In this project we will focus on oxides and hard materials. In the case of oxide materials our primary interest is on ionic and electronic transport properties. For the hard materials we focus on interface related properties. We emphasis the atomic-scale approach and electronic-structure calculations are a key component. The project is divided into two subprojects. 1. Oxide materials The aim with this subproject is to characterize different mechanisms that affect the overall ionic conductivity. Special emphasis is directed to protonic conductors. We will do this by using DFT to study the energetics of point defect and defect complexes. In order for this to be done properly it is important to have an accurately described band gap, position of valence and conduction bands. The emphasis is the effect of finite and large concentrations of defects. Cluster expansion based techniques will be used to handle finite concentrations. Machine learned potentials that achieve close to DFT accuracy will be derived to be able tot describe the defect dynamics in an accurate way. We will also determine the Raman spectra (to all orders) for barium zirconate by direct calculations of the relevant time-dependent correlation function. This will be based on our recently developed machine-learned potential using the neuroevolution potential (NEP) approach trained with density functional theory (DFT) data. Large systems can then be simulated over long time-sclaes with near DFT accuracy. 2. Hard materials The mechanisms behind grain growth inhibition during sintering of cemented carbides from the addition of transition elements like V, Cr and Ti, often added as carbides to the powder, are not yet fully understood. Formation of WC/Co interface complexions has previously been put forward as an important mechanism at quite low doping conditions. However, experimentally it has now been established that grain growth inhibition occurs also with considerably lower Ti and V doping conditions. The aim is to study possible enrichment of dopants at steps compared to flat surfaces under realistic doping conduction using DFT combined with CALPHAD modelling. The possible enrichment of dopant atoms at steps may act as a solute drag slowing down the growth.