Modelling thermodynamic properties of materials is a strong tool for predicting phase transformations during materials production and processing to save time and costs.
The Calphad method is a very powerful approach for describing different systems based on thermodynamic properties. In this method, model parameters are fitted to experimental data to develop databases. These databases can be used for predicting phase transformations in different systems and modelling microstructure evolution by time. This possibility is a valuable tool for life assessment of the materials used in harsh environment and as a result, in serious exposure to corrosion and oxidation.
Corrosion and oxidation are two main types of materials degradation mechanisms in different applications. Thus, preventing them can save a lot of cost. A great amount of experimental measurements and microscopic investigation have been done in the High Temperature Corrosion Center (HTC) at the chemistry department at Chalmers during the last two decades, to study these phenomena. However, the complicated nature of such phenomena has made it difficult to model them from computational point of view.
Recent advances in highly efficient computational schemes have enabled the calculation of thermodynamic and diffusion properties. Using these new possibilities and by the help of new techniques [1,2], the oxidation of pure Fe was shown to be accurately modeled with an excellent agreement with the experimental data . Although new advances in the computational possibilities have made is possible to model oxidation process, it is still very time consuming on an ordinary computer.
This project is a continuation of the project SNIC 2018/3-51 in which high-temperature oxidation of elements Fe, Cr, Ni, Al, their bianry and ternary systems were investigated. Agreement between the simulations results and experimental data is very promising and we attempt to study industrial cases in the next step, i.e. longer exposure times, to predict service life of the powerplant components.
We also have started to use atomistic simulation methods, i.e. DFT, to find the interaction of different elements in the mechanisms of the corrosion and provide input for thermodynamic modelling, if needed.
For these types of calculations to be time and cost wise accountable, parallel cores should be used on a very high usage/memory. For this reason, we are applying for a medium size project on the Hebbe cluster.
 H. Larsson, A. Engström, Acta Mater. 2006, 54, 2431.
 H. Larsson, L. Höoglund, Calphad 2009, 33, 495
 H. Larsson, T. Jonsson, R. Naraghi, Y. Gong, R. C. Reed and J. Ågren, Materials and Corrosion, vol. 68, No. 2, p. 133, 2017.