Crystallization behaviors of the newly developed lime-alumina-based mold fluxes for highaluminum transformation induced plasticity (TRIP) steels casting were experimentally studied, and compared with those of lime-silica-based mold fluxes. The effects of mold flux crystallization characteristics on heat transfer and lubrication performance in casting high-Al TRIP steels were also evaluated. The results show that the crystallization temperatures of lime-aluminabased mold fluxes are much lower than those of lime-silica-based mold fluxes. Increasing B 2 O 3 addition suppresses the crystallization of lime-alumina-based mold fluxes, while Na 2 O exhibits an opposite effect. In continuous cooling of lime-alumina-based mold fluxes with high B 2 O 3 contents and a CaO/Al 2 O 3 ratio of 3.3, faceted cuspidine precipitates first, followed by needlelike CaOAEB 2 O 3 or 9CaOAE3B 2 O 3 AECaF 2. In lime-alumina-based mold flux with low B 2 O 3 content (5.4 mass pct) and a CaO/Al 2 O 3 ratio of 1.2, the formation of fine CaF 2 takes place first, followed by blocky interconnected CaOAE2Al 2 O 3 as the dominant crystalline phase, and rod-like 2CaOAEB 2 O 3 precipitates at lower temperature during continuous cooling of the mold flux. In B 2 O 3-free mold flux, blocky interconnected 3CaOAEAl 2 O 3 precipitates after CaF 2 and 3CaOAE2-SiO 2 formation, and takes up almost the whole crystalline fraction. The casting trials show that the mold heat transfer rate significantly decreases near the meniscus during the continuous casting using lime-alumina-mold fluxes with higher crystallinity, which brings a great reduction of surface depressions on cast slabs. However, excessive crystallinity of mold flux causes poor lubrication between mold and solidifying steel shell, which induces various defects such as drag marks on cast slab. Among the studied mold fluxes, lime-alumina-based mold fluxes with higher B 2 O 3 contents and a CaO/Al 2 O 3 ratio of 3.3 show comparatively improved performance.

Both the Matusita equation and the modified Matusita equation for estimating the activation energy associated with non-isothermal crystallization were critically evaluated. The derivation for melts crystallization on cooling indicates that, unlike for the crystallization that occurs on heating, the term 1 À exp (ÀDG/RT) in the basic rate equation of crystal growth and the term R T s 0 exp ÀE=RT ð Þ dT depending on the initial temperature of the cooling process cannot be neglected. It is demonstrated that both the Matusita equation and its modified expression are only valid to estimate the activation energy associated with the crystallization that occurs on heating, but are inapplicable for the melt crystallization that occurs on cooling. It is suggested that the isoconversional methods of Friedman and Vyazovkin should be alternative to determine effective activation energy for melt crystallization that occurs on cooling.Various models have been developed to estimate the activation energy associated with the non-isothermal crystallization from thermal analysis data, including Kissinger equation, [1] Ozawa equation, [2] modified Ozawa-Chen equation, [3] and Matusita equations. [4,5] Among these models, Kissinger equation [1] and Matusita equations [4,5] are the most widely used approaches to determine the activation energy for the crystallization that occurs on heating. [6][7][8][9][10] Moreover, they are applied frequently for the non-isothermal crystallization of polymer melts [11][12][13][14][15] and metallurgical slags [16,17] that occurs on cooling.Matusita and Sakka [4] emphasized that the physical meaning of the crystallization activation energy determined by the Kissinger equation [1] is obscure because the crystallization of glass is advanced by both nucleation and crystal growth, rather than an n-th order reaction. Therefore, they proposed the following generalized expression to determine the activation energy for crystal growth: [4]

A kinetic study for isothermal melt crystallization of CaO-SiO 2-CaF 2-based mold fluxes with different basicity of 0.94 and 1.34 has been carried out systematically by DSC measurements. The kinetic parameters were determined by Johnson-Mehl-Avrami equation. The average Avrami exponent of cuspidine (3CaOAE2SiO 2 AECaF 2) crystallization for mold flux of lower basicity (0.94) is calculated to be 3.1, implying that the crystallization mode is instantaneous nucleation followed by 3-dimensional growth. For the mold flux of higher basicity (1.34), the average Avrami exponent of cuspidine equals to 3.4, strongly suggesting that the growth is still 3 dimensional but the nucleation should be continuous. It was found that the effective crystallization rate constant for both mold fluxes increases as the crystallization temperature decreases, showing that the crystallization rate could be governed by nucleation rate. The negative effective activation energy indicates an anti-Arrhenius behavior for crystallization of the mold fluxes studied. Therefore, it is concluded that the melt crystallization for the commercial mold fluxes will be determined by thermodynamics of nucleation which is relevant to degree of undercooling. The morphology of cuspidine crystals observed by SEM agreeds well with the isothermal crystallization kinetics results.

No abstract

No abstract