Microstructure evolution during the rapid solidification of Fe-C and Fe-C-P alloys is simulated using the phase-field model for alloys with thin interface limit parameters. Heat transfer equation is solved simultaneously to study the heat flow and the effect of latent heat generation on the microstructure. The calculations have been carried out using a double grid method and parallel computing technique. The competitive growth of growing cells is reproduced, and the cellular/dendritic transition is also observed. Since there is a negative thermal gradient in front of a leading tip, the growth can be regarded as unidirectional free dendrite growth. The microstructure changes depending on the preferred growth orientation and impurity are also studied. The secondary arms grow preferably towards inside of the melt and develop well with increase of the tilted angle. The secondary and primary arm spacing decrease by the small amount of phosphorus addition. The time change of averaged surface temperature depending on the initial undercooling shows that the surface undercooling is always observed even when the initial value is zero.KEY WORDS: Fe-C alloy; rapid solidification; micro-structure; dendrite; phase-field model.
Governing Equations
Governing Equations using Dilute Solution ApproximationThe phase-field equation with k-hold interface energy anisotropy and the solute diffusion equation for a dilute alloy are given by, 14,16) .
...(6)where h(f)ϭf 3 (15Ϫ10fϪ6f 2 ), g(f)ϭf 2 (1Ϫf) 2 , q is the angle between normal direction of interface and the x-axis, n is the magnitude of anisotropy, the superscript of e represent equilibrium state and subscripts of xx, yy and xy represent partial derivative by x or y twice, and S and L represents solid and liquid phases, respectively, R is gas constant, V m is molar volume, D(f) is solute diffusion coefficient and W, e and M are phase-field parameters which are defined later. The prime of e represents the differential by q. When the preferred growth orientation is tilted from the x-axis with q 0 , Eq. (2) is modified as follows. Note that the terms of sin(2q) and cos(2q) in Eq. (1) are substituted for the function of the partial derivative of the phase-field and they are independent of tilted angle of preferred growth direction. For a dilute ternary alloy, the governing equations and parameters are modified as in the literature. 17) In this study, temperature field is calculated simultaneously using a double grid method, in which the grid size for temperature field is set to be larger than that for phase-field and concentration field, in order to save the calculation time. The latent heat generation is estimated by summing up the change of phase-field within the corresponding temperature grid and the thermal diffusion equation is obtained as follows.where a is thermal diffusivity, L is latent heat, c p is specific heat and A represents the area ratio of phase-field grid to thermal one.The phase-field parameters of e and W are related to interface energy, s, and interface width, 2...