The viscosity of low SiO2 (10–20 mass%)-CaO-Al2O3 slag system was measured in a wide temperature range (1623–2800 K) using the rotational bob method and the aerodynamic levitation method. The influence of SiO2/Al2O3 ratio and CaO/Al2O3 ratio on the viscosity was examined. It was concluded that the SiO2/Al2O3 ratio did not affect the degree of polymerization of the aluminosilicate network in the composition range of the present study. An abnormal behaviour of the viscosity was observed at a CaO/Al2O3 ratio of 1.57 which was attributed to the formation of 12CaO·7Al2O3-like clusters. It was concluded that the overall influence on the viscosity could be expressed as the summation of the influence from the aluminosilicate network and the influence from the cluster formation of the primary precipitating solid phase. The temperature dependence of the cluster formation was coupled to the driving force of precipitation of the 12CaO·7Al2O3 phase.

The effect of in-situ precipitating particles on the grain size of Al-Ti-treated and untreated Hadfield steel cast in a pilot scale environment was studied. Hadfield steel was melted in an induction furnace and cast in Y-Block samples. Light Optical Microscopy (LOM) and the intercept method were utilized for the grain size measurements. Additionally, Thermo-Calc Software TCFE7 Steels/Fe-alloys database version 7 was used for thermodynamic equilibrium calculations of the mole fraction of particles. The planar disregistry values between the austenite and the precipitating particles were calculated. It was observed that increasing oxide content in samples with low Ti(CN) content resulted in a finer microstructure, while increasing the Ti(CN) content under similar oxide content levels led to a coarser microstructure. The potency of each type of particle to nucleate austenitic grains was determined. Spinel (MnAl₂O₄, MgAl₂O₄) particles were characterized as the most potent, followed by olivine (Mn₂SiO₄), corundum (Al₂O₃, TiO₂), and finally Ti(CN), the least potent particle.

A quantitative analysis of the amount, size and number of particles that precipitate in situ in titanium- and aluminium-treated Hadfield steel cast during pilot-scale experiments has been performed. SEM with EDS and automated particle analysis abilities was utilized for the analysis. Additionally, Thermo-Calc was used for thermodynamic calculations and Magma 5 for solidification and cooling simulations. Predicted particles sizes calculated with a model based on the Ostwald ripening mechanism were compared with the experimental data. The effect of solute availability, cooling rate and deoxidation practice on the particle population characteristics was determined. It was concluded that the amount, size and number of precipitating particles in Hadfield steel castings is possible to be controlled according to certain requirements by a careful selection of proper additives in proper amounts and also by the optimization of the casting process in aspects of deoxidation timing and control of the cooling rate of the castings.