- Email: [email protected]
Mechanical characterisation of microstructural evolution in 304 stainless steel subjected to high-pressure torsion with and without hydrogen pre-charging
Materials Science & Engineering A 661 (2016) 87–95
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Mechanical characterisation of microstructural evolution in 304 stainless steel subjected to high-pressure torsion with and without hydrogen pre-charging Yoji Mine a,n, Kaoru Koga a, Kazuki Takashima a, Zenji Horita b,c a
Department of Materials Science and Engineering, Kumamoto University Kurokami, Chuo-ku, Kumamoto 860-8555, Japan Department of Materials Science and Engineering, Kyushu University Motooka, Nishi-ku, Fukuoka 819-0395, Japan c WPI, International Institute for Carbon-Neutral Energy Research (WPI-I2CNER), Kyushu University Motooka, Nishi-ku, Fukuoka 819-0395, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 August 2015 Received in revised form 8 February 2016 Accepted 3 March 2016 Available online 4 March 2016
Micro-tensile tests were employed on a 304 metastable austenitic stainless steel to mechanically characterise the microstructures developed by processing through high-pressure torsion (HPT) with and without hydrogen pre-charging. The martensite formed by HPT processing of hydrogen-containing austenite exhibited low yield and tensile strengths but a high reduction of area compared to the one processed in the absence of hydrogen. This may be because dynamic martensite formed with hydrogen contains more retained austenite. Hydrogen charging into the austenite allowed the formation of ε– martensite, instead of deformation twinning, as an intermediate phase in the transformation to α′– martensite, which led to variation in the plastic behaviour. The inhomogeneity of the microstructure and the defects produced by deformation with hydrogen build a foundation but hardly play a crucial role in the hydrogen embrittlement (HE) of metastable austenitic steels. Excess hydrogen due to the dynamic martensitic transformation of hydrogen-containing austenite localises deformation in the retained austenite between the martensite regions formed, leading to the HE of metastable austenitic steels. & 2016 Elsevier B.V. All rights reserved.
Keywords: Mechanical characterisation Austenite Hydrogen embrittlement Martensitic transformations Twinning
1. Introduction With increasing requirements to reduce carbon dioxide emissions, the use of hydrogen as an energy carrier has recently become more practical. However, most metals used in harsh environments, e.g., in the marine, aerospace, nuclear, and chemical industries, often suffer from hydrogen embrittlement (HE). Unlike low-temperature embrittlement, which is accompanied by typical brittle fracture features, the fracture morphology after plastic deformation processes [1–4] has hindered the elucidation of the hydrogen-induced degradation of the mechanical properties. Therefore, studying hydrogen-induced plasticity is important to understand the HE mechanism. A previous study by Takai et al. [5] revealed that the defects produced by deformation with hydrogen, rather than the hydrogen itself, play an essential role in the hydrogen-induced degradation of the ductility in iron with a bodycentred cubic (bcc) structure or in Inconel 625 with a face-centred cubic (fcc) structure. For austenitic stainless steels of interest, HE becomes more intense with a lower stability of the austenitic phase [6–9]. Moreover, the change in the crystal structure from fcc n
Corresponding author. E-mail address: [email protected] (Y. Mine).
http://dx.doi.org/10.1016/j.msea.2016.03.018 0921-5093/& 2016 Elsevier B.V. All rights reserved.
to bcc through deformation, i.e., the deformation-induced martensitic transformation, complicates the HE phenomenon in metastable austenitic steels. We focused on the mechanical characteristics of the microstructure developed by deformation in a hydrogen-containing metastable austenitic steel. A previous study by Mine et al. using high-pressure torsion (HPT) processing on hydrogen-pre-charged 304 and 316L stainless steels [10] revealed that solute hydrogen reduced the deformation-induced α′–martensitic transformation, which resulted in a reduction in the microhardness. The decreased microhardness in the hydrogen-precharged microstructure was attributed to its low fraction of the α′– martensitic phase. However, the solute hydrogen not only reduced the deformation-induced martensitic transformation but might also have affected the development of the resulting microstructure, as reported by Takai et al. for stable phase metals [5]. In the case where a second phase such as α′–martensite is dynamically formed within the parent phase, it is difficult to evaluate the mechanical characteristics of each phase using conventional mechanical testing. In contrast, micro-tensile testing enables the evaluation of the strength and ductility properties of microstructural constituents on the order of several tens of micrometres [11–16]. The objective of this study was to clarify how the defects produced through deformation with hydrogen contribute to the HE process of a metastable austenitic stainless steel.
88
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 1. (a) True stress–strain curves for the martensite and two-phase specimens and (b) their respective processing histories. The arrows in (a) indicate the onsets of necking.
For this purpose, micro-tensile testing was employed for deformation-induced martensite specimens and retained austenite specimens prepared from the microstructures developed by HPTprocessing of a type 304 austenitic stainless steel, with and without hydrogen pre-charging.
2. Material and experimental methods The material used in this study was a 304 (JIS-SUS304) commercial austenitic stainless steel, which was composed of 0.05C, 18.54Cr, 8.09Ni, 0.58Si, 1.24Mn, 0.025P, and 0.003S (in mass%), with the remainder being Fe. It was received in the form of a plate that was 30 mm in thickness after solution treatment. The Vickers hardness in the as-received condition was 1767 10, where the error range represents the 95% confidence interval. Disc-shaped samples with a diameter of 19 mm and an approximate thickness of 0.8 mm were machined from the plate. Discs with and without hydrogen pre-charging were processed by HPT at room temperature in air; the corresponding discs are denoted as H-HPT and U-HPT discs. Hydrogen pre-charging was undertaken at a temperature of 543 K by exposure to hydrogen gas for 200 h at a pressure of 10 MPa. These conditions for charging were sufficient to provide a nearly uniform hydrogen distribution throughout the thin specimen. The saturated hydrogen content was determined to be 25 mass ppm. The HPT facility consisted of a pair of tool-steel anvils having a centric shallow circular cavity with a diameter of 20 mm and a depth of 0.25 mm. Shear strain was imposed by rotating the lower anvil with respect to the upper one for one turn at a rotation speed of 1 rpm under a pressure of
Fig. 2. (a, c) EBSD maps overlapped on the corresponding SEM images taken in the initial state and (b, d) fracture morphologies of the martensite specimens. Table 1 The fraction of α′–martensite, yield strength, tensile strength, and reduction of area for the martensite and two-phase specimens.
U-HPT-M H-HPT-M U-HPT-A/ M H-HPT-A/ M
α′–martensite fraction (%)
Yield strength (MPa)
Tensile strength (MPa)
Reduction of area (%)
4 99 4 99 27
1820 1690 1190
1880 1780 1340
79 87 79
59
1240
1400
86
1.5 GPa. The microstructural evolution due to HPT processing was examined by electron backscatter diffraction (EBSD) analysis. The surface of the HPT-processed discs was finished by electro-chemical polishing. The crystal orientation was determined by automatic beam scanning with a step size of 0.08–0.2 mm at an accelerating voltage of 20 kV in a field emission gun scanning electron microscope (SEM) using EBSD patterns and TSL orientation imaging microscopy software (OIM v. 7.1.0). The sample for transmission electron microscopy (TEM) was milled using a focused ion beam (FIB). The TEM observation was performed with a JEOL JEM2000FX system operated at an accelerating voltage of 200 kV. Gauge sections with dimensions of 50 mm 20 mm 20 mm of micro-tensile specimens were fabricated using FIB from the austenitic and α′–martensite phase regions and two-phase region in the H-HPT and U-HPT discs. The micro-tensile specimens of the
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
89
Fig. 3. (a, c) Phase maps with image quality overlapped on the corresponding SEM images taken in the initial state and (b, d) fracture morphologies of the two-phase specimens. The insets in (d) show EBSD maps taken on the longitudinal crosssection of the fractured specimens.
martensite, austenite, and two-phase microstructures are distinguished by the suffixes ‘-M’, ‘-A’, and ‘-A/M’, respectively. The tensile test was conducted at a displacement rate of 0.1 mm s 1, which corresponded to an initial strain rate of 2 10 3 s 1, at room temperature in atmospheric air. The gauge section of the tensile specimen was monitored during tensile testing using an optical microscope to dynamically measure the strain as a function of time. After failure, longitudinal cross-sections of some specimens were fabricated using FIB for EBSD analyses.
3. Results and discussion 3.1. Deformation behaviour of α′–martensite formed through HPT with and without hydrogen pre-charging Fig. 1 shows the true stress–strain curves for the martensite and two-phase specimens obtained from the U-HPT and H-HPT discs. The true stress, s, and true strain, ε, were calculated from the nominal stress, sn, and nominal strain, εn, using s ¼sn(1 þ εn) and ε ¼ln(1 þ εn). These true stress–strain values are not valid after the onset of necking since the geometric change during necking was
Fig. 4. (a–c) Optical micrographs showing the deformation process of the H-HPT-A/ M specimen and (d) (111) and (110) pole figures initially obtained from an austenite grain, in which yielding commenced. Solid and dotted lines indicate the traces of the activated and primary slip systems, respectively.
not considered. Fig. 1b shows a schematic indicating the processing histories of the specimens. Fig. 2 shows the EBSD maps overlapped on the corresponding SEM images in the initial state and the fracture morphologies of the martensite specimens. In both martensite specimens, necking led to fracture (Fig. 2b and d) through slight work hardening after the onset of yielding at stresses of 1690–1820 MPa (Fig. 1a). The uniform elongations were equivalent between both martensite specimens, as indicated by the arrows in Fig. 1a. The reductions of area were measured to evaluate the local ductility. As shown in Table 1, the H-HPT-M specimen exhibited low strength but a high reduction of area compared to the U-HPT-M specimen. In addition, fitting the data points between the yield strength and the ultimate tensile
90
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 5. Typical microstructures observed in the U-HPT and H-HPT discs.
Fig. 6. TEM micrographs showing the orientation relationship between austenite (γ) martensite (α′ and ε) in the H-HPT disc: (a) bright-field image, (b) electron diffraction pattern, and (c) key diagram of (b).
strength in the stress–strain curves with s ¼Kεn (K: the strength coefficient), the strain-hardening exponent n for the H-HPT-M specimen was determined to be 0.28, which was slightly lower than the value of 0.34 for the U-HPT-M specimen. A finite-element method simulation study by Maresca et al. revealed that the
austenite between the martensite laths decreased the critical shear stress in a carbon steel [17]. Taking this into consideration, the lower yield strength and lower strain-hardening exponent of the H-HPT-M specimen may be attributed to a larger amount of retained austenite between the martensite laths formed. Moreover, it can be shown in Fig. 2 that the deformation was localised in slightly coarser regions with similar crystallographic orientations in both specimens. Fig. 3 shows the phase maps prior to tensile testing of the twophase specimens and the corresponding fracture morphologies. In both two-phase specimens, although yielding commenced in the austenite regions at stresses of 1200 MPa, the final fracture occurred in the martensite regions. Therefore, the values of the reduction of area in the two-phase specimens were in accordance with their counterparts in the martensite specimens (Table 1). Unlike the martensite specimens, the two-phase specimens showed uniform elongations of several percent with moderate strain hardening (Fig. 1a). In particular, a two-step strain-hardening behaviour following yielding was observed in the H-HPT-A/ M specimen. This implies that the transition of the deformation mode occurred before necking. Fig. 4 shows the deformation process of the H-HPT-A/M specimen and the (111) and (110) pole figures initially obtained from the austenite grain, in which yielding commenced. The slip trace coincided with the trace of the slip system with a Schmid factor of 0.33 instead of the primary slip system (Schmid factor: 0.46) (compare Fig. 4b and d). This is because slip deformation on the primary system was constrained by the neighbouring martensite. After the austenite grain was fully deformed, slip transferred to the neighbouring martensite, which finally resulted in necking at the martensite (Fig. 4c). An EBSD analysis of the longitudinal cross-section of the fractured H-HPTA/M specimen revealed that the α′–martensitic transformation occurred in the deformed austenite grain, as shown in the insets of Fig. 3d. Thus, it is plausible that, when metastable austenite grains were deformed through the HPT processing with hydrogen, the formed martensite region might contain more retained austenite, while it was deformable during micro-tensile loading in the absence of hydrogen. Combined tension and hydrogen-outgassing experiments by Zhang et al. suggested [18] that dynamically formed martensite with hydrogen has a major impact on the HE of metastable austenitic stainless steels. The implications of the
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 7. (a) True stress–strain curves for the austenite specimens and (b) their respective processing histories. The arrows in (a) indicate the onsets of necking. Standard stereographic triangles show the loading directions. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
Table 2 The fraction of α′–martensite yield strength, Schmid factor, corresponding critical resolved shear stress (CRSS) for the primary slip system, tensile strength, and reduction of area for the austenite specimens. α′–martensite fraction (%)
Yield strength (MPa)
Schmid factor
CRSS (MPa)
Tensile strength (MPa)
Reduction of area (%)
U-HPT-A 4 H-HPT-A 1
938 663
0.456 0.387
427 256
1130 702
– –
169 230
0.468 0.478
79 110
487 477
90 (Chisel edge) 98 42
ST-U-A ST-H-A
α′–martensite transformation of hydrogen-containing austenite in the HE mechanism are discussed in the following section.
91
Fig. 8. (a) EBSD map overlapped on the corresponding SEM image taken in the initial state, (b) (111) and (110) pole figures of austenite and the corresponding colour-coded map, and (c) optical micrograph taken at the onset of yielding of the U-HPT-A specimen.
3.2. The development of nano-twins and slip bands and their effects on the plasticity of austenite Our previous study revealed [10] that hydrogen charging reduced the formation of α′–martensite during the subsequent process of HPT. A certain amount of austenitic phase remained even when subjected to severe plastic deformation. Fig. 5 shows typical microstructures developed by HPT processing with and without hydrogen-pre-charging. While the isolated austenite grains corresponded to a fraction of the U-HPT disc, deformation twinning occurred, which was often accompanied by lamellar plates of α′–martensite (Fig. 5a and b). In the H-HPT disc, welldeveloped slip bands were observed in most austenite grains, as shown in Fig. 5c and d. Fig. 6 shows the TEM micrographs taken on the deformed austenite grain in the H-HPT disc. Unlike the U-HPT disc, nano-twins were not observed in the H-HPT disc.
92
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 9. (a) EBSD map overlapped on the corresponding SEM image taken in the initial state, (b) (111) and (110) pole figures of austenite, and (c) optical micrograph taken at the onset of yielding of the H-HPT-A specimen.
Alternatively, there were some places where spots of α′ and ε–martensite were visible. Twinning has been reported to occur in austenitic steels with an intermediate stacking fault energy (SFE) [19,20]. Shen et al. also verified that ε–martensite and deformation twins only act as an intermediate phase in the transformation from austenite to α′–martensite in a type 304 stainless steel [20]. X-ray diffraction measurements by Pontini and Hermida revealed that hydrogen charging into a type 304 stainless steel decreased the SFE [21]. Moreover, a calculation study by Hermida and Roviglione explained that this SFE decrease allows the nucleation of the ε–martensite through H–H pair formation [22]. These facts suggest that the transformation from austenite to α′–martensite occurred via the γ–twinning preferentially in the uncharged state of the present steel, whereas solute hydrogen facilitated the formation of the ε–martensite as an intermediate phase by decreasing the SFE. Fig. 7 shows the true stress–strain curves for the austenitic phase specimens and a schematic indicating their corresponding processing histories. The U-HPT-A and H-HPT-A specimens were prepared from the microstructures shown in Fig. 5, the nanotwinned and slip-deformed microstructures, respectively. The standard stereographic triangles in Fig. 7a indicate the initial loading directions. In the case of the U-HPT-A specimen, blue and red show the parent and twinned crystals, respectively. For comparison, the results of uncharged and hydrogen-charged solution-
treated specimens [23], denoted as ST-U-A and ST-H-A, respectively, are included in Fig. 7. The ST-U-A specimen did not undergo HPT-processing or hydrogen charging (Fig. 7b). The ST-H-A specimen was tested within 3 h after hydrogen cathodic charging, which provided a saturated hydrogen content of 100 mass ppm [23]. While there might have remained very little hydrogen during micro-tensile testing in the H-HPT-A specimen, the ST-H-A specimen should have contained a considerable amount of hydrogen. Table 2 summarises the α′–martensite fraction, the yield strength, the Schmid factor, and corresponding critical resolved shear stress (CRSS) for the primary slip system, the tensile strength, and the reduction of area for the austenite specimens. The yield strengths of the U-HPT-A and H-HPT-A specimens were 940 MPa and 660 MPa, which were significantly higher than the value of 170 MPa for the ST-U-A specimen. Figs. 8 and 9 show the EBSD maps and the (111) and (110) pole figures in the initial state and the optical micrographs taken at the yielding points for the U-HPT-A and H-HPT-A specimens, respectively. In the U-HPT-A specimen, the plane of the nano-twins was arranged nearly parallel to the loading direction (Fig. 8b). The slip trace that appeared at the onset of yielding corresponded to the primary slip planes in the parent and twinned crystals (compare Fig. 8b and c). Assuming that macroscopic yielding occurred in the parent crystal, the apparent CRSS for the U-HPT-A specimen was determined to be 430 MPa, which is more than five times higher than the CRSS of 79 MPa for the ST-U-A specimen (Table 2). A micro-tensile testing study by Mine et al. using a nano-twinned stable austenitic alloy revealed [24] that introducing nano-twins with an average thickness of 20 nm mainly contributed to the strengthening of this alloy through the inhibition of slip transfer by the twin boundaries. As for the H-HPT-A specimen pre-hardened by slip deformation and/or the precipitation of ε–martensite in the 304 steel, the CRSS was determined to be 280 MPa. In addition, although the mutual interaction of dislocations introduced by the process of HPT could not be neglected, the inhibition of dislocation movement by twin boundaries was a major strengthening mechanism in the nano-twinned 304 steel. Moreover, despite the significant strengthening, a moderate ductility was retained in the U-HPT-A specimen (Fig. 7a and Table 2). This suggests that introducing nano-twins into austenitic steels can provide a good balance between the strength and ductility. The U-HPT-A specimen exhibited a two-step hardening behaviour (Fig. 7a). This might be attributed to the interaction of α′– martensite with the twin boundaries. Fig. 10 shows the EBSD maps and the (110) and (111) pole figures of the α′–martensite taken from the longitudinal cross-section of the fractured U-HPT-A specimen. Three martensite variants with habit planes parallel to the plane of the nano-twins were mainly formed (Fig. 10c and d). Fig. 11 shows the fracture morphology and the (111) and (110) pole figures of the austenite and martensite taken from the longitudinal cross-section of the fractured ST-U-A specimen. Unlike the nanotwinned specimen, two martensite variants with habit planes parallel to the primary slip plane were prevalent in the ST-U-A specimen (Fig. 11c). This finding coincided with the results of previous works by Higo et al. [25] and Kato and Mori [26]. According to their reports, assuming a double shearing mechanism, the critical first shear initiating the α′–martensitic transformation determined the habit plane of the formed martensite variants. In fact, the α′–martensite variants with habit planes parallel to the highest shearing plane were selected for the ST-U-A specimen. However, in the U-HPT-A specimen, i.e., in the presence of nanotwins, martensite variants nucleating from the twin plane were prevalent rather than ones with habit planes parallel to the highest shearing plane (Fig. 10). Nakada et al. reported [27] that three variants of the deformation-induced martensite were formed on the plane of deformation twinning with a double K-S
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
93
Fig. 10. (a) EBSD map, (b) phase map with image quality, (c) colour-coded map corresponding to (d), and (d) (110) and (111) pole figures of α′–martensite of the fractured U-HPT-A specimen.
relationship with both the parent and the twinned crystals of austenite in a cold-rolled 316 steel. Fig. 10 shows that the three martensite variants formed through tensile testing of the nanotwinned 304 steel had a double K-S relationship similar to that of the cold-rolled 316 steel. Moreover, a magnetic force microscopy study by Zhang et al. revealed [28] that deformation-induced martensite was nucleated from annealed twin boundaries in a 316 steel, which was deformed at a temperature of 200 K. Therefore, it is concluded that, regardless of deformation or annealing, introducing twin boundaries changes the route of the deformationinduced martensitic transformation in metastable austenitic steels. In the ST-H-A specimen, it is assumed that hydrogen facilitates the dislocation movement on the primary plane in the plateau stress regime following the yielding, which leads to the stress drop and premature martensite formation within the austenitic grains [23]. The stress–strain curve for the H-HPT-A specimen exhibited a plateau regime following yielding similar to that of the ST-H-A specimen (Fig. 7a). This is because the microstructure developed in
the H-HPT-A specimen contained the ε–martensite instead of the nano-twins, as shown in Fig. 6. If the deformation process proceeds in the presence of hydrogen, hydrogen can be enriched in the retained lamellar austenite owing to the martensitic transformation, resulting in quasi-cleavage fracture. However, in the present case, hydrogen had diffused out of the H-HPT-A specimen before the micro-tensile test; therefore, work hardening without hydrogen did not lead to severe hydrogen embrittlement with quasi-cleavage, as shown in the ST-H-A specimen [23]. This implies that the defects produced by deformation with hydrogen may not play a crucial role but rather build a foundation in the HE mechanism. As mentioned above, Zhang et al. reported [18] that dynamic α′–martensite formed during deformation mainly contributes to HE. It was also argued that austenite transforms into α ′–martensite during deformation with hydrogen, thereby leading to easier diffusion of hydrogen [29]. These findings supported the hypothesis that excess hydrogen due to martensitic transformation, which corresponds to the difference in the solubility between
94
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 11. (a) Fracture morphology of the ST-U-A specimen and (b, c) (111) and (110) pole figures of austenite and α′–martensite, respectively.
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
austenite and martensite, promotes the deformation localisation in the retained austenitic phase, resulting in HE in metastable austenitic steels [30].
95
acknowledges support from the ‘Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers' R2608. This study used facilities for severe plastic deformation at the International Research Centre on Giant Straining for Advanced Materials (IRC-GSAM) at Kyushu University.
4. Summary Micro-tensile testing was used in this study to mechanically characterise microstructures deformed through high-pressure torsion of hydrogen-pre-charged and uncharged samples of a 304 austenitic stainless steel. The main findings can be summarised as follows: (1) The martensite specimen formed by deformation with hydrogen exhibited low strength but a high reduction of area when compared to the one formed in the absence of hydrogen. This suggests that austenite regions are more likely to remain in martensite formed by deformation with hydrogen. (2) Deformation-twinning in the uncharged 304 austenitic steel acts as an intermediate phase in the transformation to α′– martensite, which contributes to a good balance between the strength and ductility. (3) Hydrogen charging in the 304 austenitic stainless steel facilitates the formation of ε–martensite as an intermediate phase in the transformation from austenite to α′–martensite, thereby leading to the stress drop following yielding in the stress– strain behaviour characteristic of the hydrogen-charged austenitic specimens. However, dynamic α′–martensitic transformation in the presence of hydrogen is requisite for severe HE with quasi-cleavage.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]
The defects and the inhomogeneity of the microstructure provided by deformation with hydrogen trigger hydrogen embrittlement in metastable austenitic steel but cannot play the crucial role therein.
[19] [20]
Acknowledgements
[25] [26] [27]
The authors thank Mr. R. Matsuoka, Dr. M. Tsushida, and Dr. T. Yamamuro, Kumamoto University for their assistance with the TEM observation. The present work was supported in part by a Grant-in-Aid for Scientific Research (C) 25420758 from the Japan Society for the Promotion of Science (JSPS). YM gratefully
[21] [22] [23] [24]
[28] [29] [30]
C.D. Beachem, Metall. Trans. 3 (1972) 437–451. S.P. Lynch, Acta Metall. 36 (1988) 2639–2661. H.K. Birnbaum, P. Sofronis, Mater. Sci. Eng. A 176 (1994) 191–202. I.M. Robertson, Eng. Fract. Mech. 68 (2001) 671–692. K. Takai, H. Shoda, H. Suzuki, M. Nagumo, Acta Mater. 56 (2008) 5158–5167. D. Eliezer, D.G. Chakrapani, C.J. Altstetter, E.N. Pugh, Metall. Trans. A 10 (1979) 935–941. G. Han, J. He, S. Fukuyama, K. Yokogawa, Acta Mater. 46 (1998) 4559–4570. C. San Marchi, B.P. Somerday, X. Tang, G.H. Schiroky, Int. J. Hydrog. Energy 33 (2008) 889–904. Y. Mine, T. Kimoto, Corros. Sci. 53 (2011) 2619–2629. Y. Mine, Z. Horita, Y. Murakami, Acta Mater. 57 (2009) 2993–3002. Y. Mine, K. Hirashita, M. Matsuda, M. Otsu, K. Takashima, Corros. Sci. 53 (2011) 529–533. Y. Mine, H. Fujisaki, M. Matsuda, M. Takeyama, K. Takashima, Scr. Mater. 65 (2011) 707–710. Y. Mine, K. Hirashita, M. Matsuda, K. Takashima, Metall. Mater. Trans. A 42 (2011) 3567–3571. Y. Mine, K. Hirashita, H. Takashima, M. Matsuda, K. Takashima, Mater. Sci. Eng. A 560 (2013) 535–544. Y. Mine, H. Takashima, M. Matsuda, K. Takashima, Mater. Sci. Eng. A 618 (2014) 359–367. Y. Mine, R. Maezono, H. Oda, M. Yamasaki, Y. Kawamura, K. Takashima, Mater. Trans. 56 (2015) 952–956. F. Maresca, V.G. Kouznetsova, M.G.D. Geers, J. Mech. Phys. Solids 73 (2014) 69–83. L. Zhang, Z. Li, J. Zheng, Y. Zhao, P. Xu, C. Zhou, X. Li, Int. J. Hydrog. Energy 38 (2013) 8208–8214. J.-Y. Choi, W. Jin, Scr. Mater. 36 (1997) 99–104. Y.F. Shen, X.X. Li, X. Sun, Y.D. Wang, L. Zuo, Mater. Sci. Eng. A 552 (2012) 514–522. A.E. Pontini, J.D. Hermida, Scr. Mater. 37 (1997) 1831–1837. J.D. Hermida, A. Roviglione, Scr. Mater. 39 (1998) 1145–1149. Y. Mine, K. Koga, O. Kraft, K. Takashima, Scr. Mater. 113 (2016) 176–179. Y. Mine, S. Nakamichi, K. Koga, K. Takashima, O. Kraft, To be submitted Elsewhere, 2015. Y. Higo, F. Lecroisey, T. Mori, Acta Metall. 22 (1974) 313–323. M. Kato, T. Mori, Acta Metall. 24 (1976) 853–860. N. Nakada, H. Ito, Y. Matsuoka, T. Tsuchiyama, S. Takaki, Acta Mater. 58 (2010) 895–903. L. Zhang, M. Imade, B. An, M. Wen, T. Iijima, S. Fukuyama, K. Yokogawa, ISIJ Int. 52 (2012) 240–246. J.H. Ryu, Y.S. Chun, C.S. Lee, H.K.D.H. Bhadeshia, D.W. Suh, Acta Mater. 60 (2012) 4085–4092. Y. Mine, Z. Horita, Mater. Trans. 53 (2012) 773–785.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Mechanical characterisation of microstructural evolution in 304 stainless steel subjected to high-pressure torsion with and without hydrogen pre-charging Yoji Mine a,n, Kaoru Koga a, Kazuki Takashima a, Zenji Horita b,c a
Department of Materials Science and Engineering, Kumamoto University Kurokami, Chuo-ku, Kumamoto 860-8555, Japan Department of Materials Science and Engineering, Kyushu University Motooka, Nishi-ku, Fukuoka 819-0395, Japan c WPI, International Institute for Carbon-Neutral Energy Research (WPI-I2CNER), Kyushu University Motooka, Nishi-ku, Fukuoka 819-0395, Japan b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 August 2015 Received in revised form 8 February 2016 Accepted 3 March 2016 Available online 4 March 2016
Micro-tensile tests were employed on a 304 metastable austenitic stainless steel to mechanically characterise the microstructures developed by processing through high-pressure torsion (HPT) with and without hydrogen pre-charging. The martensite formed by HPT processing of hydrogen-containing austenite exhibited low yield and tensile strengths but a high reduction of area compared to the one processed in the absence of hydrogen. This may be because dynamic martensite formed with hydrogen contains more retained austenite. Hydrogen charging into the austenite allowed the formation of ε– martensite, instead of deformation twinning, as an intermediate phase in the transformation to α′– martensite, which led to variation in the plastic behaviour. The inhomogeneity of the microstructure and the defects produced by deformation with hydrogen build a foundation but hardly play a crucial role in the hydrogen embrittlement (HE) of metastable austenitic steels. Excess hydrogen due to the dynamic martensitic transformation of hydrogen-containing austenite localises deformation in the retained austenite between the martensite regions formed, leading to the HE of metastable austenitic steels. & 2016 Elsevier B.V. All rights reserved.
Keywords: Mechanical characterisation Austenite Hydrogen embrittlement Martensitic transformations Twinning
1. Introduction With increasing requirements to reduce carbon dioxide emissions, the use of hydrogen as an energy carrier has recently become more practical. However, most metals used in harsh environments, e.g., in the marine, aerospace, nuclear, and chemical industries, often suffer from hydrogen embrittlement (HE). Unlike low-temperature embrittlement, which is accompanied by typical brittle fracture features, the fracture morphology after plastic deformation processes [1–4] has hindered the elucidation of the hydrogen-induced degradation of the mechanical properties. Therefore, studying hydrogen-induced plasticity is important to understand the HE mechanism. A previous study by Takai et al. [5] revealed that the defects produced by deformation with hydrogen, rather than the hydrogen itself, play an essential role in the hydrogen-induced degradation of the ductility in iron with a bodycentred cubic (bcc) structure or in Inconel 625 with a face-centred cubic (fcc) structure. For austenitic stainless steels of interest, HE becomes more intense with a lower stability of the austenitic phase [6–9]. Moreover, the change in the crystal structure from fcc n
Corresponding author. E-mail address: [email protected] (Y. Mine).
http://dx.doi.org/10.1016/j.msea.2016.03.018 0921-5093/& 2016 Elsevier B.V. All rights reserved.
to bcc through deformation, i.e., the deformation-induced martensitic transformation, complicates the HE phenomenon in metastable austenitic steels. We focused on the mechanical characteristics of the microstructure developed by deformation in a hydrogen-containing metastable austenitic steel. A previous study by Mine et al. using high-pressure torsion (HPT) processing on hydrogen-pre-charged 304 and 316L stainless steels [10] revealed that solute hydrogen reduced the deformation-induced α′–martensitic transformation, which resulted in a reduction in the microhardness. The decreased microhardness in the hydrogen-precharged microstructure was attributed to its low fraction of the α′– martensitic phase. However, the solute hydrogen not only reduced the deformation-induced martensitic transformation but might also have affected the development of the resulting microstructure, as reported by Takai et al. for stable phase metals [5]. In the case where a second phase such as α′–martensite is dynamically formed within the parent phase, it is difficult to evaluate the mechanical characteristics of each phase using conventional mechanical testing. In contrast, micro-tensile testing enables the evaluation of the strength and ductility properties of microstructural constituents on the order of several tens of micrometres [11–16]. The objective of this study was to clarify how the defects produced through deformation with hydrogen contribute to the HE process of a metastable austenitic stainless steel.
88
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 1. (a) True stress–strain curves for the martensite and two-phase specimens and (b) their respective processing histories. The arrows in (a) indicate the onsets of necking.
For this purpose, micro-tensile testing was employed for deformation-induced martensite specimens and retained austenite specimens prepared from the microstructures developed by HPTprocessing of a type 304 austenitic stainless steel, with and without hydrogen pre-charging.
2. Material and experimental methods The material used in this study was a 304 (JIS-SUS304) commercial austenitic stainless steel, which was composed of 0.05C, 18.54Cr, 8.09Ni, 0.58Si, 1.24Mn, 0.025P, and 0.003S (in mass%), with the remainder being Fe. It was received in the form of a plate that was 30 mm in thickness after solution treatment. The Vickers hardness in the as-received condition was 1767 10, where the error range represents the 95% confidence interval. Disc-shaped samples with a diameter of 19 mm and an approximate thickness of 0.8 mm were machined from the plate. Discs with and without hydrogen pre-charging were processed by HPT at room temperature in air; the corresponding discs are denoted as H-HPT and U-HPT discs. Hydrogen pre-charging was undertaken at a temperature of 543 K by exposure to hydrogen gas for 200 h at a pressure of 10 MPa. These conditions for charging were sufficient to provide a nearly uniform hydrogen distribution throughout the thin specimen. The saturated hydrogen content was determined to be 25 mass ppm. The HPT facility consisted of a pair of tool-steel anvils having a centric shallow circular cavity with a diameter of 20 mm and a depth of 0.25 mm. Shear strain was imposed by rotating the lower anvil with respect to the upper one for one turn at a rotation speed of 1 rpm under a pressure of
Fig. 2. (a, c) EBSD maps overlapped on the corresponding SEM images taken in the initial state and (b, d) fracture morphologies of the martensite specimens. Table 1 The fraction of α′–martensite, yield strength, tensile strength, and reduction of area for the martensite and two-phase specimens.
U-HPT-M H-HPT-M U-HPT-A/ M H-HPT-A/ M
α′–martensite fraction (%)
Yield strength (MPa)
Tensile strength (MPa)
Reduction of area (%)
4 99 4 99 27
1820 1690 1190
1880 1780 1340
79 87 79
59
1240
1400
86
1.5 GPa. The microstructural evolution due to HPT processing was examined by electron backscatter diffraction (EBSD) analysis. The surface of the HPT-processed discs was finished by electro-chemical polishing. The crystal orientation was determined by automatic beam scanning with a step size of 0.08–0.2 mm at an accelerating voltage of 20 kV in a field emission gun scanning electron microscope (SEM) using EBSD patterns and TSL orientation imaging microscopy software (OIM v. 7.1.0). The sample for transmission electron microscopy (TEM) was milled using a focused ion beam (FIB). The TEM observation was performed with a JEOL JEM2000FX system operated at an accelerating voltage of 200 kV. Gauge sections with dimensions of 50 mm 20 mm 20 mm of micro-tensile specimens were fabricated using FIB from the austenitic and α′–martensite phase regions and two-phase region in the H-HPT and U-HPT discs. The micro-tensile specimens of the
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
89
Fig. 3. (a, c) Phase maps with image quality overlapped on the corresponding SEM images taken in the initial state and (b, d) fracture morphologies of the two-phase specimens. The insets in (d) show EBSD maps taken on the longitudinal crosssection of the fractured specimens.
martensite, austenite, and two-phase microstructures are distinguished by the suffixes ‘-M’, ‘-A’, and ‘-A/M’, respectively. The tensile test was conducted at a displacement rate of 0.1 mm s 1, which corresponded to an initial strain rate of 2 10 3 s 1, at room temperature in atmospheric air. The gauge section of the tensile specimen was monitored during tensile testing using an optical microscope to dynamically measure the strain as a function of time. After failure, longitudinal cross-sections of some specimens were fabricated using FIB for EBSD analyses.
3. Results and discussion 3.1. Deformation behaviour of α′–martensite formed through HPT with and without hydrogen pre-charging Fig. 1 shows the true stress–strain curves for the martensite and two-phase specimens obtained from the U-HPT and H-HPT discs. The true stress, s, and true strain, ε, were calculated from the nominal stress, sn, and nominal strain, εn, using s ¼sn(1 þ εn) and ε ¼ln(1 þ εn). These true stress–strain values are not valid after the onset of necking since the geometric change during necking was
Fig. 4. (a–c) Optical micrographs showing the deformation process of the H-HPT-A/ M specimen and (d) (111) and (110) pole figures initially obtained from an austenite grain, in which yielding commenced. Solid and dotted lines indicate the traces of the activated and primary slip systems, respectively.
not considered. Fig. 1b shows a schematic indicating the processing histories of the specimens. Fig. 2 shows the EBSD maps overlapped on the corresponding SEM images in the initial state and the fracture morphologies of the martensite specimens. In both martensite specimens, necking led to fracture (Fig. 2b and d) through slight work hardening after the onset of yielding at stresses of 1690–1820 MPa (Fig. 1a). The uniform elongations were equivalent between both martensite specimens, as indicated by the arrows in Fig. 1a. The reductions of area were measured to evaluate the local ductility. As shown in Table 1, the H-HPT-M specimen exhibited low strength but a high reduction of area compared to the U-HPT-M specimen. In addition, fitting the data points between the yield strength and the ultimate tensile
90
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 5. Typical microstructures observed in the U-HPT and H-HPT discs.
Fig. 6. TEM micrographs showing the orientation relationship between austenite (γ) martensite (α′ and ε) in the H-HPT disc: (a) bright-field image, (b) electron diffraction pattern, and (c) key diagram of (b).
strength in the stress–strain curves with s ¼Kεn (K: the strength coefficient), the strain-hardening exponent n for the H-HPT-M specimen was determined to be 0.28, which was slightly lower than the value of 0.34 for the U-HPT-M specimen. A finite-element method simulation study by Maresca et al. revealed that the
austenite between the martensite laths decreased the critical shear stress in a carbon steel [17]. Taking this into consideration, the lower yield strength and lower strain-hardening exponent of the H-HPT-M specimen may be attributed to a larger amount of retained austenite between the martensite laths formed. Moreover, it can be shown in Fig. 2 that the deformation was localised in slightly coarser regions with similar crystallographic orientations in both specimens. Fig. 3 shows the phase maps prior to tensile testing of the twophase specimens and the corresponding fracture morphologies. In both two-phase specimens, although yielding commenced in the austenite regions at stresses of 1200 MPa, the final fracture occurred in the martensite regions. Therefore, the values of the reduction of area in the two-phase specimens were in accordance with their counterparts in the martensite specimens (Table 1). Unlike the martensite specimens, the two-phase specimens showed uniform elongations of several percent with moderate strain hardening (Fig. 1a). In particular, a two-step strain-hardening behaviour following yielding was observed in the H-HPT-A/ M specimen. This implies that the transition of the deformation mode occurred before necking. Fig. 4 shows the deformation process of the H-HPT-A/M specimen and the (111) and (110) pole figures initially obtained from the austenite grain, in which yielding commenced. The slip trace coincided with the trace of the slip system with a Schmid factor of 0.33 instead of the primary slip system (Schmid factor: 0.46) (compare Fig. 4b and d). This is because slip deformation on the primary system was constrained by the neighbouring martensite. After the austenite grain was fully deformed, slip transferred to the neighbouring martensite, which finally resulted in necking at the martensite (Fig. 4c). An EBSD analysis of the longitudinal cross-section of the fractured H-HPTA/M specimen revealed that the α′–martensitic transformation occurred in the deformed austenite grain, as shown in the insets of Fig. 3d. Thus, it is plausible that, when metastable austenite grains were deformed through the HPT processing with hydrogen, the formed martensite region might contain more retained austenite, while it was deformable during micro-tensile loading in the absence of hydrogen. Combined tension and hydrogen-outgassing experiments by Zhang et al. suggested [18] that dynamically formed martensite with hydrogen has a major impact on the HE of metastable austenitic stainless steels. The implications of the
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 7. (a) True stress–strain curves for the austenite specimens and (b) their respective processing histories. The arrows in (a) indicate the onsets of necking. Standard stereographic triangles show the loading directions. (For interpretation of the references to colour in this figure, the reader is referred to the web version of this article.)
Table 2 The fraction of α′–martensite yield strength, Schmid factor, corresponding critical resolved shear stress (CRSS) for the primary slip system, tensile strength, and reduction of area for the austenite specimens. α′–martensite fraction (%)
Yield strength (MPa)
Schmid factor
CRSS (MPa)
Tensile strength (MPa)
Reduction of area (%)
U-HPT-A 4 H-HPT-A 1
938 663
0.456 0.387
427 256
1130 702
– –
169 230
0.468 0.478
79 110
487 477
90 (Chisel edge) 98 42
ST-U-A ST-H-A
α′–martensite transformation of hydrogen-containing austenite in the HE mechanism are discussed in the following section.
91
Fig. 8. (a) EBSD map overlapped on the corresponding SEM image taken in the initial state, (b) (111) and (110) pole figures of austenite and the corresponding colour-coded map, and (c) optical micrograph taken at the onset of yielding of the U-HPT-A specimen.
3.2. The development of nano-twins and slip bands and their effects on the plasticity of austenite Our previous study revealed [10] that hydrogen charging reduced the formation of α′–martensite during the subsequent process of HPT. A certain amount of austenitic phase remained even when subjected to severe plastic deformation. Fig. 5 shows typical microstructures developed by HPT processing with and without hydrogen-pre-charging. While the isolated austenite grains corresponded to a fraction of the U-HPT disc, deformation twinning occurred, which was often accompanied by lamellar plates of α′–martensite (Fig. 5a and b). In the H-HPT disc, welldeveloped slip bands were observed in most austenite grains, as shown in Fig. 5c and d. Fig. 6 shows the TEM micrographs taken on the deformed austenite grain in the H-HPT disc. Unlike the U-HPT disc, nano-twins were not observed in the H-HPT disc.
92
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 9. (a) EBSD map overlapped on the corresponding SEM image taken in the initial state, (b) (111) and (110) pole figures of austenite, and (c) optical micrograph taken at the onset of yielding of the H-HPT-A specimen.
Alternatively, there were some places where spots of α′ and ε–martensite were visible. Twinning has been reported to occur in austenitic steels with an intermediate stacking fault energy (SFE) [19,20]. Shen et al. also verified that ε–martensite and deformation twins only act as an intermediate phase in the transformation from austenite to α′–martensite in a type 304 stainless steel [20]. X-ray diffraction measurements by Pontini and Hermida revealed that hydrogen charging into a type 304 stainless steel decreased the SFE [21]. Moreover, a calculation study by Hermida and Roviglione explained that this SFE decrease allows the nucleation of the ε–martensite through H–H pair formation [22]. These facts suggest that the transformation from austenite to α′–martensite occurred via the γ–twinning preferentially in the uncharged state of the present steel, whereas solute hydrogen facilitated the formation of the ε–martensite as an intermediate phase by decreasing the SFE. Fig. 7 shows the true stress–strain curves for the austenitic phase specimens and a schematic indicating their corresponding processing histories. The U-HPT-A and H-HPT-A specimens were prepared from the microstructures shown in Fig. 5, the nanotwinned and slip-deformed microstructures, respectively. The standard stereographic triangles in Fig. 7a indicate the initial loading directions. In the case of the U-HPT-A specimen, blue and red show the parent and twinned crystals, respectively. For comparison, the results of uncharged and hydrogen-charged solution-
treated specimens [23], denoted as ST-U-A and ST-H-A, respectively, are included in Fig. 7. The ST-U-A specimen did not undergo HPT-processing or hydrogen charging (Fig. 7b). The ST-H-A specimen was tested within 3 h after hydrogen cathodic charging, which provided a saturated hydrogen content of 100 mass ppm [23]. While there might have remained very little hydrogen during micro-tensile testing in the H-HPT-A specimen, the ST-H-A specimen should have contained a considerable amount of hydrogen. Table 2 summarises the α′–martensite fraction, the yield strength, the Schmid factor, and corresponding critical resolved shear stress (CRSS) for the primary slip system, the tensile strength, and the reduction of area for the austenite specimens. The yield strengths of the U-HPT-A and H-HPT-A specimens were 940 MPa and 660 MPa, which were significantly higher than the value of 170 MPa for the ST-U-A specimen. Figs. 8 and 9 show the EBSD maps and the (111) and (110) pole figures in the initial state and the optical micrographs taken at the yielding points for the U-HPT-A and H-HPT-A specimens, respectively. In the U-HPT-A specimen, the plane of the nano-twins was arranged nearly parallel to the loading direction (Fig. 8b). The slip trace that appeared at the onset of yielding corresponded to the primary slip planes in the parent and twinned crystals (compare Fig. 8b and c). Assuming that macroscopic yielding occurred in the parent crystal, the apparent CRSS for the U-HPT-A specimen was determined to be 430 MPa, which is more than five times higher than the CRSS of 79 MPa for the ST-U-A specimen (Table 2). A micro-tensile testing study by Mine et al. using a nano-twinned stable austenitic alloy revealed [24] that introducing nano-twins with an average thickness of 20 nm mainly contributed to the strengthening of this alloy through the inhibition of slip transfer by the twin boundaries. As for the H-HPT-A specimen pre-hardened by slip deformation and/or the precipitation of ε–martensite in the 304 steel, the CRSS was determined to be 280 MPa. In addition, although the mutual interaction of dislocations introduced by the process of HPT could not be neglected, the inhibition of dislocation movement by twin boundaries was a major strengthening mechanism in the nano-twinned 304 steel. Moreover, despite the significant strengthening, a moderate ductility was retained in the U-HPT-A specimen (Fig. 7a and Table 2). This suggests that introducing nano-twins into austenitic steels can provide a good balance between the strength and ductility. The U-HPT-A specimen exhibited a two-step hardening behaviour (Fig. 7a). This might be attributed to the interaction of α′– martensite with the twin boundaries. Fig. 10 shows the EBSD maps and the (110) and (111) pole figures of the α′–martensite taken from the longitudinal cross-section of the fractured U-HPT-A specimen. Three martensite variants with habit planes parallel to the plane of the nano-twins were mainly formed (Fig. 10c and d). Fig. 11 shows the fracture morphology and the (111) and (110) pole figures of the austenite and martensite taken from the longitudinal cross-section of the fractured ST-U-A specimen. Unlike the nanotwinned specimen, two martensite variants with habit planes parallel to the primary slip plane were prevalent in the ST-U-A specimen (Fig. 11c). This finding coincided with the results of previous works by Higo et al. [25] and Kato and Mori [26]. According to their reports, assuming a double shearing mechanism, the critical first shear initiating the α′–martensitic transformation determined the habit plane of the formed martensite variants. In fact, the α′–martensite variants with habit planes parallel to the highest shearing plane were selected for the ST-U-A specimen. However, in the U-HPT-A specimen, i.e., in the presence of nanotwins, martensite variants nucleating from the twin plane were prevalent rather than ones with habit planes parallel to the highest shearing plane (Fig. 10). Nakada et al. reported [27] that three variants of the deformation-induced martensite were formed on the plane of deformation twinning with a double K-S
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
93
Fig. 10. (a) EBSD map, (b) phase map with image quality, (c) colour-coded map corresponding to (d), and (d) (110) and (111) pole figures of α′–martensite of the fractured U-HPT-A specimen.
relationship with both the parent and the twinned crystals of austenite in a cold-rolled 316 steel. Fig. 10 shows that the three martensite variants formed through tensile testing of the nanotwinned 304 steel had a double K-S relationship similar to that of the cold-rolled 316 steel. Moreover, a magnetic force microscopy study by Zhang et al. revealed [28] that deformation-induced martensite was nucleated from annealed twin boundaries in a 316 steel, which was deformed at a temperature of 200 K. Therefore, it is concluded that, regardless of deformation or annealing, introducing twin boundaries changes the route of the deformationinduced martensitic transformation in metastable austenitic steels. In the ST-H-A specimen, it is assumed that hydrogen facilitates the dislocation movement on the primary plane in the plateau stress regime following the yielding, which leads to the stress drop and premature martensite formation within the austenitic grains [23]. The stress–strain curve for the H-HPT-A specimen exhibited a plateau regime following yielding similar to that of the ST-H-A specimen (Fig. 7a). This is because the microstructure developed in
the H-HPT-A specimen contained the ε–martensite instead of the nano-twins, as shown in Fig. 6. If the deformation process proceeds in the presence of hydrogen, hydrogen can be enriched in the retained lamellar austenite owing to the martensitic transformation, resulting in quasi-cleavage fracture. However, in the present case, hydrogen had diffused out of the H-HPT-A specimen before the micro-tensile test; therefore, work hardening without hydrogen did not lead to severe hydrogen embrittlement with quasi-cleavage, as shown in the ST-H-A specimen [23]. This implies that the defects produced by deformation with hydrogen may not play a crucial role but rather build a foundation in the HE mechanism. As mentioned above, Zhang et al. reported [18] that dynamic α′–martensite formed during deformation mainly contributes to HE. It was also argued that austenite transforms into α ′–martensite during deformation with hydrogen, thereby leading to easier diffusion of hydrogen [29]. These findings supported the hypothesis that excess hydrogen due to martensitic transformation, which corresponds to the difference in the solubility between
94
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
Fig. 11. (a) Fracture morphology of the ST-U-A specimen and (b, c) (111) and (110) pole figures of austenite and α′–martensite, respectively.
Y. Mine et al. / Materials Science & Engineering A 661 (2016) 87–95
austenite and martensite, promotes the deformation localisation in the retained austenitic phase, resulting in HE in metastable austenitic steels [30].
95
acknowledges support from the ‘Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers' R2608. This study used facilities for severe plastic deformation at the International Research Centre on Giant Straining for Advanced Materials (IRC-GSAM) at Kyushu University.
4. Summary Micro-tensile testing was used in this study to mechanically characterise microstructures deformed through high-pressure torsion of hydrogen-pre-charged and uncharged samples of a 304 austenitic stainless steel. The main findings can be summarised as follows: (1) The martensite specimen formed by deformation with hydrogen exhibited low strength but a high reduction of area when compared to the one formed in the absence of hydrogen. This suggests that austenite regions are more likely to remain in martensite formed by deformation with hydrogen. (2) Deformation-twinning in the uncharged 304 austenitic steel acts as an intermediate phase in the transformation to α′– martensite, which contributes to a good balance between the strength and ductility. (3) Hydrogen charging in the 304 austenitic stainless steel facilitates the formation of ε–martensite as an intermediate phase in the transformation from austenite to α′–martensite, thereby leading to the stress drop following yielding in the stress– strain behaviour characteristic of the hydrogen-charged austenitic specimens. However, dynamic α′–martensitic transformation in the presence of hydrogen is requisite for severe HE with quasi-cleavage.
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18]
The defects and the inhomogeneity of the microstructure provided by deformation with hydrogen trigger hydrogen embrittlement in metastable austenitic steel but cannot play the crucial role therein.
[19] [20]
Acknowledgements
[25] [26] [27]
The authors thank Mr. R. Matsuoka, Dr. M. Tsushida, and Dr. T. Yamamuro, Kumamoto University for their assistance with the TEM observation. The present work was supported in part by a Grant-in-Aid for Scientific Research (C) 25420758 from the Japan Society for the Promotion of Science (JSPS). YM gratefully
[21] [22] [23] [24]
[28] [29] [30]
C.D. Beachem, Metall. Trans. 3 (1972) 437–451. S.P. Lynch, Acta Metall. 36 (1988) 2639–2661. H.K. Birnbaum, P. Sofronis, Mater. Sci. Eng. A 176 (1994) 191–202. I.M. Robertson, Eng. Fract. Mech. 68 (2001) 671–692. K. Takai, H. Shoda, H. Suzuki, M. Nagumo, Acta Mater. 56 (2008) 5158–5167. D. Eliezer, D.G. Chakrapani, C.J. Altstetter, E.N. Pugh, Metall. Trans. A 10 (1979) 935–941. G. Han, J. He, S. Fukuyama, K. Yokogawa, Acta Mater. 46 (1998) 4559–4570. C. San Marchi, B.P. Somerday, X. Tang, G.H. Schiroky, Int. J. Hydrog. Energy 33 (2008) 889–904. Y. Mine, T. Kimoto, Corros. Sci. 53 (2011) 2619–2629. Y. Mine, Z. Horita, Y. Murakami, Acta Mater. 57 (2009) 2993–3002. Y. Mine, K. Hirashita, M. Matsuda, M. Otsu, K. Takashima, Corros. Sci. 53 (2011) 529–533. Y. Mine, H. Fujisaki, M. Matsuda, M. Takeyama, K. Takashima, Scr. Mater. 65 (2011) 707–710. Y. Mine, K. Hirashita, M. Matsuda, K. Takashima, Metall. Mater. Trans. A 42 (2011) 3567–3571. Y. Mine, K. Hirashita, H. Takashima, M. Matsuda, K. Takashima, Mater. Sci. Eng. A 560 (2013) 535–544. Y. Mine, H. Takashima, M. Matsuda, K. Takashima, Mater. Sci. Eng. A 618 (2014) 359–367. Y. Mine, R. Maezono, H. Oda, M. Yamasaki, Y. Kawamura, K. Takashima, Mater. Trans. 56 (2015) 952–956. F. Maresca, V.G. Kouznetsova, M.G.D. Geers, J. Mech. Phys. Solids 73 (2014) 69–83. L. Zhang, Z. Li, J. Zheng, Y. Zhao, P. Xu, C. Zhou, X. Li, Int. J. Hydrog. Energy 38 (2013) 8208–8214. J.-Y. Choi, W. Jin, Scr. Mater. 36 (1997) 99–104. Y.F. Shen, X.X. Li, X. Sun, Y.D. Wang, L. Zuo, Mater. Sci. Eng. A 552 (2012) 514–522. A.E. Pontini, J.D. Hermida, Scr. Mater. 37 (1997) 1831–1837. J.D. Hermida, A. Roviglione, Scr. Mater. 39 (1998) 1145–1149. Y. Mine, K. Koga, O. Kraft, K. Takashima, Scr. Mater. 113 (2016) 176–179. Y. Mine, S. Nakamichi, K. Koga, K. Takashima, O. Kraft, To be submitted Elsewhere, 2015. Y. Higo, F. Lecroisey, T. Mori, Acta Metall. 22 (1974) 313–323. M. Kato, T. Mori, Acta Metall. 24 (1976) 853–860. N. Nakada, H. Ito, Y. Matsuoka, T. Tsuchiyama, S. Takaki, Acta Mater. 58 (2010) 895–903. L. Zhang, M. Imade, B. An, M. Wen, T. Iijima, S. Fukuyama, K. Yokogawa, ISIJ Int. 52 (2012) 240–246. J.H. Ryu, Y.S. Chun, C.S. Lee, H.K.D.H. Bhadeshia, D.W. Suh, Acta Mater. 60 (2012) 4085–4092. Y. Mine, Z. Horita, Mater. Trans. 53 (2012) 773–785.