Sunteți pe pagina 1din 17

Strength, Fracture and Complexity 2 (2004) 95109 IOS Press

95

The mechanism of hydrogen embrittlement on the basis of the interaction model of dislocations and hydrogen around a crack tip
A.T. Yokobori, Jr. a, , S. Taketomi a , J.C. Ha c , T. Shoji b and G.F. Li b
Department of Nano Mechanics, Graduate School of Engineering, Tohoku University, Aoba 6-6-01, Aramaki, Aoba-ku, Sendai-shi #980-8579, Japan b Fracture Research Institute, Graduate School of Engineering, Tohoku University, Aoba 6-6-01, Aramaki, Aoba-ku, Sendai-shi #980-8579, Japan c Faculty of Science and Engineering, Teikyo University, Utsunomiya, Toyosatodai 1-1, Utsunomiya-shi #320-8551, Japan
Received 29 November 2004 Abstract. Mechanisms of dissolvent anodic chemical reaction and hydrogen embrittlement were proposed as stress corrosion cracking mechanics (SCC). The former is feasible for the case of plastic deformation dominant metals and the latter is for high strength metals such as high strength steels. However, in spite of low yield stress, a discontinuous cleavage-like fracture is sometimes observed during SCC for ductile fcc alloys. In this paper, we proposed stress corrosion cracking model on the basis of interaction of dislocation and hydrogen around a crack tip to predict discontinuous cleavage-like fracture during SCC for ductile fcc alloys. Furthermore, we conducted numerical analyses using this proposed model. Keywords: Stress corrosion, hydrogen embrittlement, dislocation dynamics, fcc alloys, cleavage-like fracture
a

1. Introduction Hydrogen embrittlement occurs for brittle materials with high yield stress [14]. For these materials, inter-granular cracking is dominant. The analyses of the hydrogen diffusion and concentration have been conducted to clarify hydrogen embrittlement [513]. On the other hand, the ductile material with low yield stress is dominated by plastic deformation induced dissolvent and chemical reaction [14] and inter-granular cracking is also dominant [15]. However, under corrosive condition even for ductile materials, there are some cases which show subcritical cleavage fracture for example such as SCC for a 316L alloy under NaOH solution and a theoretical analysis has been conducted [16]. However, in this paper [16], the criterion for cleavage fracture is not feasible to discriminate from that for ductile fracture based on dislocation theory [17]. Under static loading condition, theoretical analysis on brittle fracture was conducted based on the micromacro combined fracture mechanics [18]. It showed that a trigger point mainly exists and it is caused at the tip of dislocation pile up near a crack tip.
*

Corresponding author. E-mail: yokobori@md.mech.tohoku.ac.jp.

1567-2069/04/$17.00 2004 IOS Press and the authors. All rights reserved

96

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

Fig. 1. The comparison of the relationship between dynamical fracture toughness and trigger points experimentally obtained with that between dynamical fracture toughness and the length of dislocation free zone obtained by numerical analyses.

Under the dynamical loading, brittle fracture with cleavage fracture surface is also found to usually occur at the trigger point apart from a crack tip [19]. It is caused by inverse pile-up with high dislocation density [20,21]. The effect of applied stress rate on the trigger point of brittle fracture was found to well correlate with that of dislocation free zone obtained by numerical analysis as shown in Fig. 1 [21]. This result shows that brittle fracture with cleavage fracture surface is caused by inverse pile-up with high dislocation density at the end of dislocation free zone. This law does not contradict the fact that the trigger point is observed at nonmetallic carbides [22], because suitable nonmetallic carbides will be selected as a trigger point according to the law as shown in Fig. 1 [21]. However, a ductile material itself usually has not such inverse pile-up with high dislocation density [20]. In this paper, to clarify the mechanism of cleavage fracture for ductile materials under corrosive condition, we investigated the mechanical behavior of inverse pile-up of dislocation under corrosive condition. Furthermore, the effect of interaction of dislocation groups and hydrogen originated by chemical dissolvent anodic reaction near a crack tip on the occurrence of cleavage fracture was claried.

2. Physical model and basic equations Initially, dislocation emission from a crack tip occurs which results in pre-plastic deformation and it products a site of active path where an anodic chemical dissolvent reaction under the stress corrosion environment is induced. The hydrogen is induced by this reaction as shown Fig. 2 [23]. The hydrogen diffuses under elasticplastic stress eld interacting with the nearest dislocation groups emitted by a successive process of dislocation emission from a stressed source near a crack tip. Previously, a corresponding physical model around a crack tip was proposed as shown in Fig. 3 [24]. In this model, hydrogen was originated by chemical reaction and exists as a hydrogen cluster. Concerning slip bands, multiple slip bands will usually be originated from a crack tip. However, previously, the effect of multiple slip bands on the stress concentration as shown in Fig. 4 was theoretically analyzed. And it was compared with that of single slip band. These results show that however, the former is quantitatively different from the latter, qualitatively, they have almost same characteristics [25,26]. Therefore, one slip

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

97

Fig. 2. Chemical dissolutive reaction process around a crack tip. p is hydrostatic stress.

Fig. 3. Physical model of the interaction between dislocation groups and hydrogen around a crack tip.

band originated from a crack is physically reasonable to consider the mechanical behavior of emission and movement of dislocation groups. Furthermore, in this paper, we simplify this model into that as shown in Fig. 5 for numerical analysis. The model in Fig. 5 is essentially different from that in Fig. 3, that is, for example, types of mode I and mode II. However, these two dislocation models essentially become identical. It was assured by theoretical analyses on the static model as shown in Figs 6 and 7. In these analyses, the formula of stress intensity factor obtained by theoretical analysis in Fig. 6 [27] was found to have qualitatively same type of equations as that in Fig. 7 by Eqs (1) and (2) [28]. ag12 i , 2 i , KIIs = ag31 + ag33 2 KIn = ag11 + (1) (2)

where, KIn , KIIs are stress intensity factors at the tips of a crack and a slip band. is applied tensile stress which results in shear stress = /2 on a slip line. a is half length of a crack, gij is interacting factor as a function of d/h.

98

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

Fig. 4. Physical model of multiple slip bands originated from a crack tip.

Fig. 5. Approximated physical model of the interaction between dislocation groups and hydrogen around a crack tip.

Fig. 6. Physical model of the interaction between a crack and a coplanar slip band.

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

99

Fig. 7. Physical model of the interaction between a crack and an inclined slip band.

Concerning dislocation group dynamics emitted from a stressed source, positive and negative dislocations are emitted from the source and move in the opposite direction each other. Therefore, not only the dynamic behavior of positive dislocations but also that of negative dislocations should be noticed. Previously, the effect of negative dislocations on the mechanical behavior of positive dislocations emitted from a stressed source was investigated by numerical analysis [29]. This result shows that the effect of the mechanical behavior of negative dislocations on that of positive dislocations qualitatively does not exist and is quantitatively small. Furthermore, for the case of considering the effects of three stress singularities such as crack tip, dislocations and interactive force between dislocations and hydrogen, it is necessary to simplify the physical model from the view point of accuracy of numerical analysis. On the basis of these facts, a physical model is proposed as shown in Fig. 5. In this model a dislocation source, S is located at some nite distance of xs from the crack tip. In this model, the size of xs is taken as xs /a = 106 , where a is half of the crack length. Thus, xs amounts to order of 106 mm in the case of the a (half of the crack length) being the order of 1.0 mm. In such case, xs corresponds to the order of the core cut off of the dislocation. Therefore, the source is regarded to be located at the crack tip from the viewpoint of continuum mechanics. The hydrogen was located at the site (xh , ). Where is the range of Cottrell atmospheric distance between a atom and a solute atom [30]. Under the actual condition, hydrogen will exist in the form of hydrogen cluster at xh . Analyses were conducted for various value of xh . In this paper, hydrogen cluster was replaced by a super hydrogen and we call this a hydrogen. On the basis of this model, the analyses of the mechanical interaction between crack tip stress eld, dislocation groups and hydrogen cluster were conducted, and the behavior of dislocation distribution was analyzed. As the stress distribution near a crack tip, the stress singularity near a crack tip was taken into account under constant rate of stress application as follows, a (x, t) = t a +1 , x (3)

where = the increasing rate of stress application; t = time; x = the distance from the crack tip. Con cerning the Eq. (1), a (x, t) approaches to applied stress with increase in x. Stress singularity of 1/ x

100

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

becomes more eminent with decrease in x. Therefore, Eq. (3) is approximately valid as the macroscopic stress eld around the crack tip. The relationship between effective stress eff,i and dislocation velocity Vi is written for each individual dislocation in a coplanar array of dislocation groups by the following equation.
m Vi = M eff,i

(4)

where eff,i = the effective stress exerted on each individual dislocation in terms of applied shear stress; m = materials constant; and i = index of dislocation in order of emitting from the source. M is written as M = V0 1 0
m

(5)

where 0 = specic shear stress required to move an isolated dislocation with a velocity of 1 cm/s; and V0 = 1 cm/s. In this analysis, the effective stress exerted on the ith dislocation is written by Eq. (6). n n A a 1 1 +1 + A A xi x xj 2xi xi + xi xj j=1,i=j i j=1,i=j

eff,i = t where A=

2A (xi xh ) , [(xi xh )2 + 2 ]2

(6)

b/2(1 ) for edge dislocation, b/2 for screw dislocation.

= shear modulus; = Poissons ratio; xi = the position of the ith dislocation; and A = the constant which represents the magnitude of the interaction force between hydrogen and dislocations. The rst term of the right-hand side of Eq. (6) is macroscopic stress eld around a crack tip by applied stress. The second term is interaction stress between dislocations in the dislocation arrays, and the third and fourth terms are the image force of dislocations by the free boundary of crack surface [31]. In this model, since macroscopic crack was considered and stress eld of dislocation falls under a category of short range stress eld, the condition of a xi is satised. Therefore, the third and fourth terms in Eq. (6) is valid to use for numerical analyses [31]. The fth term is interaction stress between dislocations and hydrogen [30]. The effective stress, eff,s exerted on a dislocation source is written as follows, eff,s = t
n n a 1 1 A +1 + A A xs x xj 2xs xs + xs xj j=1 s j=1

2A (xs xh ) . [(xs xh )2 + 2 ]2 (7)

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

101

In this analyses the non-dimensional variables were used as follows: si = xi l and sa = a , l (8)

where l = 0.01 mm. When eff,s equals to the source activation stress, s , then a new mobile dislocation is introduced at the source. Material constants m and 0 for pure iron were used [32]. The detail analytic methods were written by previous literatures [33,34]. 3. Results of analysis Numerical results of dislocation distribution under the interaction between a crack, dislocation groups and hydrogen cluster are shown in Fig. 8 for pure iron. Dislocation density, fd (xi+1/2 ) is obtained by the equation of fd (xi+1/2 ) = 1/(Xi Xi+1 ). For ductile materials such as pure iron, the maximum dislocation density at the end of the dislocation free zone (DFZ) is not so high as shown by the dotted line in Fig. 8. However, when hydrogen cluster exists near the slip line, the inverse pile-up with high dislocation density occurs at the site of hydrogen cluster, as shown by the solid line in Fig. 8. The effect of the site of hydrogen on dislocation distribution is shown in Fig. 9. The site of maximum dislocation density changes according to the site of hydrogen cluster. This will cause cleavage fracture due to the dislocation pile-up. That is, when hydrogen cluster is originated by dissolvent chemical reaction at the corrosive point induced by dislocation emission near a crack tip, the sub-critical cleavage fracture can be observed even for such ductile materials due to the interaction between hydrogen cluster and dislocation groups which results in the dislocation local pile-up at the site of hydrogen cluster. The effect of site of hydrogen on the maximum dislocation density, fd max was shown in Fig. 10. Corresponding to the site of hydrogen cluster, the position of maximum dislocation density, fd max moves along the dislocation slip line as shown in Fig. 10. The relationship between fd max and the length of dislocation free zone, SDF was shown in Fig. 11. From these results, qualitatively similar characteristics were shown

Fig. 8. Numerical result of dislocation density due to the interaction of dislocations and hydrogen.

102

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

Fig. 9. Effect of the site of hydrogen on dislocation distribution.

Fig. 10. The characteristics of fd max when the site of hydrogen moves to along the dislocation slip line.

both for the case without hydrogen and that with the interaction of dislocations and hydrogen. However, even for ductile material such as pure iron, the absolute value of fd max with the interaction of dislocations and hydrogen is much larger than that for the case of brittle material without hydrogen such as Ge. This means cleavage brittle fracture occurs even for ductile materials when the interaction between dislocation and hydrogen is involved in the fracture mechanism. The magnied ratio of fd max for the case with the interaction of dislocation and hydrogen to that without hydrogen was shown in Fig. 12. These results show that the ratio takes larger value for a ductile material (high dislocation mobility) such as an edge dislocation for pure iron than that for brittle material (low dislocation mobility) such as Ge and a screw dislocation for pure iron.

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

103

Fig. 11. The relationship between fd max and the length of dislocation free zone.

Fig. 12. The magnied ratio of fd max for the case with the interaction of dislocation and hydrogen to that without hydrogen.

The numerical results of distribution of effective stress exerted on each dislocation, eff,i (x), velocity, Vi and the dislocation distribution density, fd (x) in the array obtained by our analysis were shown in Figs 13, 14 and 15. Concerning eff,i (x) and Vi , with increase in applied stress rate, the characteristic of decreasing in eff,i (x) and Vi around the trapped region of dislocations by hydrogen becomes remarkable and it localizes. Correspondingly, concerning the dislocation distribution density, fd , the maximum value of dislocation density, fd max , which appears in the trapped region of dislocations by hydrogen increases and its abrupt increasing characteristic becomes remarkable. This characteristic shows the localization of dislocations in fd max region is promoted with increase in applied stress rate, which induces cleavage fracture. The effective scale on the dislocation localization due to hydrogen equals to the order of x (x = 5 m).

104

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

Fig. 13. The distribution of effective stress exerted on each dislocation under interaction with other dislocations and hydrogen.

Fig. 14. The distribution of dislocation velocity in the dislocation array under interaction with other dislocations and hydrogen.

The analysis of effective stress eld caused by the interaction between single dislocation and hydrogen has been conducted [35]. The result obtained by this analysis [35,36] is qualitatively in agreement with that obtained by our analysis on the dislocation groups dynamics as shown in Figs 1315. However, quantitatively, effective scale on the dislocation localization of the former analysis equals to the order of burgers vector (3 107 mm) and the latter equals to the order of x = 5 m which is mezzo scale

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

105

Fig. 15. Detailed dislocation pile-up behavior in the dislocation array around hydrogen with parameter of applied stress rate.

Fig. 16. The effect of the site of hydrogen for the case when the hydrogen is located at a crack tip (lower rate of hydrogen diffusion), (a) and ahead of the leading dislocation (higher rate of hydrogen diffusion), (b).

order as are shown in Figs 1315. That is, about four orders difference in scale exists between them. Our analysis concerns the mechanical interaction between a crack, dislocation groups and hydrogen which is a dominant factor on the occurrence of cleavage fracture affected by variant feature of dislocation distribution density, effective stress and their affective scale in the dislocation array. This affective scale treated in this paper equals not to nano scale but to mezzo scale which closely concerns the dominant zone of fracture such as a trigger point of cleavage fracture which closely concerns results obtained by this paper. For each case of Sh = 0.0 and 7.6 which corresponds to the lower or higher rate of hydrogen diffusion, the inverse pile-up with high dislocation density does not occur and the characteristics of dislocation distribution equal to the case without hydrogen as shown in Figs 16(a),(b) and also Fig. 8 represented

106

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

by the dotted line. Results shown in Figs 8 and 16 show that the inverse pile-up with high dislocation density occurs when dislocation groups have competitive interaction with hydrogen cluster.

4. Consideration For ductile materials with lower yield stress, since the gradient of hydrostatic stress eld, p which derives hydrogen to the maximum stress eld becomes smaller and a hydrogen does not penetrates into the metal [12,13], the dislocation distribution is not affected by hydrogen as shown in Figs 16(a) and 17(a) and anodic dissolvent mechanism will be dominant at the corrosive active point. For brittle materials with higher yield stress, since the gradient of P becomes larger and hydrogen gets ahead of dislocation groups [12,13], the dislocation distribution is also not affected by this hydrogen as shown in Figs 16(b) and 17(c) and hydrogen embrittlement will be dominant due to hydrogen concentration at the site of maximum P [12,13]. These characteristics were in good agreement with previous experimental behaviors respectively [14,14,15]. When the rate of hydrogen diffusion to the maximum value of P has comparable value with the rate of dislocation ow, inverse pile-up with high dislocation density occurs as shown in Figs 8 and 17(b) and cleavage fracture from a trigger point where the site of maximum dislocation density appears will be dominant. This phenomenon occurs for ductile materials with large yield stress.

Fig. 17. The interaction behaviors of dislocation groups and hydrogen for ductile to brittle materials. p is hydrostatic stress.

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

107

Gao et al. investigated the fracture mechanism in ductile material under corrosive environment condition by the TEM observation [37]. He showed a micro-crack is initiated within dislocation free zone and the fracture occurs in cleavage manner. In this paper, the cleavage fracture is found to be occurred by the inverse pile-up of dislocations even for ductile materials when hydrogen exists in the dislocation group. From these results mentioned above, it is shown that sub-critical cleavage fracture for a ductile material under corrosive environment condition is a characteristic property by competitive interaction between hydrogen and dislocation groups around a crack tip. It will occur in a ductile material with rather large yield stress. This result will correspond to the experimental result obtained by TEM observation [37]. Furthermore, whether the stress corrosion cracking is dominated by anodic dissolvent mechanism, cleavage fracture mechanism and hydrogen embrittlement due to hydrogen diffusion and concentration to the site of the maximum value of P or not will depend on competitive relationship between the mobility of dislocation ow and hydrogen diffusion. When a crack growth occurs as a process of stress corrosion cracking (SCC), the problem whether its dominant factor is anodic dissolvent chemical reaction or hydrogen embrittlement mechanisms remains to be claried. Furthermore, the fact that cleavage fracture appears in the process of SCC for ductile materials makes this problem much more difcult. In this paper, the analysis of this problem was conducted under the interaction of a crack, hydrogen and dislocation groups relating the results of analysis of the mechanism of hydrogen diffusion which have been already conducted [1113]. By these results, the objective of this research that the occurrence of these three mechanisms can be predicted by our proposed physical model is accomplished without the contradiction with actual phenomenon as is shown in Fig. 17. On the other hand, from the view point of metallurgy, necessity on the construction of much more detailed physical model based on the metallurgical aspects with the effective scale of 3 107 mm may be pointed out [22]. However, the objectives of this paper concern macroscopic phenomenon of fracture. To clarify this mechanism in this paper, interactive analysis between macroscopic, mezzo and microscopic mechanical factors such as a crack, dislocation groups and hydrogen was conducted. Therefore, it is inevitably necessary for us to construct an approximated physical model with elimination of many microscopic mechanical factors which concerns metallurgical aspects from the view point of the numerical analysis. The physical validity of this approximated model should be assured how the results obtained by this model predict accurately the actual phenomenon. The results obtained by this paper were found to sufciently predict the occurrence of SCC fracture mechanisms. Therefore, as long as the physical scale which is noticed by our analyses, this approximated model is considered to be reasonable.

5. Conclusions (1) Under the corrosive condition for ductile materials, the inverse pile-up with high dislocation density occurs depending on the site of hydrogen cluster due to the interaction force between dislocation groups and hydrogen cluster. (2) This result shows that sub-critical cleavage fracture for the ductile material under corrosive environment conditions is induced by competitive interaction between hydrogen and dislocation groups around a crack tip, which is in good agreement with the experimental result obtained by other literatures.

108

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement

(3) On the other hand, for each case of the lower or higher rate of hydrogen diffusion (lower and higher yield stress), the inverse pile-up with high dislocation density does not occur and the characteristic of dislocation distribution equals to the case without hydrogen. For such case, anodic dissolvent mechanism and hydrogen embrittlement mechanism will be dominant respectively. (4) Whether the stress corrosion cracking is dominated by anodic dissolvent mechanism, cleavage fracture mechanism and hydrogen embrittlement due to hydrogen diffusion and concentration to the site of the maximum value of P or not will depend on competitive relationship between the mobility of dislocation ow and hydrogen diffusion. Acknowledgement The partial support of this work by the Grant-in-Aid for COE (Center of excellence) Research (No. 11CE2003), The Ministry of Education, Culture, Sports, Science and Technology is gratefully acknowledged. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] J.D. Landes and R.P. Wei, J. Fract. 9 (1973), 277. N.J. Petch, Phil. Mag. 1 (1956), 331. I.M. Bernstein, Mater. Sci. Eng. 6 (1970), 331. R.A. Oriani, Ann.Rev. Mater. Sci. 8 (1978), 327. H.W. Liu, Trans. ASME, J. Bas. Engng. 92 (1970), 633. H.P. Leeuwen, Corrosion 31(2) (1975), 42. H.P. Leeuwen, Engng. Fracture Mech. 10 (1978), 1. S. Ochiai, S. Yoshinaga and Y. Kikuta, Trans. ISIJ 15 (1975), 50. M. Iino, Engng. Mech. 10 (1978), 1. Y. Ueda and K. Fukuda, Matrix structure and analysis, in: Proc. 7th Japan Structure Association, 1973, p. 671. A.T. Yokobori, Jr., T. Nemoto, K. Satoh and T. Yamada, Trans. of the Japan Soc. Mech. Engineers 59 (1993), 2120. A.T. Yokobori, Jr., T. Nemoto, K. Satoh and T. Yamada, Engng. Fract. Mech. 55 (1996), 47. A.T. Yokobori, Jr., Y. Cinda, T. Nemoto, K. Satoh and T. Yamada, Corrosion Science 44 (2002), 407. T. Shoji, S. Suzuki and R.G. Ballinger, Theoretical prediction of SCC growth behavior, in: Proc. of the 7th International Symposium on Environment Degradation of Materials in Nuclear Power Systems, Breckenridge, 1995, p. 881. M. Garnkle, Metallurgical Transactions 4 (1973), 1677. T. Magnin, A. Chambreuil and J.I. Chateau, International Journal of Fracture 79 (1995), 147. N.J. Petch, J. Iron and Steel Inst. 174 (1953), 25. T. Yokobori, Y. Sawaki and S. Nakanishi, Engng. Fract. Mech. 12 (1979), 125. J. Watanabe, T. Iwadate, Y. Tanaka and T. Yokobori, Engng. Fract. Mech. 28 (1987), 589. A.T. Yokobori, Jr., T. Isogai and T. Yokobori, Acta Metall. Mater. 41 (1993), 1405. A.T. Yokobori, Jr., T. Iwadate and T. Isogai, ASTM STP 1207 24 (1994), 464. S.G. Robert, S.J. Noronha, A.J. Wilkinson and P.B. Hirsch, Acta Materialia 50 (2002), 1229. N. Ohtani, The 23th Japan National Symposium on Strength, Fracture and Fatigue, 1978, p. 17 (in Japanese). W.W. Gerberich, R.A. Oriani, M.J. Lin, X. Chen and T. Foeche, Philos. Mag., A 63 (1991), 363. M. Ichikawa and T. Yokobori, Rep. Res. Inst. Str. & Fracture of Materials, Tohoku Univ., Sendai, Japan, Vol. 1, 1965, p. 59. T. Yokobori and M. Ichikawa, Rep. Res. Inst. Str. & Fracture of Materials, Tohoku Univ., Sendai, Japan, Vol. 3, 1967, p. 1. T. Yokobori, A. Kamei and S. Konosu, Int. Journal of Fracture 10 (1974), 375. T. Yokobori, M. Yoshida, H. Kuroda, A. Kamei and S. Konosu, Engng. Fract. Mech. 7 (1975), 377. A.T. Yokobori, Jr., T. Yokobori and A. Kamei, J. Materials Science and Engineering 40 (1979), 111. A.H. Cottrell and B.A. Bilby, Proc. Phys. Soc. London 62 (1949), 19. Rui Husan Zhao, Shu-Hu Dai and J.C.M. Li, International Journal of Fracture 3 (1983), 29. A.P.L. Turner and T. Vreeland, Jr., Acta Metall. 18 (1970), 1255. A.T. Yokobori, Jr., T. Yokobori and A. Kamei, J. Appl. Phys. 46 (1975), 3720.

A.T. Yokobori, Jr. et al. / The mechanism of hydrogen embrittlement [34] [35] [36] [37] T. Yokobori, A.T. Yokobori, Jr. and A. Kamei, Phil. Mag. 39 (1974), 367. J.P. Chateau, D. Delafosse and T. Magnin, Acta Materialia 50 (2002), 1507. P. Sofronis and H.K. Birnbaum, J. Mech. Phys. Solid 43 (1995), 49. K.W. Gao, W.Y. Chu, B. Gu, T.C. Zhang and L.J. Qiao, Corrosion 56 (2000), 515.

109

S-ar putea să vă placă și