J. E. Schaffer, Fort Wayne Metals Research Products Corporation, Fort Wayne, IN
Fatigue life variability in fine diameter medical grade wire is highly dependent upon many variables including melt techniques, finish microstructural parameters, upstream and finish strain hardening conditioning, and sample preparation techniques. The ASTM F562 alloy system is used extensively as a fine wire coil or cable in the production of cardiac rhythm management leads which require excellent fatigue life, and a good understanding of life variability. In the present investigation, samples of 177 µm diameter ASTM F562, 35Co-35Ni-20Cr-10Mo wires were produced with a variety of grain sizes and strain hardening conditions. Samples were cyclically loaded to failure in rotary beam testing and preserved for fractography using high resolution scanning electron microscopy (HR-SEM). Cracks were found to initiate from three sources: intrinsic microstructural inhomogeneities, persistent slip bands, and extrinsic surface defects. A hierarchical model is proposed that accounts for mixed-mode initiation and estimates total life based on both probabilistic and deterministic input parameters. The resultant predictions were found to yield good correlation to experimental data.
Summary: The ASTM F562 alloy system is used extensively as a fine wire coil or cable in the production of cardiac rhythm management leads which require excellent fatigue life, and a good understanding of life variability. In the present investigation, samples of 0.0070 inch diameter ASTM F562, 35Co-35Ni-20Cr-10Mo wires were produced with a variety of grain sizes and strain hardening conditions. Rotary beam fatigue samples were studied using both SEM and EBSD analysis techniques. Cracks were found to initiate from three sources: intrinsic microstructural inhomogeneities, persistent slip bands, and extrinsic surface defects. A hierarchical model is proposed that determines the dominant initiation mechanism and proceeds to estimate total life based on both probabilistic and deterministic input parameters. The resultant predictions were found to yield good correlation to experimental data.
