ISIJ International, Vol. 55 (2015), No. 1, pp.

126–133

Inclusion Evolution after Calcium Addition in Low Carbon Al-Killed

Steel with Ultra Low Sulfur Content

Guangwei YANG* and Xinhua WANG

School of Metallurgical and Ecological Engineering, University of Science and Technology Beijing, Beijing, 100083 China.
(Received on June 24, 2014; accepted on September 11, 2014)

Inclusions in steel samples collected 0.5, 2, 5, 10 and 20 min after calcium addition were investigated
in Low Carbon Aluminum Killed steel (LCAK steel) with ultra low sulfur content (0.04–0.065% Al, 9 ppm
S). It’s found that inclusions change from CaO to CaO–CaS and finally to CaO–Al2O3–CaS with time. Al2O3
in inclusions linearly decreases by increasing T.Ca/T.O of the steel and disappears when T.Ca/T.O exceeds
3. CaS/CaO of the inclusions linearly increases by increasing S/T.O of the steel. Contents of CaO, Al2O3
and CaS can be estimated through T.Ca, S and T.O. CaO–CaS clusters are found at 5 min. The mechanism
of inclusion evolution and formation of CaO–CaS clusters are discussed.

KEY WORDS: CaO–CaS; CaO–Al2O3–CaS; clusters; calcium addition; LCAK Steel.

In present study, inclusions in steel samples collected 0.5,

1. Introduction
2, 5, 10 and 20 min after calcium addition were investigated
Al2O3 inclusions are considered to be harmful to steelmak- in LCAK steel with ultra low sulfur content (0.04–0.065%
ing process and final product.1–4) Proper calcium treatment Al, 9 ppm S). Contents of CaO, CaS and Al2O3 are easily
can modify Al2O3 into low melting CaO–Al2O3, through determined through T.Ca, T.O and S content in steel. The
which clogging can be avoided. Excess or insufficient calci- relation between steel and inclusion is obtained through the
um addition can produce solid CaS or CA6, CA25–7) (where fitting of the experimental data. Dissolved calcium is calcu-
C is CaO and A is Al2O3), which are also harmful in terms lated to explain the mechanism of evolution of inclusion.
of submerged entry nozzle (SEN) clogging. Thus, the com- CaO–CaS clusters were found and its formation is explained.
position of inclusion, such as CaO, CaS and Al2O3, should
be well controlled in order to get desired properties of steel.
2. Experimental Methods
There are two methods to predict the inclusion type after
calcium treatment. One is through dissolved aluminum and 2.1. Experimental Procedures
sulfur contents.8–12) However, activities of CaO, CaS and Five experiments were carried out in this study. Induction
Al2O3 are needed in this method. Moreover, controversy
exists in the literature over the choice of such data.9,10) Dif-
ferent choice may lead to different results. Thus this method
is not convenient.
The other method is through T.Ca/T.O of the steel.8,9,13)
For example, if the ratio T.Ca/T.O is just larger than 0.6,
inclusions will be a mixture of solid phase CA and liquid
phase C0.57A0.43 at 1 823 K. In this method, it’s assumed that
all calcium is bound only to oxygen in the inclusions. Only
in this case, T.Ca/T.O ratio would give a realistic picture of
inclusion type. But calcium can also react with dissolved
sulfur to form solid CaS. For this case, T.Ca/T.O ratio could
not well predict inclusions.
Contents of T.Ca, T.O and S are very important in deter-
mining the inclusion type. The above methods, however,
didn’t consider them all. Moreover, they mainly focused on
steel with relatively high sulfur content,11–13) 30–50 ppm.
Most importantly, the two methods can only predict inclu-
sion type but can’t give the composition of inclusions.

* Corresponding author: E-mail: ustb_jack@hotmail.com

DOI: http://dx.doi.org/10.2355/isijinternational.55.126 Fig. 1. Schematic diagram of experimental apparatus.

© 2015 ISIJ 126

 ISIJ International, Vol. 55 (2015), No. 1

vacuum furnace was employed as melting apparatus and its
schematic diagram is shown in Fig. 1. In each experiment,
1 000 g steel material, 0.3 g Al lump and 3 g CaSi (30%Ca-
70%Si) were firstly placed into the furnace, then the chamber
was evacuated with mechanical pumps and then backfilled
with high-purity argon to atmospheric pressure. The Al
lump was placed in the feed opening. CaSi alloy wrapped
with steel was fixed in a molybdenum rod. The chemical
composition of steel material used in the experiments is giv-
en in Table 1. After the steel had melted, the melt was held
at 1 873 K (1 600°C) for 3 minutes for homogenization.
Then Al was added, followed by CaSi addition after 5 min-
utes. After different holding time, the steel was poured into
an iron mold. The holding time was 0.5 min, 2 min, 5 min,
10 min and 20 min. It’s the only difference between the five
experiments. The flowchart of the procedure is shown in
Fig. 2. The time after CaSi addition is set as 0 min.
2.2. Chemical Analysis and Characterization of Inclu-
sions
The total calcium (T.Ca) and acid-soluble aluminum
([Al]) of the steel samples were analyzed by Inductively
Coupled Plasma-Atomic Emission Spectrometry (ICP-AES)
method. Total oxygen (T.O) and sulfur contents of the steel
were determined by infrared absorption method.
ASPEX PSEM EXPLORER with Automated Feature
Analysis system was employed to analyze inclusions,
including chemistry, morphology, number and so on. Inclu-
sions less than 1 μ m were not detected in this study. The
scan area was 20 mm2, hundreds of particles were detected.
Table 1. Chemical compositions of steel material (mass pct).
It should be noted that the inclusion composition that
ASPEX provides is the average composition of the whole
inclusion, similar to average composition when selecting an
area for EDX.
3. Experimental Results
3.1. Steel Composition
The chemical compositions of steel samples are given in
Table 2. Contents of [Al] and T.O are in the range of 0.039
to 0.064 and 0.0002 to 0.0010 pct. S content maintains 9
ppm regardless of holding time, the same as that in the steel
material. T.Ca content decreases from 0.0036 to 0.0011 pct
from 0.5 min to 5 min, and maintains nearly the same after-
wards.
3.2. Characterization of Inclusions
Typical inclusions at 0.5 min and 2 min after calcium
addition are shown in Fig. 3. At 0.5 min, the typical inclu-
sions are mainly pure CaO. The mapping of such inclusion
is shown in Fig. 3(a). Complex inclusions with CaO–Al2O3
inside and pure CaO outside are also found as shown in Fig.
3(b). It indicates Al2O3 is gradually reduced by dissolved
calcium from outside to inside. In the sample collected 2
min after calcium addition, CaO–CaS type inclusions with
CaO partially covered by CaS are observed. The mapping
of such inclusion is shown in Fig. 3(c). In the sample col-
lected 5 min after calcium addition, CaO–CaS type inclu-
sions are still the dominant inclusions. CaO–CaS clusters
are also found in this sample. The morphology and X-ray
spectrums are shown in Figs. 4(a)–4(c). Inclusions change

C Si Mn P S [Al] Table 2. Chemical compositions of steel samples (mass pct).

0.05 0.2 1.5 0.01 0.0009 0.026 Holding time, min [Al] S T.Ca T.O
0.5 0.052 0.0009 0.0036 0.0006
2 0.064 0.0009 0.0031 0.0007
5 0.039 0.0009 0.0011 0.0002
10 0.052 0.0009 0.0015 0.0010
20 0.041 0.0009 0.0014 0.0005
Fig. 2. Flowchart of experimental procedure.

Fig. 3. Typical inclusions at 0.5 min and 2 min after calcium addition.

127 © 2015 ISIJ

 ISIJ International, Vol. 55 (2015), No. 1
Fig. 4. Typical inclusions at 5 min after calcium addition.
Fig. 5. Typical inclusions at (a) 10 min and (b) 20 min after cal-
cium addition.
to spherical CaO–Al2O3–CaS for the sample collected 10
min and 20 min after calcium addition. The morphology
and X-ray spectrums are shown in Fig. 5. Ca peak at 10 min
is higher than that at 20 min, which means CaO content in
10 min sample is more than that in 20 min sample. Verma
et al.14) observed similar evolution in Fe-0.1%Al-0.0007%S
low sulfur steel after calcium addition. They found inclu-
sions in the sample collected 0.5 min after calcium treatment
were mostly CaO. In some cases, CaS was found attached
to CaO inclusions. In the sample collected 2 min after cal-
cium treatment, CaS content increased and CaO decreased.
In the sample collected 4 min after calcium treatment, the
inclusions changed to liquid CaO–Al2O3 with a low CaS
content.
Figure 6 shows the change of average composition of
inclusions. With holding time, CaO content decreases. CaS
content first increases and begins to decrease at 5 min.
Al2O3 content begins to increase at 5 min. MgO gradually
increases with time, but the content is low compared with
© 2015 ISIJ
Fig. 6. Change of inclusion composition with holding time.
others. Evolution of inclusions follows CaO→CaO–
CaS→CaO–Al2O3–CaS.
Figure 7 shows distribution of inclusion composition and
size in CaO–CaS–Al2O3 ternary system. Solid circles, open
triangles and open squares represent inclusions with diame-
ter in between 1–5 μ m, 5–10 μ m and 10–20 μ m, respective-
ly. No inclusions larger than 20 μ m are found in all samples.
The solid line is the 1 873 K liquid line, within which the
inclusions are liquid. For samples collected at 0.5 min, 2
min and 5 min, inclusions are uniformly located in CaO–
CaS line and are small. No inclusions over 10 μ m are found.
For samples collected at 10 min and 20 min, inclusions are
CaO–Al2O3 with different amounts of CaS. Big inclusions
over 10 μ m are found in these two samples, and the com-
position are mainly low melting CaO–Al2O3 with small
amount of CaS.
Figure 8 shows the change of inclusion number with
holding time. The number suddenly picks up from 2.1/mm2
to 9.5/mm2 at 2 min, reaching 10/mm2 at 5 min, then grad-
ually decreases. Similarly, Zinngrebe et al.15) found inclusion
number increased by a large factor after calcium treatment
in Al-killed steel.
4. Discussion
4.1. Mechanism on Composition Change of Inclusion
with Time
In the sample collected 0.5 min after calcium addition, the
main inclusion is pure CaO. There are two sources. Pure
Al2O3 reacts with dissolved calcium and changes into pure
CaO, as expressed by Eq. (1). In addition, dissolved oxygen
reacts with dissolved calcium and becomes pure CaO, as
expressed by Eq. (2). For steel with 0.05% Al, dissolved
oxygen in equilibrium with pure Al2O3 and CaO are approx-
imately 4.7 ppm and 0.6 ppm, as will be calculated later.
Thus (4.7–0.6=4.1) ppm dissolved oxygen and (6–4.1=1.9)
ppm undissolved oxygen (Al2O3) finally become CaO.
3[Ca ] + Al 2 O3( inclusion ) = 3CaO( inclusion ) + 2[ Al]
...... (1)
ΔGθ = −1 068 893 + 215.4 T16,17)
[Ca ] + [O] = (CaO)inclusion
.................... (2)
ΔGθ = −645 421 + 146 T17)
128
 ISIJ International, Vol. 55 (2015), No. 1

Fig. 7. Distribution of inclusion composition and size in CaO–CaS–Al2O3 ternary system.

In order to analyze change of the calcium aluminates with 2[ Al] + 3[O] = Al 2 O3( inclusion )
time, dissolved calcium contents are estimated. T.O in Table .................. (5)
2 includes dissolved oxygen ([O]) in steel and undissolved ΔGθ = −864 370 + 222.5T16 )
oxygen in oxide inclusions, as expressed by Eq. (3). T.Ca in The Al2O3 is either pure or Al2O3 in a calcium aluminate
Table 2 includes dissolved calcium ([Ca]) in steel and undis- inclusion. [O%] is given by Eq. (6).
solved calcium in CaS and CaO, as expressed by Eq. (4). 1
⎛ ⎛ ΔGθ ⎞ a Al2O3 ⎞ 3

T.O = [O] + O(oxide) ........................ (3) [O%] = ⎜ exp ⎜ ⎟ 2 ⎟ ......... (6)

T.Ca = [Ca] + Ca(CaS) + Ca(CaO).............. (4)
In an aluminum killed steel, the dissolved oxygen [O%]
is determined by the equilibrium between [Al] in the steel
and Al2O3 in the inclusions, as shown in Eq. (5).
⎜ ⎝ RT ⎠ f[ Al ] [ Al%] f[O ] ⎟
2 3
⎝ ⎠
Activity of Al2O3 in CaO–Al2O3 binary system was cal-
culated with Thermo-Calc, as shown in Fig. 9. Contents of
dissolved oxygen are obtained assuming activity coeffi-
cients of Al and O as unity, as shown in Fig. 10. The [Al]

129 © 2015 ISIJ

 ISIJ International, Vol. 55 (2015), No. 1

in this figure is 0.05%, average of [Al] contents in Table 2.
As can be seen, [O] is about 0.6 ppm when Al2O3 content
is less than 40%. The average Al2O3 contents in all samples
are around 40%, as can be read from Fig. 6. Thus [O] con-
tents for all samples are 0.6 ppm.
Fig. 8. Change of inclusion number with holding time.
For inclusions in samples collected 0.5 min, 2 min and 5
min after calcium addition, inclusions are mainly CaO and
CaS, with the total contents over 95%. The normalized com-
position is shown in Table 3. The undissolved oxygen, cal-
culated by Eq. (3), is in the form of CaO. Thus undissolved
calcium in CaO is obtained. The undissolved calcium in CaS
is also obtained based on the CaO and CaS contents in inclu-
sions. The dissolved calcium [Ca] finally is calculated by
Eq. (4). The results are shown in Table 3. Figure 11 shows
the relation between the natural logarithm of [Ca%] and
holding time. As can be seen, ln[Ca%] linearly decreases
with time. The relation is expressed as Eq. (7). Assuming
[Ca] at any time follows the same trend, [Ca] contents at 10
min and 20 min are 0.37 ppm and 0.0056 ppm.
ln[Ca%] = –0.42 × T – 6....................... (7)
Activities of [Al] and [Ca] can be calculated by Eq. (8).
Contents of C, Si, Mn and P are shown in Table 1. [Al], [O]
and [Ca] contents are shown in Tables 2 and 3. The first and
second order interaction coefficients are shown in Table 4.
lg ai = lg[i%] + lg fi

= lg[i%] + ∑ eij [ j%] + ∑ ∑ ri( ) [ j%][ k%] ( j ≤ k )

j ,k

j j k
.......................................... (8)
To analyze evolution of calcium aluminates, the stability
phase diagram of various calcium aluminates inclusion was
drawn as shown in Fig. 12. It’s based on Eq. (9). Activities
of CaO and Al2O3 are given in Fig. 9. The solid circles
denote present results. As seen in Fig. 6, the calcium alumi-
nates at 0.5 min, 2 min and 5 min are CaO. The calcium alu-
minate at 10 min are CaO-39%Al2O3, nearly the same with
C3A which contains 38%Al2O3. Thus calcium aluminates at
0.5 min, 2 min, 5 min and 10 min are CaO and C3A, which
is in accordance with the location in Fig. 12. The calcium
Fig. 9. Activity of Al2O3 and CaO in CaO–Al2O3 binary system at
1 873 K. aluminate at 20 min are CaO-43%Al2O3, locating between

Fig. 10. Dissolved oxygen in steel. Fig. 11. Change of [Ca] with time.

Table 3. Total, dissolved and undissolved oxygen and calcium of steel samples (ppm).

Time/min CaO CaS T.Ca T.O [O] O(CaO) Ca(CaO) Ca(CaS) [Ca]
0.5 95 5 36 6 0.6 5.4 13.5 0.6 21.9
2 69 31 31 7 0.6 6.4 16.0 5.5 9.5
5 39 61 11 2 0.6 1.4 3.5 4.3 3.2

© 2015 ISIJ 130

 ISIJ International, Vol. 55 (2015), No. 1

Table 4. First and second order interaction coefficients eij, rij.

j
C Si Mn P S Al Ca O
i
Al 0.091 0.056 0.0065 0.033 0.035 0.043 –0.047 –1.98
Ca –0.34 –0.096 –0.0156 –0.097 –140 –0.072 –0.002 –9 00018)
O –0.421 –0.066 –0.021 0.07 –0.133 –1.17 –3 60018) –0.174
All data are at 1 873 K (1 600°C) and the ones without notation are from Ref. 19).
rOCa = 5.7 × 10 5 ,18 ) rOO ,Ca = 2.9 × 106 ,18 ) rCa
O
= 3.6 × 106 ,18 ) rCa
Ca ,O
= 2.9 × 106 18 )

Fig. 12. Stability phase diagram of various calcium aluminates.

Fig. 13. Change of Gibbs free energies of reactions with dissolved
calcium.

C3A and C12A7 in CaO–Al2O3 binary system, which is also

in accordance with its location in Fig. 12. describe whether one reaction is favored or not. If ΔGr is
negative, this reaction can proceed spontaneously.
3[Ca ] + Al 2 O3( inclusion ) = 3CaO( inclusion ) + 2[ Al]
ΔGr = ΔGθ + RT ln Qr ..................... (12)
ΔGθ = −1 068 893 + 215.4 T16,17)
Qr is the reaction quotient. For example, reaction quotient
⎛ ΔGθ ⎞ aCaO
3
⋅ a 2Al
K1θ873 K = exp ⎜ − ⎟= = 3.75 × 1018 ... (9) of Eq. (11) can be expressed as
⎝ RT ⎠ a Al2O3 ⋅ aCa
3
aCaS
1 Qr =
⎛ 3
aCaO ⎞ 3
2 aCa ⋅ aS
aCa = ⎜ ⎟ ⋅ a Al3
⎜ a Al O ⋅ K1θ873 K ⎟ Activities of [S] and [Ca] can be calculated by Eq. (8).
⎝ 2 3 ⎠
The steel composition used in this calculation is shown in
The formation of CaS is discussed below. There are two Table 5. Contents of C, Si, Mn and P are from Table 1. Con-
sources of CaS. One is reaction product of CaO and sulfur, tent of [Al] is 0.05%, average of [Al] contents in Table 2.
as shown in Fig. 3(c). The reaction can be expressed by Eq. Content of [O] is from Table 3. The calcium aluminates in
(10). The other is product of dissolved calcium and sulfur, all samples are CaO or near C3A (Al2O3 content less than
as expressed by Eq. (11). In fact, Eq. (11) happens at 0.5 min 38%), thus the activities of CaO is unity based on Fig. 9.
after calcium addition, but most of these CaS inclusions are Activities of CaS are taken as unity due to low solubility of
too small (less than 1 μ m) that couldn’t be detected by the CaS in the calcium aluminates.21) The [Ca] contents are
SEM. They grow into bigger ones and can be detected with 21.9, 9.5 and 3.2 ppm for samples collected 0.5, 2 and 5 min
time. It can be verified by the increase of inclusions near the after calcium addition. As can be seen in Fig. 13 that Gibbs
CaS corner of the CaO–CaS–Al2O3 ternary shown in Figs. free energies of Eqs. (10) and (11) are negative, thus CaS
7(a)–7(c). can form, but the driving force of the reactions decreases
with time. Thus CaS content increases as shown in Fig. 6.
(CaO)inclusion + [S] = (CaS)inclusion + [O] Because the formation of CaS by Eq. (10) consumes previ-
......... (10)
ΔGθ = 114 521 − 29.8T17,20)) ous formed CaO, thus CaO content decreases.
For samples collected 10 min after calcium addition, the
[Ca ] + [S] = (CaS)inclusion [Ca] content is 0.37 ppm, Gibbs free energy of Eq. (11) is
................ (11) negative, thus CaS can form. However, the Gibbs free ener-
ΔGθ = −530 900 + 116.2 T 20 )
gy of Eq. (10) is positive, thus the reaction proceeds in the
Figure 13 shows the change of Gibbs free energies of reverse direction. Previous formed CaS through Eq. (10)
reactions with dissolved calcium. Gibbs free energy of reac- changes to CaO, which further changes to Al2O3 through
tion, ΔGr, can be obtained by Eq. (12). ΔGr is often used to Eq. (1) and decomposes to dissolved calcium and oxygen

131 © 2015 ISIJ

 ISIJ International, Vol. 55 (2015), No. 1

Table 5. Chemical compositions of steel for calculation (mass pct).

C Si Mn P S [Al] [Ca] [O]

0.05 0.2 1.5 0.01 0.0009 0.05 variable 0.00006

through Eq. (2) due to the decrease of dissolved calcium.

Thus CaS and CaO decreases, while Al2O3 increases, as
shown in Fig. 6.
For samples collected 20 min after calcium addition, the
[Ca] content is 0.0056 ppm, Gibbs free energies of Eqs. (10)
and (11) are positive, thus CaS can’t form. The two reac-
tions proceeds in the reverse direction. Previous formed CaS
either decomposes into dissolved calcium and sulfur
through Eq. (11) or changes into CaO through Eq. (10).
Fig. 14. Al2O3 in inclusions and T.Ca/T.O in steel.
Also, CaO will change to Al2O3 through Eq. (1) and decom-
pose to dissloved calcium and oxygen through Eq. (2) due
to the decrease of dissolved calcium, thus CaO decreases,
while Al2O3 increases, as shown in Fig. 6. Nevertheless,
CaS content is slightly higher than that at 10 min. It’s mainly
due to the higher S/T.O in this sample. And a higher S/T.O
represents a higher CaS/CaO in inclusion. The relation is
discussed later.

4.2. Relationship between Steel and Inclusions

Figure 14 shows the relation between T.Ca/T.O in the
steel and Al2O3 in CaO–CaS–Al2O3 inclusions. Results from
Numata and Higuchi22,23) are also included in this figure. It
can be seen that Al2O3 decreases with T.Ca/T.O, becoming
trace when T.Ca/T.O is over 3. Inclusions change from
CaO–Al2O3 with small amount of CaS to CaO–CaS. When
T.Ca/T.O is less than 3, a significant linear dependence is
observed between Al2O3 in the inclusions and T.Ca/T.O in Fig. 15. CaS/CaO of inclusions and S/T.O in steel.
steel. The best-fit line equation is expressed by Eq. (13). The
correlation coefficient R2 is nearly unity. Equation (13) is
mainly based on the results by Numata et al. Accordingly, Solid Al2O3 can cause SEN clogging. Calcium treatment
results from Fig. 14 are in reasonable agreement with are always employed to solve this issue. Excess calcium
Numata et al.’s. treatment can produce solid CaS, which can also give rise
to SEN clogging. Insufficient calcium treatment can pro-
When T.Ca / T.O ≤ 3, duce CA6 or CA2, which is more harmful than Al2O3 in
(Al 2 O3 %) = −30 × T.Ca / T.O + 84 R 2 = 0.9669 terms of SEN clogging. Proper calcium treatment can modify
........................................ (13) Al2O3 into low melting CaO–Al2O3, through which clogging
can be avoided. As can be read from Fig. 14, when control-
Figure 15 shows the relation between CaS/CaO of inclu-
ling T.Ca/T.O around 1.0, Al2O3 can be fully modified to
sions and S/T.O in steel. As can be seen, CaS/CaO of inclu-
low melting CaO–Al2O3 with composition near C12A7. Deng
sion linearly increases with S/T.O in steel except the sample
and Zhu13) found good castability is obtained when Ca/T.O
collected 0.5 min after calcium addition. There may be two
is controlled between 0.91 and 1.25, which is in consistent
reasons for the low CaS/CaO of inclusion. One is that 0.5
with present study.
min is too short for the system to reach equilibrium. The
It’s found that type B inclusions, which are disconnected
other is that the formed CaS particles, which are product of
oxides containing Ca, Al and S, can lead to hydrogen
dissolved calcium and sulfur, are too small (less than 1 μ m)
induced cracking. In order to prevent the formation of type
that couldn’t be detected by the SEM, leading to an under-
B inclusions in plate rolling, composition of CaO, Al2O3 and
estimate of CaS. The best-fit line equation is expressed by
CaS in inclusions must be well controlled.24) Through
Eq. (14). The correlation coefficient R2 is nearly unity.
adjusting T.Ca, S and T.O content in steel, the target com-
CaS / CaO = −0.075 + 0.4 × S / T.O R 2 = 0.9985 position of CaO, Al2O3 and CaS can be hit based on Eqs.
(13) and (14).
........................................ (14)
For LCAK steel with Al among 0.04–0.65% and S less 4.3. Formation of CaO–CaS Clusters
than 0.0010%, contents of CaO, Al2O3 and CaS in inclu- At 5 min, many CaO–CaS clusters forms. There are two
sions after calcium treatment can roughly obtained through reasons that possibly explain the formation. The first reason
Eqs. (13) and (14). It’s useful in many ways. is huge number of CaO–CaS type inclusions, and the second

© 2015 ISIJ 132

 ISIJ International, Vol. 55 (2015), No. 1
is the high aggregation tendency of CaS in inclusions.
As seen in Fig. 8, the inclusion number suddenly picks up
at 2 min, reaching the maximum at 5 min, nearly 5 times
that at 0.5 min. At 0.5 min after calcium treatment, Eqs. (1),
(2), (10) and (11) happen. Equations (1) and (10) are dis-
placement reactions, not involving the increase of inclusion
number. Equations (2) and (11) generate lots of tiny CaO
and CaS particles, these inclusions are so small that they are
beyond the scope of the detection limit, 1 μ m at least. At 2
min to 5 min after Ca treatment, previous formed tiny inclu-
sions grow into larger ones through agglomeration and
become detectable, thus the inclusion number increases.
And CaO–CaS clusters form with further agglomeration.
Wang et al.25) investigated effect of CaS formation on
the agglomerate ability of Al2O3–CaO inclusions in low-
carbon and Ca-treated steel through high-temperature con-
focal scanning laser microscopy (CSLM). They found liq-
uid-globular Al2O3–CaO inclusions don’t agglomerate with
each other, but when CaS forms outside Al2O3–CaO, these
inclusions begin to agglomerate with each other, and finally
becomes clusters. It can be inferred CaS has a strong ten-
dency to promote inclusions to agglomerate. In present
study, CaS also exists in inclusions, thus it could be another
reason for the formation of CaO–CaS clusters.
Both Ikeda et al.26) and Nafisi et al.27) found CaO–CaS
clusters in the HIC fracture surface of pipeline steel. It indi-
cates that CaO–CaS clusters are harmful for HIC resistance
property. Thus CaO–CaS clusters should be minimized. From
above discussion, it can be reasonably concluded that the
formation of tiny CaO–CaS particles can be prevented
through lowering dissolved calcium content. And CaO–
Al2O3–CaS type inclusions may be a good choice in terms
of HIC resistance. The composition of CaO, Al2O3 and CaS
can be controlled through adjust steel composition based on
Eqs. (13) and (14).
5. Conclusions
Inclusions in steel samples collected 0.5, 2, 5, 10 and 20
min after calcium addition are investigated in ultra low sul-
fur LCAK steel (0.04–0.065% Al, 9 ppm S). The following
conclusions were obtained:
(1) At 0.5 min after calcium addition, pure CaO inclu-
sions are the main inclusions. They are the reaction product
of dissolved calcium with i) Al2O3 in inclusions and ii) dis-
solved oxygen. With time, CaS content increases and inclu-
sions change to CaO–CaS. CaS content reaches the most at
5 min. They are the reaction product of dissolved sulfur with
i) CaO in inclusions and ii) dissolved calcium. Because dis-
solved calcium content decreases with time, previous
formed CaS changes to CaO or decomposes into dissolved
calcium and sulfur. With further decrease of dissolved cal-
cium, CaO changes to Al2O3 and decomposes to dissloved
calcium and oxygen. And the inclusions change to CaO–
Al2O3–CaS at 10 min after calcium treatment.
(2) Contents of CaO, Al2O3 and CaS of inclusions can
be estimated through T.Ca, S and T.O in the steel based on
following equations.
133
− 30 × T.Ca / T.O + 84 T.Ca / T.O ≤ 3
(Al 2 O3 %) =
0 T.Ca / T.O > 3
CaS / CaO = − 0.075 + 0.4 × S / T.O
(3) CaO–CaS clusters are found at 5 min after calcium
addition. There are two reasons that possibly explain the for-
mation. Firstly, dissolved calcium reacts with dissolved oxy-
gen and sulfur in the first 5 minutes, thus a huge number of
tiny CaO–CaS type inclusions are generated. Initially, they
are very small and not detectable. Later, they grow into larg-
er ones through agglomeration and become detectable, thus
the inclusion number increases. CaO–CaS clusters form
with further agglomeration. Secondly, CaS has a strong ten-
dency to promote inclusions to agglomerate.
Acknowledgments
The authors gratefully express their appreciation to the
National Basic Research Program of China (No.
2010CB630806) and National Natural Science Foundation
of China (No. 51204014 and No. 51304013) for sponsoring
this work.
REFERENCES
1) M. Burty, P. Dunand, J. Ritt, H. Soulard, A. Blanchard, G. Jeanne,
F. Penet, R. Pluquet and I. Poissonnet: Ironmaking Conf. Proc., Vol.
56, ISS, Warrendale, PA, (1997), 711.
2) C. Bonilla: Steelmaking Conf. Proc., Vol. 78, ISS, Warrendale, PA,
(1995), 743.
3) M. Nagao, K. Hiraoka and Y. Unigame: Sanyo Tech. Rep., 12 (2005),
38.
4) H. Nakasugi, S. Sugimura, H. Matsuda and T. Murata: Nippon Steel
Tech. Rep., 297 (1979), 66.
5) N. Bannenberg: Steelmaking Conf. Proc., Vol. 78, ISS, Warrendale,
PA, (1995), 457.
6) K. Ahlborg: Steelmaking Conf. Proc., Vol. 84, ISS, Warrendale, PA,
(2001), 861.
7) G. Faulring, J. Farrel and D. Hilty: Continuous Casting, Vol. 1, Iron
and Steel Society, Warrendale, PA, (1985), 57.
8) K. Larsen and R. J. Fruehan: Iron Steelmaker, 17 (1990), 45.
9) S. Choudhary and A. Ghosh: ISIJ Int., 48 (2008), 1552.
10) G. Ye, ouml, P. Jönsson and T. Lund: ISIJ Int., 36 (1996), S105.
11) W. Yang, L. Zhang, X. Wang, Y. Ren, X. Liu and Q. Shan: ISIJ Int.,
53 (2013), 1401.
12) G. Jing, C. Shu-Sen and C. Zi-Jian: ISIJ Int., 53 (2013), 2142.
13) Z. Deng and M. Zhu: Steel Res. Int., 84 (2013), 519.
14) N. Verma, P. Pistorius, R. Fruehan, M. Potter, M. Lind and S. Story:
Metall. Mater. Trans. B, 42 (2011), 720.
15) E. Zinngrebe, P. Seda, S. van der Laan and B. van Arendonk: Proc.
5th Int. Cong. on the Sci. Technol. of Steelmaking, ENGICOM
GmbH, Germany, (2012).
16) H. Itoh, M. Hino and S. Ban-Ya: Tetsu-to-Hagané, 83 (1997), 773.
17) K. Taguchi, H. Ono-Nakazato, D. Nakai, T. Usui and K. Marukawa:
ISIJ Int., 43 (2003), 1705.
18) S.-W. Cho and H. Suito: ISIJ Int., 34 (1994), 265.
19) Steelmaking Data Sourcebook, The Japan Society for the Promotion
of Science, The 19th Committee on Steelmaking, Gordon & Breach
Science Publications, New York, (1988).
20) Recommended Values of Equilibrium Constants for the Reactions in
Steelmaking, The Japan Society for the Promotion of Science, The
19th Committee on Steelmaking, JSPS, Tokyo, (1984).
21) B. Ozturk and E. Turkdogan: Met. Sci., 18 (1984), 299.
22) M. Numata and Y. Higuchi: ISIJ Int., 52 (2012), 2013.
23) M. Numata, Y. Higuchi and S. Fukagawa: Tetsu-to-Hagané, 84
(1998), 159.
24) T. Kushida, Y. Higuchi, M. Numata, A. Yamamoto, K. Ohnishi, A.
Teraguchi and J. Fujino: Sumitomo Search, 58 (1996), 24.
25) Y. Wang, S. Sridhar and M. Valdez: Metall. Mater. Trans. B, 33
(2002), 625.
26) A. Ikeda, M. Nakanishi, T. Kaneko, Y. Sumitomo and M. Takeyama:
Sumitomo Search, 26 (1981), 91.
27) S. Nafisi, J. Jordan, C. D’Souza, L. Collins and T. J. Drake: Proc.
Conf. ISSTech 2012, Association for Iron and Steel Technology,
Warrendale, PA, (2012), 1195.
© 2015 ISIJ