Reducing the greenhouse gas emissions from steel production can be done through direct reduction inside a shaft furnace using hydrogen gas as a reductant, generating water as an off gas. The temperature varies along the height of the shaft furnace, and studying the non-isothermal reduction is therefore necessary. In this work, industrial hematite pellets were non-isothermally reduced in a vertical tube furnace. Different gas mixtures containing water and hydrogen were used for reduction. The reduction gas used contained water vapor contents of 5%, 10%, and 20%, respectively, and the remaining gas was hydrogen. The experimental setup was carefully designed for the reductions to be carried out under well-controlled experimental conditions. It was clear that the water present in the reduction gas significantly decreased the reduction rate, especially at the lower temperatures. Moreover, the onset temperature of reduction was increased to around 525°C when water was present, compared to 450°C when pure hydrogen was used. Water contents above 5% lead to a low-rate stage at reduction degrees between 0.11 to 0.15. The low-rate stage ended when the wüstite phase became stable, changing the mechanism of reduction, which altered the chemical reaction rate. The reduction rate was less affected by water when the heating rate increased, since an increasing heating rate led to the reduction occurring at a higher temperature. Finally, the present study showed that the kinetics of non-isothermal reduction, using different water vapor contents, are very different from isothermal reduction.

To meet the future environmental challenges, hydrogen direct reduced iron (H-DRI) is expected to constitute the principal material for virgin steel production. For an efficient value chain, knowledge of the melting mechanism and dephosphorization mechanism of H-DRI is needed. The in situ melting behavior, the melting mechanism, and the dephosphorization mechanism during heating of H-DRI are investigated experimentally at 1773 and 1873 K. It is found that the melting rate of H-DRI varies with the reduction degree (91–99.5%), increasing with decreasing reduction degree. An autogenous slag forms during heating and flows through the pores of the H-DRI, thus increasing its effective thermal conductivity. The fraction of filled pores varies with reduction degree explaining the difference in melting rate. At this stage, the dissolution of apatite is initiated and completed upon melting of the metal phase. A gradual reversion of phosphorus from the autogenous slag to the liquid metal is observed after complete melting. The rate of reversion is discussed based on the properties of the H-DRI, for example, reduction degree and carbon addition.

Phosphorus partition between liquid iron and CaO-SiO2-MgO-FeOx slags have been investigated experimentally. Fe doped with Fe3P was equilibrated, at 1873 K, in a closed system, with a fully liquid, MgO saturated slag in a dense-sintered MgO crucible. Synthetic slags with low CaO/SiO2 ratio (1 to 1.32) and varying FeOx concentrations (10 to 30 wt pct) constituted the slag phase. P2O5 concentrations in the slag varied between 0.3 and 1.5 wt pct for added phosphorus concentrations of 0.06 to 0.316 wt pct. Phosphorus partition has been found to increase with increasing CaO/SiO2 ratios. Phosphorus partition increased with increasing oxygen potential over the investigated oxygen partial pressure range, pO2=1.4x10-5to4.8x10-5(Pa). The present experimental result has been compared with literature data. The effect of slag basicity on the dephosphorization power of slags has been discussed based on this comparison. The minimum amount of slag to achieve sufficient dephosphorization using DRI has also been calculated and discussed.

Inclusions in crude steels are linked to employed feedstock material and secondary steelmaking should be tailored correspondingly for optimized praxis. Thus, when implementing hydrogen direct reduced iron (H-DRI) as feedstock for electric arc furnace (EAF) operation, knowledge of the number, size, type of inclusions, and mechanism of formation are needed to support future process development. Crude steel samples were collected from a pilot-scale EAF, employing 100% H-DRI, to study the inclusions originating from the novel feedstock. A sampler containing aluminum was used, and the effect of deoxidation is discussed. Laboratory experiments were conducted to understand the mechanisms of formation and growth of inclusions. It was clear that the inclusions formed because oxide particles merged and grew when the pores coalesced, and the iron grains sintered during the melting of H-DRI. The composition varied between the different inclusions and a total of three inclusion types were found. The size of the inclusions varied mainly between 5 and 70 µm, but the largest inclusion had a diameter of approximately 300 µm. The smallest inclusions (5-10 µm) were solid and contained mainly magnesiowüstite with a high MgO content (Type I-3). Type I-2 was larger (3-60 µm), and contained liquid, spinel, and no or small amounts of magnesiowüstite phase. Type I-1 inclusion varied greatly in size (13-300 µm) and contained all three phases with relatively high MgO content. H-DRI was heated for different durations and quenched prior to melting to study the formation mechanism. The link between feedstock material and inclusions in crude steel is discussed.