Materials Week '97: Wednesday PM Session

Focusing on physical metallurgy and materials, Materials Week '97, which incorporates the TMS Fall Meeting, features a wide array of technical symposia sponsored by The Minerals, Metals & Materials Society (TMS) and ASM International. The meeting will be held September 14-18 in Indianapolis, Indiana. The following session will be held Wednesday afternoon, September 17.

HIGH CYCLE OF FATIGUE OF STRUCTURAL MATERIALS: Session IV: Fatigue Thresholds and Crack Growth

Section Chairs: Prof. Robert Wei, Lehigh University, Dept. of Mechanical Engineering, Bethlehem, PA 18015, Prof. Huseyin Sehitoglu, University of Illinois, Dept. of Mechanical Engineering, 1206 W Green, Urbana, IL 61801

PREFERRED LOCATIONS FOR FATIGUE CRACK INITIATION: P. Neumann, Max-Planck Institut fur Eisenforschung GmbH, Dusseldorf

Crack initiation in high cycle fatigue either at inclusions, at other faults with reduced strength or in localized slip bands. The size of inclusions and faults can be sufficiently reduced by process metallurgy in order to become irrelevant. Localized slip bands can be reduced in size by grain refinement. In a given microstructure they will form at those locations where stress concentrations are highest. Most metals are elastically anisotropic. Therefore most grain boundaries are elastically incompatible, act as stress raisers and therefore should be preferred locations for fatigue crack initiation. As an illustration 3-dimensional FEM calculations of the stress concentrations at the grain boundary in a bicrystal of NiAl will be presented. Strain controlled high cycle fatigue experiments in an austenitic and a ferritic stainless steel were performed in order to study the significance of grain boundaries for fatigue crack initiation. The relative frequency of cracks initiating at grain boundaries and grain boundary triple junctions increased with decreasing load amplitudes. In the high cycle i.e. low plastic strain regime (cycles to failure, N_f > 3.105) the most predominant initiation sites were found to be grain boundaries. In case of the austenitic steel these cracked boundaries were twin boundaries almost exclusively. In order to correlate the observation of individual events of crack initiation with local stress concentrations, the local orientation of more than 270 grains - cracked or uncracked - was measured by a modified electron channeling technique. In a simplified model these orientation data were used to calculate the local stress concentrations near each individual grain boundary. For the austenitic steel there is an excellent correlation between the values of the stress concentrations and the occurrence of crack initiation. Moreover the model explains the surprising but often observed phenomenon that cracks initiate only at every other boundary in a stack of lamellar twins.

1:25 pm INVITED

MICROSTRUCTURAL FRACTURE MECHANICS IN HIGH-CYCLE FATIGUE: E.R. de los Rios, Dept. of Mechanical Engr., Univ. of Sheffield, Sheffield S1 3JD U.K.; A. Navarro, Escuela Superior de Ingenieros de Sevilla, 41012 - Sevilla, Spain

Microstructural Fracture Mechanics principles are used to develop a model of crack growth in long life fatigue. In its simplest form microstructural modeling considers the material as a polycrystal of uniform grain size D, with a crack system divided into three zones: the crack, the plastic zone and the microstructural barrier zone. The crack may have various degrees of closure, introducing friction stresses on the crack flanks s₁, while the material's resistance to plastic deformation s₂ incorporates the work hardening characteristics of the material. The conditions for crack propagation establishes that the stress at the microstructural barrier s₃ should achieve the level of the barrier strength. The latter incorporates the effect of crystallographic orientation and the transfer of plasticity across barriers. The solution of the equilibrium equation allows for the calculation of the stresses sustained by the crack wake, plastic zone, barrier zone and elastic enclave, and the crack tip plastic displacement. Crack growth rate is calculated through a Paris type relationship in terms of f, i.e./dN = Cf_n.

1:50 pm INVITED

IN-SITU AFM OBSERVATIONS OF FATIGUE INDUCED SURFACE MICROPLASTICITY EVOLUTION: W. Gerberich, W. Geng, D. Kramer, S. Okerstrom, Depts. of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455

Evaluating the importance of surface microplasticity evolution to fatigue initiation is necessary for microplastic damage accumulation theories. Four point bending and uniaxial C-C fatigue testing are both employed in these experiments. Commercially pure and Grade IV titanium have been subjected to fully reversed cyclic loading and analyzed using atomic force and scanning electron microscopies. An in-situ AFM method has been developed here, which greatly improves the data quality from rough surfaces to make the results more reliable and repeatable. The evolution of persistent slip band spacing and height has been measured by observing the same position at intermittent periods over the fatigue life of the sample. As slip band spacings decrease, heights increase, however spacings have been observed to saturate with repeated cycling. How this fits into one damage accumulation model will be presented.

2:15 pm

FATIGUE INITIATION AS UNDERSTOOD FROM AFM OBSERVATIONS OF MICROPLASTICITY EVOLUTION: W. Geng, D. Kramer, S. Okerstrom and W. Gerberich, Depts. of Chemical Engineering and Materials Science, University of Minnesota, Minneapolis, MN 55455

Four point bending and uniaxial fatigue testing have been used to study the evolution of surface microplasticity. Commercially pure grades of titanium have been subjected to fully reversed cyclic loading and analyzed using scanning electron microscopy and atomic force microscopy. The evolution of persistent slip band spacing and height has been followed by observing the same position at intermittent periods over the fatigue life of the sample. As slip band spacings decrease, heights increase; however spacings have been observed to saturate with repeated cycling, while heights initially continue to increase. This behavior is highly dependent upon the orientation of the individual grains with respect to the applied stress axis. Grains that do not show signs of initial surface slip may activate with further cycling as existing slip bands saturate. The magnitude of surface microplastic damage is used along with a critical step size criterion to predict the number of cycles required for fatigue crack initiation.

2:35 pm

ON THE CRACK GROWTH MECHANISM IN THE THRESHOLD REGIME OF FATIGUE CRACKS: F.O. Riemelmoser, Erich-Schmid-Institt fur Festkorperphysik, Leoben, Austria

The conventional explanation of the unusual behavior of fatigue cracks in the threshold regime is that the crack grows intermittently and cleavage-like. However, the experimental results of Hong and Laird (1991) and Thompson (1996) show that the patterns on the surface of a specimen broken in the threshold regime do not necessarily, stem from a cleavage like process. Therefore we questioned whether the drop in the crack growth curve in the threshold regime can be understood when only the dislocation-nature of plasticity is taken into account. Note that the crack growth rate near the threshold is just in the order of some Burgers vectors. We simulate crack tip plasticity by means of a discrete dislocation model. For large stress intensity ranges the crack growth curve of the dislocation model runs into the Paris-line, but for small DK the crack growth rate is smaller than predicted by the Paris relation. The reason is that discrete nature of plasticity. The dislocation model also predicts a minimum striation spacing and A dislocation pattern which is in accordance with measured ones. In the presentation the intermittent model and the dislocation mode are compared.

2:55 pm INVITED

GRAIN SIZE INFLUENCE ON THE GROWTH OF LONG FATIGUE CRACKS: Basic Physics and Experimental Data: A.K. Vasudevan, Office of Naval Research, Arlington, VA 22217; K. Sadananda, Naval Research Labs, Washington, DC 20375

We observe that fatigue damage has to be represented by two driving force parameters like DK(cyclic) and K_max (static). Of these two driving forces, K_max has a strong dependence on microstructural and environmental variations. This is because K_max is directly related to the breaking of the atomic bonds at the crack tip resulting in crack extension. In this paper, we illustrate the basic physics of the problem, current interpretations and some modifications into newer ones.

3:20 pm BREAK

3:35 pm

BEHAVIOR OF SMALL SURFACE CRACKS IN HIGH-CYCLE FATIGUE OF A Ti-8Al ALLOY: IMPLICATIONS OF CRACK SHAPE AND CLOSURE: K.S. Ravichandran, Department of Metallurgical Engineering, 412 WBB, The University of Utah, Salt Lake City, UT 84112

Fatigue crack growth behavior of surface cracks were investigated in Ti-8Al alloy. The variations in crack aspect ratio induced by grain boundaries, as well as closure behavior were studied. Aspect ratios were determined from crack compliance data and the surface crack length data, collected continuously during the fatigue tests. The variations in compliance, surface length and the calculated aspect ratios as a function of crack growth were found to be sensitive to the size of the surface crack relative to the grain size of the material. Fractographic analyses of fracture surfaces were performed to correlate the growth behavior to fracture micro-mechanisms and crack aspect ratios. It is shown that when surface cracks exhibit crack shape variations, conventional methods of calculation of small-crack data, performed with an assumption of a/c=1, can result in errors in DK calculation and hence scatter in crack growth data. As an alternative, the basis for the correlation of small crack data in terms of compliance is explored. The closure behavior of small cracks and its role in explaining the relatively faster growth rates of surface cracks are also discussed.

3:55 pm INVITED

FATIGUE CRACK GROWTH--A MODELING PERSPECTIVE: H. Sehitoglu, Department of Mechanical and Industrial Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801

The fatigue crack advance in metallic materials at the micro-scale occurs as a result of localized plastic flow. Depending on the orientation of the crack plane with respect to the crystallographic orientation, the flow bands can be inclined at different angles with respect to crack growth direction. The orientation angles decide the plastic strain magnitudes, the residual stresses and the crack opening profiles and hence the crack growth rates. Despite this background of local slip at the micro-level, the current models of fatigue crack growth have relied on a description of crack tip stress fields assuming plastic strain description independent of orientation. Consequently, a quantitative description of the local orientation effects on fatigue crack growth has not been proposed. The present work addresses this issue. In the second part of the work, the irregularity of fatigue crack surfaces on the fatigue crack advance will be considered. A model is developed that accounts for the sliding and plastic flow of asperities. During this process, the contact area, contact load changes non-linearly with separation distance between crack surfaces. This nonlinear response develops even in the presence of `pure' elastic deformations. The presentation will provide an overview of micro-contact and fatigue crack closure analysis.

4:20 pm INVITED

THE INFLUENCE OF DEFORMATION-INDUCED MARTENSITE ON FATIGUE CRACK GROWTH IN METASTABLE AUSTENITIC STEELS: J.W. Morris, Jr., Department of Materials and Minerals Engineering, University of California, Berkeley, California 94720; Z. Mei, Hewlett Packard Co., 1501 Page Mill Rd., Palo Alto, CA 94303

The research reports an investigation of the influence of the mechanically induced martensitic transformation on the fatigue crack growth rate in 304-type austenitic stainless steels. The steels 304L and 304LN were used to test the influence of composition, the testing temperatures 298K and 77K were used to study the influence of test temperature, and various load ratios were used to determine the influence of the mean stress. It was found that decreasing the mechanical stability of the austenite by changing composition or lowering temperature reduces the fatigue crack growth rate and increases the threshold stress intensity for crack growth. However, this beneficial effect diminishes as the ratio increases, even though increasing the load ratio increases the martensite transformation. Several mechanisms that may affect this phenomenon are discussed, including the perturbation of the crack-tip stress field, crack deflection, and the work hardening characteristics and relative brittleness of the transformed material. The perturbation of the stress field seems the most important; by modifying previous models we develop a quantitative analysis of the crack growth rate that provides a reasonable fit to the experimental results.

4:45 pm

A REEXAMINATION OF THE EFFECTS OF OVERLOADS AND UNDERLOADS: K. Sadananda, Code 6323 Naval Research Laboratory, Washington, D.C. 20375; K.V. Jata, WL/MLLM, WPAFB, Dayton, OH 45433; A.K. Vasudevan, Office of Naval Research, Arlington, VA 22217; B. Biner, Ames National Lab, MD #208, Ames, IA 50011

It is well known that application of overloads or underloads during crack growth results in deceleration or acceleration of crack growth. Understanding and predicting these effects are necessary for fatigue life prediction of a component in service that is subjected to variable amplitudes. The retardation and acceleration effects have been, in the past, attributed to crack closure. Since our recent analysis indicated that crack closure effects are negligible, we provide an alternative explanation to account for the observed effects.

5:05 pm INVITED

CHARACTERIZATION OF FATIGUE CRACK INITIATION FROM CORROSION PITS IN Al 2024 -T3: S.I. Rokhlin, H. Nagy, Dept. of Industrial Systems & Engineering, The Ohio State University, Columbus, OH 43210

It is well known that corrosion pitting has strong effect on fatigue life of Al alloys. The statistical and empirical analyses have shown deviations from the experimental results especially at small pit sizes. This is mainly due to the poor understanding of the crack initiation from the pits. In the current work, fatigue crack initiation from actual corrosion and simulated pits has been studied. The effect of pit size was investigated using direct fractographic observations and nondestructive measurements. Ultrasonic, acoustic emission and micro-radiographic techniques were used to monitor crack initiation and propagation. The results were analyzed using fracture mechanical models including those for small cracks.

5:30 pm

ROLE OF MICROCRACKS IN HIGH CYCLE FATIGUE DAMAGE OF AN AL-Si COMPOSITE: E.Y. Chen¹, L. Lawson², M. Meshii³, ¹Physical Metallurgy Laboratory, GE Corporate Research & Development, P.O. Box 8, Schenectady, NY 12301; ²226 Interstate Parkway, Bradford, PA 16701; ³Department of Materials Science and Engineering, Northwestern University, Evansotn, IL 60208

Advanced Al-SiC composites are considered potential candidates for replacing monolithic metals in high cycle fatigue (HCF) applications such as aircraft wing skins and automotive engine connecting rods. To assess chair aptitude in such instances, this study examines the role of microcracks in the HCF damage and critical crack formation process of a X2080 Al -15 vol.% SiCp composite. Microcracks are important in fatigue since their growth (or lack of growth) greatly determines fatigue strength. In the low cycle fatigue (LCF) of this composite, the microcrack regime can dominate for over 60% of the fatigue life. In HCF, this is still often the case. While microcracks may initiate within the first 10% of the life, most retard immediately afterwords and microcrack development can exceed 70% of the life. These and other topics such as microcrack growth rates, coalescence, and fatal crack formation in HCF will be discussed in comparison to those obtained under LCP conditions. These results will emphasize the implication of microcracks when designing for fatigue strength and reliability inspectability.

5:50 pm

CORROSION FATIGUE CRACK PROPAGATION IN ALUMINUM ALLOYS: X. Zheng, R. Wang, Dept. of Materials Science and Engr., Northwestern Polytechnic University, Xi'an Shaanxi Province 710072, China

Fatigue tests were carried out to measure the corrosion fatigue crack propagation (CFCP) rates of widely used aluminum alloys in 3.5% NaCl solution environment. Test results of CFCP rates of aluminum alloys were analyzed by using the following equation: da/dN = B_cf[DK-DK_cf]² where B_cf and DK_cf are defined as the CFCP coefficient and threshold, respectively. Test results and analysis show that the corrosive environment increases the value of B_cf but has no remarkable effect on the value of DK_thcf of aluminum alloys at various stress ratios. Therefore, the corrosive environment increases the CFCP rates in the intermediate region, i.e. the range of da/dN = 10^-5 -10^-3 mm/cycle, but has little effect on the CFCP rates in the near-threshold region. Test results and analysis also show that DK_thcf decreases and the CFCP rates, especially in the near-threshold region increases with increasing stress ratio. The test results of CFCP rates of aluminum alloys obtained at two loading frequencies f= 10Hz and f= 1Hz have no appreciable difference in the intermediate region. The increase of B_cf in 3.5% NaCl environment could be attributed to the change of FCP mechanism in the intermediate region, and the effects of 3.5% NaCl environment on DK_thcf and thus the CFCP rates in near-threshold region may be thought due to the crack closure and the crack tip blunting included by the corrosion of the metal at the crack tip.