While there is a strong link between microstructure and hydrogen diffusion and embrittlement in all metallic systems, the correlation between the two is poorly documented. It is, however, sorely needed as the current research shows that by quantifying the type and density of defects along with their associated segregation energy for hydrogen (commonly called trapping or binding energy) the effective diffusivity can be estimated, which is in turn correlated to the embrittlement susceptibility. Attempts to model hydrogen interaction with defects at continuum scales have historically been plagued by the large scatter in the binding energies resulting from the experimental challenges associated to their measurement and by a lack of experimentally verified scaling laws linking the density of microstructural defects such as dislocations and carbides to the corresponding number of hydrogen trapping sites. In recent years however, advances in first principles modelling have enabled the determination of trapping energies, enabling the identification of the correlations between the density of microstructural features and the resulting density of hydrogen trapping sites.

Multiphase, high-strength steels, which are of particular importance due to their susceptibility to embrittlement and their ubiquity in applications where exposure to hydrogen is expected, are the main focus of this work. A combination of hydrogen charging, thermal desorption analysis and electrochemical hydrogen permeation was used in conjunction with continuum-scale hydrogen diffusion modelling to quantify the diffusion properties. The alloys studied were characterised using standard techniques including optical, scanning electron and transmission electron microscopy as well as X-ray diffraction. The studied microstructures were chosen such as to best isolate the effect of particular defects, namely vanadium carbides, grain boundaries, dislocations and retained austenite, on the overall rate of hydrogen diffusion.

The results from this work include the identification of scaling laws for hydrogen trap densities associated with vanadium carbides, dislocations and grain boundaries. In addition, two new mechanisms have been incorporated into a multi-trap diffusion model. The first is the effect of grain boundary diffusion under the assumption of local equilibrium between the hydrogen at the boundaries and that in the surrounding grains, which may be used for future studies on hydrogen diffusion in alloys in which grain boundary diffusion is faster than bulk diffusion, in particular Ni-based alloys and austenitic steels. The second effect is the incorporation of a boundary condition at the ferrite -- austenite interface which allows for the diffusion of hydrogen into austenite against the concentration gradient. The boundary condition is able to describe an interface with an associated energy barrier for hydrogen diffusion from one phase to another and includes the effect of an interface energy barrier. It was also demonstrated throughout that diffusion modelling is complementary to experimental work and is a useful tool to separate the effect of different types of defects on the rate of hydrogen trapping and diffusion.

The model has been shown to work on complex microstructures which include several different trap types and can furthermore treat bulk trapping in phases with higher H solubility such as austenite, whose effect has to date not been fully unravelled. The results presented here pave the way towards more accurate modelling of H diffusion problems in modern steels and other alloys. Examples include the tracking of H concentration in components exposed to H-rich environments, optimising heat treatments used for hydrogen release in the production of large casts and the heat treatment of welds which are prone to H-based cracking. Coupled with fracture-mechanics-based criteria the results can also be used to estimate the susceptibility to hydrogen embrittlement based on the information on the volume fraction and/or density of defects in the microstructure.