水素ガス中に含まれる不純物が水素助長破壊に及ぼ す影響

九州大学学術情報リポジトリ

Kyushu University Institutional Repository

水素ガス中に含まれる不純物が水素助長破壊に及ぼ す影響

薦田, 亮介

https://doi.org/10.15017/1807029

出版情報：Kyushu University, 2016, 博士（工学）, 課程博士 バージョン：

権利関係：Fulltext available.

Effect of Impurities Contained in Hydrogen Gas on Hydrogen-Assisted Degradation of Material Strength

2017

Ryosuke Komoda

Contents

Nomenclature……….1

1 Introduction………2

1.1 Recent world energy situation……….2

1.2 Hydrogen-assisted degradation of material strength………...4

1.2.1 Hydrogen-assisted degradation of fracture toughness………....5

1.2.2 Hydrogen-assisted degradation of fretting fatigue strength………..12

1.3 Impurities contained in hydrogen gas and their effect on hydrogen -assisted degradation………..………...17

1.4 Objective and content of this study………24

1.5 References………..26

2 Development of a new measurement method for relative slip during fretting fatigue test in hydrogen applying MEMS optical micro-encoder………30

2.1 Introduction………...30

2.2 MEMS optical micro-encoder………...33

2.3 Measurement method of relative slip range during fretting fatigue…..………35

2.4 Correction of the error due to elastic deformation………..………..37

2.5 Measurement result of relative slip range during fretting fatigue in hydrogen…….…...39

2.6 Conclusions………...41

2.7 References……….42

3 Effects of addition of ppm-levels of oxygen and water vapor on fretting fatigue properties in hydrogen………..44

3.1 Introduction………...44

3.2 Test procedure………...46

3.2.1 Material……….46

3.2.2 Fretting fatigue test method………..47

3.2.3 Preparation of test environment………49

3.2.4 Test environment……….…..53

3.3 Test results……….54

3.3.1 Effect of addition of ppm-levels of oxygen to hydrogen gas on fretting fatigue strength………..54

3.3.2 Effect of humidification on fretting fatigue strength in hydrogen………57

3.3.3 Fretted surface………..59

3.3.4 Position of fretting fatigue main crack……….61

3.4 Discussion……….63

3.4.1 Effect of addition of ppm-levels of oxygen on crack initiation limit at adhered spot ………..64

3.4.1.1 procedure of adhesion mimic fatigue test……….65

3.4.1.2 Finite element analysis of adhesion mimic fatigue test……….69

3.4.1.3 Crack initiation limit………..71

3.4.2 Development of fretting wear………...………73

3.4.3 XPS component analysis of fretted surface………..76

3.4.4 Relative slip range………80

3.4.5 Mechanism of reduction in the fretting fatigue strength by addition of oxygen or humidification………...82

3.5 Conclusions………...84

3.6 References……….86

3.7 Appendixes………...89

4 Consideration of gas system in fracture toughness test using open gas system and closes gas system………..91

4.1 Introduction……….………..91

4.2 Experimental procedure……….……….……..93

4.2.1 Material and specimen……..………...……….………93

4.2.2 Fracture toughness test method……….………95

4.2.3 Effect of exposure to hydrogen gas on sensor output………97

4.2.4 Preparation procedure of test environment………...98

4.2.5 Test environment……….101

4.3 Test results……….………102

4.3.1 J-a curve……….………102

4.3.2 Crack tip shape………103

4.3.3 Fracture surface….………104

4.4 Discussion………106

4.4.1 Predictive model of inhibitory effect of oxygen on hydrogen-assisted degradation of fracture toughness………….………...……….……108

4.4.2 Effect of crosshead speed on fracture toughness in air……….……..110 4.4.3 Effect of oxygen content and crosshead speed on inhibitory effect of oxygen…111 4.4.4 Effects of oxygen content and loading rate on inhibitory effect of oxygen on

hydrogen-assisted degradation of fracture toughness……….113

4.5 Conclusions……….114

4.6 References………117

4.7 Appendixes………..119

5 Inhibitory effect of carbon monoxide on hydrogen-assisted degradation of fracture toughness ………120

5.1 Introduction……….……120

5.2 Test procedure………..122

5.2.1 Material and specimen….……….………122

5.2.2 Fracture toughness test method..……….124

5.2.3 Test environment……….124

5.3 test results………125

5.3.1 Effect of carbon monoxide content on inhibitory effect……….…125

5.3.2 Crack tip shape………127

5.3.3 Fracture surface….………..……128

5.4 Discussion………...130

5.4.1 Comparison of inhibitory effects of carbon monoxide on hydrogen-assisted degradation of fracture toughness in A333 and A106………130

5.4.2 Comparison of inhibitory effects of carbon monoxide and oxygen on hydrogen-assisted degradation of fracture toughness……….131

5.4.3 Future task……….134

5.5 Conclusions……….135

5.6 References………...…136

6 Conclusion……….138

7 Acknowledgement……….142

1 Nomenclature

 elongation i electrochemical current density

 reduction of area N number of cycles

0.2 0.2 % proof stress Nf number of cycles to failure

a stress amplitude pc contact pressure

B ultimate tensile strength R stress ratio

Y proof strength V crosshead speed

wfH fretting fatigue limit AD axial direction

max maximum shear stress CT compact tension

max maximum range of shear stress FE finite element

a crack extension HD hoop direction

S relative slip range LD laser diode

Se elastic deformation MEMS micro electro mechanical systems

CO oxygen content PD Photodiode

COD crack opening displacement RD radial direction

d mesh size RH relative humidity

dc crack position SMR steam methane reforming

f test frequency WOL wedge opening loaded

F fatigue load XPS X-ray photoelectron spectroscopy

Fc compressive load

2 1. Introduction

1.1 Recent world energy situation

Since the industrial revolution, the economies and lives of developed countries have depended on fossil energy. The consumption of fossil energy is steadily increasing year after year [1].

As of 2014, the world’s energy consumption has increased to about three times as compared to that in 1965. Furthermore, it is expected that the energy consumption will be further increased in the near future due to the growth of developing countries. However, fossil energy is a finite resource [2], and we cannot continue to consume fossil energy at the current rate.

In recent years, the temperature of the earth has been rising. As of 2015, the temperature increased by 0.78℃ on average in the world as compared to 1891. Especially, the temperature increased by 0.42℃ only after 1980 [3]. One of the reasons for the global warming is carbon dioxide generated when fossil fuel is burned. Therefore, to prevent the global warming, some efforts and agreements to reduce the emission of the carbon dioxide and establish a carbon neutral society have been made worldwide [4].

Japan is also dependent for energy on fossil fuel. Since Japan has poor fossil resources, most of the fossil energy depends on importation from overseas. As of 2014, the energy self-sufficiency rate in Japan is less than 10 % [5]. Thus, the Japanese energy structure is very susceptible to changes in the energy situation worldwide and is vulnerable. Therefore, from the viewpoint of energy security, it is also required to break the dependence of Japan on fossil energy.

The use of renewable energy such as wind power and solar power is being promoted as a substitute for fossil fuel. However, there are some problems to use the renewable energy. One of them is energy instability of the renewable energy. For example, wind power and solar power are strongly affected by the weather. As a result, both excess and deficiency of generated energy can occur, making difficult to provide directly to the energy grid. For this reason, equipment such as batteries to store the power generated by renewable energies is required. However, the battery that can storage large amount of electricity still has some technology and cost issues [6]. In addition, the power generation by renewable energy is often done at remote locations away from urban areas.

Therefore, there are also problems in terms of an energy transport method.

As one of way to meet and solve these issues, hydrogen as an energy carrier attracts attentions. Hydrogen can be produced by electrolysis of water using electricity generated by

3

renewable energy, and then temporarily stored as a chemical energy. Thereafter, electric energy is electrochemically derived from the hydrogen by fuel cell. Hydrogen can be easily stored and transported in gaseous or liquid state. Therefore, the instability of renewable energy can be eliminated by using the hydrogen as an energy carrier, and it contributes to establish a carbon neutral society.

However, at the present, the amount of electricity generated by renewable energy is very small, and it takes time to establish an energy infrastructure mainly based on renewable energy sources. Even so, the use of hydrogen still has merits. Currently, most of the hydrogen is produced by the steam reforming method [1], and carbon dioxide is also emitted in the process. However, since the energy efficiency derived from hydrogen by using fuel cell is higher than that derived from fossil fuel by burning [7], the amount of carbon dioxide emission can be reduced by using hydrogen.

Also, since hydrogen can be produced from various resources, it has high stability as energy carrier.

Following the above, the use of hydrogen energy is being promoted globally in order to construct a carbon neutral society.

4 1.2 Hydrogen-assisted degradation of material strength

Hydrogen can degrade the material strength [8-11]. Hydrogen-assisted degradation may lead to catastrophic failures of the structural members of pressurized hydrogen containment systems.

Therefore, it obstructs the promotion of the use of hydrogen energy in terms of safety of pressurized hydrogen containment systems. To ensure the safety, the design and management methods of the systems are strongly restricted regional and state agency regulations [12-22].

The degree of the deterioration of the material strength by hydrogen depends on the material [23, 24]. Low carbon steels and low alloy steels, which are often used for structural components, are extremely or severely embrittled by hydrogen, which greatly limits their use.

Instead of these steels, materials having hydrogen resistance, such as SUS316L and A6061-T6, are often used for pressurized hydrogen containment systems. However, such austenitic stainless steel and aluminum alloy have lower material strength and relatively higher material cost than low carbon steels or low alloy steels. It causes an increase in the cost of pressurized hydrogen containment systems, and it also hinders the promotion of use of hydrogen energy. Therefore, in order to reduce the cost of the pressurized hydrogen containment systems while ensuring the safety, a large number of studies, investigating extensively the hydrogen-materials interactions, have been carried out on both experimental and theoretical aspects [9-11].

In next two Subsections, the effect of hydrogen on fracture toughness and fretting fatigue strength will be reviewed as the examples of the research on the hydrogen-materials interactions.

5

1.2.1 Hydrogen-assisted degradation of fracture toughness

It is known that hydrogen deteriorates fracture toughness of various materials, and the effects of influence factors on the hydrogen-assisted degradation of fracture toughness have been investigated [9-11, 25-28, 30-32].

Walter et al. reported a hydrogen gas pressure dependency on the crack arrest threshold stress intensity factors of AISI 4340 steel up to 30 MPa of the gas pressure [25]. They showed that the crack arrest threshold stress intensity factors decreased with an increase in the hydrogen gas pressure and was proportional to the hydrogen gas pressure to the 0.076 power as shown in Fig. 1.1.

According to Sievert’s low [26], hydrogen solubility into material is proportional to the hydrogen gas pressure to the 0.5 power. Walter et al. concluded that the difference in the exponent of hydrogen gas pressure dependency between their results and Sievert’s low indicates that the embrittlement resulted from a hydrogen adsorption rather than absorption of hydrogen into the material. Loginow et al. also reported a hydrogen gas pressure dependency on the crack arrest threshold stress intensity factors of various low-carbon steels and low-alloy steels up to 97 MPa of hydrogen gas pressure as shown in Fig. 1.2 [27]. For steels with yield strength from 586 to 779 MPa, the crack arrest threshold stress intensity factors were proportional to the hydrogen gas pressure to the 0.38 power. They concluded that the hydrogen gas pressure dependency of their test result is sufficiently close to the Sievert’s low, and it suggests that the embrittlement in hydrogen gas is governed by Sievert’s low for hydrogen solubility. On the other hand, for steels with yield strength from 869 to 1055 MPa, the crack arrest threshold stress intensity factors were proportional to the hydrogen gas pressure to the 0.2 power. They concluded that the difference in the exponent for the high-strength steels from the Sievert’s low suggests that there are factors other than hydrogen content in the material.

Loginow’s report also showed a material strength dependency on the crack arrest threshold stress intensity factors. The crack arrest threshold stress intensity factors decreased with an increase in the yield strength of the material. Nibur et al. reported a material strength dependency on the crack arrest threshold stress intensity factors of Cr-Mo steels, Ni-Cr-Mo steels and Hy-130 steel in 103 MPa hydrogen gas as shown in Fig. 1.3 [24]. The results showed the similar trend as Loginow’s results.

Matsumoto et al. reported a loading rate dependency on the fracture toughness of carbon steel in 0.6 MPa hydrogen gas [28]. They showed that the fracture toughness in hydrogen increased with a decrease in the crosshead speed of the fracture toughness test from 2 × 10^-3 to 2 × 10^-5 mm/s.

6

They concluded that the mitigation of the hydrogen-assisted degradation by the decreasing in the crosshead speed can be explained by the gradient of the hydrogen concentration in the material [29].

Fig. 1.1 Stress intensity at crack arrest for AISI 4340 WOL specimen at ambient temperature as a function of hydrogen pressure [25]

7

Fig. 1.3 Crack arrest threshold stress intensity factors for steels tested in 103 MPa hydrogen gas as a function of yield strength [24]

Fig. 1.2 Stress intensity at crack arrest for carbon steels and low alloy steels at ambient temperature as a function of hydrogen pressure [27]

8

Fig. 1.4 Effect of crosshead speed on hydrogen-assisted degradation of fracture toughness [28]

9

Many researchers reported the change in the fracture surface in the fracture toughness tests by hydrogen and associates the fracture surface change with hydrogen-assisted degradation of the fracture toughness. Nagumo et al. reported the fracture surface change due to hydrogen during fracture toughness test of hydrogen-charged ferrite-pearlite steels [30]. Both hydrogen charged and uncharged material showed dimple fracture, but the dimple size was enlarged by hydrogen charge and quasi-cleavage pattern were occasionally observed within the enlarged dimples as shown in Fig.

1.5. On the other hand, Wang reported that the dimple size of hydrogen charged X70 pipeline steel produced by fracture toughness test became smaller than that of uncharged specimen as shown in Fig. 1.6 [31]. The decrease in the dimple size became more significant with an increase in the current density of the hydrogen charging. Wang also reported that the fracture surface became quasi-cleavage by a dynamic hydrogen charge. Splichal et al. reported that the fracture surface of martenstic steel was changed from ductile dimple to inter-granular by hydrogen charge when the fracture toughness test was carried out at 20℃ as shown in Fig. 1.7, but the fracture surface was not changed by the hydrogen charge when the test was carried out at 120℃ [32].

Fig. 1.5 Fracture surface of uncharged specimen and hydrogen charged specimen [30]

(a) Uncharged specimen (b) Hydrogen charged specimen

10 (b) Pre hydrogen charged specimen

(i = 10 mA cm^-2)

(a) Uncharged specimen

(c) Pre hydrogen charged specimen (i = 100 mA cm^-2)

(d) Dynamic hydrogen charged specimen (i = 10 mA cm^-2)

(e) Dynamic hydrogen charged specimen (i = 100 mA cm^-2)

Fig. 1.6 Change in the fracture surface of pipeline steel by hydrogen charge [31]

11

Fig. 1.7 Hydrogen-induced inter-granular fracture [32]

12

1.2.2 Hydrogen-assisted degradation of fretting fatigue strength

Kubota and his research group reported the effect of hydrogen on fretting fatigue strength of austenitic stainless steels [33-39]. They showed that the fretting fatigue strength of solution heat-treated SUS304 was significantly reduced by gaseous hydrogen and internal hydrogen as shown in Fig. 1.7 [37].

Kubota et al. also reported unique phenomena occurring in the fretting fatigue test in hydrogen environment. After fretting fatigue test in hydrogen, the specimen and contact pad, which are originally separated, were adhered each other as shown in Fig. 1.8 [36]. In addition, when the adhered specimen and contact pad were teared off each other and the contact surface was observed, many damages like a diamond-shape were found as shown in Fig. 1.9 [35]. Kubota et al. observed the cross section of the contacting surface of the adhered specimen and contact pad in order to elucidate the reason why the specimen and contact pad were adhered during fretting fatigue test in hydrogen, and found that the contacting surfaces were locally adhered each other during the fretting fatigue test in hydrogen as shown in Fig. 1.10 [35]. Usually, adhesion between the contacting surfaces during fretting fatigue in air only occurs at the microscopic level, such as the real contact area, and cannot be recognized at the macroscopic level, such that specimen and contact pad are adhered together. This is because the accumulation of oxidized fretting wear particles between the contact surfaces separates the contact surfaces. On the other hand, oxidized fretting wear particles are not produced in hydrogen since no oxygen exists in hydrogen environment, and consequently the separation of the contacting surfaces is prevented. As the result, adhesion between the contacting surfaces occurs over a wide portion of the nominal contact area. Kubota et al. concluded that the diamond-like shaped fretting damage produced during the fretting fatigue test in hydrogen environment was caused by the local adhesion, and established the model of formulation of the fretting damage by the local adhesion as shown in Fig. 1.11 [36].

13

10⁴ 10⁵ 10⁶ 10⁷ 10⁸

0 50 100 150 200 250 300 350

Number of cycles to failure, N_f Stress amplitude, a (MPa)

Solution heat-treated SUS304

In air

Uncharge H charge In H₂

Uncharge H charge

Fig. 1.7 Effect of gaseous and internal hydrogen on fretting fatigue strength of SUS304 [37]

Fig. 1.8 Adhered specimen and contact pad after fretting fatigue test in hydrogen [36]

Fig. 1.9 Diamond-like shaped fretting damage produced during fretting fatigue test in hydrogen [35]

14

Fig. 1.11 Model of formulation of fretting damage in hydrogen: (a) occurrence of local adhesion and small crack initiation, (b) propagation of small cracks, (c) perspective view of (b) that shows crack propagation to the direction along the contact surface and (d) formulation of factory diamond-like shaped pattern when the contact surface is opened [36]

Fig. 1.10 Local adhesion between the contacting surfaces [35]

15

The local adhesion during fretting fatigue in hydrogen closely relates to the reduction in the fretting fatigue strength in hydrogen. At the adhered spot, as shown in Fig. 1.12, strain-induced martenstic transformation, severe plastic flow and grain refinement occurred [40]. These facts are evidence that severe cyclic plastic deformation occurred at the adhered spot during the fretting fatigue test. The severe cyclic plastic deformation causes small crack initiation at the adhered spot as shown in Fig. 1.13 [36]. The small crack initiation at the adhered spot is one of the reasons for the reduction in the fretting fatigue strength in hydrogen. In addition, gaseous and internal hydrogen reduces the crack initiation limit at the adhered spot as shown in Fig. 1.14 [38]. In the case of austenitic stainless steels, one of the reasons for the hydrogen-enhanced fretting fatigue crack initiation is strain-induced martensitic transformation at the adhered spot by the severe cyclic plastic deformation due to the adhesion [40]. This is because that the hydrogen diffusivity of martensite is much higher than that of austenite [42] and martensite is more sensitive to hydrogen-assisted degradation [43]. In this way, it was found that the local adhesion takes a key role in the reduction in the fretting fatigue strength in hydrogen.

Fig. 1.12 Microstructure change at adhered spot [40]

16

Fig. 1.13 Small crack initiation at the adhered spot [36]

Fig. 1.14 Reduction in the small crack initiation limit at the adhered spot gaseous and internal hydrogen [38]

17

1.3 Impurities contained in hydrogen gas and their effects on hydrogen-assisted degradation During hydrogen production process, some impurities, such as oxygen, water vapor, carbon monoxide, etc., sometime come to be mixed in the hydrogen gas [1]. Since the purity of hydrogen gas is directly linked to the cost of the gases, it should be allowable from an engineering point of view that a certain amount of impurities is contained in the hydrogen gas.

For example, the purity of hydrogen for the polymer electrolyte fuel cell (PEFC) is designated by an ISO 14687-2 standard [44] as better than 99.97 %. In other words, the hydrogen for the PEFC can contain 300 ppm impurities according to the standard. Species of impurities and level of each impurity are also provided by the standard such that hydrocarbon < 2ppm, H2O < 5 ppm, O2

< 5 ppm, He < 300 ppm, Ar, N2 < 100 ppm, CO2 < 2 ppm, CO < 0.2 ppm, S < 0.004 ppm, HCHO <

0.01 ppm, HCOOH < 0.2 ppm, NH3 < 0.1 ppm, halide < 0.05 ppm and particle < 1 mg/kg. CGA G-5.3 standard [45] determined the purity of hydrogen for various applications. For example, the purity of hydrogen gas for general industrial applications is specified as better than 99.95 %, and species of impurities and level of each impurity are also provided by the standard such that CO2 < 10 ppm, CO < 10 ppm, N2 < 400 ppm, O2 < 10 ppm, total hydrocarbon content < 10 ppm and H2O < 34 ppm.

In addition, when considering the operation condition of PEFC, the humidification of hydrogen is necessary for the proton conductivity [46]. Therefore, the hydrogen gas in PEFC contains a large amount of water vapor.

18

Impurities contained in hydrogen gas can affect hydrogen-assisted degradation of material strength. Somerday et al. reported that the addition of oxygen to hydrogen gas delayed the onset of hydrogen accelerated crack growth of pipe steel as shown in Fig. 1.15 [47]. The most significant accomplishment of their study was the establishment of a formula of a critical crack growth rate, at which hydrogen accelerated fatigue crack growth occurs in hydrogen environment containing a certain amount of oxygen. The formula shows that the inhibitory effect of oxygen is governed by the rate of bare surface creation by crack growth and the rate of repassivation of the bare surface by oxygen under given test conditions, which are hydrogen pressure, frequency and magnitude of fatigue loading and material strength. Frandsen et al. reported that oxygen, carbon monoxide and nitrous oxide have good inhibitory effects, but methane and carbon dioxide do not inhibit hydrogen-assisted fatigue crack growth of high-strength alloy steel as shown in Fig 1.16 [48].

Similar results were also reported in the reference [49]. In Ref. [49], the relationship between concentration of carbon monoxide and inhibitory effect of hydrogen-assisted fatigue crack growth of pipeline steel is discussed. The inhibitory effect of carbon monoxide diminishes with the decrease of carbon monoxide content from several vol. % to several vol. ppm as shown in Fig. 1.17. Similar to the case of fatigue crack growth, the carbon monoxide contained in hydrogen gas prevents reduction in fracture toughness due to hydrogen, but methane does not inhibit the reduction in the fracture toughness as shown in Fig. 1.18 [49]. Kussmaul et al. reported that 10 and 150 vol. ppm oxygen contained in hydrogen gas mitigates the hydrogen-assisted degradation of fracture toughness of low alloy steel, and the inhibitory effect of oxygen become more significant with higher oxygen concentration as shown in Fig. 1.19 [50].

19

Fig. 1.15 Inhibitory effect of oxygen contained in hydrogen gas on hydrogen-accelerated fatigue crack growth [47]

Fig. 1.16 Effect of species of impurity on its inhibitory effect on hydrogen-accelerated fatigue crack growth [48]

20

Fig. 1.17 Effect of inhibitor (CO) concentration in hydrogen on fatigue crack growth rate in X42 pipeline steel [49]

Fig. 1.18 Effect of species of impurity on its inhibitory effect on hydrogen-assisted degradation of fracture toughness [49]

21

Fig. 1.19 Inhibitory effect of oxygen on hydrogen-assisted degradation of fracture toughness [50]

22

Staykov et al. reported the effect of impurities on the hydrogen dissociation on a pure iron surface [51]. They calculated the potential interaction with iron surface and gas molecules by a theoretical simulation as shown in Fig. 1.20. An oxygen molecule is captured on the iron surface deeper in the gas volume (Fig. 1.20 (a)) compared to a hydrogen molecule (Fig. 1.20 (b)). The attractive interaction of an oxygen molecule with iron surface is much higher than that of a hydrogen molecule. There is an activation barrier for the hydrogen adsorption on the iron surface whereas there is no activation barrier for the oxygen adsorption on the iron surface. Consequently, an oxygen molecule is adsorbed on the iron surface faster and more efficiently than a hydrogen molecule. In addition, the oxygen dissociated on the iron surface increases the activation barrier for the hydrogen adsorption on the iron surface (Fig. 1.20 (c)). As the result, the hydrogen adsorption on the iron surface is blocked by the adsorbed oxygen. In the case of carbon monoxide, carbon monoxide also prevents the hydrogen absorption on the iron surface by the same mechanism as oxygen (Figs. 1.20 (d) and (e)). In the case of methane, there is a higher activation barrier for methane adsorption on the iron surface than that of hydrogen (Fig. 1.20 (f)). Therefore, methane does not prevent hydrogen adsorption. These calculation results correspond to the experimental result in Refs [48] and [49]

where oxygen and carbon monoxide have a good inhibitory effect on hydrogen-assisted degradation but methane does not.

Fig. 1.20 Potential interaction with iron surface and gas molecules [51]

(d) CO (e) H2 with preadsorbed CO (f) CH4

(a) O2 (b) H2 (c) H2 with preadsorbed O

23

Impurities contained in hydrogen gas also affect friction and wear properties in hydrogen environment. Fukuda et al. reported the effects of ppm-levels of oxygen or water on friction and wear properties in a sliding test in hydrogen environment [52, 53]. The ppm-levels of oxygen or water vapor increase a specific wear rate by enhancing the discharge of wear debris as shown in Fig.

1.21.

Fig. 1.20 Variations of specific wear rates at different concentration of water and oxygen in the hydrogen environment [52]

24 1.4 Objective and contents of this study

As described above, it has been clarified that some of impurities contained in hydrogen gas affect the material strength in the hydrogen environment. However, the studies on the effect of the impurities can only be found in basic fatigue crack growth test or fracture toughness test. In addition, only partial characterizations of the effect of impurities and elucidation of its mechanism have been carried out. Therefore, for deeper understanding and evaluation of the effect of the impurities on material strength properties of the pressurized hydrogen containment systems, further investigations and discussions should be done.

In this context, the effects of impurities on fretting fatigue strength and fracture toughness in hydrogen environment were studied. In the fretting fatigue, the effects of ppm-levels of oxygen and water vapor to hydrogen environment on fretting fatigue strength were characterized, and its mechanism was elucidated. In the fracture toughness, the influence factors on the effect of impurities, such as amounts and species of impurity, loading rate and material were characterized.

The detailed contents of each chapter are as follows:

In Chapter 2, a new measurement method for relative slip range during fretting fatigue test in hydrogen environment was developed. This study was done to enhance the understanding of the study in Chapter 3. An optical micro-encoder fabricated by MEMS technology was used for the relative slip sensor. By calibrating the error due to elastic deformation around the measurement point, an accurate measurement of the relative slip range in hydrogen environment was successfully done.

In addition, the measurement method showed a good reproducibility. By using this method, it was found that the relative slip range during fretting fatigue test in hydrogen environment was significantly lower than that in air.

In Chapter 3, the effects of addition of ppm-levels of oxygen and water vapor to hydrogen environment on fretting fatigue strength in hydrogen of an austenitic stainless steel SUS304 were studied. It was found that the fretting fatigue strength in hydrogen was significantly reduced by the addition of ppm-levels of oxygen and water vapor. The reduction mechanism was established by focusing on the change in the mechanical conditions at the contacting surface due to enhancing of the fretting wear by the addition of the impurities.

In Chapter 4, the effect of slight increase in the oxygen content in hydrogen environment during fracture toughness test in the hydrogen was studied. In the closed gas system, the oxygen content in hydrogen gas increased to 1 vol. ppm during the fracture toughness test. The 1 vol. ppm

25

oxygen partially inhibited the hydrogen-assisted degradation of the fracture toughness of A333 pipe steel. Based on the additional tests, it was found that the inhibitory effect of oxygen depends on the oxygen content and loading rate. This result well corresponds to the predictive model proposed by Somerday et al.

In Chapter 5, to transfer the inhibitory effect of impurities to practical application as a new barrier coating technology for hydrogen-assisted degradation, the inhibitory effect of carbon monoxide on hydrogen-assisted degradation of fracture toughness of A333 and A106 pipe steels was studied. The inhibitory effect of carbon monoxide increased with an increase in the carbon monoxide content, and more than 1000 vol. ppm carbon monoxide completely inhibited the hydrogen-assisted degradation of the pipe steels. The inhibitory effect of carbon monoxide was lower in the A106 than in the A333. This result may be due to the difference in the hydrogen sensitivity between the two pipe steels. The inhibitory effect of carbon monoxide was lower than that of oxygen. This result can be theoretically explained by the potential energy of hydrogen, oxygen and carbon monoxide molecules for iron surface.

In Chapter 6, the test results obtained by the above studies were concluded.

26 1.5 References

[1] Hydrogen energy white book, NEDO, (2015).

[2] BP statistical review of world energy 2016, BP, (2016).

[3] Secular change of the annual average temperature of the world (1891~2015), Japan Meteorological Agency, http://www.data.jma.go.jp/cpdinfo/temp/an_wld.html, (2015).

[4] A new international framework on global warming, Ministry of the Environment, (2016).

[5] Energy demand record, Agency for Natural Resources and Energy, (2014).

[6] Strategy of storage battery, Ministry of Economy, Trade and Industry, (2012).

[7] TOYOTA Motor Corporation, Well-wheel analysis of greenhouse gas emission of automotive fuels in the Japanese context, (2004).

[8] W. H. Johnson, On some remarkable changes produced in iron and steel by the action of hydrogen and acids, Proceedings of the Royal Society of London, Vo. 23, pp. 168-179 (1874).

[9] M. Nagumo, Fundamentals of hydrogen embrittlement, Uchida Rokakuho Publishing CO., LTD.

Tokyo, (2008).

[10] Y. Murakami, S. Matsuoka, Y. Kondo, S. Nishiura, Mechanism of hydrogen embrittlement and guide for fatigue design, Yokendo LTD., Tokyo, (2012).

[11] R. P. Gangloff, B. P. Somerday, Gaseous hydrogen embrittlement of materials in energy technologies, Vol. 1 and 2; Eds.; Woodhead Publishing Limited: Cambridge, (2012).

[12] JIGA, Technical standards for compressed hydrogen transport vehicle containers, JIGA-T-S/12/04 (2004).

[13] JIGA, Technical standards for compressed hydrogen transport vehicle accessories, JIGA-T-S/13/04 (2004).

[14] JARI, Technical standard for containers for compressed-hydrogen vehicle fuel device, JARI-S001 (2004).

[15] KHK, Technical standard for compressed hydrogen vehicle fuel containments with maximum filling pressure up to 70 MPa, KHK S0128 (2010).

[16] KHK, High pressure gas safety institute of Japan, KHK S0220, (2010).

[17] JIS, Structure of pressure vessel – general matter, JIS B 8265, (2010).

[18] JIS, Structure of pressure vessel – special standards, JIS B 8265, (2006).

[19] JIS, Design of pressure vessel, JIS B 8267 (2008).

[20] B. P. Somerday, C. San Marchi, K. A. Nibur, Development of methods for evaluating hydrogen

27

compatibility and suitability, International Journal of Hydrogen energy, Vol. 39, No. 35, pp.

20434-20439 (2014).

[21] ASME, ASME boiler & pressure vessel code, (2013).

[22] EN 13445 Unified pressure vessels, (2015).

[23] NASA, Safety standard for hydrogen and hydrogen systems, guidelines for hydrogen system design, materials selection, operation, storage, and transport, NSS 1740. 16 (2005).

[24] K. A. Nibur, B. P. Somerday, C. San Marchi, J. W. Foulk, M. Dadlfarnia, P. Sofronis, G. A.

Hayden, Measurement and interpretation of threshold stress intensity factors for steels in high-pressure hydrogen gas, SANDIA REPORT, SAND2010-4633 (2010).

[25] R. J. Walter, W. T. Chandler, Influence of gaseous hydrogen on metals, final report, National Aeronautics and Space Administration Report CR-124410 (1973).

[26] A. Sievert, The absorption of gasses by metals, Zeitschrift für Metallkunde, Vol. 21, pp. 37–46 (1929).

[27] A. W. Loginow, E. H. Phelps, Steels for seamless hydrogen pressure vessels, Corrosion, Vol. 31, No. 11, pp. 404-412 (1975).

[28] T. Matsumoto, H. Itoga, S. Hirabayashi, M. Kubota, A. Matsuoka, Effect of displacement velocity on elastic plastic fracture toughness of SM490B carbon steel plate in 0.7 MPa hydrogen gas, Transactions of the JSME Series A, Vol. 79, No. 804, pp. 1210-1225 (2013).

[29] M. Yoshikawa, T. Matsuo, N, Tsutsumi, H. Matsunaga, S. Matsuoka, Effects of hydrogen gas pressure and test frequency on fatigue crack growth properties of low carbon steel in 0.1-90 MPa hydrogen gas, Transactions of the JSME, Vol. 80, No. 817, pp. 80-95 (2014)

[30] M. Nagumo, H. Yoshida, Y. Shimomura, T. Kadokura, Ductile crack growth resistance in hydrogen-charged steels, Materials Transactions, Vol. 42, No. 1, pp. 132-137 (2001).

[31] R. Wang, Effects of hydrogen on the fracture toughness of a X70 pipeline steel, Corrosion Science, Vol. 51, No. 12, pp. 2803-2810 (2009).

[32] K. Splichal, J. Berka, J. Burda, M. Zmitko, Fracture toughness of the hydrogen charged EUROFER 97 RAFM steel at room temperature and 120℃, Journal of Nuclear Materials, Vol.

392, No.1, pp. 125-132 (2009).

[33] M. Kubota, N. Noyama, C. Sakae, Y. Kondo, Fretting fatigue in hydrogen gas, Tribology International, Vol. 39, No. 10, pp. 1241-1247 (2006).

[34] M. Kubota, Y. Tanaka, Y. Kondo, The effect of hydrogen gas environment on fretting fatigue

28

strength of materials used for hydrogen utilization machines, Tribology International, Vol. 42, No. 9, pp. 1352-1359 (2009).

[35] M. Kubota, T. Nishimura, Y. Kondo, Effect of hydrogen concentration on fretting fatigue strength, Journal of Solid Mechanics and Materials Engineering, Vol. 4, No. 6, pp. 816-829 (2010).

[36] M. Kubota, K. Kuwada, Y. Tanaka, Y. Kondo, Mechanism of reduction of fretting fatigue limit caused by hydrogen gas in SUS304 austenitic stainless steel, Tribology International, Vol. 44, No. 11, pp. 1495-1502 (2011).

[37] K. Mizobe, Y. Shiraishi, M. Kubota, Y. Kondo, Effect of hydrogen on fretting fatigue strength of SUS304 and SUS316L austenitic stainless steels, Proceedings of JSME / ASME 2011 International Conference on Materials and Processing (2011).

[38] R. Komoda, N. Yoshigai, M. Kubota, J. Furtado, Reduction in fretting fatigue strength of austenitic stainless steel due to internal hydrogen, Advanced Materials Research, Vol. 891-892, pp. 891-896 (2014).

[39] J. Furtado, R. Komoda, M. Kubota, Fretting fatigue properties under the effect of hydrogen and the mechanisms that cause the reduction in fretting fatigue strength, Proceedings of 13th International Conference on Fracture (2014).

[40] R. Komoda, M. Kubota, Characterization of the effect of hydrogen on the microstructure change at the adhered spot during fretting fatigue, Proceedings of Eighth International Symposium on Fretting Fatigue (2016).

[41] R. Komoda, M. Kubota, Y. Kondo, J. Furtado, Fundamental mechanisms causing reduction in fretting fatigue strength by hydrogen (Effect of hydrogen on small crack initiation at the adhered spot), Transactions of JSME Series A, Vol. 79, No. 801, pp. 536-545 (2013).

[42] G. Nelson, J. E. Stein, Gas-Phase Hydrogen Permeation through Alpha Iron, 4130 Steel and 304 Stainless Steel from less than 100℃ to near 600℃, NASA Technical note, NASA TN 0-7265 (1973).

[43] C. San Marchi, B. P. Somerday, X. Tang, G. H. Schiroky, Effects of Alloy Composition and Strain Hardening on Tensile Fracture of Hydrogen-Precharged Type 316 Stainless Steels, International Journal of Hydrogen Energy, Vol. 33, pp. 889-904 (2008).

[44] ISO 14687-2: 2012, Hydrogen fuel – product specification – part 2: proton exchange membrane (PEM) fuel cell applications for road vehicles, (2012).

29

[45] CGA G-5.3, Commodity specification for hydrogen, (2011).

[46] R. S. Yeo, Ion clustering and proton transport in nafion membranes and its application a as solid polymer electrolyte, Journal of The Electrochemical Society, Vol. 130, No. 3, pp. 533-538 (1983).

[47] B. P. Somerday, P. Sofronis, K. A. Nibur, C. San Marchi, R. Kirchheim, Elucidating the variables affecting accelerated fatigue crack growth of steel in hydrogen gas with low oxygen concentrations, Acta Materialia, Vol. 61, pp. 6153-6170 (2013).

[48] J. D. Frandsen, H. L. Marcus, Environmentally assisted fatigue crack propagation in steel, Metallurgical Transactions A, Vol. 8A, pp. 265-272 (1977).

[49] J. H. Holbrook, H. J. Cialone, E. W. Collings, E. J. Drauglis, P. M. Scott, M. E. Mayfield, In gaseous hydrogen embrittlement of materials in energy technologies, Vol. 2; R. P. Gangloff, B.

P. Somerday, Eds.; Woodhead Publishing Limited: Cambridge, 2012; pp.129-153.

[50] K. Kussmaul, P. Deimel, H. Fischer, E. Sattler, Fracture mechanical behavior of the stees 15 MnNi 6 3 in argon and in high pressure hydrogen gas with admixtures of oxygen, International Journal of Hydrogen Energy, Vol. 23, No. 7, pp. 577-582 (1998).

[51] A. Staykov, J. Yamabe, B. P. Somerday, Effect of hydrogen gas impurities on the hydrogen dissociation on iron surface, International Journal of Quantum Chemistry, Vol. 114, pp. 626-635 (2014).

[52] K. Fukuda, Y. Kurono, N, Izumi, J. Sugimura, Influence of trace water and oxygen in a hydrogen environment on pure FE friction and wear, Tribology Online, Vol. 5, No. 2, pp. 80-86 (2010).

[53] K. Fukuda, M. Hashimoto, J. Sugimura, Friction and wear of ferrous materials in a hydrogen gas environment, Tribology On line, Vol. 6, No. 2, pp. 142-147 (2011).

30

2. Development of a new measurement method for relative slip during fretting fatigue test in

hydrogen applying MEMS optical micro-encoder

2.1 Introduction

Fretting is a cyclic small amplitude relative slip motion between contacting surfaces of joined structures subjected to a vibration or fatigue load. Fretting is induced by the difference in the elastic deformation between the contacting bodies. Therefore, the amount of the relative slip is quite small. In fact, there are reports that the fretting with a 1-m relative slip range creates fretting fatigue cracks [1, 2].

Fretting fatigue properties are strongly affected by the relative slip range between the contacting surfaces [2-4]. Physical meaning of “slip” associated with fretting fatigue crack initiation is still not fully understood, however, the role of the slip can be interpreted in association with friction. At the small relative slip range, the tangential force acting on the contacting surfaces increases with an increase in the relative slip range. As a result, the fretting fatigue strength decreases with the increase in the relative slip range. At the middle relative slip range, the fretting fatigue strength becomes constant independently of the relative slip range because the increase in the tangential force is saturated and takes a constant value. At the large relative slip range, the fretting fatigue strength changes to increase with an increase in the relative slip range because a large amount of fretting wear relieves the stress concentration at the contacting part and removes small cracks.

Measurement of the relative slip range is important to characterize fretting fatigue properties and understand the mechanisms, however, there are some difficulties for the measurement of the relative slip during fretting fatigue test. Since the relative slip range during the fretting fatigue is at most several tens m, and sometimes less than 1 m, a high resolution is required for the measurement. In addition, the relative slip range is equivalent to an elastic deformation around the contact edge. Therefore, the elastic deformation included in the measured relative slip range should not be ignored as follows. When the relative slip between contacting bodies is measured, reference points are needed to each body. Figure 2.1 shows a schematic of the measurement error due to elastic deformation. When the reference point on the specimen is placed just at the contact edge so that two reference points aligned with the contact edge (points O and A), the measured value is in

31

agreement with the true relative slip at the contact edge (O-A’). On the other hand, when the reference point on the specimen is placed with some distance from the contact edge (point B), the measured relative slip includes an elastic deformation of the specimen between the points A’ and B’.

However, sensors have a certain dimension, the setting of the specimen side reference point tends to be set at B. Again, the relative slip range is the same level as the amount of elastic deformation. Thus, the error due to the elastic deformation should not be ignored. Therefore, the reference points for the relative slip measurement should be placed as close as possible to the contact edge in order to measure the relative slip accurately.

Fig. 2.1 Error in the measurement of relative slip range of fretting due to elastic deformation around the reference points

Tensile load O

O A

A’

B

B’

Elastic deformation between A’ and B’

True relative slip Measured value

= true relative slip

Measured value

32

For an accurate measurement of the relative slip range during fretting fatigue, a variety of measurement methods have been developed. For example, small displacement sensors [5-7], synchronized laser displacement sensors [8] and a digital image correlation method [9] and so on are found in the literature.

For the fretting fatigue in hydrogen environment, the elucidation of the mechanisms why the fretting fatigue strength is significantly lower than that in air [10-16] is necessary from the viewpoint of relative slip behavior. It is considered that one of the reasons for the reduction in the fretting fatigue strength in hydrogen is local adhesion between the contacting surfaces and subsequent many small crack initiations at the adhered spot [14]. When considering the local adhesion between the contacting surfaces during the fretting fatigue in hydrogen, the relative slip range must be reduced, and it could result in changes in the stress condition at the contacting surface.

Therefore, identification of the slip condition during fretting fatigue in hydrogen environment is necessary to achieve the deeper understanding of the mechanism of the reduced fretting fatigue strength in hydrogen.

In addition to the accuracy of the measurement depending on the sensor setting, there are some barriers for the measurement in hydrogen environment. For example, the output of the sensor element drifts due to hydrogen absorption into the sensor material [17], and the space inside the hydrogen chamber is too narrow for commercial sensors.

Incidentally, micro electro mechanical system (MEMS) technology has been developed this 50 years [18], and it is applied to a variety of devices, such as sensors, actuators etc. Prof.

Sawada in Kyushu University studies and develops small sensors fabricated by MEMS technology.

There is an optical micro encoder in one of the sensors Prof. Sawada developed. Since this optical micro encoder has a high accuracy and the size is small, it was considered that the relative slip range during fretting fatigue in hydrogen can be measured accurately by using the optical micro encoder.

In this context, a new method for the measurement of the relative slip range during the fretting fatigue test in hydrogen was developed in this study by appling the MEMS optical micro-encoder in collaboration with Prof. Sawada’s laboratory. Especially, this study was carried out with the great help of Mr. Morita who is a member of Prof. Swada.

33 2.2 MEMS optical micro-encoder

An optical micro-encoder fabricated by MEMS technology [19] was used for the relative slip sensor in this study. This optical micro-encoder was provided by Professor Sawada in Kyushu University. The structure and working principle of the optical micro-encoder are shown in Fig. 2.2.

The size of the micro-encoder chip is 2.8 mm square and 1.0 mm in thickness. The micro-encoder consists of two layers, which are a silicon layer and glass layer. The silicon layer has laser diodes (LDs) mounted on the bottom of the silicon layer and Au mirrors mounted on the slant sidewall of the silicon layer. The glass layer has two photodiodes (PDs), a phase shifter and two micro aspherical lenses. Two laser beams are oppositely emitted from the LDs. The laser beams are reflected by the Au mirrors, and then collimated by pathing thorough the micro-aspherical lens. Each beam runs slantingly toward the diffraction grating scale. After the laser beams are diffracted at the grating scale, the two diffracted beams run parallel to each other at right angles to the grating scale and interfere on the PD. As the result of the interference, an optical intensity detected by the PD changes depending on the relative displacement between the grating scale and micro-encoder as shown in Fig. 2.3. The relative displacement can be correlated with the wavenumber of the PD output. The wavenumber corresponds to 1.6 m of the relative displacement between the micro-encoder chip and grating scale in this study. This is substantially sufficient to measure the

m-order relative slip range in the fretting fatigue test. Since the inside of the micro-encoder chip is separated from the external environment, this MEMS micro-encoder is not affected by the hydrogen environment. The environmentally pressure capacity of this encoder chip is 1.0 MPa. The micro-encoder chip is mounted on a pre-amp circuit as shown in Fig. 2.2 (b). The size of the pre-amp circuit board is 14 mm in length, 12 mm in width and 2 mm in thickness.

34

Fig. 2.3 Relationship between sensor output and relative displacement Time

Sensor outputDisplacement 1.6 m

Line symmetry

Reverse motion

Fig. 2.2 MEMS micro-encoder used for micro-relative displacement measurement (c) Structure and working principle of optical MEMS sensor

LD

Glass layer

Si layer Through-silicon via

Au mirror

PD Phase shifter

Lens

 

Laser

Grating scale

First-order diffraction beam

Displacement (b) Pre-amp circuit 5 mm

Micro-encoder chip

(a) Micro-encoder chip Electrodes for

PDs

PD(A) PD(B) Lens

1 mm

35

2.3 Measurement method of relative slip range during fretting fatigue

Figure 2.4 shows the setting of the MEMS sensor on the fretting fatigue test specimen. The fretting fatigue test apparatus was the same as that is shown in chpter 3. The loading frequency was 20 Hz and the stress ratio was -1. The test environment was air and hydrogen. The oxygen and water vapor content in the hydrogen was 0.088 vol. ppm and 0.89 vol. ppm, respectively. The hydrogen pressure was 0.1 MPa in gauge pressure. The material of the specimen and contact pad, which was SUS304 austenitic stainless steel, was the same as that in chapter 3.

Fig. 2.4 MEMS sensor setting for fretting fatigue test Specimen

Contact load Fatigue load

Bar spring Contact pad

Tightening bolt

MEMS optical encoder and grating scale

36

Figure 2.5 shows the details of the installation of the MEMS micro-encoder to the fretting fatigue testing setup. The pre-amp circuit with the micro-encoder chip was installed in the specimen using a T-shaped holder. The T-shaped holder was attached on the extension line of the contact edge with 1 mm × 2 mm of contact area. The grating scale was installed in the contact pad using an L-shaped holder. The L-shaped holder was attached to the contact edge face with 2 mm × 2 mm of contact area.

(b) Installing of encoder holder Contact area (1 mm × 2 mm) 1 2

Contact pad

Specimen Encoder holder

Fig. 2.5 Setup of MEMS slip sensor

(c) Installing of grating scale holder Contact area (2 mm × 2 mm) 2 2

Contact pad

Grating scale holder Grating scale (a) Installation position of micro-encoder and grating scale

Micro-encoder chip Diffraction grating scale

Contact pad Specimen

Contact pad

Pre-amp circuit

Fatigue load

Fatigue load

Contact load

Contact load

37 2.4 Correction of the error due to elastic deformation

The output of the MEMS slip sensor includes the relative displacement between the specimen and contact pad due to elastic deformation around the material sensor attached. This component of the elastic deformation should be removed from the measured value to obtain the true value of the relative slip range. In this study, the error due to the elastic deformation was experimentally calibrated prior to the relative slip measurement.

For the calibration, a special specimen, which has the integrated shape of the specimen and contact pad as shown in Fig. 2.6, was used. Since no relative slip occurs in this integrated specimen, the sensor output of this specimen can be regarded as the elastic deformation of the material around the sensor attached. To verify this experimental calibration, the amount of the elastic deformation was also obtained computationally by a finite element (FE) analysis.

Fig. 2.6 Integrated specimen for calibration of elastic deformation 36

A 20

10

44

10

100

34.4 6 12.5

4-8.2

5 14

34.4

12.5 6 3

Detail of A

0.5

38

Figure 2.7 shows the results of the measurement of the elastic deformation and FE analysis.

The amount of the elastic deformation linearly increased with an increase in the stress amplitude. To verify the reproducibility of the measurement, the measurement of the elastic deformation was duplicated twice. The results of the two measurements well agreed with each other, and the measurements also agreed with the result of the FE analysis. Based on these results, the true relative slip was defined as the subtracted value of the elastic deformation from the sensor output in this study.

0 100 200

0 5 10 15 20

Stress amplitude, _a (MPa) Elastic deformation, Se (m) SUS304

R = -1 f = 20 Hz p_c = 100 MPa

S_e = 0.0271_a Experimental FE analysis

Fig. 2.7 Measurement of the elastic deformation of the material around sensor

39

2.5 Measurement of relative slip range during fretting fatigue in hydrogen

Figure 2.8 shows one of the measurement results of the relative slip range during the fretting fatigue test in the high-purity hydrogen. The stress amplitude was 120 MPa. In this case, 3.25 waves were included in one cycle of the fatigue load. Therefore, the true relative slip range (1.9

m) was calculated by the following equation.

1.9 m = 3.25 cycles × 1.6 m/cycle – 0.0271 m/MPa × 120 MPa (2.1)

Fig. 2.8 Sensor output measured during fretting fatigue test in hydrogen at a = 120MPa and f = 20 Hz (True relative slip range was 1.9 m)

-200 -100 0 100 200

Fatigue load (MPa)

0 0.02 0.04 0.06 0.08

-0.1 0 0.1

Time (s)

Output (V)

Load increasing Load decreasing

40

Figure 2.9 shows the relative slip range measured by the MEMS optical encoder during the fretting fatigue test in air and hydrogen. The relative slip range in the hydrogen was significantly lower than that in air. During the fretting fatigue in hydrogen, local adhesion occurred between the contacting surfaces [14]. The local adhesion prevents relative slip between contacting surfaces, however, plastic deformation of the adhered spots arises. Therefore, the relative slip range in the hydrogen was reduced, but not zero. The result of the measurement can be reasonably explained by the local adhesion during the fretting fatigue test in hydrogen.

0 100 200

0 5 10

Stress amplitude, _a (MPa)

Relative slip range,S (m)

In laboratory air

In high-purity hydrogen

(0.088 vol. ppm O₂, 0.89 vol. ppm H₂O) SUS304

R = -1 f = 20 Hz p_c = 100 MPa

Fig. 2.9 Relative slip range during fretting fatigue in hydrogen and air measured by the optical MEMS sensor

41 2.6 Conclusions

To realize an accurate measurement of the relative slip range during fretting fatigue in hydrogen, a new measurement method applying MEMS optical encoder was developed. The results are as follows:

1. The system and procedure of the measurement of the relative slip range during the fretting fatigue test in hydrogen were established.

2. The relative slip range in the hydrogen is significantly lower than that in air. It was presumed that the reduction was caused by the local adhesion between the contacting surface produced during the fretting fatigue in hydrogen.

42 2.7 References

[1] R. Komoda, M. Kubota, Y. Kondo, J. Furtado, Effect of oxygen addition on fretting fatigue strength in hydrogen of JIS SUS304 stainless steel, Tribology International, Vol. 76, pp. 92-99 (2014).

[2] J. Ding, D. Houghton, E. J. Williams, S. B. Leen, Simple parameters to predict effect of surface damage on fretting fatigue, International Journal of Fatigue, Vol. 33, No. 3, pp. 332-342 (2011).

[3] K. Nishioka, K. Hirakawa, Fundamental investigation of fretting fatigue – Part 5. The effect of relative slip amplitude, Transaction of the JSME, Vol. 34, No. 268, pp. 2068-2073 (1968).

[4] Y. Mutoh, T. Nishida, I. Sakamoto, Effect of relative slip amplitude and contact pressure on fretting fatigue strength, Journal of The Society of Materials Science, Japan, Vol. 37, No. 417, pp. 649-655 (2009).

[5] Y. Kondo, M. Bodai, Study on fretting fatigue crack initiation mechanism based on local stress at contact edge, Transaction of the JSME, Series A, Vol. 63, No. 608, pp. 669-676 (1997).

[6] B. U. Wittkowsky, P. R. Birch, J. Dominguez, S. Suresh, An apparatus for quantitative fretting fatigue testing, Fatigue & Fracture of Engineering Materials & Structure, Vol. 22, No. 4, pp.

307-320 (1999).

[7] Y. Kondo, C. Sakae, M. Kubota, T. Nagasue, S. Sato, Evaluation of fretting fatigue limit based on local stress at the contact edge, Journal of The Society of Materials Science, Japan, Vol. 51, No. 9, pp. 1017-1022 (2002).

[8] Y. Kondo, M. Bodai, Critical slip for fretting fatigue microcrack initiation, Transaction of the JSME, Series A, Vol. 63, No. 610, pp. 1178-1182 (1997).

[9] J. De Pauw, W. De Waele, R. Hojjati-Talemi, P. De Baets, On the use of digital image correlation for slip measurement during coupon scale fretting fatigue experiments, International Journal of Solids and Structure, Vol. 51, No. 18, pp. 3058-3066 (2014).

[10] M. Kubota, N. Noyama, C. Sakae, Y. Kondo, Fretting fatigue in hydrogen gas, Tribology International, Vol. 39, No. 10, pp. 1241-1247 (2006).

[11] M. Kubota, Y. Tanaka, Y. Kondo, Fretting fatigue strength of SCM435H steel and SUH660 heat-resistant steel in hydrogen gas environment, Tribotest, Vol. 14, No. 3, pp. 177-191 (2008).

[12] M. Kubota, Y. Tanaka, Y. Kondo, The effect of hydrogen gas environment on fretting fatigue strength of materials used for hydrogen utilization machines, Tribology International, Vol. 42, No. 9, pp. 1352-1359 (2009).

43

[13] M. Kubota, T. Nishimura, Y. Kondo, Effect of hydrogen concentration on fretting fatigue strength, Journal of Solid Mechanics and Materials Engineering, Vol. 4, No. 6, pp. 816-829 (2010).

[14] M. Kubota, K. Kuwada, Y. Tanaka, Y. Kondo, Mechanism of reduction of fretting fatigue limit caused by hydrogen gas in SUS304 austenitic stainless steel, Tribology International, Vol. 44, No. 11, pp. 1495-1502 (2011).

[15] R. Komoda, N. Yoshigai, M. Kubota, J. Furtado, Reduction in fretting fatigue strength of austenitic stainless steel due to internal hydrogen, Advanced Materials Research, Vol. 891-892, pp. 891-896 (2014).

[16] R. Komoda, M. Kubota, J. Furtado, Effect of addition of oxygen and water vapor on fretting fatigue properties of an austenitic stainless steel in hydrogen, International Journal of Hydrogen Energy, Vol. 40, No. 47, pp. 16868-16877 (2015).

[17] T. Matsuo, J. Yamabe, H. Fukuyama, K. Seki, S. Watanabe, S. Matsuoka, Development of new strain gauge for high-pressure hydrogen gas use, Experimental Mechanics, Vol. 54, pp. 431-442 (2014).

[18] H. C. Nathanson, W. E. Newell, R. A. Wickstrom, J. R. Davis, The resonant gate transistor, IEEE Transactions on Electron Devices, Vol. 14, No. 3, pp. 117-133 (1967).

[19] E. Higurashi, D. Chino, T. Suga, R. Sawada, Au-Au surface-activated bonding and its application to optical microsensors with 3-D structure, IEEE Journal of Selected Topics in Quantum Electronics, Vol. 15, pp. 1500-1505 (2009).

44

3. Effects of addition of ppm-levels of oxygen and water vapor on fretting fatigue properties in

hydrogen

3.1 Introduction

Fretting fatigue is a composite phenomenon of fatigue and frictional contact which may occur at the contact part of the mechanical component subjected to a vibration or alternating load.

When fretting fatigue occurs, the fatigue strength at the contact part is significantly reduced compared to that of the smooth specimen. The reduction in the fatigue strength due to fretting can reach 1/30 of the fatigue limit of a smooth specimen [1]. Therefore, there are many examples of fretting fatigue failure of machines and structures [2-5]. In this regard, fretting fatigue is one of the most important factors in the design of structural members involving contact parts.

For pressurized hydrogen containment systems, fretting fatigue is also important similar to conventional machines and structures because components in gas system, such as valve, joint, seal, etc., have many contact part. In addition, hydrogen can affect both fatigue [6-8] and wear properties [9]. In fact, fretting fatigue strength in hydrogen is significantly lower than that in air [10-16].

Therefore, to ensure the safety of pressurized hydrogen containment systems, characterization of fretting fatigue strength in hydrogen and elucidation of the mechanisms by which hydrogen affects fretting fatigue strength are required.

One of the reasons for the reduction in the fretting fatigue strength in hydrogen is local adhesion between the contacting surfaces [14]. Usually, adhesion between the contacting surfaces during fretting fatigue in air only occurs at the microscopic level, such as the real contact area, and cannot be recognized at the macroscopic level, such that specimen and contact pad are adhered together. This is because the accumulation of oxidized fretting wear particles between the contact surfaces separates the contact surfaces. On the other hand, oxidized fretting wear particles are not produced in hydrogen since no oxygen exists in hydrogen environment, and consequently the separation of the contacting surfaces is prevented. As the result, adhesion between the contacting surfaces occurs over a wide portion of the nominal contact area.

The enhancement of adhesion during fretting fatigue in hydrogen closely relates to the reduction in the fretting fatigue strength in hydrogen. At the adhered spot, severe cyclic plastic deformation occurs [18], and it causes many small cracks at the adhered spot [14]. In addition,

45

hydrogen accelerates the early crack propagation under fretting conditions [16]. In the case of austenitic stainless steels, one of the reasons for the hydrogen-enhanced fretting fatigue crack initiation is strain-induced martensitic transformation at the adhered spot by the severe cyclic plastic deformation due to the adhesion [18]. In this way, the local adhesion takes a key role in the reduction in the fretting fatigue strength in hydrogen.

As described above, the local adhesion in hydrogen is caused by no oxide production due to the absence of oxygen in the hydrogen environment. Therefore, when some amount of oxygen or other impurities that can produce wear particles exists in the hydrogen environment, the impurities may affect the adhesion behavior, and it can affect the fretting fatigue strength in hydrogen. If the addition of impurities changes the fretting fatigue strength in hydrogen, the effect of impurities in a hydrogen environment on the fretting fatigue properties should be elucidated in order to enable the safe design of contact parts in pressurized hydrogen containment systems. This study focused on the effects of ppm-level oxygen and water vapor were focused on, and the effects of addition of ppm-level oxygen and water vapor were characterized. For this purpose, a new test method, which can precisely control the oxygen level or humidity in the hydrogen gas, was established. Then, fretting fatigue tests were carried out in different hydrogen environments containing impurities.

46 3.2 Test procedure

3.2.1 Material

The test material was an austenitic stainless steel JIS SUS304. The chemical composition is shown in Table 3.1. The material was solution-heat treated at 1303 K for 3.9 ks followed by rapid cooling. The mechanical properties are shown in Table 3.2.

Table 3.2 Mechanical properties and hardness

% proof stress

0.2 (MPa)

Ultimate tensile strength

B (MPa)

Elongation

(%)

Reduction of area

(%)

Vickers Hardness, HV1

237 771 72.3 81.2 174

Table 3.1 Chemical composition (mass %)

C Si Mn P S Ni Cr

0.05 0.53 0.81 0.026 0.006 8.10 18.19