Healthy cortical bone tissue is both tough and strong and has a unique ability to resist fracture. One reason is the hierarchical structure of the tissue where toughening mechanisms at all length scales act to slow down or stop a propagating crack. The most potent toughening mechanisms arise at the microscale when cracks interact with the osteonal microstructure and deflect along weak interfaces. However, cortical bone is a living material and the tissue properties change over time. With aging the properties are known to deteriorate. Yet, the link between age-related structural and compositional changes and impaired fracture resistance in old bone is not fully known. This is key for understanding and being able to predict the increased risk for fracture with age and in patients with osteoporosis.

The aim of this thesis is to understand the role of microstructure for crack propagation in cortical bone. Both experimental and numerical techniques have been used to evaluate the importance of mechanical properties and microstructural distributions for how cracks interact with the microstructure. In the experimental part, in situ loading in combination with digital image correlation and small- or wide-angle x-ray scattering was used to simultaneously measure deformation at meso- and nanoscale in cortical bone. Micro-CT analysis of the bone samples was performed after the tests and showed that the crack trajectory to a large extent followed the microstructure. In the numerical part, the extended finite element method (XFEM) was adopted to explicitly simulate crack propagation in cortical bone at the microscale. The key feature is a new interface damage model in 2D that captures crack deflections at osteon boundaries, as seen in experiments and which previous XFEM models were not able to predict. The modelling framework has been applied to simplified geometries comprising one osteon with different orientations to look at the effect of the microstructural distribution. These models have also been used in a parameter study to identify important material properties for the fracture behavior using a statistical framework with a Design of Experiments approach. The results identified factors related to the cement line to influence the crack propagation, where the interface strength was important for the ability to deflect cracks. Furthermore, the results illustrated how reduced matrix toughness promoted crack penetration of the cement line. However, the cement line properties are not well determined experimentally and need to be better characterized. Additionally, the models have been applied to more realistic scenarios, where crack paths were simulated in microstructural geometries based on microscopy images of cortical bone. The aim was to mimic the aging process in cortical bone and investigate how increased porosity and reduced fracture energy in the tissue influenced the fracture behavior. Both factors resulted in straighter crack paths with cracks penetrating osteons, similar to what is seen experimentally for old cortical bone. However, only the latter predicted a more brittle failure behavior. In the final study, the interface model was applied at the mesoscale to simulate crack propagation in larger domains compared to the previous models. In this case, orientation maps based on micro-CT images from the experimental study were converted to FE-models to introduce the microstructure in the models. The simulated crack paths captured the trends seen in experiments with more irregular cracks predicted for crack propagation perpendicular to the osteon orientation.

In summary, the work in this thesis illustrates the important role of cortical bone microstructure in providing alternative crack paths in the tissue. It also highlights the need of techniques to study local damage both in experimental and numerical settings, and the gain of using computational models as complementary methods to experimental techniques to identify and distinguish effects of different damage mechanisms in cortical bone.