The Influence of Surface Features on High Cycle Fatigue in Metallic Additive Manufacturing

High cycle fatigue (HCF) testing of Inconel 718 made by laser powder-bed fusion (LPBF) and then hot isostatically pressed and heat treated demonstrates that surface features exert primary control on crack initiation and fatigue life. When machined before testing, uniform-gauge dogbone and notched LPBF 718 specimens had similar fatigue behavior to wrought material, regardless of as-printed specimen orientation (horizontal or vertical). In contrast, specimens tested with as-built surfaces have a fatigue limit at 10 million cycles that is typically 30-40% lower than the machined specimens. Surface roughness alone, however, is not adequate to capture the types of features at which fatigue cracks initiate. Horizontal and vertical notched fatigue specimens with as-built surfaces have similar fatigue behavior despite considerable differences in apparent surface roughness. Uniform gauge round dogbone specimens printed with contour scans showed a systematic bias in fatigue crack initiation within the specimen circumference relative to their azimuthal position within the printer. Life-limiting features were 50-µm individual pores within the contour scans that would not register on a surface roughness measurement. Specimens printed without contour scans using either continuous or island scanning showed strong influence of the scanning strategy, which is also due to surface conditions. Examination of fracture surfaces reveals characteristic features where fatigue cracks initiate, allowing quantitative estimates of stress intensity associated with fatigue crack initiation on as-built surfaces. These are likely to be different from those derived from testing machined standard fatigue-crack growth specimens for which threshold stress intensity is defined based on linear elastic fracture mechanics conditions that are different from those associated with as-built surfaces of LPBF materials.