A correlative approach to evaluating the links between local microstructural parameters and creep initiated cavities

The study and modeling of material degradation processes, such as the initiation and growth of creep cavities in high-temperature applications, require a correlative and comprehensive knowledge of the microstructure. However, individual microscopy is limited to a small region and specific microstructural information of the specimen. This work demonstrates a novel correlative microscopy approach for characterising creep cavitation and establishing correlations with local microstructural parameters in a statistical manner. This approach combines datasets
from stitched higher-resolution backscattered electron (BSE) images, XeF2 FIB images, and backscattered electron
diffraction (EBSD) maps with advanced image correlation techniques. Deep-learning image segmentation techniques
and statistical analysis are applied to find relations between creep cavitation and local microstructural environment. This approach is demonstrated in a cyclic creep-tested 316H stainless steel specimen with extensive creep cavities. The results show that in this material, strain localization, grain boundary misorientation, and substantial precipitation dominate the nucleation of cavities, whereas other microstructural properties such as grain size and Schmid factor play smaller roles. This study presents the use of the correlative microscopy approach to provide new insights into creep cavitation behaviour and its implications for establishing creep cavitation damage models.