} slip plane. Therefore, LC locks can retard the fall of strain hardening values (the onset of stage III), explaining the discrepancy between corresponding strains of the peak value of martensite transformation rate and the peak value of strain hardening rate. At higher strains, the nucleation of a9-martensite becomes more difficult, and Shockley partials receive enough stress to recombine and pass the LC locks more easily, resulting in the decrease in strain hardening and the appearance of stage III.

1832


2011

VOL

27

NO

12

Conclusions In the present work, the strengthening mechanisms of an AISI 304L austenitic stainless steel during cryogenic deformation was studied. In addition to a9-martensite, contributions of e-martensite and LC locks at different levels of deformation were discussed. The authors’ findings revealed that these factors can play an important role in the flow behaviour of austenitic stainless steels at low temperatures. This suggests that, considering their contributions in flow, models can provide more accurate predictions of low temperature deformation behaviour of austenitic stainless steels.

References 1. W. F. Smith: ‘Structure and properties of engineering alloys’, 256– 261; 1981, New York, McGraw-Hill Inc. 2. M. R. Rochaa and C. A. S. Oliveira: Mater. Sci. Eng., A, 2009, A517, 281–285. 3. I. Me´sza ´ros and J. Proha ´szka: J. Mater. Process. Technol., 2005, 161, 162–168. 4. J. A. Lichtenfeld, M. C. Mataya and C. J. V. Tyne: Metall. Mater. Trans. A, 2006, 37A, 147–61. 5. J. Talonen and H. Ha¨nninen: Acta Mater., 2007, 55, 6108–6118. 6. K. Mumtaz, S. Takahashi, J. Echigoya, L. Zhang, Y. Kamada and M. Sato: J. Mater. Sci. Lett., 2003, 22, 423–427. 7. Q. X. Dai, X. N. Cheng, X. M. Luo and Y. T. Zhao: Mater. Charact., 2003, 49, 367–371. 8. Q. X. Dai, X. N. Cheng, Y. T. Zhao, X. M. Luo and Z. Z. Yuan: Mater. Charact., 2004, 52, 349–354. 9. K.Tomimura, S. Takaki, S.Tanimotoand Y.Tokunaga:ISIJInt., 1991,31, 721–727. 10. M. Eskandari, A. Najafizadeh and A. Kermanpur: Mater. Sci. Eng., A, 2009, A519, 46–50. 11. A. DI Schino, I. Salvatori and J. M. Kenny: J. Mater. Sci., 2002, 37, 4561–4565. 12. M. Shimojo, T. Inamura, T. H. Myeong, K. Takashima and Y. Higo: Metall. Mater. Trans. A, 2001, A32, 261–265. 13. K. Spencer, J. D. Embury, K. T. Conlon, M. Ve´ron and Y. Bre´chet: Mater. Sci. Eng., A, 2004, A387–389, 873–881. 14. M. Eskandari, A. Najafizadeh, A. Kermanpur and M. Karimi: Mater. Des., 2009, 30, 3869–3872. 15. R. D. K. Misra, S. Nayak, S. A. Mali, J. S. Shah, M. C. Somani and L. P. Karjalainen: Metall. Mater. Trans. A, 2010, 41A, 3–12. 16. J. Talonen, P. Nenonen, G. Pape and H. Ha¨nninen: Metall. Mater. Trans. A, 2005, 36A, 421–432. 17. S. Asgari, E. El-Danaf, S. Kalidindi and R. D. Doherty: Metall. Mater. Trans. A, 1997, 28A, 1781–1795. 18. E. El-Danaf, S. R. Kalidindi and R. D. Doherty: Metall. Mater. Trans. A, 1999, 30A, 1223–1233. 19. P. Huang and H. F. Loez: Mater. Lett., 1999, 39, 244–248. 20. F. Hamdi and S. Asgari: Metall. Mater. Trans. A, 2008, 39A, 294–303. 21. K. Datta, R. Delhez, P. M. Bronsveld, J. Beyer, H. J. M. Geijselaers and J. Post: Acta Mater., 2009, 57, 3321–3326. 22. J. W. Brooks, M. H. Loretto and R. E. Smallman: Acta Metall., 1979, 27, 1829–1838. 23. P. Hedstro ¨ m, U. Lienert, J. Almer and M. Ode´n: Mater. Lett., 2008, 62, 338–340. 24. L. E. Murr and K. P. Staudhammer: Metall. Trans. A, 1982, 13A, 627–635.