Designing steels for impurity tolerance

Steel is the world’s most important engineering material, with the often stated “If it’s not made of steel, it’s made using steel”. The need to reduce CO2 emission from the steel industry is patently obvious and is driving global investment in steel production; the UK is making an abrupt change from the blast furnace to electric arc furnaces, which primarily use scrap steel as a feedstock and emit just a fraction of the CO2 content. This brings marked problems from the increased residual (“tramp”) elements (Cu and Sn) in the final steel. Daehn et al. forecast that by 2050 all new steel will contain more Cu and other tramp elements than can be tolerated and will thereby be rendered useless. This is a ticking time bomb in the steel industry. Whereas these elements are hitherto regarded as deleterious to hot shortness, ductility and toughness, we can make a virtue out of a necessity and turn these intruders to our advantage through alloy design for impurity tolerance. The metallurgical reasons for the effect of residual elements are far from clear. Residual elements have a complex effect on the mechanical properties of the steel, with significant contradictions in the literature, and a very curious observation that both the proof stress and tensile strength increase linearly with Cu content with the same slope with a small decrease in ductility. We urgently need an abrupt advance in knowledge to mitigate the role of residuals, but this will not be brought about by the classic “knowledge informed trial and error” approach: the number of variables is simply too large. The only way forward is to understand the fundamental role of residuals at the atomic scale and how they interact with other alloy additions, with migrating and transforming boundaries. This will open up robust strategies for mitigation, and indeed turn the presence of residuals to an advantage. Therefore, we will have to develop theory and modelling to understand the role of single and multiple solutes, both in the lattice and at interfaces. The use of fast and accurate machine learned atomic cluster expansion (ACE) interatomic potentials will enable the calculation of these interactions in a way that was inaccessible to previous computational techniques. Preliminary modelling work has shown an interesting interaction between Cu and Ni in Fe, indicating that Cu on its own will segregate to the boundary, but when combined, the presence of Ni will reduce the tendency of Cu to segregate. This brings a very different view on the role of Ni in Cu bearing steels than hitherto considered. Model alloys will be manufactured with targeted tramp element contents based on the outcomes of the theory and on initial proof of concept work. High temperature simulations will be made of solubility, grain boundary segregation, work of separation with a view to concentrate on the synergistic and competitive behaviour of tramp elements (Cu, Sn) and additions (C, Mn, Ni, Si, Cr) in combination. Laboratory experiments will accurately simulate the full industrial process, with site-specific high spatial resolution techniques through a combination of TEM and APT to comprehensively characterise the grain boundary segregation, solute clustering, crystal structure of precipitates, composition and coherency at the atomic scale. This new understanding will feed into industrial case studies and plant trials to deliver new tramp element tolerant green steel.