Ceramics International xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Ceramics International
journal homepage: www.elsevier.com/locate/ceramint

Influence of ladle slag composition in the dissolution process of the

dicalcium silicate (C2S) layer on doloma-C refractories
⁎
R.T. Cruz , W.V. Bielefeldt, S.R. Bragança
Federal University of Rio Grande do Sul, LACER/UFRGS, Av. Osvaldo Aranha, 99/711, Porto Alegre, RS 90135-190, Brazil

A R T I C L E I N F O A BS T RAC T

Keywords: Understanding the role of dicalcium silicate (C2S) in the corrosion mechanism of doloma-C refractory during
Dolomitic refractories the steel refining process is very important to maximize the performance of these refractories and to ensure the
Coating technical requirements of quality and operational safety. In industrial practice the spontaneous formation of
Corrosion dicalcium silicate phase (C2S) on doloma-C brick protects the refractory from slag penetration and is an example
Steelmaking
of protection by passive corrosion. However, fluxing compounds present in the slag, such as, Al2O3, CaF2 and
FactSage
FeOn can accelerate the dissolution of the C2S layer by the formation of phases of lower melting temperature.
Specifically, the influence of the fluxing oxide quantities and the reaction temperature of the phase formation,
which leads to melting/dissolution of the C2S, was studied. The thermodynamic simulation (FactSage software)
was performed to better understand the reactions aforementioned, and this data was validated through testing
in a laboratory furnace with precise control of temperature and atmosphere. The results provide important
information to the understanding of the phenomena involved and the optimization of industrial practice.

1. Introduction Herbert [4] studied the effect of fluorite (CaF2) and lime (CaO)
concentrations in the wear rate of magnesia bricks at 1650 °C. The
In a competitive and integrated world market, a deep knowledge of results of this study discuss the importance of the slag saturation in
refractories is necessary for their correct selection and application to CaO. Correspondingly, this indicates the risk of the excessive refractory
reduce the associated costs, and to ensure a greater useful life of the wear due to the uncontrolled fluorite use, which decreases both the
refractories in operation. melting temperature and viscosity of the slag, making it more
Doloma-C refractories are widely used as a ladle lining for the aggressive. Therefore, the formation/maintenance of the C2S layer
steelmaking process, especially in plants that make silicon killed steel. depends on the slag saturation in CaO and the control of fluorite level.
They exhibit excellent compatibility with basic slag, are resistant to Lee and Zhang [3] reviewed the types of refractory corrosion by
thermal shock and have high refractoriness [1]. The great advantage of liquid slag attack. According to the authors, the corrosion can be direct
these refractories is due to the spontaneous formation of the dicalcium or indirect. If the reaction product is soluble or dissociates directly into
silicate (2CaO.SiO2 or C2S) on the brick surface, a layer that protects liquid slag, it is established that a direct attack (active corrosion) will
the refractory from slag penetration, and it is an example of passive occur that may continue, destroying the refractory. However, if the
corrosion [2]. reaction product is not completely soluble in the liquid slag, a barrier
In the steelmaking process, the addition of fluxes agents (Al2O3, can occur at the slag-refractory interface, and the attack is slow. In this
FeOn and CaF2) is necessary to improve the physical properties of slags. case, the attack is said to be indirect (passive corrosion), with the
Fluxing agents will modify the amount of liquid fraction and will form corrosion rate becoming controlled by the chemical reaction forming
compounds of low melting temperatures. They will decrease the effective the C2S layer, diffusion through the layer or diffusion through the slag.
viscosity of the slag and increase its capacity to remove steel impurities. There are many studies concerning the performance of refractories
However, the use of these oxides requires special care. After formation of used in the steelmaking process, but the number of publications related
the C2S layer, the presence of these fluxing agents can dissolve this to C2S layer formation on doloma-C refractory surface in contact with
protective layer by the formation of phases with lower melting tempera- the slag is limited. Lee et al. [2,3] mentioned the formation of this layer
ture, such as Ca-aluminates and Ca-silicates. In this case, the dissolution in order to reduce the dissolution rate of the refractory. Jansson [5]
of the refractory can be accelerated by direct corrosion [3]. evaluated the dissolution rate of doloma refractory through the

⁎
Corresponding author.
E-mail address: thomedacruz@gmail.com (R.T. Cruz).

http://dx.doi.org/10.1016/j.ceramint.2017.08.076
Received 2 March 2017; Received in revised form 10 August 2017; Accepted 10 August 2017
0272-8842/ © 2017 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Please cite this article as: Thomé da Cruz, R., Ceramics International (2017), http://dx.doi.org/10.1016/j.ceramint.2017.08.076
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

immersion of samples in crucibles with synthetic slag; controlling in thermodynamic simulation were based on the chemical composition
chemical composition, temperature and rod rotation speed. Although (Table 2) of the C2S sample removed of the steel ladle.
some studies have related the C2S layer to the corrosion process and to Only oxides and fluorite constituents of the mixtures were con-
the performance of the doloma-C refractory, there is a lack of a specific sidered in the thermodynamic calculations. The presence of the gas
study about the characteristics, formation and dissolution of this layer. phase and the metal bath were not included in the calculations.
In a previous study by Cruz and Bragança [6], the C2S layer was The following databases were used and are described in the
investigated by characterizing chemical composition, phase formation documentation program [9]: The solutions database FToxid
and microstructure. This study showed the chemical and thermal (FToxid53Soln.sda) contains solutions oxides evaluated by the group
factors that can lead to the break-up or dissolution of the C2S layer that developed the program. The database FToxid (FToxid53Base.cdb)
formed on doloma-C brick surfaces. In the present work, the previous contains stoichiometric solid and liquid oxides. The database group
studies were complemented. The mechanism of C2S reaction/dissolu- FToxid includes data for pure oxides and 20 solution elements, and for
tion was evaluated as a function of the main fluxing compounds present diluted solutions of S, SO4, PO4, H2O/OH, CO3, F, Cl, I, C, N and CN in
in the steelmaking slag – Al2O3, CaF2 and FeOn. This research solutions molten slag (liquid). The FactSage databases were selected
investigated the influence of the phase formation and the correspond- using XRD and chemical analysis of the components studied. The
ing reaction temperature, which leads to melting/dissolution of the results obtained through the thermodynamic simulation software were
C2S. FactSage software was used to analyze the thermodynamics verified by practical tests conducted in the laboratory and bibliography
reactions of the compounds aforementioned. The simulated data was studies. All simulations were performed with temperatures varying
validated through tests in a laboratory furnace with precise control of from 0 to 1600 °C in steps of 50 °C.
temperature and atmosphere. The practical tests with the formulations from Table 1 were
performed in a furnace with an inert atmosphere (argon), a heating
rate of 300 °C/h until the temperature reaches 1600 °C (temperature
2. Experimental procedure usually employed in secondary refining of steel) with 30 min of dwell
time at 1600 °C. The cooling of the samples was performed with a rate
Dicalcium silicate (C2S layer) samples used in this work were at 13 °C/min at 1600 °C and 6.5 °C/min at 1000 °C. Therefore, the
removed from a ladle with doloma-C lining in normal process condi- samples were studied by X-ray diffraction analysis after the cooling
tions after a significant number of runs. The samples were obtained (room temperature). The phases present in the initial composition of
from a mini-mill that produces low carbon steels mainly used in civil the samples were incorporated into the slag through heating until high
construction. The C2S sample was named "C2S STD". temperatures. Using a slow cooling rate, the same high temperatures
Different formulations of C2S STD with increasing levels of fluxing phases were found to remain at room temperature, with exceptions of
agents (Al2O3, CaF2 and FeOn) were studied (Table 1, test A–K). Thus, C2S and FeOn polymorphisms, which will be discussed later.
the dissolution potential of the C2S layer was analyzed according to C2S It was not possible to measure the oxygen partial pressure inside
STD reaction with different concentrations of industrial fluorite (CaF2, the furnace. Wear of the crucibles did not happen during the experi-
Nossa Senhora do Carmo Mining Co., > 88%), alumina (Al2O3, Almatis ments, which is an indication that the slag formed did not chemically
Co., > 97%) and iron (III) oxide (Fe2O3, Vetec Fine Chemicals Co., > react with the material. In addition, there was no presence of metallic
98%). In addition to the individual analysis between fluxing com- iron in the slag after the experiments. The liquid slags in the FeO-MnO-
pounds and C2S, the dissolution of C2S was evaluated considering the CaO-SiO2-Al2O3 system at 1500 °C has a contact angle between 110
CaF2 concentration in the slag ladle (Table 1, test L to O). In this work, and 132° with graphite. Usually, contact angles greater than 90° are
all components used in the different formulations (Table 1) were in a called low wetting liquids. In other words, as the contact angle
powder state. increases, the liquid separates from the solid [10]. The high purity of
The evaluation of the thermal behavior of the compositions shown the gas used as a protective atmosphere also helps to minimize the
in Table 1 was carried out by using the thermodynamic simulation formation of reducing gases.
software (FactSage 6.3) [7]. The calculations performed by the program All tested formulations were weighed (mass of each test: 25 g) and
are based on the minimization of the free energy of the system and placed in graphite crucibles (Fig. 1), which provide high thermal
provide information about the phases formed, their proportions and conductivity and were found to be inert in the range of temperatures
compositions, the individual activities of each chemical component and and atmosphere used during the tests. The experimental laboratorial
thermodynamic properties for various compositions, pressures and corrosion tests were static (no rotation or other movement).
temperatures [8]. It is important to note that all calculations performed The furnace (Fig. 2) consists of a vertical graphite tube (where the
crucible is placed) surrounded by a graphite tubular heating element.
Table 1 This set is covered with a thermal insulator layer (alumina and carbon
Proportion (weight %) of the *C2S STD and fluxing agents evaluated in this study.
fibers composite), situated in a stainless steel cube, which forms the
Test System Composition (weight %) frame furnace. The tube and the chamber receive high purity argon
(supplied by Air Liquid Co. Purity 99.999% with N2 < 3 ppm, O2 <
A C2S 100 1 ppm, CO < 0.5 ppm, THC < 0.1 ppm, CO2 < 0.5 ppm, H2O < 2 ppm)
B C2S – CaF2 95 – 5
injection during all the test (heating and cooling).
C C2S – CaF2 90 – 10
D C2S – CaF2 80 – 20 After the heating and cooling cycle, the samples were studied by X-
E C2S – Al2O3 95 – 5 ray diffraction analysis (Philips X′Pert). Because of the slower cooling
F C2S – Al2O3 90 – 10 rate, equilibrium phases at room temperature are expected. Using the
G C2S – Al2O3 80 – 20 cooling technique, it was possible to compare the phases formed with
H C2S – Fe2O3 90 – 10
I C2S – Fe2O3 70 – 30
thermodynamic simulations as a function of flux content. The chemical
J C2S – Fe2O3 – Al2O3 90 – 5 – 5 analysis of the C2S STD and slag (Table 2) were performed by X-ray
K C2S – Fe2O3 – Al2O3 80 – 10 –10 fluorescence (XRF spectrometer Shimadzu-1800).
L C2S – Slag 50 – 50
M C2S – CaF2 – Slag 47.5 – 5 – 47.5
3. Results and discussion
N C2S – CaF2 – Slag 45 – 10 – 45
O C2S – CaF2 – S 40 – 20 – 40
Table 2 presents the chemical composition of the C2S STD and slag
*C2S = (2CaO·SiO2). (after deoxidation) used in this work. The “C2S STD” characterization

2
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Table 2
Chemical composition in wt% of the *C2S STD and slag after deoxidation.

Weight (%) CaO SiO2 Al2O3 Fe2O3 MgO MnO Cr2O3 P2O5 SO3 TiO2

C2S STD 59.92 34.85 0.57 0.15 3.32 … … … 1.19 …

Slag 47.32 23.62 9.83 0.71 17.04 0.52 0.01 0.01 0.48 0.46

*C2S = (2CaO·SiO2).

(chemical, physical and mineralogical analysis) can be found else-

where, (Cruz and Bragança [6]).
The chemical analysis of the C2S STD shows that there are some
impurities (MgO, SO3, Al2O3, Fe2O3) in addition to the predominant
calcium and silicon oxides. According to Gomes [11], C2S can have 4–
6 wt% of other elements in solid solution. For example, the C2S crystal
tends to concentrate sulphur in its structure in cases involving excess
sulphur during the steel refining. This explains the presence of SO3 in
the C2S sample.
The C2S dissolution is explained below considering the different
systems with increasing levels of fluxing agents (CaF2, Al2O3 and
FeOn). As mentioned before (Item 2), all FactSage calculations were
based on the chemical composition (Table 2) of the C2S sample
removed of the steel ladle.

3.1. C2S + CaF2 system

The thermodynamic predictions made during the investigation of Fig. 2. Schematic drawing of the furnace inner chamber used in the practical tests.
the C2S layer dissolution as a function of CaF2 content (according
Table 1, test A–D) are shown in Fig. 3. Some differences between literature data (Fig. 4) and thermody-
The thermodynamic calculation results for 100 wt% C2S STD (Fig. 3 namic calculation results of C2S STD (Fig. 3A) were observed. They can
A) show the polymorph γ-C2S (orthorhombic system) stable at room be explained by the fact that the program considers the thermodynamic
temperature. After heating between 500 and 800 °C, the polymorph equilibrium calculations, therefore, metastable phases such as β
transformation of the γ-C2S to α′-C2S occurs, and above 1425 °C, the polymorph are not considered by the program.
α′-C2S becomes α polymorph. The addition of 5 wt% fluorite (test B) increases the liquid phase
The phase transformation of the C2S STD (test A) without addition of (slag) to 55 wt% at 1600 °C (Fig. 3B). For this composition, the
fluorite (Fig. 3 A) shows that there is 15 wt% of liquid phase (slag) and increase in the liquid amount is explained by the formation of the
85 wt% of solid phase (C2S) at 1600 °C. According Eriksson [12], the cuspidine phase (Ca4Si2F2O7), which has a low melting point, around
solidus temperature (first liquid formation) and liquidus (melting tempera- 1407 °C, [14].
ture) of the pure α-C2S are about 1464 and 2154 °C, however, according to The increase of fluorite content in the C2S-CaF2 system supports
Fig. 3 A, above 1200 °C the formation of the liquid phase begins. This cuspidine formation and, consequently, the formation of liquid at lower
difference can be explained by the presence of impurities in the sample. temperatures. When 10 wt% fluorite is added to the standard mix (test C),
According to Mumme et al. [13], the cooling of C2S below 670 °C the amount of liquid phase (slag) rises to 90 wt% at 1600 °C (Fig. 3 C). In
produces the β polymorph (monoclinic; metastable phase), which is not this composition, the formation of liquid phase begins below 1000 °C. The
formed during heating from the γ polymorph. When the β polymorph is formation of the CaF2 phase (about 1 wt%) can be attributed to the
cooling to temperatures below 500 °C, a reconstructive and irreversibly beginning of fluorite saturation in the C2S – 10 wt% CaF2 system.
transformation produces the γ polymorph (Fig. 4).

Fig. 1. Image of graphite crucibles (high purity) into the vertical graphite tube.

3
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Fig. 3. Evolution of the phases formed in the system (C2S + CaF2) as a function of the temperature. See formulations in Table 1. (A) 100 wt% C2S; (B) C2S + 5 wt% CaF2; (C) C2S + 10 wt
% CaF2; (D) C2S + 20 wt% CaF2. The presence of the gas phase and the metal bath were not included in the calculations. C2S composition used in the calculations is shown in Table 2.
Data generated in FactSage 6.3.

Despite the great industrial importance, the utilization of fluorite as

a fluxing agent in the slag must consider the dissolution of the C2S
layer. The results indicate that even small concentrations of CaF2 are
able to increase extensively the proportion of liquid in the system:
10 wt% of CaF2 is sufficient to obtain 90 wt% liquid in the C2S-CaF2
system at 1600 °C. According to the fluorite content, the dissolution of
the C2S layer, which is formed on doloma brick surface, leading to
Fig. 4. C2S polymorphic transformations temperatures, under atmospheric pressure direct refractory corrosion after the C2S layer is removed, causing
[13]. possible destruction of the refractory brick.

3.2. C2S + Al2O3 system

When the fluorite content is increased to 20 wt% (test D), 100% of a
liquid phase (slag) exists between 1200 and 1600 °C (Fig. 3D). The thermodynamic results of the C2S layer dissolution as a
Therefore, the fluorite content is supersaturated, and besides, with function of alumina concentration (according Table 1, tests E, F and
the presence of cuspidine, about 10 wt% of unreacted fluorite was G) are shown in Fig. 6.
formed. Fig. 6E presents the test E (95 wt% C2S – 5 wt% Al2O3). The result
The results of the phase formation in the laboratory tested C2S- of this simulation indicates that the liquid phase formation begins
CaF2 system (Fig. 5) indicated similar results with the phases obtained around 1050 °C. The proportion of liquid phase (slag) increases
by thermodynamic simulation. According to XRD analysis, the follow- continuously with temperature, and by 1600 °C, there is about 40 wt
ing phases were identified in the experimental test: γ-C2S (crystallizes % of liquid (slag) in the system. The addition of alumina to C2S STD
in the orthorhombic system with a structure similar to the mineral favors gehlenite (Ca2Al2SiO7) crystallization. According to Zigo et al.
olivine), cuspidine (Ca4Si2F2O7), periclase (MgO), fluorite (CaF2) and [15], this phase has a melting point of 1590 °C. This explains the
calcium magnesium silicate (Ca3Mg(SiO4)2) as a minor phase. The significant amount of liquid at 1600 °C, since the increase of the liquid
phases γ-C2S, periclase and cuspidine are present in all compositions phase occurs by the incorporation of gehlenite to the liquid slag.
and fluorite increments. The increased intensity of cuspidine peak (at The results of this simulation indicate that the total mass of
29°) indicates a higher proportion of this phase due to the fluorite alumina added to the C2S-Al2O3 system reacts with silica and calcium
additions. Moreover, the intensity of fluorite (at 28°) increases with the (C2S) to form gehlenite. Thermodynamic studies indicated that in-
fluorite content, indicating the presence of unreacted fluorite. creasing alumina from 0 to 20 wt%, more gehlenite would be formed.

4
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Fig. 5. Evolution of the phases formed in the system (C2S-CaF2) as a function of the fluorite content. Samples subjected to heating at 1600 °C in an inert atmosphere with subsequent
cooling to room temperature.

Therefore, in the C2S-Al2O3 system, when the alumina content is trolled during the secondary refining of steel process; otherwise the
elevated, a higher liquid percentage is expected at the steel process protective C2S layer can be destroyed.
temperature. Fig. 7 shows the mineralogical analysis by XRD of the experimental
According to Fig. 6F (10 wt% of alumina) and G (20 wt% of tests A, E, F and G (Table 1) that were carried out to verify the results
alumina), there are 62 wt% and 100 wt% of liquid phase at 1600 °C, obtained through the FactSage program.
respectively. In other words, the alumina percentage must be con- The analyzes of the phases formed in the C2S – Al2O3 system

Fig. 6. Evolution of the phases formed in the system (C2S + Al2O3) as a function of the temperature. See formulations in Table 1. (E) C2S + 5 wt% Al2O3; (F) C2S + 10 wt% Al2O3; (G)
C2S + 20 wt% Al2O3. The presence of the gas phase and the metal bath were not included in the calculations. C2S composition used in the calculations is shown in Table 2. Data generated
in FactSage 6.3.

5
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Fig. 7. Evolution of the phases formed in the system (C2S-Al2O3) as a function of the alumina content. Samples subjected to heating at 1600 °C in an inert atmosphere with subsequent
cooling to room temperature.

(Fig. 7) indicates the increase of gehlenite and decrease of γ-C2S phase there is a saturation of this compound in the C2S-Fe2O3 system. For
as a function of rising alumina content. This occurs because the Ca and this system, it was observed that the liquid phase starts at 1050 °C, and
Si present in the γ-C2S react with alumina to form gehlenite. The by 1600 °C constitutes 58 wt% of the phases in the system.
analysis of the test G (20 wt% Al2O3), for example, indicates a complete Increasing Fe2O3 to 30 wt% in the C2S-Fe2O3 system (Fig. 8I) does
reaction of the alumina added to the C2S STD to form gehlenite, which not modify the andradite content, which remains about 20 wt%. In
is the majority phase in this composition. Besides gehlenite, the equilibrium conditions, the excess iron oxide does not react with the Si
analysis identified the presence of MgO and γ-C2S. The results of the and Ca from C2S. At this concentration, the Fe2O3 content is super-
practical tests E, F and G (Table 1) indicate a strong similarity with the saturated in the system, occurring as excess Fe2O3. The liquid phase in
data obtained by thermodynamic simulation. this composition begins around 1050 °C, and by 1400 °C, there is
100 wt% of liquid (slag). The large increase in the liquid percentage at
3.3. C2S + Fe2O3 system temperature below 1600 °C is explained by the incorporation of Fe2O3
and α'-C2S to the slag.
Fig. 8 presents the C2S dissolution as a function of Fe2O3 content At equilibrium conditions Fe2O3 is stable until ~ 1200 °C, where it
(Table 1, test H and I). rapidly begins to be incorporated in the liquid phase (slag). The authors
According to FactSage simulations, the addition of 10 wt% of Fe2O3 assumed: no reaction with carbon, and no influence of gas phase (both
to the C2S STD (Fig. 8H), favors the formation of andradite are the settings in FactSage). However, during the lab tests, a non-
(Ca3Fe2(SiO4)3), that has a melting point around 1200 °C. Thus, the equilibrium condition, part of the Fe2O3 (or FeO) did not react to form
Andradite crystallization supports the liquid (slag) formation and the slag, and was found in the DRX analysis, as Fe2O3.
consequently the dissolution of the C2S layer in industrial practice. The comparative crystallographic analysis of the practical tests H
Moreover, the presence of unreacted hematite (Fe2O3) indicates that and I (Fig. 9) shows that the behavior of both compositions is very

Fig. 8. Evolution of the phases formed in the system (C2S + Fe2O3) as a function of the temperature. See formulations in Table 1. (H) C2S + 10 wt% Fe2O3; (I) C2S + 30 wt% Fe2O3. The
presence of the gas phase and the metal bath were not included in the calculations. C2S composition used in the calculations is shown in Table 2. Data generated in FactSage 6.3.

6
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Fig. 9. Evolution process of the phases formed in the system (C2S – Fe2O3) as a function of the iron oxide content. Samples subjected to heating at 1600 °C in an inert atmosphere with
subsequent cooling to room temperature.

similar. Although test ‘I’ possess a significantly higher Fe2O3 content, (slag) at lower temperature occurred as a function of fluxing agents
the crystallographic analysis did not identify this difference. The additions. Fig. 10K shows 100 wt% of the liquid phase at 1600 °C.
explanation can be found in the ternary diagram of CaO-SiO2-FeOn Fig. 11 shows the crystallographic analyses of practical tests J and K
system [16], whose corresponding area to α-C2S phase is very large, (Table 1).
even for high levels of FeOn (where FeOn means ferrous or ferric iron), Test J (addition of 5 wt% Fe2O3 and 5 wt% Al2O3 in the C2S STD)
causing similar behavior between the crystallographic analysis of the shows the presence of γ-C2S, gehlenite, periclase and andradite phases
tests H and I. (Fig. 11). The addition of 10 wt% of Al2O3 and 10 wt% of Fe2O3 to the
C2S STD (test K) also showed data that confirms the thermodynamic
simulations (Fig. 10), i.e., gehlenite as the main phase in this
3.4. C2S + Al2O3 + Fe2O3 system
composition, and the same amount of andradite. Although not
identified, the peaks corresponding to the hematite phase may be
Fig. 10 shows the addition effect of Al2O3 and Fe2O3 in the C2S STD,
covered by andradite and γ-C2S phases, which have more intense peaks
according to the compositions in Table 1 (Tests J and K).
in the same 2 Theta position.
Gehlenite and andradite phases (Fig. 10J) are formed as a
consequence of the addition of 5 wt% Fe2O3 and 5 wt% Al2O3 to the
C2S STD. Increasing the two compounds to 10 wt% (Fig. 10K) favors 3.5. Influence of fluorite content in the slag
the formation of hematite and raises the amount of gehlenite, while
andradite remains constant. This behavior is analogous to the fluxing Fig. 12 shows the thermodynamic simulations considering a slag
agents individual simulations showed previously. Comparing the with increasing levels of CaF2 (Table 1, tests L, M, N and O). The
systems J and K (Fig. 10), a significant increase of the liquid content calculations performed in thermodynamic simulation were based on

Fig. 10. Evolution of the phases formed in the system (C2S + Al2O3 + Fe2O3) as a function of the temperature. See formulations in Table 1. (J) C2S + 5 wt% Al2O3 + 5 wt% Fe2O3; (K)
C2S + 10 wt% Al2O3 + 10 wt% Fe2O3. The presence of the gas phase and the metal bath were not included in the calculations. C2S composition used in the calculations is shown in
Table 2. Data generated in FactSage 6.3.

7
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Fig. 11. Crystallographic analyses in the system C2S – Fe2O3- Al2O3. Tests J (5 wt% Fe2O3 and 5 wt% Al2O3 in the C2S STD) and K (10 wt% Fe2O3 and 10 wt% Al2O3 in the C2S STD).
Samples subjected to heating at 1600 °C in an inert atmosphere with subsequent cooling to room temperature.

Fig. 12. Evolution of the phases formed in the system (C2S + Slag + CaF2) as a function of the temperature. See formulations in Table 1. (L) 50 wt% C2S + 50 wt% slag; (M) 47,5 wt%
C2S + 47,5 wt% slag + 5 wt% CaF2; (N) 45 wt% C2S + 45 wt% slag + 10 wt% CaF2; (O) 40 wt% C2S + 40 wt% slag + 20 wt% CaF2. The presence of the gas phase and the metal bath were
not included in the calculations. C2S and slag composition used in the calculations are shown in Table 2. Data generated in FactSage 6.3.

8
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

Fig. 13. Evolution process of the phases formed in the system (C2S–Slag–CaF2). Tests (L) 50 wt% C2S + 50% Slag, (M) 47,5 wt% C2S + 47,5 wt% Slag + 5 wt% CaF2, (N) 45 wt% C2S +
45 wt% Slag + 10 wt% CaF2, (O) 40 wt% C2S + 40 wt% Slag + 20 wt% CaF2. Samples subjected to heating at 1600 °C in an inert atmosphere with subsequent cooling to room
temperature.

the chemical composition of the C2S STD and slag (C2S and slag Ca3Mg(SiO4)2 reacts with fluorite to form cuspidine.
composition are shown in Table 2). The FactSage databases were selected using XRD and chemical
The test L (C2S 50% – 50% slag), shows that at 1600 °C, the system analysis of the components studied. The Ca3Mg(SiO4)2 is not present in
is 46 wt% liquid (Fig. 12L). The addition of 5 wt% of fluorite (Fig. 12M) the FactSage results, but a small peak was identified by XRD. According
increases the amount of liquid phase to 85 wt% at 1600 °C. This can be to the literature, the diagram CaO-MgO-SiO2 present Ca3Mg(SiO4)2 as a
explained by the crystallization of cuspidine (Ca4Si2F2O7), which has a stable phase. This demonstrates the importance of analyzing the data by
low melting point (~ 1407 °C), favoring the increase of the liquid crossing information obtained in FactSage, DRX and literature.
content at 1600 °C. Therefore, considering all information (FactSage, DRX and literature),
The amount of liquid phase at lower temperatures is significantly a small amount of Ca3Mg(SiO4)2 is present at room temperature.
higher when 10 wt% of fluorite is added to C2S-Slag-CaF2 system
(Fig. 12N). The system is 100 wt% liquid above1450°C.
When the fluorite content is increased to 20 wt% in the C2S-Slag- 3.6. Comparative analysis of the C2S dissolution tests evaluated in
CaF2 system (Fig. 12, O), 100 wt% liquid slag phase is obtained above this work
1223 °C. The amount of cuspidine phase remains unchanged, but the
presence of unreacted fluorite is observed, indicating that the system is The data of the thermodynamic simulations for the different
supersaturated in fluorite. The excess of fluorite explains the develop- systems studied in this work are summarized in Table 3. This data is
ment of the liquid phase at lower temperatures. According to the very important to estimate the C2S protective layer dissolution as a
ternary diagram CaO-MgO-SiO2 [17], the higher concentration of CaF2 function of the type and concentration of the fluxing agent used.
favors a larger amount of liquid at 1600 °C. Table 3 shows that fluorite is the most effective compound for the
Fig. 13 shows the phases formed in the system C2S-Slag-CaF2 due dissolution of C2S. For example, considering tests C, F and H (10 wt% of
to the increase of the fluorite content (laboratory tests L, M, N and O). the fluxing agents CaF2, Al2O3 or Fe2O3 combined with C2S, respectively),
The experimental results are comparable to the data simulated in the it is observed that at 1600 °C there is 90, 62 and 58 wt% of liquid.
FactSage. As stated before, Ca and Si present in the γ-C2S and The influence of the fluorite is higher considering the C2S–Slag
system. The additions of only 5 wt% of CaF2 into C2S–Slag system (test

Table 3
Liquidus and solidus temperatures and liquid percentage at 1600 °C for the systems evaluated in this work.

Test Composition (wt%) Liquidus temperature (°C) Solidus temperature (°C) wt% liquid (1600 °C)

A 100% – C2S > > > 1700 1200 15

B 95% C2S – 5% CaF2 > > 1700 1150 55
C 90% C2S – 10% CaF2 1650 950 90
D 80% C2S – 20% CaF2 1200 950 100
E 95% C2S – 5% Al2O3 > > 1700 1050 39
F 90% C2S – 10% Al2O3 > 1700 1050 62
G 80% C2S – 20% Al2O3 1555 1050 100
H 90% C2S – 10% Fe2O3 > > 1700 1100 58
I 70% C2S – 30% Fe2O3 1400 1100 100
J 90% C2S – 5% Fe2O3 – 5% Al2O3 > > 1700 1200 60.5
K 80% C2S – 10% Fe2O3 – 10% Al2O3 1581 1200 100
L 50% C2S – 50% Slag > > 1700 1100 46
M 47.5% C2S – 5% CaF2 – 47.5% Slag 1691 1000 85
N 45% C2S – 10% CaF2 – 45% Slag 1450 1000 100
O 40% C2S – 20% CaF2 – 40% Slag 1223 1000 100

9
R.T. Cruz et al. Ceramics International xxx (xxxx) xxx–xxx

M) nearly doubled the percentage of liquid at 1600 °C. This data is very The thermodynamic program cannot accurately represent all the
important because it means that the fluorite becomes even more experimental conditions during the tests. However, it shows important
aggressive to C2S; considering the interaction with the oxides SiO2, information considering the parameters settled which correlated well
CaO, MgO, and FeO present in the slag. with DRX analysis.

4. Conclusions Acknowledgements

According to the thermodynamic simulation, the development of The authors would like to acknowledge the financial contribution
the liquid phase in the slag as a function of the fluxing agents (CaF2, provided by the Brazilian Coordination for the Improvement of Higher
Al2O3 and Fe2O3) occurred through the formation of low melting Level Personnel (Capes).
phases that can compromise the stability of the C2S.
In the tests with incremental amounts of fluorite, cuspidine References
formation is responsible for the increased liquid formation. In the
tests involving increased amounts of Al2O3 and Fe2O3, the formation of [1] C.A. Schacht, Refractories Handbook, Marcel Decker, 2004.
the gehlenite and andradite phases, respectively, can be related to the [2] W.E. Lee, S. Zhang, Direct and indirect slag corrosion of oxide and oxide-c
refractories, in: VII International Conference on Molten Slags Fluxes and Salts, The
increased liquid phase at 1600 °C. South African Institute of Mining and Metallurgy, 2004.
The comparison among fluxing agents CaF2, Al2O3 and Fe2O3 [3] W.E. Lee, S. Zhang, Melt corrosion of oxide-carbon refractories, in: International
clearly indicates that fluorite is the most powerful compound with Materials Reviews, Communications Ltd and ASM International, Sheffield, UK, vol.
44(3), 1999, pp. 309–320.
respect to the dissolution of C2S. When in contact with C2S, even small [4] W. Herbert, et al., Refractory Wear and Steel Desulfurization during Vacuum
quantities of CaF2 are able to increase the liquid proportion consider- Degassing with Controlled Fluorspar Additions, Metallurgical Plant and
ably. 5 wt% and 10 wt% of CaF2 in the system CaF2-C2S are able to Technology International, 1991.
[5] S. Jansson, V. Brabie, P. Jonsson, Corrosion mechanism of commercial doloma
form 55 wt% and 90 wt% of liquid, respectively, at 1600 °C.
refractories in contact with CaO-Al2O3-SiO2-MgO slag, Ironmak. Steelmak. 35 (2)
Furthermore, the presence of fluorite decreases considerably the initial (2008) 99–107.
melting temperature during the heating process. The fluorite becomes [6] R.T. da Cruz, S.R. Bragança, Evaluation of the protective C2S layer in the corrosion
process of doloma-C refractories, Ceram. Int. 41 (2015) 4775–4781.
even more aggressive to C2S, considering the interaction with the
[7] 〈http://www.factsage.com〉.
oxides SiO2, CaO, MgO, and FeO present in the slag. The additions of [8] C.W. Bale, P. Chartrand, S.A. Degterov, G. Eriksson, K. Hack, R. Ben Mahfoud,
only 5 wt% of CaF2 into C2S–Slag system (test M) nearly double the J. Melançon, A.D. Peltron, S. Petersen, FactSage thermochemical software and
percentage of liquid at 1600 °C. Therefore, despite the great industrial databases, Calphad 26 (2) (2002) 189–228.
[9] FactSage Database Documentation. Part Integrant of FactSage Software Version 6.3
importance, fluorite use requires special cares. and Revised in 2012.
The use of Al2O3 and Fe2O3 compounds also requires attention, [10] A.W. Cramb, I. Jimbo, Interfacial Considerations in Continuous Casting – I & SM,
although they are less strong than fluorite. 62 wt% and 58 wt% of liquid June 1989, pp. 43–55.
[11] S.D. Gomes, Incorporation of Sulfur in Dicalcium Silicate Portland Cement Clinker:
are formed at 1600 °C, when 10 wt% of Al2O3 and Fe2O3 are added to Influence on the Stability of Polymorphs and the Hydraulic Reactivity
C2S (Tests F, H), respectively. Thus, greater amounts of these (Dissertation), São Paulo University, 2007.
compounds in the slag can promote the C2S layer dissolution on the [12] G. Eriksson, P. Wu, M. Blander, A.D. Pelton, Critical evaluation and optimization of
the thermodynamic properties and phase diagrams of the MnO-SiO2 and CaO–
brick surface, which means that the dissolution of the dolomitic SiO2 systems, Can. Metall. Q. 33 (1994) 13–21.
refractory can be accelerated by direct corrosion. On the other hand, [13] W.G. Mumme, R.J. Hill, G. Bushnell-Wye, E.R. Segnit, Rietveld crystal structure
well-calculated quantities of Fe2O3 or Al2O3 may be interesting requirements, crystal chemistry and calculated powder diffraction data for the
polymorphs of dicalcium silicate and related phases, N. Jahrb. Mineral. Abh. Stuttg.
alternatives to fluorite in actions that require the increase of the liquid 169 (1995) 35–68.
phase in the slag. [14] T. Watanabe, H. Fukuyama, K. Nagata, Stability of cuspidine (3CaO 2SiO2·CaF2)
The simultaneous addition of Fe2O3 and Al2O3 to C2S STD (J and K and phase relations in the CaO-SiO2-CaF2 system, ISIJ Int. 42 (5) (2002) 489–497.
[15] O. Zigo, K. Adamkovicová, L. Kosa, I. Nerád, I. Proks, Determination of the heat of
tests) did not increase the dissolved amount of C2S. For these tests, it
fusion of 2CaO-A12O3-SiO2 (gehlenite), Chem. Pap. 41 (2) (1987) 171–181.
was observed that the behavior was analogous to the individual [16] E.B. Pretorius, R.C. Carlisle, Foamy slag fundamentals and their practical appli-
simulations of these fluxing agents. Therefore, there is no synergistic cation to electric furnace steelmaking, Iron Steelmak. ISS 26 (10) (1999) 79–88.
action of these two oxides.

10