Systematic atomistic simulations of homo-and hetero-phase boundaries have been carried out to quantify interphase boundary energies in iron including ¤-phase and £-phase grain boundaries and ¤/£, ¤/liquid and £/liquid interfaces. Due to structural mismatch between body centered cubic (BCC) and face centered cubic (FCC) structures of the ¤ and £ phases, the minimum interface energy of the ¤/£ interface is as high as 0.41 J/m 2 , much higher than the minimum interface energies of the ¤/¤ and £/£ homo-phase interfaces, which are zero, suggesting that the high interface energy is one of the key factors that lead to the massive-like phase transformation from the ¤ phase to the £ phase observed by in situ radiography. Although the minimum ¤/£ interface energy is not significantly higher than the ¤/liquid interface energy that determines the ¤ nucleation upon solidification, it is yet high enough for the small entropy change upon the phase transformation to inhibit £ nucleation at a given critical radius until more than one orders of magnitude higher undercooling is achieved according to the classical theory of homogeneous nucleation.

Effective interface energies of various homo-and hetero-interfaces of iron were calculated with an aid of phase-field modeling, taking into account geometric constraints by competition among grains or interfaces. Calculated effective interface energies for ¤/£, ¤/¤, and £/£ interfaces are 0.56, 0.44 and 0.37 J/m 2 , respectively. Using two simple geometric models for nucleation on or off an interface in the matrix, the optimal shape of a nucleus at a given radius and undercooling, a critical radius and an energy barrier for nucleation for each possible circumstance were numerically calculated. It is found that, although the energy barrier for £-phase nucleation in homogeneous ¤-phase matrix is more than three orders of magnitude greater than that for homogeneous solidification of ¤-phase, the £ nucleation on a ¤/¤ grain boundary in the solidifying matrix suppresses the energy barrier, increasing a nucleation rate. Furthermore, it is found that the £-phase nucleation on an existing £ nucleus halves undercooling needed with smaller critical radius. This suggests that, once £ nucleation is initiated, then following £ nucleation is promoted by doubled driving force, enabling multiple £ nucleation as in chain reaction. These findings are sufficient to explain experimentally observed phenomena during the ¤-£ massive-like phase transformation even if other factors such as solute re-distribution or transformation is neglected.

No abstract

No abstract

No abstract