Detection of Non-metallic Inclusions in Steel Continuous

Casting Billets
YING REN, YUFENG WANG, SHUSEN LI, LIFENG ZHANG, XIANGJUN ZUO,
SIMON N. LEKAKH, and KENT PEASLEE

This work applied automated particle analysis to study non-metallic inclusions in steel. Com-
pared with traditional methods, the approach has the advantage of capturing the morphology,
measuring the size, recording the original positions, and identifying the composition of inclu-
sions on a selected area in a short time. The morphology and composition of typical inclusions
were analyzed using partial acid extraction and discussed through thermodynamic calculation.
Steel samples were collected from the entire cross section of billets cast during times of steady
state and ladle change. The spatial distribution of inclusions agreed well with the measurement
of the total oxygen. The spatial distribution of inclusions was plotted to represent the entrap-
ment positions of inclusions on the casting strand and their concentration on the cross section of
the billet. Also, regarding the different size and type of inclusions, the spatial distribution of
classified inclusions was explored such as the distribution of sulfide, oxide, and high sodium and
potassium content inclusions. The sufficient information could be used to identify the source of
inclusions and guide the steel refining process.

DOI: 10.1007/s11663-014-0042-y
 The Minerals, Metals & Materials Society and ASM International 2014

I. INTRODUCTION amount, morphology, size distribution, and spatial

distribution of inclusions; and identifying their chemical
CHARACTERIZING non-metallic inclusions is one composition. The research group of the current author
of the most important aspects to assuring clean steel. has extensively investigated the different aspects of
Inclusions remaining in the final product can damage inclusions, including the experimental investigation,
steel properties and degrade its quality.[1,2] To produce industrial trials, modeling, and literature review.[1,3–12]
high-performance steel, non-metallic inclusions must be Zhang and Thomas[1] reviewed more than twenty
controlled. Generally, non-metallic inclusions can be methods of evaluating impurities and inclusions in steel.
classified as either indigenous or exogenous. Indigenous Among these methods, a metallographic microscope[13] is
inclusions are a result of alloying elements within the widely applied to explore the two-dimensional morphol-
steel reacting with dissolved gas (generally oxygen) to ogy and size of inclusions in steel. However, it cannot
form solid inclusions in the cast steel. The inclusion identify the chemical composition of inclusions. Scanning
could be formed during deoxidation, reoxidation, or electron microscopy[14] permits exploration of morphol-
solidification from reduced gas species solubility in the ogy of inclusions in the microscale. Further, by coupling
solid state. Exogenous inclusions come from sources with energy-dispersive spectroscopy, scanning electron
outside the liquid steel, such as slag entrainment or microscopy permits identification of the elemental com-
refractory damage. The evaluation of non-metallic position of inclusions and their relative proportions.
inclusions in steel is of great interest to metallurgists Counting a large number of inclusions with this method,
and materials scientists and includes exploring the total however, is too time-intensive to be practical. Acid
extraction method can obtain the stereoscopic morphol-
ogy of inclusions, but destroy sulfide inclusions.[15–18] For
the extraction using bromine-methanol, iron bromide
YING REN, Ph.D. Student, and LIFENG ZHANG, Professor, are
Beijing Key Laboratory of Green Recycling and Extraction of Metals
dissolves readily and in the electrolytic approach, in
(GREM) and School of Metallurgical and Ecological Engineering, general, a chelant must be added to dissolved the iron,
University of Science and Technology Beijing, No. 30, Xueyuan Road, which helps to reveal the three-dimensional morphology
Haidian District, Beijing 100083, P.R. China. Contact e-mail: and keep original composition of the inclusions.[19]
[email protected] YUFENG WANG, Engineer, is with the Ultrasonic inspection[20,21] and X-ray detection[22] are
R&D Group, SSAB, Muscatine, IA. SHUSEN LI, Assistant General
Manager, Qian’an Steelmaking Co. Ltd., Shougang Group, Qian’an more efficient than these methods, and can be employed
064404, Hebei, P.R. China. XIANGJUN ZUO, Engineer, is with the for online analysis; however, the morphology and com-
Continuous Casting Department, CISDI Engineering Co. Ltd., position data are difficult to obtain. Other indirect
Chongqing, P.R. China. SIMON N. LEKAKH, Research Professor, methods, like total oxygen (T.O.) measurement[23,24] or
and KENT PEASLEE, Professor, are with the Department of
Materials Science & Engineering, Missouri University of Science and
continuous excitation of spark spectrum, have the limi-
Technology (Missouri S&T), Rolla, MO. tations to determine both the distribution and the
Manuscript submitted June 28, 2011. composition of inclusions.[25,26]

METALLURGICAL AND MATERIALS TRANSACTIONS B

 Table I. Chemical Composition of the Steel (Percentage)

C Si Mn S Ti Mg Al Ca
0.20 0.23 0.48 0.006 0.002 0.0006 0.011 0.0002

In the present study typical inclusions after different

time acid extractions in steel continuous casting billets
were observed, which was in accordance with thermo-
dynamic calculation. Moreover, the size distribution,
spatial distribution, and composition of non-metallic
inclusions in steel continuous casting billets were inves-
tigated using automatic particle analyzer—ASPEX. This
method permits quick determination of the morphology,
size distribution, spatial distribution, and composition
of all inclusions in a selected area. The inclusions results
were compared with measured total oxygen analyzed by
a Leco oxygen analyzer.

II. METHODOLOGY
A. Experimental Procedures
Billet samples were obtained from an industrial plant
casting aluminum-killed medium carbon steel in a four
strand, 150 9 150 mm2 billet caster. The composition of Fig. 1—Sampling from a steel billet.
the steel produced is given in Table I. The tundish was
lined with refractory material high in magnesium oxide observation of inclusions was performed using an
with vertical straight nozzles submerged 100 mm below automated SEM–EDS—ASPEX 1020 system.[27]. The
the meniscus of the mold. The casting speed was 1.5 to automated particle analysis is a frame-based analysis.
2.0 m/min resulting in a standard casting time of about First, a selected area was divided into several small fields
45 minutes per heat with 2 to 3 minutes required to under the microscope. A focused electron beam was
change a ladle. A casting sequence typically pours 4 to 6 positioned on one field, and the microscope did a fast
heats with 60 ton per heat, depending on the lifetime of search of the field with large search steps. The intensity
the submerged entry nozzle and the casting schedule. of the back-scatter electrons was then recorded and
Billets samples were collected at both steady state and transferred to a computer. Once a particle was found
unsteady state. One sample was taken from the second based on thresholding of the intensity of the back-
strand of the billet caster after 3 heats of pouring at a scatter electrons, the step size was reduced to measure
casting speed of 1.5 to 1.7 m/min to represent the more precisely the dimensions of the particle. The center
steady-state (S1) operation. Another sample was of the particle was identified using a bisected chord
obtained during ladle change to reflect the unsteady method, and the rotating chord algorithm was employed
state of the casting process (S2). These samples were cut to measure the particle. Finally, the beam was again
into sections, as shown in Figure 1. A billet sec- positioned in the center of the particle to acquire the
tion 5 mm in thickness was used for oxygen analysis characteristic X-ray. After combination of these steps in
to map the three-dimensional distribution of oxide one field, the beam moved to other fields and repeated
inclusions over the billet cross section. the analysis.[28,29]
For the oxygen samples, a 5-mm-thick billet slice was In the present study, using ASPEX 1020 system, the
sawed into 45 rectangle samples with the dimension of size distribution, composition, and rough morphology
30 9 15 9 5 mm3, as shown in Figure 2(a). Each sam- are automatically obtained. Although the images of
ple was machined into three cylindrical samples, with a inclusions obtained using ASPEX are not high in
diameter of 5 mm for the oxygen analyzer. For the resolution, the ASPEX has the advantage of fast
metallographic sample, a 20 mm billet slice was divided scanning and automatic feature analysis resulting in a
into 49 samples, each measuring 20 mm by 20 mm, as large number of particles in a sample being analyzed in
illustrated in Figure 2(b). several hours. A sample measuring 20 9 20 mm2 was
selected for analysis on each metallographic specimen
and the electron beam scanned the random area on the
B. Analysis Methods 20 9 20 mm2 sample until a 9.8 mm2 scanned area was
The analysis of oxygen was performed using a Leco analyzed. The minimum particle size was set at 0.6 lm,
Oxygen Analyzer in which the sample was melted inside which means that the ASPEX 1020 was sensitive to only
a graphite crucible and the product gas quantified by a inclusions larger than 0.6 lm. The types of inclusions
detector and converted to the total oxygen content. The were categorized according to the weight percentage of

METALLURGICAL AND MATERIALS TRANSACTIONS B

Fig. 2—Sample preparation.

Table II. Inclusions Classification Rules (Weight Percentage)

FeO stain/spots/pollution Fe ‡ 30 and O ‡ 2.5
Microdefects Fe ‡ 97 and O < 2.5
MnS Mn > 30 and S > 20
CaS Ca > 30 and S > 20
Ca-Mn-S Ca > 10, Mn > 10, and S > 20
Other sulfides S > 20
Al2O3 Al > 20, Mn < 20, Ca < 10, Si < 20, S < 20, and Mg < 5
MgO Mg > 5, Al < 10, Mn < 20, Ca < 10, Si < 20, and S < 20
Al-Mg-O Mg > 5, Al > 10, Mn < 20, Ca < 10, Si < 20, and S < 20
CaO Ca > 20, S < 20, Al < 10, Si < 20, Mn < 20, and Mg < 5
Ca-Al-O Ca > 10, Al > 10, Si < 20, Mn < 20, Mg < 5, and S < 20
Ca-Mg-Al-O Ca > 10, Mg >5, Al > 10, S < 20, Si < 20, and Mn < 20
SiO2 Si > 20, Mn < 20, Al < 10, Ca < 20, and S < 20
MnO Mn > 20, Al < 20, S < 20, and Si < 20
Mn-Si-O Mn > 20, Si > 10, Al < 10, and S < 20
Mn-Al-O Mn > 20, Al > 10, S < 20, and Si < 20
ZrO2 Zr > 10, Al < 10, Ca < 20, Si < 20, S < 20, and Mn < 20
TiO2 Ti > 20, Al < 10, Ca < 20, Si < 20, Mn < 20, and S < 20
Na2O + K2O Na + K > 3

elements detected in them. The rule file is shown in indicates the morphology of inclusions by ASPEX
Table II. In order to observe the three-dimensional (Figure 3). The typical inclusions in the steel billets are
morphologies of inclusions, typical morphologies of alumina-based clusters, alumina magnesia (Al2O3-MgO)
inclusions were revealed using partial acid extraction. A spinel single or cluster inclusions, silicon oxide (SiO2)
36 pct HCl solution was used as the solvent. Morphol- bearing spherical inclusions, separate sulfide-based
ogies and compositions of inclusions on a steel sample inclusions, and irregularly shaped slag inclusions con-
(32 9 25 9 15 mm3) at different extraction times were taining calcium oxide (CaO), potassium oxide (K2O),
compared. and sodium oxide (Na2O). Alumina (Al2O3) inclusions
mainly stem from deoxidation by aluminum or reoxi-
dation of the residual aluminum in the steel. Inclusions
III. MORPHOLOGY AND COMPOSITION high in magnesium oxide are a result of the reaction
OF INCLUSIONS between molten steel and magnesium refractory, espe-
cially the tundish lining refractory. Inclusions with
Figure 3 shows the typical morphology and compo- sodium (Na) and potassium (K) are from the entrain-
sition of inclusions in the billet sample, which clearly ment of the mold slag. Sulfide inclusions mainly

METALLURGICAL AND MATERIALS TRANSACTIONS B

 CaS MnS Al2O3 SiO2
Average Diam. = 5.7 µm Average Diam. = 2.5 µm Average Diam. = 11 µm Average Diam. = 32.9 µm
Ca=50.3%, S=36%, Al=7.9%, Mn=66.6%, S=31.2%,Ti=2.2% Al=71.3%,Mn=7.9%, Si=61.7%, Ca=11.9%,
Si=2.2%, Mn=2.3% Mg=1.4% Ca=5.1%, Na=1.6%,Mg=1.1%, Na=21.3%, Zr=2.3%,
Si=5.5%, Ti=4.7%, Al=1.1%, Mn=1.6%
Zr = 2.8 %

Al2O3-SiO2 Al2O3-MgO Al2O3-CaO Na-K

Average Diam. = 55.2 µm Average Diam. = 12.7 µm Average Diam. = 4.9 µm Average Diam. = 21.3 µm
Si=47.1%, Al=29.9%, Al=49.8%, Mg=29.8%, Ca=35.5%,Al=24.7%, K=52.3%, Mg=8.0%, Ti=7.6%,
K=16.8%, Zr = 3.8% S=7.2%, Si=2.8% S=17.9%, Mn=12%, Mg=6.6%, Mn =1.8% Na=1.2%, Al=1.0%
Mg=1.2%, Mn=1.2% Ca=6.7%, Mn=2.1% Si=3.4%
Zr = 1.6%

Fig. 3—Inclusion morphology and composition (in mass pct) observed by SEM (polished sample).

Fig. 4—Inclusion morphology and composition (atomic pct) observed by SEM (etched by acid).

METALLURGICAL AND MATERIALS TRANSACTIONS B

 Compositions
No. Inclusions after different time acid extraction
(atom %)

O 69.1%
(a) Al 28.0%
Ca 2.9%

O 67.2%
Mg 6.0
(b)
Al 25.9%
Ca 0.9%

O 60.0%
(c) Al 39.1%
Ca 0.9%

O 58.5%
(d) Al 26.2%
Ca 15.3%

O 68.8%
Al 8.7%
(e)
Si 15.0%
Mn 7.4%

Fig. 5—Typical inclusions after different time acid extractions (the numerical value in the top right corner is the acid extraction acid time).

METALLURGICAL AND MATERIALS TRANSACTIONS B

Fig. 6—Al2O3-CaO-MgO inclusions in steel.

Fig. 7—Clogging materials (mainly Al2O3-CaO-MgO inclusions) at the submerged entry nozzle.

precipitate during cooling and solidification of the steel. analytical time and storage space for thousands of
The optical resolution from ASPEX was set for provid- counted inclusions.
ing sufficient information to determine the diameter, The morphology and composition of typical inclu-
area, and perimeter of inclusions with reasonable sions etched by acid were detected using SEM–EDS

METALLURGICAL AND MATERIALS TRANSACTIONS B

 of the inclusions were compositions with high melting
points, the submerged entry nozzle was seriously
clogged in the industrial trial, as shown in Figure 7,
resulting in early termination of casting.
Figure 8 shows the calculated stability diagram of
Al-Ca-O system in the molten steel at 2146 K (1873 C),
which gives the stability regions for various inclusions at
equilibrium as a function of [Al], [Ca], and [O] on the
molten steel. The equilibrium constant logK of the
reactions and the interaction activity coefficient were
different from each other in the literatures,[30–34] assum-
ing unity activity of the oxide products. The values near
the dotted circles show the oxygen content [mass ppm
O]. To avoid nozzle clogging, the [Ca] and [Al] content
of steel should be controlled within the shade liquid
Fig. 8—Calculated stability diagram of Al-Ca-O system in the mol- region in Figure 8. Among the various Al-Ca-O inclu-
ten steel at 2146 K (1873 C). sions, only Al2O3-3CaO and 7Al2O3-12CaO inclusions
are liquid at 2146 K (1873 C). The [Ca] and [Al]
(Figure 4). Figures 4(a) through (d) are plate-like, contents of the steel shown in Table I were located in the
angular, spherical, and irregular Al2O3-based inclusions. Al2O3-rich region, indicating that the calcium treatment
Precipitation of different shape individual alumina during the ladle refining in the current industrial trial
particles along with the decrease of supersaturation was insufficient and more Ca-Si wire was needed.
degree of deoxidizing elements in the initial stage of The Equilibrium precipitation of inclusions during
holding time after deoxidation occurred.[18] Meanwhile, solidification of steel with a composition of Fe-0.2C-
plenty of Al2O3-MgO spinel inclusions were generated, 0.23Si-0.48Mn-0.006S-0.002Ti-0.0006Mg-0.0018O-
as shown in Figure 4(e). There were spherical calcium 0.007N-0.011Al-Ca (in mass pct) was calculated using
aluminate and silicon aluminate inclusions (Figures 4(f) the thermodynamic software, FACTSAGE (version 6.4)
and (g)). Spherical Si-Mn-O inclusions and hexahedral- using ‘‘FACT53’’, ‘‘Ftoxid,’’ and ‘‘FSstel’’ databases[35]
like TiN inclusions were also found, as shown in and is shown in Figure 9. When [Ca] in the steel was
Figures 4(h) and (i). After the acid extraction, CaS 2 ppm (Figure 9(a)), inclusions were mainly spinel and
and MnS disappeared since sulfide reacted with the hardly liquid in the molten steel and they caused clogging
acid.[15,17] of the submerged entry nozzle during continuous casting.
The three-dimensional morphologies of inclusions During solidification of the steel, liquid inclusions were
were revealed using partial acid extraction (Figure 5). converted to CaMg2Al16O27 and CaS. After the temper-
After acid extraction, separate Al2O3 and Al2O3-MgO ature reached approximately 1573 K (1300 C), plenty of
single particles were revealed to big clustered inclusions MnS and small amounts of TiN inclusions would be
(Figures 5(a) and (b)). In Figure 5(c), the diameter of precipitated. With [Ca] content reaching 5 ppm first and
spherical Al2O3-based inclusions after 300s acid extrac- then to 10 ppm (Figures 9(b) and (c)), the amount of
tion was larger than that under 180s acid extraction spinel inclusions decreased and then disappeared, and
condition. Inclusions of Al2O3-CaO and Al-Si-Mn-O there were only liquid inclusions in the molten steel. The
shown in Figures 5(d) and (e) were nearly spherical due inclusions of MnS were modified to CaS by the calcium
to the fact that their melting temperature was lower than treatment. If more calcium was added (Figures 9(d) and
the temperature of the molten steel. (e)), both CaS and liquid inclusions were generated in
Figure 6 shows the ternary diagram of oxides based molten steel, and CaS would possibly clog the submerged
on the chemical composition of Al2O3, CaO, and MgO. entry nozzle at this condition. Therefore, the lack of
Each individual spot in this ternary diagram represents control of the composition of inclusions during steel-
the composition of an inclusion. In continuous casting, making and refining has a catastrophic consequence on
solid particles can attach to the nozzle and may block the casting process.
the path of the molten steel. To prevent nozzle clogging,
inclusions should be in the liquid state. However, the
composition distribution indicates that most inclusions
were solid under casting temperature. Only a few were in IV. SIZE AND SPATIAL DISTRIBUTION
the liquid state, as illustrated by the shaded region OF INCLUSIONS IN BILLETS
[temperature <1773 K (1500 C)] in the ternary dia-
gram. The samples from both steady state (S1) and ladle A. Size Distribution of Inclusions
change (S2) show that many particles were close to the Figure 10 shows the two-dimensional size distribution
Al2O3 region. Excess MgO would raise the melting point of the oxide inclusions directly obtained from the
of inclusions. Several inclusions on the ternary diagram inclusion size measurement by ASPEX. Most of the
had magnesium oxide contents greater than 30 pct, inclusions in the billets were 0.6 to 5.0 lm. Inclusions
indicating that the melting temperature of these particles larger than 20 lm were typically complex and stemmed
was over 2573 K (2300 C), and thus these inclusions from the refractory lining, entrained slag, or a combina-
were solid throughout the casting process. Since many tion of deoxidation and reoxidation products.[7] These

METALLURGICAL AND MATERIALS TRANSACTIONS B

Fig. 9—Equilibrium precipitation of inclusions during solidification for steel of composition: Fe-0.2C-0.23Si-0.48Mn-0.006S-0.002Ti-0.0006Mg-
0.0018O-0.007N-0.011Al-Ca in mass pct.

large inclusions are particularly harmful to the mechan- B. Spatial Distribution of Inclusions
ical properties, acting as stress concentrators or casing The distribution of inclusions with various composi-
surface cracking. Figure 10 indicates that there were more tions in the billet had special significance and it provided
>20 lm inclusions measured on the cross section of the information about the distribution of alumina-based
billet during ladle change (S2) than that at the steady state inclusions stemming from deoxidation,[18,36] reoxidation,[37]
(S1) due to the fact that more slag was entrained and more entraining refining slag,[38] or refractory materials.[39,40]
air was absorbed at the unsteady state than steady state.

METALLURGICAL AND MATERIALS TRANSACTIONS B

The inclusions of Na/K-bearing were from mold slag (S1), there were fewer sulfide and oxide inclusions than
entrainment[1]; most of sulfide precipitated during the that during the ladle change period (S2). Sulfide
cooling and solidification of the steel.[41] The distribu- inclusions measured tend to accumulate at the edge of
tion of sulfide inclusions and oxide inclusions on the the billet (Figure 11). As shown in Figure 12, for
billet cross section as measured with the ASPEX measured oxide inclusions, there were more inclusions
inclusion analyzer is shown in Figures 11 and 12, close to the inner radius, but there was no clear
respectively, which are based on the number fraction. accumulation at the center of the billets; on cross
On the billet cross section under steady casting state section analyzed of the billet during ladle change period
(S2), there were more inclusions accumulated in the
center and the inner radius side of the billet. Figure 13
shows the spatial distribution of Al-Ca-Mg spinel
inclusions detected on the cross section of the billets.
These inclusions are scattered over the entire cross
section and concentrated slightly in the center of the
billet, with some deviation to the right side. The
distribution of inclusions with high Na and K content
is illustrated in Figure 14, no obvious distribution
regularity was found in the measured cross section of
the billet. Although ‘‘steady-state’’ case is steady casting,
it has oscillatory behavior. Therefore, the distribution
and the source of the inclusions can be determined more
exactly by examining more different cross sections.
The number of small size inclusions was high, but
small inclusions have little impact on the quality of
steels. Only large size inclusions can directly lead to the
failure of steel products. Therefore, the distribution of
oxide inclusions measured was analyzed by different
sizes. Figure 15 shows the distribution of oxide inclu-
sions measuring 1, 5, and>5 lm under steady casting. It
indicates that 1 lm oxide particles scatter uniformly
over the entire cross section, whereas 5 and >5 lm
particles tend to be concentrated at the center and on the
inner radius of the billet. This pattern of entrapment of
non-metallic inclusions in steel continuous casting billets
was verified using mathematical simulation in our
Fig. 10—Two-dimensional size distribution of oxide inclusions. previous study.[42] The simulation result of distribution

Inner Radius (mm) Inner Radius (mm)

150 150
2.5 Num% Num%
135 135 2.5
0 0
2.5 1.5 2.0 1.0 1.5 2.0 2.5
120 120
1.0 1.0
105 105
1.5 2.0 1.0 1.5 1.5
Thickness (mm)

Thickness (mm)

2.0 2.0
90 90 2.0

75 3.0 75 1.0 3.0

1.5 1.5 0.5 2.0 3.0
60 1.5 4.0 60 4.0
1.5
2.5 3.5 2.5 2.0 1.0 2.5 3.5
45 45
2.0
30 2.5 30
1.0 1.5 1.5
2.0
15 15
2.5 2.5 3.0
3.0 4.0
0 0
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm) Outer Radius (mm)
(a) Steady state (S1) (b) Ladle change (S2)
Fig. 11—Distribution of sulfide inclusions on the cross section of the billet.

METALLURGICAL AND MATERIALS TRANSACTIONS B

 Inner Radius (mm) Inner Radius (mm)
150 3.5 Num% 150
3.5 Num%
1.5 1.5
135 2.5 0 135 2.5
5.0 3.5 2.5 0
1.5 1.5 1.5
120 3.5 1.0 120
3.0 1.0
3.0 2.0
2.5
Thickness (mm)

Thickness (mm)
105 2.0 105 2.5 2.0
2.5 1.5 1.5 1.0
2.0
90 3.0 90 2.0 3.0
75 4.0 75 2.5 4.0 5.0 3.5 1.0
2.0 4.0
60 5.0 60 1.0 2.5 5.0
1.5 2.0 3.0 2.5 1.5 1.5 4.5 4.0
45 45 2.5
3.0 3.5
30 1.5 30 3.0
1.5 1.5 0.5 2.5 3.0
15 15
1.0 2.5
2.5
0 0
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm) Outer Radius (mm)
(a) Steady state (S1) (b) Ladle change (S2)
Fig. 12—Distribution of oxide inclusions on cross section of the billet.

Inner Radius (mm)

Inner Radius (mm)
150 1.5 3.0 % 150
4.0 1.5 1.5 2.0 Num%
135 3.5 4.0
0 135 1.5 1.5
1.5 0.5 2.5 0
120 2.0 1.0 2.0 0.5
1.0 120 0.5
2.0 1.0 1.0
105 1.5
Thickness (mm)

Thickness (mm)

2.5 3.0 105 2.5 1.5 1.5

2.0 1.5 2.5 2.0
90 2.5 1.5 1.5
90 2.5
2.0 3.0 3.5
75 1.5 1.0 3.0
3.5 75 1.0 2.0 0.5 3.5
60 4.0 2.5 1.5 4.0
60 3.0
1.5 2.0 1.0
45 45 2.5
2.5
30 30 3.0
2.5 1.5 1.0 1.0
15 15
3.5
2.0
0 0
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm) Outer Radius (mm)
(a) Steady state (S1) (b) Ladle change (S2)
Fig. 13—Distribution of inclusion with Al, Mg, and Ca on billet cross section.

of 5 lm oxides on the cross section of the billet under As is well known, the total oxygen can reflect the
steady state (S1) is shown in Figure 16. At the initial level of inclusions and cleanliness of billet. The total
solidification, the distance between the dendrite arms oxygen results directly detected by the Leco oxygen
can only provide space for smaller inclusions, whereas analyzer are shown in Figure 17. Under steady-state
the approaching larger inclusions might be pushed casting (S1), the total oxygen concentrated at the
away. As the dendrites grow, the dendrite spacing has center of the billet and the average total oxygen was
the capacity to entrap much larger inclusions. 10.3 ppm. During the ladle change period, the total

METALLURGICAL AND MATERIALS TRANSACTIONS B

 Inner Radius (mm) Inner Radius (mm)
150 6 Num% 150 Num%
135 2 2 135
4 2
120 4
120 4
16 6 6

Thickness (mm)

Thickness (mm)
105 8 105 8
14 2 2 10 4
90 18 10
12 90 6 12
2
6 14 14
75 16 75 2 16
18 8 18
60 2 60 2 18
45 2 12
45 6 16
30 2 30 12 8 6 10 16
2
15 15 10
4
0 0
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm) Outer Radius (mm)
(a) Steady state (S1) (b) Ladle change (S2)

Fig. 14—Distribution of inclusion with Na and K on billet cross section.

Inner Radius (mm) Inner Radius (mm)

150
150 Num%
3.5 Num% 4
1.5 135
135 2.5 0
0 6 1
3.5 3.0 1.5 120 5 1 2
120 1
1.0 1 2
105 2.5 105 4
2.0 3.0 2.5 4 5 2
Thickness (mm)

Thickness (mm)

2.0 3
90 2.5 90 4
6
3.0 75 3 5
75 1.5 2.0 1.5
2.0 4 6
60 4.0 60
1 1 2 4
2.5 2.0 1.5 2.5 2 3
45 3.0 45
30 2.5 30 2
2.5 2.0 1.5 1.5 1
15 15 1
1.0
0 0
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm) Outer Radius (mm)
(a) 1 µm oxides (b) 5 µm oxides
Inner Radius (mm)
150 Num%
4 2
135 1
0
120 4 1
2
Thickness (mm)

105 2 6 3
90
5 4
75 5
6
60 4
3 5
45
1
30 3
15 3
2 1
0
0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm)
(c) >5 µm oxides

Fig. 15—Distribution of different size oxide inclusions on the measured cross section of the billet under steady state (S1).

METALLURGICAL AND MATERIALS TRANSACTIONS B

oxygen scattered over the entire cross section and with V. CONCLUSIONS
an average value of 15.7 ppm, which was 52 pct higher
than that during the steady state—an indication of Automated particle analysis was applied to study the
increased reoxidation. Comparing the distribution non-metallic inclusions in the entire cross section of a
inclusions by ASPEX, inclusions by calculation and billet. The following was concluded:
the measured total oxygen, it was indicated that there (1) Automated particle analysis overcomes several
was a high oxide peak close to the center of on billet weaknesses of traditional measures. It provides an
cross section. accurate and fast scanning in a selected area to
collect information on particles, including mor-
phology, size, spatial distribution, and composition.
(2) Three-dimensional morphologies of typical inclu-
Inner Radius (mm) sions were revealed by acid extraction. The mor-
75 phology of inclusions after different time acid
60 extractions was obtained.
45
(3) The experimental results and thermodynamic cal-
culation indicated that most oxide inclusions were
30
Thichness (mm)

0.40 solid under casting temperature with only a few in

15 the liquid state since the refining process did not
0.40
properly modify inclusions. Inclusions with high
0 0.30
melting point can block the nozzle and have a det-
-15 0.40 rimental effect on the continuous casting operation.
Num. %
-30 0.30 (4) On the cross section of the billet under steady cast-
0.0
0.050 ing state, there were fewer sulfide and oxide inclu-
-45 0.10
0.20 0.15 sions than that during the ladle change period. The
0.20
-60 0.25 distributions of inclusions on the entire cross section
0.10 0.20 0.20 0.30
-75 0.35 of the billet were plotted according to inclusions
0.40
-75 -60 -45 -30 -15 0 15 30 45 60 75 0.45 type and their diameter.
0.50
Outer Radius (mm) 0.55 (5) On the measured cross section, total oxygen value of
0.60
the ladle change period is higher than that of steady
Fig. 16—Distribution of 5 lm oxides on the cross section of the bil-
casting state, indicating that more air is absorbed
let under steady state (S1) by mathematical simulation[42]. during ladle change period.

Inner Radius (mm) Inner Radius (mm)

150 150
T.O. (ppm) 18 17
9 15 T.O. (ppm)
135 11
135 16
5 16 16 5
10 14 17
120 10 120
12
9 8 17
8
105 11 11 8 8 11 105 16 14 15 11
Thickness (mm)

Thickness (mm)

12 7 18
14 14
90 12 11 90
13 17 14 17 15 16 17
75 75 17 15
17 20 16 16 20
60 15
16 60
18
10 10 14 16 15
45 45 16 16
15

30 11
30 16
14
18 17
16 17 16
15 15
15 15
0 0
0 15 30 45 60 75 90 105 120 135 150 0 15 30 45 60 75 90 105 120 135 150
Outer Radius (mm) Outer Radius (mm)
(a) Steady state (S1) (b) Ladle change (S2)

Fig. 17—Distribution of the T.O. in the billet by oxygen analyzer.

METALLURGICAL AND MATERIALS TRANSACTIONS B

 ACKNOWLEDGMENTS 17. L. Zhang, S. Li, J. Wang, and X. Zuo: Iron Steel (Chin.), 2009,
vol. 44 (3), pp. 75–80.
The authors are grateful for support from the 18. W. Yang, X. Wang, L. Zhang, and W. Wang: Steel Res. Int., 2013,
National Science Foundation China (Grant Nos. vol. 84 (9), pp. 878–91.
19. J. Tan and P.C. Pistorius: AISTech 2013 Iron and Steel Technology
51274034 and 51334002), Beijing Key Laboratory of Conference, May 6, 2013–May 9, 2013, Pittsburgh, PA, U.S.A.,
Green Recycling and Extraction of Metals (GREM), Association for Iron and Steel Technology, AISTECH, 2013, vol.
the Laboratory of Green Process Metallurgy and 1, pp. 1301–11.
Modeling (GPM2), and the High Quality steel Consor- 20. P.C. Glaws, R.V. Fryan, and D.M. Keener: 74th Steelmaking
tium (HQSC) at the School of Metallurgical and Eco- Conference Proceedings, ISS, Warrendale, PA, 1991, vol. 74,
pp. 247–64.
logical Engineering at University of Science and 21. M. Iwasaki, N. Suzuki, T. Ohshiro, H. Utsumi, K. Miyake, and K.
Technology Beijing (USTB), China. Sahara: R&D Res. Dev. Kobe Steel Eng. Rep., 1985, vol. 35 (3),
pp. 73–76.
22. R.C. Sussman, M. Burns, X. Huang, and B.G. Thomas: 10th
Process Technology Conference Proceedings, Iron and Steel Soci-
REFERENCES ety, Warrendale, PA, 1992, vol. 10, pp. 291–304.
23. C. Bonilla: 78th Steelmaking Conference Proceedings, ISS, War-
1. L. Zhang and B.G. Thomas: ISIJ Int., 2003, vol. 43 (3), pp. 271– rendale, PA, 1995, vol. 78, pp. 743–52.
91. 24. H. Gao: Steelmaking (Chin.), 2000, vol. 16 (2), pp. 38–43.
2. T. Ehara, Y. Kurose, and T. Fujimura: 79th Steelmaking Confer- 25. M. Goransson, F. Reinholdsson, and K. Willman: I Smaker, 1999,
ence Proceeding, ISS, Warrendale, PA, 1996, vol. 79, pp. 485–86. vol. 26 (5), pp. 53–58.
3. L. Zhang and B.G. Thomas: 7th European Electric Steelmaking 26. Q. Zhang, L. Wang, and X. Wang: ISIJ Int., 2006, vol. 46 (10),
Conference, Venice, Italy, Associazione Italiana di Metallurgia, pp. 1421–26.
Milano, 2002, vol. II, pp. 277–86. 27. http://www.aspexcorp.com/Solutions/Software/MQAtrade.aspx.
4. L. Zhang, B.G. Thomas, K. Cai, L. Zhu, and J. Cui: ISSTech2003, 28. F. Schamber: Introduction to Automated Particle Analysis by
ISS, Warrendale, PA, 2003, pp. 141–56. Focused Electron Beam, ASPEX Corporation, Report, 2009.
5. L. Zhang: Steel Res. Int., 2006, vol. 77 (3), pp. 258–69. 29. V. Singh, S. Lekakh, and K. Peaslee: 62nd SFSA Technical and
6. L. Zhang: J. Iron. Steel Res. Int., 2006, vol. 13 (4), pp. 1–8. Operating Conference, 2008.
7. L. Zhang and B.G. Thomas: Metall. Mater. Trans. B, 2006, 30. The Japan Society for the Promotion of Science: Steelmaking Data
vol. 37B, pp. 733–61. Sourcebook, Gordon and Breach Science, New York, 1988.
8. S. Liu, S. Niu, M. Liang, C. Li, X. Zuo, L. Zhang, and X. Wang: 31. B. Hallstedl: J. Am. Ceram. Soc., 1990, vol. 73 (1), pp. 15–23.
Proceedings of AISTech 2007 Iron & Steel Technology Conference 32. H. Itoh, M. Hino, and S. Ban-Ya: Metall. Mater. Trans. B, 1997,
and Exposition, AIST, Warrendale, PA, 2007, vol. II, pp. 771–80. vol. 28B, pp. 953–56.
9. S. Liu, X. Zuo, L. Zhang, S. Niu, M. Liang, C. Li, and X. Wang: 33. H. Ohta and H. Suito: Metall. Mater. Trans. B, 1997, vol. 28B,
Clean Steel 2007, 2007, pp. 272–82. pp. 1131–39.
10. X. Zuo, M. Long, J. Gao, Y. Wang, and L. Zhang: Iron Steel 34. K. Taguchi, H. Ono-Nakazat, T. Usui, K. Marukawa, K. Katogi,
Technol., 2010, vol. 7 (10), pp. 65–76. and H. Kosaka: ISIJ Int., 2005, vol. 45 (11), pp. 1572–76.
11. Y. Ren, Y. Chen, X. Yang, S. Yang, L. Zhang, X. Ding, J. Li, S. 35. http://www.factsage.com/.
Li, and F. Liu: AISTech 2012 Proceedings, 2012, pp. 1171–77. 36. K. Mukai, H. Sakao, and K. Sano: ISIJ Int., 1969, vol. 9,
12. Y. Ren, L. Zhang, S. Yang, and W. Yang: AISTech 2013 Pro- pp. 196–203.
ceedings, 2013, pp. 1159–66. 37. C. Wang, N. Verma, Y. Kwon, W. Tiekink, N. Kikuchi, and S.
13. R. Kiessling: Met. Sci., 1980, vol. 15 (5), pp. 161–72. Sridhar: ISIJ Int., 2011, vol. 51 (3), pp. 375–81.
14. R. Rastogi and A.W. Cramb: 2001 Steelmaking Conference Pro- 38. D. Bouris and G. Bergeles: Metall. Mater. Trans. B, 1998,
ceedings, ISS, Warrendale, 2001, vol. 84, pp. 789–829. vol. 29B, pp. 641–49.
15. S. Li, L. Zhang, and X. Zuo: Proceedings of Materials Science and 39. V. Brabie: ISIJ Int., 1996, vol. 36 (Supplement), pp. S109–12.
Technology (MS&T) 2008, AIST, Warrendale, PA, 2008, pp. 1259–69. 40. V. Brabie: Steel Res. Int., 1997, vol. 2, pp. 54–60.
16. K. Kawamura, S. Watanabe, and M. Yamada: Tetsu-to-Hagane, 41. M. Imagumbai: ISIJ Int., 1994, vol. 34 (11), pp. 896–905.
1972, vol. 58 (14), pp. 2060–66. 42. Y. Wang and L. Zhang: JOM, 2012, vol. 64 (9), pp. 1063–74.

METALLURGICAL AND MATERIALS TRANSACTIONS B