Fig. 4 - steel-grips.comsteel-grips.com/articles/2011/g02360.pdf · 2020. 12. 30. · foaming index of the slag [1; 9…11]. In addition, solid particles can serve as nucleation sites

reviewed paper from R&D

9 (2011) Process technology 355

Saman Mostafaee and Pär G. Jönsson:

Influence of slag properties on the amount of solid precipitatesin high-chromium EAF steelmaking

The total amount of solid particles as a parameter influencing the viscosity and the foaming properties of the slag was investi-gated. In this context, the amount of these particles for different process conditions was quantified. In addition, the effect of theprocess conditions on the total amount of the particles was studied. More specifically, some parameter studies were carried outin order to determine the influence of chromium oxide (Cr2O3) content, calcium oxide (CaO) content, basicity and temperatureon the total amount of solid particles. The interactions between the process conditions were also taken into account.The results were visualized in diagrams and thereafter discussed in details. It is shown that the total amount of solid particles isgenerally influenced by the chromium oxide content, basicity and slag temperature. An increased chromium oxide contentgives rise to the precipitation of solid chromium containing complexes. In addition, the amount of dicalcium silicate precipitatesis a function of the slag basicity and temperature. Higher basicity and lower slag temperature promote dicalcium silicate pre-cipitation.

The influence of solid precipitates on the foamability ofelectric arc furnace (EAF) slags in stainless steelmakingwas expressed by several researchers. Vidacak et al. [1]showed that the slags with a lower amount of chromiumcontaining precipitates have better foamability than slagswith a higher amount of solid particles. The results wereunexpected, since the latter contained more solid precipi-tates that could increase the apparent viscosity and foamingindex of the slag [2; 3].

Kerr and Fruehan [4] proposed a dependency betweenthe foamability of the slag and the solid constituents withinthe slag. They reported that when the solubility limit forchromium oxide is exceeded, chromium containing second-phase particles form in the slag. These solid particles in-crease the apparent viscosity of the slag. This, in turn, leadsto a higher foaming index. However, Kerr and Fruehan [4]observed that when the amount and the size of chromiumcontaining precipitates exceeded a particular limit, thefoamability of the slag contrarily decreased. The suggestedtrend is illustrated in fig. 1.

The solid part of the EAF stainless steelmaking slagsmainly consists of magnesiochromite spinels (MgCr2O4)[5…7]. However, Mostafaee et al. [8] showed that, de-pending on the slag basicity (CaO/SiO2) and temperature, adicalcium silicate phase (2CaO·SiO2) can also precipitateduring the EAF stainless steelmaking. Higher basicities andlower temperatures lead to higher amounts of dicalciumsilicate precipitates. In addition, high-chromium stainlesssteelmaking slags can also contain small amounts (around 3

mass contents in %) of a calcium chromite phase(CaO·Cr2O3). Formation of this phase depends on thechromium oxide content of the slag. This phase only oc-curred in the slag samples taken before ferrosilicon (FeSi)addition to the EAF. At this stage of the process, the chro-mium oxide content of the slag is on its highest level [7].

Fig. 2 illustrates a typical phase distribution diagram forthe slag samples taken during the refining stages in EAFhigh-chromium steelmaking. The diagram was calculatedfor different slag temperatures [8]. The vertically dashedline in the diagram represents the slag temperature (1726oC) at sampling. According to this diagram, the slag con-sists of a magnesiochromite phase (mass contents of around6 %) and a liquid part (mass contents of around 94 %) atprocess temperature.

As referenced above, it is generally accepted that the ad-dition of solid particles normally improves foaming char-acteristics. Moreover, Ito and Fruehan [2; 3] showed thatthe existence of the solid particles as occurring in the slag

Fig. 1: Relationship between foaming index and second-phase particles [4]

Fig. 2: A typical phase distribution in the slag during theEAF stainless steelmaking [8]

from R&D reviewed paper

9 (2011) Process technology356

has a large effect on the foam stability. More specifically, itwas shown that these solid particles lower the drainage rateof the liquid from the slag films that separate the bubbles[3]. Solid particles also increase the bulk viscosity of theslag. An increased viscosity, in turn, results in an enhancedfoaming index of the slag [1; 9…11]. In addition, solidparticles can serve as nucleation sites for gas bubbles. Thisleads to the nucleation of a large number of small gas bub-bles, which, in turn, leads to improved slag foaming [12;13].

The present paper focuses studying the effect of the proc-ess parameters on the amount of the solid precipitateswithin the slag during the EAF high-chromium stainlesssteelmaking. More specifically, the influence of slag ba-sicity, temperature and composition on the total amount ofsolid particles is illustrated and discussed.

Methods

Sampling and petrographical analysis. The samplingmethod and procedure were explained and illustrated indetail in [7]. Sampling was carried out at different processstages for each EAF heat. Simultaneously to slag sampling,the temperature of the slag samples was also measured.Samples were taken from seven EAF duplex stainless steelheats (mass contents of 21.5 - 22.5% Cr, 1.6 - 5.7% Ni and0.3 - 3.2% Mo).

In the present study, the chemical composition of the slagsamples taken before ferrosilicon (FeSi) addition is used asa base for the calculations. However, the performed pa-rameter studies include different contents of chromium ox-ide and calcium oxide, slag basicitiesand temperatures. Thus, the achievedresults can be valid even for otherEAF process stages, such as the refin-ing and temperature increasing period.A plant description for the meltshop isalso given in an earlier paper [7].

Table 1 presents an average chemical composition of theslag samples [8]. The averaged temperature of the slag isalso presented. Fig. 3 illustrates the phase distribution dia-gram of a slag with the composition and temperature pre-sented in table 1 [8].

The vertical dashed lines in the diagram (fig. 3) representthe sampling temperature (1664 oC). At this temperature,the sample contains magnesiochromite spinels (line 2). Theamount of the spinel mass contents at sampling temperatureis around 15%. In addition, the slag contains a smallamount of calcium chromite phase (CaO·Cr2O3) at thistemperature (mass contents of around 3%).

Fig. 3 shows that the crystallization of dicalcium silicatephase initializes at around 1600 oC. However, the authorsof [8] reported that, at higher basicities (in this case higherthan 1.55), the solid part of the slag can contain dicalciumsilicate precipitates as well.

Thermodynamic calculations. The thermo-Calc soft-ware package [14] was used as to calculate the most stablephase assemblage at different compositions and tempera-tures of the slags. More specifically, the version of this toolwas used which is equipped with TCMSI1 (thermo-calcmetal slag interaction) database [15]. In the calculations,the database includes thermodynamic data regarding steel-making slags equilibriums and metal/slag interactions.

In order to use this tool for determining the equilibriums,it is assumed that the EAF slag is homogeneous. Moreover,it is assumed that the phases within the slag are in equilib-rium with each other. In addition, the slag temperature is

Table 1: Averaged chemical composition of the slag samples (mass contents in%), averaged temperature (in oC) and the basicity (CaO/SiO2) [8]SiO2 Cr2O3 Al2O3 CaO MgO TiO2 MnO FeO basicity

(averaged)

temperature

(averaged)

29.8 14.7 1.9 42.5 3.4 1.1 3.8 2.5 1.44 1664

Fig. 3: Mass contents of the phases in % in the slag samples(averaged composition) versus temperature [8]

Fig. 4: Amount of precipitated phases and total amount ofsolid particles versus mass contents of Cr2O3- and CaO

reviewed paper from R&D

9 (2011) Process technology 357

considered to be 50 oC higher than the measured steel tem-perature. This assumption is based on the personal commu-nication with engineers working in the EAF meltshop [7].

Results and discussions

As discussed and referenced in the introduction, the totalamount of precipitated phases within the slag has a large ef-fect on the slags´ physical properties, such as, the viscosity.Viscosity, in turn, has a considerable effect on thefoamability of the slag [4]. In this context, the influence ofsome of the different process parameters on the total vol-ume of the precipitates was investigated. More specifically,the effects of the chromium- and calcium oxide contentswere studied. Moreover, the impacts of the slag basicity andtemperature on the precipitates amount were observed. Theresults are presented and discussed in the following sec-tions.

Total amount of solid particles vs. Cr2O3 content. Infig. 4, the lower X-axis shows the mass contents of thechromium oxide content of the slag. The correspondingCaO content is presented on the upper X-axis. The diagramis plotted for the averaged composition presented in table 1.The basicity of the slag is assumed to be equal to 1.6. Thisassumption is based on the fact that at this basicity thesample also contains solid dicalcium silicate. In calculatingthe diagram, it is assumed that the composition (except forCaO, SiO2 and Cr2O3) and the basicity of the slag are con-stant. Thus, it is obvious that an increase in the Cr2O3 con-tent corresponds to a decrease (non-linear) in the CaO con-tent. One should consider that this can also be the case inpractice. This is due to the limited ranges of the reasonablebasicities and slag compositions for production conditions.

Fig. 4 also indicates that as the mass content of Cr2O3

rises from zero to higher values, the precipitation of theMgCr2O4 spinels is gradually increased (line 2). Simulta-neously, the CaO content of the slag is reduced. This leadsto a slight decrease in the amount of the precipitated dical-cium silicates (2CaO·SiO2) (line 1). On the whole, the totalamount of precipitated phases (line 4) continuously in-creases during this period. The precipitation of calciumchromite (CaO·Cr2O3) phase initializes when the masscontent of Cr2O3 becomes higher than about 0.14 (line 3).As the volume of the precipitated calcium chromites (line3) is increased, the amount of the solid dicalcium silicateconsiderably drops (line 1). This happens simultaneously toa decrease in the formation rate of the magnesiochromitespinels (line 2). These changes lead to a fall in the totalamount of solid particles (line 4) within the slag at masscontents of Cr2O3 and CaO of about 0.14 and 0.44, respec-tively. In other words, the total solid particles curve (line 4)has a local maximum at these values. As the mass contentsof Cr2O3 exceeds 0.19, the dicalcium silicate phase (line 1)totally dissolves in the slag. Thereafter, solid particleswithin the slag only consist of calcium chromite particles(line 3) and magnesiochromite spinels (line 2). At thisstage, the total solid particles´ amount begins to rise again.This results in a local minimum on the total solid particlescurve at a Cr2O3 mass content of about 0.19 and of 0.415for CaO, respectively. Afterwards, the overall amount of

the solid precipitates (line 4) rises continuously with a fur-ther increase in the mass contents of Cr2O3 and a decreasein CaO mass content. One should take into account that thedrawn diagram represents the changes in the total amountof solid particles for a specific case of slag composition,temperature and basicities. Depending on these factors, thisdiagram may look different for other cases. In the follow-ing, the influence of basicity and temperature on the re-markable trend of changes in overall solid particles´amount is discussed.

Influence of temperature on the total amount of solidparticles. Fig. 5 illustrates the amount of the total solidprecipitates versus chromium oxide content of the slag. Theslag composition is an averaged composition of the samples(table 1) and the basicity of the slag has the fixed value of1.44 (table 1). The diagram is plotted for four different slagtemperatures (line 1-4). Line 3 (T = 1937 K) represents anaverage of the real temperature of the samples (table 1).

The maximum and minimum points which were dis-cussed in the previous section (total solid particles vs.Cr2O3 content) only occur on lines 1 and 2. At higher tem-peratures (lines 3 and 4), the total amount of precipitatedphases continuously rises as the Cr2O3 content is increased.This is due to the fact that the solubility of the dicalciumsilicate in the slag is raised with increased slag temperature.In other words, dicalcium silicate does not form at highertemperatures (lines 3 and 4). This is in agreement with datareported earlier [8].

Influence of basicity on the total amount of solid par-ticles. Fig. 6 shows a diagram illustrating the total amountof solid precipitates versus the chromium oxide content ofthe slag at different slag basicities, namely, 1.44, 1.60 and1.70. The slag temperature is assumed to be constant andequal to 1937 K. In addition, the slag composition corre-sponds to an averaged composition illustrated in table 1. Ascan be seen in fig. 6, the total amount of solid particles doesnot show any extremum when the slag basicity equals 1.44

Fig. 5: Total solid particles amount versus mass contents ofCr2O3 at different slag temperatures

from R&D reviewed paper

9 (2011) Process technology358

(line 1). In contrast, the curves show maximum and mini-mum points at higher basicity (lines 2 and 3). This is due tothe precipitation of dicalcium silicate at these basicities(1.60 and 1.70). Considering line 3 reveals that at thechromium oxide mass content of about 0.14, the slag maycontain solid-phase mass contents of more than 25 %. Onthe contrary, at the same chromium oxide content, the slagwith a lower basicity of 1.44 does not contain mass con-tents of solid particles of more than 10 % (line 1). The dia-gram also indicates that when the chromium oxide masscontent exceeds about 0.20, lines 1 and 2 almost coincide.Moreover, at higher mass contents of Cr2O3 (around 0.25),all three curves coincide. As discussed in the previous sec-tion, an increased Cr2O3 content corresponds to a decreasedCaO content. This is due to the constancy of slag basicityand slag composition (except for CaO, SiO2 and Cr2O3) inthe calculations. In addition, it was shown that, dependenton the slag temperature, lower CaO contents lead to less oreven no precipitation of dicalcium silicate [8]. Thus, in thiscondition, the solid part of the slag only consists ofchromite phases. On the other hand, it was shown in [6]that in the operating range of Cr2O3 content, the slag ba-sicity has almost no influence on the volume of magnesio-chromite spinels in the slag. This could be an explanationfor the observed coincidence (fig. 6).

Concluding discussion. Several studies have alreadybeen carried out to investigate the impact of second-phaseparticles on the foamability of slags. Ito and Fruehan [2; 3]experimentally showed that the foaming index of the CaO-FeO-SiO2 slag system is improved by an increased amountof second-phase precipitates. In addition, it was shown thatfoam stability is also enhanced as a result of the existenceof solid particles within the slag.

Some different theories regarding the role of solid parti-cles in improving the foamability of slags are reported. Ac-cording to Ito and Fruehan [3], the presence of solid parti-cles reduces the rate of the drainage of the films that sepa-

rate the bubbles within the foam. It was also expressed thatthe precipitated particles can serve as nucleation sites forgas bubbles. This clearly contributes to the foamability ofthe slag by increasing the number of small gas bubbleswithin the slag [12; 13]. Finally, additions of solid particlesto a solution were found to lead to an increased apparent(bulk) viscosity of the solution [16]. According to equation(1) [17], an increased viscosity of the slag directly resultsinto an enhanced foaming index:

C

D

µΣρ

= (1)

where � is the foaming index, � is the bulk viscosity, � isthe density of the slag and D is the bubble diameter. Con-stant C is dependent on the nature of the slag system.

Equation (1) correlates the foamability of the slag to itsphysical properties and the average size of the foaming gasbubbles, by a dimensional analysis. However, some studiesshowed that the viscosity is probably the most importantparameter affecting the foaming index of the slag [9; 18].The authors of [19] quantified the above mentioned physi-cal properties for a high-chromium stainless steelmakingslag. They showed that the bulk viscosity of a slag is a de-cisive physical parameter which influences its foamability.

As discussed above, it is generally accepted that the ex-istence of the solid particles within a slag system has a largeeffect on foaming characteristics. This indicates that theknowledge of the total amount of these particles is of ab-solutely high importance for evaluating both, the apparentviscosity and the foamability of the slag. In this regard, thetotal amount of solid particles at different conditions of ba-sicity, temperature and chromium-oxide content is pre-dicted.

In [7; 8] it was shown that, depending on the chromiumoxide content, basicity and temperature at different stagesof EAF high-chromium stainless steelmaking, the solidparticles may be different. More specifically, the slag con-tains magnesiochromite (MgCr2O4) spinels at all processstages. However, before FeSi injection, the slag may alsocontain calcium chromite (CaO·Cr2O3) precipitates. In ad-dition, at higher basicity around 1.5 or more, dicalcium sili-cate (2CaO·SiO2) particles can also form within the slag [7;8].

Conclusions

The present paper focuses the total amount of solid parti-cle as a parameter affecting the viscosity and the foamingproperties of slag. Here, the amount of these particles isquantified for different process conditions. In addition, theeffect of the process conditions on the total amount of theparticles is studied. More specifically, the influence of thecontents of chromium-oxide (Cr2O3) and calcium oxide(CaO), as well as of basicity and temperature on the totalamount of solid particles is also visualized and discussed.

One should consider that the process parameters may de-pend on each other and can, sometimes, interact with eachother. In a specific parameter study case, changes in theslag´s chromium oxide content at constant basicity lead tochanges in the calcium oxide content. This interaction is

Fig. 6: Total solid particles amount versus mass contents ofCr2O3 at different slag basicities

reviewed paper from R&D

9 (2011) Process technology 359

also taken into account in the performed parameter study.The most important conclusions from this study may besummarized as follows:⇒ at a constant basicity and temperature (higher basicities

and lower temperatures which can lead to the formationof dicalcium silicate phase, for instance; basicity = 1.6and T = 1664 oC), an increased chromium oxide contentof the slag results in an increased total amount of solidparticles (chromium containing complexes and dical-cium silicate phase). However, the amount of solid par-ticles initially has a maximum (at mass contents aroundCr2O3 = 0.14 and CaO = 0.44), but further increase ofthe chromium oxide content (simultaneously with a de-crease in the CaO content) results in a decreased totalamount of solid particles. This is due to the dissolutionof dicalcium silicate phase into the slag as the CaOcontent decreases;

⇒ at lower basicities and higher temperatures where nosolid dicalcium silicate exists in the slag (for instance;basicity = 1.44 and T = 1564 oC), an increased chro-mium oxide content directly results in an increased totalamount of solid precipitates (solid chromium containingcomplexes);

⇒ the total amount of solid particles is generally influ-enced by the chromium oxide content, basicity and slagtemperature. An increased chromium oxide contentgives rise to the precipitation of solid chromium con-taining complexes. In addition, the amount of dicalciumsilicate precipitates is a function of the slag basicity andtemperature. More specifically, a higher basicity andlower temperature of the slag contribute to the precipi-tation of dicalcium silicate.

Acknowledgements

The authors wish to thank the Swedish Energy Agency(Energimyndigheten) and the Swedish Steel Producers As-

sociation (Jernkontoret, Project No.: JK23028) for financialsupport. The authors are also grateful to the Thermo-Calcsoftware AB and especially to Dr. Malin Selleby of KTHfor her assistance in performing thermodynamic calcula-tions.

References

[1] Vidacak, B. et al.: Scand. Journ. Met. 31 (2002) No. 5, p. 321/27.[2] Ito, K.; Fruehan, R. J.: Metallurg. Mater. Trans. 20B (1989) No. 4, p.

509/14.[3] Ito, K.; Fruehan, R. J.: Iron & Steelmaker 16 (1989) No. 8, p. 55/60.[4] Kerr, J. J.; Fruehan, R. J.: Iron & Steelmaker 29 (2002) No. 4, p.

45/52.[5] Durinck, D.: Steel Research Int. 78 (2007) No. 2, p.125/35.[6] Mostafaee, S.; Andersson, M.; Jönsson, P. G.: Ironmak. Steelmak. 37

(2010) No. 6, p. 425/36.[7] Mostafaee, S.; Andersson, M.; Jönsson, P. G.: Ironmak. Steelmak. 38

(2011) No. 2, p. 90/100.[8] Mostafaee, S.; Andersson, M.; Jönsson, P. G.: Computational ther-

modynamics as a tool to study the microstructural evolution of EAFduplex stainless steelmaking slags, Ironmak. Steelmak. to be pub-lished.

[9] Lee, Y. E.; Gornerup, M.; Klevan, O. S.: 57th Electric Furnace Conf.Proc., 1999, Pittsburgh, PA, USA, p. 413/24.

[10] Ito, K; Fruehan, R. J.: Metallurg. Mater. Trans. 20B (1989) No. 4, p.515/21.

[11] Gornerup, M.; Jacobsson, H.: Electric Furnace Conf. Proc., 1997,Chicago, IL, USA, p. 57/69.

[12] Pretorius, E. B.; Nunnington, R. C.: Ironmak. Steelmak. 29 (2002)No. 2, p.133/39.

[13] Pretorius, E. B.; Carlisle, R. C.: Iron Steelmaker 26 (1999) No. 10, p.79/88.

[14] Andersson J. O. et al.: CALPHAD 26 (2002) No. 2, p. 372/12.[15] Database provided by Thermo-Calc Software AB, Stockholm, Swe-

den, 2009.[16] Happel, J: Journ. Appl. Phys. 28 (1957) No. 11, p. 1288/92.[17] Kapilashrami, A.: Metallurg. Mater. Trans. 37B (2006) No. 1, p.

109/17.[18] Zhang, Y.; Fruehan, R. J.: Metallurg. Mater. Trans. 26B (1995) No. 4,

p. 803/12.[19] Mostafaee, S.: A study of EAF high-chromium stainless steelmaking

slags characteristics and foamability, Stockholm, 2011 (doctoral thesis).

Saman Mostafaee, PhD Professor Pär G. Jönsson

during this investigation: Licentiate of Engineering, with KTH

now with: Process Development Division Division of Applied Process Metallurgy

Ovako Hofors AB Steel Mill KTH Royal Institute of Technology

Hofors, Sweden Stockholm, Sweden