In-situ micromechanical testing of interfaces for multiscale modeling of fracture

Research Project, 2016 – 2019

Whereas computational tools and method has increased the efficiency of many other parts of the technology or product development process, the development and qualification of new materials is still largely based on the empirical trial-and-error approach. Therefore, materials development has in many cases gone from being a facilitator of new and improved technologies to a rate-limiting step. Multiscale materials modeling is a promising tool to address the immediate need for an accelerated materials development process. However, in order to develop and implement reliable and robust multiscale models, it is essential be able to calibrate and/or validate all components of the models on their respective length scales. The purpose of the proposed research is to provide a missing link in the chain for multiscale modeling of materials where interfacial fracture plays an important role. This is done by developing an experimental method to test the cohesive properties of individual grain or phase boundaries, which allow continuous imaging of the fracture process. The method is based on the manufacturing of micron-sized cantilevers using a focused ion beam. The cantilevers will contain single interfaces of known structure and orientation, and will be subjected to bending until failure inside an electron microscope using micro-robots. In the project, the method will be developed and demonstrated on a Ni-based superalloy known to suffer from quasi-brittle grain boundary fracture. A large number of interfaces with varying structure and orientation will be tested, and the tests will be simulated using combined crystal plasticity and cohesive zone models in a finite element framework. In this way, the correlation between interface microstructure and cohesive properties will be determined, and the models will be calibrated accordingly. Finally, the calibrated model will be used to simulate macroscopic tests of poly-crystal specimens to validate the method. The immediate result of such an approach is that the effect of observed variations in the microstructure on the macroscopic properties can be predicted. It can thus be used for judging the suitability of materials and/or processes, both for virtual property optimization and for addressing the effect of changes in e.g. manufacturing processes, directly contributing to reductions in materials development times. In the long term it is intended to also be used for validation of quantum mechanical or atomistic models, so that their accuracy can be assessed before they are included in a multiscale modeling framework. In this way, the proposed method will be an integral contribution to solving the problem of experimental component model validation on all relevant length scales, thus facilitating the development of robust and reliable multiscale models suitable for industrial use. The main part of the project will be the development of equipment and methods in order to achieve repeatable and efficient testing.