Anisotropy in Hydrogen Embrittlement Resistance of Drawn Pearlitic Steel Wire Evaluated by in-situ Miniature Test and Scanning Electron Microscopy during Plasma Hydrogen Charging

Abstract

In drawn pearlitic steel wire, hydrogen embrittlement cracks are preferentially formed along cementite/ferrite interfaces perpendicular to the tensile stress, predicting an anisotropy in the hydrogen embrittlement resistance of the wire. To validate this prediction, notched miniature tensile test specimens were extracted from the wire with their orientation changed, and in-situ miniature tensile tests were conducted during plasma hydrogen charging in a scanning electron microscope. The fracture process was also investigated from secondary electron images simultaneously obtained during these in-situ miniature tensile tests. As a result, it was revealed that, in accordance with the prediction, the fracture stress is decreased by the plasma hydrogen charging for the specimens with the tensile direction perpendicular to the drawing direction of the wire whereas it is hardly changed for the specimen with the tensile direction parallel to the drawing direction. In the secondary electron images, a stationary subcrack was observed before the fracture only for the specimen with the tensile direction parallel to the drawing direction. Based on finite element method analysis results, the subcrack was found to be formed by the triaxial stress due to the notch. The difference in the presence and absence of the subcrack depending on the specimen orientation was also well explained from the preferential hydrogen embrittlement crack formation along the cementite/ferrite interfaces perpendicular to the tensile stress.

1. Introduction

It is well known that hydrogen embrittlement (HE) resistance decreases with increase in steel strength. On the other hand, drawn pearlitic steel wire (DPSW), obtained by cold-drawing pearlitic steel, has a superior HE resistance compared to the other steel with the same strength such as martensitic one.^1,2,3,4) As schematically shown in Fig. 1, the DPSW has a fiber structure and is composed of lamellar cementite (θ) and lamellar ferrite (α) parallel to the drawing direction.^5,6,7,8) To explore the origins of this superior HE resistance, the authors fabricated micro-cantilevers with different orientations to the θ/α interface by focused ion beam and conducted in-situ microbending tests during cathodic hydrogen charging by electrochemical nanoindentation.⁹⁾ Stationary HE cracks formed in the microbending tests were also observed by scanning/transmission electron microscopy. Consequently, it was clarified that the θ/α interface perpendicular to the tensile stress is a preferential crack formation site. It was also clarified that the lamellar θ prevents the HE cracks from propagating and the θ/α interface parallel to the tensile stress is difficult to fracture. Based on these results, the authors concluded that the directional alignment of the θ/α interfaces exhibits the superior HE resistance of the DPSW.

Fig. 1. Schematic of the microstructure of DPSW. (Online version in color.)

The above findings mean that the HE resistance of the DPSW is anisotropic. Namely, the HE resistance is much lower in the radial direction of the DPSW than in the drawing direction because the θ/α interfaces perpendicular to the tensile stress, easily fractured by the hydrogen atoms, are present for the radial direction as shown in Fig. 1. It is known that, when the DPSWs are strained in the drawing direction during the hydrogen charging, they exhibit delamination or tearing fractures,^10,11,12,13) which also implies the anisotropy in the HE resistance. However, no experimental study has been conducted so far that directly shows this anisotropy by measuring strength or ductility in mechanical tests. This is because the typical diameter of the DPSW is a few mm and macroscopic specimens cannot be extracted from the DPSW with their orientation changed. The in-situ microbending test⁹⁾ described above also cannot be used for this purpose because its evaluation area is too small and a few μm². Namely, the in-situ microbending test can evaluate the HE resistance of colonies, where the θ/α interfaces align in the same direction, but cannot that of the DPSW which is aggregates of the colonies. On the other hand, T. Fujita et al. extracted miniature tensile test specimens with a few mm size from the DPSW in different orientations and showed an anisotropy in tensile strength of the DPSW in air; the tensile strength is higher in the drawing direction of the DPSW than in the radial direction.¹⁴⁾ However, due to the lack of waterproof or pressure resistant capabilities of commercial miniature tensile testing machines, in-situ miniature tensile tests during the hydrogen charging have rarely been conducted.^15,16) It is noted that the in-situ miniature tensile test is essential because the hydrogen atoms desorb immediately from such thin specimens.

Hydrogen is charged to steel when it is exposed to hydrogen plasma.^15,17,18,19) The authors injected hydrogen microplasma jet onto the strained martensitic steel and in-situ observed its HE cracks by scanning electron microscopy (SEM).²⁰⁾ It is very likely that, if a miniature tensile testing machine is installed on this SEM apparatus (hydrogen plasma SEM: HP-SEM), miniature tensile tests can be conducted during the hydrogen charging and the detailed fracture process can be also investigated from the in-situ obtained SEM images. Thus, in the present study, to validate the anisotropy in the HE resistance of the DPSW, the miniature tensile test specimens were extracted from the DPSW in different orientations, and the miniature tensile tests were conducted during the plasma hydrogen charging in the HP-SEM.

2. Experiments

2.1. Equipment

The present experiment was conducted using the HP-SEM apparatus in Nippon Steel Corporation, which has almost the same structure as that in Kochi University of Technology used in the previous authors’ study.²⁰⁾ A schematic of the apparatus is shown in Fig. 2 and a secondary electron (SE) image of its measurement part in Fig. 3. A commercial miniature tensile testing machine (DEBEN, MT200) with a maximum load of 180 N and a double-opening chucks was mounted on the sample stage of the SEM, and a stainless-steel gas nozzle with an orifice diameter of 40 μm was placed at a tilt angle of 45° to the tensile axis. Piezo actuators were also installed to move the gas nozzle with respect to the specimen in the triaxial directions at nanometer steps.

Fig. 2. Schematics of the HP-SEM apparatus. (Online version in color.)

Fig. 3. SE image of the measurement part of the HP-SEM.

To generate plasma from hydrogen gas by glow discharge, the gas nozzle was connected to a hydrogen generator (Horiba, OPGU-7200) with a hydrogen gas purity of 99.999%, and a mass flow controller was installed in the flow path. The gas nozzle and the specimen were connected to a DC pulse power supply so that the specimen corresponds to a cathode. As a ballast resistor to sustain a stable glow discharge, that of 200 kΩ was inserted in the circuit. The output of the power supply and the discharge current were monitored with a digital oscilloscope. It is essential for the in-situ SEM observations to apply pulse voltage like this and to synchronize its frequency with the frame rate of the SEM because the electrons in the generated hydrogen plasma saturate the signals of the SE detector of the SEM as described in detail elsewhere.²⁰⁾

When the pressure in the vacuum chamber of the HP-SEM becomes more than 1 Pa, the glow discharge occurs unexpectedly in spaces other than that between the gas nozzle and the specimen. To avoid this, the vacuum chamber was evacuated using two turbo molecular pumps; in addition to the existing pump of the SEM apparatus, another pump was installed. For safety, the hydrogen gas exhausted from these pumps were diluted with a large amount of air and vented outdoor.

2.2. Materials and Specimens

The JIS SWRS82B steel with chemical compositions shown in Table 1, used in the authors’ previous study,²¹⁾ was also used in the present study. Namely, a rod with a diameter of 13 mm was heated at 880°C for 600 s, and then subjected to patenting treatment in a lead bath at 530°C to obtain pearlitic structures. The rod was cold-drawn into a wire with a diameter of 5 mm followed by annealing at 300°C for 600 s, and finally the sample DPSW was obtained. The tensile strength of the DPSW in the drawing direction was measured to 2000 MPa by conventional macroscopic tensile tests with unnotched specimens in air.

Table 1. Chemical composition of the sample material. The unit is mass%.

Fe	C	Si	Mn	P	S
Bal.	0.84	0.25	0.73	0.017	0.005

The miniature tensile test specimens with a U-shape notch shown in Fig. 4 were extracted from the DPSW by wire electrical discharge machining. The notch radius is 0.1 mm, and the width and thickness of the specimen at the notch position are 0.4 mm and 0.2 mm, respectively. Both sides of the specimen were finished by machining. To evaluate the anisotropy in the HE resistance of the DPSW, three kinds of specimens with different orientations (type-I to -III) were prepared as shown in Fig. 5; the drawing direction is parallel to the thickness direction for the type-I specimen, the width direction for the type-II specimen, and the tensile direction for the type-III specimen.

Fig. 4. Shape and dimension of the miniature tensile test specimen. The unit is mm.

Fig. 5. Orientations of the miniature tensile tests specimen in the DPSW.

2.3. Experimental Procedures

2.3.1. Miniature Tensile Tests

The miniature tensile tests during the plasma hydrogen charging in the HP-SEM apparatus were conducted as follows.

Relative position of the gas nozzle to the specimen varied slightly with the chuck opening of the miniature tensile testing machine. Thus, the stroke was increased stepwise in the displacement control mode as schematically shown in Fig. 6, and the gas nozzle position was adjusted at each step. Namely, the stroke was increased at a speed of 1.5 mm/s by the displacement corresponding to 10 N. Subsequently, the stroke was fixed, the gas nozzle position was quickly adjusted within 30 s, and hydrogen plasma was injected to the center of the miniature tensile specimen for more than 300 s. This routine of the stroke increase, the gas nozzle position adjustment, and the hydrogen plasma injection were repeated until the fracture of the specimen, and the fracture stress was measured. During the test, SE images of the specimen were simultaneously obtained at a frame rate of 1.5 Hz in the high-vacuum mode.

Fig. 6. Schematic of time variation of the stroke during the miniature tensile test. (Online version in color.)

In Fig. 7, actual time variations of the discharge current and the applied pulse voltage measured with the digital oscilloscope are shown. Hydrogen content of the specimen is likely to be the same for the same discharge current. Thus, the discharge current was fixed in the present study. Namely, throughout the tests, the hydrogen gas flow was fixed at 80 sccm, and frequency, peak voltage, and duty of the pulse voltage at 1.5 Hz, 1.3 kV, and 2%, respectively. The frequency of 1.5 Hz is the same as the frame rate of the SEM described above. After the stroke increase was finished, only the gas nozzle position was quickly adjusted by the piezo actuators so that the discharge current commonly becomes 200 μA.

Fig. 7. Actual time variations of the applied voltage and the discharge current. (Online version in color.)

The duration time t until the hydrogen content in the specimen reaches a steady state is estimated to 112 s from the equation,^22,23,24)

  

$$t=\frac{0.28}{D_H{L}^2}$$(1)

where L and D_H are the thickness of the specimen (0.2 mm) and the hydrogen diffusion coefficient of the DPSW (1.0×10⁻¹⁰ m²/s ²⁵⁾), respectively. Thus, it seems that the plasma injection time of 300 s during the stroke holding described above is long enough to cause HE for the present miniature tensile specimen.

From SE images obtained by tilting the specimen, the distance between the gas nozzle orifice and the specimen along the gas injection axis was measured to approximately 500 μm in the present discharge condition. On the other hand, it is known that when C₂H₂ plasma is injected onto a Si substrate from a similar gas nozzle, a C–H film is deposited thereon.²⁶⁾ Based on the C–H film deposition area, it is estimated that the specimen surface inside a cone with an open-angle of about 60° under the gas nozzle orifice is affected by the plasma. Thus, as schematically shown in Fig. 8, it seems that an egg-shape region with a major axis of 1200 μm and a minor one of 600 μm on the specimen surface is affected by the hydrogen plasma. Namely, it seems that hydrogen is charged to the entire region across the width of the specimen at the notch position.

Fig. 8. Schematic of the region affected by the hydrogen plasma. (Online version in color.)

Three specimens were subjected to the miniature tensile test during the hydrogen plasma charging for the type-I and -II whereas four specimens for the type-III. For comparison, fracture stress without the hydrogen charging was also measured by straining the specimen in air at a speed of 0.1 mm/s until the fracture. One specimen was subjected to the miniature tensile test in air each for the type-I to -III.

2.3.2. Finite Element Method Analysis

As described later, a subcrack was formed in the tensile direction in the miniature tensile test of the type-III specimen. To investigate the formation process of the subcrack, stress distributions around the notch of the specimen were also analyzed by finite element method (FEM). The miniature tensile specimen was created as a two-dimensional, quarter-symmetrical model of the specimen gauge section using Abaqus CAE 2018. The geometry and boundary conditions of the model are shown in Fig. 9. The model was meshed using a 4-node plane strain element (CPE4). The total number of nodes and elements were 17868 and 17588, respectively. Because no available data exists, material properties were assumed to be isotropic and those in the drawing direction were employed. That is, Young’s modulus of 200 GPa, Poisson’s ratio of 0.3, and the true stress-plastic strain curve in the drawing direction derived from the tensile test using unnotched specimens were employed. Details of the tensile test have been presented elsewhere.²⁷⁾

Fig. 9. Geometry and boundary conditions for the FEM model. The unit is mm.

3. Results

3.1. Miniature Tensile Test Results in Air

In Fig. 10, stress-stroke relations obtained in the miniature tensile tests in air are shown, where the stress is nominal one obtained by dividing the load by the cross-sectional area at the notch position (0.08 mm²). For the type-I and -II specimens, the fracture stresses were 2063 MPa and 1638 MPa, respectively. On the other hand, the type-III specimen endured without fracture up to the maximum load limit of the miniature tensile testing machine (180 N), and the fracture stress was more than 2250 MPa. Thus, consistent with the miniature tensile tests by T. Fujita et al.,¹⁴⁾ the tensile strength in the drawing direction of the DPSW was higher than that in the radial direction in air. For the type-III specimen, the obtained tensile strength (>2250 MPa) is somewhat larger than that obtained in the conventional macroscopic tensile test (2000 MPa).²¹⁾ This is likely due to apparent stress increase caused by plastic constraint resulting from the notch of the specimen.^28,29)

Fig. 10. Stress-stroke relations obtained by the miniature tensile tests in air. (Online version in color.)

In Fig. 11, SE images of the type-I and -II specimens after the fracture are shown. The fracture surface profile was largely jagged for the type-I specimen and slightly jagged for the type-II specimen. In Fig. 12, the SE images of the fracture surfaces of type-I and -II specimens in Fig. 11 are shown. The upper sides of the specimens correspond to the SEM observation surfaces in Fig. 11. For the type-I specimen, striations were observed in the drawing direction, that is, the thickness direction entire fracture surface. For the type-II specimen, the striations were also observed in the drawing direction, that is, the width direction. However, compared to the type-I specimen, the striations were less distinct and there were also some regions without the striations.

Fig. 11. SE images of the (a) type-I and (b) type-II specimens after the miniature tensile tests in air.

Fig. 12. SE images of the fracture surfaces of the (a) type-I and (b) type-II specimens formed in air.

3.2. Miniature Tensile Test Results during Plasma Hydrogen Charging

In Fig. 13, time variations of stress in the miniature tensile tests during the plasma hydrogen charging are shown for the representative type-I to -III specimens. During the stroke holding, stress relaxation occurred, and the stress gradually decreased with time. Most of the specimens fractured either during the stroke holding like the type-II or the stroke increase like the type-I and -III in the figure.

Fig. 13. Time variations of stress for the representative specimens in the miniature tensile tests during the plasma hydrogen charging. (Online version in color.)

In Table 2, the fracture stress and the fracture timing of all the tested specimens are summarized as well as the fracture stresses in air described in the section 3.1. For intuitive understanding, the fracture stresses are also compared in a graph in Fig. 14. For the type-I, all three specimens fractured during the stroke increase. For the type-II, two specimens fractured during the stroke increase and one specimen during the stroke holding. The incubation time until the fracture after the stroke increase was about 200 s. For the type-III, two specimens were fractured during the stroke increase, one specimen during the stroke holding. The incubation time was 2 s, and the fracture occurs soon after the stroke increase was finished. The remaining one specimen endured without fracture up to the maximum load limit of the miniature tensile testing machine (180 N). The fracture stress was well reproduced for the same specimen type, and the average was 892 ± 127 MPa, 796 ± 55 MPa, and >2108 MPa for the type-I, -II, and -III specimens, respectively. Namely, the plasma hydrogen charging decreased the fracture stress by about half for the type-I and -II specimens, and no significant difference was identified in the fracture stress during the hydrogen charging between these two types. On the other hand, the plasma hydrogen charging did not have a significant effect on the fracture stress for the type-III specimen.

Table 2. Fracture stress and fracture timing of the tested specimens.

Specimen type	Test environment	Fracture timing	Fracture stress (MPa)	Average fracture stress (MPa)
Type I	Air	–	2063	2063
	H plasma	Increasing	746	892 ± 127
		Increasing	955
		Increasing	976
Type II	Air	–	1638	1638
	H plasma	Increasing	847	796 ± 55
		Holding	738
		Increasing	801
Type III	Air	No fracture	>2250	>2250
	H plasma	No fracture	>2250	>2108
		Increasing	1948
		Holding	2084
		Increasing	2150

Fig. 14. Fracture stress of all the tested specimens. (Online version in color.)

From the SE images simultaneously obtained during the miniature tensile tests, the fracture process was also investigated. For the type-I and -II, all the specimens fractured instantaneously, and no stationary crack was observed until the fracture. The SE images of the specimens after the fracture are shown in Fig. 15. For the type-I specimen, the fracture surface profile was jagged similarly to in air. On the other hand, the type-II specimen fractured cleanly, and the fracture surface profile was more linear than in air.

Fig. 15. SE images of the (a) type-I and (b) type-II specimens after the miniature tensile tests during the plasma hydrogen charging.

On the other hand, for the type-III, a stationary subcrack was observed before the fracture for one of the three fractured specimens. This subcrack was formed during the stroke increase toward the value corresponding to 160 N (2000 MPa). Soon after the next stroke increase toward the value corresponding to 170 N (2125 MPa) was finished, the fracture occurred at 2084 MPa. The SE image during the stroke holding after the stroke increase toward the value corresponding to 160 N is shown in Fig. 16(a). During the glow discharge, the electrons in the plasma saturate the signals of the SE detector as described above, resulting in bright band area at the bottom of the image.²⁰⁾ On the other hand, in the other areas, the specimen surface was imaged clearly, and the subcrack was observed in the tensile direction in the center region of the specimen. In Fig. 16(b), the SE image obtained after the fracture is shown. The fracture surface profile included the subcrack and was sharply pointed in the tensile direction.

Fig. 16. SE image of the type-III specimen (a) just before and (b) after the fracture during the plasma hydrogen charging.

In Fig. 17, SE images of the fracture surfaces of all the type -I to -III specimens in Figs. 15 and 16 are shown. The upper side of the specimen correspond to the plasma injection side. For the type-I specimen, striations were observed in the drawing direction, that is, in the thickness direction. No significant difference was identified in the fracture surface morphology between in air and during the plasma hydrogen charging. For the type-II specimen, the striations were also observed in the drawing direction, that is, in the width direction. Compared to in air, the striations were more distinct. On the other hand, for the type-III specimen, the fracture surface was relatively smooth, and the striation was absent.

Fig. 17. SE images of the fracture surfaces of the (a) type-I, (b) type-II, and (c) type-III specimens formed during the plasma hydrogen charging.

3.3. Stress Distributions in the Specimen Analyzed by Finite Element Method

When the notched tensile specimens are strained, axial stress also appears in the width direction although its magnitude is small.^28,29) Thus, it is very likely that the initial crack formation causing the fracture in the type-II specimen and the subcrack formation in the type-III specimen during the hydrogen charging occur at the same microstructure in the same mechanism because of the symmetry of these two specimens. To confirm this situation, distributions of the axial stress in the tensile direction σ₂₂ at the average fracture stress of type-II specimen (796 MPa) and those in the width direction σ₁₁ at the subcrack formation stress of the type-III specimen (2000 MPa) were analyzed by FEM.

In Fig. 18, the analyzed distributions of σ₂₂ and σ₁₁ are shown. The σ₂₂ value was highest at the notch bottom and 2674 MPa. On the other hand, the σ₁₁ value was high at the center region of the specimen (2490 MPa) and the curvature change point of the notch (2687 MPa). The high σ₁₁ at the center region is likely to form the HE crack than that at the curvature change point because hydrogen content is higher at the center region which is the injection center of the hydrogen microplasma jet as shown in Fig. 8. The σ₁₁ value at the center region is close to the σ₂₂ value at the notch bottom, indicating that the initial crack formation in the type-II specimen and the subcrack formation in the type-III occur at the same microstructure in the same mechanism.

Fig. 18. Distributions of (a) σ₂₂ at the average fracture stress of the type-II specimen and (b) σ₁₁ at the subcrack formation stress of the type-III specimen. (Online version in color.)

4. Discussion

The fracture stress during the plasma hydrogen charging was higher for the type-III specimen than for the type-I and -II specimens (Table. 2 and Fig. 14). The tensile direction is perpendicular to the drawing direction for the type-I and -II specimens and parallel for the type-III specimen (Fig. 5). Thus, it was successfully demonstrated by the in-situ miniature tensile tests during the plasma hydrogen charging that the HE resistance of the DPSW is anisotropic, and higher in the drawing direction than in the radial direction. Given that the θ/α interfaces perpendicular to the tensile stress is a preferential crack formation site as reported in the previous authors’ study,⁹⁾ the detailed fracture behaviors of each type of specimen are also well explained as schematically shown in Fig. 19.

Fig. 19. Schematic of the fracture process of the (a) type-I, (b) type-II, and (c) type-III specimens. (Online version in color.)

The type-I specimen has the θ/α interfaces perpendicular to its tensile direction, and such interfaces at the notch bottom become an initial crack site. Large stress concentration around the initial crack tip forms a new crack along the adjacent θ/α interface perpendicular to the tensile direction, and the new and the initial cracks are connected. This formation and connection of the cracks along the θ/α interfaces occurs repeatedly, the crack propagates, and the specimen is fractured. In the in-situ microbending tests by the authors, the θ/α interface was fractured with only minimal plastic deformation.⁹⁾ It was also clarified that fracture toughness of the θ/α interface is very small and 1.7 MPa·m^1/2, which is almost the same as values of ceramics. Because of these small toughness and fracture toughness, the stress of the unfractured ligament area and hence the stress concentration around the crack tip increases with the crack propagation during the tensile test. Consequently, the HE crack propagates immediately without arresting, leading to the unstable fracture especially for the present case where the all the type-I specimens fractured during the stroke increase (Table 2).³⁰⁾ The HE crack propagates preferentially along the θ/α interfaces perpendicular to the tensile direction like this. Thus, the fracture surface profile becomes jagged (Fig. 15(a)), and the striations appear on the fracture surface in the drawing direction, that is, the thickness direction of the specimen (Fig. 17(a)). Because the specimen undergoes unstable fractures like this, no stationary crack is observed in the in-situ obtained SE images.

The type-II specimen also has the θ/α interfaces perpendicular to its tensile direction. Similarly to the type-I specimen, the type-II specimen undergoes unstable fracture from the θ/α interfaces perpendicular to the tensile direction at the notch bottom. The HE crack propagates preferentially along the θ/α interfaces perpendicular to the tensile direction, leading to the linear fracture surface profile (Fig. 15(b)) and the striations on the fracture surface in the drawing direction, that is, the width direction of the specimen (Fig. 17(b)). No stationary crack is again observed in the in-situ obtained SE images because the specimen undergoes unstable fractures.

For the type-I and -II specimens, the fracture stress is equivalent to the crack initiation stress. On the other hand, the crack initiation site is commonly the θ/α interfaces perpendicular to the tensile direction. Thus, there is no significant difference in the fracture stress between these two specimen types (Table 2 and Fig. 14).

On the other hand, the type-III specimen has no θ/α interfaces perpendicular to its tensile direction and endures up to the higher stresses. At such higher stresses, the axial stress in the width direction of the specimen is no longer ignored as indicated by the FEM analysis (Fig. 18) and it delaminates the θ/α interfaces perpendicular to the width direction in the center region of the specimen. This crack does not lead to the fracture directly because it is parallel to the tensile direction and the higher σ₁₁ region is also localized in the center of the specimen. Consequently, the stationary subcrack is observed in the in-situ obtained SE images before the fracture (Fig. 16).

No subcrack was formed before the fracture for two of three fractured type-III specimens. This can be attributed to the spatial variations of the θ/α interface orientation. The θ/α interface is orientated at random around the tensile axis for the type-III specimen (Fig. 1). It is very likely that, statistically, the θ/α interface perpendicular to the width direction is absent in the higher σ₁₁ region for these specimens.

5. Summary

The HE cracks are preferentially formed along the θ/α interfaces perpendicular to the tensile stress in the DPSW. This predicts that the HE resistance of the DPSW is anisotropic; the HE resistance is much higher in the drawing direction than the radial direction. In the present study, to validate this prediction, the notched miniature tensile test specimens were extracted from the DPSW with their orientation changed, and the in-situ miniature tensile tests during the plasma hydrogen charging were conducted in the HP-SEM apparatus. The fracture process was also investigated from the SE images simultaneously obtained during the miniature tensile tests. Consequently, the following findings were obtained.

(i) For the type-I and II specimens with the tensile direction perpendicular to the drawing direction, the fracture stress was decreased by about half by the plasma hydrogen charging. On the other hand, for the type-III specimen with the tensile direction parallel to the drawing direction, the fracture stress was hardly changed by the plasma hydrogen charging. Thus, it was clarified that the HE resistance of the DPSW is higher in the drawing direction than in the radial direction in accordance with the prediction.

(ii) For the type-I and -II specimens, no stationary crack was observed in the SE images until the fracture. On the other hand, for the type-III specimen, the stationary subcrack parallel to the tensile direction was observed in the center region of the specimen in the SE images before the fracture.

(iii) The fracture surfaces of the type-I and -II specimens formed during the plasma hydrogen charging had the striations parallel to the drawing direction. On the other hand, the fracture surfaces of the type-III had no striation and relatively smooth.

(iv) It was clarified by the FEM analysis that, at the higher fracture stresses of the type-III specimen, the axial stress in the width direction σ₁₁ due to the notch becomes significant. Its magnitude approximately matched to that of the axial tress in the tensile direction σ₂₂ at the notch bottom at fracture stresses of the type-II specimen.

(v) For the type-I and -II specimens, it was considered that the HE crack initiates along the θ/α interfaces perpendicular to the tensile direction at the notch bottom and propagates immediately, leading to the unstable fracture without the stationary crack. On the other hand, for the type-III specimen, it was considered that the stationary subcrack is formed along the θ/α interface parallel to the tensile direction by the triaxial stress due to the notch before the fracture. The preferential HE crack formation at the θ/α interfaces also well explained the observed difference in the fracture stress, the fracture surface morphology, and the presence or absence of the stationary subcrack depending on the specimen type.

To the authors’ knowledge, the present study is the first one that simultaneously conducted in-situ miniature tensile test and in-situ SEM observations of the HE cracks during the hydrogen charging. Historically, the miniature tensile tests have been used to measure strength of local areas such as a grain boundary, a weld region, and a surface film, which sometimes play important roles in the HE of steel. On the other hand, the in-situ SEM observation enables us to understand more deeply what changes in the material lead to the mechanical properties obtained by the miniature tensile test as demonstrated in the present study. Thus, the developed novel method is highly beneficial for elucidation of the HE mechanism and exploration of microstructure with higher HE resistance and expected to can be applied to various types of steel including the DPSW.

Statement for Conflict of Interest

Authors are requested to declare any conflicts of interest related to the conduct of this research.

References

• 1)   B.  Marandet: NACE, 5 (1977), 774.
• 2)   T.  Fujita and  T.  Yamada: NACE, 5 (1977), 736.
• 3)   S.  Matsuyama: Okurehakai (Delayed Fracture), Nikkankogyoshinbunsha, Tokyo, (1989), 1.
• 4)   K.  Takai,  J.  Seki and  Y.  Homma: Tetsu-to-Hagané, 81 (1995), 1025 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.81.10_1025
• 5)   J. D.  Embury and  R. M.  Fisher: Acta Metall., 14 (1966), 147. https://doi.org/10.1016/0001-6160(66)90296-3
• 6)   V. K.  Chandhok,  A.  Kasak and  J. P.  Hirth: Trans. Am. Soc. Met., 59 (1966), 288.
• 7)   J.  Tribio and  E.  Ovejero: Mater. Sci. Eng. A, 234–236 (1997), 579. https://doi.org/10.1016/S0921-5093(97)00231-1
• 8)   T.  Tarui and  S.  Yamasaki: Plastos, 1 (2018), 111 (in Japanese). https://doi.org/10.32277/plastos.1.2_111
• 9)   K.  Tomatsu,  T.  Amino,  T.  Chida,  S.  Uji,  M.  Okonogi,  H.  Kawata,  T.  Omura,  N.  Maruyama and  Y.  Nishiyama: ISIJ Int., 58 (2018), 340. https://doi.org/10.2355/isijinternational.ISIJINT-2017-461
• 10)   J.  Toribio and  E.  Ovejero: Corr. Sci., 49 (2007), 3539. https://doi.org/10.1016/j.corsci.2007.03.016
• 11)   J.  Toribio,  D.  Vergara and  M.  Lorenzo: Eng. Frac. Mech., 174 (2017), 243. https://doi.org/10.1016/j.engfracmech.2016.12.020
• 12)   T.  Chida,  M.  Kosaka,  M.  Kubota,  T.  Tarui and  T.  Omura: Tetsu-to-Hagané, 105 (2019), 655 (in Japanese). https://doi.org/10.2355/tetsutohagane.TETSU-2019-002
• 13)   S.-H.  Yu,  H.-B.  Jeong,  J.-S.  Lee and  Y.-K.  Lee: Acta Mater., 225 (2022), 117567. https://doi.org/10.1016/j.actamat.2021.117567
• 14)   T.  Fujita,  Y.  Yamada and  T.  Yamada: Tetsu-to-Hagané, 59 (1973), S255 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.59.4_S255
• 15)   D.  Wan,  Y.  Deng and  A.  Barnoush: Scr. Mater., 151 (2018), 24. https://doi.org/10.1016/j.scriptamat.2018.03.038
• 16)   P.  Tao,  W.  Zhou,  X.  Miao,  J.  Peng and  W.  Liu: Metals, 13 (2023), 1753. https://doi.org/10.3390/met13101753
• 17)   N.  Narita: Scr. Metall., 18 (1984), 985. https://doi.org/10.1016/0036-9748(84)90274-6
• 18)   A.  Kimura and  H. K.  Birnbaum: Scr. Mater., 21 (1987), 53. https://doi.org/10.1016/0036-9748(87)90406-6
• 19)   A.  Massone,  A.  Manhard,  W.  Jacob,  A.  Drexler,  W.  Ecker,  A.  Hohenwarter,  S.  Wurster and  D.  Kiener: Rev. Sci. Instrum., 91 (2020), 043705. https://doi.org/10.1063/1.5142043
• 20)   K.  Tomatsu,  T.  Aoki,  K.  Kobayashi,  T.  Omura and  A.  Hatta: Rev. Sci. Instrum., 94 (2023), 023707. https://doi.org/10.1063/5.0129618
• 21)   T.  Omura,  T.  Chida,  K.  Tomatsu,  T.  Manabe,  T.  Teshima,  K.  Sugita,  M.  Mizuno,  H.  Araki and  Y.  Shirai: Tetsu-to-Hagané, 108 (2022), 224 (in Japanese). https://doi.org/10.2355/tetsutohagane.TETSU-2021-101
• 22)   T.  Tsuru: Zairyo-to-Kankyo, 63 (2014), 3 (in Japanese). https://doi.org/10.3323/jcorr.63.3
• 23)   M.  Sakairi: Zairyo-to-Kankyo, 67 (2018), 191 (in Japanese). https://doi.org/10.3323/jcorr.67.191
• 24)   J.  McBreen,  L.  Nanis and  W.  Beck: J. Electrochem. Soc., 113 (1966), 1218. https://doi.org/10.1149/1.3087209
• 25)   M.  Truschner,  J.  Pengg,  B.  Loder,  H.  Köberl,  P.  Gruber,  M.  Moshtaghi and  G.  Mori: Int. J. Mater. Res., 114 (2023), 439. https://doi.org/10.1515/ijmr-2022-0376
• 26)   K.  Matra,  H.  Furuta and  A.  Hatta: Micromachines, 8 (2017), 211. https://doi.org/10.3390/mi8070211
• 27)   R.  Konno,  T.  Manabe,  D.  Hirakami and  K.  Takai: Tetsu-to-Hagané, 104 (2018), 36 (in Japanese). https://doi.org/10.2355/tetsutohagane.TETSU-2017-036
• 28)   Y.  Moriwaki and  K.  Hayashi: Steel Const. Eng., 2 (1995), 1 (in Japanese). https://doi.org/10.11273/jssc1994.2.8_1
• 29)   R.  Qu,  P.  Zhang and  Z.  Zhang: J. Mater. Sci. Eng., 30 (2014), 599. https://doi.org/10.1016/j.jmst.2014.04.014
• 30)   N.  Uemura,  T.  Chiba,  K.  Saito and  K.  Takai: ISIJ Int., 64 (2024), 678. https://doi.org/10.2355/isijinternational.ISIJINT-2023-351