The author previously reported the fatigue crack initiation process in a low carbon steel, and clarified that the fatigue cracks originated in ferrite grains by an intrusion mechanism. In this paper, the mechanism of fatigue crack initiation in a high strength steel was investigated using a Ni-Cr-Mo steel (SNCM 439) consisted of sorbitic structure, that is, mixed structure of farrite and granular carbide. The observations of fatigue crack initiation process were made on the surface, the longitudinal section and the fracture surface of specimens using optical and scanning electron microscopes. Then, crystallographic investigations were made on the fracture surface by the use of etch pits. Based on the experimental results, the conclusions are summarized as follow: (1) The feature of Stage I cracking was found in the fatigue crack initiation for all the specimens observed, and the initiation process could be explained by an intrusion mechanism. (2) The fatigue crack originated at a part of ferritic matrix in which granular carbides hardly precipitated, and no influence of inclusion was recognized on the initiation process. (3) Therefore, it was assumed that mechanism of a fatigue crack initiation in high strength steels with ferritic structure was essentially the same as that of a low carbon steel.
Fracture morphologies are often formed as a result of plastic deformation at the crack tip. Therefore, it has been found that sizes of fracture morphologies (striation spacings, stretched zone width etc.) are correlated with fracture mechanics parameters. Such relationship makes possible the quantitative analysis of service failures of machines or structures. In this paper, striation spacings were measured by transmission microscope fractography on the replica obtained from fatigue failed machine parts (gears, shaft and crane hook), and the values of applied stress amplitude or cycles given to propagate the crack until failure occurred were estimated using the relation between the striation spacing and stress intensity factor, ΔK. Fatigue tests or fatigue crack growth tests were carried out with the specimens taken from actually failed machine parts to know the fatigue properties for the parts. The combined fractographic fracture mechanics method was found very valuable in the quantitative analysis of fatigue failures of machine parts.
Tensile properties, fracture toughness, charpy impact value and rotating bending fatigue strength were obtained for ductile, vermicular and gray cast irons with ferritic matrix specially melted for this investigation. Detailed SEM observations were conducted on fracture surfaces of these failed specimens. A step like pattern at the crack initiation area and ductile striation at the crack propagation area were characteristically observed on the fatigue fracture surface of ductile or vermicular cast iron. On the contrary, a step like pattern was not observed on either fractured surface of crack initiation and propagation areas of gray cast iron. Dimple and river patterns were predominant on tensile and charpy impact fracture surface, respectively. However, a step like pattern was not found on the both fracture surfaces. Since the step like pattern could be identified by SEM observations with a relatively lower magnification and was not observed on tensile or charpy impact fracture surface, it is considered that the step like pattern might be a useful key pattern to identity fatigue fracture in the service failure of machine components made of ferritic cast iron. The stretched zone could be clearly observed on the fracture surface of ductile cast iron but it became harder to be identified as graphite nodularity decreased, and the stretched zone could not be identified on the fracture surfaces of gray cast iron.
The elastic-plastic fracture toughness JIc tests by the stretched zone method were performed to obtain JIc values for machine structural steels. The validity of the JIc values obtained was evaluated, and the effects of the heat treatment conditions and microstructure of steels on the JIc values were investigated by use of streched zone width, SZW. The results obtained are summarized as follows. (1) The inclination of the blunting line in SZW versus J curve is hardly affected by the condition of heat treatment for SCM440 steels, while it is affected obviously for S45C steels. (2) The relation between SZWC and σB varies with the difference of microstructure and with the ratio of ferrite area to pearlite area in ferrite-pearlite steels. (3) SZWC versus JIc/σfs diagram (in log. scale) can be arranged by two linear lines according to the difference of microstructure.
Profiles of stretched zones were observed at various stages during crack tip blunting and stable crack growth in JIc tests on five steels with yield strength levels ranging from 765 to 333MPa. During the stable crack growth, the steel with a high yield strength of 765MPa indicated little deformation in both the stretched zone and the dimple region, while the medium and, particularly, low strength steels showed considerable deformation there. Based on the observations in this study, several proposals were made for improving the stretched zone width method in the JSME standard of test for JIc.
The comparison in crack tip plastic stretch between the fatigue and the ideal cracks was made with special attention to the plasticity induced crack closure phenomenon during fatigue crack growth. By using this result, a model to evaluate transitional behavior from the fatigue crack to the ideal crack was proposed. The results obtained are as follows: (1) For a given stress intensity factor, K, the ratio of the fatigue crack tip plastic stretch to that of the ideal crack is nearly 0.13. With this value, the ratio of the crack opening stress intensity, Kop, to the maximum stress intensity, Kmax, can be predicted as Kop/Kmax=0.58. (2) When a single peak overload is applied to the fatigue crack, transitional behavior from the fatigue crack to the ideal crack occurs. A model to evaluate this behavior is proposed using the predicted Kop/Kmax value. Predicted values from this model showed a good agreement with the experimental results for several stress ratios, R. (3) This model and the experimental results suggest that a fatigue pre-crack condition for the fracture toughness test adopted in the standards of the Japan Scociety of Mechanical Engineers, S001, and of the ASTM, E399, is reasonable.
A fractographic investigation on mode II fatigue crack growth has been made on 7075-T6 and 2017-T4 aluminum alloys. Not like in mode I growth which is known to show a stage I-stage II transition, no transition has been observed in mode II growth but there has been a gradual change in fracture appearance with an increase in ΔKII, whose growth is characterized by crystallographic growth in the lower ΔKII region while in the higher ΔKII region the smooth curved surface which seems to be formed by glide plane decohesion becomes dominant. Another characteristic pattern, which is often observed regardless of the magnitude of ΔKII, is what looks like piled-up thin lamellas. Stress ratio R and superimposed static KI have no influence on the appearance of mode II growth. Mode II growth rate depends on the range of the stress for both cases of R=0 and-1. A model which is to explain these characteristics of mode II growth has been proposed.
The influences of grain size and stress ratio on near threshold fatigue crack growth behavior have been studied by means of fracture mechanics and fractography on annealed 60-40 brass where the grain size was varied from 15 to 33 and 92μm. The results obtained are summarized as follows. (1) For the mid-range of crack growth rates above 5×10-9m/cycle, the dominant fracture appearance was striation formation regardless of grain size and stress ratio. On the other hand, for the near threshold crack growth rates below 5×10-9m/cycle, the striation-like appearance a specified crystallographic plane was mixed with the microstructure sensitive fracture appearance such as cleavage-like mode and intergranular facet. In particular, the intergranular facets were observed on only the smallest grain size. (2) According to the changes in dominant fracture appearance, various effects of grain size and stress ratio on crack growth rate were observed. In particular, the resistant for near threshold crack propagation and the ΔKth decreased with decreasing grain size. If the microstructure was before recrystallization, the ΔKth remarkably depended on grain size as the stress ratio increased. On the other hand, the ΔKth of recrystallized structure had almost the same dependence on grain size regardless of stress ratio. This fact may be well explained by the difference in fracture appearance and the concept of fracture surface roughness induced crack closure. (3) As the stress ratio increased, the near threshold crack growth rate increased and the ΔKth decreased regardless of grain size. The ΔKth also depended remarkably on stress ratio with decreasing grain size. However, when the microstructure was the recrystallized one, the ΔKth had almost the same dependence on stress ratio regardless of grain size.
The fatigue crack propagation path is not always straight nor vertical to the stress axis, when a hard region lying ahead of the crack tip is inclined to the fatigue crack propagation direction. In this report, the effect of such hard region to the propagation direction and the mechanism for the change of fatigue crack path were investigated using the specimens having a plain soft-hard boundary which was induced by induction hardening. The fatigue crack got inclined toward the relatively softer side and propagated in the soft region parallel to the soft-hard boundary. The residual stress seemed not to have much effect on the change of propagation direction but the hardness heterogeneity had a marked effect on it. The fractographic observation revealed the symmetric striation at the place where the crack propagated straight, and the asymmetric striation and many secondary cracks growing out of the fissures at the places where the crack inclined. From these findings, it seems that the change of propagation direction is caused by the formation of the asymmetric striation and the micro crack branching.
Fatigue crack propagation under periodic overstressing i.e. intermittent application of a very small number of cycles of overstress during a cycling of very large number of understress below Kth, was studied with the normalized iron specimens and those stretched to 10% strain. Remarkable acceleration of crack propagation rate of about 200 times was observed in the former but it amounted about 100 times in the latter. Fracture surface observation revealed characteristic clam-shell markings corresponding to the varying stress. Etch pit study in the area where the markings are obvious suggested that the fracture surface formed by understressing has some bearing with crystallographic orientation. Direct observation of dislocation structure adjacent to the crack tip showed that acceleration of crack growth under the varying stressing is originated from cooperation of over- and understressing: Large crack tip opening displacement accompanied with formation of radial bands of dislocation structure emanating from crack tip on overstressing, and crack propagation along the band occurring on the way to recovery of the structure during understressing. Some characteristics of acceleration under the periodic overstressing are discussed in the light of the new findings.
Fatigue tests under a constant average load with load spectrum of conditional distribution, which resembles a usual service load in structural parts, were carried out on CT test specimens of stainless steel (SUS 304) and the striation spacings formed on the fracture surface were measured by micro-fractography. As preliminary tests, five kinds of block loadings were applied to the test specimens, and the relationship between the applied load and striation spacings was investigated. Fatigue tests and fractography on two kinds of random loadings simulating a usual service load were then carried out, and the possibility to evaluate the random load by striation spacing measurements was examined. It is suggested that fractography is useful for evaluation of main service load spectrum responsible to fatigue crack growth.
An investigation has been made into the analysis of the fracture morphology of hydrogen embrittlement for a quenched and tempered Cr-Mo steel. The Cr-Mo steel showed the tempered martensite embrittlement by 300°C tempering treatment. The susceptibility to hydrogen embrittlement was highest for the as-quenched specimen, and decreased with increasing tempering temperature. The ratio of intergranular fracture in the fractured surface was highest for the specimen tempered at 300°C. The fracture morphology of tempered martensite embrittlement was intergranular fracture with flat surface, but the morpholgy of hydrogen embrittlement was intergranular fracture with tear ridges on the surface. The characteristic morphology of hydrogen embrittlement was quasi-cleavage fracture due to hydrogen (QCHE), which had the same small facet size as martensite lath morphology, with secondary cracks along the martensite lath. This fracture morphology was the same as that of hydrogen embrittlement of HT80 steel previously indicated.
Fracture mechanics studies on delayed fracture have been done extensively, but they are mainly concerned in a through crack in CT or WOL specimens and little concerned in a three dimensional surface crack. The surface cracks are often seen in actual structures. In the present work, surface crack growth in delayed fracture in tap water with cathodic charging has been examined by using the bend type specimens of AISI 4340 steels (JIS SNCM 439) tempered at 200°C and 400°C. The results obtained are summarized as follows; (1) The shape of growing surface crack is semi-elliptic. Here, the crack length takes a maximum value at a little inside of the specimen surface. Comparing the shape of the delayed crack with the one of the fatigue crack, both are nearly the same for the material tempered at 400°C, but the shape of the delayed crack becomes flatter than the fatigue crack for the material tempered at 200°C. The reason should be attributed to the fact that the material tempered at 200°C is more sensitive to hydrogen than the one tempered at 400°C. (2) Reflecting the result of (1), for the material tempered at 200°C, the crack growth rate in the direction of surface, db/dt, shows a higher value than the one in the direction of depth, da/dt. (3) The values of delayed fracture toughness, Kdc, in the region of fast fracture and of threshold stress intensity factor in crack growth, KSCC, are lower in the case of surface crack than in the case of through crack.
This paper describes the environmentally induced ductility change in the heat-affected-zone (HAZ) of high strength low alloy steel. Tensile tests were carried out on the samples with or without hydrogen absorption at elevated temperature. Even for QT-type steel there was a possibility of ductility loss due to dynamic strain aging in the HAZ. Under the temperature conditions which promoted dynamic strain aging, the ductility decreased at the temperature range between room temperature and 250°C. This was due to an increase in free nitrogen and the coarsening of pre-austenite grain size. The ductility loss was significant at the HAZ with coarse pre-austenite grain size or high free nitrogen content. When the materials susceptible to hydrogen embrittlement were used in the environments which promote dynamic strain aging, the ductility decreased remarkably at elevated temperature by the independent but simultaneous effects of dynamic strain aging and hydrogen embrittlement.
Fatigue properties of soluted and sensitized SUS 304 stainless steel have been investigated in high pressure hydrogen up to 4.02MPa and atmospheric argon at room temperature. The results obtained are as follows: (1) The number of cycles to failure of both the soluted and the sensitized steel decreased with increasing hydrogen pressure. Large decrease was observed on the sensitized steel in comparison with the soluted steel. (2) The fatigue limits of the soluted and the sensitized steel decreased with increasing hydrogen pressure. Little difference was observed on the fatigue limits between the soluted and the sensitized steel. (3) Typical fatigue fracture modes were observed on the fracture surface in fatigue crack growth area (Area II) depending on the type of gas, that is argon or hydrogen. In this area, striations were observed on the surface fractured in argon and brittle transgranular fracture in hydrogen.
The fatigue crack growth behavior in hydrogen gas environment was examined using compact tension specimens of five kinds of materials, i. e., low alloy steels (Ni-Mo-V, Ni-Cr-Mo-V), austenitic alloy steels (18Mn-5Cr, SUH660) and a copper alloy (Be-Cu). The fatigue crack growth rates in laboratory air and hydrogen gas were obtained and the relation between the fatigue crack growth rate and fracture morphology was discussed based on the fractographical observations. The main results obtained are as follows. (1) The effect of hydrogen gas on the fatigue crack growth rate was classified into following four types. (a) The effect of hydrogen gas was not observed (18Mn-5Cr, SUH660). (b) A slight decrease in fatigue crack growth rate was observed (Be-Cu). (c) A considerable increase in fatigue crack growth rate was observed, but the effect of frequency was not observed (Ni-Mo-V). (d) A considerable increase in fatigue crack growth rate was observed, and the effect of frequency was also observed (Ni-Cr-Mo-V). (2) In the materials of types (c) and (d), the cleavage fracture was observed and the area percentage of the intergranular fracture on the fracture surface increased in hydrogen gas. The change of fracture morphology in hydrogen gas had a direct influence on the fatigue crack growth rate. (3) A good correspondence between the striation spacing and the fatigue crack growth rate was observed in both air and hydrogen gas except for Ni-Cr-Mo-V steel.
Sustained load tests were performed on a sensitized high-strength aluminum alloy with repetitional stresses of small amplitude and of constant low strain rate superimposed in 3.5% NaCl solution. The crack growth rate of stress corrosion cracking (SCC) remarkably accelerated under repetitional stresses of a high strain rate K (dynamic SCC), whilst SCC crack growth rate greately decreased under repetitional stresses of a low K owing to the dissolution-induced crack tip blunting. SCC crack growth rate varied much even under a constant K condition. Especially da/dt varied surprisingly under repetitional stresses of a low strain rate K, which may be explained by such a process competition model that one of the processes of fatigue crack growth by repetitional stresses, dissolution-induced crack tip blunting and SCC crack growth by a damage of passive films occurs exclusively.
Propagation of surface cracks under creep-fatigue conditions at elevated temperatures was studied with a SUS 304 steel and a 21/4 Cr-1Mo steel. The shape of cracks was semi-circular in most cases and the propagation rate was the same in the surface direction as in the depth direction. The creep crack propagation rates of surface cracks under constant stresses agreed with those of through-thickness cracks in the J diagram. Microscopic fracture morphology was predominantly intergranular creep type. In the cases of repeated stresses, similarly to the case of through-thickness crack previously reported, crack propagation was controlled by ΔK in the low stress region and by J in the high stress region, microscopic fracture morphology being fatigue striation type and creep dimple type, respectively. This is true, however, only for large cracks and, in the case of small cracks, crack propagation was controlled by J even in the low stress region, microscopic fracture morphology being still fatigue striation type.