**2. Materials and Methods**

Prior to the viscosity measurement, MgO solubility in the CaO–SiO2–FeO–Al2O3 system at 1823 K was determined by using a thermochemical equilibrium technique [11]. The slag sample was prepared using reagent-grade CaO, SiO2, Al2O3, FeO, and MgO. CaO was obtained by the calcination of CaCO3 at 1273 K for 6 h. The powder was mixed in an agate mortar to obtain a homogeneous mixture. Afterward, approximately 5 g of the powder mixture was placed in a MgO crucible (99% purity) and heated in an electric resistance furnace equipped with MoSi2 heating elements under an Ar atmosphere. The equilibration time was determined as 8 h in a previous study [11]. After 8 h, the samples were removed from the furnace and quenched by blowing Ar gas. The slag was separated from the MgO crucible and ground using a pulverizing ball mill to less than 100 μm for chemical analysis. The slag composition was analyzed using X-ray fluorescence (XRF, S4 Explorer; Bruker AXS, Madison, WI, USA). Table 1 shows the pre- and post-experiment slag compositions. Although the pre-experiment compositions of FeO and Al2O3 were 10, 20, 30, and 40 wt%, the post-experiment contents varied because of the different MgO solubilities. For convenience, the pre-experiment concentrations of FeO and Al2O3 were used to identify the samples in the present study.


**Table 1.** Experimental results of MgO solubility for CaO–SiO2–FeO–Al2O3 slags at 1823 K.

Referring to the MgO-saturated compositions in the CaO–SiO2–FeO–Al2O3 slag system at 1823 K, as shown in Table 1, the slag mixture was prepared using reagent-grade CaO, SiO2, FeO, Al2O3, and MgO. Approximately 120 g of the homogeneous powder mixture ground in an agate mortar was placed in a Pt–10% Rh crucible (outer diameter: 41 mm, inner diameter: 40 mm, and height: 65 mm). The crucible was placed in an electric resistance furnace at 1873 K under an Ar atmosphere. After maintaining the conditions for 1 h to achieve thermal equilibrium, the viscosity was measured by using a rotating cylinder method. The viscosity and torque data were recorded each second using a digital viscometer (DV2TLV; Ametek Brookfield, Middleboro, MA, USA) calibrated with silicone oil at room temperature. Figure 1 shows the schematic of the viscosity measurement apparatus. To evaluate the temperature dependency, the viscosity was measured by decreasing the temperature by 25 K at 5 K/min and by maintaining each temperature for 30 min during viscosity measurement.

After the viscosity measurement, the temperature was increased to 1873 K and the crucible was removed from the furnace. The molten slag was quenched on a water-cooled Cu plate. No characteristic X-ray diffraction (XRD) peaks were observed from the quenched sample, indicating that it was in an amorphous state. The obtained sample was crushed and ground to a particle size of less than 100 μm for structural analysis. The intermediate-range order of the slag structure was analyzed using FT-IR spectroscopy (Spectra 100; Perkin-Elmer, Shelton, CT, USA) and Raman spectroscopy (LabRaman HR, Horiba Jobin-Yvon, France). More details of the structural analysis procedure utilizing FT-IR spectroscopy and Raman spectroscopy have been explained elsewhere [22–24].

**Figure 1.** Schematic of the viscosity measurement apparatus.

#### **3. Results and Discussion**
