Fatigue crack propagation in carburized X-2M steel

The growth rates of fatigue cracks propagating through the case and into the core have been studied for carburized X-2M steel (0.14 C, 4.91 Cr, 1.31 Mo, 1.34 W, 0.42 V). Fatigue cracks were propagated at constant stress intensities, ΔK, and also at a constant cyclic peak load, and the crack growth rates were observed to pass through a minimum value as the crack traversed the carburized case. The reduction in the crack propagation rates is ascribed to the compressive stresses which were developed in the case, and a pinched clothespin model is used to make an approximate calculation of the effects of internal stress on the crack propagation rates. We define an effective stress intensity, K_e = K_a + K_i, where K_a is the applied stress intensity, K_i = σ_id _i ^1/2 , σ_i is the internal stress, and d_i is a characteristic distance associated with the depth of the internal stress field. In our work, a value of d_i = 11 mm (0.43 inch) fits the data quite well. A good combination of resistance to fatigue crack propagation in the case and fracture toughness in the core can be achieved in carburized X-2M steel, suggesting that this material will be useful in heavy duty gears and in aircraft gas turbine mainshaft bearings operating under high hoop stresses.     

Fatigue crack propagation in carburized high alloy bearing steels

Fatigue cracks were propagated through carburized cases in M-50NiL (0.1 C,4 Mo, 4 Cr, 1.3 V, 3.5 Ni) and CBS-1000M (0.1 C, 4.5 Mo, 1 Cr, 0.5 V, 3 Ni) steels at constant stress intensity ranges, ΔK, and at a constant cyclic peak load. Residual compressive stresses of the order of 140 MPa (20 Ksi) were developed in the M-50NiL cases, and in tests carried out at constant ΔK values it was observed that the fatigue crack propagation rates,da/dN, slowed significantly. In some tests, at constant peak loads, cracks were stopped in regions with high compressive stresses. The residual stresses in the cases in CBS-1000M steel were predominantly tensile, probably because of the presence of high retained austenite contents, andda/dN was accelerated in these cases. The effects of residual stress on the fatigue crack propagation rates are interpreted in terms of a pinched clothespin model in which the residual stresses introduce an internal stress intensity, K_i where K_i, = σ_id _i ^1/2 (σ_i = internal stress, d_i = characteristic distance associated with the internal stress distribution). The effective stress intensity becomes K_e = K_a + K_i where K_a is the applied stress intensity. Values of K_i were calculated as a function of distance from the surface using experimental measurements of σ_i and a value of d_i = 11 mm (0.43 inch). The resultant values of K_e were taken to be equivalent to effective ΔK values, andda/dN was determined at each point from experimental measurements of fatigue crack propagation obtained separately for the case and core materials. A reasonably good fit was obtained with data for crack growth at a constant ΔK and at a constant cyclic peak load. The carburized case depths were approximately 4 mm, and the possible effects associated with the propagation of short cracks were considered. The major effects were observed at crack lengths of about 2 mm, but the contributions of short crack phenomena were considered to be small in these experiments, since the two steels were at high strength levels, and short cracks would be expected to be of the order of 10 μm. Also, the two other steels behaved differently and in a way which followed the residual stress patterns. Both M-50NiL and CBS-1000M have a high fracture toughness, with K_lc = 50 MPa · m^1/2 (45 Ksi · in^1/2), and the carburized cases exhibit excellent resistance to rolling contact fatigue. Thus, M-50NiL, carburized, may be useful for bearings where high tensile hoop stresses are developed, since fatigue cracks are slowed in the case by the residual compressive stresses, and fracture is resisted by the relatively tough core.     

Near-threshold fatigue crack growth behavior in copper

Near-threshold fatigue crack growth rate data were developed in annealed, quarter-hard, and full-hard copper at various load ratios, (R = σ_min/σ_max). Increasing theR value decreases the resistance to threshold crack growth. At a fixed value ofR, annealed copper has the slowest near-threshold crack propagation rate while full-hard copper has the fastest crack growth rate. Waveform (sine and triangle) and specimen geometry (WOL, CT, and CCT) do not appear to affect the rates of near-threshold crack propagation. The influences of load ratio and material strength on threshold crack growth behavior can be rationalized by crack closure.     

Metallography of fatigue crack initiation in an overaged high-strength aluminum alloy

Crack initiation was observed by optical microscopy using Nomarski interference contrast during fatigue cycling of an overaged 2024 aluminum alloy. The number of cracks more than five microns long at any given fraction of the fatigue life, and the distribution of cracks among various possible initiation sites, both depend on the applied stress amplitudeσa). The crack density at failure falls from approximately 300/mm² when σ^a is 90 pct of the yield strength, to less than I/mm² when σ^a is less than 60 pct of the yield strength. Cracks may begin in the matrix, in grain boundaries, or at constituent particles. At all stress amplitudes, however, the most common initiation sites areβ (Al₇Cu₂Fe) constituent particles. At low stress amplitudes in particular, fatigue cracks develop from the interface between closely-spaced fragments of β particles broken during prior processing (cluster sites). The stress-raising effect of voids which often occur at cluster sites may be responsible for their effectiveness in initiating fatigue cracks. Formerly Graduate Students in the Department of Metallurgical Engineering at Wayne State University During 1982–83 he is on leave as Associate Director of the Metallurgy Program, Division of Materials Research, National Science Foundation, Washington, DC 20550.     

Initiation and growth of small fatigue cracks in a Ni-base superalloy

This article reports research on the initiation and growth of small fatigue cracks in a nickel-base superalloy (produced commercially by INCO as INCOLOY* 908) at 298 and 77 K. The experimental samples were square-bar specimens with polished surfaces, loaded in fourpoint bending. The crack initiation sites, crack growth rates, and microstructural crack paths were determined, as was the large-crack growth behavior, both at constant load ratio (R) and at constant maximum stress intensity (K _max). Small surface cracks initiated predominantly at (Nb,Ti)_xC_y, inclusion particles, and, less frequently, at grain boundaries. Small cracks grew predominantly along {111} planes in individual grains and were perturbed or arrested at grain boundaries. For values of ΔK above the large-crack threshold, ΔK _th, the average rate of smallcrack growth was reasonably close to that of large cracks tested under closure-free conditions. However, short-crack growth rates varied widely, reflecting the local heterogeneity of the microstructure. The threshold cyclic stress (Δσ_th) and the threshold cyclic stress intensity (ΔKσ_th) for small surface cracks were measured as functions of the crack size, 2c. The results suggest that a combination of the fatigue endurance limit and the threshold stress intensity for closure-free growth of large cracks can be used to define a fatigue-safe load regime. formerly with Lawrence Berkeley Laboratory     

The Effect of Grain Size on Fatigue Growth of Short Cracks

The influence of alloy grain size on growth rates of surface cracks 20 to 500 μm in length was studied in Al 7075-T6 specimens prepared in 12 and 130 μn grain sizes. Grain boundaries temporarily interrupt the propagation of cracks shorter than several grain diameters in length. Linear elastic fracture mechanics is inadequate to describe resulting average growth rates which must instead be characterized as a function of cyclic stress amplitude, σ_a, and alloy grain size as well as stress intensity range, σK. These observations are rationalized using two models, one that relates crack closure stress to alloy grain size, and a second that relates the development of microplasticity in a new grain in the crack path to grain size. In addition, growth rates were found to be faster in fully reversed loading than in tension-tension loading, especially in the large grained material. Evidence is presented to demonstrate that this is a consequence of the fatigue induced development of a compressive residual surface stress during tension-tension loading. These complex effects, and the role of grain size in determining short crack growth, are discussed.     

The fatigue of pseudoelastic polycrystalline β-cuznsn

The fatigue of polycrystalline pseudoelastic β-CuZnSn has been studied by cycling specimens to fixed stress. The fatigue life was found to decrease with increasing initial strain and decreasing specimen grain size. In both cases the results gave similar stress -vs - fatigue life curves, indicating that stress is the primary parameter controlling fatigue life. The results fitted a curve of the form △ε.N^B _f = constant, whereβ = 0.32 for the total initial strain, andβ = 0.29 for the initial elastic strain. The fatigue life appeared to be independent of strain rate. Fatigue cracks nucleated in the first cycle at three grain intersections and grew along grain boundaries until adjacent cracks linked up. In the later stages of crack growth, some intergranular cracking occurred when there were no suitably oriented grain boundaries. Both the intergranular and transgranular regions showed somewhat ill-defined fatigue striations.     

The growth of small fatigue cracks in A286 steel

Fatigue crack propagation in high-strength A286 steel was studied by comparing crack growth rates determined from: (1) conventional long-crack propagation tests, (2) closure-free long-crack tests at constant K_max, and (3) small-crack propagation tests. Small-crack growth rates were measured by following the growth of surface cracks in samples cycled from near-zero stress to 0.5 or 0.8σ_y. While most of the surface cracks became dormant shortly after nucleation, some grew into long cracks, and some of these propagated at cyclic stress intensities below the long-crack threshold, ΔK_th (or ΔK _th ^eff , the threshold cyclic stress intensity after crack closure effects have been removed). Surface cracks grew more rapidly than long cracks at the same ΔKor ΔK_eff. The small-crack effect disappeared when the crack-tip plastic zone size became greater than the grain size. The results show that the absence of crack closure is only one of several factors that influence short-crack growth in A286 steel. Both peak stress and microstructural effects are important. Microstructural effects are apparently responsible for subthreshold crack growth; the cracks that grow at ΔK < ΔK _th ^eff form and grow in statistically weak regions of the microstructure.     

An analysis of non-propagating fatigue cracks

An analysis for the formation of nonpropagating fatigue cracks at (the base of V-shaped) notch roots, based on the considerations of the extent of the critically stressed region ahead of a notch or a crack tip, and the resulting volumetric strength effect, is developed. Assuming that the minimum local cyclic stress required for crack initiation from a notch root is equal to the unnotched fatigue limit, σ_e, and that the minimum local cyclic stress required for the propagation of the crack is equal to the theoretical strength of the material, σ_e, a model of notch fatigue limit is proposed that shows that nonpropagating cracks should form at the notch base if ρ≤ ρ⁰, a critical root radius, provided the notch is sufficiently deep,i.e. d ≥ ρ₀. The radius ρ₀ is a material constant and can be estimated from known material properties. The estimated values of ρ₀ are in fairly good agreement with available experimental values for steels and pure copper. For stresses near the notch fatigue limit it is suggested that p₀ be regarded as a radius above which notch fatigue limit is essentially initiation controlled and below which essentially propagation controlled. The notch fatigue limit based on complete fracture can then be estimated more accurately with mild as well as sharp notches. D. N. LAL, formerly a Graduate Assistant in Materials Science, Syracuse University     

Fatigue crack propagation in 99.99+ and 1100 aluminum at 298 and 77 k

The fatigue crack propagation rate,dc/dN, in cold-rolled and annealed 99.99+ Al is about 80 times slower at 77 K than at 298 K. In annealed 1100 Al which contains constituent particles,dc/dN decreases by a factor of 20 on cooling from 298 to 77 K. At 298 and 77 K, annealed 99.99⁺ Al and 1100 Al cyclically harden but the amount is greater at 77 K. Cold-rolled 99.99⁺ Al cyclically hardens at 77 K but cyclically softens at 298 K. The much slower fatigue crack propagation rate at 77 K in aluminum is attributed in part to the increase in cyclic yield stress, σ_y′, on cooling. At 77 K the high rate of work hardening at large strains is also thought to result in high plastic work per unit area of fatigue crack thereby reducing the fatigue crack propagation rate. Rice’s theory for a Mode I plane stress crack predicts the measured plastic zone size if the local stress corresponding to zero plastic strain in the cyclic stress-strain curve is employed in the formula.     

Effect of testing frequency on the corrosion fatigue of a squeeze-cast aluminum alloy

The corrosion fatigue crack propagation behavior of a squeeze-cast Al-Si-Mg-Cu aluminum alloy (AC8A-T6), which had been precracked in air, was investigated at testing frequencies of 0.1, 1, 5, and 10 Hz under a stress ratio (R) of 0.1. Compact-toughness specimens were precracked about 6 mm in air prior to the corrosion fatigue test in a 3 pct saline solution. At some near-threshold conditions, these cracks propagated faster than would be predicted by the mechanical driving force. This anomalous corrosion fatigue crack growth was affected by the initial stress-intensity-factor range (ΔK _i), the precracking conditions, and the testing frequency. The initial crack propagation rate was as much as one order of magnitude higher than the rate for the same conditions in air. This rapid rate was associated with preferential propagation along the interphase interface in the eutectic structure. It is believed that a chemical reaction at the crack tip and/or hydrogen-assisted cracking produced the phenomenon. Eventual retardation and complete arrest of crack growth after this initial rapid growth occurred within a short period at low ΔK values, when the testing frequency was low (0.1 and 1 Hz). This retardation was accompanied by corrosion product-induced crack closure and could be better explained by the contributory stress-intensity-factor range (ΔK _cont) than by the effective stress-intensity-factor range (ΔK _eff).     

Elevated-temperature fatigue crack growth

The fatigue crack growth behavior of MAR-M200 single crystals was examined at 982 °C. Using tubular specimens, fatigue crack growth rates were determined as functions of crystallographic orientation and the stress state by varying the applied shear stress range-to-normal stress range ratio. Neither crystallographic orientation nor stress state was found to have a significant effect on crack growth rate when correlated with an effective ΔK which accounted for mixed-mode loading and elastic anisotropy. For both uniaxial and multiaxial fatigue, crack growth generally occurred normal to the principal stress direction and in a direction along which ΔK _II vanished. Consequently, the effective ΔK was reduced to ΔK_I and the rate of propagation was controlled by ΔK _I only. The through-thickness fatigue cracks were generally noncrystallographic with fracture surfaces exhibiting striations in the [010], [011], and [111] crystals, but striation-covered ridges in the [211] specimen. These fracture modes are contrasted to crystallographic cracking along slip bands observed at ambient temperature. The difference in cracking behavior at 25 and 982 °C is explained on the basis of the propensity for homogeneous, multiple slip at the crack tip at 982 °C. The overall fracture mechanism is discussed in conjunction with Koss and Chan’s coplanar slip model.     

The debonding and fracture of Si particles during the fatigue of a cast Al-Si alloy

Constant-amplitude high-cycle fatigue tests (σ_max=133 MPa, σmax/σy=0.55, and R=0.1) were conducted on cylindrical samples machined from a cast A356-T6 aluminum plate: The fracture surface of the sample with the smallest fatigue-crack nucleating defect was examined using a scanning electron microscope (SEM). For low crack-tip driving forces (fatigue-crack growth rates of da/dN<1 × 10⁻⁷ m/cycle), we discovered that a small semicircular surface fatigue crack propagated primarily through the Al-1 pct Si dendrite cells. The silicon particles in the eutectic remained intact and served as barriers at low fatigue-crack propagation rates. When the semicircular fatigue crack inevitably crossed the three-dimensional Al-Si eutectic network, it propagated primarily along the interface between the silicon particles and the Al-1 pct Si matrix. Furthermore, nearly all of the silicon particles were progressively debonded by the fatigue cracks propagating at low rates, with the exception of elongated particles with a major axis perpendicular to the crack plane, which were fractured. As the fatigue crack grew with a high crack-tip driving force (fatigue-crack growth rates of da/dN>1 × 10⁻⁶ m/cycle), silicon particles ahead of the crack tip were fractured, and the crack subsequently propagated through the weakest distribution of prefractured particles in the Al-Si eutectic. Only small rounded silicon particles were observed to debond while the fatigue crack grew at high rates. Using fracture-surface markings and fracture mechanics, a macroscopic measure of the maximum critical driving force between particle debonding vs fracture during fatigue-crack growth was calculated to be approximately K _max ^tr ≈6.0 MPa √m for the present cast A356 alloy.     

Critical assessment of the local cleavage stress σ ƒ * in notch specimens of C-Mn steel

An elastic-plastic finite element method (FEM) was used to calculate the stress and strain distributions ahead of notches with various root radii in a bending specimen of C-Mn steel with grain sizes of 10 and 30 μm. By accurately measuring the distance of the cleavage initiation site from the root of the notch, the local cleavage stress σ _ƒ ^* was measured. When the notch radius increased from 0.25 to 1.0 mm, the distribution of high stress had a definite variation but the σ _ƒ ^* remained relatively constant. In notch specimens with different root radii, the critical fracture event is identical,i.e., propagation of a ferrite grain-sized crack into the neighboring matrix. Therefore, the σ _ƒ ^* is mainly determined by the length of the critical microcrack, here, the size of ferrite grain instead of the high stress volume for finding an eligible brittle particle. The critical strain for initiating a crack was about 1 pct. The cleavage site ahead of a notch was related to the relative distributions of stress and strain and the random distribution of the weakest grains. The higher fracture load of the fine-grain material can be attributed to its higher value of σ _ƒ ^* /σ_o as compared with the coarse-grain. The σ _ƒ ^* /σ_o is a potential engineering parameter for toughness assessment in notch specimens.     

Plastic work of fatigue crack propagation in steels and aluminum alloys

The plastic work per unit area of fatigue crack propagation,U, is one of the parameters controlling the rate of fatigue crack propagation,dc/dN. The equation,dc/dN = A ΔK ⁴/(σf_y ²μ U), was previously shown to fit the data for 7 iron and aluminum base alloys for the range of thedc/dN vs ΔK curve where the Paris relation is valid. Values ofU are now available for 6 additional alloys covering a much wider range of σ_y 42 to 868 MN/m². For the total populationA = (2.8 ± 0.9) X 10^-3 where 2.8 is the mean and 0.9 is the standard deviation. In this equation, σ_y is the 0.2 pct offset cyclic yield stress and μ is the shear modulus. The parameterU is related to microstructure and should be of interest to the metallurgist. Generally,U varies oppositely to σ_y due to decrease in the plastic zone size; however, the plastic strain amplitude and degree of localization of the plastic strain in the plastic zone are also important.     

Effect of surface roughness on low-cycle fatigue behavior of type 304 stainless steel

The effects of surface roughness on the low-cycle fatigue life of Type 304 stainless steel at 593°C in air have been investigated. It is observed that, at a strain rate of 4 × 10⁻³ s⁻¹ and a total strain range of 1 pct, the fatigue life (N _f cycles) decreases with an increase in surface roughness. Information on crack growthvs strain cycles has been generated, as a function of surface roughness, by the measurement of striation spacing on fractured surfaces of specimens tested to failure. Crack propagation follows the Ina ∞N (wherea is the crack length afterN strain cycles) relation for longer specimen fatigue lives (N_f > 2700 cycles) and departs from Ina ∞N for shorter fatigue lives. A quantitative estimate is made of the number of cycles N_o(r) to generate a crack length equal to 0.1 mm (≈ 1 grain diam). The initial surface roughness significantly affects only the initiation component of specimen life time. The effect of roughness on crack initiation is described byN ₀ (R) = 1012R^−0.21, whereR is the surface roughness (root-mean-square value) in microns.     

Intrinsic fatigue crack growth

The influences of microstructure and deformation mode on inert environment intrinsic fatigue crack propagation were investigated for Al-Li-Cu-Mg alloys AA2090, AA8090, and X2095 compared to AA2024. The amount of coherent shearable δ (Al₃Li) precipitates and extent of localized planar slip deformation were reduced by composition (increased Cu/Li in X2095) and heat treatment (double aging of AA8090). Intrinsic growth rates, obtained at high constantK _max to minimize crack closure and in vacuum to eliminate any environmental effect, were alloy dependent;da/dN varied up to tenfold based on applied ΔK or ΔK/E. When compared based on a crack tip cyclic strain or opening displacement parameter (ΔK/(σ_ys E)^1/2), growth rates were equivalent for all alloys except X2095-T8 which exhibited unique fatigue crack growth resistance. Tortuous fatigue crack profiles and large fracture surface facets were observed for each Al-Li alloy independent of the precipitates present, particularly δ, and the localized slip deformation structure. Reduced fatigue crack propagation rates for X2095 in vacuum are not explained by either residual crack closure or slip reversibility arguments; the origin of apparent slip band facets in a homogeneous slip alloy is unclear. Better understanding of crack tip damage accumulation and fracture surface facet crystallography is required for Al-Li alloys with varying slip localization.     

The influence of interparticle spacing on cyclic deformation and fatigue crack propagation in an aluminum-4 Pct copper alloy

A high purity Al-4 pct Cu alloy has been overaged for two different times at 400°C giving interparticle spacings (λ) of about 0.53 and 1.37 μm. Cyclic plasticity of the alloy with the smaller interparticle spacing can be explained in terms of plastic deformation behavior controlled by the structure whereas that for the alloy with the larger interparticle spacing is controlled by the matrix. The fatigue lives of the weaker alloy (λ = 1.37 μm) may be accurately predicted using the models of Coffin-Manson and Tomkins, however, these models are not applicable to the stronger alloy (λ = 0.53 μm). It was found that the crack tip opening displacement at the threshold stress intensity range (ΔK_th) was equivalent to the interparticle spacing. ΔK_th is related to the cyclic yield stress, σ_cy and the interparticle spacing in the following manner: ΔK_th ≈ (2 Eλσ_cy)^1/2, whereE is the modulus of elasticity. In the present case, the term λσ_cy is constant, giving the impression that ΔK_th is independent of the mechanical properties and microstructure. At very low growth rates, however, the fatigue crack growth is independent of these parameters and also the method of cyclic deformation. A transition to higher crack growth rates occurs when the plastic zone size reaches approximately one-seventh of the specimen thickness, allowing a nonplanar crack front to be developed. The value of the stress intensity range (ΔK_T) at this transition was found to be dependent upon the interparticle spacing according to the relation: ΔK_Tλ = 9.6 Pa-m^3/2. Formerly Lecturer and Research Associate, Department of Mechanical Engineering, University of Waterloo     

Effect of hydrogen on fatigue crack propagation in vanadium

The influence of hydrogen on fatigue crack propagation in unalloyed vanadium and several hydrogen-charged vanadium alloys has been investigated. The Paris-Erdogan equation,da/dN =C(Δ.K)^m, was approximately obeyed for all alloys. Crack growth rates were lowest in vanadium and dilute vanadium-hydrogen alloys, and were not very sensitive to volume fraction of hydrides in more concentrated alloys. The crack growth exponent,m, is inversely proportional to the cyclic strain hardening rate,n′, and the rate constantC is inversely proportional to the square of the ultimate tensile stress, σ_UTS: Metallographic examination showed hydride reorientation and growth in the originally hydrided alloys. No stress-induced hydrides were observed in V-H solid-solution alloys. Fractures in hydrided materials exhibited cleavage-like features, while striations were noted in unalloyed vanadium and dilute solid-solution alloys.     

Mechanisms of fatigue in mill-annealed Ti-6Al-4V at room temperature and 600°F

The mechanisms of the fatigue of mill-annealed Ti-6A1-4V were studied at 600°F and room temperature. Early crack initiationN ₀/N _f< 0.14) was found to occur in hcpα-grains by a slip-band mechanism under all but the least severe conditions of cyclic stress. Under cyclic stresses near the fatigue limit at room temperature, fatigue cracks began much laterN ₀/N _f ∼ 0.4) at the interface between hcpα and bccβ grains without detectable slip. Under all conditions, Stage I fatigue crack growth occupied 50 to 80 pct of the total life. Although mechanical twins were produced in profusion near the growing Stage II fatigue cracks, they appeared to play no role in crack initiation or Stage I crack growth; nor did they facilitate Stage II growth. None of the observations could be interpreted as evidence for a metallurgical instability or strain-induced phase transformation which might be harmful to the fatigue resistance of the alloy. J. C. GROSSKREUTZ, formerly with Midwest Research Institute