http://www.paper.edu.

cn

Cleanliness of Low Carbon Aluminum-Killed Steels

during Secondary Refining Processes
Wen Yang, Xinhua Wang, Lifeng Zhang, Qinglin Shan, and Xuefeng Liu

Efficient secondary refining process is necessary for massive and stable production of low
carbon aluminum-killed (LCAK) steels. Plant trials were performed to investigate the
cleanliness of steels. Characteristics of composition, cleanliness, and inclusions of LCAK
steel during different secondary refining processes, including LF, CAS, RH-LIT, and RH,
were studied and compared. The results showed that CAS, RH-LIT, and RH processes had
better control of low carbon, silicon, and nitrogen than LF process. High cleanliness of
LCAK steels could be achieved by all the mentioned refining processes. It was concluded
that the total oxygen (T.O.) should be <35 ppm to reduce the amount of inclusions and
reach the level of clean steels. The removal rate of inclusions during RH-LIT and RH
processes was much higher than that of LF and CAS. The T.O. content and the amounts of
inclusions during CAS, RH-LIT, and RH could be quickly decreased to a low value within
10 min. The results showed that CAS and RH-LIT as well as RH refining processes can
produce LCAK steels that meet the requirements of high efficient, low cost, clean, and
stable production, while LF is more suitable for the heats with poor control of end point of
BOF, and for the process with calcium treatment to control sulfur content and lower the
clogging of Submerged Entry Nozzle during thin slab continuous casting.

1. Introduction process is the important step to link up steelmaking and

continuous casting. The LT gas stirring treatment was
Low carbon aluminum-killed (LCAK) steels are well used developed approximately in 1951.[1] This process is simple
for production of cold rolled sheets used for automobiles, and low cost,[2] however, slags are needed to be deoxidized
household appliances, etc. It requires that LCAK steel to prevent the reoxidation of molten steel and prevent the
should have low carbon, low silicon, and low nitrogen to increase of inclusion amount. Moreover, the molten steel is
obtain good forming property, and the total oxygen (T.O.) easy to be reoxidized by air when the stirring is too strong.
and the size of inclusions of LCAK steels should be strictly As the tight quality requirement of the steel, LT is hardly
controlled to guarantee high surface quality of the product. used alone or little used nowadays. The CAS refining proc-
The secondary refining processes of LCAK steels grades ess was developed in 1975 by Nippon Steel[3] and expanded
include LT (Ladle Treatment by Gas Stirring), CAS gradually at the end of 1970s. Comparing the LT process,
(Composition Adjustment by Sealed Argon Bubbling), the CAS process utilizes a snorkel, which hinders the
RH (RuhrstahlHeraeus), and LF (Ladle Furnace treat- explosion of the molten steel to air and slag and therefore
ment). Besides improving steel quality, secondary refining lowers the reoxidation of the molten steel. Thus, the stir-
ring intensity can be strengthened to promote the removal
of inclusions, leading to the short refining time and rela-
W. Yang, Prof. X. Wang, Prof. L. Zhang tively high cleanliness of steel and high yield of alloy. RH
State Key Laboratory of Advanced Metallurgy and School of refining process was developed by the two companies
Metallurgical and Ecological Engineering, University of Science and
Technology Beijing, Beijing 100083, China of Ruhrstahl and Heraeus, so called RH, in 1959.[4] The
Email: zhanglifeng@ustb.edu.cn advantage of RH is degassing and removal of inclusions
Q. Shan
Steelmaking Department of Shougang Jingtang United Iron & Steel
utilizing high vacuum and circulation flow. Due to the
Co., Ltd, Tangshan 063200, China using of vacuum chamber, the RH process can well avoid
X. Liu the reoxidation of molten steel from air and slag, and the
School of Materials Science and Engineering, University of Science
and Technology Beijing, Beijing 100083, China stirring of RH can be very strong. Multiple functions of CAS
and RH were expanded in the 1970s and 1980s, such as
DOI: 10.1002/srin.201200213 CAS-OB, RH-OB, RH-KTB, RH-PB for additional function

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 473

FULL PAPER http://www.paper.edu.cn

of reheating, decarburization, and desulfurization.[5] into the molten steel to decrease the dissolved oxygen to
Besides, the RH light treatment (RH-LIT) was proposed several ppm. Synthetic slag was added into the ladle during
in 1975 to reduce the burden of steelmaking furnace.[5] tapping to reduce the FeO and MnO in the slag. During LF
Different with RH route in which deoxidation is performed refining, high basicity slag was used and aluminum was
during BOF tapping, in RH-LIT route there is no deoxida- added to improve deoxidation. Steel samples were taken
tion performance during tapping, and the first stage of RH- before LF refining, at the earlier stage, the middle stage, the
LIT refining process is decarburization by the dissolved end of LF refining, after Ca treatment, and after soft
oxygen in vaccum condition, following by deoxidation with blowing.
aluminum addition, which leads to a lower dissolved oxy- In CAS refining route, during tapping of BOF, a certain
gen before deoxidation and less alumina generated after amount of Al and synthetic slag were added into the molten
deoxidation. As a result, RH-LIT process allows relatively steel as well. After tapping, slag modifier was added into
high carbon and low dissolved oxygen at the end point of the slag to reduce the slag oxidizability. At CAS station, the
BOF. LF refining process was developed in 1971,[5,6] the snorkel was submerged into the molten steel after the slag
main characteristic is steel heating by electric arc, thus it surface was blown open to a bright face, followed by a
can melt ferroalloy by electric arc heating, adjust compo- strong stirring by argon gas to promote the removal of
sition, and temperature, perform desulfurization and deox- inclusions. The temperature was decreased by adding steel
idation by a reducing slag, and control the cleanliness and scraps and be raised by adding aluminum and blowing
inclusions in the molten steel. The typical advantages of oxygen. Steel samples were taken before CAS refining,
the LF refining process include high cleanliness of steel, during process of argon blowing with time interval of
modification of inclusions, and stable temperature, thus, it approximately 7 min, and after argon stop.
is widely used nowadays although it was developed later In the RH refining route, the end point of BOF was
than other refining processes. controlled under common practice, deoxidation, and
There have been many studies on the cleanliness of alloying were performed during tapping of BOF. During
LCAK steel at different refining processes,[723] the T.O. the RH refining, a large argon gas flow rate was employed
content in LCAK steels summarized in Table 1.[714] to promote the recirculation and removal of inclusions
From these studies, it is known that the high cleanliness from the molten steel. While in the RH-LIT route, the
of LCAK steel can be achieved by all the refining processes carbon content at the end point of BOF was relatively
discussed above. In recent years, the LCAK steel pro- higher than the normal RH process, and no deoxidation
duction faces many challenges, including the require- or alloying were performed during the tapping of BOF.
ments of high quality of the steel product and low Decarburization was firstly performed during RH-LIT
production costs, less environment pollution, and low process under a certain vacuum degree. After decarburiza-
energy consumption. Thus, high efficient, low cost, and tion, aluminum was added for deoxidation. Then, the
stable production of high quality steel is necessary for the generated inclusions were removed by flow transport
sustainable development of steel enterprises. However, it is under excellent dynamic conditions. Steel samples were
lack of comparison between different refining processes of taken after deoxidation. Since the temperature drop was
LCAK steel, leading to considerable divergences on the large during RH process, aluminum addition, and oxygen
choice of refining process for LCAK steel. Accordingly, in blowing were performed to increase the temperature of
the current paper, characteristics of steel composition, the molten steel.
cleanliness, and inclusions of each refining process includ- The schematic operation and sampling scheme of
ing LF, CAS, RH, and RH-LIT were studied and compared each secondary refining process were shown in Figure 1.
to provide a basis for the choice of refining processes. Temperature of the molten steel was measured several
times during the refining process of each heat. The dis-
solved oxygen content was measured by a zirconia sensor
during RH and RH-LIT refining processes and at the begin-
2. Experimental and Analysis Methodology ning of other refining processes.
The contents of C, S, and T.O. of the steel samples were
Industrial trials for the production of LCAK steel were analyzed using infrared analysis. The concentration of
performed. The production routes are KR (to remove silicon was analyzed using ICP-AES and the nitrogen con-
sulfur from the hot metal) ! 300 t BOF ! LF Ca treat- tent was analyzed using the thermal conductance method.
ment/CAS/RH/RH-LIT ! continuous casting. The LF Inclusions on the cross-section of each steel sample
refining with Ca treatment process is for the production were detected using the automated SEM/EDS inclusion
of SPHC steel while the other refining processes are for the analysis (ASPEX). ASPEX is a computer-controlled scan-
production of DC01 steel. The compositions of SPHC and ning electron microscope that is designed for the auto-
DC01 steel are listed in Table 2. mated imaging and elemental analysis of a wide spectrum
In this work, calcium treatment was assumed to be a of surfaces and particulates. The system provides a fully
part of LF refining process. In the LF refining route, during integrated SEM and EDX platform for addressing the
tapping of BOF, a certain amount of aluminum was added microscale visualization needs. It can simultaneously

474 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn

Steel works Refining Ladle Tundish Slab Year Ref.

America
Inland No.4 BOF shop RH-OB 6080 840 2003 [7]

Dofasco, Canada EAF, LMF 14 2003 [7]

BOF, LMF 23 2003 [7]

Europe
Dillinger, Germany Tank Degas 1015 2000 [7]

Sidmar, Belgium 37 2000 [7]

RH 2056 2002 [8]

Corus strip Products, Port Talbot Works, UK CAS-OB, or RH 1015 2001 [7]

Teesside Technology Center, Corus, UK 6 2002 [7]

Voest-Alpine Stahl GmbH, Linz, Austria RH-OB 1036 25 2003 [7]

China
Panzhihua Iron and Steel Co. RH-MFB 2024 2000 [7]

Ar stirring 15.2 2000 [7]

Maanshan Iron and Steel ASEA-SKF 1824 2000 [7]

BaoSteel CAS 73100 3853 1417 2000 [9]

WISCO, No.2 Works RH (Pressure vessel steel 2834 2426 1219 2000 [7]

WISCO, No.3 Works RH 16 20 15 2002 [7]

Anshan Iron and Steel ANS 3556 3335 25 2003 [10]

Lianyuan Iron and Steel LF Ca treatment 3036 3442 3743 2006 [11]

476 476
steel BaoSteel, int. Meishan
researchsteel research works
84 (2013)int.
No. 84
5 (2013) No. 5 LF Ca treatment 34
2012 WILEY-VCH Verlag50
2012 WILEY-VCH
GmbHVerlag 16
& Co. GmbH 2006
KGaA, [12]
KGaA, Weinheim
& Co. Weinheim

Qianan Iron and Steel CAS 1924 1720 1115 2007 [13]

Tangshan Iron and Steel LF Ca treatment 1639 1946 1238 2009 [14]

Table 1. Total oxygen content (ppm) reported during various processing steps for LCAK steel after the year of 2000.

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 475
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

Steel C Si Mn P S Alt N

SPHC 0.05 0.03 0.2 0.015 0.012 0.0150.05 0.006

DC01 0.03 0.03 0.2 0.015 0.012 0.0150.05 0.006

Table 2. Composition of SPHC and DC01 steel (%).

Figure 1. Key operation and sampling scheme of each refining process (Sam., sampling; meas., measuring; add., adding).

476 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

detect the size, shape, area, number, location, and elemen- 0.06 LF RH-LIT
tal composition of inclusions in steel samples. The working CAS RH
magnification was set at 225 and the minimum detect- 0.05

Carbon content (%)

able inclusion was 1.0 mm in the current study.
A statistical parameter, area fraction, is defined as the 0.04
following formula to characterize the amount of non-
metallic inclusions. 0.03

0.02
Ainclusion
AF (1)
Atotal 0.01 Carburant addition
in RH-LIT

where, AF is the area fraction of inclusions, ppm; Ainclusion 0.00

is the total area of detected inclusions, mm2; and Atotal is the 0 10 20 30 40 50 60
sample detection area, mm2. Time of refining (min)
Since an inclusion with irregular shape especially a
cluster might be cut into two or more particles on one Figure 2. Variation of carbon content during refining processes.
cross section, AF is more accurate than the number density
to describe the amount of the irregular shape inclusions
since the area contains the information of both number
and size.
In the current study, in order to reveal the three-dimen- the carbon content increased to the target value. Since
sional morphology of inclusions in the steel produced by decarburization can be performed in RH-LIT process
CAS and RH refining processes, inclusions were extracted route, the carbon content at the end point of BOF can
from steel samples.[24,25] After the surface of the steel be controlled in a higher value and wider range than the
sample being polished, about 3 g of each steel sample normal RH process, leading to an easier control to
was dissolved in hydrochloric acid solution (1:1 in volume) reach the target value at the end point of BOF. While in
at 708C. After the iron matrix was completely dissolved, other refining routes, since no decarburization was per-
deionized water was added to dilute the solution. Then, the formed and might occur carbon pick-up during refining, it
inclusions were filtered out using a polycarbonate filter is required that the carbon content in steel should be
paper propelled by a vacuum pump. The pore size of the controlled to a low value and in a narrow range at the
filter was 0.2 mm in order to capture most of the inclusions. end point of BOF, which increases the difficulty of BOF
As alumina-based inclusions are insoluble to HCl acid, control and increases the damage possibility of the
they will remain on the filter paper. The extracted BOF refractory.
inclusions were then observed using SEM after depositing
carbon powder over the filter paper under vacuum
conditions.
3.2. Silicon Content

To obtain good forming property of LCAK sheets, [Si] in

LCAK steel should be <0.03%. The variation of [Si] in each
3. Results and Discussions
refining process as well as the comparison of [Si] content
and [Si] pick-up after different refining processes are
The content of carbon, silicon, sulfur, T.O., nitrogen, and
shown in Figure 3, indicating that the order from high
inclusions in steel samples taken during steel refining
to low [Si] content by different refining processes is:
processes were evaluated.
LF > CAS, RH > RH-LIT, while the order from large to small
silicon pick-up during refining is: LF > CAS, RH, and RH-
LIT.
3.1. Carbon Content
The reaction inducing the increasing of silicon is
Figure 2 shows the carbon content in steel during
secondary refining processes. Since the carbon content 3SiO2 4Al 2Al2 O3 3Si (1)
of SPHC is a little higher than that of DC01, the carbon
content in steel during LF refining process is higher than The larger [Si] pick-up in LF refining process is due to the
other processes. This figure shows that the carbon content active reaction between slag and molten steel. During the
during LF, CAS, and RH was slightly changed during LF refining process, the activity of SiO2 in slag is improved
refining. While at the initial stage of RH-LIT process by decreasing the oxidizability of the slag. Besides, the
the carbon content decreased significantly to <0.01% stirring of steel caused by argon blowing provides a good
due to the decarburization, after the addition of carburant, dynamic condition to the reaction.

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 477
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

(a) (b)
0.035 0.030

[Si] and [Si] pick-up after refining (%)

LF [Si] after refining
0.030 CAS 0.025 [Si] pick-up
RH-LIT
0.025 RH
[Si] in steel (%)

0.020
0.020

0.015 0.015
0.010
0.010
0.005

0.000 Deoxidation in RH-LIT 0.005

0 10 20 30 40 50
0.000
Time of refining (min) LF CAS RH-LIT RH

Figure 3. Variation of [Si] during refining processes (a), and comparisons of [Si] content and [Si] pick-up after different refining
processes (b).

3.3. Desulfurization content is needed to improve the efficiency and stability of

the calcium treatment. Thus, the desulfurization during
The casting speed of the thin slab continuous casting is different refining processes is discussed as well.
higher than that of the traditional continuous casting proc- The [S] content in steel and the efficiency of desulfur-
ess. Low sulfur content is required to avoid slab breakout ization of each secondary refining process are shown
during casting. Besides, calcium treatment is usually used in Figure 4, which indicates that the LF refining process
to modify inclusions from solid to liquid to prevent the has the best desulfurization efficiency and with the lowest
submerged entry nozzle from being clogged, and low sulfur [S] content in steel after refining.

(a) (b)
0.014 0.014
LF RH-LIT
[S] content after refining (%)

CAS RH
0.012 0.012

0.010
[S] in steel (%)

0.010

0.008 0.008
0.006
0.006
0.004
0.004
0.002
0.002
0.000
0 10 20 30 40 50 60
0.000
Time of refining (min) LF CAS RH-LIT RH

(c)
70
Efficiency of desulfurization (%)

60

50

40

30

20

10

0
LF CAS RH-LIT RH

Figure 4. Variation of [S] content during refining processes (a), [S] content in steel after secondary refining (b), and efficiency of
desulfurization (c) of each refining process.

478 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

1680 7

Average temp. drop rate (C/min)

Temperature of molten steel ()
LF
1660 CAS 6
RH-LIT
1640
RH 5
slag
4
1620
3
1600
2
1580
1
1560
0 10 20 30 40 50
0
Time of refining (min) LF CAS RH-LIT RH
(a) (b)
Figure 5. Variation of temperature (a) and the average temperature drop rate (b) of each refining process.

The desulfurization reaction is expressed by[26] The heating methods of the refining processes are
different from each other. The attribute of LF refining
3CaO 3S 2Al 3CaS Al2 O3 (2) process is electric arc heating of steel, thus the temperature
control is much easier and more accurate, which is good
According to Equation 2, in order to increase the efficiency for discharging the burden for the furnace, slagging during
of desulfurization, the activities of CaO and [Al] should be the refining, and keeping suitable temperature of the melt
higher. Since the steel grade decides the [Al] content in to caster, but it may induce carbon pick-up and nitrogen
steel, the desulfurization mainly depends on the activity of pick-up. While the temperature of the molten steel during
CaO in slag. Because a high basicity slag was used in LF CAS or RH refining is raised by aluminum addition and
refining process in the current study, the activity of CaO oxygen blowing, which increases the T.O. and the amount
was high, leading to the highest desulfurization efficiency. of inclusions in steel.
On the other hand, LF process is designed to desulfurize
because the heating method is more suitable for the use of
high basicity slag. 3.5. Total Oxygen and Nitrogen

In most refining processes except RH-LIT, deoxidation

3.4. Temperature Control were performed by adding aluminum during tapping of
BOF, thus the oxygen content at the end point of
Larger temperature drop rate requires more temperature BOF is assumed to be the dissolved oxygen before
compensation to ensure castability of the molten steel, deoxidation. Figure 6 shows the contents of [C] and [O]
leading to more cost and possible worse steel cleanliness. before deoxidation in CAS, RH, and RH-LIT process, and at
Thus, small temperature drop rate is usually respected.
In the current study, the temperature drop rate is
defined as the average drop rate in the range from the 2000
CAS
beginning of refining to the first temperature rising oper-
RH
ation except for the LF refining process, in which the 1600 RH-LIT at end
parameter is defined as average temperature drop in the point of BOF
[O ] c o n te n t (p p m )

refining time except the time period of slag addition, elec- RH-LIT after
1200 decarburization
trical arc heating, and soft blowing. Figure 5 presents the
C-O equilibrium
variation of temperature of molten steel during refining
1225

processes and the average temperature drop rate of each 800

refining process. The arrows in Figure 5a show the first
645

temperature rising operations during refining processes. It 400

is indicated that the order of temperature drop from large
to small during different refining processes is: RH > RH-
0
LIT > LF and CAS. The large temperature drop rate during 0.00 0.01 0.02 0.03 0.04 0.05 0.06
RH and RH-LIT refining which is owing to the strong argon
[C] content (%)
blowing and the larger circulation rate of the molten steel
will increase the burden of BOF for higher temperature Figure 6. Contents of [C] and [O] of steel in different refining
requirement. processes.

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 479
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

(a) (b)
100 70
LF CAS
60
T.O. and N in steel (ppm)

T.O. and N in steel (ppm)

80 T.O. T.O.
N Al N
50
addition
60
40
Ca treatment
40 30

20
20
10 Argon blowing

0 0
0 10 20 30 40 50 0 10 20 30 40 50
Time of refining (min) Time of refining (min)

(c) (d)
20
220 80 RH
RH-LIT

T.O. and N in steel (ppm)

200 T.O. T.O.
16
N N
T.O. in steel (ppm)

60
N in steel (ppm)
180
12
120 40
90 8
20
60
Heating by Al 4
30 addition and O2 blowing Heating by Al addition and O2 blowing
0
0 0
16 20 24 28 32 36 40 0 10 20 30 40 50
Time of refining (min) Time of refining (min)

Figure 7. Variation of T.O. and N during refining processes, a) LF, b) CAS, c) RH-LIT, and d) RH.

the end point of BOF in RH-LIT process is shown in the oxygen content at the end point of BOF might be due to the
figure as well. The dissolved oxygen content increased with excessive oxidation of melt by oxygen blowing. Lower
the decreasing of carbon at the end of BOF, and smaller carbon content induces higher oxygen content and easier
carbon content induced more deviation of dissolved oxy- excessive oxidation. Thus, it is harmful to pursue too low
gen from its equilibrium value, which is calculated at the carbon content at the end point of BOF.
pressure of 1 atm. In RH-LIT refining route, the [O] content Figure 7 shows the variation of T.O. and N contents
before deoxidation was smallest, 215 ppm. In RH refining during each secondary refining process. At the earlier stage
route, the value was 645 ppm higher than that of RH-LIT. of refining processes of CAS and RH, the content of
The [O] content before deoxidation was highest in CAS T.O. decreased significantly, with a decreasing speed of
route, 1225 ppm higher than that in RH-LIT. The high 4 ppm min1, as shown in Figure 8a. During LF refining

(a) (b)
30 60
T.O. after secondary refining (ppm)
- d T .O ./d t o f L F /C A S /R H ( p p m /m in )

-dT.O./dt of RH-LIT (ppm/min)

6 Heating of RH
20 50
3
10 40
0
0 30
-3
LF -10 20
CAS
-6 RH-LIT -40
Heating of RH-LIT
RH 10
-9 -50
0 10 20 30 40 50 0
Time of refining (min) LF CAS RH-LIT RH

Figure 8. The decreasing rate of T.O., dT.O./dt of each refining process (a) and T.O. contents (b) after secondary refining.

480 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

40 3.6.1. Composition and Morphology

35 N after refining Figure 10 shows the composition of inclusions at the end
N pick-up of each refining process, detected by ASPEX. The curve in
30
the figure represents the boundary of liquid inclusion
N and N pick-up after

25 region. In LF route, after calcium treatment the oxide

refining (ppm)

inclusions transferred to CaOMgOAl2O3 (Figure 10a),

20
and many of them were located in the liquid region, which
15 is good for preventing nozzle clogging and improving the
10 castability of the steel. However, Figure 10b represents that
after calcium treatment a vast amount of CaS generated,
5
most of them were precipitaed on the oxides, this would
0 decrease the modification effect of alumina inclusions.
LF CAS RH-LIT RH Thus, lower sulfur content is required to improve the
inclusion modification. At the end of CAS, RH-LIT, and
Figure 9. N content and N pick-up during each refining
process. RH refining processes, since there was no calcium treat-
ment, inclusions were solid alumina, which would induce
nozzle clogging during casting. Thus, on the point of view
of the castability of the molten steel, LF refining (including
process, the T.O. decreased gradually, while during CAS calcium treatment) is better than CAS, RH-LIT, and RH.
and RH refining, after a fast decreasing at the first 89 min, The transformation of inclusions during LF refining
the T.O. content kept almost constant. During RH-LIT and process was owing to the reaction of calcium treatment.
RH refining processes, the heating of steel by the addition High basicity reductive slag was used during LF refining,
of aluminum and oxygen blowing caused a sharp increase the reaction between the slag and the molten steel changed
of T.O. At the end of each refining process, the T.O. the composition of inclusions, especially after calcium
decreased to a low value 33 ppm by LF, 34 ppm by addition. Inclusions were modified to liquid ones mainly
CAS, 45 ppm by RH-LIT, and 28 ppm by RH, as shown according to the following reaction.
in Figure 8b, indicating that high cleanliness of steel can
be achieved by all the refining processes discussed. 7Al2 O3 12Ca 12O 12CaO  7Al2 O3 (3)
However, T.O. during RH-LIT was relatively higher than
that of other refining processes due to the shorter net time As mentioned above, oxide inclusions after LF refining was
for the removal of inclusions. mainly MgOCaOAl2O3 that could be partially dissolved
Figure 9 shows the nitrogen content and nitrogen pick- by hydrochloric acid, while after CAS and RH refining
up for each refining process. It is indicated that the order inclusions were mainly alumina that would not be dis-
of nitrogen content from high to low is: LF > CAS and solved by hydrochloric acid. Accordingly, the morphology
RH > RH-LIT, and the nitrogen pick-up was largest during of inclusions after LF refining was directly observed by
LF refining process, which was as high as 12 ppm and SEM, and inclusions after CAS and RH refining were firstly
mainly occurred during Ca treatment, as shown in extracted using HCl acid and then observed by SEM to
Figure 7a, which might be attributed to the air absorption reveal the three-dimensional morphology of alumina
caused by the boiling of molten steel on the near-surface of inclusions.
the molten pool at time of CaFe wire injection that was Figure 11 shows the morphology of inclusions at the
induced by the high steam pressure of calcium. end of each refining process. Since inclusions after RH
refining were the same as those after RH-LIT refining,
inclusions only after RH-LIT refining are shown here.
After LF refining, inclusions were transformed to spherical,
3.6. Inclusions in Steel during Refining and most of them were small. CaS was precipitated on the
gap part of the spheres shown in Figure 11a. However, after
Inclusions in LCAK steel after deoxidation are mainly CAS and RH-LIT refining, there were still many irregular
alumina-based. Large alumina-based inclusions have a inclusions, and large clusters and aggregates were
detrimental effect on the castability of the molten observed, which were very detrimental to the subsequent
steel[2729] and the surface quality of the steel sheet[3032]. casting process. Besides clusters and aggregates, many
Currently, large alumina-based inclusions are either modi- small spherical and polyhedral particles were observed
fied into low melting ones by calcium treatment,[9,33,34] or as well in the steel samples after CAS and RH refining.
removed by flow transport,[35,36] or become defects of the
steel product.[26] 3.6.2. Amount and Removal Rate
Some aspects of inclusion characteristics, including The amount of inclusions is characterized by their area
composition, morphology, amount, size, and remove rate fraction (AF). Figure 12 shows the variation of AF of
in each refining process, were analyzed. inclusions during each secondary refining process.

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 481
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

Figure 10. Composition of inclusions at the end of each refining process: a,b) LF, c) CAS, d) RH-LIT, e) RH.

482 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

Figure 11. Morphology of inclusions after refining processes a) LF, b) CAS, and c) RH-LIT.

During LF refining process, before calcium treatment, the increased the AF of inclusions due to the massive gener-
AF of inclusions decreased, while after calcium treatment, ation of new alumina inclusions.
the amount of inclusions increased sharply, which was The relationship between the AF of alumina-based
caused by the massive generation of new inclusions especi- inclusions and the T.O. is shown in Figure 13. It is clearly
ally sulfide (as shown in Figure 10b) and slight reoxidation indicated that the AF of alumina-based inclusions
of molten steel (as shown in Figure 7a) during calcium increased with the increasing of T.O. in steel, and it
treatment. During other refining processes, the variation of increased slowly when the T.O. was <35 ppm but increased
AF was consistent with the tendency of T.O. At the early exponentially after the T.O. was over 35 ppm. Thus, the
stage of refining process (10 min), the AF of inclusions T.O. in steel should be controlled to <35 ppm to reduce the
decreased significantly, and then decreased slowly. During amount of inclusions. As discussed by Zhang,[22,37] the T.O.
RH-LIT and RH refining processes, the heating of molten only represents the inclusions smaller than 50 mm. How to
steel by adding aluminum and blowing oxygen extremely control large inclusions that are mainly exogenous ones

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 483
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

300
300
LF
250 CAS
250
AF of inclusions (ppm)

AF of inclusions (ppm)
200 200
Ca treatment
150 150 Argon blowing

100 100

50 50

0 0
10 20 30 40 50 60 0 10 20 30 40 50
Time of refining (min) Time of refining (min)

1000 700
Heating by adding Al
and blowing O2 RH-LIT 600 RH
800
AF of inclusions (ppm)

AF of inclusions (ppm)
500 Heating by adding Al and blowing O2
600
400

400 300

200
200
100
0 0
16 20 24 28 32 36 40 44 0 10 20 30 40 50
Time of refining (min) Time of refining (min)

Figure 12. AF of inclusions during each secondary refining process.

from slag and lining refractory materials is another import- CAS refining and before calcium treatment in LF refining
ant task of steelmakers. process. However, after calcium treatment, the AF was
The AF of inclusions at the end of each refining process sharply increased to 232 ppm owing to the massive
is shown in Figure 14, where LF(Ca) means before generation of CaS and slight reoxidation of the steel during
calcium treatment and LF(Ca) means after calcium treat- the practice of calcium treatment.
ment. The AF of inclusions was smallest by refining of The minimum AF of inclusions characterizes the
RH-LIT or RH, 37 ppm, around 52 ppm at the end of removal ability of inclusions from the molten steel, while

240
400
AF of inclusions (ppm)
AF of inclusions (ppm)

220
300
200
200 60

40
100

20
0
0 20 40 60 80 100
0
T.O. content in steel (ppm) LF(- Ca) LF(+ Ca) CAS RH-LIT RH

Figure 13. Relationship between the AF of inclusions and the Figure 14. AF of inclusions at the end of each secondary
T.O. of the steel. refining.

484 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

3.6.3. Size Distribution

Figure 17 shows the average size and size distribution of
inclusions during refining processes. The number density
(ND) of inclusions is defined as the number of inclusions
per unit detection area. The tendency of average size of the
inclusions was similar to the AF of inclusions. During LF
refining, before calcium treatment, the size of inclusions
decreased gradually because of the removal of >2 mm
inclusions. Immediately after calcium treatment, the aver-
age size decreased to only 1.53 mm owing to the massive
generation of 12 mm inclusions by the calcium treatment.
As the time increasing during soft blowing, inclusions
aggregated and their average size increased. During CAS
Figure 15. Schematic of removal rate of inclusions. refining, during argon blowing process, the average size
decreased, while after argon blowing was off, the average
size increased firstly due to the aggregation of small
inclusions and then decreased due to the floating removal
the removal efficiency of inclusions is characterized by of large inclusions. During RH-LIT refining, at the early
average removal rate of the AF at the fast decreasing stage stage of steel heating, many large alumina generated which
of AF, defined as Equation 4. led to a large average size. At the later stage of heating,
it quickly deceased due to the fast removal of large
AFin  AFlow inclusions. During RH refining, at the early stage, it
hIR (4)
Dt IR decreased significantly due to the removal of inclusions,
and increased obviously during the heating stage for the
where, hIR is the removal rate of inclusions in AF, ppm/
same reason as that during RH-LIT.
min; [AF]in is the initial AF of inclusions, ppm; [AF]low is the
The average size of inclusions at the end of each refining
AF of inclusions at the end of fast decreasing stage shown
process is shown in Figure 18, indicating that the average
in Figure 12, ppm; DtIR is the time of the fast decreasing
size of inclusions after LF refining was smallest due to the
stage, min. The schematic diagram is shown in Figure 15.
modification by slagsteel reaction and calcium treatment.
The calculated removal rate of inclusions during each
The actual size of inclusion in CAS or RH refining may be
refining process is shown in Figure 16, indicating that the
larger than the measured one because of the dominant
removal rate of inclusions in AF in RH-LIT and RH was far
irregular and clustered alumina inclusions.
larger than that in LF and CAS owing to the quick circu-
lation flow of molten steel. The very low removal rate
during LF refining was due to the relatively small bottom
argon blowing rate. Another reason is that the initial AF of
3.7. Total Refining Time
inclusions considered in the calculation (at 14 min) might
not be included in the actual initial fast decreasing stage
The secondary refining process is a connection link
that is generally during the initial 10 min of the refining
between the preceding steelmaking and the following
process.
casting, thus besides functions mentioned previously,
the secondary refining process is also a regulation of the
70
production tempo. To improve the production efficiency
and reduce manufacturing costs, high speed continuous
60 casting is expected, which requires shorter secondary
Removal rate of inclusions

refining time.
50
The refining time in the current work is defined as the
in AF (ppm/min)

40 time from start to the end of treatment, e.g., for LF and CAS
refining it is the time from the start of to the end of argon
30 blowing, and for RH-LIT and RH refining it is the time from
the start to the end of vacuum. The refining time of each
20
process is shown in Figure 19, indicating that the refining
10 time of LF was longest and that of CAS was shortest. Here,
the LF refining process included calcium treatment. RH-
0
LIT process contained a decarburization stage and a steel
LF CAS RH-LIT RH
heating process, leading to a relatively long refining time.
Figure 16. Removal rate of inclusions during each refining Although there was no decarburization stage during RH
process. process, however, there were two steel heating processes,

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 485
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

Figure 17. Average size and size distribution of inclusions during each refining process.

486 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

Average size of inclusions (P)

70
4.0
3.5 60

Total refining time (min)

3.0 50
2.5
40
2.0
30
1.5
20
1.0
0.5 10

0.0 0
LF CAS RH-LIT RH LF CAS RH-LIT RH

Figure 18. Average size of inclusions at the end of each refining Figure 19. Refining time of each secondary refining process.
process.

which induced much longer time than expected. same. The slags added during the tapping of BOF were
Therefore, the temperature rising operation should be nearly the identical except that the slag modifier added
shortened as far as possible to lower the total refining time. after tapping in RH-LIT route was less. Thus, the materials
This requires a compact and reasonable plants layout with large consumption mainly includes aluminum metal
and more scientific management of the manufacturing for deoxidation and alloying such as AlFe, Al gravels, other
process. alloy such as CaFe, slags added during refining process
such as synthetic slag and fluorite, argon gas for stirring,
and materials for temperature rising such as electrical
3.8. Material Consumption power during LF refining and aluminum and oxygen
during RH-LIT and RH refining.
Low cost manufacturing is necessary and urgent nowadays Table 3 shows the consumption of the main auxiliary
for the environmentally friendly and sustainable develop- materials for each refining process, the power for vacuum
ment of steel industries. The key of low cost manufacturing pumps in RH-LIT and RH was not considered. It indicates
is to produce high quality products with the least consump- that the material consumption during LF refining process
tion of resource raw materials. In the current study, material was much larger than those during other processes.
consumption of each refining process was analyzed. The comparison on the aspects analyzed above among
Material consumption mainly includes the consump- the four secondary refining processes is summarized
tion of: alloys, slags, argon gas, oxygen gas, power, and so in Table 4, indicating that CAS and RH-LIT as well as
on. For the current study, the steel grade was similar and RH can well meet the requirements of high efficient, low
the consumption of MnFe for alloying was almost the cost, and stable production of LCAK steels, while LF is

Materials LF CAS RH-LIT RH

AlFe (kg t1-steel) 3.45 6.94 0 5.55

1
Al grain (kg t -steel) 1.29 0.30 1.11 0.70
1
CaFe (kg t -steel) 3.82 0 0 0
1
Synthetic slag (kg t -steel) 8.77 0 0 0
Fluorite (kg t1-steel) 4.62 0 0 0
3 1
Argon (Nm t -steel) 0.091 0.088 0.408 0.387
1
Electricity for heating (kWh t -steel) 23.37 0 0 0
1
Al grain for heating (kg t -steel) 0 0 1.34 0.87
3 1
Oxygen for heating (m t -steel) 0 0 1.08 0.91

Table 3. Main materials consumption of each secondary refining process.

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 487
 http://www.paper.edu.cn
www.steel-research.de
FULL PAPER

Aspects Items LF CAS RH-LIT RH

Composition and Resist [C] fluctuation at the end of BOF D D * D

temperature control
Resist temperature fluctuation at the end of BOF * D D D
Decarburization   * D
Desulfurization * D  
Avoid Si pick-up  * * *
Temperature drop * *  
Temperature compensation * D D D
Cleanliness [O] before deoxidation   * D
T.O. after refining * * D *
[N] after refining  D * D
Avoid [N] pick-up D * * *
Inclusions Modification *   
Amount  D * *
Average size * D D D
Removal rate  D * *
Castability of steel * D D D
Efficiency and cost Refining time  * D D
Production cost  D * D
Construction cost D *  
Note: *, Good; D, Fair; , Poor.

Table 4. Comparison of different refining processes.

more suitable for the case with a low level of refining (2) High cleanliness of LCAK steels could be achieved by
technology, or for the case with poor control of end points all the mentioned refining processes. The T.O., e.g., was
of BOF steelmaking, or for thin slab continuous casting 33 ppm for LF, 34 ppm for CAS, 45 ppm for RH-LIT, and
process. 28 ppm for RH. The T.O. should be <35 ppm to reduce
the amount of inclusions.
(3) The amount of inclusions characterized by the AF after
each refining was: 52 ppm before calcium treatment
4. Conclusions and 232 ppm after calcium treatment for LF process,
52 ppm for CAS, 37 ppm for both RH-LIT and RH proc-
Characteristics of the composition, cleanliness, and esses. The higher amount of inclusions after calcium
inclusions of LCAK steels by different secondary refining treatment was due to the massive generation of CaS
processes were studied based on the plant trials, and then a and slight reoxidation during calcium treatment.
comprehensive comparison was performed. The following (4) The removal rate of inclusions during RH-LIT and RH
conclusions were obtained: was much higher than during LF and CAS processes.
The T.O. content and the amount of inclusions during
(1) The processes of CAS, RH-LIT, and RH could keep low CAS, RH-LIT, and RH could be quickly decreased to a
values of carbon, silicon, and nitrogen which are low value within 10 min refining.
required for the production of LCAK steels, while (5) According to the comprehensive comparison, it is
higher pick-up of silicon and nitrogen were found known that CAS and RH-LIT as well as RH can well
for LF refining process. meet the requirements of high efficient, low cost, and

488 steel research int. 84 (2013) No. 5 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
 http://www.paper.edu.cn
www.steel-research.de

stable production of LCAK steels, while LF is more [17] L. Zhang, X. Wang, Shandong Metall. (in Chinese)
suitable for the case with low level of refining technol- 2005, 27(1), 1.
ogy, or the case with poor control of end points of BOF, [18] L. Zhang, X. Wang, Shandong Metall. (in Chinese)
or for thin slab continuous casting process. 2005, 27(2), 1.
[19] L. Zhang, J. Iron Steel Res. Int. 2006, 13(3), 1.
Acknowledgments [20] J. Wei, Z. Tian, L. Zhang, K. Cai, Y. Zhou, Inclusions
in the Low Carbon Al-killed Steel Produced by a
The authors are grateful for support from the National CSP Thin Slab Casting Process at Handan Steel, in
Science Foundation China (grant no. 51274034), the Proccedings of AISTech 2005 Iron & Steel Technology
High Quality Steel Consortium at University of Science Conference and Exposition, Vol II, AIST, Warrandale,
and Technology Beijing (China), and Shougang Jingtang PA (May 912, 2005, Charlotte, NC), 2005, pp. 585
United Iron & Steel Co. (China) for industrial trials. 592.
[21] L. Zhang, B. G. Thomas, K. Cai, L. Zhu, J. Cui,
Received: August 21, 2012;
ISSTech 2003 Conf. Proc. 2003, 2, 141.
Published online: December 5, 2012
[22] L. Zhang, B. G. Thomas, ISIJ Int. 2003, 43(3), 271.
[23] L. Zhang, B. G. Thomas, X. Wang, K. Cai, Evaluation
Keywords: cleanliness; inclusion; LCAK steel; secondary and control of steel cleanliness Review, 85th
refining Steelmaking Conference Proceedings, March 10,
2002 March 13, 2002, (Nashville, TN, United
states), Iron and Steel Society, 85, 2002, 431452.
References
[24] S. Li, L. Zhang, X. Zuo, Observation of Inclusions in
[1] E. Spire, AIME Electric Furnace Steel Proc. 1951, 9, 75. the Steel Using Partial Acid Extraction and SEM in
[2] J. Szekely, H. J. Wang, K. M. Kiser, Metall. Trans. B Proccedings of Materials Science and Technology
(Process Metall.) 1976, 7B, (2) 287. (MS&T), AIST, Warrandale, PA, (Pittsburgh) 2008,
[3] H. Haraguchi, K. Okobira, U. Muneo, Tetsu Hagane pp. 12591269.
1975, 61(4), S135. [25] L. Zhang, S. Li, J. Wang, X. Zuo, Iron Steel (Chinese)
[4] H. Thielmann, H. Maas, Stahl Eisen 1959, 79(5), 276. 2009, 44(3), 75.
[5] H. Kajioka, Ladle Refining-Challenges to Mass [26] G. Ye, P. Jonsson, T. Lund, ISIJ Int. 1996, 36 (Suppl),
Production Systems of Various Grade and High S105.
Quality Steels, Metallurgical Industry Press, Beijing [27] B. Wang, Y. Pan, L. Tian, K. Hou, W. Wang,
2002. Steelmaking (in Chinese) 2008, 24(6), 41.
[6] H. Ushiyama, G. Yuasa, T. Yajima, Ironmaking [28] R. Dekkers, Ph.D. Thesis Thesis, Katholieke
Steelmaking 1978, 5(3), 121. University, Leuven, 2002.
[7] L-f. Zhang, J. Iron Steel Res. Int. 2006, 13(4), 1. [29] R. Dekkers, B. Blanpain, P. Wollants, F. Haers,
[8] R. Dekkers, B. Blanpain, P. Wollants, ISSTech 2003 B. Gommers, C. Vercruyssen, Steel Res. 2003, 74(6),
Conf. Proc. 2003, 2, 197. 351.
[9] G. Shi, L. Zhang, Y. Zheng, J. Zhi, Iron Steel [30] G. Zhu, H. Yu, W. Wang, X. Wang, Iron Steel (in
(in Chinese) 2000, 35(3), 12. Chinese) 2004, 39(4), 54.
[10] X. Yuan, H. Li, X. Wang, Iron Steel (in Chinese) 2003, [31] W. Gao, Adv. Mater. Res. 2012, 402, 221.
38(11), 15. [32] P. Rocabois, J. Pontoire, ISSTech 2003 Conf. Proc.
[11] Q. Peng, Y. Li, C. Yang, B. Chen, Steelmaking 2003, 2, 995.
(in Chinese) 2007, 23(3), 45. [33] Y. Higuchi, M. Numata, S. Fukagawa, K. Shinme, ISIJ
[12] C. Ren, Research on the Cleanliness of Continuous Int. 1996, 36 (Suppl), S151.
Casting Slabs of Low Carbon Al-Killed Steel, Thesis, [34] N. Verma, P. Pistorius, R. Fruehan, M. Potter,
University of Science and Technology, Beijing, 2007. M. Lind, S. Story, Metall. Mater. Trans. B 2011,
[13] T. Zhang, Y. Bao, H. Cui, F. Yue, Y. Jin, Special Steel 42(4), 720.
(in Chinese) 2008, 29(2), 60. [35] W. Zhu, D. Jin, B. Li, Iron Steel (in Chinese) 1991,
[14] W. Yang, J. Cao, X-h. Wang, Z-r. Xu, J. Yang, J. Iron 26(2), 22.
Steel Res. Int. 2011, 18(9), 6. [36] Z. Xue, Y. Wang, L. Wang, Z. Li, J. Zhang, Acta
[15] S. Yang, L. Zhang, J. Li, K. Peaslee, Metall. Mater. Metall. Sin. (in Chinese) 2003, 39(4), 431.
Trans. B 2012, 43B(4), 731. [37] L. Sun, L. Zhang, J. Li, A. Dong, Y. Chen, S. Yang,
[16] L. Zhang, J. Iron Steel Res. Int. 2006, 13(4), 01. Iron Steel Technol. 2012, 9(4), 20.

2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim steel research int. 84 (2013) No. 5 489