Visualization of Hydrogen in Stress and Strain Fields Using SIMS

Abstract

In order to clarify the mechanism of hydrogen embrittlement in high strength steel, it is necessary to evaluate local hydrogen concentration in stress and strain fields where hydrogen embrittlement is considered to occur. There are several ways to visualize hydrogen, but SIMS is advantageous in that it can detect hydrogen directly without reaction. In this study, visualization of hydrogen accumulated in stress and strain field was tried using SIMS. Hydrogen flux intensity in the bent portion of the U-bend specimen, where the stress gradient exists, changed according to the stress gradient, and the hydrogen flux intensity was higher on the outside of the bending than on the inside, with the center of the plate thickness at the boundary. On the shearing surface with strain field, the hydrogen flux intensity had a positive correlation with the stress. On the other hand, it remained constant with respect to the strain. From these observations, the hydrogen partitioning behavior in steel could be visualized semi-quantitatively by using the isotope labeling method with SIMS.

1. Introduction

With the aim of improving the fuel consumption and collision safety of vehicles by reducing the body weight, a high strength steel exceeding 1180 MPa is being used extensively in vehicles. Above this strength grade, “delayed fracture (hydrogen embrittlement),¹⁾” apparently caused by hydrogen absorbed by steel while these automotive parts are in use, is involved. Hydrogen embrittlement is caused by three factors: material, environment, and stress. Many studies have been energetically carried out from each viewpoint. Among them, hydrogen, which is an environmental factor, is the smallest element and is difficult to detect and quantify.²⁾

In order to study the hydrogen embrittlement mechanism of high strength steel, it is important to understand hydrogen behavior in steel. Evaluation based on hydrogen concentration has been the mainstream^3,4) in the research on hydrogen embrittlement until now. In particular, the widespread use of TDA (Thermal Desorption Analysis) has revealed the presence of hydrogen caused by precipitates and inclusions in steel, and hydrogen caused by lattice defects such as vacancies and dislocations.^5,6,7,8,9) Based on these experimental results, HELP (Hydrogen-Enhanced Localized Plasticty)¹⁰⁾ and HESIV (Hydrogen-Enhanced Strain-Induced Vacancies)¹¹⁾ have been proposed as hydrogen embrittlement mechanisms related to dislocations and vacancies in materials.

However, the problem of evaluating based on the hydrogen concentration is that the hydrogen analysis represented by the TDA method has to be evaluated for a fixed volume (approximately 2–3 mm³) due to the limitation of the measuring method, for example, the detection sensitivity and the handling of the analysis sample. In order to clarify the local hydrogen behavior in which hydrogen embrittlement actually occurs, it is necessary to carry out a quantitative evaluation in a smaller region (less than 1 μm³). Various methods have been studied to evaluate and observe (visualize) hydrogen distribution and hydrogen embrittlement in steel. Various methods have been proposed to evaluate the local hydrogen content, for example, by calculation using a hydrogen concentration diagram with a stress induced diffusion term added to the diffusion term of Fick’s first law,¹²⁾ local approach method¹³⁾ considering local stress obtained by FEM and Weible analysis and power law¹⁴⁾ etc. Various techniques for observing local hydrogen (visualization) have also been proposed, such as, HMT (Hydrogen Microprint Technique),^14,15,16) Ag-crystaldecoration technique,¹⁷⁾ TAR (Tritium Auto Radiography),¹⁸⁾ SIMS (Secondary Ion Mass Spectrometry),¹⁹⁾ 3D-AP (Three-dimensional Atom Probe).²⁰⁾ These detailed techniques are described in detail in the original book and in the commentary.²¹⁾

Among the local hydrogen observation (visualization) techniques, SIMS is considered to be one of the most suitable techniques for local hydrogen analysis in considering hydrogen embrittlement given the following points: (1) Direct detection of ionic species without chemical reactions associated with hydrogen atoms. (2) The sub-μm resolution required for considering the hydrogen embrittlement mechanism on the assumption of fracture phenomena can be analyzed with high sensitivity in the order of between ppb and ppm. Therefore, by observing hydrogen distribution in steel using SIMS, it may be possible to correctly understand the hydrogen behavior at the point where hydrogen embrittlement actually occurs. As examples of local analysis of hydrogen in steel using conventional SIMS, it has been reported that hydrogen localized around inclusions such as hydrogen accumulated around CaS¹⁹⁾ and hydrogen accumulated around MnS,²⁴⁾ and hydrogen existing in γ phase²⁵⁾ were analyzed. However, when considering actual hydrogen embrittlement, it is necessary to examine the hydrogen behavior in steel in terms of the distribution of the effect of the stress and strain, but there are few examples of visualizing hydrogen distribution at these sites.

In this study, using a model specimen which assumed two places of stress and strain field assumed to be the locus of hydrogen accumulation in steel, visualization of hydrogen existing in these places in steel was tried by combining isotope standard methods using SIMS.

2. Experimental Procedure

2.1. Material

We used low carbon martensitic steel sheet mainly containing C, Si and Mn whose strength level was adjusted to 1500 MPa grade by water quenching from the austenite region and low temperature tempering below 300°C. The microstructure of this steel was confirmed to be a single martensite structure by SEM observation.

2.2. Model Specimen for Hydrogen Visualization

The specimens used to observe the hydrogen distribution in the steel were prepared to evaluate the stress gradient and strain concentration, respectively. As a model specimen of the stress gradient, we used a U-bend specimen²⁶⁾ which can easily generate the stress gradient in the thickness of the steel plate as an observation field in SIMS. Figure 1 shows the manufacturing procedure of the U-bend specimen. A piece of steel of 1.2 mm thick × 34 mm × 6 mm was cut out from a steel sheet, and the thickness was reduced by polishing from both sides until the plate thickness became 0.5 mm. Then, a U-shaped bend was performed using a punch (R = 5 mm) and die (distance between dies = 11.5 mm). Subsequently, M3 bolts and nuts were used to fasten the holes of 22 mm intervals provided in the U-shaped legs, thereby applying stress to the bent portions. The applied stress was controlled by the tightening amount (distance “d” of the inboard leg of the U-bend specimen) of the bolt, and d = 11.5 mm and 10.0 mm.

Fig. 1.

Manufacturing procedure of the U-bend specimen.

The end of steel sheet was cut by shearing, to provide a model specimen of strain concentration. This caused a concentration of plastic deformation.²⁷⁾ Figure 2 shows a schematic diagram of shearing. It was performed on a steel sheet which was 1.2 mm thick, 50 mm width and 12 mm long, with a cutting edge shape of 90 degrees, a plate-holding pressure of 22 kN, and a clearance (The gap between the upper and lower punches divided by the plate thickness) of 10% at a cutting speed of 60 mm/sec. In the observation of hydrogen in steel, the vicinity of the rupture surface on the free end side (an unsheathed surface) was used as an evaluation site.

Fig. 2.

Schematic diagram of shearing.

2.3. Hydrogen Visualization by SIMS

Hydrogen was introduced to the steel as follows: (1) the specimen was embedded in an epoxy resin, (2) the plate-thickness-direction cross section (the observation surface) was polished to a mirror finish by wet-polishing, and (3) hydrogen was introduced using the cathodic electrolysis method with an applied current density of 0.1 × 10^-3 A/cm² in a 3% NaCl D₂O solution containing 30 g/L of NH₄SCN as a poison. By using deuterium (D) instead of hydrogen (H), the background resulting from naturally existing water (H₂O) can be reduced. A sample embedded in an epoxy resin and a platinum counter electrode were arranged, and deuterium was introduced into the steel by energizing the sample with a galvanostat (Hokuto Denko: HA-151A) as a cathode and a platinum counter electrode as an anode in the aforementioned solution at a current density of 0.1 × 10⁻³ A/cm² for 24 hours. After deuterium was introduced into the specimen, the observed surface was immediately wet-polished with colloidal silica and analyzed using SIMS.

Sector-type SIMS (CAMECA IMS-7F) was used for hydrogen analysis. After the surface of the deuterium-introduced specimen was electroconductive coated with Au, the sample was introduced into the analysis chamber which was pre-evacuated for 1 hour. SIMS measurement points (grid of about 1 mm square) were sputtered in advance for cleaning purposes. Negative deuterium ions (²D⁻) were obtained in the scanning ion imaging mode by using Cs⁺ as an ion source and analyzing a 600 μm square region under 15 keV primary ion conditions. Note that the difference in the secondary ion yield, depending on the crystal orientation of the substrate Fe, affects the apparent intensity of ²D⁻.²⁸⁾ In order to clarify the hydrogen existing position, ⁷²FeO⁻ was acquired as the background of Fe and the intensity of ²D⁻ obtained was divided pixel-by-pixel using ⁷²FeO⁻ to normalize the secondary ion image (²D⁻/⁷²FeO⁻). Since ⁵⁶Fe⁻ has low sensitivity under the condition of negative ion detection, ⁷²FeO⁻ with high sensitivity was used. In this paper, deuterium is referred to as hydrogen unless otherwise specified.

2.4. Stress and Strain Analysis by FEM

The stress and strain field generated in each model sample was analyzed by a two-dimensional plane strain model using FEM. Abaqus Explicit was used for all analysis programs. The stress gradient was determined with a tightening amount d = 11.5 mm between the legs in the case of the U-bend specimen, and the strain concentration was determined where the end face of shearing with a clearance of 10%.

3. Results

3.1 Evaluation in Models with Stress Gradients

Figure 3 shows the appearance of the U-bend specimen (d= 11.5 mm) and hydrogen flux intensity (²D−/⁷²FeO⁻ image) using SIMS in the cross section in the thickness direction of the bending apex where bolt tightening quantity d = 11.5 mm. Hydrogen flux intensity was stronger on the outside of bending (upper part of the field of view) than on the inside (lower part of the field of view) with the approximate center of the plate thickness as the boundary. On the outer side of the bent portion, hydrogen flux intensity was stronger in the part which entered inside about 100 μm on than the material surface. Figure 4 shows line analysis of hydrogen distribution in the bent-portions of U-bending specimens (d = 11.5 mm). Line analysis was performed in the SIMS signal, averaging 300 μm wide around the parietal area and assessed according to the direction of the arrows in the figure. Hydrogen flux intensity increased from the outermost layer of the bent portion to the center of the plate thickness, and reached a maximum in the area of 200–250 μm from the outside of the bent to the inside of the bend plate, that is, in the vicinity of the center of the plate thickness. From there, hydrogen flux intensity becomes weaker as it proceeds toward the inside of the bend. The strength ratio between the weakest inner bend and the strongest central part of the sheet thickness was about 1.2 times. The hydrogen flux intensity ratio was determined from the average values of the 100 μm sections (86 points) between 150 and 250 μm and between 350 and 450 μm as the high and low signal sections, respectively.

Fig. 3.

Appearance of the U-bend specimen (d = 11.5 mm) and hydrogen flux intensity (²D⁻/⁷²FeO⁻ image).

Fig. 4.

Line analysis of hydrogen distribution in the bent-portions of U-bend specimen (d = 11.5 mm).

Next, in order to examine the effect of tensile stress generated on the outside of the bending, the same evaluation was carried out for the specimen where the bolt tightening amount d = 10.0 mm. In this specimen, in order to examine the effect of the position of the bent part, the same analysis was carried out at 3 points of the bent top (Point 1), 3 mm from the bent top (Point 2), and 6 mm from the bent top (Point 3). As a reference point for comparing the effects of plastic deformation and tensile stress, the same analysis was also carried out for the portion (Point 4) outside the bolt used for tightening to apply stress. Figure 5 shows the appearance of the U-bend specimen (d = 10.0 mm) and hydrogen flux intensity (²D⁻/⁷²FeO⁻ image) using SIMS at point 1–4. As in the case of d = 11.5 mm, hydrogen flux intensity was strong from the outside of the bend to the center of the sheet thickness in any bent portion (Point 1–3), and hydrogen flux intensity became weaker as the bending proceeded from the center of the sheet thickness to the inside of the bend. In the region where plastic deformation and tensile stress were not applied (Point 4), the hydrogen flux intensity was uniform throughout the plate thickness, and distribution of hydrogen in the steel was not observed. Figure 6 shows the result of line analysis of hydrogen distribution from the outside to the inside of the bend of the point 1–4 with d = 10.0 mm as shown in Fig. 5 (The method of line analysis and the comparison of hydrogen flux intensity were the same technique as in Fig. 4). As in the case of the above-described bolt tightening amount d = 11.5 mm, the strength ratio between the inside of the bend where the hydrogen flux intensity is the weakest and the center of the plate thickness where the hydrogen flux intensity is the strongest was about 1.2 times at any of the points 1–3 of the bent portion. When the average hydrogen flux intensity at the point 4 (in Figs. 5 and 6: The average hydrogen flux intensity was calculated as the average of 86 points at 100 μm sections between 150 and 250 μm.) was compared with the hydrogen flux intensity outside the bending process, the ratio was about 1.4 times.

Fig. 5.

Appearance of the U-bend specimen (d = 10.0 mm) and hydrogen flux intensity (²D⁻/⁷²FeO⁻ image).

Fig. 6.

Line analysis of hydrogen distribution in different sites of U-bend specimen (d = 10.0 mm).

Figure 7 shows the stress and strain distribution at the bent top portion obtained by FEM. The distribution of the tensile stress, which thought considered to affect the hydrogen embrittlement, reached a maximum value in the area about 0.2 mm inward from the outside of the bending (outermost surface), and the tensile stress rapidly decreased as it advanced to the center of the sheet thickness, reaching almost 0 MPa at the center of the plate thickness. From there, a negative value, i.e., compressive stress, was obtained as the bending inner side was advanced, and after the minimum value was obtained at the position of 0.3 mm, the compressive stress became smaller as the bending inner side was advanced. On the other hand, the strain distribution was 0 at the center of the plate thickness, and it became symmetrical with respect to the outside and inside of the bending. Figure 8 shows the stress distribution extracted from the bent top portion of the U-bend specimen.

Fig. 7.

Stress and Strain distribution at the bent top portion (d = 11.5 mm) obtained by FEM.

Fig. 8.

Stress distribution extracted from the center of the U-bend specimen (d = 11.5 mm).

3.2. Evaluation in a Model with Strain Concentration

Figure 9 shows the evaluation results of hydrogen flux intensity on the shearing surface (free end side) cut with a clearance of 10%. There was a region where the hydrogen flux intensity was locally strong in a portion corresponding to the surface of the original sheet on the right side of the field of view, in a shear region about 300 μm away from the end on the right side of the field of view toward the center of the plate thickness, and in the final rupture on the left side of the field of view. The tendency of the average hydrogen flux intensity in the observed field was stronger in the rupture area on the left side of the field than in the shear area on the right side of the field. In the vicinity of the final rupture part, there was a part with strong hydrogen flux intensity according to the deformation (flow associated with cutting) of the microstructure caused by elongation in the final process of cutting. In order to take into account the effect of shearing, the hydrogen flux intensity was evaluated even in the un-deformed part (7.5 mm from the cut end face) sufficiently far from the fracture surface as a reference. In Fig. 9, the ratio of the intensity of the high-signal part (the part to the right of the shearing part) to the low-signal part (the part to the right of the unprocessed part) is about five times, indicating that the signal change in the shearing part is significant.

Fig. 9.

Hydrogen flux intensity on the shearing surface (free end) cut with a clearance of 10%.

Figure 10 shows the results of FEM analysis of stress and equivalent plastic strain on the shearing surface (free end side) cut with a clearance of 10%. The stress was high in the area corresponding to the surface of the base plate on the right side of the field of view and in the vicinity of the final rupture at the end of the left side of the field of view, especially in the area 0.5 mm inward from the surface of the base plate. The stress near the rupture surface was lower than those at these sites. On the other hand, the equivalent plastic strain was concentrated in the area slightly inside the cut surface. The area where the equivalent plastic strain became the highest was near the final rupture.

Fig. 10.

FEM analysis of stress and equivalent plastic strain on the shear plane (free end side) cut with a clearance of 10%.

4. Discussion

4.1. Effect of Stress Field on Hydrogen Distribution

As shown in Figs. 4 and 6, the distribution of hydrogen flux intensity in the U-bend specimen with a stress gradient in the specimen increased from the outside of the bend toward the center of the thickness, and reached a maximum in the region of 0.20–0.25 mm from the surface layer, that is, near the center of the thickness. From there, the hydrogen flux intensity becomes weaker toward the inner side of the plate thickness. Since hydrogen distribution the steel was reported to follow the stress gradient,²⁶⁾ the relationship between the calculated and experimental results was compared for the hydrogen distribution in the bent portion of the U-bend specimen.

Figure 11 shows the calculated results of hydrogen distribution according to the stress distribution obtained from FEM at the bent top portion of the U-bend specimen. The calculation conditions were based on the previous report^12,13,26) and were calculated according to H* = Hexp(σ_mΔV/RT). Where H * is the hydrogen concentration when stress is applied, H is the hydrogen concentration when stress is not applied, σ_m is the hydrostatic stress (Regard the tensile stress as positive), ΔV is 2 × 10⁻⁶ m³/mol for the molar volume change of hydrogen, R is the gas constant, and T is the temperature (Assuming T = 298 K). H was also calculated assuming 1 ppm for simplicity. After calculation, as in the previous report, the hydrogen distribution in the steel followed the stress distribution obtained by FEM shown in Fig. 11, and the hydrogen quantity took the maximum value at the position of 0.15–0.20 mm from the outside of the bending, and then decreased as the bending proceeds. After that, hydrogen content reached a minimum value at the position of 0.32 mm where the stress became minimum, and then increased further toward the inside of the bend.

Fig. 11.

Calculated results of hydrogen distribution according to the stress distribution obtained from FEM.

The results were compared with the hydrogen flux intensities shown in Figs. 4 and 6. The hydrogen flux intensity obtained by SIMS showed a larger decrease than the calculated value on the outside of the bend, but the region where the maximum value was obtained almost agreed. However, the hydrogen flux intensity did not rapidly decrease as much as the calculated value in the area inside the bend, that is, in the area where the tensile stress decreases and turns into the compressive stress, and it became the minimum in the area inside the area where the compressive stress reaches a maximum, and the value became almost constant from there to the inside of the bend. From this result, it is considered we were able to demonstrate that the hydrogen distribution in the steel in the model specimen in which the stress gradient occurs is basically accumulated in the place of high tensile stress as in previous reports. in actual steel does not follow the stress gradient as the conventional calculation result, and the possibility of having the distribution around the stress maximum point was indicated. The hydrogen distribution ratio (Maximum divided by Minimum) corresponding to the stress gradient in the bent section was about 1.2 times that obtained from SIMS of this experiment, but when calculated using the conventional calculation the hydrogen distribution ratio was about 2.3 times, and dissociation occurred. The experimental hydrogen distribution was smaller than the calculated value. It should be noted that the experimental value (Figs. 4 and 6) and calculated value (Fig. 11) were obtained using a different method, so a quantitative comparison between them may not be appropriate. However, the reason for the difference between the two trend is the effect of the increase in dislocations associated with plastic strain. It seems likely that the difference in hydrogen distribution between the inside and outside of the bend was reduced as a result of the increase in the hydrogen concentration, which is the standard, due to the presence of hydrogen in the dislocations, which are the locations where hydrogen is present, due to plastic deformation even on the inside of the bend. This can be explained from the relationship between the hydrogen flux intensity at point 4 (Figs. 5 and 6) of the un-deformed portion. On the other hand, the previous calculation did not take into account the growth of dislocations due to plastic deformation, so it can be assumed that the difference was larger than the experimental result.

When these factors are considered together, it is likely that the hydrogen distribution calculated from the conventional hydrostatic stress and the equation based on Fick’s first law does not properly evaluate the hydrogen enrichment in the stress concentration part, even if the effect of plastic deformation is taken into consideration.

4.2. Effect of Strain Field on Hydrogen Distribution

In order to investigate the effects of stress and strain, the relationship between the results of FEM stress and strain analyses and the strength of hydrogen flux intensity in SIMS was compared as shown in Fig. 10. Stress and strain were taken as the average values in the region of 40 × 40 μm in the FEM analysis, and the corresponding hydrogen flux intensity was evaluated as the average values in the locations shown in Fig. 12 (using the symbols ○, △, □: Each point selected in terms of stress and strain). The accuracy of FEM analysis of extremely large machined parts, such as sheared sections, was reported by Matsuno²⁹⁾ and Fukumura.³⁰⁾ According to these reports, each of the points used in this analysis is several tens of μm or more apart from the shearing surface, and has sufficient accuracy.

Fig. 12.

Comparison of stress/strain and hydrogen flux intensity in shearing specimens.

Figure 13 shows the relationship between the stress and strain in the shearing specimen and the hydrogen flux intensity. As the stress increases, the hydrogen flux intensity increases and there is a positive correlation between the stress and the hydrogen flux intensity. On the other hand, the relationship between the hydrogen flux intensity and the strain shows that the hydrogen flux intensity remains almost constant even when the strain increases. The point (△) where the strength of the hydrogen flux is high near the strain of near 0 (0.03) is the point where the stress is high (1439 MPa) in comparison with other points. Therefore, in the steel subjected to plastic deformation, the actual distribution of hydrogen in the steel was determined to be affected more by stress than by strain.

Fig. 13.

Relationship between the stress and strain in the shearing specimen and hydrogen flux intensity.

Hydrogen existence sites in steel have been reported to be lattice defects such as dislocations and vacancies.^5,6,7,8,9) Shamsujjoha has studied the change of dislocation density with deformation in martensitic steel at low C, and reported that dislocation density saturates at 2% plastic strain.³¹⁾ When dislocations grown with plastic deformation are hydrogen existence sites, the relationship between the plastic strain and the hydrogen flux intensity shown in Fig. 13 coincides with the trend. On the other hand, Takai et al. investigated dislocation density and existence of hydrogen with deformation using pure iron, and reported that the intensity of emission spectrum of hydrogen at TDA increased monotonically up to 20% of plastic strain.³²⁾ This result is different from that shown in Fig. 13, but it appeared to be due to the difference in the initial dislocation quantity present in the material. Generally, the martensite structure has a high dislocation density from the initial stage, so that the amount of dislocations does not increase so much when the martensite structure is processed. On the other hand, since the initial dislocation density of ferrite is low, the amount of dislocation increases when processing is added. As a result, it can be concluded considered that the difference of the structural form affected the relationship between the amount of plastic strain and hydrogen. As for the stress, as previously reported and discussed in Section 4.1, it is clear that the hydrogen flux intensity changed according to the tensile stress because hydrogen accumulated in the place with high tensile stress.

5. Conclusions

• As a result of evaluating the hydrogen flux intensity in the bending part of the U-bend specimen as a model with stress gradient, in accordance with the stress gradient existing in the bending part, the hydrogen flux intensity which is stronger than that in the inside was higher in the outside of the bending with the center of the plate thickness as a boundary. The hydrogen distribution calculated from the conventional hydrostatic stress and the equation based on Fick’s first law may not be able to properly evaluate the hydrogen enrichment in the stress concentration part, even if the effect of strain with plastic deformation is taken into consideration.

• As a result of evaluating the hydrogen flux at the shearing surface as a model with strain concentration, it was found that the strength of hydrogen flux intensity increased as the stress increased, and the correlation was almost positive. On the other hand, the hydrogen flux intensity remained almost constant even when the strain increased. This suggests that the substantial distribution of hydrogen in steel is affected more by stress than by strain.

• Using a highly sensitive hydrogen visualization method that combines secondary ion mass spectrometry and isotope-labeling techniques, we were able to semi-quantitatively visualize the hydrogen distribution behavior between the tensile stress generator and the compressive stress generator, as well as the hydrogen distribution behavior in the region where a large strain distribution occurs.

References

• 1)   S.  Matsuyama: Okurehakai (Delayed Fracture), Nikkankogyoshimbunsha, Tokyo, (1989), 1 (in Japanese).
• 2)   M.  Nagumo: Zairyo-to-Kankyo, 55 (2006), 42 (in Japanese).
• 3)   E.  Akiyama: ISIJ Int., 52 (2012), 307.
• 4)   M. Q.  Wang,  E.  Akiyama and  K.  Tsuzaki: Corros. Sci., 48 (2006), 2189.
• 5)   K.  Takai: Trans. Jpn. Soc. Mech. Eng. A, 70 (2004), 1027.
• 6)   K.  Takai and  A.  Nozue: J. Jpn. Inst. Met., 64 (2000), 669.
• 7)   K.  Takai,  J.  Seki and  Y.  Homma: Tetsu-to-Hagené, 81 (1995), 1025 (in Japanese).
• 8)   K.  Takai,  G.  Yamauchi,  M.  Nakamura and  M.  Nagumo: J. Jpn. Inst. Met., 62 (1998), 267.
• 9)   K.  Takai: Zairyo-to-Kankyo, 60 (2011), 230 (in Japanese).
• 10)   M.  Nagumo: Zairyo-to-Kankyo, 56 (2007), 343 (in Japanese).
• 11)   M.  Nagumo: Mater. Sci. Technol., 20 (2004), 940.
• 12)   A.  Nozue: Okurehakai-Kaimei-no-Shintenkai (Advances in Delayed Fracture Solution), ISIJ, Tokyo, (1997), 197 (in Japanese).
• 13)   S.  Takagi,  T.  Inoue,  T.  Hara,  M.  Hayakawa,  K.  Tsuzaki and  T.  Takahashi: Tetsu-to-Hagané, 86 (2000), 689 (in Japanese).
• 14)   T. E.  Perez and  J.  Ovejero-Garcia: Scr. Metall., 16 (1982), 161.
• 15)   J.  Ovejero-Garcia: J. Mater. Sci., 20 (1985), 2623.
• 16)   A.  Nagao,  S.  Kuramoto,  M.  Kanno and  T.  Shiraga: Tetsu-to-Hagané, 86 (2000), 24 (in Japanese).
• 17)   T.  Schober and  C.  Dieker: Metall. Trans. A, 14 (1983), 2440.
• 18)   J. P.  Laurent and  G.  Lapasset: Int. J. Appl. Radiat. Isot., 24 (1973), 213.
• 19)   K.  Takai,  J.  Seki and  Y.  Homma: Mater. Trans., JIM, 36 (1995), 1134.
• 20)   J.  Takahashi,  K.  Kawakami,  Y.  Kobayashi and  T.  Tarui: Scr. Mater., 63 (2010), 261.
• 21)   E.  Akiyama: Bull. Iron Steel Inst. Jpn., 21 (2016), 630.
• 22)   M.  Koyama,  M.  Rohwerder,  C. C.  Tasan,  A.  Bashir,  E.  Akiyama,  K.  Takai,  D.  Raabe and  K.  Tsuzaki: Mater. Sci. Technol., 33 (2017), 1481.
• 23)   K.  Takai: Zairyo-to-Kankyo, 49 (2000), 271 (in Japanese).
• 24)   F.  Yuse,  T.  Nakayama and  T.  Tuchida: CAMP-ISIJ, 11 (1998), 1239.
• 25)   R.  Silverstein,  O.  Sobol,  T.  Boellinghaus,  W.  Unger and  D.  Eliezer: J. Alloy Compd., 695 (2017), 2689.
• 26)   Y.  Toji,  S.  Takagi,  M.  Yoshino,  K.  Hasegawa and  Y.  Tanaka: Tetsu-to-Hagané, 95 (2009), 887 (in Japanese).
• 27)   M.  Usuda: The 96th Plastic Processing Symp., “Evaluation method in sheet metal forming,” The Japan Society for Technology of Plasticity, Tokyo, (1985), 25.
• 28)   S.  Yabu,  M.  Kawamori,  J.  Kinugasa and  F.  Yuse: Proc. JSSE Conf. on Springs and Restitution Effect, Japan Society of Spring Engineers, Tokyo, (2016), 13.
• 29)   T.  Matsuno,  Y.  Kuriyama,  S.  Yonezawa,  H.  Murakami and  N.  Ohtake: J. Jpn. Soc. Technol. Plast., 53 (2012), 836.
• 30)   M.  Fukumura,  Y.  Zaizen,  T.  Omura,  K.  Senda and  Y.  Oda: J. Jpn. Soc. Technol. Plast., 59 (2018), 65.
• 31)   M.  Shamsujjoha: Mater. Sci. Eng. A, 776 (2020), 139039.
• 32)   K.  Takai: Zairyo-to-Kankyo, 60 (2011), 230 (in Japanese).