Paper 11-037.

pdf, Page 1 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA

Calcium Wire Ladle Treatment to Improve

Cleanliness of Centrifugally Cast Steel

E. Y. Martinez, K. D. Peaslee, and S. N. Lekakh

Missouri University of Science and Technology, Rolla, MO

Copyright 2011 American Foundry Society

Most of the alumina inclusions are small and remain in

ABSTRACT
Centrifugal casting provides the unique opportunity of
combining improved melt refining techniques with large
centrifugal forces and directional solidification during
casting to reduce the non-metallic inclusions in the final
product. In this paper, the effects of calcium wire
injection on the number, size, composition and
morphology of non-metallic inclusions in centrifugally
cast products are evaluated. Calcium-bearing wire was fed
into the liquid steel in the ladle prior to centrifugal
casting. Samples were collected throughout the
steelmaking and casting processes starting at argon
oxygen decarburization (AOD) and continuing through
the centrifugally cast product. Samples were analyzed on
the Aspex PICA 1020 for quantitative inclusion analysis.
The effects of calcium treatment combined with
centrifugal casting on inclusions and mechanical
properties are discussed.
Keywords: Calcium-bearing wire, centrifugally cast steel,
automated inclusion analysis
INTRODUCTION
Low alloy cast steels suitable for high pressure service are
often manufactured using centrifugal casting. To meet the
mechanical specifications for pressure vessels, low alloy
cast steels undergo extensive processing steps to improve
steel cleanliness. One primary defect reducing the quality
of steel castings is non-metallic inclusions produced
during the steelmaking and solidification processes. Non-
metallic inclusions degrade the quality or cleanliness of
the cast steel with increasing volume fraction, diameter,
or size, and with uneven or localized distribution, as well
as amorphous morphology. Therefore, the manufacturing
process of low alloy steel must be carefully controlled to
improve the mechanical properties.
the molten steel. The purpose of calcium treatment after
aluminum deoxidation is to modify the composition of
alumina inclusions and form low melting point calcium
aluminates that float at a faster rate and produce cleaner
liquid steel. Low melting aluminates, consisting of
3CaOAl2O3, 12CaO7Al2O3 and CaOAl2O3, melt in the
temperature range of 1395-1605C (2543-2921F) (Figure
1)1. The melting temperature for calcium aluminates is
significantly lower than the alumina inclusions
(2072C[3762F]).
Low melting point non-metallic inclusions are easier to
remove from the molten steel because they have a strong
possibility of recalescence. In addition, Ca inclusion
modification reduces the tendency for the modified
inclusions to attach to the refractory surface preventing
nozzle blockage during teeming and continuous casting.
In addition, manganese and calcium sulfides ((Mn, Ca) S)
can be formed as bi-phase inclusions during calcium
treatment. A bi-phase inclusion is composed of an oxide
prime phase surrounded by a sulfide secondary phase.
Typically, (Mn, Ca) S forms around the CaO Al2O3. The
oxisulfide inclusions form round Type I inclusions,
providing better mechanical properties when compared to
other types of inclusions2.

The cleanliness of cast steel products can be improved

through inclusion flotation. After aluminum deoxidation,
the aluminum oxide rich inclusions start floating to the
surface where they are incorporated into the ladle slag.
However, in industrial practice, the time interval between
aluminum deoxidation of the liquid steel in the ladle and
pouring the mold is not sufficient for inclusion flotation,
especially for small (<2 m) non-metallic inclusions. Fig. 1. CaO-Al2O3 phase diagram.3
 Paper 11-037.pdf, Page 2 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA
EXPERIMENTAL
Medium carbon, low alloy cast steel (0.21 %C, 0.21 %Si,
0.63 %Mn, 2.6 % Cr, 0.12 %Ni, 0.79 %Mo, 0.012 %P,
and 0.001 %S) was melted in an EAF followed by AOD.
After the final additions in the AOD, the molten steel was
centrifugally cast into large diameter tubes (~17 Outside
Diameter, ~12 Inside Diameter, 22 long). Three heats
were tested and analyzed. These heats included the Base
Heat A with no calcium addition, and calcium treated
Heats B and C. Heats B and C were calcium treated using
a steel clad pure calcium wire in the ladle prior to
centrifugal casting. Table 1 shows the composition of the
9 mm diameter calcium wire used during the experimental
calcium treated heats. The calcium wire was injected
using a PC-Micro wire feeder with approximate 17 ft/min
injection speed.
Table 1. Composition (wt. %) of
4
Ca Wire Filler
Ca Al Mg
98.5 0.5 0.5
min. max max
The aim Ca additions were based on treating steel
containing 0.06 wt. % dissolved aluminum [Al]dis in the
ladle to produce low melting CaOAl2O3 and
12CaO7Al2O3 phases. The treatment was done to
decrease inclusions through additional floatation and
modify the shape of the remaining inclusions to enhance
the mechanical properties. Table 2 shows the ladle
capacity for each heat and the aim % calcium added in the
ladle prior to centrifugal casting. Heat B was fed with
relative low calcium content (0.014% Ca) due to the rapid
thermal losses in the ladle during the calcium treatment.
Table 2. Aim % Calcium Added in the Ladle Prior to
Centrifugal Casting and Ladle Capacity.
Heat Caadded, Ladle,
wt.% lbs
A Base heat 10,935
B 0.014 6,785
C 0.077 8,130
Lollipops samples were collected directly from the molten
steel using a submerged sampler for inclusion analysis.
Centrifugally as-cast samples from the end of the finished
tube were also collected for inclusion analysis and impact
testing. A schematic representation of the as-cast tube
samples showing the location from the inside to the
outside of the cast product is given in Figure 2. The
inclusion analysis results of each as-cast tube were
divided into three regions. The three regions represent a
part of a whole (1/4 or 0.25, or 0.50, or 0.75). The
final product region was the average of three samples at
x=0.25, x=0.50 and x=0.75. The inside diameter (ID) at
x=0 and the outside diameter (OD) at x=1 are typically
removed through machining and not incorporated in the
final product.
OD
Width
X=1.00
X=0.75
Thickness
X=0.50
X=0.25
X=0.00
ID
Fig. 2. Schematic representation of the as-cast tube
sample taken from the end of the finished tube
showing the location from the inside to the outside of
the cast product.
The specimens were polished to 1,200 grit and final
polished with 0.1 m diamond paste using the Abramin
auto-polishing machine. An approximate area of 9.8 mm2
was used for the inclusion analysis on each lollipop and
as-cast tube specimen using the Aspex PICA 1020.
Duplication of results was performed by taking the
average value of three tests performed in the same
location of the specimen.
Centrifugally as-cast tube samples from each heat were
heat treated before mechanical testing. The heat treatment
was based on the standard specification for steel castings
suitable for pressure service (ASTM A-487). The
centrifugally cast samples were austenitized at 954C
(1750F), water quenched followed by tempering at 613C
(1135F) and furnace cooling to below 316C (600F).
Charpy impact tests were performed on Heat A (base
case), Heat B (0.014 %Ca) and Heat C (0.077 %Ca) using
three samples for each test.
RESULTS
The number of inclusions/ mm2 and percent area covered
by inclusions for the ladle specimens and final product
(average of the three samples at x=0.25, x=0.50 and
x=0.75) of each as-cast tube sample are displayed in
Figures 3 and 4. Note that there are only limited samples
for some of the locations. Missing samples does not
indicate an absence of inclusions.
The number of oxides/mm2 for all heats, especially the
aluminum containing oxides, regardless of calcium
addition decreased from the AOD to the final product
(Figure 3). Heat B (0.014% Ca) was the case where it
was possible to perform inclusion analysis at all stages of
the ladle treatment and casting process. This heat
demonstrated that the small calcium addition did not
change significantly the population of the oxide
inclusions. The Ca-treatment for Heat B had a minor
effect on the inclusion composition in the ladle. However,
it does demonstrate the decrease in the number of
 Paper 11-037.pdf, Page 3 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA

inclusions from AOD through ladle treatment to the final
product as a result of natural inclusion flotation. In both
Heat B (0.014% Ca), and Heat A (Base Case), an increase
in the number of Mn rich inclusions was observed in the
final product, an indication of reoxidation.
The number and volume of alumina inclusions/mm2
decreased dramatically from the ladle to the final product
for both Heats A (Base Case) and C (0.077% Ca). Heat C
was treated with the highest Ca addition (0.077% Ca) and
exhibited different trends in comparison to the no Ca
treated heat A. Despite the fact of a larger amount of
inclusions after the final additions in AOD for Heat C, the
Ca-treatment modified the inclusion composition.
Simultaneously, the Ca-treatment decreased the number
of inclusions/ mm2 and the total inclusion volume in the
final casting in comparison to the ladle after the final
additions in AOD and before treatment (Figures 3 and 4).
The alumina inclusion population decreased in the final
casting by more than five times.
45
AOD (After Final
40 Additions)
35
No. of Inclusions/ mm2
outside diameter (x=1) (Figures 5 and 6). Theoretically,
the inside diameter of the centrifugally as-cast tube should
contain a larger amount of low-density inclusions than the
rest of the cast product because of the centrifugal forces.
However, it was observed in Figure 5 that the highest
amount of inclusions/mm2 was concentrated in both the
outside diameter (OD) and inside diameter (ID).
The large amount of inclusions, especially at the OD,
might be due the fact that the refractory surface tends to
oxidize at the beginning of the casting. The superficial
oxidation produces carbon monoxide between the
refractory and molten steel interface (Cs +1/2 O2=COg).
The carbon monoxide from the refractory and other
emitted gasses from the molten steel and environment
reoxidize pre-existing particles that form new inclusions.
On the other hand, the number of inclusions/mm2 for heat
C, especially Al2O3 and Mn-Al-O inclusions, decreased
dramatically in the final product of the calcium treated
Heat C (0.077 % Ca). In addition, most of the Al2O3
Ladle (After Ladle (After Final Product
inclusions were transformed into calcium aluminate
Tapping & Ca Treatment) inclusions.
Before Ca
Treatment) CaO

30 Ca-Al-O
Machining Allowance: x=0
25 Mn-Si-O
300 CaO
20 Mn-Al-O
No. of Inclusions/mm2

Ca-Al-O
250
15 Al2O3
MnSiO
10 200
Mn-Al-O
5 150 Al2O3
0
100

50

0
Heat A (Base Case) Heat B (0.014% Ca) Heat C (0.077% Ca)

2
Fig. 3. Comparison of the inclusions number per mm
from tap to the final product in different heats. 300
Machining Allowance: x=1

CaO
No. of Inclusions/mm2

250
0.060 Ca-Al-O
AOD (After Ladle (After Ladle (After Final Product 200 MnSiO
0.050
Final Tapping & Ca Treatment)
Additions) Before Ca CaO 150 Mn-Al-O

Treatment)
% Area Covered by Inclusions

0.040 Ca-Al-O Al2O3

100
Mn-Si-O
0.030 50
Mn-Al-O

0.020 0
Al2O3
Heat A (Base Case) Heat B (0.014% Ca) Heat C (0.077% Ca)
0.010

0.000 Final Product: x=0.25, x=0.50 and x=0.75

300
CaO
No. of Inclusions/mm2

250 Ca-Al-O

200 MnSiO

Mn-Al-O
150
Fig. 4. Comparison of the percent area covered by Al2O3

100
inclusions from tap to the final product in different
heats. 50

0
Inclusion analysis was performed on the as-cast tube Heat A (Base Case) Heat B (0.014% Ca) Heat C (0.077% Ca)

samples which were split into regions representing the

final product (average of the three samples x=0.25, Fig. 5. Average number of inclusions per mm2 at
x=0.50 and x=0.75), the inside diameter (x=0) and the different locations in the as-cast tube samples.
 Paper 11-037.pdf, Page 4 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA

Machining Allowance: x=0 indicates that Ca-treatment provided better mechanical

MnS
0.10 properties through the modification and transformation of
CaO
0.09
detrimental non-metallic inclusions.
% Area Covered by Inclusions

0.08 Ca-Al-O

0.07 Mn-Si-O
0.06
Mn-Al-O
0.05 94
Al2O3
0.04
92
0.03

Impact Energy (ft*lb)

0.02 90
0.01
0.00 88 Heat A (Base
Heat A (Base Case) Heat B (0.014% Ca) Heat C (0.077% Ca) Case)
86 Heat B (0.014
% Ca)
84 Heat C (0.077
Machining Allowance: x=1 % Ca)
MnS
0.10
CaO
82
% Area Covered by Inclusions

0.09
Ca-Al-O 80
0.08
0.07 Mn-Si-O 27 27 28 28 29 29 30
0.06 Mn-Al-O HRC
0.05 Al2O3
0.04
Fig. 7. Comparison of the Charpy impact energy and
0.03
0.02
the HRC for the three heats.
0.01
0.00 MODELING
Heat A (Base Case) Heat B (0.014% Ca) Heat C (0.077% Ca)

During solidification, centrifugal castings experience

Final Product: 0.25, x=0.50 & x=0.75 forces that are orders of magnitude larger than the force of
MnS
0.10
0.09
CaO gravity that is experienced in static sand molds. This
% Area Covered by Inclusions

0.08 Ca-Al-O difference can be presented by the Gfactor which is

0.07 Mn-Si-O calculated using the peripheral velocity (V, m/sec),
0.06
0.05
Mn-Al-O casting radius (r, m), and gravity acceleration (g, m/s2):
Al2O3
0.04
0.03
0.02
0.01
Gfactor=V2/rg Equation 1
0.00
Heat A (Base Case) Heat B (0.014% Ca) Heat C (0.077% Ca)

The Gfactor could vary from 50 to 200 for large

Fig. 6. Percentage area covered by inclusions at centrifugally cast tubes having a tremendous effect on
different locations in the as-cast tube samples. non-metallic inclusion flotation. Figure 8 shows the
solidification kinetics using a tube wall thickness of 100
Centrifugally cast tube samples were heat treated prior to mm (4) and an internal layer of packed sand. The packed
Charpy impact testing. Heat treatment included sand was 25.4 mm (1) thick and the steel flask was 25.4
austenitization at 954C (1750F), water quenching mm (1) thick. The dashed line represents the position of
followed by tempering at 613C (1135F), and furnace the liquidus whereas the solid line exhibits the solidus.
cooling to below 316C (600F). Average Charpy impact The solidification kinetics is plotted versus the cooling
results (Table 3) show that the impact energies for Heats time of the molten steel poured at 50C superheat. Most
B (0.014% Ca) and Heat C (0.077% Ca) were larger than of the casting volume solidified unidirectional from the
the base Heat A. This means that some degree of OD with development of a wide mushy zone. The rest of
transformation and reduction of non-metallic inclusions the casting volume, which is composed of a very thin
by Ca-treatment positively influenced the impact energy. layer, solidified from the inside diameter (ID) during the
final stage of the casting process. These conditions can be
Table 3. Average Charpy Impact Energy (ft*lb) of Steel varied depending on the intensity of the heat transfer at
in the Three Calcium Injection Experiments. the ID. For example, a thin layer of slag forming on the
Heat A (Base Heat B (0.014% Heat C (0.077% ID can provide sufficient insulation to allow full
Case) Ca) Ca)
unidirectional solidification.
83.1 (SD 1.01) 89.7 (SD 3.57) 84.9 (SD 1.05)

Figure 7 is a quality assessment of the centrifugally cast

products after heat treatment and Charpy impact testing.
The combination of higher hardness (strength) and higher
impact values for Heats B and C indicates a better overall
quality combination than the untreated Heat A. This
 Paper 11-037.pdf, Page 5 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA

100 inclusions growing in the liquidus-solidus mushy zone

(endogenous inclusions) would be trapped in the casting.
The main hydrodynamic reason for flotation deceleration
Distance from OD, mm

80 Liquid
of non-metallic inclusions is the rapid increase of
60 viscosity in the mushy zone. When the liquid phase
contains some amounts of solid phase, the effective (e) of
40 Mushy
the system will increase. The effective or apparent
Solid viscosity (e) of a heterogeneous system depends
20
significantly on the fraction of a solid phase (f), as well as
the liquid phase viscosity (). These parameters can be
calculated using the Einsteins5 (solid line), Jefferys6
0
0 1000 2000 3000 4000 5000 6000 7000
(dashed line) equations or the equation used by Pretorius7
Time, sec (dotted line) in Figure 10.
0.06
Fig. 8. Solidification kinetics in a centrifugally cast
tube using a wall thickness of 100 mm (4). [1]
0.05

Dynamic viscosity, kg/ms

[2]
The cooling time from Figure 8 was used for flotation 0.04 [3]
analysis on non-metallic inclusions during the casting
process. Fluent Software was used to model two- 0.03

dimensional inclusion flotation. A stagnant melt was

0.02
chosen after pouring into a steel flask mold using the
Discrete phase model. Figure 9 shows flotation 0.01
dynamics assuming spherical non-metallic inclusions
through a wall thickness of 100 mm (4). The stagnant 0
melt was cooled in a sand mold (1G) and in centrifugal 0 0.1 0.2 0.3 0.4

mold (100G). The period for superheat removal, in which Solid fraction

the melt is above the liquidus, is also shown by the shaded Fig. 10. Effect of solid phases on the apparent
area in Figure 9. At this period of time, the non-metallic dynamic viscosity (e) of molten steel.5, 6, 7
inclusions smaller than 10 m diameter were not removed
from the sand casting. In contrast, centrifugal forces The estimated effect of the apparent liquid phase viscosity
accelerated inclusion flotation and particles with a () on inclusion flotation in the mushy zone in the
diameter larger than 2 m were removed completely from centrifugally cast tube at 100G is given in Figure 11. The
the melt in a short period of time. These particles were simulated wall thickness was 100 mm (4) and the
removed before the melt could cool from the pouring diameter of the non-metallic inclusions was 5 m. The
temperature to the liquidus temperature. In reality, some data shows that even under the influence of large
melt motion is produced in the sand mold due to thermal centrifugal forces, the flotation of inclusions is
convection. In centrifugal casting, the melt can be dramatically decreased in the mushy zone due to the
displaced through mold vibration. In both cases, the apparent liquid phase viscosity () increase. This
particle movement forces are less when compared to the indicates that a significant number of small non-metallic
simulated centrifugal forces at 100G. inclusions would not float to the tube ID but be trapped in
1
the casting due to the solidified melt with a large apparent
liquid phase viscosity ().
0.8
Rem oved inclusions

5 ( 1G )
1
0.6 1 0 (1G )
5 ( 100 G )
0.8
Removed inclusions

2 ( 100 G )
0.4
0.005
0.6
0.2
0.05
0.4
0
0 5 00 1 0 00 1 5 00 20 00
Time , s ec 0.2

Fig. 9. Fraction of non-metallic inclusions removed by 0

floatation with diameters from 2 m to 10 m in sand 0 200 400 600 800
mold (1G) and centrifugal casting (100G). Time, sec

Fig. 11. Effect of the apparent liquid phase viscosity

If the centrifugal forces have the possibility of non- (= 0.005 kg/ms without solid fraction, ap= 0.05 kg/ms
metallic inclusion removal from the molten steel when with 25-30% solid fraction) on flotation dynamics
above the liquidus in a short period of time, only the using non-metallic inclusions of 5 m in diameter.
 Paper 11-037.pdf, Page 6 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA

Calculation of the non-steady state solidification with
different apparent liquid phase viscosities () in the
mushy zone of the centrifugal cast product was performed
in Figure 12. Non-metallic inclusions 5 m in diameter
formed at different temperatures were evaluated. The non-
metallic inclusions located in the casting region at a
temperature above the liquidus had sufficient time to be
released while a significant amount of non-metallic
inclusions located in the mushy zone were trapped in the
centrifugally cast tube.
Group 2 includes endogenous non-metallic inclusions
growing directly in the molten steel, which could be
subdivided into three subgroups:
a) Prime non-metallic inclusions, formed in the molten
steel as a result of de-oxidation, de-sulfurization and other
reactions with active components in the melt (e.g.
nitrides). The prime non-metallic inclusions are smaller
than the exogenous group. The prime nonmetallic
inclusions are forced to the ID and OD due to the
centrifugal forces in the casting process.

b) Re-oxidation products, formed when the molten steel is

exposed to air during mold filling. Typically, these
inclusions are small and can be removed from the melt
before formation of the mushy zone, while some part of
these inclusions could be trapped near the OD surface of
centrifugally cast tube.

c) Secondary non-metallic inclusions, formed near the

liquidus temperature and inside the mushy zone. Two
main mechanisms can promote the formation of
secondary inclusions. The first one is due to the
thermodynamic affinity of free radicals in the molten steel
that interact with the remained impurities (sulfur and
a) oxygen). The second mechanism includes inclusion
segregation of free radicals in the remaining portion of the
molten steel during dendrite solidification. Figure 13
shows the inclusion segregation mechanism for MnS
using the Scheil model. It was observed that more stable
inclusions (Al2O3, Al-Ca-O and CaS) were mainly formed
above the liquidus temperature and their concentrations
did not significantly changed during solidification. Sulfur
segregation and increasing activity of Mn and Si radically
promoted MnS, Mn-Si-O and Mn-Si-Al-O complex
inclusion in the mushy zone during solidification. This
subgroup of inclusions could be trapped by the mushy
zone located in the centrifugally cast tube.
b)
Fig. 12. Initial distribution of 5 m non-metallic 100 0.08
inclusions located in the centrifugal cast tube (a) and Al-Ca-O
CaS 0.07
calculated final distribution of 5 m non-metallic %
80 ,t
inclusions (b) in the centrifugal cast tube (3000 sec). h
0.06 g
% i
t Al2O3 e
h w
DISCUSSION ig
e 60 Liquid
0.05
,
s
n
w
, io
e Prime (Al) s
s 0.04 u
l
Based on the experimental results, modeled solidification a Prime (Al+Ca) c
h
p 40 in
kinetics and flotation dynamics, all types of non-metallic s
i
Secondary (Al)
0.03 ic
ll
inclusions in the centrifugal casting process could be u MnS Secondary (Al+Ca) a
t
iq Al-Si-O e
L
classified into two main groups: Mn-Si-O 0.02 -m
20 Mn-Al-Si-O n
Group 1 comprises exogenous non-metallic inclusions, o
N
Al2O3 0.01
from different outside melt sources including large Al-Ca-O
particles from the slag, refractory and product reactions 0 0.00
with mold materials. This group of inclusions may not 1450 1500 1550 1600
affect the mechanical properties of the centrifugally cast Temperature, 0C
tube because they will be forced to the ID or OD due to
Fig. 13. Thermodynamic prediction (Scheil model) of
the centrifugal forces in the casting process. The ID and complex inclusion formation (MnS, Mn-Si-O and Mn-
OD of the centrifugally cast tube are usually removed by Si-Al-O) in the mushy zone during solidification
secondary machining. (FACTSAGE software).
 Paper 11-037.pdf, Page 7 of 7
AFS Proceedings 2011 American Foundry Society, Schaumburg, IL USA

The sand casting products may be composed of a forced through intensive flotation from the melt to the ID
combination of exogenous and endogenous non-metallic and OD, leaving behind the secondary non-metallic
inclusions in the final product. Figure 14 shows that the inclusions in the mushy zone of the cast tube. These
centrifugal forces could drive most of the exogenous, processes were also simulated.
prime and re-oxidation non-metallic inclusions to the ID
and OD of centrifugally cast tube. This reduces Large calcium additions in the molten steel prior to
significantly the amount of these detrimental inclusions in casting (based on the aluminum containing oxides), less
the final product of the centrifugal cast tube, especially injection time, better mold refractory, less fluid
when the molten steel is calcium treated prior to casting. turbulence in the pouring basin, among other factors will
Therefore, the only type of inclusions formed in the final increase the potential for success in transforming alumina
product of the centrifugally cast tube is the secondary inclusions into lower melting calcium aluminates
non-metallic inclusions. (CaOAl2O3 and 12CaO7Al2O3) to enhance the impact
properties of the centrifugally cast tube.
0.1
Sand casting (Al +Ca)
ACKNOWLEDGEMENTS
Area covered by inclusions, %

Centrifugal casting (AOD, Al+Ca)

0.08
Forged steel The work for this project was made available through
0.06
funding provided by U.S. Army Benet Labs Award
W15QKN-07-2-0004 and the funding for the ASPEX
inclusion analyzer was made available through U.S. Army
0.04
DURIP Grant W911NF-08-1-0267. The Ca-wire feeder
was provided through a grant from P.C. Campana. The
0.02
authors also acknowledge the support of the Steel
Founders Society of America and the member companies
0 that participated in this research.
AOD Al-killed Ca-treated Final product

Fig. 14. Comparison of the evolution of non-metallic

inclusions during sand mold gravity casting process REFERENCES
and AOD-Ca treated centrifugal casting process.
1. Finardi, J., Effect of Calcium Treatment of
CONCLUSIONS
Steel for Castings, AFS Transactions, pp. 409-
417 (2003).
Inclusion analysis and impact testing was performed on
2. Lynch, D.W.P., and Robison, J.W., Jr., Calcium
centrifugally cast large diameter tubes. The effects of the
wire ladle treatment: Cleanliness and ductility
Ca-wire ladle treatment prior to casting using 0.014% and
improvement of foundry steel, Pfizer Inc., pp.
0.077% Ca were compared to the base heat with no
303-311.
calcium treatment. It was observed that the Ca treatment
3. Singh, V., Inclusion modification in steel
modified to some extent the types of inclusions and
castings using automated inclusion analysis,
influenced the inclusion population at different locations
[MS Thesis]: Missouri University of Science and
of the centrifugally cast tube. In particular:
Technology (2009).
4. Alloy cored wire specifications, PC Campana
The Al2O3 and Mn-Al-O inclusions were reduced and
(2006).
transformed into low melting calcium aluminate
5. Einstein, A., A new determination of molecular
inclusions in the final product of Heat C (0.077% Ca)
dimensions, Ann. Phys., 19 (1906).
whereas there was minor inclusion modification in Heat B
6. Jeffery, G. B., The motion of ellipsoidal
with a low calcium ladle addition (0.014% Ca). The
particles immersed in a viscous fluid, Proc.
majority of the percentage area covered by inclusions in
Roy. Soc. A., 102, 161-179 (1922).
the final product of Heat C (0.077% Ca) corresponded to
7. Pretorius, E. B. and Carlisle, R. C., Foamy Slag
modified Ca-Al-O inclusions.
Fundamentals and Their Application to Electric
Furnace Steelmaking, Proceedings of the 56th
Both Ca-treated heats showed increased toughness while
Electric Furnace Conference, New Orleans, LA,
the impact energy for Heat B (0.014% Ca) was larger
pp. 275-291 (1998).
than heat C (0.077% Ca) due to independent effect of
refining practices in AOD.

The Al2O3 and Mn-Al-O inclusions were increased in

terms of number and near the OD and ID locations for
calcium treated heats. This indicates that the exogenous,
prime and re-oxidation non-metallic inclusions were