Viscosity prediction on iron-bearing slags during pyrometallurgical recycling: structure-based modeling of CaO−‘FeO’−MgO−Al2O3−SiO2 system and its subsystems

A structure-based modeling of the CaO−‘FeO’−MgO−Al2O3−SiO2 system and its subsystems was investigated based on iron extraction from nickel slag by aluminum dross. Parameters optimization in the present model indicated that the coefficient of free O2− in FeO, $$a_{{{\text{O}}_{{{\text{FeO}}}}^{{2 - }} }}$$ , on the lengths of network linkage had the largest value and $${\text{O}}_{{{\text{FeO}}}}^{{2 - }}$$ (free O2− in FeO) had the largest mobility. The coefficients of bridging oxygen (aSi−O−Al and aAl−O−Al) were lower than those of non-bridging oxygen and free oxygen (O2−). Viscosity prediction for the CaO−‘FeO’−(8 wt.%) MgO−Al2O3−SiO2 system was conducted at a fixed slag basicity, which indicated that the predicted viscosity changed monotonously with the FeO content. However, the non-monotonous evolution with Al2O3 content reflected the amphoteric behavior of Al2O3. In addition, the performances of the present model in predicting viscosity from binary (‘FeO’−SiO2) to quinary (CaO−‘FeO’−MgO−Al2O3−SiO2) system were analyzed and a comparison with the established models was made.