Effect of Hydrogen on Deformation Behavior of Carbon Steel S25C —Measurement of Time Evolution of Strain Distribution until Crack Initiation Using Digital Image Correlation Method—

Abstract

To explain hydrogen embrittlement, it is important to understand the effect of hydrogen on the plastic deformation of materials. In this study, we measured the plastic deformation process until crack initiation in hydrogen-charged and hydrogen-uncharged carbon steel S25C using the digital image correlation method. As a result, we found that the equivalent strain at crack initiation decreased at the stress-concentrated regions owing to the presence of hydrogen, whereas the size of regions with a high equivalent strain rate increased at an earlier stage. Comparing the equivalent strain rate in regions with roughly the same equivalent strain, we found that there is no significant difference between hydrogen-charged and hydrogen-uncharged specimens for small equivalent strain; however, the equivalent strain rate increased rapidly for large equivalent strain in hydrogen-charged specimens.

1. Introduction

Despite the accumulated and ongoing research, hydrogen embrittlement remains a topic of debate. One reason put forward is the complex interaction of hydrogen with lattice defects. Typical mechanisms for hydrogen embrittlement include: 1) the hydrogen-enhanced localized plasticity (HELP),^1,2,3) which assumes that hydrogen increases the mobility of dislocations, and thus locally facilitates the formation of ductile fractures; 2) the hydrogen-enhanced strain-induced vacancy model (HESIV),^4,5) which assumes that hydrogen promotes the generation and cohesion of atomic vacancies accompanying plastic deformation, facilitating ductile failure; and 3) hydrogen-enhanced decohesion (HEDE),^6,7,8) which assumes that cracking occurs at grain boundaries owing to the accumulated hydrogen.

Using atomistic simulations, we have been studying the interaction between hydrogen and individual lattice defects in alpha iron.^{9,10,11,12,13,14,15,16)} We demonstrated that dislocation mobility varies because of hydrogen,^9,10,11) vacancy multiplication occurs owing to plastic deformation,¹²⁾ and cohesive energy decreases because of the hydrogen atoms trapped at grain boundaries.¹³⁾ Similar phenomena have also been reported via microscopic experiments.^17,18) These studies can serve as steps for all the abovementioned mechanisms (HELP, HESIV, HEDE). Thus, it is important to have discussions based on the premise that a lot of hydrogen-affected lattice defects exist and interact at macroscopic scales and that there should be competition between different mechanisms. Furthermore, because the density of each type of lattice defect changes depending on the amount of deformation and deformation conditions, principal hydrogen effect may change during deformation.

In this study, we experimentally investigated in detail how hydrogen effects appear during plastic deformation of carbon steel S25C. First, while conducting slow strain-rate testing (SSRT) of hydrogen-charged and hydrogen-uncharged specimens, we isochronally photographed the regions around the notch introduced at the center section of the specimens. By applying the digital image correlation method (DICM)^19,20,21) to the corrected images, we visualized the deformation field, and then discussed the principal effect of hydrogen on plastic deformation considering the accumulated equivalent strain and the process of accumulation.

2. Experimental

2.1. Specimen

A rod of carbon steel S25C that consists of ferrite and pearlite was machined into a 0.5-mm thick specimen, with the shape as shown in Fig. 1, using wire electrical discharge machining. The specification of the chemical composition of S25C is given in Table 1. The direction of the thickness in the specimen and longer direction of the rod coincided. To limit the origination point of plastic deformation, a narrow notch was introduced at the center section using 0.2 mm wire. The width of introduced notch was 0.26 ± 0.02 mm owing to machining accuracy. We confirmed that the influence of the notch-width difference and the surface asperity on the results is small.

Fig. 1.

Geometry of the specimen.

Table 1. Material composition of the carbon steel S25C.

	C	Si	Mn	P	S
S25C	0.22–0.28	0.15–0.35	0.30–0.60	< 0.030	< 0.035

The specimen was annealed in Ar atmosphere at 900°C for 10 min, and then cooled slowly back to room temperature in the furnace. After heat treatment, both surfaces of the sheet specimen were polished with emery paper (#240, #400, #800, #1500, #2000) and then buffed (diamond paste: 3 μm and 1 μm, and alumina: 0.3 μm and 0.05 μm).

2.2. Hydrogen Charging by Acid Immersion

Hydrogen charging for 24 h and 48 h was conducted at 30°C using a FIP (Federation International de la Precontrainte)²²⁾ immersion bath (20% by mass of NH₄SCN). Taking into consideration the deterioration of the FIP bath solution in 48 h, the solution was replaced after 24 h. After hydrogen charging, the specimen was placed in a tensile tester and the test was started within 5 min.

2.3. Slow Strain Rate Testing (SSRT)

SSRT was performed using a small-scale, custom-made, Imoto tensile testing machine mounted on an active vibration isolation table. The specimen was fixed using the holes made on both sides and tensile load was applied. The tensile rate was 0.03 mm/min and the nominal strain rate for the gauge section of the specimen was ε_x = 6.25 × 10^–5 s^–1. During the tensile test, the vicinity of the notch bottom was photographed at 15-s intervals with a digital microscope (Keyence Digital Microscope VHX-900) at a magnification of 300× (0.63 μm/pixel, image resolution 1600 × 1200 pixels).

2.4. Digital Image Correlation Method (DICM)

The application of the DICM to the collected images during SSRT allows visualizing the deformation field near the notch bottom. The correlation function used to search for equivalent points is the same as in Ref. 21). In the paper, we verified that our system can measure the deformation of S25C with high accuracy.

Neglecting the deformation in the thickness direction and out of the plane, the equivalent strain increment between two successive images is calculated by the following equation.   

$$\mathrm{\Delta }{\epsilon }_{\text{eq}}^{}=\frac{2}{3}\sqrt{\mathrm{\Delta }{\epsilon }_x^2+\mathrm{\Delta }{\epsilon }_y^2-\mathrm{\Delta }{\epsilon }_x^{}\mathrm{\Delta }{\epsilon }_y^{}+\frac{3}{4}\mathrm{\Delta }{\gamma }_{xy}^2}$$(1)

The equivalent strain was evaluated by taking the summation of equivalent strain increments in a coordinate system that considers the movement of the material points, i.e., material coordinate system.²¹⁾ The size of the subset used in the DICM was 35 × 35 pixels.

To measure the strain using DICM, an appropriate pattern must be formed on the material surface for pattern matching. The pattern formed on the surface owing to corrosion by acid immersion (Fig. 2(a)) was used to measure the deformation of the charged material. On the other hand, because no appropriate pattern was formed on the surface of the uncharged material after polishing, a surface pattern was etched using a nital solution (5% HNO₃, 95% C₂H₅OH) before performing SSRT (Fig. 2(b)). Thus, because a pattern was imprinted on both the charged and uncharged specimens at the same scale, the deformation should be measured with roughly the same accuracy. The hydrogen content in the uncharged specimens following etching was measured using thermal desorption analysis (TDA), described in the next section, and no hydrogen penetration from etching was found.

Fig. 2.

Surface pattern around the notch bottom before SSRT.

2.5. Hydrogen Thermal Desorption Analysis (TDA)

The hydrogen content of the specimen was established by TDA, which uses a gas chromatograph (Shimadzu Gas Chromatograph GC-2014) equipped with the thermal conductivity detector for hydrogen detection. The sample was heated in a sample tube from 25°C to 350°C at 5°C/min. The released hydrogen was stored in the sample tube, and sent to the detector at 5-min intervals. Ar was used as a carrier gas.

The hydrogen content was measured immediately before starting SSRT. Therefore, an approximately 2-mm wide fragment to be used in TDA was cut from one end of each specimen before hydrogen charging, and the fragment was immersed with the specimen for SSRT in the same FIP bath. TDA was started within 5 min after charging. The specimen for SSRT was left in the FIP bath during TDA. The additional hydrogen charged during this period was negligible.

2.6. Experimental Procedure

The experimental procedure is summarized in Fig. 3. The hydrogen-uncharged specimens are designated as No_charged, and the 24- and 48-h hydrogen-charged specimens are referred to as Charged_24 h and Charged_48 h, respectively. The experiments were repeated fifth for the no-charged specimens and twice for the 24- and 48-h hydrogen-charged specimens each and similar results were obtained; thus, only representative results are given in this paper.

Fig. 3.

Experimental procedure.

3. Experimental Results and Discussion

3.1. Hydrogen Content

The hydrogen that diffuses at relatively low temperature has a strong effect on hydrogen embrittlement.^23,24) In this study, the hydrogen released at ≤200°C is defined as diffusible hydrogen, and the hydrogen content of the specimen is represented by this amount.

The hydrogen desorption spectra from each individual specimen by TDA are shown in Fig. 4. We confirmed that only hydrogen from the charged specimens was detected as being concentrated around 100°C and stops being released at 200°C. No release of hydrogen in this temperature range was found for the uncharged material. The diffusible hydrogen concentration was 0.01 mass ppm for No_charged, 0.66 mass ppm for Charged_24 h, and 0.75 mass ppm for Charged_48 h. Comparing the hydrogen desorption spectra between Charged_24 h and Charged_48 h, a major difference is confirmed in hydrogen content which was released around 75°C; Charged_48 h contains about 0.08 mass ppm more hydrogen with high diffusibility than Charged_24 h.

Fig. 4.

Hydrogen desorption spectra. (Online version in color.)

3.2. Stress–displacement Relation

The nominal stress curves are shown in Fig. 5. The horizontal and vertical axes are the crosshead displacement and nominal stress at the notched section, respectively. The shift in the start point and gradual increase of the stress immediately after the experiment starts (shaded region in the figure) are attributed to the fixing method of the specimen. Here we define the upper yield point on the nominal stress curve as a yielding point, and focus on the deformation behavior at plastic range. a, b, and c in the figure indicate yielding point for No_charged, Charged_24 h, and Charged_48 h, respectively. At each point, (A–C) denoted by the arrows for No_charged, Charged_24 h, and Charged_48 h, cracks were generated at the notch bottom and started to grow immediately afterward. We call these points as crack initiation points. The nominal stress was 420 MPa for No_charged, 405 MPa for Charged_24 h, and 377 MPa for Charged_48 h. The nominal stress at crack initiation point decreased with increasing diffusible hydrogen. On the other hand, large difference in yielding stress and work-hardening behavior until crack initiation point could not be observed.

Fig. 5.

Nominal stress – crosshead displacement relation: A, B, and C indicate crack initiation points for No_charged, Charged_24 h, and Charged_48 h, respectively. Similarly, a, b, and c indicate yielding points. (Online version in color.)

3.3. Distribution of Equivalent Strain and Equivalent Strain Rate

3.3.1. Equivalent Strain Distribution at Crack Initiation

Figure 6 shows the distribution of accumulated equivalent strain near the notch bottom from the start of the deformation and up to the crack initiation point. To show the correspondence of strain distribution with texture, the strain distribution maps were made semi-transparent, and they were superimposed on the optical microscope images. For uncharged specimens, a maximum equivalent strain of roughly 1.6 had accumulated at the notch bottom. On the other hand, maximum equivalent strains of about 1.2 and 1.0 had accumulated for Charged_24 h and Charged_48 h, respectively. The cracks initiated from these maximum strain regions. Under the measurement resolution of our experiments, we observed that for charged specimens, the equivalent strain when cracks initiated decreased. We also found that the higher the diffusible hydrogen content, the smaller the equivalent strain at crack initiation.

Fig. 6.

Distribution of equivalent strain around the notch bottom (immediately before crack initiation).

As to whether ferrite or pearlite phase is affected by hydrogen, we need to study using higher magnification. Although we observed a tendency for the deformation to localize in the ferrite phase, this was not clear.

3.3.2. Comparison of the Distribution of Equivalent Strain Rate during Deformation

To evaluate the effect of hydrogen in the deformation process, we compared the distribution of the equivalent strain rate ( ${\epsilon }_{\text{eq}}^{\cdot }$ = Δε_eq/Δt, Δt = 90 s). The image immediately following the yield point was called the yield point image and we labeled it the 0th image. The distribution of equivalent strain rate calculated from the 0th to the 6th image and the equivalent strain rate distribution calculated from the 24th to the 30th image are shown in Figs. 7 and 8, respectively. These two divisions correspond to the deformation immediately after yielding and prior to the crack initiation for Charged_48 h specimen (point C in Fig. 5), respectively. The elongation in these periods is 0.045 mm.

Fig. 7.

Distribution of equivalent strain rate immediately after yielding (Image No. 0–No. 6).

Fig. 8.

Distribution of equivalent strain rate prior to the crack initiation for Charged_48 h specimen (Image No. 24–No. 30).

No considerable difference in the distribution of equivalent strain rate immediately after yielding is observed as shown in Figs. 7(a), 7(b), and 7(c). However, when comparing Figs. 8(a), 8(b), and 8(c), we find that the region of the largest strain rate has expanded for Charged_48 h specimen. This suggests that the effect of hydrogen is considered as localization of the deformation as deformation progresses. In the following sections, we concentrate on the evolution of regions that exhibit high strain rates.

3.4. High Equivalent Strain Rate Regions

We define HS region as a high strain-rate area where the equivalent strain rate is ≥1.1 × 10^–3 s^–1 (corresponding to roughly 20 times the equivalent strain rate loaded onto the specimen). The evaluation results for the temporal changes in the HS region until crack initiations are shown in Fig. 9. From this figure, we can observe that the HS region increases as the deformation progresses in all specimens. Although there is little difference in the behavior immediately after yielding, the results for charged specimens start to differ relative to uncharged specimens, and area of the HS region increases very rapidly when elongation after yielding exceeds approximately 0.15 mm. Because the experiments were performed with the same tensile rate for all specimens, the early-stage increase in the HS region suggests that the deformation progresses locally. Moreover, another noteworthy characteristic is that the areas of HS region at the last state are 2.20 × 10⁵ μm² for the No_charged, 1.76 × 10⁵ μm² for the Charged_24 h, and 1.47 × 10⁵ μm² for the Charged_48 specimens, which verified the reduction in the area of HS region at the crack initiation for charged specimens.

Fig. 9.

Change of high strain-rate (HS) region. (Online version in color.)

3.5. Evolution of Equivalent Strain Rate at Local Point

In this section, we concentrate on the relation between the equivalent strain and equivalent strain rate at local material point. As shown in Figs. 6, 7, and 8, the temporal variation of the equivalent strain and equivalent strain rate were found for an arbitrary point within the evaluation region. We picked up all measured points with equivalent strain over a constant range (ε_eq ± 0.05) from all the deformation data from yielding to the crack initiation point, and then $\overline{{\epsilon }_{\text{eq}}^{\cdot }}$ was evaluated as the average of equivalent strain rate at those points. The equivalent strain rate at each point was evaluated using Δt = 15 s.

The relationship between equivalent strain and $\overline{{\epsilon }_{\text{eq}}^{\cdot }}$ is shown in Fig. 10. We can observe that in uncharged specimens, the increase in the deformation rate accompanying the progress of deformation is approximately constant. On the other hand, though there is no noticeable difference for small equivalent strain (<0.8 for Charged_24 h and <0.2 for Charged_48 h), when the deformation progresses in charged specimens, the deformation rate corresponding to the same state of deformation (the amount of equivalent strain) increases very rapidly. For example, comparing the average equivalent strain rates for the region with a total accumulated strain of 0.9, the rates for the Charged_24 h and Charged_48 h specimens are 1.3 and 2.0 times greater than that for the No_charged specimen, respectively. It is to be noted that the difference of diffusible hydrogen concentration between Charged_24 h and Charged_48 h specimens is about 0.1 mass ppm. This small difference of hydrogen concentration made the large difference of the localized equivalent strain rate. The hydrogen distribution caused by, for example, hydrostatic stress effect^25,26,27) probably intensifies the localized deformations.

Fig. 10.

Relationship between local equivalent strain and averaged equivalent strain rate. (Online version in color.)

4. Conclusions

We conducted SSRT for three different types of carbon steel S25C specimens (No_charged, Charged_24 h, Charged_48 h). Then, we measured the plastic deformation using DICM, and we were able to investigate the effect of hydrogen on plastic deformation. Our findings can be summarized as follows:

(1) For hydrogen-charged specimens, the equivalent strain when cracks initiated decreased. Furthermore, we found that the higher the diffusible hydrogen content, the smaller the equivalent strain at crack initiation.

(2) For hydrogen-charged specimens, the area of the region of large equivalent strain rate increases more rapidly than that of uncharged specimens. Because the tensile rate is the same for all specimens, it is reasonable to argue that the plastic deformation becomes localized more rapidly because of hydrogen. This difference occurs long before differences in the macro response, such as that observed in nominal stress curves.

(3) If we compare the equivalent strain rates for regions in which the local equivalent strain is roughly the same, we find that there is no significant difference between the hydrogen-charged and hydrogen–uncharged specimens for small equivalent strain; however, the equivalent strain rate increases very rapidly in hydrogen-charged specimens for increasing equivalent strain. Therefore, as the deformation progresses in the stress-concentrated regions, the effect of hydrogen is manifested as an increase in the strain rate. The enhanced localized plastic deformation by hydrogen was clearly observed through the straightforward measurement of deformation fields.

There is no doubt that hydrogen induces localized plastic deformation; however, it is not at all clear why localized plastic deformation occurs. We understand that there are changes in the stability and mobility of the various lattice defects owing to hydrogen as mentioned in the Introduction. Here, we revealed that the effect of hydrogen became apparent where the equivalent strain accumulated, and the equivalent strain at the crack initiation decreased by hydrogen. We need to consider complex mechanisms that rely not wholly on the dislocation mobility.

Acknowledgments

This study was partially supported by Grant-in-Aid for Young Scientists (A), 23686022, by the Ministry of Education, Culture, Sports, Science and Technology, Japan.

References

• 1)   C. D.  Beachem: Metall. Trans., 3 (1972), 441.
• 2)   P. J.  Ferreira,  I. M.  Robertson and  H. K.  Birnbaum: Acta Mater., 47 (1999), 2991.
• 3)   S.  Matsuoka,  N.  Honma,  H.  Tanaka,  Y.  Fukushima and  Y.  Murakami: J. Jpn. Inst. Met., 70 (2006), 1002 (in Japanese).
• 4)   M.  Nagumo: Mater. Sci. Technol., 20 (2004), 940.
• 5)   K.  Takai,  H.  Shoda,  H.  Suzuki and  M.  Nagumo: Acta Mater., 56 (2008), 5158.
• 6)   S.  Matsuyama: Tetsu-to-Hagané, 80 (1994), 679 (in Japanese).
• 7)   S.  Hinotani,  F.  Terasaki and  K.  Takahashi: Zairyo-to-Kankyo, 64 (1978), 899 (in Japanese).
• 8)   X.-Y.  Liu,  J.  Kameda,  J. W.  Anderegg,  S.  Takaki,  K.  Abiko and  C. J.  McMahon: Mater. Sci. Eng. A, 492 (2008), 218.
• 9)   S.  Taketomi,  R.  Matsumoto and  N.  Miyazaki: J. Mater. Sci., 43 (2008), 1166.
• 10)   S.  Taketomi,  R.  Matsumoto and  N.  Miyazaki: Acta Mater., 56 (2008), 3761.
• 11)   S.  Taketomi,  R.  Matsumoto and  N.  Miyazaki: J. Mater. Res., 26 (2011), 1269.
• 12)   R.  Matsumoto,  N.  Nishiguchi,  S.  Taketomi and  N.  Miyazaki: J. Soc. Mater. Sci. Jpn., 63 (2014), 182 (in Japanese).
• 13)   M.  Riku,  R.  Matsumoto,  S.  Taketomi and  N.  Miyazaki: J. Soc. Mater. Sci. Jpn., 59 (2010), 589 (in Japanese).
• 14)   S.  Seki,  R.  Matsumoto,  Y.  Inoue,  S.  Taketomi and  N.  Miyazaki: J. Soc. Mater. Sci. Jpn., 61 (2012), 175 (in Japanese).
• 15)   R.  Matsumoto,  Y.  Inoue,  S.  Taketomi and  N.  Miyazaki: Scr. Mater., 60 (2009), 555.
• 16)   R.  Matsumoto,  S.  Taketomi,  S.  Matsumoto and  N.  Miyazaki: Int. J. Hydrogen Energ., 34 (2009), 9576.
• 17)   T.  Tabata: Scr. Metall., 17 (1983), 947.
• 18)   A.  Barnoush and  H.  Vehoff: Acta Mater., 58 (2010), 5274.
• 19)   M. A.  Sutton: Image Correlation for Shape, Motion and Deformation Measurements, Springer, New York, (2009).
• 20)   N.  Shishido,  T.  Ikeda,  N.  Miyazaki,  K.  Nakamura,  N.  Miyazaki and  T.  Sawatari: J. Soc. Mater. Sci. Jpn., 57 (2008), 83 (in Japanese).
• 21)   R.  Matsumoto,  M.  Kubota and  N.  Miyazaki: Exp. Tech., (2014). in press, DOI: 10.1111/ext.12040
• 22)  Technical Report, Reports on Prestressing Steel: 5. Stress Corrosion Cracking Resistance Test for Prestressing Tendons, Fédération Internationale de la Précontrainte, Slough, (1980).
• 23)   K.  Takai and  R.  Watanuki: ISIJ Int., 43 (2003), 520.
• 24)   S.  Suzuki,  N.  Ishii,  T.  Miyagawa and  H.  Harada: Tetsu-to-Hagané, 79 (1993), 227.
• 25)   P.  Sofronis and  R. M.  McMeeking: J. Mech. Phys. Solids., 37 (1989), 317.
• 26)   H.  Kotake,  R.  Matsumoto,  S.  Taketomi and  N.  Miyazaki: Int. J. Pres. Ves. Pip., 85 (2008), 540.
• 27)   K.  Takatama,  R.  Matsumoto,  S.  Taketomi and  N.  Miyazaki: Int. J. Hydrogen Energ., 36 (2011), 1037.