Date of Award

Spring 5-2018

Embargo Period


Degree Type

Dissertation (CMU Access Only)

Degree Name

Doctor of Philosophy (PhD)


Materials Science and Engineering


P. Chris Pistorius


The advancement in computational thermodynamics can help researchers to test their hypotheses regarding complex steelmaking operations in a more quantified manner. The main aim of the current work was to use develop a kinetic model that can predict changes in steel, slag and inclusions during ladle refining and use this model as a tool to develop better understanding of the steelmaking process itself. The important reactions during ladle refining are: steel-refractory reaction, slag-refractory reaction, flotation of inclusions to slag, steel-inclusion reaction, steel-slag reaction and inclusions originating from slag. The chemical reactions between two phases were considered to be mass transfer controlled. The macro-processing feature in FactSage was used to do multiple equilibrium calculations and calculate the change in steel, slag and inclusion composition. Targeted experiments and industrial trials were conducted to find model parameters. For laboratory experiments, the rate of magnesium-transfer to oxide inclusions in steel due to steel-crucible and steel-slag reaction was studied. It was concluded that the presence of spinel layer on MgO crucible at the steel-crucible reaction can help in significantly reduce the rate of Mg pick-up due to steel-crucible reaction. For industrial trials, a comparison between the rate of steel-slag reaction and inclusion flotation rate showed that the steel-slag reaction could be significantly slowed due to slag inhomogeneity. The kinetic model was also used to identify artifacts in steel and slag sampling during ladle refining. One of the main limitations of the kinetic model was the over-prediction of calcium pick-up in steel due to steel-slag reaction. Induction furnace experiments were conducted using MgO, ZrO2 and CaO crucible with different slag composition and silicon concentration to study the extent of calcium pick-up due to steel-slag and steel-crucible reactions. The steel-CaO crucible equilibrium experiment was used to estimate Ca-O interaction parameter. The equilibrated steel was reoxidized with known amount of oxygen to allow all the dissolved calcium to precipitate as oxide inclusions. Inclusion analysis of sample taken after reoxidation was used to estimate dissolved calcium in steel. The measured dissolved calcium was used to estimate Ca-O interaction parameter. A private database for liquid steel was created in FactSage and used for kinetic modeling of laboratory scale steel-slag-crucible experiments. The use of private database for kinetic model helped in avoiding excess calcium pick-up in steel due to steel-slag reaction. However, the model and database should be tested for conditions where significant calcium pick-up is experimentally observed. In the present work, the inclusion removal was assumed to be first order reaction with fixed rate constant. In practice, the inclusion removal is expected to be a more complicated process of agglomeration and flotation. Similarly, the steel-inclusion reactions were considered in equilibrium for each time step of calculation. Sometimes, the composition difference inside single inclusions was found. Some characterization tools were used that could be useful in future to study the agglomeration of inclusions and composition differences inside single inclusion. The agglomeration behavior of inclusions at the steel-argon interface inside confocal laser scanning microscope was compared to the agglomeration in bulk samples from laboratory and industrial steel samples. The size and morphology of inclusion clusters were studied using X-ray micro CT. The composition and morphology of single inclusion was studied using focused ion beam methods: Ga-FIB instrument and plasma-FIB instrument.