- Email: [email protected]
Mechanical characterisation of hydrogen-induced quasi-cleavage in a metastable austenitic steel using micro-tensile testing
Scripta Materialia 113 (2016) 176–179
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat
Regular Article
Mechanical characterisation of hydrogen-induced quasi-cleavage in a metastable austenitic steel using micro-tensile testing Yoji Mine a,⁎, Kaoru Koga a,1, Oliver Kraft b, Kazuki Takashima a a b
Department of Materials Science and Engineering, Kumamoto University, 2-39-1 Kurokami, Chuo-ku, Kumamoto 860-8555, Japan Institute for Applied Materials, Karlsruhe Institute of Technology, Hermann-von-Helmhotz-Platz 1, Eggenstein-Leopoldshafen 76344, Germany
a r t i c l e
i n f o
Article history: Received 19 September 2015 Received in revised form 3 November 2015 Accepted 10 November 2015 Available online 19 November 2015 Keywords: Micromechanical characterisation Austenitic steels Hydrogen embrittlement Martensitic phase transformation Metastable phase
a b s t r a c t Micro-tensile tests were performed using single- and twinned bi-crystals with and without hydrogen precharging in a metastable austenitic steel to elucidate the occurrence of hydrogen-induced quasi-cleavage. The hydrogen-charged specimens exhibited in the stress–strain behaviour a stress decrease in the plateau region after yielding. In this stage, a lamellar microstructure of martensite was formed within the parent austenite. Fracture occurred as quasi-cleavage along the habit planes of the martensite lamellae. A mechanism is proposed where the excess hydrogen generated by the martensitic transformation of hydrogen-containing austenite concentrates into the retained austenite between the martensite laths, promoting the localised deformation to fracture. © 2015 Elsevier Ltd. All rights reserved.
Austenitic stainless steels are widely used in hydrogen-related industrial applications because of their low susceptibility to hydrogen embrittlement (HE), which depends largely on the stability of the austenitic phase [1–5]. When exposed to gaseous hydrogen at a pressure of 100 MPa, 316L and 310S stable austenitic steels absorb hydrogen at a content of up to approximately 100 mass ppm. Solute hydrogen promotes shear deformation, which results in an extended slant surface in the periphery of the cup-and-cone fracture surface [5]. In this case, 316L and 310S steels, even having such a high content of solute hydrogen, exhibited only moderate decreases in ductility as characterised by a less pronounced reduction of cross-sectional area at fracture. As for the 316 steel with a medium stability of austenite, the specimen containing ~100 mass ppm of hydrogen exhibited a fully shear fracture [5]. Thus, it is reasonable to attribute the hydrogen-induced degradation in the ductility of the stable austenitic steels to a mechanism that involves hydrogen-enhanced localised plasticity (HELP) [6–8]. Quasi-cleavage and flat-faceted features are particularly characteristic of severe HE in metastable austenitic steels with low austenite stabilities, such as 301 and 304 steels [9–13]. The formation of quasi-cleavage may be related with deformation-induced martensitic transformations. Previous studies revealed that cracking in the martensitic phase induced by deformation is responsible for the HE in these steels [13–16]. However, it has been reported that martensite that existed prior to loading was resistant to HE [17,18]. Therefore, the role of martensite in the ⁎ Corresponding author. E-mail address: [email protected] (Y. Mine). 1 Currently: RYOBI LIMITED.
http://dx.doi.org/10.1016/j.scriptamat.2015.11.013 1359-6462/© 2015 Elsevier Ltd. All rights reserved.
mechanism of quasi-cleavage in metastable austenitic steels is still controversial. Therefore, it is the main goal of this study to analyse the formation of quasi-cleavage in a metastable austenitic stainless steel from a crystallographic perspective. We employ micro-tensile testing that allows for testing small single crystals and twinned bi-crystals with defined crystallographic orientations. The material used in this study is a commercial type of 304 stainless steel that is composed of 0.05 C, 18.54 Cr, 8.09 Ni, 0.58 Si, 1.24 Mn, 0.025 P, and 0.003 S (in mass%) with the remainder being Fe. Coarse-grained samples with grain sizes of several tens of micrometres were obtained through solution treatment at a temperature of 1403 K for 2 h followed by water quenching. The coarse-grained samples were thinned to a thickness of less than 30 μm through grinding with emery paper and diamond paste. For electron backscatter diffraction (EBSD) analysis, the surfaces were electro-chemically polished. A field-emission scanning electron microscope (SEM) with an EBSD analyser was used. The EBSD analysis was carried out using the TexSEM Laboratories orientation imaging microscopy software (OIM v. 7.1.0). Single- and twinned bi-crystalline specimens, denoted as SC and TW specimens, respectively, with gauge sections of about 20 μm × 20 μm × 50 μm were fabricated using a focused ion beam (FIB). The loading directions (LDs) for the SC specimen were chosen for single slip as shown in the stereographic triangle in Fig. 1. The TW specimen had a similar orientation and twin boundary parallel to the LD (Fig. 4). A set of specimens in a foil form with gauge part supporting members was soaked in a H2SO4 aqueous solution (pH = 3.5) held at a temperature of 353 K and electrochemically charged with hydrogen for 5 h at a current density of 27 A m−2. Based on an estimate of the diffusion velocity [5], the
Y. Mine et al. / Scripta Materialia 113 (2016) 176–179
177
using FIB to make a longitudinal cut. The crystal orientations were determined using automatic beam scanning with a step size of 0.08 μm at an accelerating voltage of 20 kV. Fig. 1 shows the shear stress–shear strain curves and deformation processes of the hydrogen-charged and uncharged single-crystalline specimens, SC-H and SC-U, respectively. The shear stress and shear strain, τ and γ, were calculated from the nominal stress and strain, σ and ε, using: τ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi σ cosϕ ε2 þ 2ε þ cos2 λ ; 1þε
ð1Þ
and γ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ε2 þ 2ε þ cos2 λ− cosλ ; cosϕ
ð2Þ
where ϕ and λ are the angles of the LD with the normal of the slip plane and the slip direction, respectively. The SC-U specimen exhibits a three-step work hardening behaviour after yielding at a shear stress of ~79 MPa (Fig. 1a) through the activation of the primary slip system
Fig. 1. (a) Shear stress–shear strain curves and (b and c) optical micrographs showing the deformation processes of the uncharged and hydrogen-charged single-crystalline specimens, SC-U and -H, respectively.
charging time was chosen such that the 20 μm-thick specimens are saturated with solute hydrogen. The saturated hydrogen content under the corresponding hydrogen charging conditions was measured to be 101 mass ppm by thermal desorption spectrometry. Tensile tests were started within 3 h after hydrogen charging through cutting the gauge part supporting members using a laser beam. Tensile tests were performed at room temperature under laboratory atmospheric conditions using micro-gluing for gripping and a piezoelectric actuator at a displacement rate of 0.1 μm s−1, corresponding to a strain rate of 2 × 10−3 s−1. The setup has been described in more detail in [19]. The gauge section of the tensile specimen was monitored during tensile testing using an optical microscope in order to dynamically measure the strain as a function of time. For EBSD analysis after failure or after the interruption of loading, the specimens were further fabricated
{111} b110 N with a Schmid factor of 0.47. This is confirmed by the slip steps that are visible in the optical micrograph in Fig. 1b. The critical resolved shear stress (CRSS) was in agreement with the value obtained by loading in the [111] direction [20]. In the stress–strain curve, a stage I deformation up to a shear strain of about 0.52 can be identified. It is argued that easy glide takes place with the primary slip system being activated (Fig. 1c), although some hardening occurs. When the secondary slip system is activated owing to crystal rotation, work hardening becomes more pronounced (stage IIa) induced through mutual interactions between the dislocations on the two slip systems. The hardening rate is somewhat reduced in stage IIb and, subsequently, the workhardening rate increases again in stage III. The appearance of the reduced hardening region of stage IIb can be presumably attributed to the formation of martensite, as previously reported by Tsurui et al. [21]. In fact, in-situ observations made during micro-tensile testing of the SC-U specimen revealed that gradual undulation, which differs from slip steps, appeared in region in stage IIb (compare Fig. 1c and d). The increase in the work-hardening rate in stage III can be caused by work hardening of the formed martensite. While the shape of the stress–strain curve in the SC-H specimen is similar to the one for the SC-U specimen, the stage I is different with a plateau region after the onset of yielding at about 110 MPa, which is 39% higher than that in the SC-U specimen. Overall, the solute hydrogen increases the flow stresses but shortens the length of each stage, resulting in a smaller strain-to-failure. While the trace of the primary slip plane appeared at the onset of yielding (Fig. 1e), the stage I deformation in the SC-H specimen was localised within a short length of the specimen gauge part, as shown in Fig. 1f. In the stage II, deformation was widely spread throughout the entire gauge part through the activation of the multiple slip systems (Fig. 1g). Fig. 2 compares the fracture morphology of the single-crystalline specimens with and without hydrogen pre-charging. The uncharged specimen exhibits a ductile fracture with strong necking (Fig. 2a and b), whereas the fracture surface of the hydrogen-charged specimen is characterised by a typical hydrogen-induced quasi-cleavage (Fig. 2c and d). The insets in Fig. 2b and d show the EBSD maps of SC-U and SC-H specimens after fracture, respectively. The analysis reveals that martensite variants developed in both specimens, mainly with their habit planes parallel to the primary slip planes. In the SC-H specimen, variants with their habit planes parallel to the secondary slip plane were also observed (Fig. 2d). This can be induced by plastic constraint at the shoulder part of the specimen. Also, it is confirmed that the quasicleavage facets were formed parallel to the two martensite habit planes (Fig. 2c and d).
178
Y. Mine et al. / Scripta Materialia 113 (2016) 176–179
Fig. 2. SEM micrographs comparing the fracture morphology of the uncharged (a and b) and hydrogen-charged (c and d) single-crystalline specimens. The insets in (b and d) show the EBSD maps taken in the areas marked by the boxes, after the broken specimens were further prepared by FIB, making a longitudinal cut.
In the stress–strain curve of the hydrogen-charged specimen (Fig. 1a), a slight decrease in the stress was observed just after the onset of yielding. It is argued that this relates to the martensitic transformation which is accompanied by a stress decrease. In order to analyse this in more detail, EBSD was applied after interrupting the tensile test at the end of stage I for a hydrogen-charged single-crystalline specimen. The orientation of the LD is close to the [110]-direction of the crystal. Fig. 3 shows the shear stress–shear strain curves (a), and the deformation morphology (b). The flow stress in the plateau regime is about 130 MPa, which is higher than that for the SC-H specimen. This may be related to a higher amount of hydrogen retained as the increase in the flow stress due to hydrogenation depends on the hydrogen content [7]. Observation of the top (Fig. 3b) and side (not shown) surfaces revealed that primary and secondary slips were activated. Fig. 3c and d shows the result of an EBSD analysis on the cross-section of the specimen including the (111)- and (110)-pole figures of austenite (c) and
martensite (d), and the corresponding colour-coded maps. Three martensite variants with their habit planes parallel to the primary slip plane developed, while some fraction of the parent austenite crystal was retained. In other words, the martensite/austenite lamellar microstructure formed parallel to the primary slip plane. The hydrogen solubility in austenite is approximately an order of magnitude higher than that in martensite [22]. If hydrogen-containing austenite is transformed to martensite, excess hydrogen corresponding to the difference in solubility between the two phases is generated. Therefore, the excess hydrogen diffuses out of the martensite and into the austenite. It can be further argued that the hydrogen concentrates close to the interface in the austenite because the hydrogen diffusivity is significantly lower in austenite than in martensite [23]. As the martensite formation proceeds, the volume of retained austenite between the martensite laths decreases, leading to local supersaturation of hydrogen. Thus, it is concluded that quasi-cleavage in a metastable
Fig. 3. (a) Shear stress–shear strain curves for hydrogen-charged single-crystalline specimens. The test was interrupted prior to fracture. (b) SEM micrograph of the deformation morphology of the deformed tensile specimen. (c and d) EBSD analysis showing colour-coded maps of the area marked by the box in (b) for austenite and martensite, respectively, and the corresponding (110) and (111) pole figures. Circles and triangles show the habit planes and directions, respectively.
Y. Mine et al. / Scripta Materialia 113 (2016) 176–179
179
Fig. 4. (a) Shear stress–shear strain curves and (b and c) fracture surfaces for the uncharged and hydrogen-charged twinned bi-crystalline specimens, TW-U and TW-H, respectively.
austenitic steel is caused by excess-hydrogen-assisted slip localisation in the retained austenite between the deformation-induced martensite laths along their habit planes. This conclusion is in agreement with the fact that in our previous work [24] retained austenite was found by transmission electron microscopy in the vicinity of the fracture surface of a hydrogen-charged micro-tensile specimen. Also, it has been found that in lath martensite single packet structure slip occurs preferentially parallel to the habit plane [25], and that this anisotropic slip behaviour might stem from the presence of the retained austenite between the martensite laths as indicated by finite element simulations [26]. Also, it is important to note that the proposed mechanism is consistent with the observation that pre-existing martensite does not promote HE because it does not introduce excess hydrogen in the austenite. Fig. 4a shows the shear stress–shear strain curves of an uncharged and a hydrogen-charged twinned bi-crystalline specimen, TW-U and TW-H, respectively. Unlike the single-crystalline specimens, the workhardening rate did not decrease but rather increased in the transition from stage II to III. This may be because the presence of the twin boundary promotes the activation of multiple-slip glides in an early stage and therefore results in the formation of more martensite nuclei. For the TW-H specimen, the yield strength was higher than that of the TW-U specimen. Like the single-crystalline specimens, hydrogen shortened the length of each stage and decreased the elongation-to-failure. Fig. 4b and c shows the fracture surfaces of the TW-U and TW-H specimen, respectively. While the TW-U specimen exhibited a necking almost symmetric with respect to the twin boundary (Fig. 4b), quasicleavages appeared with different fracture planes in each crystal of the TW-H specimen (Fig. 4c). The formed martensite variants had habit planes parallel to the primary slip planes (not shown in the figure). This indicates that the quasi-cleavage occurs in each crystal along the primary slip system based on the same mechanism as in the singlecrystalline specimen. Micro-tensile tests were performed on single crystals and twinned bi-crystals with and without hydrogen pre-charging in a 304 metastable austenitic steel to elucidate the occurrence of quasi-cleavage induced by hydrogen. In all specimens, the stress–strain curves were composed of three work-hardening stages. The hydrogen pre-charge shortened the length of each stage, resulting in a smaller strain-to-failure compared to the uncharged specimen. A quasi-cleavage fracture, whose facets correspond to the habit planes of martensite formed during deformation, was observed in the hydrogen-charged specimens. The hydrogencharged specimens also exhibited a small stress drop after the onset of yielding. In this stage, a lamellar microstructure of martensite within the parent austenite is formed. Based on these observations, the following mechanism is suggested for the hydrogen-induced quasi-cleavage.
Excess hydrogen is generated by the deformation-induced martensitic transformation since the solubility of hydrogen is smaller in martensite. As a result, hydrogen concentrates in the retained austenite between the martensite laths. The high hydrogen concentration leads to localised shear in the retained austenite along the habit plane of the martensitic transformation, leading to quasi-cleavage fracture in metastable austenitic steels. 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), and in part by ‘Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers’ (R2608).
References [1] D. Eliezer, D.G. Chakrapani, C.J. Altstetter, E.N. Pugh, Metall. Trans. A 10 (1979) 935. [2] L. Zhang, M. Wen, M. Imade, S. Fukuyama, K. Yokogawa, Acta Mater. 56 (2008) 3414. [3] T. Kanezaki, C. Narazaki, Y. Mine, S. Matsuoka, Y. Murakami, Int. J. Hydrog. Energy 33 (2008) 2604. [4] C. San Marchi, T. Michler, K.A. Nibur, B.P. Somerday, Int. J. Hydrogen Energy 35 (2010) 9736. [5] Y. Mine, T. Kimoto, Corros. Sci. 53 (2011) 2619. [6] H.K. Birnbaum, P. Sofronis, Mater. Sci. Eng. A 176 (1994) 191. [7] D.P. Abraham, C.J. Altstetter, Metall. Mater. Trans. A 26 (1995) 2849. [8] D.P. Abraham, C.J. Altstetter, Metall. Mater. Trans. A 26 (1995) 2859. [9] C.L. Briant, Metall. Trans. A 10 (1979) 181. [10] H. Hänninen, T. Hakkarainen, Metall. Trans. A 10 (1979) 1196. [11] R. Liu, N. Narita, C. Altstetter, H. Birnbaum, E.N. Pugh, Metall. Trans. A 11 (1980) 1563. [12] D. Sun, G. Han, S. Vaodee, S. Fukuyama, K. Yokogawa, Mater. Sci. Technol. 17 (2001) 302. [13] G. Han, J. He, S. Fukuyama, K. Yokogawa, Acta Metall. 46 (1998) 4559. [14] R.M. Vennett, G.S. Ansell, Trans. Am. Soc. Met. 60 (1967) 242. [15] R.B. Benson Jr., R.K. Dann, L.W. Roberts Jr., Trans. Metall. Soc. AIME 242 (1968) 2199. [16] A. Inoue, Y. Hosoya, T. Matsumoto, Trans. Iron Steel Inst. Jpn. 19 (1979) 170. [17] J.R. Buckley, D. Hardie, Corros. Sci. 34 (1993) 93. [18] L. Zhang, Z. Li, J. Zheng, Y. Zhao, P. Xu, C. Zhou, X. Li, Int. J. Hydrog. Energy 38 (2013) 8208. [19] T. Ito, Y. Mine, M. Otsu, K. Takashima, J. Jpn. Inst. Met. Mater. (2016) in press. [20] Y. Mine, K. Koga, K. Takashima, O. Kraft, Unpublished data (2015), Kumamoto University. [21] T. Tsurui, S. Inoue, K. Matsuda, H. Ishigaki, K. Murata, K. Koterazawa, J. Soc. Mater. Sci., Jpn. 50 (2001) 1115. [22] Y. Mine, C. Narazaki, T. Kanezaki, S. Matsuoka, Y. Murakami, in: E.E. Gdoutos (Ed.), Fracture of Nano and Engineering Materials and Structures, Springer, Netherlands 2006, pp. 907–908. [23] K. Sakamoto, H. Katayama, J. Jpn. Inst. Metals 46 (1982) 805. [24] Y. Mine, K. Hirashita, M. Matsuda, K. Takashima, Metall. Mater. Trans. A 42 (2011) 3567. [25] Y. Mine, K. Hirashita, H. Takashima, M. Matsuda, K. Takashima, Mater. Sci. Eng. A 560 (2013) 535. [26] F. Maresca, V.G. Kouznetsova, M.G.D. Geers, J. Mech. Phys. Solids 73 (2014) 69.
Contents lists available at ScienceDirect
Scripta Materialia journal homepage: www.elsevier.com/locate/scriptamat
Regular Article
Mechanical characterisation of hydrogen-induced quasi-cleavage in a metastable austenitic steel using micro-tensile testing Yoji Mine a,⁎, Kaoru Koga a,1, Oliver Kraft b, Kazuki Takashima a a b
Department of Materials Science and Engineering, Kumamoto University, 2-39-1 Kurokami, Chuo-ku, Kumamoto 860-8555, Japan Institute for Applied Materials, Karlsruhe Institute of Technology, Hermann-von-Helmhotz-Platz 1, Eggenstein-Leopoldshafen 76344, Germany
a r t i c l e
i n f o
Article history: Received 19 September 2015 Received in revised form 3 November 2015 Accepted 10 November 2015 Available online 19 November 2015 Keywords: Micromechanical characterisation Austenitic steels Hydrogen embrittlement Martensitic phase transformation Metastable phase
a b s t r a c t Micro-tensile tests were performed using single- and twinned bi-crystals with and without hydrogen precharging in a metastable austenitic steel to elucidate the occurrence of hydrogen-induced quasi-cleavage. The hydrogen-charged specimens exhibited in the stress–strain behaviour a stress decrease in the plateau region after yielding. In this stage, a lamellar microstructure of martensite was formed within the parent austenite. Fracture occurred as quasi-cleavage along the habit planes of the martensite lamellae. A mechanism is proposed where the excess hydrogen generated by the martensitic transformation of hydrogen-containing austenite concentrates into the retained austenite between the martensite laths, promoting the localised deformation to fracture. © 2015 Elsevier Ltd. All rights reserved.
Austenitic stainless steels are widely used in hydrogen-related industrial applications because of their low susceptibility to hydrogen embrittlement (HE), which depends largely on the stability of the austenitic phase [1–5]. When exposed to gaseous hydrogen at a pressure of 100 MPa, 316L and 310S stable austenitic steels absorb hydrogen at a content of up to approximately 100 mass ppm. Solute hydrogen promotes shear deformation, which results in an extended slant surface in the periphery of the cup-and-cone fracture surface [5]. In this case, 316L and 310S steels, even having such a high content of solute hydrogen, exhibited only moderate decreases in ductility as characterised by a less pronounced reduction of cross-sectional area at fracture. As for the 316 steel with a medium stability of austenite, the specimen containing ~100 mass ppm of hydrogen exhibited a fully shear fracture [5]. Thus, it is reasonable to attribute the hydrogen-induced degradation in the ductility of the stable austenitic steels to a mechanism that involves hydrogen-enhanced localised plasticity (HELP) [6–8]. Quasi-cleavage and flat-faceted features are particularly characteristic of severe HE in metastable austenitic steels with low austenite stabilities, such as 301 and 304 steels [9–13]. The formation of quasi-cleavage may be related with deformation-induced martensitic transformations. Previous studies revealed that cracking in the martensitic phase induced by deformation is responsible for the HE in these steels [13–16]. However, it has been reported that martensite that existed prior to loading was resistant to HE [17,18]. Therefore, the role of martensite in the ⁎ Corresponding author. E-mail address: [email protected] (Y. Mine). 1 Currently: RYOBI LIMITED.
http://dx.doi.org/10.1016/j.scriptamat.2015.11.013 1359-6462/© 2015 Elsevier Ltd. All rights reserved.
mechanism of quasi-cleavage in metastable austenitic steels is still controversial. Therefore, it is the main goal of this study to analyse the formation of quasi-cleavage in a metastable austenitic stainless steel from a crystallographic perspective. We employ micro-tensile testing that allows for testing small single crystals and twinned bi-crystals with defined crystallographic orientations. The material used in this study is a commercial type of 304 stainless steel that is composed of 0.05 C, 18.54 Cr, 8.09 Ni, 0.58 Si, 1.24 Mn, 0.025 P, and 0.003 S (in mass%) with the remainder being Fe. Coarse-grained samples with grain sizes of several tens of micrometres were obtained through solution treatment at a temperature of 1403 K for 2 h followed by water quenching. The coarse-grained samples were thinned to a thickness of less than 30 μm through grinding with emery paper and diamond paste. For electron backscatter diffraction (EBSD) analysis, the surfaces were electro-chemically polished. A field-emission scanning electron microscope (SEM) with an EBSD analyser was used. The EBSD analysis was carried out using the TexSEM Laboratories orientation imaging microscopy software (OIM v. 7.1.0). Single- and twinned bi-crystalline specimens, denoted as SC and TW specimens, respectively, with gauge sections of about 20 μm × 20 μm × 50 μm were fabricated using a focused ion beam (FIB). The loading directions (LDs) for the SC specimen were chosen for single slip as shown in the stereographic triangle in Fig. 1. The TW specimen had a similar orientation and twin boundary parallel to the LD (Fig. 4). A set of specimens in a foil form with gauge part supporting members was soaked in a H2SO4 aqueous solution (pH = 3.5) held at a temperature of 353 K and electrochemically charged with hydrogen for 5 h at a current density of 27 A m−2. Based on an estimate of the diffusion velocity [5], the
Y. Mine et al. / Scripta Materialia 113 (2016) 176–179
177
using FIB to make a longitudinal cut. The crystal orientations were determined using automatic beam scanning with a step size of 0.08 μm at an accelerating voltage of 20 kV. Fig. 1 shows the shear stress–shear strain curves and deformation processes of the hydrogen-charged and uncharged single-crystalline specimens, SC-H and SC-U, respectively. The shear stress and shear strain, τ and γ, were calculated from the nominal stress and strain, σ and ε, using: τ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi σ cosϕ ε2 þ 2ε þ cos2 λ ; 1þε
ð1Þ
and γ¼
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ε2 þ 2ε þ cos2 λ− cosλ ; cosϕ
ð2Þ
where ϕ and λ are the angles of the LD with the normal of the slip plane and the slip direction, respectively. The SC-U specimen exhibits a three-step work hardening behaviour after yielding at a shear stress of ~79 MPa (Fig. 1a) through the activation of the primary slip system
Fig. 1. (a) Shear stress–shear strain curves and (b and c) optical micrographs showing the deformation processes of the uncharged and hydrogen-charged single-crystalline specimens, SC-U and -H, respectively.
charging time was chosen such that the 20 μm-thick specimens are saturated with solute hydrogen. The saturated hydrogen content under the corresponding hydrogen charging conditions was measured to be 101 mass ppm by thermal desorption spectrometry. Tensile tests were started within 3 h after hydrogen charging through cutting the gauge part supporting members using a laser beam. Tensile tests were performed at room temperature under laboratory atmospheric conditions using micro-gluing for gripping and a piezoelectric actuator at a displacement rate of 0.1 μm s−1, corresponding to a strain rate of 2 × 10−3 s−1. The setup has been described in more detail in [19]. The gauge section of the tensile specimen was monitored during tensile testing using an optical microscope in order to dynamically measure the strain as a function of time. For EBSD analysis after failure or after the interruption of loading, the specimens were further fabricated
{111} b110 N with a Schmid factor of 0.47. This is confirmed by the slip steps that are visible in the optical micrograph in Fig. 1b. The critical resolved shear stress (CRSS) was in agreement with the value obtained by loading in the [111] direction [20]. In the stress–strain curve, a stage I deformation up to a shear strain of about 0.52 can be identified. It is argued that easy glide takes place with the primary slip system being activated (Fig. 1c), although some hardening occurs. When the secondary slip system is activated owing to crystal rotation, work hardening becomes more pronounced (stage IIa) induced through mutual interactions between the dislocations on the two slip systems. The hardening rate is somewhat reduced in stage IIb and, subsequently, the workhardening rate increases again in stage III. The appearance of the reduced hardening region of stage IIb can be presumably attributed to the formation of martensite, as previously reported by Tsurui et al. [21]. In fact, in-situ observations made during micro-tensile testing of the SC-U specimen revealed that gradual undulation, which differs from slip steps, appeared in region in stage IIb (compare Fig. 1c and d). The increase in the work-hardening rate in stage III can be caused by work hardening of the formed martensite. While the shape of the stress–strain curve in the SC-H specimen is similar to the one for the SC-U specimen, the stage I is different with a plateau region after the onset of yielding at about 110 MPa, which is 39% higher than that in the SC-U specimen. Overall, the solute hydrogen increases the flow stresses but shortens the length of each stage, resulting in a smaller strain-to-failure. While the trace of the primary slip plane appeared at the onset of yielding (Fig. 1e), the stage I deformation in the SC-H specimen was localised within a short length of the specimen gauge part, as shown in Fig. 1f. In the stage II, deformation was widely spread throughout the entire gauge part through the activation of the multiple slip systems (Fig. 1g). Fig. 2 compares the fracture morphology of the single-crystalline specimens with and without hydrogen pre-charging. The uncharged specimen exhibits a ductile fracture with strong necking (Fig. 2a and b), whereas the fracture surface of the hydrogen-charged specimen is characterised by a typical hydrogen-induced quasi-cleavage (Fig. 2c and d). The insets in Fig. 2b and d show the EBSD maps of SC-U and SC-H specimens after fracture, respectively. The analysis reveals that martensite variants developed in both specimens, mainly with their habit planes parallel to the primary slip planes. In the SC-H specimen, variants with their habit planes parallel to the secondary slip plane were also observed (Fig. 2d). This can be induced by plastic constraint at the shoulder part of the specimen. Also, it is confirmed that the quasicleavage facets were formed parallel to the two martensite habit planes (Fig. 2c and d).
178
Y. Mine et al. / Scripta Materialia 113 (2016) 176–179
Fig. 2. SEM micrographs comparing the fracture morphology of the uncharged (a and b) and hydrogen-charged (c and d) single-crystalline specimens. The insets in (b and d) show the EBSD maps taken in the areas marked by the boxes, after the broken specimens were further prepared by FIB, making a longitudinal cut.
In the stress–strain curve of the hydrogen-charged specimen (Fig. 1a), a slight decrease in the stress was observed just after the onset of yielding. It is argued that this relates to the martensitic transformation which is accompanied by a stress decrease. In order to analyse this in more detail, EBSD was applied after interrupting the tensile test at the end of stage I for a hydrogen-charged single-crystalline specimen. The orientation of the LD is close to the [110]-direction of the crystal. Fig. 3 shows the shear stress–shear strain curves (a), and the deformation morphology (b). The flow stress in the plateau regime is about 130 MPa, which is higher than that for the SC-H specimen. This may be related to a higher amount of hydrogen retained as the increase in the flow stress due to hydrogenation depends on the hydrogen content [7]. Observation of the top (Fig. 3b) and side (not shown) surfaces revealed that primary and secondary slips were activated. Fig. 3c and d shows the result of an EBSD analysis on the cross-section of the specimen including the (111)- and (110)-pole figures of austenite (c) and
martensite (d), and the corresponding colour-coded maps. Three martensite variants with their habit planes parallel to the primary slip plane developed, while some fraction of the parent austenite crystal was retained. In other words, the martensite/austenite lamellar microstructure formed parallel to the primary slip plane. The hydrogen solubility in austenite is approximately an order of magnitude higher than that in martensite [22]. If hydrogen-containing austenite is transformed to martensite, excess hydrogen corresponding to the difference in solubility between the two phases is generated. Therefore, the excess hydrogen diffuses out of the martensite and into the austenite. It can be further argued that the hydrogen concentrates close to the interface in the austenite because the hydrogen diffusivity is significantly lower in austenite than in martensite [23]. As the martensite formation proceeds, the volume of retained austenite between the martensite laths decreases, leading to local supersaturation of hydrogen. Thus, it is concluded that quasi-cleavage in a metastable
Fig. 3. (a) Shear stress–shear strain curves for hydrogen-charged single-crystalline specimens. The test was interrupted prior to fracture. (b) SEM micrograph of the deformation morphology of the deformed tensile specimen. (c and d) EBSD analysis showing colour-coded maps of the area marked by the box in (b) for austenite and martensite, respectively, and the corresponding (110) and (111) pole figures. Circles and triangles show the habit planes and directions, respectively.
Y. Mine et al. / Scripta Materialia 113 (2016) 176–179
179
Fig. 4. (a) Shear stress–shear strain curves and (b and c) fracture surfaces for the uncharged and hydrogen-charged twinned bi-crystalline specimens, TW-U and TW-H, respectively.
austenitic steel is caused by excess-hydrogen-assisted slip localisation in the retained austenite between the deformation-induced martensite laths along their habit planes. This conclusion is in agreement with the fact that in our previous work [24] retained austenite was found by transmission electron microscopy in the vicinity of the fracture surface of a hydrogen-charged micro-tensile specimen. Also, it has been found that in lath martensite single packet structure slip occurs preferentially parallel to the habit plane [25], and that this anisotropic slip behaviour might stem from the presence of the retained austenite between the martensite laths as indicated by finite element simulations [26]. Also, it is important to note that the proposed mechanism is consistent with the observation that pre-existing martensite does not promote HE because it does not introduce excess hydrogen in the austenite. Fig. 4a shows the shear stress–shear strain curves of an uncharged and a hydrogen-charged twinned bi-crystalline specimen, TW-U and TW-H, respectively. Unlike the single-crystalline specimens, the workhardening rate did not decrease but rather increased in the transition from stage II to III. This may be because the presence of the twin boundary promotes the activation of multiple-slip glides in an early stage and therefore results in the formation of more martensite nuclei. For the TW-H specimen, the yield strength was higher than that of the TW-U specimen. Like the single-crystalline specimens, hydrogen shortened the length of each stage and decreased the elongation-to-failure. Fig. 4b and c shows the fracture surfaces of the TW-U and TW-H specimen, respectively. While the TW-U specimen exhibited a necking almost symmetric with respect to the twin boundary (Fig. 4b), quasicleavages appeared with different fracture planes in each crystal of the TW-H specimen (Fig. 4c). The formed martensite variants had habit planes parallel to the primary slip planes (not shown in the figure). This indicates that the quasi-cleavage occurs in each crystal along the primary slip system based on the same mechanism as in the singlecrystalline specimen. Micro-tensile tests were performed on single crystals and twinned bi-crystals with and without hydrogen pre-charging in a 304 metastable austenitic steel to elucidate the occurrence of quasi-cleavage induced by hydrogen. In all specimens, the stress–strain curves were composed of three work-hardening stages. The hydrogen pre-charge shortened the length of each stage, resulting in a smaller strain-to-failure compared to the uncharged specimen. A quasi-cleavage fracture, whose facets correspond to the habit planes of martensite formed during deformation, was observed in the hydrogen-charged specimens. The hydrogencharged specimens also exhibited a small stress drop after the onset of yielding. In this stage, a lamellar microstructure of martensite within the parent austenite is formed. Based on these observations, the following mechanism is suggested for the hydrogen-induced quasi-cleavage.
Excess hydrogen is generated by the deformation-induced martensitic transformation since the solubility of hydrogen is smaller in martensite. As a result, hydrogen concentrates in the retained austenite between the martensite laths. The high hydrogen concentration leads to localised shear in the retained austenite along the habit plane of the martensitic transformation, leading to quasi-cleavage fracture in metastable austenitic steels. 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), and in part by ‘Program for Advancing Strategic International Networks to Accelerate the Circulation of Talented Researchers’ (R2608).
References [1] D. Eliezer, D.G. Chakrapani, C.J. Altstetter, E.N. Pugh, Metall. Trans. A 10 (1979) 935. [2] L. Zhang, M. Wen, M. Imade, S. Fukuyama, K. Yokogawa, Acta Mater. 56 (2008) 3414. [3] T. Kanezaki, C. Narazaki, Y. Mine, S. Matsuoka, Y. Murakami, Int. J. Hydrog. Energy 33 (2008) 2604. [4] C. San Marchi, T. Michler, K.A. Nibur, B.P. Somerday, Int. J. Hydrogen Energy 35 (2010) 9736. [5] Y. Mine, T. Kimoto, Corros. Sci. 53 (2011) 2619. [6] H.K. Birnbaum, P. Sofronis, Mater. Sci. Eng. A 176 (1994) 191. [7] D.P. Abraham, C.J. Altstetter, Metall. Mater. Trans. A 26 (1995) 2849. [8] D.P. Abraham, C.J. Altstetter, Metall. Mater. Trans. A 26 (1995) 2859. [9] C.L. Briant, Metall. Trans. A 10 (1979) 181. [10] H. Hänninen, T. Hakkarainen, Metall. Trans. A 10 (1979) 1196. [11] R. Liu, N. Narita, C. Altstetter, H. Birnbaum, E.N. Pugh, Metall. Trans. A 11 (1980) 1563. [12] D. Sun, G. Han, S. Vaodee, S. Fukuyama, K. Yokogawa, Mater. Sci. Technol. 17 (2001) 302. [13] G. Han, J. He, S. Fukuyama, K. Yokogawa, Acta Metall. 46 (1998) 4559. [14] R.M. Vennett, G.S. Ansell, Trans. Am. Soc. Met. 60 (1967) 242. [15] R.B. Benson Jr., R.K. Dann, L.W. Roberts Jr., Trans. Metall. Soc. AIME 242 (1968) 2199. [16] A. Inoue, Y. Hosoya, T. Matsumoto, Trans. Iron Steel Inst. Jpn. 19 (1979) 170. [17] J.R. Buckley, D. Hardie, Corros. Sci. 34 (1993) 93. [18] L. Zhang, Z. Li, J. Zheng, Y. Zhao, P. Xu, C. Zhou, X. Li, Int. J. Hydrog. Energy 38 (2013) 8208. [19] T. Ito, Y. Mine, M. Otsu, K. Takashima, J. Jpn. Inst. Met. Mater. (2016) in press. [20] Y. Mine, K. Koga, K. Takashima, O. Kraft, Unpublished data (2015), Kumamoto University. [21] T. Tsurui, S. Inoue, K. Matsuda, H. Ishigaki, K. Murata, K. Koterazawa, J. Soc. Mater. Sci., Jpn. 50 (2001) 1115. [22] Y. Mine, C. Narazaki, T. Kanezaki, S. Matsuoka, Y. Murakami, in: E.E. Gdoutos (Ed.), Fracture of Nano and Engineering Materials and Structures, Springer, Netherlands 2006, pp. 907–908. [23] K. Sakamoto, H. Katayama, J. Jpn. Inst. Metals 46 (1982) 805. [24] Y. Mine, K. Hirashita, M. Matsuda, K. Takashima, Metall. Mater. Trans. A 42 (2011) 3567. [25] Y. Mine, K. Hirashita, H. Takashima, M. Matsuda, K. Takashima, Mater. Sci. Eng. A 560 (2013) 535. [26] F. Maresca, V.G. Kouznetsova, M.G.D. Geers, J. Mech. Phys. Solids 73 (2014) 69.