High-Temperature Rupture of Microstructurally Unstable 304 Stainless Steel Under Uniaxial and Triaxial Stress States HO-KYUNG KIM, FARGHALLI A. MOHAMED, and JAMES C. EARTHMAN Specimens of 304 stainless steel were tested to failure under two different stress states, uniaxial tension using smooth bar specimens and triaxial tension using notched bar specimens. The tests were conducted at a temperature that gives rise to carbide particle growth which, in turn, leads to microstructural softening. Rupture times are compared for uniaxial and triaxial stress states with respect to multiaxial stress parameters that are directly related to physical mechanisms. The success of the parameters is judged according to how well the rupture times of notched specimens can be predicted using the rupture data for specimens under uniaxial tension. The data indicate that the rupture time is not governed by deformation processes, despite evidence for substantial softening by particle coarsening. The results further suggest that the creep rupture process is dominated by cavitation that is coupled with localized shear deformation along the inclined grain boundaries. I. INTRODUCTION INTERGRANULAR fracture by cavity growth and co- alescence has long been known to be an important failure mechanism in high-temperature components. Most in- vestigations of this mode of failure have been performed with specimens tested under uniaxial tension. Although uniaxial stress experiments have led to a better under- standing of the physical processes involved, they do not provide sufficient information to predict cavity growth and creep rupture under multiaxial stress states. Bending and torsion are examples of loading conditions that can cause multiaxial stresses in smooth components. Notches and other geometric irregularities typical of engineering components can also produce multiaxial stresses when the remote loading condition is purely uniaxial. Thus, multiaxial stresses must be addressed for accurate pre- dictions of high-temperature rupture in many engineer- ing structures. Hayhurst and his colleagues 11-41 proposed that for a smooth cylindrical specimen subjected to uniaxial ten- sion, the rupture lifetime at a given temperature can be expressed as tf = M6r -x [11 where ~r is the applied uniaxial stress and M and X are constants that characterize the evolution of damage at the temperature in question. The motivation for considering multiaxial stresses is demonstrated by the fact that Eq. [ 1 ] does not correctly predict creep rupture of notched bars even if the nominal stress in the notch is used in the expression. [2] Consequently, representative stress pa- rameters are substituted for or in Eq. [1 ] to predict creep life data under'multiaxial stress states with rupture data obtained with specimens under uniaxial stresses. Several parameters which contain adjustable "constants" have HO-KYUNG KIM, Graduate Research Assistant, FARGHALLI A. MOHAMED, Professor, and JAMES C. EARTHMAN, Assistant Professor, are with the Materials Section, Department of Mechanical and Aerospace Engineering, University of California, Irvine, CA 92717. Manuscript submitted December 7, 1990. been proposed to correlate rapture times for different stress states, p,5-71 However, the present work only considers multiaxial stress parameters that are rigid in the sense that their determination does not involve adjustable terms. These parameters were chosen because each is linked to a particular set of physical mechanisms that could con- trol the rupture process. Evaluation of the validity of these nonadjustable stress parameters has been performed in the present investigation to assist in the development of a better understanding of the mechanisms that control high-temperature rupture. These stress parameters are briefly discussed in the following. A. The Maximum Principal Stress It is well founded that the diffusive growth of inter- granular cavities is driven by the tensile stresses acting normal to the grain boundaries. This, combined with the general observation that intergranular fracture usually occurs first on grain boundaries that are perpendicular to the maximum principal stress, suggests that this com- ponent of the stress state plays a central role in high- temperature rupture. The role of the maximum principal stress will dominate if cavity nucleation is easy and the cavities are homogeneously distributed on all grain boundaries. In this case, cavitation would not be con- strained by creep deformation, and cavity growth rates would be limited only by the magnitude of the tensile stresses driving the diffusive cavity growth process. B. The von Mises Effective Stress Microstructural evidence reported by several investi- gators has shown that cavitation in many metals and al- loys is not homogeneously distributed and that cavitating grain boundary facets are often isolated from one an- other. Dyson t8] first suggested that for inhomogeneously distributed cavitation, the overall rate of the diffusive cavity growth process is governed by the creep rate of the surroundings. In this case, the shear stresses that drive creep deformation should also have a role in governing the rate of cavity growth. 191 This provides the rationale METALLURGICAL TRANSACTIONS A VOLUME 22A, NOVEMBER 1991 2629