JOM

DOI: 10.1007/s11837-017-2371-1 Ó 2017 The Minerals, Metals & Materials Society

The Role of Hydrogen-Enhanced Strain-Induced Lattice Defects on Hydrogen Embrittlement Susceptibility of X80 Pipeline Steel M. HATTORI,1 H. SUZUKI,2 Y. SEKO,3 and K. TAKAI

2,4

1.—Graduate School of Science and Technology, Sophia University, 7-1 Kioi-cho, Chiyoda-ku, Tokyo 102-8554, Japan. 2.—Department of Engineering and Applied Sciences, Faculty of Science and Technology, Sophia University, 7-1 Kioi-cho, Chiyoda-ku, Tokyo 102-8554, Japan. 3.—Earthquake Engineering and Fracture Mechanics Laboratory, Fundamental Technology Research Institute, Tokyo Gas Co., Ltd., Yokohama 230-0045, Japan. 4.—e-mail: [email protected]

Studies to date have not completely determined the factors influencing hydrogen embrittlement of ferrite/bainite X80 pipeline steel. Hydrogen embrittlement susceptibility was evaluated based on fracture strain in tensile testing. We conducted a thermal desorption analysis to measure the amount of tracer hydrogen corresponding to that of lattice defects. Hydrogen embrittlement susceptibility and the amount of tracer hydrogen significantly increased with decreasing crosshead speed. Additionally, a significant increase in the formation of hydrogen-enhanced strain-induced lattice defects was observed immediately before the final fracture. In contrast to hydrogen-free specimens, the fracture surface of the hydrogen-charged specimens exhibited shallower dimples without nuclei, such as secondary phase particles. These findings indicate that the presence of hydrogen enhanced the formation of lattice defects, particularly just prior to the occurrence of final fracture. This in turn enhanced the formation of shallower dimples, thereby potentially causing premature fracture of X80 pipeline steel at lower crosshead speeds.

INTRODUCTION The use of hydrogen in multiple industrial applications necessitates the establishment of substructures to produce, transport, and store hydrogen. Pipelines are primarily used to transport large amounts of hydrogen at reasonable costs. However, due to the possibility of hydrogen embrittlement (HE), the safety of these pipelines must be maintained even during large earthquakes that can induce a wide range of strain rates on the pipelines. HE and stress corrosion cracking caused by cathodic corrosion protection systems have been reported for natural gas and petroleum pipelines.1–4 Previous studies have shown that the interaction between hydrogen and dislocations in pure iron,5 tempered martensitic steel6 and cold-drawn pearlitic steel7 leads to the formation of hydrogenenhanced strain-induced lattice defects during plastic deformation. Neeraj et al.8 concluded that the accumulation of hydrogen-stabilized vacancies and coalescence of nanovoids in X65 and X80 steels plays a prominent role in failure pathways for HE.

However, the role of hydrogen-enhanced straininduced lattice defects on the HE susceptibility of X80 pipeline steel has not yet been clarified. Therefore, the present study involved the preparation of X80 pipeline steel specimens with a ferrite/ bainite microstructure to investigate the role of hydrogen and hydrogen-enhanced strain-induced lattice defects, produced during plastic deformation, in the HE process. EXPERIMENTAL Dependence of HE Susceptibility on Crosshead Speed In the present study, X80 pipeline steel with a ferrite-bainite microstructure was used. Tensile test specimens with gage dimensions of 20 9 10 9 1.0 mm were machined from the actual pipelines. They were electrochemically precharged with hydrogen in an aqueous solution of 0.4 mass% NaOH and 1.0 mass% NH4SCN with a current density of 5 A m 2 at 30°C for 72 h. Furthermore, hydrogen of 0.75 mass ppm was charged into the

Hattori, Suzuki, Seko, and Takai

Fig. 1. Stress–strain curves for specimens with/without hydrogen at different crosshead speeds (a). Relationship between relative failure strain and crosshead speed (b).

Fig. 2. SEM images of fractured surfaces of (a), (b) hydrogen-free specimens and (c), (d) hydrogen charged specimens at a crosshead speed of (a), (c) 0.005 mm min 1 and (b), (d) 10 mm min 1.

specimens prior to tensile testing. The tensile tests were carried out at crosshead speeds ranging from 0.005 mm min 1 to 100 mm min 1 at 30°C, with/ without hydrogen precharging. The HE susceptibility was evaluated based on the relative fracture strain (i.e., the fracture strain with hydrogen divided by the fracture strain without hydrogen). Hydrogen charging was conducted concurrently with the tensile tests after precharging in order to keep constant hydrogen content. For the hydrogenfree specimens, the tensile tests were conducted in air at 30°C. The surfaces of the fractured specimens obtained in the tensile tests were observed using optical microscopy and scanning electron microscopy (SEM). In order to evaluate the characteristics of the fracture surface, the ratio of the number of dimples with a nucleus to the total number of dimples was calculated from three fields observed by SEM. Lateral surfaces near the fracture surfaces were also observed via SEM.

Formation of Lattice Defects Enhanced by Straining With/Without Hydrogen The formation of hydrogen-enhanced strain-induced lattice defects was investigated by first deforming the specimens to a plastic strain of 8% at crosshead speeds ranging from 0.005 mm min 1 to 100 mm min 1, with/without hydrogen precharging. The specimens were subsequently kept in a thermostatic chamber in air at 30°C for 24 h to desorb the majority of diffusible hydrogen. This was followed by charging the specimens again with hydrogen (as a tracer for lattice defects) by immersing them in a 20 mass% NH4SCN solution at 30°C for 24 h. A thermal desorption analysis (TDA) was used to measure the tracer hydrogen content, corresponding to the amount of lattice defects. The TDA was performed at temperatures from room temperature to 200°C at a heating rate of 5°C min 1. Moreover, an in-depth examination of the lattice defect formation process was performed

The Role of Hydrogen-Enhanced Strain-Induced Lattice Defects on Hydrogen Embrittlement Susceptibility of X80 Pipeline Steel

by measuring the tracer hydrogen content for specimens subjected to various amount of strain up to failure at a crosshead speed of 0.01 mm min 1. RESULTS AND DISCUSSION Dependence of Hydrogen Embrittlement Susceptibility on Crosshead Speed Figure 1a shows stress–strain curves for specimens strained with/without hydrogen at different crosshead speeds. For the hydrogen-free specimens, only the result for a crosshead speed of 0.1 mm min 1 is shown. For the specimens strained in the presence of hydrogen, the results indicate that the hydrogen had no effect on the yield stress or the ultimate tensile stress. Both the uniform and local elongation to failure strongly decrease with decreasing crosshead speed. Figure 1b shows the change in the relative fracture strain with

Fig. 3. Relationship between ratio of number of dimples with nuclei to total number of dimples and crosshead speed for hydrogencharged and hydrogen-free specimens.

crosshead speed. As can be seen, as the crosshead speed is reduced, the relative fracture strain significantly decreases, resulting in increased HE susceptibility. These results are attributed to the interaction between hydrogen atoms and dislocations, since previous studies suggested that hydrogen atoms are dragged by dislocations at low crosshead speeds.9,10 Figure 2 shows SEM images of fractured surfaces for (a), (b) hydrogen-free specimens and (c), (d) hydrogen-charged specimens at a crosshead speed of (a), (c) 0.005 mm min 1 and (b), (d) 10 mm min 1. All specimens revealed features indicating ductile fracture, such as dimples. Furthermore, in Fig. 2a and b, nuclei (such as secondary phase particles) or traces appear over the entire region. These results are reasonably consistent with those obtained in previous studies, which found that ductile fractures occur as a result of the generation, growth, and coalescence of voids nucleated around inclusions or secondary phase particles.11 Such dimples with nuclei are also seen in Fig. 2d. In contrast, in Fig. 2c, shallow flat dimples with no nuclei appear. As mentioned previously, ductile fracture is a consequence of the coalescence of voids. Hence, there is a possibility of the presence of alternative nuclei that are separate from the secondary phase particles. As for the fact that fracture surfaces show no quasi-cleavage or intergranular fracture, the X80 specimens contain a smaller amount of hydrogen, about 0.75 ppm, than specimens previously reported in the literature12 and show no surface damage, such as corrosion pits due to the NaOH solution used for hydrogen charging. By setting the conditions strictly and increasing the amount of hydrogen, obvious brittle areas like quasi-cleavage or intergranular fracture may be observed.13 Figure 3 shows the relationship between the ratio of the number of dimples with secondary phase nuclei or their vestiges to the total number of dimples and the crosshead speed for specimens with/without hydrogen. This was calculated from

Fig. 4. SEM images of lateral surfaces of fractured hydrogen-charged specimens subjected to strain at a crosshead speed of (a1), (a2) 0.005 mm min 1 and (b1), (b2) 10 mm min 1.

Hattori, Suzuki, Seko, and Takai

Fig. 5. Hydrogen desorption profiles for prestrained specimens with/without hydrogen subjected to strain at a crosshead speed of (a) 0.005 mm min 1and (b) 10 mm min 1 followed by tracer hydrogen charging until saturation.

Fig. 4a1. Additionally, several voids connected in a direction perpendicular to the tensile direction are also observed, as shown in Fig. 4a2. In contrast, for the specimens subjected to a high crosshead speed (10 mm min 1), several isolated round voids are observed in Fig. 4b1. Some of the voids displayed dimples filled with inclusions, as shown in Fig. 4b2. The type of dimples in Fig. 4b1 and b2 was also observed in the hydrogen-free specimens. It should be noted that, for hydrogen-charged specimens subjected to strain at low crosshead speeds, the shallow flat dimples in Fig. 2 probably form as a consequence of voids without inclusions such as secondary phase particles. Formation of Lattice Defects Enhanced by Straining With/Without Hydrogen

Fig. 6. Relationship between tracer hydrogen content and crosshead speed of specimens applied at 8% plastic strain with/without hydrogen.

the fracture surfaces in Fig. 2. The ratio is seen to decrease with decreasing crosshead speed and is accompanied by an increase in HE susceptibility. Figure 4 shows SEM images of the lateral surfaces of fractured hydrogen-charged specimens. In Fig. 4a1 and a2, the crosshead speed is 0.005 mm min 1, and in Fig. 4b1 and b2, it is 10 mm min 1. All observations were conducted in the vicinity of the fracture surface. For specimens subjected to a low crosshead speed (0.005 mm min 1), the majority of voids grew perpendicular to the tensile direction as observed in

This section focuses on the formation of hydrogenenhanced strain-induced lattice defects. Figure 5 shows hydrogen desorption profiles for pre-strained specimens with/without hydrogen deformed to 8% plastic strain at crosshead speeds of (a) 0.005 mm min 1 and (b) 10 mm min 1 followed by tracer hydrogen charging until saturation. Defect formation was assessed by comparing the tracer hydrogen content of the specimens subjected to plastic strain of 8% with hydrogen [H + 8% strain] and without hydrogen [8% strain]. As seen in Fig. 5a, there are significant differences between the two specimens for a low crosshead speed of 0.005 mm min 1. In contrast, the profiles appear similar for a high crosshead speed of 10 mm min 1, as seen in Fig. 5b. These results, in conjunction with the fact that there was a similar increase in dislocation density for the [8% strain] and [H + 8% strain] specimens with increasing amount of prestrain, indicate that prestraining in the presence of hydrogen enhances the formation of new lattice defects at low crosshead speeds. The increase in the temperature limit for hydrogen evolution in Fig. 6a

The Role of Hydrogen-Enhanced Strain-Induced Lattice Defects on Hydrogen Embrittlement Susceptibility of X80 Pipeline Steel

Fig. 7. Schematic illustration of fracture process of X80 pipeline steel during plastic deformation with/without hydrogen.

also indicates the formation of newly generated defects. These results are in good agreement with those previously reported for pure iron,5 tempered martensitic steel6 and cold-drawn pearlitic steel.7 Figure 6 reveals that the tracer hydrogen content corresponding to the amount of defects formed increased with decreasing crosshead speed. In additional experiments, a significant increase in the amount of defects was observed immediately prior to fracture. These findings indicate that hydrogen and strain both play a role in enhancing the formation of lattice defects immediately prior to fracture at low crosshead speeds. This could be due to interactions between dislocations and hydrogen. Previous studies showed that vacancy-type defects in pure iron were eliminated by annealing at 200°C.5,14 In order to investigate the type of strain-induced defects that were enhanced by hydrogen at lower crosshead speeds, the [H + 8% strain] and [8% strain] specimens were annealed at 200°C for 2 h and hydrogen was again introduced as a tracer for lattice defects. The results showed that the amount of defects was the same in both specimens following annealing, indicating that the defects enhanced by straining with hydrogen at lower crosshead speeds were vacancies and vacancy clusters. Therefore, the findings from the present study indicate that hydrogen plays a significant role in promoting the formation of vacancy type defects during plastic deformation immediately prior to fracture at low crosshead speeds. Low crosshead speeds, i.e., low dislocation velocities enhance lattice defect formation and stabilization. It is likely that the presence of hydrogen causes stabilization and accumulation of vacancies generated by cutting of screw dislocations and combination of edge dislocations during straining. Using firstprinciples calculations, Matsumoto et al. 15 also reported that hydrogen stabilizes vacancies and that locally accumulated vacancies promote ductile fracture. Furthermore, vacancy type defects grow preferentially and connect together to form the nuclei of dimples, resulting in a ductility loss as shown in Fig. 7.

CONCLUSION The role of strain-induced lattice defects enhanced by the presence of hydrogen on hydrogen embrittlement susceptibility of X80 pipeline steel was investigated. The results obtained in the present study can be summarized as follows. Hydrogen embrittlement susceptibility and the amount of hydrogen-enhanced strain induced lattice defects increased with decreasing crosshead speed. Following annealing, specimens strained with/without hydrogen exhibited the same amount of tracer hydrogen, and this was caused by the annihilation of vacancies and clusters. The fracture surfaces of samples with/without hydrogen appeared to exhibit conventional ductile fracture with a dimpled structure. However, this was actually only the case for the hydrogen-free specimens, which had conventional dimples containing nuclei such as secondary phase particles. In contrast, the hydrogen-charged specimen exhibited shallow flat dimples with no particles acting as nuclei. The dimples with nuclei were generated due to the formation of voids around the secondary phase particles, whereas the empty dimples were generated by the coalescence of voids, which in turn could be nucleated by the coalescence of vacancy type defects. The presence of a series of interconnected voids aligned perpendicular to the tensile direction in the hydrogen-charged specimens also suggests that final fracture occurred after the conjugation of voids nucleated by the coalescence of vacancy type defects. REFERENCES 1. B. Saleem, F. Ahmed, M. Asif Rafiq, M. Ajmal, and L. Ali, Eng. Fail. Anal. 46, 157 (2014). 2. B.T. Lu, J.L. Luo, and P.R. Norton, Corros. Sci. 52, 1787 (2010). 3. P. Liang, X. Li, C. Du, and X. Chen, Mater. Des. 30, 1712 (2009). 4. Z.Y. Liu, X.Z. Wang, C.W. Du, J.K. Li, and X.G. Li, Mater. Sci. Eng. A 658, 348 (2016). 5. K. Takai, H. Shoda, H. Suzuki, and M. Nagumo, Acta Mater. 56, 5158 (2008). 6. T. Doshida, M. Nakamura, H. Saito, T. Sawada, and K. Takai, Acta Mater. 61, 7755 (2013). 7. T. Doshida and K. Takai, Acta Mater. 79, 93 (2014).

Hattori, Suzuki, Seko, and Takai 8. T. Neeraj, R. Srinivasan, and J. Li, Acta Mater. 60, 5160 (2012). 9. C. Hwang and I.M. Berstein, Acta Metall. 34, 1001 (1986). 10. H. Shoda, H. Suzuki, K. Takai, and Y. Hagihara, ISIJ Int. 50, 115 (2010). 11. H. Qiu, H. Mori, M. Enoki, and T. Kishi, ISIJ Int. 39, 358 (1999).

12. P. Fassina, F. Bolzoni, G. Fumagalli, L. Lazzari, L. Vergani, and A. Sciuccati, Proc. Eng. 10, 3226 (2011). 13. M. Wang, E. Akiyama, and K. Tsuzaki, Scr. Mater. 52, 403 (2005). 14. T. Doshida, H. Suzuki, K. Takai, N. Oshima, and T. Hirade, ISIJ Int. 52, 198 (2012). 15. R. Mastumoto, N. Nishiguchi, S. Taketomi, and N. Miyazaki, J. Soc. Mater. Sci. Jpn. 39, 182 (2014).