Slag Solidification Modeling Using the Scheil–Gulliver Assumptions Dirk Durinck, w Peter Tom Jones, Bart Blanpain, and Patrick Wollants Department of Metallurgy and Material Science, Katholieke Universiteit Leuven, 3001 Heverlee, Belgium Gilles Mertens and Jan Elsen Department of Geology, Katholieke Universiteit Leuven, 3001 Heverlee, Belgium In this paper, the Scheil–Gulliver approach on solidiﬁcation modeling is applied to slag solidiﬁcation. The Scheil–Gulliver model assumes no diffusion in the solid phases, inﬁnitely rapid diffusion in the liquid phase, and local equilibrium at the solid/ liquid interface. For two distinct CaO–MgO–SiO 2 slags, solidi- ﬁcation is simulated with the Scheil–Gulliver model and with the commonly used thermodynamic equilibrium model. The simula- tions show that in the Scheil–Gulliver model, as opposed to the equilibrium model, compositional gradients in the solid phases can exist and peritectic reactions do not occur. Solidiﬁcation experiments are performed under laboratory conditions to val- idate the simulation results. The experimental results show a better correspondence with the Scheil–Gulliver model than with the equilibrium model. However, quantitative differences in min- eralogy persist. I. Introduction O WING to the ever-growing economic and environmental im- plications associated with slag dumping, slag valorization has become increasingly important for the pyrometallurgical in- dustry over the last decades. Valorization of blast furnace slag and steel slag has already developed to such an extent that in the Western world, more than 90% of all slag is being reused, most- ly as a secondary resource for road construction or cement pro- duction. 1 The main challenge is now to ﬁnd higher value applications for steel slag. In the case of stainless-steel slag, fer- rous alloy slag, and non-ferrous metallurgical slag, valorization targets the same ﬁelds of application but is basically still in its infancy. 2 Valorization requires the slag product to possess certain mechanical, physical, and chemical properties. Often, these properties are directly linked to the slag mineralogy. Examples are:  Chromium leaching from stainless-steel slag or ferro- chrome slag can be limited by the presence of spinel minerals, which bind the chromium in a stable structure. 3  Slag disintegration during cooling is caused by the b- to g-phase transformation of dicalciumsilicate grains (2CaO  SiO 2 or C 2 S). 4  Swelling of the slag product is associated with the hydra- tion of free MgO or free CaO in the slag’s microstructure. 4 Although this list is far from exhaustive, it illustrates the im- portance of the slag mineralogy for valorization purposes. In order to improve the quality of the ﬁnal slag product, the min- eralogy should be optimized. Consequently, solidiﬁcation of the slag is a critical step in the valorization chain. Efforts have been made to investigate experimentally the effects of composition and cooling path on the mineralogy of the cooled slag product. Typically, a laboratory furnace is used to impose a cooling path on a synthetic slag sample, which is subsequently characterized with respect to mineralogy and microstructure. 5–7 Multiple solidiﬁcation tests are required to investigate certain effects of slag composition and cooling path. Recently, a new technique was developed that allows to study the inﬂuence of the cooling rate on the slag microstructure in a single experiment. 8 Studies have also been performed to observe solidiﬁcation phenomena in situ using confocal scanning laser microscopy (CSLM) 9,10 or the double-hot thermocouple tech- nique (DHTT). 11 Investigation of the solidiﬁcation behavior of metallurgical slags by experimental techniques, however, has been proven to be time-consuming and expensive. Therefore, few experimental data on slag solidiﬁcation are available in the scientific literature. Apart from experimental data, modeling of solidiﬁcation also provides a tool to investigate the effects of composition on the mineralogy of the cooled slag product. The common approach to solidiﬁcation modeling in oxide systems is based on the assumption of global thermodynamic equilibrium. Park 7 and Verscheure et al. 8 claim to have obtained a good qualitative correspondence between their experimental and modeling results when they simulated thermodynamic equilibrium at different temperature intervals during cooling. For most slags, however, the quantitative amounts of phases predicted by the equilibrium model differ substantially from the ex- periment. An alternative approach to solidiﬁcation modeling is based on the Scheil–Gulliver assumptions. 12 This model assumes in- ﬁnitely rapid diffusion in the liquid phase, no diffusion in the solid phases, and local equilibrium at the solid/liquid interface. In this way, the slow diffusion rates of components in solid crystalline phases can be taken into account. This modeling ap- proach is widely applied for studying the solidiﬁcation of metals. Some recent work includes modeling the solidiﬁcation of Al al- loys, 13,14 Mg alloys, 15 and Pb-free solders. 16 For the solidiﬁca- tion of steel and cast iron, the Scheil–Gulliver assumptions have already been further extended. Allowing back-diffusion of in- terstitials in the solid phases results in better modeling results for these alloys. 17 To the best of the authors’ knowledge, however, this modeling approach has not yet been used to model the solidiﬁcation of slags or any other oxidic systems. In the present paper, the Scheil–Gulliver model is applied to slag solidiﬁcation. Two distinct CaO–MgO–SiO 2 slags are used to validate the modeling results. The synthetic slag samples are equilibrated at 16401C, cooled at a rate of 11C/min, and ana- lyzed with electron probe microanalysis (EPMA) and X-ray dif- fraction (XRD). Quantitative results are obtained by reﬁning the diffraction patterns adopting Rietveld’s method. 18,19 These experimental results are compared with the results of both the R. Snyder—contributing editor This work has been performed with the ﬁnancial and technical support of U&A Belgium and the IWT (project no.050715). The second author is currently working as an Aspirant of the Flemish Fund for Scientific Research (FWO), which is kindly acknowledged for its ﬁnancial support. w Author to whom correspondence should be addressed. e-mail: dirk.durinck@ mtm.kuleuven.be Manuscript No. 22336. Received October 9, 2006; approved January 10, 2007. J ournal J. Am. Ceram. Soc., 90 [4] 1177–1185 (2007) DOI: 10.1111/j.1551-2916.2007.01597.x r 2007 Katholieke Universiteit Leuven 1177