Available online at www.sciencedirect.

com

Engineering Fracture Mechanics 75 (2008) 551559 www.elsevier.com/locate/engfracmech

Inuence of yield strength levels on crack growth mode in delayed fracture of structural steels
Yuki Komatsuzaki, Haengsik Joo, Kunihiro Yamada
*
Department of Mechanical Engineering, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan Received 17 November 2006; received in revised form 29 January 2007; accepted 3 February 2007 Available online 21 February 2007

Abstract A mechanistic aspect of the susceptibility to the delayed fracture is studied with an emphasis on the critical behaviour of the subsurface growth of Quasi-Cleavage (QC) and Inter-Granular (IG) cracks. The materials employed are 0.35%C plain carbon steel and boron added bolt steel which were quenched and tempered to have various levels of yield strength ranging from 500 to 1400 MPa. Fractographic analysis shows us that QC + IG cracking process can be an essential mode in the delayed fracture of steels. A low susceptibility to delayed fracture can be explained by the crack growth behavior when the crucial blunting occurs at the crack tip. 2007 Elsevier Ltd. All rights reserved.
Keywords: Structural steel; Fracture mechanisms; Hydrogen damage; Crack growth; Fractography

1. Introduction There still remains a lot of discussion about the issue of premature fracture of structural components due to hydrogen degradation under corrosive environments. Since this sort of fracture, that develops at an extremely low range of stresses lower than the yield strength level, is quite well known for high strength steels [17], the material selection for a structural component involving high strength steels would meet with highly complicated problems. System design of pipelines for crude oil and natural gas etc, still encounters tough problems of premature fracture in SCC (Stress Corrosion Cracking) [8], hydrogen induced cracking (HIC) [9,10] associated with the hydrogen degradation [11] of the materials, even for low strength steels. An inter-granular (IG) type of fracture is notable in high strength steels while no appreciable trace of IG fracture can be observed in the hydrogen related fracture of low strength steels [11,12]. According to these evidences of the fracture morphologies, there still remain several critical issues on the subsurface crack growth in both high and low strength steels under hydrogen attack.

Corresponding author. Fax: +81 45 566 1495. E-mail address: kymd@mech.keio.ac.jp (K. Yamada).

0013-7944/$ - see front matter 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.engfracmech.2007.02.009

552

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

To discuss about the issue on the hydrogen degradation of structural components with an attention on the crack growth behaviour, two types of structural steels such as plain carbon steel and boron added steel were used to prepare specimens having various levels of yield strength ranging from 500 MPa to 1400 MPa. These specimens were then put into sustained load fracture tests with concurrent hydrogen charging to examine basic nature of hydrogen degradation with a special emphasis placed on the subsurface crack growth behaviour. 2. Experimental procedure The materials employed in this study were a plain carbon steel S35C and a boron added bolt steel Bolten110N with chemical compositions as shown in Table 1 and with mechanical properties as shown in Table 2 together with the results of delayed fracture tests. These materials were machined into unnotched shape with a gauge length and a mid-section diameter of 10 mm and 5 mm, respectively. To obtain martensitic structure with an average prior austenite grain diameter of about 20 lm, the Bolten110N specimen was rst austenitized at 1223 K for 10 min in electro-furnace, followed by oil quenching. The S35C steel was austenitized at 1173 K for 10 min in electro-furnace, followed by oil quenching to arrange a quenched martensitic structure. These specimens were mechanically polished, and then tempered at 473 K, 523 K, 573 K, 623 K, 673 K, 773 K, 873 K and 973 K, respectively, for 1 h in vacuum furnace with 1.3 104 Pa to have a variety of yield strength levels as shown in Fig. 1 and Table 2. Magnitudes of mechanical properties in Table 2 are averaged results of at least two specimens. The threshold stresses of the delayed fracture were estimated from the fracture test results of 510 specimens, as shown later in Figs. 3 and 4. The specimen surface was covered with a shielding tape to prevent damage from an attack by H2SO4, except at the gauge area. Sustained load type fracture tests were carried out using a 3 ton creep-testing machine as is shown in Fig. 2 under the conditions, in which the mechanical loading began with concurrent hydrogen charging in H2SO4solution of 0.05 M concentration and 500 A/m2 current density during the test. This can charge a fracture specimen with 12 ppm hydrogen. To trace the crack path on the fractured surface, a fractographic analysis was then made employing scanning electron microscope (SEM).

Table 1 Chemical composition of specimens Materials S35C Bolten110N C 0.38 0.20 Si 0.19 0.16 Mn 0.70 0.77 P 0.015 0.012 S 0.009 0.010 Ni 0.06 0.03 Cr 0.14 0.63 Mo 0.01 Cu 0.06 0.02 B 0.0002 0.0022 Al 0.002 0.066

Table 2 Mechanical properties of the specimens Materials Tempering temperature (K) 473 573 673 723 773 973 473 523 573 673 723 973 Yield strength (MPa) 1330 1224 1170 1071 970 650 1205 1204 1271 978 785 462 U.T.S. (MPa) 1630 1471 1307 1179 1096 697 1372 1367 1308 1197 998 707 Micro-Vickers hardness (P = 500 g, 30 points) 494 443 406 362 334 215 440 434 446 368 302 186 Threshold stress rth (MPa) 150 350 500 650 600 475 600 350 250 400 450 410 Elapsed loading time (min) 104 104 104 104 104 104 104 104 104 104 104 104

S35C

Bolten110N

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

553

1400 1200 Yield Strength (MPa) 1000 800 600 400 200 0 0 200

S35C Bolten110N

400 600 800 1000 Tempering temperature (K)

1200

Fig. 1. Inuence of tempering temperature on yield strength.

Fig. 2. Schematic illustration of delayed fracture test.

3. Results and discussion 3.1. Results of delayed fracture test A delayed fracture test under the above hydrogen charging condition was carried out for all kinds of specimens having wide ranges of yield strength levels. Fig. 3a shows a typical example of the delayed fracture test for S35C specimens quenched and tempered at 473 K. All results of S35C specimens tested are then given in Fig. 3b. Fig. 4a shows a typical result for Bolten 110N specimens tempered at 473 K. All results of Bolten110N specimens are shown in Fig. 4b. The fracture data in Figs. 3b and 4b level o at about 103 min so that the applied load was discontinued at 4 10 min for evaluating a threshold stress of each specimen. The threshold stress (rth) was then dened as the maximum stress at which the fracture did not take place after 104 min of sustained load time duration. All these threshold stress values (rth) are summarized in Table 2 and then plotted against the yield strength level (ry) after normalization with the yield strength (rth/ry), as is shown in Fig. 5. The results show that the normalized delayed fracture strength decreases with the increase in the yield strength, and a remarkable change in fracture susceptibility can be seen at about the yield strength level of

554

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

a
Fracture stress (MPa)

700 600 500 400 300 200 100 0 1

473K tempering

10 100 1000 102 103 Time to fracture (min)

10000 104

b
Fracture stress (MPa)

900 800 700 600 500 400 300 200 100 0 1 10 1002 10 1000 103

473K 573K 673K 723K 773K 973K

104 10000

Time to fracture (min)

Fig. 3. Delayed fracture curve of hydrogen charged specimen (a) S35C 473 K tempering and (b) S35C.

8001000 MPa. Fig. 5 also shows that very high strength steels, even with the unnotched specimen geometry, have surprisingly low threshold stress levels in delayed fracture. A question arises as to how the hydrogen damage accumulates and develops at such wide ranges of stress levels as the above ones. Then, to examine the detailed feature of crack development during the hydrogen related fracture process of specimens in all range of yield strengths, a fractographic analysis was made with SEM. 3.2. Dierence in morphology of crack growth between high and low strength steels As is well-known, the feature of crack growth under hydrogen attack can be mainly inuenced by a dierence of the strength level of specimens, such that IG crack is noticeable in high strength steels while no IG crack is appreciable in low strength steels [3]. A fractographic analysis gave us valuable information that the QC crack, not the IG crack rst appears in the subsurface of the specimen in both high and low strength steels, as shown in Figs. 6 and 7. This QC crack growth continues until reaching the transition stage where the growth mode changes into IG mode with an unstable crack growth for high strength steel specimens, while no transition to IG mode was recognized in the crack growth process for low strength steel specimens. These characteristics of crack growth in high and low strength steels are illustrated in Fig. 8 with typical fracture facets known as QC, IG and MVC. Accordingly, it can be concluded from the mechanistic aspects of fracture that the susceptibility to delayed fracture is related to the event that the fracture process of high strength steels prefers the IG crack growth

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

555

a
Fracture stress (MPa)

1200 1100 1000 900 800 700 600 500 400 300 200 100 0

473K tempering

10

102

10 3

104

Time to fracture (min)

b
Fracture stress (MPa)

1200 1100 1000 900 800 700 600 500 400 300 200 100 0 1

473K 523K 573K 673K 723K 973K

10
102 10 3 Time to fracture (min) 104

Fig. 4. Delayed fracture curve of hydrogen charged specimen (a) Bolten110N 473 K tempering and (b) Bolten110N.

1.00
Normalized delayed fracture strength ( th/ y)

S35C Bolten110N

0.90 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00 0 200 400 600 800 1000

1200

1400

Yield strength (MPa)

Fig. 5. Relationship between resistance of delayed fracture and yield strength (Unnotched specimen.).

triggered by the QC crack resulting in a fast fracture, while no IG crack appears in the fracture process of low strength steels, i.e., low susceptibility. To elucidate the mechanistic aspects of the susceptibility of specimens in delayed fracture ranging from high to low yield strengths, nature of the development of QC crack, which triggers the onset of IG crack growth, should be carefully examined as in the following.

556

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

Fig. 6. Fracture origin and non-metallic inclusion in the center of fracture origin (High strength steels; S35C 473 K tempering, r = 650 MPa, t = 312 min.).

Fig. 7. Fracture origin (Low $ medium strength steels; S35C 773 K tempering, r = 700 MPa, t = 1780 min.).

Fig. 8. Illustration of the morphology of crack propagation in delayed fracture.

3.3. Crack tip blunting and IG crack growth A dierence of the fracture morphology between high and low strength steels seems to be obvious so that a fractographic analysis is made to examine how the dierence of the yield strength level of specimens aect on

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

557

the mode of crack growth. A comparison between the crack growth processes of high and low strength steel specimens is shown in Fig. 8 with three micrographs of the typical facets such as IG, QC and MVC. The QC crack always starts at the beginning of the cracking process in both high and low strength steels. The QC crack growth is interrupted by the event of transition of cracking mode to the IG crack followed by MVC in high strength steels, while no trace of transition of cracking mode from QC to IG crack is observed in the fracture surface of low strength steels. The QC crack developed in low strength steels grows with coalescence of a number of neighboring QC cracks during a relatively long period of time, followed by a development of MVC, not the IG crack growth, leading to cause fast fracture. The QC cracks should remain as non-propagation cracks (NPC) when the applied stress is not enough to boost further crack coalescence and cause fast fracture. Furthermore, this transitional event of development of NPC is encouraged by the plastic deformation at the crack tip, namely, the crack tip blunting. If a substantial amount of plasticity is built up at the tip of QC crack in low strength steels, the hydrogen diusion and concentration should remarkably decrease [12] compared with the high strength steels, where no appreciable plasticity is expected near the crack tip. In high strength steels, a sharp crack tip conguration is maintained as sharp with little plastic deformation resulting in the fast fracture accompanying IG crack propagation [4,5]. A reduction in the stress intensity and reduced hydrogen accumulation at the crack tip may result in reducing the possibility of the occurrence of the onset of IG crack growth from the QC crack front. Bolten110N specimen tempered at 473 K shows slightly dierent crack propagation behaviour from that in S35C specimens. The QC crack developed at the beginning continues to grow making QC facet and QC 0 facet. Such a QC 0 facet is characterized by a deformed QC facet having rough and uneven fracture surface compared with the ordinary QC facet, showing an appreciable plastic deformation. A question arises as to whether the crack tip blunting actually works, as we expected that the localized plasticity at the crack tip could suppress the fatal growth of the QC crack. A conguration of the tip of the QC crack has then been examined on the cross section of the specimens which had already been loaded for 4.32 103 min to develop QC crack under sustained load at the stress levels of 150 MPa and 600 MPa for S35C specimen and of 250 MPa and 600 MPa for Bolten110N specimen, respectively. These applied stress levels correspond to the respective threshold stresses of each delayed fracture specimen. A metallographic examination on a cross section of the gauge area of the above specimens has been made after mechanical polishing to examine the crack tip conguration, as is shown in Fig. 9(a)(d). These micrographs (Fig. 9) show that two dierent crack tip morphologies of the initial QC cracks have developed in both types of materials tested. The crack tip is very sharp with no appreciable plasticity in S35C steel tempered at 473 K, as is shown in Fig. 9(a), and also in Bolten110N tempered at 573 K in Fig. 9(c). On the other hand, in the cases of S35C steel tempered at 773 K and Bolten110N tempered at 473 K, both cases show substantial plastic deformation at the tip and the ank of the crack, as is shown in Fig. 9(b) and (d). The reason for the diculty for the onset of IG crack growth from the QC crack in the low strength steels may be explained by these evidences: the crack tip blunting is easily built up in the case of S35C having low strength level, and this may reduce the hydrogen accumulation near the crack tip [13] resulting in the disappearance of IG crack growth which should be triggered by the sharp QC crack in the crack growth process. This explanation can also be applicable to the case of crack tip blunting of Bolten110N tempered at 473 K though it has relatively high yield strength compared with that of S35C specimen. 3.4. Model of crack development in delayed fracture of steels By analyzing both the fracture surface and the specimen cross section, characteristics of the crack growth process can be summarized as a simple model in Table 3. It is obvious that the initial crack is always the QC crack irrespective of the yield strength levels of the specimens. The main aspect of the crack growth process involves the QC crack, IG crack, and MVC in the delayed fracture of structural steels. However, in the low susceptibility case, the IG cracking process disappears due probably to the crack tip blunting resulting in diculty of attaining a substantial amount of hydrogen concentration at the crack front and diculty of reaching the suciently high stress intensity to cause fracture. The model of the sequence of crack growth mode QCIGMVC can be considered as an essential aspect in the crack propagation mechanisms regardless of the level of the yield strength in steels subjected to delayed

558

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

Fig. 9. The optical micrographs of a crack shape.

Table 3 Essential aspects of the crack growth process Materials Tempering temperature (K) 473 773 473 573 Yield strength (MPa) 1330 970 1207 1271 Normalized delayed fracture strength (rth/ry) 0.11 0.62 0.50 0.20 Crack propagation model QC ! IG ! MVC QC ! MVC QC ! QC 0 ! MVC QC ! IG ! MVC Crack shape

S35C Bolten110N

Sharp Blunt Blunt Sharp

fracture test. This model of crack growth can explain the crack growth process in the high susceptibility case for high strength steels, and the absence of IG crack mode for low strength steels, i.e., low susceptibility case. This event of disappearance of IG crack in the low susceptibility case can be well explained by the role of plasticity developed around the QC crack front. 4. Conclusions A mechanistic aspect of the susceptibility to delayed fracture of structural steels is studied using unnotched specimens with an emphasis on the mechanisms of the subsurface crack growth, using two kinds of steels which have various yield strength levels ranging from 500 MPa to 1400 MPa. Results obtained are summarized as follows: (1) A remarkable decrease in the delayed fracture strength can be seen at the particular strength level about ry = 8001000 MPa in S35C and Bolten110N steels. The development of the subsurface QC crack is a common and fundamental phenomenon in the crack growth process in both high and low strength steels under hydrogen attack. (2) The crack growth mode QCIGMVC can be regarded as an essential event to characterize the development of subsurface fracture damage in unnotched specimen under hydrogen attack. This model of

Y. Komatsuzaki et al. / Engineering Fracture Mechanics 75 (2008) 551559

559

crack growth fully appears in the high susceptibility case, while the IG crack disappears in the low susceptibility case. This dierence can be well explained by the appearance of the critical event of crack tip plasticity. References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] Troiano AR. The role of hydrogen and other interstitials in the mechanical behaviour of metals. Trans ASM 1960;52:5480. Beachem CD. A new model for hydrogen assisted cracking (Hydrogen, Embrittlement). Metallurg Trans 1972;3:43751. Hirth JP. Eects of hydrogen on the properties of iron and steel. Metallurg Trans A 1980;11(6):86190. Ishikawa M, Nakatani Y, Yamada K. Mechanistic aspects of hydrogen cracking process in high strength steel. In: Turnbull A, editor. Hydrogen transport and cracking in metals. UK: The Institute of materials; 1995. Tsuboi K, Yatabe H, Yamada K. Hydrogen induced cracking in high strength steel. Mater Sci Technol 1996;12:4004. Yatabe H, Yamada K, de los Rios ER, Miller KJ. Formation of hydrogen assisted intergranular crack in high strength steels. Fatigue Fract Engng Mater Struct 1995;18(3):37784. Tsay LW, Lin ZW, Shiue RK, Chen C. Hydrogen Embrittlement susceptibility of laser hardened 4140 steel. Mater Sci Engng A 2000;290:4654. Wang JQ, Atrens A. SCC initiation for X65 pipeline steel in the high pH carbonate/bicarbonate solution. Corros Sci 2003;45:2199217. a Domizzi G, Anteri G, Ovejero-Garc J. Inuence of sulpher content and inclusion distribution on the hydrogen induced blister cracking in pressure vessel and pipeline steels. Corros Sci 2001;43:32539. Zhang T, Chu WY, Gao KW, Qiao LJ. Study of correlation between hydrogen-induced stress and hydrogen Embrittlement. Mater Sci Engng A 2003;347:2919. Stress Corrosion Cracking and Hydrogen Embrittlement of Iron Base Alloys, 1977; NACE-5. Nakatani Y, Matsuoka H, Yamada K. Formation and characteristics of quasi-cleavage crack in hydrogen damaged low strength steel. In: Brown MW, de los Rios ER, Miller KJ, editors. Proceeding of the ECF 12. Fracture from Defects, vol. III. The European Structural Integrity Society; 1998. p. 11216. Yokobori Jr T, Nemoto T, Sato K, Yamada T. Numerical analysis on hydrogen diusion in solid around the crack tip. Trans Jpn Soc Mech Engng A 1993;59:21207 [in Japanese].

[13]