Journal Information
Download PDF
More article options
Original Article
DOI: 10.1016/j.jmrt.2018.05.022
Open Access
Available online 28 July 2018
Experimental and thermodynamic analysis of MgO saturation in the CaO–SiO2–Al2O3–MgO slag system melted in a laboratory resistive furnace
Vinicius Cardoso da Rocha
Corresponding author

Corresponding author.
, Pedro Cunha Alves, Julio Aníbal Morales Pereira, Letícia Pegoraro Leal, Wagner Viana Bielefeldt, Antônio Cezar Faria Vilela
Department of Metallurgy (DEMET), Federal University of Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
This item has received

Under a Creative Commons license
Article information
Full Text
Download PDF
Figures (7)
Show moreShow less
Tables (4)
Table 1. Chemical composition of the slag samples studied.
Table 2. Technical data of the operation in the resistive furnace adopted in this study.
Table 3. Input data of the slag compositions in the Equilib module of FactSage™ 7.1.
Table 4. Composition ranges in MgO (wt.%) and phases formed in the slag for the temperatures of 1500°C (1773K), 1550°C (1823K) and 1600°C (1873K) calculated by FactSage™ 7.1.
Show moreShow less

Slags from the CaO–SiO2–Al2O3–MgO system (CSAM) are commonly employed for the secondary refining treatment of special steels. Those slags must be designed to work in the refining reactions, presenting properties that maximize their capability to capture impurities from the liquid steel. With this in context it becomes relevant to study the phases of those slags in the operating temperatures. In the present study, four slags from the CSAM system, with increments of MgO (5.98–23.43wt.%) and binary basicity (1.96–2.48), were melted in a high temperature resistive electric furnace. The objective was to analyze the phases present in the slags in secondary refining temperatures (1500°C (1773K) to 1600°C (1873K)), depending on their chemical composition. Fluorescence (XRF), X-ray diffraction (XRD) and scanning electron microscopy (SEM) with dispersive energy spectrometer (EDS) were used for the slag analyses. The experimental results were compared with thermodynamic calculations obtained by using the FactSage™ 7.1. As the MgO contents increased in the slag (>14.96wt.%), the precipitation of MgO solids were predicted by thermodynamic calculations, indicating the saturation of the slag in this compound. In the SEM/EDS analysis, this saturation was confirmed by the presence of dark and dendritic phases rich in MgO in the slag microstructure. However, for MgO contents up to 7.93wt.%, the microstructures obtained were mainly composed of C2S (dicalcium silicate) phases dispersed in the liquid matrix, identifying the unsaturated condition for the MgO compound.

Slag phases
Resistive furnace
MgO saturation
Computational thermodynamics
Full Text

Secondary refining slags have important functions during steel production, especially those for the automotive industry, which demands exceptional quality standards. For many steel research centers, the slags are seen as a study priority, in the continuous search for a better understanding of the phenomena involved in the refining process of special steels. In this sense, worldwide efforts are being made to seek clarification on the phenomena involving ladle slags. Recently published studies [1–4] confirm the researchers’ commitment in the explanation of some effects of slags during the treatment of liquid steel.

In the steel industry, ladle slags are designed to maximize their refining capacity, including the removal of impurities such as sulfur and non-metallic inclusions from the liquid steel [5]. In this way, an understanding of the properties of the slag becomes of great importance, since from this it is possible to establish study areas focused on the clarification of the main effects promoted by the variation of these properties in its steel production and refining capacity. These properties can be interpreted in the thermodynamic and thermophysical investigation of the slags [6].

The CaO–SiO2–Al2O3–MgO (CSAM) quaternary slag system is known for its common application in secondary refining stations for special steels. One of the main reasons for the application of this system is the high compatibility with the refractories used in the ladle, mainly composed of magnesia (MgO–C) [7,8]. Considering the strong dissemination of the CSAM system, many studies have been developed involving this system for the analyses about MgO saturation [8,9], viscosity [10–13], non-metallic inclusions removal capacity [4,14,15] and sulfur removal [16–18]. At laboratory scale, the phases present in the slags have also instigated several studies [7,19–21]. The knowledge of the slag phases formed for different chemical compositions ranges is important for the understanding with respect to slag/refractory and slag/steel reactions or even in the area of slag reuse, through a deep evaluation of the chemical and mechanical properties regarding the formed phases.

For the study of slags, it is essential to establish an accurate experiment that is capable in producing reliable data, since those data are pre-requisite for the development and improvement of extrapolation models and, even, to the direct contribution to steelmakers and researchers in the area. This laboratory study has the objective to analyze the phases observed in the slag of the CSAM system considering the secondary refining temperatures ranging from 1500°C (1773K) to 1600°C (1873K), with MgO increments from 5.98 to 23.43wt.% and binary basicity (B2=CaO/SiO2), from 1.96 to 2.48. The analysis is developed from the experimental and theoretical (thermodynamic) point of view, providing an in-depth discussion regarding the formation of the different phases in the slags.

2Materials and methods

The analyzed slags comprehend exclusively the CSAM quaternary system, being considered four chemical compositions, as detailed in Table 1. The slag samples were prepared based on the mixing of CaO, SiO2, Al2O3 and MgO components (99% pure) with laboratory mixer equipment for a period approximately of 2h in order to ensure a good homogenization of the samples. Each slag sample corresponds to approximately 20g.

Table 1.

Chemical composition of the slag samples studied.

Slag sample  Chemical composition (wt.%)
  CaO  SiO2  Al2O3  MgO  B2 
E1  47.43  24.22  22.37  5.98  1.96 
E2  47.01  21.51  23.54  7.93  2.19 
E3  43.14  18.49  23.41  14.96  2.33 
E4  39.80  16.02  20.75  23.43  2.48 

The chemical composition of the slags (Table 1) was measured by X-ray fluorescence (XRF) technique using Philips equipment model PW2600. The compositions were designed to present an increase in the MgO content in the slag and also in the binary basicity (B2). After mixing, each slag sample was added in graphite crucibles (99.9% C) with internal dimensions of 90mm in height and 32mm in diameter. The outside diameter of the crucible has 100mm in height and, in diameter, 45mm. Afterward, all the crucibles were placed in a support, also made of graphite, and then introduced into the resistive furnace, as illustrated by Fig. 1. This support (Fig. 1(a)) acts as basket, where it is possible to fit up to four crucibles. In addition, its central rod assists in the removing process of the samples from the furnace tube at high temperature.

Fig. 1.

(a) Side view of the graphite support for the packaging of the crucibles containing slag for (b) application within the resistive furnace (top view).


The slags melting were carried out in a resistive electric furnace, illustrated in Fig. 2. It is a high temperature furnace model HT-2100-Vac-Graphit-Special from Linn High Therm manufacturer. It is interesting to note that this furnace is one of the only models available in Brazil in the current configuration. The main operating details adopted in this work are presented in Table 2. Previous works, involving steel cleanliness [22] and evaluation of refractory materials for ladles [23] made use of this same resistive furnace.

Fig. 2.

Resistive furnace HT-2100-Vac-Graphit-Special and its components.

Table 2.

Technical data of the operation in the resistive furnace adopted in this study.

Technical information
Graphite tube dimensions  Diameter: 150mm
Total height: 470mm
Working height (with insulation system coupled): 350mm 
Graphite tube volume  4.4
Maximum operation temperature  1650°C 
Heating rate  5°C/min (for T> 25°C) 
Atmosphere  Inert 
Protection gas  Ar – 99.999% with O2<1ppm and N2<3ppm; Flow: 82l/h (constant) 
Temperature measurement  Mantel-thermocouple WRe 6/25 
Thermal insulation  Carbon fiber mantles 
Cooling system  Water (inlet: ∼13°C/outlet: ∼37°C) 
Voltage/frequency  3×380V/50–60Hz 

The thermocouple identified by Fig. 2 measures the temperature near the electrical resistance surrounding the graphite tube. It is reasonable to assume a difference between the temperature of the slag samples (inside the graphite tube) and the temperature measured by the furnace thermocouple. In fact, in previous measurements [24], a difference of up to −80°C was found inside the graphite tube with respect to the furnace resistance region. Thus, the thermocouple indicated in Fig. 2 was used to measure the temperature of the slag systems, and this difference was taken into account in the thermodynamic observations of this work.

The four slag compositions were melted in the resistive furnace to the temperature of 1650°C (1923K), based on the following operations: heating rate of 5°C/min (from 25°C) for 5h and 30min up to 1650°C (1923K), remaining in this temperature for 1h and 30min. After this residence period, the furnace was shut down and opened for the removal of the graphite support, containing the four crucibles with the slag samples. In this meantime, the slag samples are considered to have their temperatures reduced for values in the range of 1500°C (1773K) to 1600°C (1873K). This decrease in temperature is justified by two reasons: the lower temperature observed inside the graphite tube and the time interval until the effective cooling process with liquid nitrogen. The temperature difference inside the crucibles was commented previously. The time interval includes the furnace shutdown, opening process and removing the crucibles from within the graphite tube. From the furnace shutdown, its opening process for withdraw samples takes around one minute. However, it should be add the time to effectively remove all the crucibles out of the graphite tube, resulting in another one minute. Then, in these two minutes, the contact of the crucibles at 1650°C (given by the furnace thermocouple) with the atmospheric air at room temperature (23°C) should promote a loss in temperature of up to 150°C. Immediately after the crucible support removal, the slags were cooled with liquid nitrogen (−196.5°C (76.6K)). Subsequently, the slags already cooled, were removed from the crucibles to initiate the metallographic procedures necessary for SEM/EDS analysis, in order to identify the slag phases formed and their morphologies. Those procedures consisted in the inlaying process of the slag samples in resin material with cold cure. Afterwards, these samples were prepared using grinding sequences of 200, 300, 400, 1000 and 1200 in the presence of water as lubricant. The SEM analysis requires the sample material to be electric conductive. Due to the low conductivity of the slag samples, it became necessary to use a methodology to obtain a conductive surface. Thus, a thin layer of carbon was deposited on the surface of the samples. This coating process is developed in the vacuum with the precipitation of a micrometric conductive film on the polished surface of the sample, improving the emission level of electrons and thus guaranteeing good resolution images. For this process, samples were placed in a pressure chamber at about 0.1–0.05mbar where the target is bombarded with inert gas atoms, such as argon. The SEM/EDS equipment used was from Shimadzu manufacturer, model SSX-550 Superscan, with application of the following parameters: acceleration voltage of 15kV, current of 1.0nA and a working distance of 17mm. For this analysis, the backscattering electrons (BSE) counting was applied. The image obtained by this detector allows filtering information of composition and topography, being generated by the emission of backscattered electrons that demonstrate the compositional differences in the ionized region. The volume of the ionized region depends on the average atomic number (Z) in the interaction zone of the material with the electron beam. The compositional difference is then demonstrated by distinct tones of gray, where the light tones correspond to the portions made up of elements with average Z relatively higher than those with darker tones. Together with the SEM, the EDS system provides qualitative and semi-quantitative data of the composition in the analyzed phases, from the emission of characteristic X-rays. The limit of detection is in the order of 1% and may vary according to the specifications used during the analysis.

In addition to the melting experiments performed in the resistive furnace, a thermodynamic study of the slags was also developed, aiming to characterize the slags regarding their formed phases, solid and liquid fractions and the saturation degree in MgO. For this purpose, the FactSage™, a computational thermodynamics program was used, in its latest version, 7.1 [25]. In the present study, the modules used in the program were Equilib and Phase Diagram with the FToxid and FactPS databases selection. The Equilib module was applied to obtain the solid and liquid fractions and respective compositions of the phases formed in the slag, considering the thermodynamic equilibrium. The Phase Diagram module was used for the pseudoternary diagrams calculation with isothermal cuts. The temperatures considered for all thermodynamic calculations were 1500°C (1773K), 1550°C (1823K) and 1600°C (1873K), in order to ensure that all the results are studied within the range immediately before the cooling process with liquid nitrogen. In addition, an X-ray diffraction (XRD) analysis of the slag samples was performed to confront with the thermodynamic results. For this, the samples were ground and sieved in a 325 mesh number, and the diffraction technique was performed in the Philips X’Pert equipment, applying a scan of 5 to 75°. The diffraction results were analyzed using the X’Pert HighScore program.

The insertion of the CSAM slag systems data in the Equilib module of FactSage™ 7.1 follows the format shown in Table 3, considering the compositions obtained through the XRF (Table 1).

Table 3.

Input data of the slag compositions in the Equilib module of FactSage™ 7.1.

Slag sample  Input data 
E1  MgO=<A>
E2  MgO=<A>
E3  MgO=<A>
E4  MgO=<A>

<A>, “Alpha” which varies from 0 to 25wt.%.

The input data provided by Table 3 were inserted into the Equilib module of FactSage™ 7.1, where the term <A> denotes the MgO content, ranging from 0 to 25wt.%, thus, it is possible to identify the saturation point of the slag in this component (MgO). The Al2O3 contents inserted into the FactSage are the same from Table 1. The CaO and SiO2 contents are always inserted as a function of the term <A>. It is important to note that the sum of all terms (components) of each slag should be equal to 100, being in agreement with Table 1. From these input data (Table 3) it is possible to obtain the phases formed for the different compositions presented by slags E1, E2, E3 and E4, in the equilibrium condition.

3Results and discussion3.1Slags after melting and cooling

The visual appearance (macroscopic) of molten slags after being cooled in liquid nitrogen can be seen in Fig. 3.

Fig. 3.

Visual appearance of slag samples (E1, E2, E3 and E4) after rapid cooling with liquid nitrogen.


Analyzing Fig. 3 it is possible to identify different physical characteristics in the cooled slags. The crucible containing the sample E1, after cooling, broke and it was identified the presence of a completely pulverized slag, as illustrated. In this context, for the subsequent SEM/EDS analysis the sample E1 was not used. The formation of this powder in sample E1 is associated with the chemical composition of the slag. According to Pontikes et al. [26], the disintegration or dust formation is caused by the presence of dicalcium silicate (C2S) in the slag, which undergoes a series of polymorphic transformations accompanied by volumetric expansions, causing internal tensions in the structure and, finally, promoting the disintegration of the slag. On the other hand, samples E2, E3 and E4 presented more similar appearance with each other, with the formation of a gray, porous and fragile solid.

In order to identify the phases formed in the slag samples, an X-ray diffraction (XRD) analysis was performed. The results of the analysis, for each sample, are shown in Fig. 4.

Fig. 4.

Identified phases by the XRD technique for slags (a) E1, (b) E2, (c) E3 and (d) E4.


The qualitative results of X-ray diffraction show the formation of different phases for the four slag samples considered in this study. For the sample E1 (Fig. 4(a)) there are only phases 1, 2 and 3 which are, respectively, Ca2SiO4, Ca2Al(Al,Si)2O7 and Ca2MgSi3O12. The presence of the dicalcium silicate, showed by the diffraction, reinforces the evidence of the formation of powder in sample E1. In Fig. 4(b), referring to sample E2, in addition to the phases reported in sample E1, a fourth phase (4), Ca2Al(Al,Si,O7), appears. In samples E3 and E4 two new phases (5 and 6) were detected: Ca3Mg(SiO4)2 and MgO. The presence of phases 5 and 6 are strongly related to the precipitation of MgO in the slag E3 and E4, reflected by the increase in the MgO content in the chemical composition of these slags (from 14.96wt.%), as reported in Table 1.

3.2SEM/EDS analysis

The slag samples E2, E3 and E4 were analyzed by SEM/EDS and the results are shown in Fig. 5(a), (b) and (c), respectively. As commented previously, the sample E1, after cooling, presented the formation of a fine powder and, for that reason, was not submitted to the microstructural analysis.

Fig. 5.

SEM/EDS results with oxide phase composition (wt.%) for slag samples (a) E2, (b) E3 and (c) E4.


In the sample E2 (Fig. 5(a)) the presence of two phases, consisting of light (points 1 and 2) and gray colors (points 3 and 4) (matrix) is observed. The lighter phase (points 1 and 2) represents the solids rich in CaO and SiO2 formed. The gray phase (points 3 and 4) is the portion of liquid that was cooled rapidly, being richer in CaO and Al2O3, according to the EDS results. In samples E3 and E4 (Fig. 5(b) and (c)), the MgO content increases and, thus, darker morphologies approaching to a dendritic format, rich in the MgO compound (points 5 and 6), are identified. Beskow and Sichen [27] also observed the presence of MgO dendrites in some ladle slags. According to the authors, the presence of these dendrites has a strong relation with the transformation of the liquid phase to solid, through a rapid cooling process. In addition, the growth of MgO dendrites is associated with the MgO precipitation in the slag, indicating a supersaturation condition in MgO.

Fig. 5(b) shows, in addition to the MgO rich dark phases, a region also rich in this compound, represented by the EDS analysis at point 2 (88.9wt.% MgO). The solid at point 2 shows a clearer region very similar to those rich in CaO and SiO2. However, a characteristic difference noticed is the shape of the solid, being more irregular and faceted. Differently, point 1, which is rich in CaO and SiO2, presents a more rounded morphology, with softer and less angular edges. Thus, the MgO rich portions are presented in two specific forms in the analyzed samples: dendritic forms in dark colors and in lighter phases with angular or faceted shapes.

From the point of view of the refining processes it is important to guarantee the saturation of the slag in MgO, in order to minimize reactions with the refractory wall [8]. However, it is also interesting not to exceed the saturation point with much higher values in MgO, since the greater presence of MgO precipitates promotes an increase in the slag viscosity, reducing its ability to interact with the liquid steel and, thus, affecting the refining reactions [1,4,14,15].

3.3Thermodynamic calculations

In order to identify the saturation condition of the slags in MgO, calculations were developed with the FactSage™ 7.1 application, conducted in the Equilib and Phase Diagram modules. The input data in the program were applied as shown in Table 3. For analysis purposes, the results of the thermodynamic calculations for the solid fractions of the slag were plotted as a function of the MgO content in the slag (Fig. 6).

Fig. 6.

Solid fractions and phases formed in the slags (a) E1, (b) E2, (c) E3 and (d) E4 as a function of the MgO content. Predicted by FactSage™ 7.1.


In Fig. 6 it is possible to identify the phases and respective amounts formed in the different slag samples, in the equilibrium condition, at 1500°C (1773K), 1550°C (1823K) and 1600°C (1873K) temperatures. It is noted that at each inflection in the curves illustrated in Fig. 6 a new phase is formed and another is dissolved. It is observed that for all slags, immediately before the formation of a totally liquid slag, the solid fraction is always higher for the lower temperature of 1500°C (1773K), as expected. However, it is not observed a liquid slag when the E4 sample is considered at 1500°C (1773K), where MgO precipitation is observed occurring together with the dissolution of the C2S solid phase, from 6.32wt.% in MgO.

After the liquid window region (when it is formed), the solid fractions are very similar at temperatures of 1550°C (1823K) and 1600°C (1873K). However, when the temperature decreases to 1500°C (1773K) the solid fraction is slightly higher than in the higher temperatures, showing the curves a little further apart after the fully liquid phase, especially in samples E1 and E2 (Fig. 6(a) and (b)). For all four slags, immediately after the liquid window ends, the saturation process in MgO begins, promoting a gradual increase in the slags solid fraction, due to the precipitation of this compound. Immediately at the point of saturation, the slag is completely liquid in practically all four samples (this is not valid only for the sample E4 at 1500°C (1773K), where it is not seen a fully liquid condition). The presence of liquid slag, or liquid window, is more noticeable as the temperature increases. In this context, it is possible to observe that at 1600°C (1873K) the liquid window extension is always larger when compared to the lower temperatures of 1500°C (1773K) and 1550°C (1823K).

With the increase of the MgO content in the slags up to 25wt.%, different phases are formed, as exposed before. In general, from 1550°C (1823K) in the slags shown in Fig. 6, the sequence of phases formed is C2S, liquid slag and saturated MgO. However, there are some exceptions, which can be observed mainly in the lower temperature of 1500°C (1773K), where there is formation of other phases. For this lower temperature, in the slag samples E1, E2 and E3 the presence of melilite phase together with solid C2S is noted with lower MgO contents range. In addition, in the 25wt.% MgO range, the formation of spinel phases in slag E1, E2 and E3 is also observed, being more noticeable at 1500°C (1773K). As the temperature increases, the spinel phase disappears, requiring a higher amount of MgO to form this compound. For example, for slag E1 at 1500°C (1773K), there is spinel from 8.33wt.% in MgO, however, as the temperature increases to 1550°C (1823K), the spinel begins to form only from 18.43wt.% in MgO. The melilite phase appears in its largest quantity in the slag E1, from 0 to 3.27wt.% in MgO, in the temperature of 1500°C (1773K). For slag E2, there is no melilite presence at 1550°C (1823K), even with lower MgO levels. However, by lowering the temperature in 50°C, the formation of melilite is noticed up to almost 2wt.% in MgO. With the increase of binary basicity in the slag (E4>E3>E2>E1), the melilite phase disappears and begins to precipitate only C2S in the initial range of MgO. In other slag samples (E3 and E4) at 1550°C (1823K) and 1600°C (1873K), in the entire range of 0 to 25wt.% in MgO, the phases that are formed follows the sequence already mentioned: C2S, liquid slag and saturated MgO.

Slags E1 and E2 are in the condition of fully liquid slag, for 1550°C (1823K) and 1600°C (1873K), according to the MgO contents identified in the slag samples, by XRF (Table 1) and as visualized in Fig. 6(a) and (b). However, when it is considered the temperature of 1500°C (1773K), only the slag E2 would be completely liquid, because the slag E1 would be in the region of the solid phase C2S. However, for the compositions of samples E3 and E4, in the three temperatures, there is no liquid slag, but MgO in the precipitation process, making them saturated in this compound and in agreement with the results of Figs. 4(c), (d) and 5(b), (c). The temperature of 1500°C (1773K), although very low in relation to the temperature at which the samples were taken from the furnace, showed a better coherence between the thermodynamic calculations with comparison to the experimental results. This observation is plausible when considering the possible temperature reduction detailed previously on Materials and methods caption.

The results of the X-ray diffraction confirmed the C2S phase presence in the E1 sample, as identified by Fig. 4(a). Gran et al. [28], who also studied the CSAM system, reported the formation of dust during slag cooling. According to the authors [28], high amounts of calcium silicate (2CaO·SiO2) and absence of liquid phase in the slags are common factors related to the powder formation. However, in this study, according to Fig. 6(a), the formation of C2S was given together to a liquid portion. In other words, by the thermodynamic calculations, the powder verified for sample E1 was obtained even with presence of liquid fraction. This information is adverse to that proposed by Gran et al. study [28], reporting that when C2S and liquid coexist, the formation of dust is not observed. This divergence may be attributed to the cooling rate or to the change in slag chemistry, which can prevent the polymorphic transformations [26] commented previously.

The E2 sample showed C2S phases in the XRD and CaO and SiO2 precipitates in the SEM/EDS analysis. However, as already explained, FactSage™ did not predict the solid phase formation at the three analyzed temperatures (1500, 1550 and 1600°C) for this slag sample, considering the MgO content in Table 1. In this context, there was no agreement in the result from the thermodynamic calculation regarding E2 sample when compared to the experimental achievement. This divergence of the experimental result in relation to the thermodynamic calculations can be explained by the reasonable fact that the thermodynamic equilibrium has not been reached. In spite of the calculations well predicted the MgO saturation of slags E3 and E4.

As a complement and for better visualization, the exact calculated MgO ranges, where the phases described in Fig. 6 prevail, are shown in Table 4.

Table 4.

Composition ranges in MgO (wt.%) and phases formed in the slag for the temperatures of 1500°C (1773K), 1550°C (1823K) and 1600°C (1873K) calculated by FactSage™ 7.1.

Slag sample  Temperature  Melilite+C2C2Liquid slag  MgO  Spinel+MgO 
E1  1500°C (1773K)  0–3.27  0–6.53  6.53–8.34  >9.73  >9.73 
E2    0–1.88  0–6.40  6.40–8.68  >8.68  >10.5 
E3    0–0.77  0–6.77  6.77–8.04  >8.04  >16.85 
E4      0–15.67    >6.32   
E1  1550°C (1823K)  0–0.64  0–5.05  5.05–10.31  >10.31  >18.43 
E2      0–4.84  4.84–9.47  >9.47  >22.08 
E3      0–5.18  5.18–8.82  >8.82   
E4      0–7.03  7.03–7.41  >7.41   
E1  1600°C (1873K)    0–3.43  3.43–11.13  >11.13   
E2      0–3.20  3.20–10.32  >10.32   
E3      0–3.54  3.54–9.65  >9.65   
E4      0–5.51  5.51–8.17  >8.17   

From Table 4, it is possible to clearly observe that the saturation point of the slag in MgO is progressively reduced as it is advanced toward E4 sample in the three temperatures considered for the thermodynamic calculations. This phenomenon is associated to the slight increase in the binary basicity of the slag (Table 1), which reduces the MgO solubility [4,29]. The pseudoternary diagram, shown in Fig. 7, illustrates the displacement of slag samples in the direction of MgO saturation region through the CaO–SiO2–MgO system with Al2O3 fixed at 22.45wt.%. This Al2O3 content was calculated by taking the average of this compound from all slag samples (Table 1).

Fig. 7.

Pseudoternary diagram of the CMAS system (Al2O3=22.45wt.%) with isothermal cuts in 1500°C (1773K) and 1600°C (1873K). Calculated by FactSage™ 7.1.


As expected, the diagram of Fig. 7 shows the size difference of the liquid region (Slag-liq) for the two temperatures considered. For the higher temperature (1600°C (1873K)), the liquid region presented a small increase in relation to the lowest temperature (1500°C (1773K)). The saturation line of MgO (delimited by Slag-liq+Monoxide region) practically does not change with the increase of the temperature of 1500°C (1773K) to 1600°C (1873K). However, more precisely, in Table 4, it can be seen that the values of MgO where the saturation starts are slightly increased with temperature, with the highest increase recorded by slag E2, from 6.32 to 8.17wt.% in MgO, i.e., an increase of 1.85wt.%, when the temperature increases from 1500°C (1773K) to 1600°C (1873K). It is observed that the samples E1 and E2 are outside the saturation line in MgO, delimited by the light gray region (Fig. 7), with E1 being located in the zone of two phases (dark gray) containing liquid slag and C2S (Slag-liq+aCa2SiO4), considering the 1500°C (1773K) isotherm. With the increase of the MgO content in the slags (E4>E3>E2>E1) the saturation process is developed, being the slags E3 and E4 completely saturated in MgO, remaining within the light gray region (Slag-liq+Monoxide), where there is presence of liquid slag and MgO solids. This result is in accordance with previous comments involving Figs. 5(b), (c) and 6(c), (d).

The MgO saturation of the E3 and E4 slags indicated by the thermodynamic calculations with FactSage™ is in agreement with the results obtained experimentally and analyzed by the SEM/EDS technique. As already discussed, in Fig. 5(b) and (c), the formation of dark and dendritic-like morphologies suggests precipitation of the MgO compound in the slag. This fact shows that there is a good accuracy of the thermodynamic computational method to obtain the saturation point of the slags in MgO.


An experimental and thermodynamic investigation was developed with respect to the slag phases formed between the temperatures of 1500°C (1773K) and 1600°C (1873K) in the CaO–SiO2–Al2O3–MgO system, considering MgO increments. The increase of the MgO content in the slags of this system, from 14.96wt.%, identified the saturation condition in this compound and, for all slags studied, the binary basicity had effect of reducing MgO solubility. Regarding the evaluation of the MgO saturation in the slag, the thermodynamic results show consistency with the experimental achievements. Saturated slags in MgO showed dark precipitate formation in SEM/EDS images, often in the form of dendrites. Clear precipitates visualized in SEM/EDS rich in CaO and SiO2 were identified in the liquid matrix of the slag in the sample with 7.93wt.% in MgO (unsaturated condition) and identified in XRD as C2S phases.

Conflicts of interest

The authors declare no conflicts of interest.


The authors are grateful for financial support provided by the Coordination for the Improvement of Higher Education Personnel (CAPES) and National Council of Technological and Scientific Development (CNPq) from Brazil. To the Dr. Eng. Ricardo Thomé, from the Laboratory of Ceramic Materials (LACER/UFRGS), for the support in the diffraction analyzes. Special thanks also to the Laboratory of Physical Metallurgy (LAMEF/UFRGS) and the Chemical Laboratory of Gerdau Charqueadas.

Y. Sui, C. Yue, B. Peng, C. Wang, M. Guo, M. Zhang, et al.
Optimization of slag chemistry toward inclusion control for 28CrMo47 drill pipe steel based on viscosity and equilibration studies.
Steel Res Int, 87 (2016), pp. 752-760
H.L. Yang, J.S. Ye, X.L. Wu, Y.S. Peng, Y. Fang, X.B. Zhao.
Effect of top slag with low basicity on transformation control of inclusions in spring steel deoxidized by Si and Mn.
ISIJ Int, 56 (2016), pp. 108-115
W.J. Ma, Y.P. Bao, M. Wang, D.W. Zhao.
Influence of slag composition on bearing steel cleanness.
Ironmak Steelmak, 41 (2014), pp. 26-30
V.C. Rocha, J.A.M. Pereira, A. Yoshioka, W.V. Bielefeldt, A.C.F. Vilela.
Evaluation of secondary steelmaking slags and their relation with steel cleanliness.
Metall Mater Trans B, 48 (2017), pp. 1423-1432
R.J. Fruehan.
Unique functions of slags in steelmaking.
VII international conference on molten slags fluxes and salts, pp. 263-270
S. Seetharaman, L. Teng, M. Hayashi, L. Wang.
Understanding the properties of slags.
ISIJ Int, 53 (2013), pp. 1-8
J. Gran, B. Yan, D. Sichen.
Experimental determination of the liquidus in the high-basicity region in the Al2O3(25 mass pct)–CaO–MgO–SiO2 and Al2O3(35 mass pct)–CaO–MgO–SiO2 systems.
Metall Mater Trans B, 42 (2011), pp. 1008-1016
A.P. Luz, F.C. Leite, M.A.M. Brito, V.C. Pandolfelli.
Slag conditioning effects on MgO–C refractory corrosion performance.
Ceram Int, 39 (2013), pp. 7507-7515
C. Brüggmann, J. Pötschke.
MgO saturation in secondary metallurgical lime-aluminate and lime-silicate slags.
Steel Res Int, 82 (2011), pp. 422-427
M. Song, Q. Shu, D. Sichen.
Viscosities of the quaternary Al2O3–CaO–MgO–SiO2 slags.
Steel Res Int, 82 (2011), pp. 260-268
J. Xu, L. Su, D. Chen, J. Zhang, Y. Chen.
Experimental investigation on viscosity of CaO–MgO–(–Al2O3)–SiO2 slags and solid–liquid mixtures.
J Iron Steel Res Int, 22 (2015), pp. 1091-1097
L. Pengcheng, N. Xiaojun.
Effects of MgO/Al2O3 ratio and basicity on the viscosities of CaO–MgO–SiO2–Al2O3 slags: experiments and modeling.
Metall Mater Trans B, 47 (2016), pp. 446-457
A.V. Costa e Silva.
Estimating viscosities in iron and steelmaking slags in the CaO–Al2O3–MgO–SiO2–(TiO2) system with basis on a thermodynamic model.
J Mater Res Technol, 1 (2012), pp. 154-160
M. Jiang, X.H. Wang, W.J. Wang.
Study on refining slags targeting high cleanliness and lower melting temperature inclusions in Al killed steel.
Ironmak Steelmak, 39 (2012), pp. 20-25
B.H. Reis, W.V. Bielefeldt, A.C.F. Vilela.
Efficiency of inclusion absorption by slags during secondary refining of steel.
ISIJ Int, 54 (2014), pp. 1584-1591
H.L. Yang, X.L. Wu, J.S. Ye, Y. Fang, X.B. Zhao.
Control of sulphur for spring steel killed by Si–Mn and with low basicity slag.
Ironmak Steelmak, (2017), pp. 1-7
H. Yu, J. Xu, J. Zhang, X. Wang.
Effect of Al2O3 content on metallurgical characteristics of refining slag.
Ironmak Steelmak, 43 (2016), pp. 607-615
D. Roy, P.C. Pistorius, R.J. Fruehan.
Effect of silicon on the desulfurization of Al-killed steels: Part II. Experimental results and plant trials.
Metall Mater Trans B, 44 (2013), pp. 1095-1104
O.Y. Sheshukov, I.V. Nekrasov, M.A. Mikheenkov, D.K. Egiazar’yan, I. Ovchinnikova, L.A. Kashcheev, et al.
Effect of refining slag phase composition on ladle furnace unit lining life.
Refract Ind Ceram, 57 (2016), pp. 109-116
L. Kriskova, Y. Pontikes, L. Pandelaers, O. Cizer, P.T. Jones, K.V. Balen, et al.
Effect of high cooling rates on the mineralogy and hydraulic properties of stainless steel slags.
Metall Mater Trans B, 44 (2013), pp. 1173-1184
M. Tossavainen, F. Engstrom, Q. Yang, N. Menad, M.L. Larsson, B. Bjorkman.
Characteristics of steel slag under different cooling conditions.
Waste Manage, 27 (2007), pp. 1335-1344
W.V. Bielefeldt, A.C.F. Vilela.
Study of inclusions in high sulfur Al-killed Ca-treated steel via experiments and thermodynamic calculations.
Steel Res Int, 85 (2014), pp. 1-11
R.T. Cruz, G.F. Pelisser, W.V. Bielefeldt, S.R. Brangança.
Free opening performance of steel ladle as a function of filler sand properties.
Mater Res, 19 (2016), pp. 408-412
W.V. Bielefeldt.
Calcium treatment of non-metallic inclusions in SAE 1141 and SAE 8620 steels.
Federal University of Rio Grande do Sul, (2009),
(Dr. thesis)
What's New in FactSage 7.1. [cited in 2017 June 24]. Available from:
Y. Pontikes, P.T. Jones, D. Geysen, B. Blanpain.
Options to prevent dicalcium silicate-driven disintegration of stainless steel slags.
Arch Metall Mater, 55 (2010), pp. 1167-1172
K. Beskow, D. Sichen.
Ladle glaze: major source of oxide inclusions during ladle treatment of steel.
Ironmak Steelmak, 31 (2004), pp. 393-400
J. Gran, M. Thunman, D. Sichen.
Effects of slag composition and cooling rate on formation of glaze on MgO refractory.
Ironmak Steelmak, 37 (2010), pp. 27-34
M.A. Tayeb, A.N. Assis, S. Sridhar, R.J. Fruehan.
MgO solubility in steelmaking slags.
Metall Mater Trans B, 46 (2015), pp. 1112-1114
Copyright © 2018. Brazilian Metallurgical, Materials and Mining Association
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.