INCLUSION ENGINEERING FOR IMPROVED PROPERTIES

IN STEEL CASTINGS

Kent D. Peaslee1, Vintee Singh

2 and Simon N. Lekakh

1

1Missouri University of Science & Technology

2Nucor Steel - Hickman

ABSTRACT

One of the most important factors affecting the toughness of steel castings is steel cleanliness.

Steel cleanliness is determined by the number, size, shape, and composition of the final

nonmetallic inclusions in the steel matrix. This paper reports on the effects of changing the melt

and ladle practices (calcium treatment, argon stirring, etc.) in steel foundries on the cleanliness

and resulting toughness of the final steel castings. Practice changes are correlated with inclusion

characterization using automated inclusion analysis and with final casting properties. FLUENT

modeling was used to understand the effects of various ladle parameters on the inclusion

flotation. This paper will correlate the toughness of steel castings with inclusion volume, shape

and spacing as influenced by ladle treatment before casting.

KEY WORDS: inclusions, reoxidation, toughness, calcium treatment

INTRODUCTION

Higher quality steel is being required in many new applications forcing steelmakers to ensure that

their steel products meet more stringent “cleanliness” standards. Mechanical properties are

affected by the volume fraction, size, distribution, composition, and morphology of inclusions.

The toughness of steel is very important in many critical applications where fracture resulting in

failure could produce catastrophic effects. Hence, the determination of the exact composition,

morphology and location of non-metallic inclusions is essential to the success of research aimed

at increasing toughness of steel parts.

Most inclusions in steel castings are a product of deoxidation in the ladle or reoxidation during

processing. During deoxidation, the dissolved oxygen content of the steel is reduced by adding

elements that have a greater affinity for oxygen than carbon and form thermodynamically more

stable oxides than iron oxide. The most common deoxidizer for steel castings is aluminum,

which produces solid particles of Al2O3. Alumina inclusions are dendritic when formed in a high

oxygen environment such as reoxidation and often coalesce to create irregular shaped “alumina

clusters” as a result of the collision of smaller particles1. These clusters significantly affect the

mechanical properties of steel, especially fracture sensitive properties such as toughness and

fatigue life, and may also result in the generation of surface defects2.

Exogenous inclusions arise from unintentional chemical and mechanical interaction of liquid steel

with its surroundings. They are generally deleterious to fracture sensitive mechanical properties

because of their large size and location near the surface3. The majority of these inclusions are

formed by reoxidation in which liquid steel, having "free" deoxidants (Al, Si, Mn or Ca) dissolved

in the molten steel, picks up oxygen from contact with the air during pouring and transportation

through the gating system. In addition, inclusions can be formed by reaction of the liquid steel

with gases or water vaporizing from the molding sands and debris in the gating system.

Exogenous inclusions are usually entrapped accidentally during tapping, pouring and

solidification resulting in a random distribution throughout steel castings. These inclusions act as

heterogeneous nucleation sites for precipitation of new inclusions during their motion in molten

steel4.

In a cast steel product, non-metallic inclusions are one of the primary sites at which void

nucleation occurs. Voids nucleated at an inclusion, either by decohesion of the inclusion-steel

interface or by inclusion fracture, grow until they coalesce by impingement or by the process of

void sheet coalescence.5 Void sheet coalescence requires fracture of the ligament between the

voids created at the larger non-metallic inclusions. Fracture toughness is the ability of a material

to arrest an existing crack and prevent the onset of rapid crack propagation at stress levels below

the yield stress.6 The characteristic inclusion volume fraction and the inclusion spacing have been

shown to greatly influence the fracture toughness of steel. Recent research has shown that

decreasing the volume fraction of inclusions that induce void nucleation and increasing the

inclusion spacing result in significant improvement in toughness.7 Hahn et al

8 found that fracture

toughness of steel (K1C) inversely varied with the volume fraction of the inclusions according to

Equation 1:

EdVK yieldfI 3

1

6

1

62

(1)

where: KI is fracture toughness, Vf is volume fraction of the inclusions, σyield

is

yield strength, E

is Young’s modulus, and d is particle diameter.

The fracture toughness varies directly with the inclusion spacing according to Equation 2:

21

1 )2( EsK yield

(2)

where: K1 is fracture toughness, σyield is

yield strength, E is Young’s modulus, and s is average

spacing between the inclusions.

One technique used to diminish the harmful effects of inclusions in steel is calcium treatment.

When Ca-treatment is effective, alumina inclusions are converted to molten calcium aluminates

which are globular in shape. The calcium aluminate inclusions retained in the steel suppress the

formation of harmful MnS inclusions during the solidification of steel by modifying MnS

inclusions to spherical CaS inclusions. When alumina is modified to calcium aluminate, the

reaction sequence with additional calcium additions (illustrated in Figure 1) is: Al2O3 -> CA6 ->

CA2 -> CA -> C12A7 , where C and A denotes CaO and Al2O3, respectively. The presence of

liquid calcium aluminates, CA2, CA, C12A7 at steelmaking temperatures (~ 1600 oC) results in

inclusions that are much easier to float than the solid alumina inclusions and also reduce the

tendency of blocking ladle and casting nozzles.

Although there has been a significant work done in larger steelmaking shops, very little research

has been done in proving the effects of calcium treatment in smaller ladles (less than 10 tons).

The purpose of this research was to investigate the effectiveness of calcium treatment in steel

foundries and determine the benefits on toughness. In order to increase the speed and efficiency

of inclusion evaluation, this research uses an automated inclusion analyzer, ASPEX PICA-1020.

Figure 1: Change in stable inclusion composition in Al-killed medium-carbon cast steel

with Ca-treatment, calculated with FACTSAGE software at 1600°C.

This equipment provides a rapid and accurate method for determining the composition, size,

number, spacing and distribution of inclusions present in the steel samples. It is an integrated

SEM and EDS system and allows for automated characterization of all the inclusions (1 μm to 5

mm) in a microscopic specimen including the volume fraction, size and shape, and complete

inclusion identification. In order to measure the spacing between the inclusions, all of the x-y

inclusion coordinates provided by the ASPEX inclusion analyzer were transported to a Microsoft

Excel file and a code written in Visual Basic to calculate the distance of an inclusion from each of

the other inclusions and determine the distance to the closest neighbor. These distances are

averaged over all inclusions to determine the average spacing between the inclusions in the

specimen in the final inclusions analysis.

CFD MODELING OF INCLUSION BEHAVIOR

Computational fluid dynamic (CFD) modeling using commercial software (FLUENT CFD

Version 6.3.26) was performed to predict inclusion flotation in typical foundry ladles as affected

by ladle capacity, inclusion type, inclusion size and argon stirring. Non-steady state heat transfer

was solved by using a “coupled” thermal boundary between the liquid steel (1600°C) and the

preheated ladle lining (900°C) assuming that there was radiation and convection from the open

steel surface at the top. Free convection flow of the liquid metal was solved using the Boussinesq

density model for thermally expanded liquids. Inclusions were injected into the liquid steel

through the central vertical plane. Heat transfer and flow equations were solved by applying the

specific drag forces assuming that solid particles were either spherical or non-spherical (shape

factor of 3.0). To simulate the formation of liquid inclusions (calcium injection), inclusions were

introduced in the liquid metal as liquid droplets with the possibility of collision and coalescence.

It was assumed that the injected particles were reflected back into the metal from the bottom of

the ladle while the particles escaped reaching the top or the sides of the metal surface.

The flotation of different size spherical inclusions by gravity forces alone and by natural

convection in a ladle is compared in Figure 2. It can be observed that although large spherical

particles float and escape in approximately the same time for both the cases, micron-sizes

particles were not floated out by gravity, but were partially removed from the liquid metal by

natural convection.

0

0.01

0.02

0.03

0.04

0.05

0 0.01 0.02 0.03 0.04 0.05 0.06

Incl

usi

on

s fo

rme

d, w

eig

ht

%

Ca additive, weight %

Al2O3

MnS

CaS

CACA6

CaO

C3A

CA2

CA

Figure 2: Comparison of inclusion flotation due to gravity and natural convection in the ladle

Figure 3a shows the effect of the inclusion shape factor and density on inclusion flotation in the

ladle, in the presence of natural convection in the liquid steel. Particles with lower density floated

faster than higher density particles. Non-spherical particles floated faster than spherical particles

because of the drag force, which is greater on the non-spherical particles and helps particles float

out under natural convection. Liquid droplets resulted in the highest flotation rate. As the number

of liquid calcium aluminate droplets increase with time, the possibility of droplet collision and

coalescence also increase with time. Droplet coalescence results in larger particles and a higher

flotation rate. Therefore, calcium treatment that modifies solid alumina and forms liquid calcium

aluminate inclusions has the opportunity of reducing the total number of inclusions.

The flotation of the injected spherical particles in the molten metal, in the presence of only

natural convection, for different ladle capacities is shown in Figure 3b. Ladle size has a negligible

effect on particle flotation in the model. Larger ladles had longer flotation distances compared to

the smaller ladle, but more intensive convection, thus reducing the flotation time. Because of

these two factors, the limited hold-time of the molten steel in the ladle is the main factor limiting

inclusion flotation.

The CFD modeling suggested two possible methods for increasing steel cleanliness. The first

method is modification of solid inclusions into liquid droplets by changing the composition

during Ca-treatment. The second method is to increase the intensity of the flow of the molten

metal by Ar-stirring in the ladle. Figure 4 shows the particle motion of spherical inclusions (50

µm diam.) in a ladle during Ar-stirring from the bottom. It can be seen that most of the particles

escaped the ladle after a 60 second Ar-stir. Figure 5 compares three mechanisms of inclusion

flotation, gravity, natural convection and Ar-stirring showing graphically the advantage of Ar-

stirring. The applied model did not take into account the attachment of particles to the bubble

surface which would increase the flotation efficiency.

EXPERIMENTAL PROCEDURE

Five heats (10 ladles) trials were completed at an industrial foundry to determine the effects of

Ca-wire feeding and Ar-stirring in the ladle on inclusion content in medium carbon cast steel. The

steel was melted in a one ton induction furnace and tapped twice into 1000 lb ladles. A

submerged lance was used for Ar-stirring in the ladle.

0

20

40

60

80

100

0 1 2 3 4 5

Par

ticl

e n

um

be

r (%

)

Time (min)

0.005 mm (convection)

0.005 mm (gravity)

0.2 mm (convection)

0.2 mm (gravity)

1 mm (convection)

1 mm (gravity)

9100 kg ladle

a) b)

Figure 3: Effect of a) particle geometry, density and form and b) ladle size on the inclusion

escape rate under the influence of natural convection

1 sec 4 sec 20 sec 60 sec

Figure 4: Effect of Ar-stirring (3.5 cfm in center of bottom) on distribution of 50 µm spherical

inclusions in a 9100 kg capacity ladle

Figure 5: Comparison of inclusion flotation rates in a ladle under different flotation mechanisms

0

10

20

30

40

50

60

70

80

90

100

40

50

60

70

80

90

100

0 1 2 3 4

Co

ale

sce

d d

rop

lets

(%

)

Pa

rtic

le n

um

be

r (%

)

Time (min)

Spherical solid

Nonspherical solid (factor 3)

Escaped coalesced droplets

Spherical solid, 1.7 g/cm3

Coalesced droplets

9100 kg ladle, 0.005 mm

40

50

60

70

80

90

100

0 2 4 6 8 10

Time (min)

Per

cen

t o

f p

arti

cle

rem

ain

ing

60 kg

335 kg

2700 kg

9100 kg

0.005 mm

inclusins

60 kg

335kg

2700kg

9100kg

2.7 g/cc

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

Furnace

before

deoxidants

Furnace after

Al addition

Ladle after

deoxidants

Ladle end of

pour

Final CastingFra

cti

on

of

are

a c

ov

ere

d b

y in

clu

sio

ns

others

TiO2

CA

Al2O3

MnO

MnSiO3

CaS

MnS

SulfidesSulfidesSulfidesSulfides

Oxides

OxidesOxides

Oxides

Sulfides

Oxides

After the steel was melted in the induction furnace, it was deoxidized with an addition of 0.04

weight % Al and alloyed to meet the final chemistry specification (aim in wt%: 0.25%C,

0.75%Mn, <0.01%P, <0.020%S, and 0.40%Si). During tap, an additional 1 lb of Al, 0.66 lbs of

Fe70Ti and 0.88 lbs of Fe51Si35Zr were added into the 1000 lb ladle. Ladle treatment options

included CaSi wire additions, Ca-wire additions Ar-lance stirring in the ladle for 1 min, and

combined Ar-stirring and CaSi wire additions.

Steel chemistry samples were collected from the furnace, ladle and mold using submerged

chemistry samplers. The samples were collected before and after deoxidation and post-tap

treatment. In addition, samples were cut from standard keel blocks cast from the same ladle.

Microscopic specimens were prepared from these samples and a 10 mm2 area was analyzed in

each specimen for inclusions using the Aspex PICA-1020 automated inclusion analyzer.

In order to conduct Charpy impact testing, cast keel blocks were poured during each trial. The

keel blocks were normalized by holding at 1650°F (900°C) for 1.5 hours and then air-cooled.

Charpy V-notch specimens of dimensions 10mm×10mm×55mm were prepared according to

ASTM E23 standards. The Charpy impact tests were conducted at a temperature of -40°C.

RESULTS AND DISCUSSION

Standard Practice: One ladle was treated using the standard melting practice with no additional

post-tap treatment. Figure 6 compares the total and dissolved oxygen and the inclusion area

fraction at the various stages of liquid processing. The dissolved oxygen dropped after

deoxidation, resulting in the formation of a large number of oxide inclusions and an increase in

the total oxygen. The total area of the oxide inclusions and the total oxygen increased during the

pour and also the casting had more inclusions than in the ladle, suggesting that there is significant

reoxidation during pouring and liquid steel transport. The area of Al2O3 and TiO2 inclusions

increased after the Al and FeTi additions in the ladle. The average size of the TiO2 inclusions in

the samples was 1.7 μm, significantly smaller than alumina inclusions which averaged 2.8 μm in

diameter.

a) b)

Figure 6: a) Oxygen content and b) area of inclusions by composition in the steel at various

stages of the casting process (standard practice with no Ca-wire or Ar-stir)

Ca-Treatment: To study the effects of Ca-treatment, ladles were fed with either CaSi-wire or

pure Ca-wire in the ladle. Figure 7 compares the total and dissolved oxygen in the liquid steel

and the area fraction covered by inclusions in calcium-treated steel (0.028 wt. % Ca added using

CaSi-wire). After Al-treatment in the furnace, there was an increase in the volume of alumina

0

20

40

60

80

100

120

140

160

180

200

Furnace

before

deoxidants

Furnace after

Al addition

Ladle after

deoxidants

Ladle end of

pour

Final Casting

Am

ou

nt

of

oxyg

en

(p

pm

)

Dissolved O

Total O

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Wt% of Ca added

Fra

cti

on

of

are

a c

overe

d b

y i

nclu

sio

ns

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

Ladle after

deoxidants

Ladle before

CaSi

Ladle after

CaSi

Ladle end of

pour

Final CastingFra

cti

on

of

are

a c

ov

ere

d b

y i

nc

lus

ion

s

others

TiO2

CA

Al2O3

MnO

MnSiO3

CaS

MnS

SulfidesSulfidesSulfidesSulfides

OxidesOxides

Oxides

Oxides

Sulfides

Oxides

0

20

40

60

80

100

120

140

160

180

200

Ladle after

deoxidants

Ladle before

CaSi

Ladle after

CaSi

Ladle end of

pour

Final Casting

Am

ou

nt

of

oxyg

en

(p

pm

)

Dissolved O

Total O

0

20

40

60

80

100

120

140

160

180

200

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

Wt% of Ca added

Am

ou

nt

of

ox

yg

en

(p

pm

)

inclusions and the total oxygen. The composition and number of inclusions changed after the

CaSi-treatment in the ladle with most of the alumina inclusions forming calcium aluminates (CA)

and the MnS inclusions converting to CaS inclusions. The Si added in the CaSi-wire promoted the

formation of MnSiO3 inclusions. The Ca-wire addition also dropped the total oxygen content.

a) b)

Figure 7: a) Oxygen content and b) area of inclusions by composition in the steel at various

stages of the casting process (0.028% Ca in CaSi-wire, no Ar-stir)

The amount of Ca-added was varied between 0.02 and 0.06 wt% with no Ar-stirring during the

trials. Figure 8 shows the effect of varying the Ca-addition on the total oxygen content and the

amount of inclusions, as measured from the keel blocks obtained during the various heats. Both

the total oxygen and the volume of inclusions decreased linearly with increasing amounts of Ca-

added. Thus, Ca-treatment was shown to be an excellent method of improving the cleanliness of

cast steel products. Also, it can be observed that the actual zero point of both curves, that is, the

amount of inclusions or total oxygen with no Ca-addition, is above the best-fit line projected to

zero calcium. This provides evidence that the initial addition of calcium significantly decreases

inclusions.

a) b)

Figure 8: Effect of Ca-wire additions on a) total oxygen and b) the inclusion fraction

Ar-stirring: In one trial ladle, no Ca was added and Ar-stirring was performed with a submerged

lance for 1 minute. As observed in Figure 9, Ar-stirring increased inclusion floatation with the

volume of inclusions and the total oxygen decreasing after the Ar-stirring. This observation is

supported by the CFD modeling discussed earlier.

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

Furnace after Al

addition

Ladle after

deoxidants

Ladle after Ar-

stirring

Ladle end of

pour

Mold Final Casting

Fra

cti

on

of

are

a c

overe

d b

y in

clu

sio

ns

others

TiO2

CA

Al2O3

MnO

MnSiO3

CaS

MnS

sulfidessulfidessulfidessulfidessulfides

oxidesoxides

oxides

oxides

oxides

oxides

sulfides

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

Ladle after

deoxidants

Ladle

before CaSi

Ladle after

CaSi

Ladle after

Ar-stirring

Mold Final

Casting

Fra

cti

on

of

are

a c

overe

d b

y in

clu

sio

ns

others

TiO2

CA

Al2O3

MnO

MnSiO3

CaS

MnS

SulfidesSulfidesSulfidesSulfidesSulfides

OxidesOxides

Oxides

OxidesOxides

Sulfides

Oxides

0

20

40

60

80

100

120

140

160

180

200

Furnace after Al

addition

Ladle after

deoxidants

Ladle after Ar-

stirring

Ladle end of

pour

Mold Final Casting

Am

ou

nt

of

ox

yg

en

(p

pm

)Dissolved O

Total O

0

20

40

60

80

100

120

140

160

180

200

Ladle after

deoxidants

Ladle

before CaSi

Ladle after

CaSi

Ladle after

Ar-stirring

Mold Final

Casting

Am

ou

nt

of

oxyg

en

(p

pm

)

Dissolved O

Total O

a) b)

Figure 9: Effect of Ar-stirring on a) dissolved and total oxygen and b) inclusion fraction

To study the combined effect of Ca-treatment with Ar-stirring, a ladle trial was conducted in

which CaSi-wire (0.043 wt. % Ca) was added in the ladle followed by Ar-stirring for 1 minute

(see Figure 10). As observed in the previous cases, both Ar-stirring and Ca-additions helped in

inclusion removal. The dissolved oxygen also decreased after Ar-stirring which indicates that the

Ca-reaction with the liquid steel continued during stirring. But a greater decrease in inclusions

and oxygen was observed in the ladle sample immediately after Ca-addition, as compared to the

sample taken after Ar-stirring. This shows that Ar-stirring is not as effective as Ca-treatment for

inclusion removal, but the overall combination of both Ca-addition and Ar-stirring was more

beneficial for inclusion removal than either alone.

a) b)

Figure 10: Effect of CaSi-wire addition (0.043 wt. % Ca) and Ar-stirring in the ladle on

a) dissolved and total oxygen and b) volume of inclusion by composition

Comparison of all the treatments: Figure 11 compares the total oxygen content and the volume of

inclusion and composition for all of the heats, as measured in samples taken from cast keel

blocks. CaSi-wire additions were effective in reducing the alumina and MnS inclusions and

forming calcium aluminate and CaS inclusions. Better results were achieved as the amount of Ca-

added increased. Increasing calcium helped reduce the total oxygen content of the steel castings.

However, the silicon content in CaSi wire led to an increase in MnSiO3 inclusions with increasing

Ca. Ar-stirring in the ladle, after CaSi-wire treatment, helped in flotation and reduction of the

inclusions and the total oxygen content. But Ar-stirring alone was not as effective as Ca-

additions.

0

0.0002

0.0004

0.0006

0.0008

0.001

0.0012

0.0014

0.0016

0.0018

0.00 % Ca

0.00 ft/min

No Stir

Ladle Stir 0.024% Ca

12.5 ft/min

No Stir

0.028% Ca

12.5 ft/min

No Stir

0.032% Ca

12.5 ft/min

No Stir

0.043% Ca

12.5 ft/min

No Stir

0.043% Ca

12.5 ft/min

Ladle Stir

0.05% Ca

12.5 ft/min

No Stir

0.06% Ca

20 ft/min

No Stir

Fra

cti

on

of

are

a c

overe

d b

y in

clu

sio

ns Others

TiO2

CA

Al2O3

MnO

MnSiO3

CaS

MnS

0

20

40

60

80

100

120

140

160

180

200

0.00 % Ca

0.00 ft/min

No Stir

Ladle Stir 0.024% Ca

12.5 ft/min

No Stir

0.028% Ca

12.5 ft/min

No Stir

0.032% Ca

12.5 ft/min

No Stir

0.043% Ca

12.5 ft/min

No Stir

0.043% Ca

12.5 ft/min

Ladle Stir

0.05% Ca

12.5 ft/min

No Stir

0.06% Ca

20 ft/min

No Stir

Am

ou

nt

of

ox

yg

en

(p

pm

)

a)

b)

Figure 11: Comparison of a) total oxygen and b) volume and composition of inclusions

in cast keel blocks (from all trial ladles)

Toughness Results: Standard Charpy V-notch specimens were machined from normalized

(900oC for 1.5 hours) keel blocks according to ASTM E23 standards. Three Charpy tests were

completed at -40°C for each trial heat and the average plotted in the attached figures. As shown

in Figure 12, the toughness increased with the amount of calcium. Ar-stirring was also found to

be beneficial in increasing the toughness. The highest toughness average was for the trial with the

combined CaSi-addition and Ar-stirring.

As given in Equation 1 earlier, fracture toughness theoretically varies linearly with (Vf)-1/6

, where

Vf is the volume fraction of inclusions, assuming all of the other properties are constant. In this

work, the area fraction of inclusions (Af) was measured which is related to the volume fraction.

On this basis, toughness, as measured by the Charpy impact energy data, was plotted versus

inclusion (Af)-1/6

(see Figure 13). As observed from this figure, the Charpy impact toughness

varies linearly with the inclusion (Af)-1/6

verifying this relationship and the adverse affect of

inclusions on the toughness of cast steel.

Figure 12: Average Charpy impact energy absorbed for each of the different ladle treatments

Figure 13: Correlation between the toughness and the inclusions content

Figure 14 shows the effect of the (Af)-1/6

of a) oxide and b) sulfide inclusions on the Charpy

impact energy absorbed, as measured for all the cast samples. A linear relationship is observed

between the toughness and the oxide inclusion (Af)-1/6

showing that decreasing the volume of

oxide inclusions directly increases the toughness. The toughness data did not show quite as close

of a linear relationship when plotted against the (Af)-1/6

of sulfide inclusions, however, toughness

definitely increases with decreased sulfide inclusion volume. These results clearly suggest that

both types of inclusions are harmful to the toughness of cast products and must be avoided if

toughness is critical.

0

2

4

6

8

10

12

14

16

18

0.00 % Ca

0.00 ft/min

No Stir

Ladle Stir 0.024% Ca

12.5 ft/min

No Stir

0.028% Ca

12.5 ft/min

No Stir

0.032% Ca

12.5 ft/min

No Stir

0.043% Ca

12.5 ft/min

No Stir

0.043% Ca

12.5 ft/min

Ladle Stir

0.05% Ca

12.5 ft/min

No Stir

0.06% Ca

20 ft/min

No StirCh

arp

y Im

pact

En

erg

y A

bso

rbed

(ft

-lb

s)

0

2

4

6

8

10

12

14

16

18

20

2.8 2.9 3 3.1 3.2 3.3 3.4

(Fraction of Area Covered by Inclusions)^(-1/6)

Ch

arp

y I

mp

act

En

erg

y A

bso

rbed

(ft-

lbs)

(Fraction of Area Covered by Inclusions * 10^3)

2.075 0.6470.7740.9311.1271.3721.681

a)

b)

Figure 14: Correlation between the toughness and the inclusions content for a) oxides and b)

sulfides, as measured for all the cast samples

The toughness of the casting was also found to have a direct correlation with the aspect ratio of

the inclusions (Figure 15a) and the average spacing (Figure 15b) between inclusions. As

expected, the castings with inclusions that were rounder and less irregular in shape (smaller

aspect ratio) resulted in higher toughness. This verifies the observation that Ca-treatment helped

modify the inclusions to a round shape, which in turn increased the toughness of the casting. As

given in Equation 2 earlier, the fracture toughness is expected to varies linearly with s1/2

, where s

is average inclusion spacing, assuming all other properties for the system are constant. Using the

program written in Excel with the x-y location data provided by the ASPEX inclusion analysis,

the average inclusion spacing was calculated for each of the trial heats and plotted as (inclusion

spacing) 1/2

versus the average Charpy impact energy absorbed (see Figure 15b). The relationship

0

2

4

6

8

10

12

14

16

18

20

2.9 3 3.1 3.2 3.3 3.4

(Fraction of Area Covered by Oxide Inclusions)^(-1/6)

Ch

arp

y I

mp

act

En

erg

y A

bso

rbed

(ft

-lb

s)

(Fraction of Area Covered by Inclusions * 10^3)

0.6470.7740.9311.1271.3721.681

0

2

4

6

8

10

12

14

16

18

20

3.5 3.7 3.9 4.1 4.3 4.5 4.7 4.9 5.1

(Fraction of Area Covered by Sulfide Inclusions)^(-1/6)

Ch

arp

y I

mp

act

En

erg

y A

bso

rbed

(ft

-lb

s)

(Fraction of Area Covered by Inclusions * 10^3)

0.0720.0930.1200.1580.2110.2840.3900.544 0.057

was linear which verifies the theoretical relationship in Equation 2 and implies that it is not only

the volume of inclusions but the spacing between inclusions that affects cast steel toughness.

a)

b)

Figure 15: Effect of the a) average aspect ratio of inclusion and b) (average inclusion spacing) ½

on the Charpy impact energy absorbed as measured for all the cast samples

SUMMARY AND CONCLUSIONS

The use of modern tools such as automated inclusion analyzers to identify the number, shape,

size, composition and spacing of inclusions in steel casting is critical to studying the effects of

practice changes on steel quality. Automated inclusion analysis allows for collection of more

data in one hour than could be collected manually in several days.

Calcium treatment was found to be beneficial for inclusion modification and control. In all the

heats conducted with Ca-additions in the ladle, the fraction of area covered by inclusions and the

total oxygen was found to decrease after the Ca-treatment. The major reduction was found in

alumina and MnS inclusions. The inclusions present after Ca-treatment consisted primarily of

calcium aluminate and CaS inclusions. The shape factors of calcium aluminate and CaS

inclusions was close to 1 and their SEM images also showed a round shape, suggesting that they

are not as harmful to mechanical properties of cast steel as non-spherical inclusions. The volume

of inclusions and total oxygen were found to be directly decreased by the amount of Ca-added.

0

2

4

6

8

10

12

14

16

18

20

2 2.5 3 3.5 4 4.5 5

Average Aspect Ratio of the Inclusions

Ch

arp

y I

mp

ac

t E

ne

rgy

Ab

so

rbe

d (

ft-

lbs

)

0

2

4

6

8

10

12

14

16

18

20

0.14 0.145 0.15 0.155 0.16 0.165 0.17 0.175 0.18 0.185 0.19

(Inclusion Spacing)^0.5 (mm^0.5)

Ch

arp

y I

mp

act

En

erg

y A

bso

rbed

(ft-

lbs)

The average Charpy impact energy absorbed for cast products were found to be linearly

proportional to the volume fraction of inclusions to -1/6 power, in agreement with the relationship

in Equation 1. Samples with lower aspect factors (closer to round) resulted in higher toughness.

Increased spacing between inclusions improved toughness according to the relationship in

Equation 2. The Charpy impact energy absorbed was found to increase with increased amounts

of Ca-added in the ladle. The highest toughness was obtained in samples with high levels of

calcium added and then stirred with Ar.

ACKNOWLEDGEMENTS

The work for this project was made available through 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 DURIP Grant W911NF-08-1-0267. The Ca-wire feeder was

provided through a grant from P.C. Campana. The authors also acknowledge the support of the

Steel Founders Society of America and the member companies that participated in this research.

REFERENCES

1. R.W. Rastogi and A.W. Cramb, 2001 Steelmaking Conf. Proc., ISS, Warrendale, PA, Vol.

84, 2001, pp. 789-829.

2. R.A. Rege, E.S. Szekeres and W. D. Forgeng, Met. Trans., AIME, Vol.1, No. 9, 1970, pp.

2652.

3. L. Zhang, B.G. Thomas, “Literature Review: Inclusions in Steel Ingot Casting,” Met and

Materials Trans B, Vol.37B, No.5, 2006, pp.733-761.

4. T.B. Braun, J.F. Elliott, and M.C. Flemings, Met and Materials Trans B, Vol.10B, 1979, pp.

171-84.

5. T.B. Cox and J.R. Low, “Investigation of the Plastic Fracture of AISI 4340 and 18Ni-200

Grade Maraging Steels,” Met and Materials Trans A, Vol. 5A, 1974, pp.1457-70.

6. P. Kenny and J D Campbell, Prog. Mat. Sci., Vol. 13, 1967, pp. 135-181.

7. W.M. Garrison, “Controlling Inclusion Distributions to Achieve High Toughness in Steels,”

AIST Trans, Vol. 4, No. 5, 2007, pp.132-139.

8. G.T. Hahn, M.F. Kanninen, and A.R. Rosenfield, Annual Review of Materials Science, Vol.

2 , 1972, pp. 381-404.