A study of the processes during high temperature oxidation that control surface hot shortness in copper-containing low carbon steels

Copper is a problematic residual element in electric arc furnace steel production because it leads to "surface hot shortness," a cracking defect that occurs during hot rolling of steel. The cracking arises from a liquid, copper-rich phase that penetrates into and embrittles the austenite grain boundaries. The liquid forms because copper is nobler than iron and enriches at the oxide/metal interface during oxidation of iron after casting and reheating prior to hot rolling. This cracking can be reduced or eliminated by controlling the distribution of the copper-rich layer, i.e. preventing it from penetrating down the austenite grain boundaries.;This study investigated the effect of alloy chemistry on the oxidation behavior and copper-rich liquid phase evolution. Alloy compositions were selected such that effects of copper, nickel, and reactive impurities (manganese, aluminum, and silicon) can be isolated. Industrially produced low carbon steels with varying copper, nickel and silicon contents were also studied. Alloys were oxidized in air or water vapor for times up to one hour at 1150°C. Oxidizing heat treatments were conducted in a thermogravimetric setup where the weight change could be measured during oxidation. Scanning electron microscopy was used to investigate in detail the oxide/metal interfaces.;The modeling work focused on describing the enrichment and subsequent growth of the copper-rich layer. A fixed grid finite difference model was developed that predicts the evolution of the enriched region from given oxidation kinetics. The model predictions were validated under a variety of conditions using an iron - 0.3 wt% copper alloy. Deviations from the model predictions in these alloys suggest a critical amount of separated copper is necessary for substantial grain boundary penetration to occur and the required amount decreases when the gas contains water vapor.;The parabolic oxidation rate for the iron-copper alloy did not differ from that of pure iron, but the parabolic rate for the nickel-containing alloys decreased by a factor of two. The microstructure of the iron-copper alloy consisted of a thin, copper-rich layer at the oxide/metal interface. Both nickel-containing alloys had perturbed oxide/metal interfaces consisting of alternating solid/liquid regions. The perturbed interfaces arise from unequal copper and nickel diffusivities in the ternary alloy. These diffusion effects are discussed in detail. The oxidation rate decrease is justified by the interface microstructure assuming that iron can only be rapidly supplied to the oxide through the liquid regions.;Additions of manganese or aluminum to an iron-copper-nickel alloy did not lead to significant changes in behavior. Oxidation kinetics, amount of separated material, and interface roughness were unchanged. There was slightly more material occluded in the samples containing manganese and aluminum due to increased internal oxidation. These internal oxides do not affect the oxidation behavior because manganese can dissolve in wustite and the aluminum internal oxides are extremely small and heterogeneously dispersed near the oxide/metal interface.;Additions of silicon, however, to an iron-copper-nickel alloy led to a significant decrease in oxidation rate, amount separated, and amount occluded. The differences in behavior are attributed to the formation of a fayalite layer at the oxide/metal interface. This layer blocks iron transport in the wustite layer, decreasing the oxidation rate and therefore the enrichment rate. Formation of the fayalite layer was found not to depend on the amount of nickel in the samples.;The results above were then used to explain the oxidation behavior of low carbon steels containing copper, nickel, silicon, manganese, and aluminum. Steels containing high amounts of silicon had lower oxidation rates and higher amounts occluded. The amount of occluded material is much higher in the steels than in the iron-copper-nickel-silicon alloys. This is attributed to competition among the easily oxidizable impurities resulting in smaller internal oxide particles that assist occlusion but prevent formation of the continuous fayalite layer.