Engineering
Procedia Engineering 00 (2009) 000–000www.elsevier.com/locate/procedia
Fatigue 2010
Degradation of fatigue properties in high pressure gaseous hydrogen
environment evaluated by cyclic pressurization tests
Jun Nakamura
a*, Mitsuo Miyahara
b, Tomohiko Omura
a,
Hiroyuki Semba
a, Masayuki Wakita
aand Youhei Otome
aa
Corporate Research and Development Laboratories, Sumitomo Metal Industries, Ltd.
b
Railway Automotive and Machinery Parts Company, Sumitomo Metal Industries, Ltd.
Received 24 February 2010; revised 10 March 2010; accepted 15 March 2010
Abstract
Fatigue properties in high pressure gaseous hydrogen environment were investigated for pipe materials used in fuel cell vehicles and hydrogen stations. Cyclic pressurization tests were conducted using a tubular specimen filled with hydrogen pressurized up to 90MPa. Tested materials were types 316L, 304 and A286 stainless steels, low alloy steels such as JIS SCM435 (1.0Cr-0.2Mo), Cr-Mo-V steels. The fatigue life in hydrogen was compared with that in inert gas to evaluate the effect of gaseous hydrogen on fatigue properties. The fatigue life in hydrogen was slightly shorter than that in argon in the case of a stable stainless steel type 316L. In contrast, a metastable stainless steel type 304 showed a remarkable degradation of the fatigue life in the hydrogen environment. Although the fatigue lives in hydrogen of type 316L and 304 stainless steels decreased with the increase in the
cycle times, the fatigue lives remained unchanged over 102 of the cycle time. The fatigue life of low alloy SCM435 steel in
hydrogen extremely decreased. The fatigue life of high-strength austenitic steel A286 in hydrogen is much better than that of low alloy SCM435 steel at the same tensile strength. The Cr-Mo-V steels showed longer fatigue lives than JIS SCM 435 at same strength levels in hydrogen. Fracture surfaces revealed transgranular cracking for the Cr-Mo-V steels, while intergranular cracking was observed for SCM435 with tensile strength more than 1.2GPa. It was assumed that the carbide precipitation affected the fracture morphology.
Keywords: Gaseous hydrogen; High pressure; Fatigue; Crack propagation; Stainless steel; Low alloy steels; Carbide
1. Introduction
Material properties in a high-pressure gaseous hydrogen environment have to be clarified in order to choose the appropriate materials for ensuring the safe operation of the metallic components, such as tubing and pressure vessels exposed to high pressure gaseous hydrogen in fuel cell vehicles and hydrogen stations.
“Establishment of codes and standards for hydrogen economy society” project started in 2005 as a national project in Japan. This project addresses the targets on the developments of hydrogen production, storage,
* Corresponding author. Tel.: +81-6-6489-5733; fax: +81-6-6489-5794. E-mail address: nakamura-jun@sumitomometals.co.jp
c 2010 Published by Elsevier Ltd. Procedia Engineering 2 (2010) 1235–1241 www.elsevier.com/locate/procedia 1877-7058 c 2010 Published by Elsevier Ltd. doi:10.1016/j.proeng.2010.03.134
Open access under CC BY-NC-ND license.
transportation and safe use systems, and started to investigate mechanical properties of several candidate materials by tensile or fatigue evaluations in high pressurized hydrogen gas environments.
The susceptibility to hydrogen embrittlement and the fatigue properties of several alloys were evaluated by SSRT (Slow Strain Rate Test) and a cyclic pressurization test using tubular specimens filled with hydrogen pressurized up to 90MPa [1-7]. In this study, the fatigue properties in hydrogen were evaluated by comparing them with those in argon gas. The degradation mechanism is discussed from the viewpoint of fracture morphologies as well as microstructural features.
2. Experimental
2.1. Materials
Tested materials were types 316L (0.014C-16Cr-12Ni-2.1Mo), 304 18Cr-9Ni-0.1Mo) and A286 (0.05C-15Cr-24.1Ni-0.3V-2.1Ti) stainless steels, low alloy steel JIS SCM435 (0.4C-1Cr-0.26Mo) andCr-Mo-V steels (Steel A; 0.27C-0.5Cr-0.7Mo-0.1V and Steel B; 0.4C-1Cr-0.7Mo-0.25V). The tensile strengths of types 316L and 304 were adjusted by solution heat treatments, to 551MPa and 569MPa respectively. As for type 316L large scale tensile specimens with 40mm thickness were taken from hot-rolled plate, then 30% plastic strain was applied by tensile tests at room temperature. The tensile strength of the cold worked type 316L was 723MPa.
Many kinds of materials have to be applied for the wide use of fuel cell vehicles although only aluminum alloy A6061-T6 and type 316L are recommended as liner material of FRP (Fiber Reinforced Plastics) tanks in fuel cell vehicles under High Pressure Gas Safety Act in Japan. A286 and cold worked type 316L are expected to be used for weight saving with an increase in hydrogen pressure in future. Additionally it is necessary to clarify the applicability of economic steels such as type 304 and low alloy steels. SCM435 steel has been used for accumulators in hydrogen stations.
2.2. Internal and external fatigue tests
In the internal pressure fatigue test the tubular specimen is filled with hydrogen gas or inert gas, and its pressure is cyclically varied (Fig.1 (a)). Therefore the stress is applied by pressure fluctuation, so it is comparatively easy to evaluate fatigue properties in very high pressure hydrogen gas. In the internal pressure fatigue test the internal pressure was cyclically varied from minimum pressure 10 MPa to maximum pressure 85 MPa within a frequency range between 0.00167 and 0.0167 Hz. Either hydrogen or nitrogen gas was used as the inner gas; the purity of the hydrogen gas was 99.99999%. On the other hand in the external pressure fatigue test the internal pressure is kept constant and the outside of the specimen is cyclically pressurized by water(Fig.1(b)) with a frequency of 0.05Hz at room temperature with a triangle wave. Either hydrogen or argon gas was used as the inner gas. The value of external waterpressure was determined so that the stress ratio (ratio of minimum stress to maximum stress at the inner surface) was equal to zero.
Notch Notch
(a) Internal pressure fatigue test (b) External pressure fatigue test Fig.1. Schematic view of internal and external pressure fatigue tests
When the fatigue crack penetrated the outer surface, the internal pressure decreased by leakage of the inner gas in these tests. The number of cycles to reach the leakage from the beginning of the test is defined as the fatigue life in this study. Pre-cracked specimen, shown in Figure 2, as well as a smooth specimen was used. The depth and the length of the notch in the pre-cracked specimen were 1mm and 5mm, respectively.
po㧔water pi( H2or Ar ) pi(const.) po Time Pressure pi( H2or Ar ) Pressure pi Time
J. Nakamura et al / Procedia Engineering 00 (2010) 000–000 3 b=1 2a=5 200 50 R30 24 3 4
Fig.2 Example of dimensions of pre-cracked specimen.
3. Results
The relationship between stress amplitude and the fatigue life of pre-cracked specimens by external fatigue test is shown in Figure 3. The fatigue life in hydrogen was slightly shorter than that in argon in the case of stable stainless steels types 316Land cold worked 316L. In contrast, A286 and a metastable 304 type stainless steel showed a remarkable degradation of the fatigue life in the hydrogen environment. The relationship between stress amplitude and the fatigue life of smooth specimens is shown in Figure 4. Type 304 and A286 showed a remarkable degradation of the fatigue life in the hydrogen environment even without the pre-crack.
Figure 5 shows the effect of tensile strength on fatigue life ratio of low alloy steels. The fatigue life ratio was defined as the fatigue life in hydrogen over that in argon. The fatigue life ratios of low alloy steels were degraded with an increase of the tensile strength. The fatigue life ratios of the Cr- Mo-V steels were larger than that of SCM435.
The relationship between fatigue lives of austenitic stainless steels types 316L and 304 and cycle times by internal fatigue test are shown Figure 6. Although the fatigue lives in hydrogen of type 316L and 304 stainless steels decreased with the increase in cycle time, the fatigue lives remained unchanged over 102 of the cycle time. The fatigue lives in nitrogen of these stainless steels were not affected by the cycle time.
0 50 100 150 200 250 300 103 104 105 Type 316L Cold worked type 316L A286 Type 304 S tr e ss amplitud e, σa
Number of cycles to failure, Nf
open:in Ar solid:in H2
Pre-cracked specimen 2a=5, b=1mm
100 200 300 400 500 600 103 104 105 Type 304 A286 S tr es s am pl it u d e, σa ( MPa)
Number of cycles to failure, Nf
open:in Ar solid:in H2
Smooth specimen
Fig.4 Results of fatigue tests for smooth specimen. Fig.3 Results of fatigue tests for pre-cracked specimen.
10-2 10-1 100 800 900 1000 1100 1200 1300 SCM435 Steel A Steel B Pmax = 85MPa F at ig u e lif e ra ti o ( H2 /Ar )
Tensile strength, T.S. (MPa)
103 104 10 102 103 SUS316L : Pi=10~85MPa SUS304: Pi=10~70MPa Cycle time (s) Nu mb er of c y cl e s to fa ilu re, N f solid : in H2 open : in N2
Fig.5 Effect of tensile strength on fatigue lives Fig.6 Effect of cycle time on fatigue lives
of low alloy steels. of stainless steels
4. Discussion
4.1. Fatigue degradation mechanism of type 304 stainless steel
It is generally reported that the fatigue degradation in hydrogen depends upon strain-induced martensitic transformation in metastable austenitic stainless steel [8, 9]. Therefore, the microstructure at the crack tip of type 304 was observed by Electron Back Scatter Diffraction Pattern (EBSD). Figure 7 and 8 show EBSD images near the crack tip of type 304 after an external fatigue test, which was interrupted. The martensitic phase was observed around the fatigue crack tip and only the each side of the crack propagation. The martensitic phase near the crack tip was observed on the slip line. It was expected to clarify that fatigue crack in hydrogen environment and the strain-induced martensitic phase as increasing the strain-introduced martensitic phase in the large strain case, hence in order to clarify the relationship between fatigue crack in hydrogen and the martensitic phase clearly the sample after an external fatigue test was observed by EBSD. Figure 9 and 10 show EBSD images near the crack tip after an external fatigue test. Fatigue crack propagation was observed along the interface between the martensitic phase and the austenitic phase. Hydrogen diffuses more rapidly in the martensitic phase than in the austenitic phase. Therefore it was considered that the base metal near the crack tip was embrittled due to hydrogen accumulation [10], which was supplied through the strain-induced martensitic phase.
200μm Area 200μm 200μm Area 110 111 100 110 111 100
Fig.8 Inverse pole figure map (martensite) by EBSD of crack tip of SUS304
(H2 gas, pi=50MPa,po=5-44.8MPa, interrupted test).
α’ phase Crack
Fig.7 Inverse pole figure map by EBSD of crack tip of SUS304
J. Nakamura et al / Procedia Engineering 00 (2010) 000–000 5 100μm 100μm Area 1 Area 2 110 111 100 110 111 100
Fig.9 Inverse pole figure map by EBSD of crack tip Fig.10 Inverse pole figure map (martensite) of SUS304 after fatigue test (H2 gas, pi=50MPa, by EBSD of crack tip of SUS304 after
po=5-44.8MPa). fatigue test (H2 gas, pi=50MPa).
4.2 Fatigue degradation mechanism of A286 steel
Examples of the fracture surface observed by using a scanning electron microscope (SEM) for A286 steel are shown in Figure 11.The fracture surface of the smooth specimen reveals intergranular fracture, while the fracture surface of pre-cracked specimen shows secondary cracking. The fatigue degradation of A286 in hydrogen could not be explained based on martensitic transformation, because the high-nickel content of A286 makes it resistant to strain-induced phase transformation. Two mechanisms concerning the fatigue degradation in hydrogen are considered. The grain of A286 is strengthened by precipitated γ’- phase (Ni3(Al, Ti)). The γ’-phase can localize the
slip, resulting in the occurrence of secondary cracking across the main slip. Furthermore, η-phase (Ni3Ti)
precipitated along grain boundaries, as shown in Figure 12, should promote intergranular fatigue degradation.
Fig.11 SEM images of fracture surface of A286 after external (b)Pre-cracked specimen (a)Smooth specimen
pressure fatigue test (H2 gas, pi=85MPa).
4.3 Fatigue degradation mechanism of low alloy steels
Examples of fracture surface observed in a SEM for SCM435 and Steel B are shown in Figure 13. Transgranular fracture was observed in Steel B while intergranular fracture was observed in the SCM435 at the same strength level. Figure 14 shows TEM images near the grain boundary ofthe Steel B and SCM435. In the Steel B, the carbides were spheroidized and dispersed uniformly both on the grain boundaries and within grains. In SCM435, carbides were needle-like along the grain boundaries. The difference in carbides precipitation should affect the fracture morphology. Figure 15 shows pattern diagram of mechanism of hydrogen embrittlement. Since carbides particles act as trapping sites, dislocation involved hydrogen concentrates near the carbides precipitation sites. It is considered that hydrogen accumulates at the interface of needle-like carbides, resulting in intergranular fatigue fracture.
Fig.14 Comparison of carbide morphology between SCM435 and the SteelB.
(b)Steel B (TS: 875MPa) 1μm (a) SCM435 (TS: 854MPa)
1μm
Fig.13 SEM images of fracture surface at SCM435 and the Steel B at the same strength level (H2 gas, pi=85MPa).
(a)SCM435 (TS: 1268MPa)
30μm 30μm
(b)Steel B (TS: 1245MPa)
Dislocation involved hydrogen atom
H Fe3C Cementite H H H H Fe3C H H H H H H Fe3C Fe3C H H H H H GB GB H H H H H HH H H Fe3C Fe3CH H H H H H H H H H H Grain Grain Grain Grain
Dislocation involved hydrogen atom
H Fe3C Cementite H H H H Fe3C H H H H H H Fe3C Fe3C H H H H H GB GB H H H H H Fe3C H H H H H H Fe3C Fe3C H H H H H H H H H H Fe3C H H H H H H Fe3C Fe3C H H H H H GB GB H H H H H HH H H Fe3C Fe3CH H H H H H H H H H H Grain Grain Grain Grain H H H H HH H H Fe3C Fe3CH H H H H H H H H H H Grain Grain Grain Grain Grain Grain Grain Grain
Fig.15 Mechanism of hydrogen embrittlement of low alloy steels.
J. Nakamura et al / Procedia Engineering 00 (2010) 000–000 7 Conclusion
The external and internal pressure fatigue tests were conducted with tubular specimens of types 316L, 304 and A286 austenitic stainless steels and low alloys SCM435 and Cr- Mo-V steels. Fatigue properties in high pressure hydrogen gas were compared with those in argon. The mechanism of fatigue properties were discussed from the viewpoint of microstructural alteration and fracture surface morphologies. Following results were obtained. 1) As for stable austenitic stainless steels types 316L and cold worked 316L, the difference in the fatigue life between hydrogen and argon environments was small. In contrast, as for A286 and metastable austenitic stainless steel type 304, the fatigue lives were shorter than those in argon. The degradation in fatigue properties of type 304 steel was supposed that the base metal near the crack tip was embrittled due to hydrogen accumulation, which was supplied through the strain-induced martensitic phase. The grain of A286 is strengthened by precipitated γ’-phase (Ni3 (Al, Ti)), which can localize slip, resulting in the occurrence of secondary cracking across the main slip.
Furthermore η-phase (Ni3Ti) precipitated along grain boundaries, resulting in the intergranular fatigue degradation.
2) As for low alloy steels, the fatigue life in hydrogen was shorter than in argon with an increase in the tensile strength. In Cr-Mo-V steels, carbides were spheroidized and dispersed both on the grain boundaries and within grains. In SCM435, carbides were needle-like along the grain boundaries. It is assumed that the difference in carbide precipitation affected the fracture morphology, resulting in the difference in the fatigue properties.
Acknowledgment
Part of the study was carried out through the project “Establishment of codes and standards for hydrogen economy society” in Japan, administrated by New Energy and Industrial Technology Development Organization (NEDO).
References
[1] T. Omura, K. Kobayashi, M. Miyahara and T. Kudo: Zairyo-to-kankyo, 55(2006), 537. [2] T. Omura, K. Kobayashi, M. Miyahara and T. Kudo: Zairyo-to-Kankyo, 55(2006), 139. [3] T. Omura, M. Miyahara and H. Semba: Journal of high pressure institute of Japan, 46(2008), 205. [4] T. Omura, H. Hirata, M. Miyahara and T. Kudo: Zairyo-to-Kankyo 57(2008), 30.
[5] M. Miyahara, T. Omura, H. Okada, K. Ogawa, M. Igarashi and T. Kudo: Proceeding of the WHTC2005.
[6] M. Miyahara, T. Omura, K. Okada, M. Igarashi, K. Ogawa and T. Kudo: Proceeding of the Society of Materials Science, Japan, 55(2006), 65.
[7] J. Nakamura, M. Miyahara, E. Nakayama, T. Omura, H. Semba and M. Wakita: Proceeding of Camp ISIJ, 21(2008), 1289. [8] S. Fukuyama, K. Yokogawa and M. Araki: Journal of the Society of Materials Science, Japan, 36(1986), 506.
[9] S. Fukuyama, K. Yokogawa and K. Kudo: Journal of the Society of Materials Science, Japan, 32(1986), 430. [10] Y. Fukai et al: Hydrogen and metal , Uchidaroukakuho (1998), 113.
