AISTech 2019 — Proceedings of the Iron & Steel Technology Conference

6–9 May 2019, Pittsburgh, Pa., USA

DOI 10.1000.377.124

Optimization of Ca Additions for Inclusion Modification

Y. Kacar1, B.A. Webler2, P.C. Pistorius2
1
ATI Flat Rolled Products
1300 Pacific Ave, Natrona Heights, PA, 15065
2
Center for Iron and Steelmaking Research, Carnegie Mellon University
5000 Forbes Ave, Pittsburgh, PA, USA, 15213
Email: [email protected] (corresponding author)
Phone: 412-268 2645

Keywords: inclusion, inclusion modification, Ca treatment, total oxygen

INTRODUCTION

Calcium treatment helps to convert solid alumina and spinel inclusions to low melting point calcium aluminates
that are liquid in steelmaking temperatures and remain spherical in morphology during rolling conditions. When
calcium is added to liquid steel, it can modify oxide inclusions, desulfurize steel, and control the shape of sulfide
inclusions. Total oxygen plays a crucial role in inclusion modification. Whereas higher total oxygen in steel can
cause partially modified solid calcium aluminate inclusions, lower total oxygen promotes CaS inclusions (for
a Ca given Ca addition). A large amount of CaS inclusions can worsen the castability of steel and cause erosion
of the ladle slide gate plate and stopper rod tip. Therefore, Ca additions should be optimized based on
steel composition, particularly total oxygen content of steel.

BACKGROUND

When calcium is added into liquid steel, it will react with oxygen and sulfur and modify oxide and
sulfide inclusions [1, 2, 3]. If the steel has a low sulfur concentration, alumina and spinel inclusions are
apparently directly modified by injected calcium (although there are likely intermediate reaction steps). The net
effect is that some aluminum (and magnesium) from the alumina (and spinel) inclusions are returned to the steel
and the CaO forms a reaction product with alumina (and magnesium oxide) (reaction 1). If steel has a
higher sulfur concentration, desulfurization (reaction 2) can be significant and can compete with the oxide
modifying reaction. The critical question is whether calcium added to steel will react with sulfur by reaction
(2) and form CaS or convert Al2O3 inclusions to liquid calcium aluminates (reaction 1). The degree of
modification of Al2O3 inclusions by calcium treatment depends on the competitive reaction between O and S
with Ca in liquid steel [4, 5, 6].
[Ca] + (x+1/3)Al2O3 = (CaO·xAl2O3)+ 2/3[Al] (1)
[Ca] + [S] = CaS (2)

Insufficient calcium addition results in partially modified solid calcium aluminate inclusions which have high
melting points. These are detrimental for castability of steel and formability in final product. On the other hand,
excess calcium addition typically causes CaS formation; these may also decrease castability and may have
adverse effects on the physical and mechanical properties of the final product [3, 7, 8].

The preferred level of addition of calcium depends on the total oxygen content of the steel. In aluminum-killed
steel (with low dissolved oxygen), most of the oxygen is present as oxide inclusions. Modification of alumina
and spinel into liquid inclusions requires an appropriate Ca:Al ratio in the inclusions after treatment. This implies

© 2019 by the Association for Iron & Steel Technology. 1225

a required ratio of added calcium to total oxygen. To illustrate this point, equilibria after calcium modification
were calculated with FactSage 7.2 (using the databases FToxid, FTmisc and FactPS) [9]. Recent kinetic analyses
showed that the associate model for liquid steel in the FactSage FTmisc database greatly overestimates the
concentration of dissolved calcium in liquid steel [10]. This was remedied by disabling the Ca*O associates in
the liquid-steel solution model.

Fig. 1 shows calculation results for two different levels of total oxygen (TO); 15 part per million (ppm) by mass,
and 30 ppm. The requirement for a larger calcium addition to the steel with a higher TO is evident. Replotting
the results in terms of the ratio of Ca:TO (Fig. 2) shows that fully liquid inclusions are obtained for
approximately the same Ca:TO ratios (approximately 0.5-0.8 in this case). In all cases, the calculations predicted
a dissolved oxygen concentration of approximately 3 ppm, both before and after calcium treatment.

In these examples, a "liquid window" (fully modified inclusions) is obtained for Ca:TO ratios between ~0.5 to
0.8. Any Ca amount outside the liquid window would increase the proportion of solid inclusions. Excess Ca in
steel typically leads to solid CaS inclusions, whereas insufficient Ca yields partially modified solid calcium
aluminate inclusions (CA6, CA2, and unmodified spinel).

1
Liquid fr action in inclusions

0.8

0.6

0.4

0.2 TO = 15 ppm
TO = 30 ppm
0
0 10 20 30 40
Total Ca (ppm )

Fig. 1. The effect of total oxygen content on inclusion modification by Ca-treatment (1550 oC). (Shaded regions
show the "liquid windows": Ca addition ranges that yield fully liquid inclusions).
(C: 0.25%-wt, Mn: 1.05%-wt, Si: 0.25%-wt Al: 0.0250%-wt, S: 0.0050%-wt, Mg: 3 ppm-wt)

Compared with the large effect of differences in total oxygen, the effects on the liquid window of
other compositional variables is small, for variations within the ranges encountered in the plant trials.
This is illustrated by Fig. 3. As expected, CaS formation is promoted by higher concentrations of dissolved
aluminum and sulfur in the steel, because of reaction (3). With a higher total Mg concentration in the steel
(largely as MgO in spinel before Ca treatment) initial liquid formation is enhanced slightly (because spinel can
coexist with CaO-Al2O3-MgO liquid [11]). CaS formation is promoted slightly by the higher Mg concentration,
apparently because MgO increases the CaO activity in the CaO-MgO-Al2O3 liquid (modified inclusion).

(CaO) + (2/3)[Al] + [S] = <CaS> + (1/3)(Al2O3) (3)

1226 © 2019 by the Association for Iron & Steel Technology.

 1

Liquid fr action in inclusions

0.8

0.6

0.4

0.2 TO = 15 ppm
TO = 30 ppm
0
0 0.2 0.4 0.6 0.8 1 1.2
Ca/TO r atio

Fig. 2. The effect of Ca/TO ratio on inclusion modification (1550oC; replotted version of Fig.1)

1
(a)
Liquid fr action in inclusions

0.8

0.6

0.4

50 ppm S
0.2
80 ppm S
0
0 10 20 30 40
Total Ca (ppm )

1
(c)
Liquid fr action in inclusions

0.8

0.6

0.4

0.2 3 ppm Mg
5 ppm Mg
0
0 10 20 30 40
Total Ca (ppm )

Fig. 3. The effect of (a) [ppm S], (b) [%Al] and (c) [ppm Mg] on inclusion modification (1550oC).
Steel composition is based on that shown below, with changes as indicated on each graph.
(C: 0.25%-wt, Mn: 1.05%-wt, Si: 0.25%-wt, Al: 0.0250%-wt, S: 0.0050%-wt, T.O: 15 ppm-wt, Mg: 3 ppm-wt)

TRIAL CONDITIONS AND SAMPLE ANALYSIS

Plant trials were conducted at a CSIR member company. The process route was EAF-LF-CC during the
trials. Ladle and tundish capacities are 97 ton and 14 ton respectively. The aim of the trials was to test
whether the calcium addition could be reduced without affecting castability. This aim was based on
previous inclusion analyses that showed the presence of CaS, indicating possible over-treating.

© 2019 by the Association for Iron & Steel Technology. 1227

During the optimization trials, similar ladle practices were applied for each heat. Best desulfurization in ladle is
achieved with CaO-saturated, low FeO+MnO (< 2%) slags [12]. Accordingly, the ladle slag was deoxidized and
adjusted at the beginning of the ladle process after 3 minutes of heating (Table 1).

Table 1. Ladle slag composition before Ca treatment (%-wt)

CaO Al2O3 SiO2 MgO FeO MnO P2O5
55.00 28.00 8.00 6.00 0.80 0.40 0.50

The ladle was first stirred with a higher flow rate (0.5-0.6 Nm3/min) of argon through a porous plug to
desulfurize steel. Then steel chemistry was adjusted with additional alloys, if needed. FeTi (wire) was added 6
min. before CaSi injection. Alloying order and sampling times are given at Fig. 4. The 2nd steel sample was taken
3 minutes before CaSi injection to determine the pre-existing inclusion chemistry and total oxygen amount in
steel. Three minutes of bottom stirring with low intensity (0.1 Nm3/min Ar) was applied for each heat after
calcium treatment, then the 3rd steel sample was taken. Tundish steel samples were compared in terms of
inclusion chemistry and total oxygen amount to optimize Ca additions.

Slag
Addition

FeSiMn Former+
FeMn FeCr Al Lump Al cone FeTi CaSi
Sampling

1st 2nd 3th 4th

Tapping Ladle Treatment Tundish

Fig. 4. Steelmaking process flow during the plant trials, indicating steel sampling times.

Steel composition during the trials is given in Table 2; the steel grade is a medium carbon, Al killed,
Si containing Ca treated steel. Trials were performed in which the total calcium additions in the steel (measured
in the tundish) was decreased by approximately 20% and 50%, by decreasing the CaSi addition. These are
referrred to as "20% reduced" and "50% reduced" calcium heats in the rest of this paper. Other heats that were
examined were one with a higher total Al concentration (0.058%, compared with the typical 0.025%) and one
with higher S (74 ppm; other heats were around 50 ppm).

Lollipop samples were prepared metallographically and investigated with automated scanning electron
microscopy (FEI ASPEX Explorer) at 10 kV accelerating voltage, with other parameters as described in detail
elsewhere [13]. Total oxygen values were estimated based on the analyzed concentrations of Al, Ca, Mg, S, Mn
in the inclusions. The inclusion analyses are presented in ternary diagrams (mole fractions), with compositions
normalized to exclude elements not shown in that diagram. Symbol sizes are scaled proportional to the area
fraction of inclusions with a particular composition. Thin grey lines in the ternary plots indicate the approximate
boundaries of compositions that are at least 50% liquid at 1550 °C.

1228 © 2019 by the Association for Iron & Steel Technology.

 Table 2. Tundish steel composition during the plant trials (%-wt)

Trial Condition # of heats [%C] [%Mn] [%Al] [%Si] [ppm S] [ppm Ca] [ppm Otot]
Baseline 6 0.25 0.72 0.026 0.25 42 28 24
CaSi Reduced 20% 3 0.25 1.08 0.025 0.25 50 22 20
CaSi Reduced 50% 3 0.25 1.05 0.024 0.24 54 14 18
Higher Al 3 0.21 0.69 0.058 0.22 51 18 22
Higher S 3 0.2 0.7 0.027 0.2 74 19 28

RESULTS AND DISCUSSIONS

Mainly spinel inclusions were present in the steel before CaSi injection (Fig. 5). The total oxygen level
and inclusion area fraction were reasonably low, 24 ppm and 82 ppm respectively.

Area Fraction: 82 ppm

36 inclusions per mm2
Ototal: 24 ppm

Fig. 5. Inclusion chemistry in steel before Ca treatment

In tundish samples (Fig. 6), CaS and Ca-rich calcium aluminate inclusions were present. Although [S] in steel is
reasonably low (42 ppm), a large proportion of CaS formation can be explained by a high Ca/Ototal ratio
(approximately 1.6), far larger than needed to transform just the oxide inclusions. Reflecting inclusion removal,
the total oxygen is lower than it was before calcium treatment.

Area Fraction: 132

43 inclusions per m
Ototal: 18 ppm
Ca/Ototal: 1.56

Fig. 6. Baseline heat: inclusions in tundish sample composition (Catot: 28 ppm). The prominent CaS population
is circled.

© 2019 by the Association for Iron & Steel Technology. 1229

For heats in which the total calcium was ~20% lower, a slight decrease in the CaS inclusion population was
observed (Fig. 7), but the spread in inclusion compositions remained similar to the baseline; the Ca addition
remained higher than required for oxide modification, with Ca/Ototal ≈ 1,2.

Area Fraction: 112 p

38 inclusions per mm
Ototal: 18 ppm
Ca/Ototal: 1.22

Fig. 7. Heat with 20% less Ca: tundish sample inclusions composition (Catot: 22 ppm)

In heats with 50% less calcium (in the tundish samples), the inclusion chemistry was significantly different, with
much less CaS in the inclusions, and adequate modification: most inclusion compositions were in the >50%
liquid region, with no alumina or spinel inclusions remaining (Fig. 8).

Area Fraction: 71 ppm

21 inclusions per mm2
Ototal: 15 ppm
Ca/Ototal: 0.93

Fig. 8. Heat with 50% less Ca: tundish sample inclusion composition (Catot: 14 ppm)

In the higher-Al heats, some alumina-rich calcium aluminates and a small proportion of spinel inclusions were
found, but most inclusions were in the >50% liquid region (Fig. 9). There were no CaS inclusions (the total Ca
was 18 ppm), indicating a higher total oxygen. The higher total oxygen likely reflects some reoxidation, as also
indicated by the presence of spinels.

1230 © 2019 by the Association for Iron & Steel Technology.

 Area Fraction: 106
29 inclusions per m
Ototal: 28 ppm
Ca/Ototal: 0.64

Fig. 9. Higher-Al heat: inclusions in tundish sample (Catot: 18 ppm, [Al]: 0.0579%)

In the higher-S heat, CaS and Ca-rich calcium aluminate inclusions were prevalent, without spinels or
alumina inclusions. The distribution of inclusion compositions was similar to that of the heat with 20% less
Ca (Fig. 7, which yielded a similar Ca/Otot ratio and similar total oxygen).

Area Fraction: 78 ppm

Intensity: 28 inc./mm2
Ototal: 15 ppm
Ca/Ototal: 1.27

Fig. 10. Higher-S heat: inclusions in tundish sample (Catot: 19 ppm, [S]: 74 ppm)

CONCLUSIONS

Thermodynamic investigations and plant trials were conducted to optimize Ca additions for inclusion
modification in medium carbon Al killed, Si containing steels. According to these results;

- The total oxygen content of the steel steel has a critical role in the success of inclusion modification:
ideally, the Ca additions should be tailored to the expected total oxygen content.

- The Ca/Ototal mass ratio that is required for inclusion modification is approximately 0.5 to 0.8. Higher
Ca/Ototal ratios lead to CaS formation, whereas alumina or spinel inclusions would not be fully modified
at lower Ca/Ototal ratios.

© 2019 by the Association for Iron & Steel Technology. 1231

 - In the plant trials, the heats with 50% less Ca than the baseline contained successfully modified (largely
liquid) inclusions.

- For the range of compositions studied, the effects of higher Al or higher S concentrations were
secondary in importance to differences in total oxygen content – as shown by calculations, and by plant
samples.

ACKNOWLEDGEMENTS

Support of this work by industrial members of the Center for Iron & Steelmaking Research is
gratefully acknowledged. The authors would like to acknowledge use of the Materials Characterization
Facility at Carnegie Mellon University under grant # MCF-677785

REFERENCES

1. Fruehan, R.J., Ladle Metallurgy Principles and Practices. Iron and Steel Society, Warrendale, PA, 1985.
2. Turkdogan, E.T., Metallurgical Consequences of Calcium Retention in Liquid and Solid Steel, 1st
International Calcium Treatment Symposium, pp. 3-13, 1988.
3. Holappa, L., Hamalainen, M., Liukkonen, M., Lind, M., Thermodynamic Examination of Inclusion
Modification and Precipitation from Calcium Treatment to Solidified Steel, Clean Steel 6 Conference, 12
August 2002.
4. Kor, G.J.W., Glaws, P.C., Ladle Refining and Vacuum Degassing. In: The Making, Shaping and
Treating of Steel, Steelmaking and Refining Volume, 11th Edition, Chapter 11. AISE Steel Foundation,
Pittsburgh, 1998.
5. Lind, M. and Holappa, L., Transformation of Alumina Inclusions by Calcium Treatment, Metallurgical
and Materials Transactions B, vol. 41B, 2010, pp. 359-366.
6. Turkdogan, E.T., Fruehan, R.J., Fundamentals of Iron and Steelmaking. In: The Making, Shaping and
Treating of Steel, Steelmaking and Refining Volume, 11th Edition, Chapter 2. AISE Steel Foundation,
Pittsburgh, 1998.
7. Choudhary, S.K., Gosh A., Thermodynamic Evaluation of Formation of Oxide–Sulfide Duplex
Inclusions in Steel, ISIJ International, vol. 48, 2008, pp. 1552-1559 8
8. Ototani, T., Calcium Clean Steel, Springer, New York, 1986.
9. Bale, C.W., Bélisle, E., Chartrand, P., Decterov, S.A., Eriksson, G., Gheribi, A.E., Hack, K., Jung, I.-H.
Kang, Y.-B. Melançon, J., Pelton, A.D., Petersen, S., Robelin, C., Sangster, J., Spencer, P., Van Ende,
M-A., FactSage thermochemical software and databases, 2010-2016, CALPHAD, vol. 54, 2016, pp. 35-
53.
10. Kumar, D., Pistorius, P.C., AISTech 2016 Proceedings, Association for Iron and Steel Technology,
Warrendale, 2016, pp. 1151–59.
11. Verma, N., Pistorius, P.C., Fruehan, R.J., Potter, M.S., Oltmann, H.G., Pretorius, E.B., Calcium
Modification of Spinel Inclusions in Aluminum-Killed Steel: Reaction Steps. Metallurgical and
Materials Transactions B, vol. 43B, 2012, pp. 830-840.
12. Holappa, L., Kacar Y., Slag Formation - Thermodynamic and kinetic aspects and mechanisms. In:
Reddy R.G., Chaubal P., Pistorius P.C., Pal U. (eds) Advances in Molten Slags, Fluxes, and Salts:
Proceedings of the 10th International Conference on Molten Slags, Fluxes and Salts 2016, pp 1017-
1024. Springer, Cham
13. Tang, D., Ferreira, M.E., Pistorius, P.C., Automated inclusion microanalysis in steel by computer-based
scanning electron microscopy: accelerating voltage, backscattered electron image quality, and analysis
time. Microscopy and Microanalysis, vol. 23, 2017, pp. 1082-1090.

1232 © 2019 by the Association for Iron & Steel Technology.