j m a t e r r e s t e c h n o l .

2 0 1 8;7(3):283–299

Available online at www.sciencedirect.com

www.jmrt.com.br

Review Article

Non-metallic inclusions in steels – origin

and control

André Luiz Vasconcellos da Costa e Silva

EEIMVR-UFF, Volta Redonda, RJ, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: Non-metallic inclusions (NMIs) play a key role in many important properties of steel,
Received 11 January 2018 influencing both processing and application of steel products. In this work, the current
Accepted 19 April 2018 understanding of the origin and classification of NMIs is reviewed, highlighting the dra-
Available online 24 May 2018 matic development of the last decades. This includes the discussion of the thermodynamics
of inclusion formation and the current knowledge on the effects of melt shop processing
Keywords: variable on NMIs composition, amount and size distribution. The development of inclusion
Non-metallic inclusions engineering – tailoring the process to obtain the desired NMIs is highlighted and the devel-
Ladle metallurgy opment in selected areas – tire cord, springs and bearing steels as well as prevention of
Steelmaking nozzle clogging – is used to illustrate this development. The promising field of “oxide met-
Computational thermodynamics allurgy” is also discussed in the context of inclusion engineering. Finally, the difficulties in
Kinetics meaningfully characterizing and quantifying NMIs are briefly commented. In summary it
is concluded that inclusion control in steels has evolved significantly in the last decades.
This is due to the progress in understanding the interplay between thermodynamics, steel
and slag chemical composition as well as melt shop processing. This made possible the tai-
loring of non-metallic inclusions via processing, to optimize steel properties. Nonetheless,
some important problems remain and must still be solved to improve inclusion control and
optimization.
© 2018 Brazilian Metallurgical, Materials and Mining Association. Published by Elsevier
Editora Ltda. This is an open access article under the CC BY-NC-ND license (http://
creativecommons.org/licenses/by-nc-nd/4.0/).

E-mail: andre.costaesilva@gmail.com
https://doi.org/10.1016/j.jmrt.2018.04.003
2238-7854/© 2018 Brazilian Metallurgical, Materials and Mining Association. Published by Elsevier Editora Ltda. This is an open access
article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
284 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
André Luiz Vasconcellos da Costa e Silva
graduated as Metallurgical Engineer from
the Military Institute of Engineering (IME),
Rio de Janeiro, in 1976. Obtained his
M.A.Sc. from the University of British
Columbia, Vancouver, in 1979 and his
Ph.D. from the University of Florida,
Gainesvile, in 1994. His carreer started as
an engineer in Eletrometal Aços Finos S.A.
(currently Villares Metals S.A.). Has been
the Materials Engineering Manager at CBV
Industria Mecanica S.A. (Currently TechnipFMC plc) and has
taken part on the pioneer projects on deepwater oilfield com-
pletion and production. Has been the Technical Director of
the Brazilian Institute for Nuclear Quality, where he has been
certified both as Level III Inspector and Level III Expert in the
Engineering Materials area. He is Professor at the Metallurgical
Engineering School in Volta Redonda, of Universidade Federal
Fluminense. Andre was the Chairman of the Alloy Phase Dia-
gram International Commission (APDIC) from 2008 to 2013. He
was awarded the 2014 Hume Rothery Award, for distinguished
achievements in relation to phase transformations by IOM3 ,
England and in 2017 became the 15th recipient of the trien-
nial “Hubertus Colpaert” Silver Medal for his contribution in
metallography and physical metallurgy by ABM, Brazil. He is a
member of ASM, ABM and TMS.
1. Introduction
Until some 50 years ago, it was not uncommon for a metal-
lurgy student to be told that non-metallic inclusions (NMIs)
in steel came mostly from eroded pieces of refractories and
entrapped slag. Albeit inclusions in steel were already a theme
of important studies at that time, the knowledge of their
importance to steel behavior was certainly not as perva-
sive as it is today. Since NMIs are non-metallic phases in
the steel matrix, perhaps it is also understandable that they
were not attractive to many physical metallurgists, mostly
concerned with understanding the metal behavior and its
phase transformations. As steel was challenged by different
alternative materials and by more demanding applications, it
became clearer that an important part of its behavior is inti-
mately connected to the type, size and distribution of NMIs
present in the matrix. This led to intense development in
understanding their origin, properties and behavior during
product processing. Much improvement also took place in
NMIs control and quantification and in the correlation of its
features with their effect on properties. Tailoring inclusions
to improve properties and performance became an important
feature of steelmaking and in the 1980s the term inclusion
engineering was coined to describe this process [1]. Inclu-
sion engineering starts with the definition of the desirable
properties that inclusions should have. Then, through ther-
modynamics and adequate processing, a product where these
desirable inclusions are dominant is designed and made. In
the present work, studies that characterize the origin and
control of NMIs are reviewed together with the current under-
standing of their behavior during processing. A brief overview
of the work on quantification of inclusions and character-
ization of their distribution is presented. The most recent
work on inclusion engineering is also reviewed. The effects
of NMIs on steel properties are not discussed in depth in
the current review. As literature in the field is growing at
a fast pace, the reader is directed, whenever possible, to
recent reviews on specific topics. The classical references for
the subject are the seminal book by Kiessling [2] and the
proceedings of the series of International Clean Steel Con-
ferences organized, every 3–5 years, by the Hungarian Mining
and Metallurgical Society (OMBKE). The interested reader is
directed to these sources for invaluable information on the
subject and on the many excellent reviews mentioned in the
text.
2. Origin and classification of non-metallic
inclusions
2.1. NMIs classification
When modern steelmaking technology started developing, it
was soon noticed that oxygen, widely used as refining agent,
and sulfur, a ubiquitous impurity coming most from coal and
coke, both extensively soluble in the liquid metal had very
little solubility in the solid phases of iron. Thus, upon solid-
ification, they were rejected to the liquid and finally formed
low melting compounds “FeO” (wustite) and FeS and eutectics
involving these compounds. This rendered steel inadequate
for hot working [3]. The mutual solubility of the oxides and
sulfides of iron and some alloying elements (e.g. Mn) fur-
ther complicates the problem [4,5]. Hence, the content of
oxygen and sulfur in solution must be decreased to pro-
duce quality steel. Many elements commercially available and
acceptable as part of the steel composition such as silicon,
manganese and aluminum, have a high affinity for oxygen
and can thus be used as deoxidizers readily forming non-
metallic deoxidation products when added to the liquid steel
[6]. These products can become important oxide NMIs. In
the case of sulfur, on the other hand, only elements with
low solubility in iron (such as Ca and Mg) or rare-earths
have sufficiently high affinity to sulfur to form non-metallic
sulfides at the liquid metal temperatures [6]. Thus, most
of the sulfur in steel must be removed from solution by
slag refining and the rest, by precipitation reactions occur-
ring mostly during solidification. These differences introduce
two of the possible forms of classifying non-metallic inclu-
sions: by their chemical composition (oxides, sulfides, etc.)
and by the stage in the process in which they are formed,
with reference to the start of the solidification. Inclusions
formed before the start of steel solidification are called pri-
mary and those formed after solidification starts, secondary.
Besides these, another common classification that has con-
tributed significantly to confusion in the field, is concerned
with the source of the inclusions. Those originating from
the steelmaking process (such as the oxides and sulfides
discussed above) are classified as “endogenous” and those
appearing from “external” sources (fragments of refractories,
entrapped slag, etc.) are classified as “exogenous”. Rarely how-
ever, will an “exogenous” volume of material survive long
enough in the steel without suffering extensive reaction with
the melt and thus changing significantly. In the past there
was an important contribution from “exogenous” sources to
the inclusionary population, in special large inclusions from
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
Wire
Inclusion
Acc.V Spot Magn Det WD
15.0 kV 6.0 318x BSE 8.2
Fig. 1 – Tire cord wire broken during wire drawing operation due to the presence of large hard inclusion. Inclusion is
indicated by arrow. SEM, scanning electron micrograph; BE, backscattered electrons [12].
refractories and casting pit material, for instance [7]. This,
however, has been significantly reduced in modern steels.
Nonetheless, it is still a problem that when students are
introduced to the subject sometimes they end up believing
that the most important sources of non-metallic inclusions
are “exogenous”. This does not mean that the interactions
between the metal and refractories, for instance, are not
relevant [8,9]. These interactions can indeed have a signifi-
cant influence on the inclusions present in the steel. If the
definition of “exogenous” includes also, for instance, inclu-
sions originated from reoxidation from contact with air and
those resulting from entrained slag interacting with the liquid
metal, as proposed in [10], the range of “exogenous” inclu-
sions expands considerably. This is probably a more adequate
definition of “exogenous” inclusions. However, not only these
inclusions are changed to a substantial extent during the
process but, in some cases, it can be argued that they are
“part of the process”. Thus, it is debatable to which extent
the “exogenous”–“endogenous” classification helps or further
confuses the issue.
Finally, a common way of classification of NMIs is related
to the inclusion size: thus, inclusions can be classified as
macro and micro inclusions. One must be aware that, to
have a sensible cutoff between sizes, Kiessling’s proposal
should be used, when adopting this classification [2,11]: i.e.,
an inclusion is a macro inclusion if it is large enough to
cause immediate failure of the product either during pro-
cessing or use, as for instance, the inclusion shown in Fig. 1.
All other inclusions should be classified as micro inclusions.
Thus, while important, this is a difficult classification to apply.
Frequently, sizes are defined arbitrarily to separate macro and
micro inclusions and the classification becomes difficult to
justify.
285
100µm
2.2. Thermodynamic basis
As steelmaking temperatures are high, inclusion formation
reactions approach equilibrium. For this reason, thermody-
namics is an important tool in understanding their formation.
The thermodynamic fundamentals of NMI formation have
been the subject of study for a long time [13,14] and are
quite well known, e.g. refs. [6,15–18]. Nonetheless, it is worth
mentioning that there are still important discrepancies in the
thermodynamic data related to the formation of, for instance,
alumina inclusions [19,20]. This may not be a problem for the
usual range of aluminum application as deoxidant, but it does
introduce important uncertainties in the processing of the
new generation of high Al high Mn steels, for instance [21].
The problem is also complicated in the case of elements with
limited solubility in iron such as Mg and Ca [18,22]. For these
elements, significant discussion still prevails with respect to
key thermodynamic data relevant for the study of reactions
in steelmaking. The application of computational thermody-
namics has provided, in recent decades, important insight in
the solution of complex problems. These used to be rather
difficult to handle with more conventional calculation tech-
niques. General aspects and examples of the application of
this technique to steels and steelmaking have been discussed,
for instance, in refs. [23–27]. Examples of the application of the
technique to problems related to NMIs can be found in many
publications such as refs. [28–32].
3. Formation, removal and control of NMIs
The classification as primary and secondary is useful in the
discussion of inclusion formation and removal. Removing
286 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299

0.16 a
0.14
Oxygen in steel (%)

0.12 Oxygen a
%0i
in
0.1 solution

0.08 48
Total oxygen (Ot) 54
0.06

0.04 b
%0t
0.02 %0s c

0
%Als %Aladded
-2 0 2 4 6 8 10 12 14 16 b
Minutes after silicon addition 16

14

Inclusion level in cast slab

Fig. 2 – Change in total oxygen and dissolved oxygen in
steel as a function of time after addition of silicon to 12

deoxidize the bath. Total oxygen was measured in four 10

different heats. Data from [34]. 8 Ar stirring
Power
6 injection
4
primary NMIs from the liquid steel is, in principle, possible. Vacuum
2 treatment
Conversely, secondary inclusions, that form during solidifi-
cation, are probably impossible to remove and can only be 0
2 3 4 5 6 7 8
“engineered” to result in the less detrimental possible effect
Ot content in ladle (10-30%)
on the product. c 10
Ot from pin samples, ppm

3.1. Primary inclusions

8

3.1.1. Nucleation and morphology

Primary inclusions form in the liquid steel and follow closely
the thermodynamics of the process. Thus, two important
aspects should be considered: their nucleation and the cor-
respondent morphology.
In general, nucleation is rapid when deoxidants are added.
This may be due to the large supersaturations observed dur-
ing the dissolution of the deoxidant additions. Nonetheless,
Sigworth and Elliott, for instance, carefully evaluating the
conditions for nucleation of silica [33] observed the need for
supersaturation. The results of Miyashita [34] and industrial
observations do not indicate, however, that supersatura-
tion is significant in silicon deoxidation during steelmaking.
Miyashita also observed that “comparison of changes of the
dissolved and the total oxygen shows that the rate of decrease
of the total oxygen is determined by the rate of removal of the
deoxidation products”, Fig. 2.
The dependence of the amount of oxide inclusions on
the difference between total oxygen and soluble oxygen in
steel has been highlighted in many deoxidation studies (for
instance ref. [20]). This is shown schematically in Fig. 3.
Suito and co-workers [37] discussed the possibility of super-
saturation in an experimental study of the formation of
alumina in iron. As in the case of silicon deoxidation, this
does not seem to be a problem in steelmaking. In actual steel
processing, heterogeneous nuclei will probably be abundant.
Hence, the laboratory conditions observed by [33,37] cannot be
reproduced in the industrial environment.
Significant work has been done on the nucleation and
growth morphology of alumina in steel, because of the impor-
tance of aluminum killed steel to the industry. Several authors
[38–43] have looked in detail into the evolution of alumina
6
4
2
0
1 1.5 2.5 3
Index of oxide inclusions (chemical samples)
Fig. 3 – (a) A simple example of the dependence of the
amount of oxide NMIs on the difference between total
oxygen and soluble oxygen. In the example, deoxidation
starts in point “a” with the addition of Al to a steel with an
initial content of dissolved oxygen %Oi . In the absence of
nucleation barriers, alumina will be formed to bring the
dissolved oxygen and aluminum contents to point “c”. The
total oxygen content in the steel will correspond to the
oxygen in solution (%Os ) added to the oxygen in the
alumina inclusions still present in the bath, corresponding
to point “b”. From “a” to “c” deoxidation proceeds
according to the stoichiometry of the reaction
2Al + 3O = Al2 O3 . Note: Usually the difference between
soluble Al (%Als ) and total Al is small and difficult to
measure. (b) The relationship between the amount of
inclusions and total oxygen (%Ot ) in steel determined by
[35]. When inclusion and total oxygen content is lower,
extreme care is necessary to obtain reliable measurements,
as shown in (c). (c) Relationship between oxide inclusion
density in chemical samples versus total oxygen in pin
samples collected after completion of ladle furnace
treatment. Oxide inclusions quantified by automated SEM
(adapted from [36]).
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299 287

Increasing activity of oxygen

Increasing activity of deoxidant

Fig. 4 – Oxide growth shapes as a function of the local activities of oxygen and aluminum (adapted from [40]).
Estimated average local oxygen activity (ppm)

High

600

400

200

100

0 0.01 0.05 0.1 1 High

Estimated local AI concentration (%)

Fig. 5 – A summary of morphologies of alumina observed

in steel with different estimated local O activities and Al
concentrations. Adapted from [44]. For the source of the
individual observations, see ref. [44].

morphology and its relation to supersaturation and processing

time. Fig. 4 presents the classical view developed by Steinmetz
and co-workers [40]. As expected, supersaturation plays an
important role in growth morphology. Recently, Tiekink and
co-workers [44] have attempted to summarize alumina mor-
phologies observed as a function of oxygen and aluminum
supersaturation, as presented in Fig. 5. The complexity of the
task is evident.
Alumina morphology can have a significant impact on the
final product properties (e.g., ref. [45]). It is particularly inter-
esting to note that primary inclusions, forming and growing
in the liquid have the possibility of “free” growth in several Fig. 6 – Some examples of alumina inclusion
different morphologies (Figs. 4 and 6) and to agglomerate due morphologies. Inclusions extracted by dissolving the
to collisions (Fig. 7). Shape changes may occur due to surface as-cast matrix. In (a) the dendritic structure is circles with a
energy aspects if the processing time is sufficiently long. dashed line [45]. The fibers of the filter used to retain the
inclusions during the dissolution process are visible in the
3.1.2. Removal background. (b) Polygonal alumina inclusion, courtesy of A.
The possibility of floating out inclusions formed in a stag- Mitchell, UBC [12].
nant liquid bath can be ruled out by simple calculations,
mostly due to the limitations of Stokes’ law [38]. Small inclu-
sions reach limit floating velocities that require unrealistic occurring during clustering of inclusions has been made pos-
processing times for usual ladle depths. Clustering of the sible by in situ observation [39,47]. Emi and co-workers [47,48]
inclusions [38,39] thus significantly helps on the possibility observed in situ the behavior of inclusions at the steel–gas
of floating. The importance of clustering, in special in the interface. They demonstrated that in this condition, alumina
case of alumina inclusions, has been noted for a long time clusters readily. Calcium aluminate inclusions, on the other
[40]. Important progress in understanding of the processes hand, seem to have little tendency to move closer to each
288 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
Fig. 7 – Alumina cluster in sample taken from ladle, deep
etched with Nital [46]. Courtesy C. Cicutti, Tenaris Siderca.
other, clustering only when a full collision happens in the
liquid. Wikström and co-workers (e.g. [49,50]) expanded the
in situ observations to the steel–slag interface and to inclu-
sions within the slag. They confirmed the results of Emi for
the metal–gas interfaces. Emi and co-workers [47] observed
that while these observations are relevant for phenomena at
gas–steel interface (for instance, at bubble surfaces) they do
not give direct indication about how agglomeration happens
within the liquid steel. This is especially important for liq-
uid inclusions, where other forces may be relevant. In any
case, whether the inclusions are solid, or liquid plays a role on
agglomeration. For a long time now, it has been established
that stirring is essential to promote clustering but, perhaps
more important, to bring the inclusions to their most impor-
tant sinks, the slag and the ladle walls. The work of Lindskog
and co-workers [51,52] using radioactive tracers has been very
important in understanding the importance of these sinks.
With the current limitations on the use of radioactive tracers,
BaO has been adopted as a tracer when it comes to evaluating
the eventual entrapment of slags and fluxes [53] and its influ-
ence on steel cleanness. The use of BaO tracers has also been
very effective in determining the effect of ladle glaze on sub-
sequent heat cleanness [53,54]. IRSID has developed the use
of lanthanum as a tracer for oxide inclusions. Since La2 O3 is
very stable, the alumina inclusions present at the moment of
La addition were marked by this element. Hence, re-oxidation
inclusions formed afterwards could be identified [55]. In the
case of “exogenous” inclusions originated from mold flux, the
presence of alkali oxides can be used as “tracer”. Mold fluxes
are usually the only additions in steelmaking that contain sig-
nificant amounts of alkali oxides [56].
The overall process of inclusion removal in ladle pro-
cessing, where most of primary deoxidation inclusions are
formed, involves at least three steps: the growth and/or
clustering of individual inclusions, the movement of the
inclusions to the sinks (ladle slag and ladle walls) and their
attachment to these sinks. Two factors are important for the
movement of the inclusions: bath stirring, which is probably
the dominant factor, and dragging by rising bubbles. It has
been shown that in most ladle processes, the gas bubbles in
steel are too large to effectively remove a significant number
of inclusions, unless a very large volume of gas is used [57].
The calculations of Zhang and Taniguchi [57] indicate that this
O, ppm
100
80
60
800A
40
20 1250A
1000A
0 2 4 6 8 10 12 14
Time, min
Fig. 8 – Total oxygen content as a function of stirring
current and treatment time in an ASEA-SKF ladle furnace.
Final oxygen content is dependent on residual aluminum
content in each heat. Adapted from [60].
tool is only effective in regions where the steel velocity is high
and fine bubbles are possible. This is typically the case for
submerged entry nozzles (SEN) [58] or similar configurations
such as the ladle nozzle shroud [59]. Some advantage is also
observed in the ascending leg of RH vessels.
Stirring has been recognized to promote the agglomera-
tion and the transport of the inclusions or their clusters to
sinks. Fig. 8, [60] is one of the first studies that clearly showed
that increasing the stirring energy (in this case the current in
electromagnetic stirring) increases the kinetic constant for the
removal of inclusions (measured by total oxygen content).
Industrial observations seem to indicate that the efficiency
of inclusion removal passes through a maximum at a certain
agitation energy. The first report of this observation is proba-
bly the one by Suzuki and co-workers [61]. Their results were
presented as a function of specific work of agitation, but they
recognize the importance of agitation energy: “The reason for
the decrease [of inclusion removal] seems to be due to the
erosion of brick or entrapment of the slag. . ., because CaO and
MgO type inclusions are found increased with stronger stirring
energy.” Their results are shown in Fig. 9.
Later, Neifer and co-workers [62], and Ek and co-workers
[63] used computational fluid dynamics and physical model-
ing to study the removal of oxides as a function of the rate of
gas flow in the ladle. The modeling results of Neifer [62] indi-
cated an optimum gas flow for the removal of oxide NMIs. They
observed that the decrease in total oxygen was less efficient
with increased gas flow rate. They ascribed this to reoxidation
from slag or from the atmosphere. These results are consistent
with the results of Suzuki and co-workers [61]. The results of
Ek and co-workers [63] showed relatively low influence of the
argon flow rate on the rate of inclusion removal. The authors
suggested that rinsing (for inclusion removal) is best achieved
with a low argon flow rate. Interestingly, the industrial mea-
surements of Neifer and co-workers [62] indicated that the
rate of total oxygen reduction decreased with the increase in
gas flow. They suggested that natural convection gave the best
results in their industrial tests. As there is very limited data

100
Low Si
Al fully killed
80 Si killed
Degree of deoxidation (%)
Ot <20 ppm
60
40
20
0 Fig. 11 – Effect of ladle stirring on average chemistry of all
0 20
Fig. 9 – Effect of stirring work on the degree of deoxidation.
Circled points have reached less than 20 ppm total oxygen
content. Adapted from [61].
Ar bubbling
Oxygen removal rate constant (min-1)
100
10-1
10-2
10-3
100
Fig. 10 – Oxygen removal constant k in equation d%O dt
t
= −kt
as a function of specific stirring power in different
secondary metallurgy reactors. Data mostly from [60,66,67]
adapted from [64].
on the industrial heats where measurements were made, it is
difficult to speculate on the causes of this observation.
More recently, Zhang and Thomas [64] collected many
results of the kinetic constant for the removal of oxide inclu-
sions as a function of agitation energy in a single graph
(Fig. 10). They included an envelope around the measure-
ments, to stress the existence of an optimum agitation. They
also performed simulations that reproduced the expected
behavior [65] represented by the envelope in Fig. 10.
The existence of an optimum agitation may be caused by
reoxidation. Excessive stirring leads to the opening of an “eye”
in the slag causing steel exposure to the atmosphere [64] and
j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299 289
60
Average elemental content (all NMIs) %
Strong stirring
50
Soft stirring
40
30
20
10
0
Mg Ai Ca Ti
40 60 80 100 120 140 inclusions observed in samples collected from ladle, mold
Specific stirring work (103 j/t) and slabs. According to Kaushik and co-workers, strong
stirring was performed during desulfurization and probably
promoted slag emulsification as evidenced by the higher Ca
and Mg content in the inclusions. Adapted from ref. [71].
slag entrainment, mostly at the edge of the “eye” [68–70], as
ASEA-SKF suggested by Suzuki and co-workers [61]. Fig. 11 illustrates
RH the change in NMIs chemistry when strong stirring is used
VOD
for desulfurization. The increase of Ca and Mg content in the
inclusions is evidence of probable slag emulsification [71].
In the absence of a slag, excessive agitation with argon
was observed to cause re-entrainment of alumina inclusions
[70]. Thus, it seems clear that rinsing operations aimed at the
removal of NMIs from the liquid metal in the ladle, should be
performed at optimum selected agitation energies.
The importance of reoxidation for steel cleanness cannot
be overemphasized. Nadif and co-workers recently reviewed
the importance of reoxidation and the measures to control it
[72]. In recent decades, manufacturers have added many mea-
sures to control the most important sources of reoxidation
in and after ladle metallurgy: flat products producers regu-
101 102 103 104
larly use submerged entry nozzles (SEN) and ladle shroud to
Specific stirring power (W/t)
prevent contact of the steel stream with air. Tundish inerti-
zation [73] and mold protection with inert gas, in special in
long products casting, are becoming standard for high qual-
ity steels. Special attention is also given to shroud and valve
design to control air ingress caused by negative pressures in
valve systems [74].
The attachment of the inclusions to the sinks depends
mostly on surface tension and, on a secondary level, on the
ability of the sink to dissolve the inclusions. Most steelmaking
slags, regardless of their composition, will wet oxide inclu-
sions. This is due to the magnitude of the difference in the
surface energy between inclusion and steel and inclusion and
slag. This issue has been discussed for a long time, at least
since the work of [75] and others and has been summarized
by Olette in ref. [76]. The importance of the liquid fraction
of the slag in promoting NMIs removal was clearly under-
stood very early [75] and was confirmed by clever experiments
[77]. There remains, however, some controversy on the effect
of the slag viscosity. Nakajima and Okamura [78] proposed a
290 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
a Steel-slag
Z Steel-slag interface
inclusion
interface
Slag
0
Metal
Inclusion
Preliminary Contact
approach with slag
time
Re < 1
Steel-slag
b Steel-slag interface
Thin
inclusion
Z interface
Slag film
S0 S
Metal
Preliminary Inclusion Steel film Steel film rupture
approach drainage contact with slag
time
Re > 1
Fig. 12 – Schematic representation of two types of behavior
when inclusions cross the metal–slag interface, according
to Nakajima [78]. The Reynolds number (Re) as the
inclusion reaches the interface would define the behavior.
Adapted from ref. [78].
model to describe the process of inclusion transfer across the
metal–slag interface. This has been later discussed in detail
and used in many studies [79–82] and has been considered
in a recent review on the topic of absorption of inclusions by
slags [83]. Nakajima and Okamura proposed that under certain
conditions the inclusion transfer from the metal to the slag
may include the drainage of a film of metal from the interface,
while in other cases, in special in the case of solid inclusions,
such a film may not exist [80] (Fig. 12).
They concluded that the viscosity of slag and the relevant
surface energies were the critical parameters to define the effi-
ciency of the passage of the inclusion, and of reducing the
risk of it being re-entrained in the metal. This could be sum-
marized by a condition defined by the Reynolds number at
the inclusion approach. More recently Sridhar and co-workers
[81,82,84,85] have concluded that their in situ observations
indicated film drainage to be a pervasive phenomenon, delay-
ing the passage of inclusions to the slag in most cases.
Once the inclusions leave the metal, their dissolution has
been considered immediate, in the case of liquid inclusions
(see, for instance, ref. [65]). With the possibilities of in situ
observation, interesting work has been performed in the dis-
solution kinetics of solid inclusions [86–88]. In some cases, the
kinetics of dissolution is controlled by transport (“diffusion in
a boundary layer”). In other cases, such as MgO inclusions, the
formation of intermediate layers that depends on slag compo-
sition may hinder the dissolution introducing a chemical step
in the process. This confirms previous results [89] obtained
by more conventional techniques. More recently, Yan and
co-workers [90] assessed the dissolution rate of MgO in slags,
indicating that all available data agrees with mass transfer
control.
Holappa and co-workers [91] investigated the develop-
ment of active tundish slags in which inclusion absorption
is a “key property”. They observed that the efficiency with
which a given slag will dissolve solid non-metallic inclusions
depends on the complex interplay between slag composition,
thermodynamics, surface tension and viscosity. Holappa and
co-workers concluded, “further systematic research is needed
to get better knowledge. . ..and to develop optimal slags”. The
attachment of NMIs to ladle refractories can be favorable [51]
but ladle glaze from previous heats may also be a source of
inclusions [53], depending on the slag composition of the heat
previously processed in that ladle [54,92]. Furthermore, inclu-
sion attachment to valve refractories can be disastrous if it
happens during casting causing nozzle clogging, for instance.
The subject has been well reviewed (e.g. refs. [93,94]) and will
be discussed later. It is interesting to note, however, that, in
the case of clogging caused by alumina, the possible important
effect of FeO formed during reoxidation on the agglomeration
of the inclusions was clearly demonstrated [95].
A significant number of publications on the possibilities
of removal of primary inclusions by casting powders or mold
fluxes, both in conventional ingot casting [96,97] and in con-
tinuous casting [56,98–100], is available but the main issues
are not different from those discussed above. Casting pow-
ders or mold fluxes (similar to tundish slags) must perform a
multitude of functions [56] and flow conditions in the mold
are defined by various constrains, including the avoidance of
powder entrainment on the product surface [101]. This some-
what limits the extent of inclusion removal and retention by
the mold casting powders [102]. Optimizing the flow of steel
in the tundish or in the mold to enhance the chances of inclu-
sions reaching the slag surface can conflict with other relevant
metallurgical objectives [103]. It seems that the conventional
approach of eliminating inclusions as much as possible in the
ladle and then preventing the formation of new ones (by reox-
idation, for instance) may still be the most promising one for
clean steel.
An additional concern is the movement of primary
inclusions inside the strand. The asymmetry of inclusion dis-
tribution in the cross section of products of curved casting
machines, for instance, is well established [12]. Asymmetry of
inclusion distribution can also appear due to partial nozzle
clogging [103].
Sichen [104] has recently reviewed the efforts to model
secondary steelmaking, in special ladle processing. The
attempts at modeling slag-metal interfacial reactions, eye
opening, inclusion generation, nucleation, growth and sep-
aration/removal have been carefully reviewed indicating the
most promising techniques. Sichen [104] cautions, however,
about the difficulties of relating important ladle process vari-
ables (mass transfer coefficients, inclusion removal rate, open
eye size, etc.) to the stirring gas flow rate since “industrial ladle
involves many uncertainties, e.g., the blockage of the porous
plug and the leakage of the gas line.” The difficulties in con-
trolling and measuring gas flow rates in real industrial ladles
were discussed in detail in [105,106] for instance. The possibili-
ties of controlling eye opening via cameras and image analysis
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
and the promising option of controlling ladle stirring via vibra-
tion measurement [107] were evaluated. These techniques are
being adopted in some plants.
3.2. Secondary inclusions
The driving force for precipitation of secondary inclusions
increases as solutes segregate during solidification and the
solubility of oxides and sulfides in steel diminishes as tem-
perature decreases.
The precipitation of inclusions due to changes in solubility
in steel has been discussed for a long time (see, for instance,
the discussion in [108]). Around the beginning of the 1960s the
denomination primary and secondary inclusions was firmly
established [109] and the relations between segregation and
inclusion precipitation were becoming evident. At that time,
the first models to describe the phenomenon were estab-
lished [110]. Turkdogan (e.g. refs. [111–113]) and Flemings (e.g.
refs. [114–116]) made important contributions to the under-
standing of the combined effects of segregation and change
of solubility with temperature on the genesis of secondary
inclusions. In the 1980–90s more sophisticated models were
developed by Nippon Steel [117–121] and by IRSID [122–126].
The same techniques are also applied successfully to the pre-
cipitation on microalloying nitrides during the solidification of
HSLA steels (see [127,128], for instance). These models opened
the way to the development of inclusion engineering, as it
is known today, by coupling thermodynamic databases, and
later kinetic ones, to solidification modeling and the calcula-
tion of inclusions formation. These calculations started from
the knowledge of the desired steel composition for the precip-
itation of the envisaged inclusions followed by the design of
slag constitution to properly achieve these steel compositions
during ladle processing.
One important issue is the interaction of inclusions formed
in the interdendritic liquid and the advancing solidification
front. In situ observation [129] has shown that the solidifica-
tion conditions play an important role on whether inclusions
formed ahead of the interface are pushed or engulfed. Theo-
ries have been formulated and adjusted to properly describe
these results, considering additionally the effects of sur-
face tension and density [130]. The theoretical developments
[131–133] were more focused on metal matrix composites,
but presented good agreement with the results on NMIs in
steels. The results indicate that the critical growth veloc-
ity, V of the advancing interface for the transition between
engulfment and pushing of inclusion of radius R may be
approximated by V = k/R where k is dependent of the inclusion
type.
The morphology of secondary inclusions can be specially
influenced by reactions occurring during their precipitation.
The case of sulfides is probably the best characterized since
Sims observed the effect of deoxidation on sulfide morphology
in the 1930s [108] and later proposed a classification into three
main types of sulfides. This has been addressed in detail by
many authors (e.g. [134]). More recently, Ishida and co-workers
[135–137] concluded that, besides the type of reaction associ-
ated with sulfide formation, surface tension plays a critical
effect on sulfide morphology.
291
4. Inclusion engineering
Gaye and co-workers [124] present one of best overviews of
the conceptual process in the application of thermodynamics
to inclusion engineering in steel. Fig. 13 presents the con-
cept in a simplified way, over two ternary isotherms. Once
the desired inclusions are defined, the steel composition
that will generate these inclusions is calculated. Then, the
ladle slag that will establish this steel composition through
steel-slag equilibrium is calculated and used in the ladle
furnace. The classical aphorism “Look after the slag, and
the metal will look after itself” [138] is fully considered, in
these cases. The concept has been successfully applied to
a wide range of steels. These include steels in which hard
alumina and alumina containing inclusions (such as spinel)
must be avoided, as bearing steels, and steels where the
inclusions must act as nuclei for phase transformation dur-
ing steel cooling. Calcium treatment to form liquid inclusions
Inclusion at 1470 °C SiO2
SiO2
AI, and O in Si-Mn steel AI=2 ppm
AI=1 5
20
0.35%Si 3 8 15
1.0%Mn
O=10 ppm
5ppm
CaO AI2O3
MnO AI2O3
Ladle slag at 1520 °C
Fig. 13 – A schematic representation of the inclusion
engineering process, from the thermodynamic point of
view, for a Si–Mn killed steels in which low melting low
Al2 O3 inclusions are desired. The desired inclusions are
represented in a simplified form on the ternary system
MnO–SiO2 –Al2 O3 . The diagram on the left presents this
system isotherm at 1470 ◦ C. The region of 100% liquid
phase is delimited by the thinner lines on the diagram, as
indicated. The aluminum content of steels in equilibrium
with liquid inclusions in this system is given by the thick
“iso-aluminum” lines at the selected temperature. The
dashed line indicates the composition of inclusions (as a
function of the Al content of the steel) for a 0.35%Si, 1%Mn
steel composition, at this temperature. From this diagram,
it is evident that the steel aluminum content should not go
above the composition of the gray circle (8 ppm) if liquid
inclusions are desired. The diagram on the right presents a
simplified representation of the ladle slag in the
CaO–SiO2 –Al2 O3 system. The 1520 ◦ C isotherm corresponds
to the ladle furnace conditions and shows the region of
100% liquid slag delimited by the thinner lines. The
aluminum content of steels in equilibrium with these slags
is given by the thick “iso-aluminum” lines at the selected
temperature. The gray lines give the corresponding oxygen
content of the steel being studied. Ladle slag compositions
must be selected among those that will guarantee steels
with less than 8 ppm aluminum, indicated in the diagram,
if the desired liquid inclusions (in diagram on the left) are
to be formed. Adapted from [124].
292 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299

and avoid clogging during casting along with sulfide mod- 102
ification with calcium are also discussed in this section,
albeit they may be considered simpler cases of inclusion ppm O

ppm O, ppm AI, ppm Mg

engineering.
101

4.1. Tire cord and spring steels ppm AI

Tire cords are usually made of silicon manganese steel wires. 100

Many automotive springs are made from similar steels. The

presence of hard NMIs (in general alumina or other high
alumina NMIs) is very detrimental to the wire drawing oper- 10-1
ppm Mg
ation of tire cords (Fig. 1) and has an important influence
on the fatigue life of springs [139–141]. Steel composition
must be adjusted to prevent the formation of alumina or 10-2
alumina-rich NMIs, and to favor low melting, plastic sil- 7 8 9 10 11 12 13 14 15 16

icates. This requires rigid control of the alumina content %MgO in slag
of the slag, control of raw materials to prevent undesired 10
C2 0
C1 0
aluminum additions together with low slag binary basic-
ity, which contradicted the prevailing ladle furnace practice
C3 0

ppm AI and O (calculated)

when the problem was first solved [140]. There are many
excellent examples and reviews focusing on the thermody-
namics of inclusion engineering in tire cord and spring steel
[140,142–147].
4.2. Bearing steel
While there is still controversy concerning the effect of the
individual types of NMIs on the fatigue life of bearing steels
[141,148–151] it is clearly established that the size and quantity
of the NMIs present in steel is of paramount importance for
the fatigue performance of the bearings [151]. Some general
agreement seems to prevail in that calcium aluminates and
spinel NMIs are detrimental. There is also agreement that very
low total oxygen, sulfur and aluminum contents are required
to guarantee small size and low volume fraction of NMIs. Fur-
thermore, reversion of magnesium from the slag, which may
lead to the formation of spinel inclusions, must be avoided
[152]. Different processing conditions to achieve these objec-
tives are used at different plants [10,149,152,153]. The control
of slag composition is, however, always one of the key points in
controlling NMIs in bearing steels. Fig. 14 presents calculations
of the effect of slag composition on aluminum, oxygen and
magnesium contents in 100Cr6 (AISI 52100) steel and a com-
parison of calculated and measured aluminum and oxygen
contents in this steel [154].
4.3. Calcium treatment and alumina clogging
prevention
Calcium addition to steel was initially developed to modify
sulfides and control anisotropy in hot rolled and forged parts
[155]. The inclusion modification achieved with calcium treat-
ment has also been used to control machinability [156]. One
can argue, however, that in recent decades calcium has found
most extensive application in modifying alumina inclusions to
prevent nozzle clogging in continuous casting. In this respect,
the operational complexities associated with calcium solu-
bility, variabilities in yield when added to steel and “fading”
effects possibly associated to calcium high vapor pressure
have been extensively studied (e.g. [157–160]). The mechanism
8
6
C3 AI
4
C2 AI
C1 AI
2
0
0 2 4 6 8 10
ppm AI and O (analysis)
Fig. 14 – (a) Effect of %MgO in slag with constant
%Al2 O3 = 5%, %CaO = 48% on Al, O and Mg content in
bearing steel in equilibrium with slag. Calculated with
®
Thermo-calc and SLAG2 database at 1540 ◦ C. (b)
Comparison of calculated and measured Al and O content
in three heats of bearing steel after ladle furnace
®
processing. Calculated with Thermo-calc and SLAG2
database [154].
of inclusion modification and the correct amount of calcium
needed to achieve the desired changes have also been stud-
ied to a large extent [161–165] and are currently relatively well
understood. The inclusions formed are complex, frequently
forming layers of sulfur rich compounds around oxides as well
as other distributions, as exemplified in Fig. 15.
Since the objective of the treatment is to have a significant
fraction of the inclusions in the liquid state as they pass the
casting valves and nozzles, the process becomes more difficult
as steels are cast at lower temperatures [166]. The distribution
of phases in the formed NMIs is further complicated by their
solidification, as shown in Fig. 16.
The amount of calcium needed to modify oxides depends
on the total oxygen content in the steel. To date, there is no
method of determining total oxygen in time to use it for the
decision on the amount of calcium to be added. This poses an
important problem in industrial practice. The use of thermo-
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299 293

Fig. 15 – Complex calcium aluminate inclusion found in slabs treated with calcium to ensure good castability. Sulfide and
MgO were probably dissolved in the liquid inclusion during casting. Mg source is probably reduction from MgO in slag (see
discussion in item 4.2). Courtesy of C. Cicutti, Tenaris-Siderca.

5.0
A+CA6
4.5
CA2+CA6

4.0
CA2+L
3.5
ppm O

3.0 Liquid
inclusions
2.5

2.0 Ot=30ppm

Ot=15ppm Ot=20ppm
1.5
Fig. 16 – Complex calcium aluminate inclusion in fractured
sample from as-cast slab. Dendritic solidification is visible 1.0
0 5 10 15 20 25 30
in the outer layer of the inclusion. Courtesy of C. Cicutti,
Tenaris Siderca. ppm ca

Fig. 17 – Calculated oxygen in solution steel containing

dynamics to understand clogging and define the “castability
window” is routine [28,163,164,167–169]. As soluble oxygen
can be readily measured, it can be used as tool to control the
0.025%Al, 0.01%S, and various contents of Ot (15, 20 and
30 ppm, left to right) as a function of the Ca content at
1540 ◦ C. Non-metallic phases present indicated for each
region of composition. Calculated with Thermo-Calc and
®

efficiency of calcium treatment, as shown in Fig. 17. SLAG3 database. Points are experimental measurements of
oxygen in solution in heats without clogging events.
4.4. Oxide metallurgy
Adapted from [45].

One of the concerns in near net-shape continuous casting is

the need for techniques of microstructural control that do
not rely heavily on thermomechanical processing. The use
of tailored NMIs as nuclei for ferrite in weld metal has been
well demonstrated [170–172]. Based on the current theories for
the effect of inclusions on nucleation, inclusions that would
deplete the surrounding matrix on manganese seem to be
effective. Thus, the development of NMI oxides that have high
solubility for or nucleate MnS and will hence cause the desired
effect has been demonstrated [121,173–175]. Furthermore, the
titanium deoxidation of silicon manganese steels, changing
inclusions to titanium oxides [176] or oxides and nitrides [177],
has also been demonstrated to be highly effective for the
nucleation of ferrite. The work of Koseki, Inoue, Suito and Park
demonstrated that titanium nitride can be a powerful nucle-
ation agent promoting the formation of a larger equiaxed grain
region in stainless steels both in cast products [30,178,179] and
in welds [180,181]. Park and Kang [182] have recently reviewed
the developments in this area. It is evident that alloy and
process design in oxide metallurgy can greatly benefit from
computational thermodynamics [183–185] and modeling [186].
5. Inclusion quantification
By the last decades of the XX Century it was clear that the
classical methods of measuring and quantifying inclusions
by comparison of images with charts were not fulfilling the
needs of the steel community. To improve inclusion quan-
294 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
tification (size, volume fraction and composition information)
several new methods and practices have been proposed and
have been reviewed in [64,187–192]. It is apparent that in many
cases more than one practice must be used to properly charac-
terize and understand the effect of process on NMIs and then
to correlate their effect with properties. This is understandable
since some properties seem to depend on the overall distribu-
tion of inclusions and others, such as fatigue, depend on the
maximum sized inclusion. Furthermore even excluding low-
end products, the range of cleanness of steel products varies
significantly. Low-carbon aluminum killed (LCAK) steel have
around 40 ppm total oxygen content while bearing steels with
around 5 ppm total oxygen are now typical. This results in a
dramatic difference in volume fraction of oxide inclusions, not
to mention sulfides. Extreme value statistics and its applica-
tion are of paramount importance in fatigue. These methods
are seldom included in usual reviews. For a review including
these methods, the reader is referred to ref. [188]. The method
of inclusion rating using statistics of extremes adopted in the
procedure proposed by Murakami [193] is widely used in the
area of fatigue, with very good results. It must be mentioned
that, in this method, there would be no limit to maximum
inclusion size, since it increases with the volume of steel used
in the calculations [188]. This is somewhat incompatible with
the expectations of steelmakers.
6. Conclusions
Never the classical aphorism “Look after the slag, and the
metal will look after itself” has been so well understood
and followed in the processing of steels. In the last decades,
the understanding of the effects of NMIs in steel properties
has caused steelmaking process to evolve from preventing
“unavoidable” contamination to the optimization of NMIs
composition, amount and distribution in the steel products.
This has affected all steps in melt shop processing. Raw
material selection (e.g., to avoid aluminum contamination),
slag composition design, optimization of secondary refining
conditions (e.g., time of processing and fluid dynamic condi-
tions), rigorous control of tundish and mold operations and
rigorous measures to control reoxidation in all steps of pro-
cessing have become standard in quality steelmaking.
In this development, thermodynamics has played a deci-
sive role. The understanding of the interplay between
thermodynamics, steel and slag chemical composition, and
melt shop processing conditions has evolved dramatically.
The modeling tools available have also improved significantly,
so that important knowledge for inclusion control in steels
can now be developed on a scientific basis. These techniques
are now widely used in process development and in tailor-
ing NMIs. This ability places a continuous demand on the
improvement of the understanding of slag and NMIs proper-
ties.
Inclusion engineering is now a reality in a great number
of steels and steel plants and steel cleanness has improved at
least an order of magnitude in the last decades. This improve-
ment has posed significant challenges to those involved in
NMI qualification and quantification, essential to relate quan-
titatively NMI population to product properties and behavior.
This is a vast field and will be discussed in a future publication.
Thus, notwithstanding all the advances summarized in
this review, important problems and questions remain if the
quality of steels is to continue improving in the next decades.
Conflicts of interest
The author declares no conflicts of interest.
Acknowledgements
The author thanks all his former and present students for
inspiration and important contributions to the work on non-
metallic inclusions in steels. The collaboration of colleagues
in many steelmakers and research organizations is gratefully
acknowledged.
The support of CNPq, CAPES and FAPERJ during different
stages on the author’s work is also acknowledged.
references
[1] Kay DAR. Inclusions engineering. CIM Bull 1982;75(842):130.
[2] Kiessling R. Non-metallic inclusions in steels. London: The
Iron and Steel Institute; 1968.
[3] Sauveur A. Metallography and heat treatment of iron and
steel. 4th ed. New York, NY: McGraw-Hill; 1935.
[4] Turkdogan ET, Kor GJW. Sulfides and oxides in Fe–Mn alloys:
Part I. Phase relations in Fe–Mn–S–O system. Metall Trans
1971;2(June):1561–70.
[5] Kor GJW, Turkdogan ET. Sulfides and oxides in Fe–Mn alloys:
Part III. Formation of oxysulfides during freezing of steel.
Metall Trans 1972;3(June):1269–78.
[6] Gaye H. Chapter 3, Inclusion formation in steels. In: Cramb
A, editor. The making, shaping and treating of steel, casting
volume. 11th ed. Pittsburgh, PA: AISE; 2003.
[7] Kohn A. The use of radioisotopes in metallurgical research
and industry. Metall Rev 1958;3(1):143–201.
[8] Poirier J. A review: influence of refractories on steel quality.
Metall Res Technol 2015;112(4):410–1.
[9] Lehmann J, Boher M, Kaerlé MC. An experimental study of
the interactions between liquid steel and MgO-based
tundish refractories. CIM Bull 1971;90(1013):69–74.
[10] Ruby-Meyer F, Hénault E, Rocher-Bakour M, Merchi F.
Improvement of inclusion cleanness in bearing steels and
Ca-treated steels. In: 7th international conference on Clean
Steel 4-662007 Balatonfüred, Hungary OMBKE & IOM 2007.
2007.
[11] Kiessling R. The influence of non-metallic inclusions on the
properties of steel. J Met 1969;21(October):47–54.
[12] Colpaert H, Costa e Silva A. Metalografia dos produtos
siderúrgicos comuns. 4th ed. São Paulo: Editora Blucher;
2008. p. 672.
[13] Chipman J. Another look at the problem of steel
deoxidation. Met Prog 1949;(August):211–21.
[14] Ward RG. An introduction to the physical chemistry of iron
& steelmaking. London: Edward Arnold; 1962.
[15] IISI. IISI study on clean steel. Brussels, Belgium:
International Iron and Steel Institute; 2004.
[16] Turkdogan ET. Principles of steelmaking. London: The
Institute of Materials; 1996.
[17] Oeters F. Metallurgy of steelmaking. Duesseldorf: Verlag
Stahleisen; 1989.
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
[18] Hino M, Ito K, editors. Thermodynamic data for
steelmaking, the 19th committee for steelmaking, JSPS.
Sendai: Tohoku University Press; 2010.
[19] Kang Y, Thunman M, Sichen D, Morohoshi T, Mizukami K,
Morita K. Aluminum deoxidation equilibrium of molten
iron-aluminum alloy with wide aluminum composition
range at 1873 K. ISIJ Int 2009;49(10):1483–9.
[20] Costa e Silva A, Beneduce F, Avillez RR. Liquid iron
deoxidation by aluminum: a brief review of experimental
data and thermodynamic description. Tecnol Metal Mater
Min 2015;12(4):267–73.
[21] Paek M-K, Do K-H, Kang Y-B, Park J-H, Pak J-J. Inclusion
thermodynamics for high-Al high-Mn advanced high
strength steels. In: Proceeding of clean steel in future – Prof
Hae-Geon Lee symposium. Pohang, Korea: Postech; 2012.
[22] Costa e Silva A. Calcium and magnesium thermodynamics
in steel and its impacts on secondary steelmaking: a
computational thermodynamics approach. Cah Inf
Tech/Rev Metall 2008;105(4):181–93.
[23] Kang Y-B, Kim HS, Zhang J, Lee H-G. Practical application of
thermodynamics to inclusions engineering in steel. J Phys
Chem Solids 2005;66(2–4):219–25.
[24] Jung I-H, Decterov SA, Pelton A. A thermodynamic model
for deoxidation equilibria in steel. Metall Mater Trans B
2004;35B(June):493–504.
[25] Jung I-H. Overview of the applications of thermodynamic
databases to steelmaking processes. Calphad
2010;34(3):332–62.
[26] Costa e Silva A. Thermodynamic aspects of inclusion
engineering in steels. Rare Met 2006;25(5):412–9.
[27] Costa e Silva A. Applications of multicomponent databases
to the improvement of steel processing and design. J Phase
Equilibria Diffus 2017;38(6):916–27.
[28] Park J-H, Kim DS, Lee S-B. Inclusion control of ferritic
stainless steel by aluminum deoxidation and calcium
treatment. Metall Mater Trans B 2005;36B(February):67–73.
[29] Kang Y-B, Lee H-G. Inclusions chemistry for Mn/Si
deoxidized steels: thermodynamic predictions and
experimental confirmations. ISIJ Int 2004;44(6):1006–15.
[30] Park J-H. Effect of inclusions on the solidification structures
of ferritic stainless steel: computational and experimental
study of inclusion evolution. Calphad 2011;35(4):455–62.
[31] Gaye H, Riboud PV, Welfringer J. Use of slag model to
describe slag-metal reactions and precipitation of
inclusions. Ironmak Steelmak 1988;15(6):319–22.
[32] Riboud PV, Gatellier C, Gaye H, Pontoire JN, Rocabois P.
Examples of physical chemistry approach to high quality
steel processing. ISIJ Int 1996;36:S22–5.
[33] Sigworth GK, Elliott JF. The conditions for nucleation of
oxides during the silicon deoxidation of steel. Metall Trans
1973;4:105–13.
[34] Miyashita Y. Change of the dissolved oxygen content in the
process of silicon deoxidation. Tetsu Hagane – J Iron Steel
Inst Jpn 1966;52(7):1049–57.
[35] Ishihara S.Rep Proc 15th Annual Conf. International Iron
and Steel Institute; 1981. p. 224–53, cited by Lange KW.
Thermodynamics and kinetic aspects of secondary
steelmaking processes. International Materials Reviews.
1988;33(2):53–89.
[36] Kaushik P, Pielet H, Yin H. Inclusion characterisation; tool
for measurement of steel cleanliness and process control:
Part 2. Ironmak Steelmak 2009;36(8):572–82.
[37] Suito H, Inoue H, Inoue R. Aluminum oxygen equilibrium
between CaO–Al2 O3 melts and liquid iron. ISIJ Int
1991;31(12):1381–8.
[38] Asano K, Nakano T. Deoxidation of molten steel with
deoxidizer containing aluminum. Tetsu Hagane – J Iron
Steel Inst Jpn 1971;57(13):1943–52.
295
[39] Braun TB, Elliott JF, Flemings MC. The clustering of alumina
inclusions. Metall Trans B 1979;10(2):171–84.
[40] Steinmetz E, Lindenberg HU. Morphology of inclusions at
aluminum deoxidation. Arch Eisenhuttenwes
1976;47(4):199–204.
[41] Steinmetz E, Lindenberg HU, Morsdorf W, Hammerschmid
P. Growth shapes of aluminum-oxides in steels. Arch
Eisenhuttenwes 1977;48(11):569–74.
[42] Dekkers R [Doctor of Science thesis] Non-metallic
inclusions in liquid steel. Belgium: Katholieke Universiteit
Leuven; 2002.
[43] Van Ende M-A [Applied Science doctor thesis] Formation
and morphology of non-metallic inclusions in aluminium
killed steels. Belgium: Katholieke Universiteit Leuven; 2010.
[44] Tiekink W, Boom R, Overbosch A, Kooter R, Sridhar S. Some
aspects of alumina created by deoxidation and reoxidation
in steel. Ironmak Steelmak 2010;37(7):488–95.
[45] Batista RP, Martins AA, Costa e Silva A. The effects of
deoxidation practice on the quality of thin foil low-carbon
steel. J Min Metall Sect B: Metall 2017;53(3B):357–63.
[46] Cicutti C, Valdez M, Pérez T, Ares R, Panelli R, Petroni J.
Optimisation of calcium treatment to improve castability.
In: 18th process technology conference. ISS; 2001. p. 871–82.
[47] Yin H, Shibata H, Emi T, Suzuki M. Characteristics of
agglomeration of various inclusion particles on molten
steel surface. ISIJ Int 1997;37(10):946–55.
[48] Yin H, Shibata H, Emi T, Suzuki M. In-situ observation of
collision agglomeration and cluster formation of alumina
inclusion particles on steel melts. ISIJ Int 1997;37(10):936–45.
[49] Wikström J, Nakajima K, Shibata H, Tilliander A, Jönsson P.
In situ studies of the agglomeration phenomena for
calcium–alumina inclusions at liquid steel–liquid slag
interface and in the slag. Mater Sci Eng A
2008;495(1–2):316–9.
[50] Wikström J, Nakajima K, Shibata H, Tilliander A, Jönsson P.
In situ studies of agglomeration between Al2 O3 –CaO
inclusions at metal/gas, metal/slag interfaces and in slag.
Ironmak Steelmak 2008;35(8):589–99.
[51] Lindskog N, Sandberg H. Studies on the deoxidation
mechanism in a stirred steel bath using radioactive
isotopes. Scand J Metall 1973;2:71–8.
[52] Lindskog N. Studies of Al deoxidation in ASEA-SKF furnaces
using radioactive Zr. Scand J Metall 1975;4:153–60.
[53] Fuhr F, Torga G, Medina F, Cicutti C. Application of slag
tracers to investigate source of non-metallic inclusions.
Ironmak Steelmak 2007;34(6):463–70.
[54] Song M, Nzotta M, Sichen D. Effect of ladle slag of previous
heat on number of non-metallic inclusions in ladle
treatment of tool steel. Ironmak Steelmak 2011;(8):584.
[55] Burty M, Dunand P, Ritt J, Soulard H, Blanchard A, Jeanne G,
et al. Control of DWI steel cleanness by lanthanum tracing
of deoxidation inclusions, ladle slag treatment and a
methodical approach. In: Ironmaking conference
proceedings. 1997. p. 711–7.
[56] Mills KC, Fox AB, Li Z, Thackray RP. Performance and
properties of mould fluxes. Ironmak Steelmak
2005;32(1):26–34.
[57] Zhang L, Taniguchi S. Fundamentals of inclusion removal
from liquid steel by bubble flotation. Int Mater Rev
2000;45(2):59–82.
[58] Yokoya S, Takagi S, Iguchi M, Asako Y, Hara S. Removal of
inclusion through bubble curtain created by swirl motion in
submerged entry nozzle. ISIJ Int 1998;38(10):1086–92.
[59] Wang L, Lee H-G, Hayes P. A new approach to molten steel
refining using fine gas bubbles. ISIJ Int 1996;36(1):17–24.
[60] Tivelius B, Sohlgren T, Wretlind C. The processing of z-steel
in an ASEA-SKF ladle furnace. In: Secondary steelmaking:
proceedings of a conference/organized by Activity Group
296 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
Committee V (Ironmaking and Steelmaking) of the Metals
Society, and held at the Royal Lancaster Hotel, London, on
5–6 May 1977. London, England: The Metals Society; 1977.
[61] Suzuki K, Kitamura K, Takenouchi T, Funazaki M, Iwanami
Y. Manufacturing of high-purity steels by ladle refining. Iron
Steelmak 1982;9(7):33–7.
[62] Neifer M, Rodl S, Bannenberg N, Lachmund H.
Untersuchungen zur Bewegung und Abscheidung von
Einschlüssen in gasgerührten Stahlschmelzen. Stahl Eisen
1997;117(5):55–63.
[63] Ek M, Wu L, Valentin P, Sichen D. Effect of inert gas flow rate
on homogenization and inclusion removal in a gas stirred
ladle. Steel Res Int 2010;81(12):1056–63.
[64] Zhang L, Thomas BG. State of the art in evaluation and
control of steel cleanliness. ISIJ Int 2003;43(3):271–91.
[65] Zhang L, Thomas BG. Alumina inclusion behavior during
steel deoxidation. In: 7th European electric steelmaking
conference. Milano, Italy: Associazione Italiana di
Metallurgia; 2002. p. 2–77.
[66] Matsuno M, Kikuchi Y, Komatsu M, Arai M, Watanabe K,
Nakashima H. Development of a new deoxidation
technique for RH degassers. Iron Steelmak 1993;20(7):35–8.
[67] Ogawa K. Slag refining for production of clean steel. In:
143rd–144th Nishiyama memorial seminar. Tokyo: ISIJ;
1992. p. 137–66.
[68] Engh TA. A fluid mechanical model of inclusion removal.
Scand J Metall 1975;4:49–58.
[69] Kim S-H, Fruehan RJ. Physical modeling of liquid/liquid
mass transfer in gas stirred ladles. Metall Trans B
1987;18(2):381–90.
[70] Kim J-S, Kawakami M, Tanida K. Effect of Ar bubbling on
removal rate of alumina inclusion in Al deoxidized steel.
Tetsu Hagane – J Iron Steel Inst Jpn 1995;81(3):167–72.
[71] Kaushik P, Pielet H, Yin H. Inclusion characterisation – tool
for measurement of steel cleanliness and process control:
Part 1. Ironmak Steelmak 2009;36(8):561–71.
[72] Nadif M, Burty M, Soulard H, Boher M, Pussé C, Lehmann J,
et al. Control of steel re-oxidation and CC nozzle clogging.
In: IISI study on clean steel. Brussels, Belgium:
International Iron and Steel Institute; 2004. p. 87–164.
[73] Lachmund H, Xie Y. High purity steels: a challenge to
improved steelmaking processes. Ironmak Steelmak
2003;30(2):125–9.
[74] Arth G, Viertauer A, Hackl G, Krumpel G, Petritz B, Meure D.
Tundish technology and processes: ladle to mould systems
and solutions (Part I). RHI Bull 2016;(1):39–44.
[75] Repetylo O, Olette M, Kozakevitch P. Deoxidation of liquid
steel with aluminum and elimination of the resulting
alumina. J Met 1967;19(5):45–9.
[76] Olette M. Surface phenomena: a cornerstone of iron and
steelmaking processes. ISIJ Int 1993;33(10):1113–24.
[77] Rocabois P, Lehmann J, Gatellier C, Teres JP. Non-metallic
inclusion entrapment by slags: laboratory investigation.
Ironmak Steelmak 2003;30(2):95–100.
[78] Nakajima K, Okamura K. Inclusion transfer behavior across
molten steel–slag interface. In: Proc 4th Int Conf on Molten
Slags and Fluxes. Sendai, Japan: ISIJ; 1992. p. 505–10.
[79] Bouris D, Bergeles G. Investigation of inclusion
re-entrainment from the steel–slag interface. Metall Mater
Trans B 1998;29(3):641–9.
[80] Strandh J, Nakajima K, Eriksson R, Jönsson P. Solid
inclusions transfer at a steel–slag interface with focus on
tundish conditions. ISIJ Int 2005;45(11):1597–606.
[81] Shannon G, Wang Y, Vantilt S, Coletti B, Blanpain B, Sridhar
S. Observation of behavior of oxide inclusions at molten
slag/steel interfaces. In: Proceedings of the VII international
conference on molten slags, fluxes and salts. 2004. p. 571–6.
[82] Shannon G, White L, Sridhar S. Modeling inclusion
approach to the steel/slag interface. Mater Sci Eng A
2008;495(1–2):310–5.
[83] Reis BH, Bielefeldt WV, Vilela ACF. Absorption of
non-metallic inclusions by steelmaking slags – a review. J
Mater Res Technol 2014;3(2):179–85.
[84] Valdez M, Shannon GS, Sridhar S. The ability of slags to
absorb solid oxide inclusions. ISIJ Int 2006;46(3):450–7.
[85] Shannon GN, Sridhar S. Modeling Al2 O3 inclusion
separation across steel–slag interfaces. Scand J Metall
2005;34(6):353–62.
[86] Lee SH, Tse C, Yi KW, Misra P, Chevrier V, Orrling C, et al.
Separation and dissolution of Al2 O3 inclusions at
slag/metal interfaces. J Non-Cryst Solids 2001;282:41–8.
[87] Park J-H, Jung I-H, Lee H-G. Dissolution behavior of Al2 O3
and MgO inclusions in the CaO–Al2 O3 –SiO2 slags: formation
of ring-like structure of MgAl2 O4 and Ca2 SiO4 around MgO
inclusions. ISIJ Int 2006;46(11):1626–34.
[88] Park S-C [Doctor of Philosophy thesis] Inclusion control in
steels: dissolution behavior of MgO and MgAl2 O4 particles
into the CaO–Al2 O3 –SiO2 –MgO slags and interfacial tension
between molten iron and CaO–SiO2 –MgO–Al2 O3 –FeO slag
system. Pohang, Korea: Pohang University of Science; 2008.
[89] Zhang P, Seetharaman S. Dissolution of MgO in
CaO–“FeO”–CaF2 –SiO2 slags under static conditions. J Am
Ceram Soc 1994;77(4):970–6.
[90] Yan P, Webler BA, Pistorius PC, Fruehan RJ. Nature of MgO
and Al2 O3 dissolution in metallurgical slags. Metall Mater
Trans B 2015;46(6):2414–8.
[91] Holappa L, Kekkonen M, Louhenkilpi S, Hagemann R,
Schröder C, Scheller P. Active tundish slag. Steel Res Int
2013;84(7):638–48.
[92] Beskow K, Jia J, Lupis CHP, Sichen D. Chemical
characteristics of inclusions formed at various stages
during the ladle treatment of steel. Ironmak Steelmak
2002;29(6):427–35.
[93] Frank LA. Castability – from alumina to spinels (and more).
In: 2001 Electric Arc Furnace Conf. ISS-AIME; 2001.
p. 403–16.
[94] Cramb AW, Rastogi R, Maddalena RL. Nozzle clogging. In:
Cramb A, editor. The making, shaping and treating of steel,
casting volume. 11th ed. Pittsburgh, PA: AISE; 2003.
[95] Mizoguchi T, Ueshima Y, Sugiyama M, Mizukami K.
Influence of unstable non-equilibrium liquid iron oxide on
clustering of alumina particles in steel. ISIJ Int
2013;53(4):639–47.
[96] Zhang L, Thomas BG. State of the art in the control of
inclusions during steel ingot casting. Metall Mater Trans B
2006;37(5):733–61.
[97] Ragnarsson L, Ek M, Eliasson A, Sichen D. Flow pattern in
ingot during mould filling and its impact on inclusion
removal. Ironmak Steelmak 2010;37(5):347–52.
[98] Mills KC. Structure and properties of slags used in the
continuous casting of steel: Part 1. Conventional mould
powders. ISIJ Int 2016;56(1):1–13.
[99] Zhang L. Nucleation, growth, transport, and entrapment of
inclusions during steel casting. JOM 2013;65(9):1138–44.
[100] Thomas BG. Modeling of continuous casting defects related
to mold fluid flow. Iron Steel Technol 2006;3(7):
127.
[101] Sumida M. Control of steel re-oxidation and CC nozzle
clogging. In: IISI study on clean steel. Brussels, Belgium:
International Iron and Steel Institute; 2004. p. 183–207.
[102] Zhang L, Wang Y, Zuo X. Flow transport and inclusion
motion in steel continuous-casting mold under submerged
entry nozzle clogging condition. Metall Mater Trans B
2008;39(4):534–50.
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
[103] Zhang L, Thomas BG. Inclusions in continuous casting of
steel. In: XXIV National Steelmaking Symposium. Citeseer;
2003. p. 28.
[104] Sichen D. Modeling related to secondary steel making. Steel
Res Int 2012;83(September (9)):825–41.
[105] Burty M, Europäische Kommission, org. Development of
advanced methods for the control of ladle stirring process:
contract no. 7210-PR/330, final report. Luxembourg: Office
for Official Publ. of the European Communities; 2007. p. 139.
[106] Dannert C, Research Fund for Coal and Steel. Improvement
of ladle stirring to minimise slag emulsification and
reoxidation during alloying and rinsing (Stimprove).
Bruxelles: European Commission; 2012.
[107] Burty M, Pussé C, Bertoletti C, Wetta P, Cariola E. KETTLOR:
efficient stirring in ladle metallurgy. Cah Inf Tech/Rev
Metall 2006;103(11):493–9.
[108] Sims CE, Lillieqvist GA. Inclusions – their effect, solubility
and control in cast steel (with discussion). Trans AIME
1932;100:154–95.
[109] Pickering FB. Formation of inclusions in rimming steels.
Steel Times 1964;189:686–92.
[110] Nilles P. Theoretical study of solidification of rimming steel.
J Iron Steel Inst 1964;202(7):601.
[111] Turkdogan ET. Causes and effects of deoxidation occurring
during cooling and solidification of steel. Trans Metall Soc
AIME 1965;233:2100–12.
[112] Turkdogan ET. Deoxidation of steel – what happens from
tap to solidification. J Met 1967;(1):38–44.
[113] Turkdogan ET, Grance RA. Microsegregation in steel. J Iron
Steel Inst 1970:482–94.
[114] Flemings MC. Solidification processing. New York:
McGraw-Hill; 1974.
[115] Yarwood JC, Flemings MC, Elliott J. Inclusion formation in
the Fe–O,S system. Metall Trans 1971;2:2573–82.
[116] Flemings MC. (B) Formation of oxide inclusions during
solidification. Int Metall Rev 1977;22(1):201–7.
[117] Matsumiya T, Kajioka H, Mizoguchi S, Ueshima Y, Esaka H.
Mathematical analysis of segregations in continuously-cast
slabs. Trans Iron Steel Inst Jpn 1984;24(11):873–82.
[118] Ueshima Y, Mizoguchi S, Matsumiya T, Kajioka H. Analysis
of solute distribution in dendrites of carbon steel with
delta/gamma transformation during solidification. Metall
Trans B 1986;17B(December):845–59.
[119] Ueshima Y, Komatsu N, Mizoguchi S, Kajioka H. Effects of
alloying elements on interdendritic microsegregation of
carbon steel. Tetsu Hagane – J Iron Steel Inst Jpn
1987;73(11):1551.
[120] Matsumiya T. Mathematical analyses of segregations and
chemical compositional changes of nonmetallic inclusions
during solidification of steels. Mater Trans JIM
1992;33(9):783–94.
[121] Mizoguchi S [Ph.D. thesis] A study on segregation and oxide
inclusions for the control of steel properties. University of
Tokyo; 1996.
[122] Gaye H, Gatellier C, Nadif M, Riboud PV, Saleil J, Faral M.
Reaction metal-latier et controle de la composition des
inclusions residuelles en metallurgie secondaire. Cah Inf
Tech/Rev Metall 1987;(November):759–71.
[123] Gatellier C, Gaye H, Lehmann J, Bellot J, Moncel M. Inclusion
control in low-aluminum steels. Cah Inf Tech/Rev Metall
1992;(April):362–9.
[124] Stouvenot F, Gaye H, Gatellier C, Lehmann J. Secondary
steelmaking slag treatment for inclusions control in
semi-killed steels. In: Electric furnace conference
proceedings. Warrendale, PA: ISS; 1994. p. 423–8.
[125] Wintz M, Bobadilla M, Lehmann J, Gaye H.
Experimental-study and modeling of the precipitation of
297
nonmetallic inclusions during solidification of steel. ISIJ Int
1995;35(6):715–22.
[126] Wintz M, Bobadilla M, Lehmann J. Macrosegregation and
precipitation of nonmetallic inclusions during solidification
of steel: a modelling and experimental study. In: Beech J,
Jones H, editors. Solidification processing 1997 – 4th
Decennial international conference on solidification
processing. Sheffield, UK: University of Sheffield; 1997. p.
226–9.
[127] Capurro C, Cicutti C. Characterization of titanium nitrides
formed during steel solidification. In: Anais do Seminário
de Aciaria, Fundição e Metalurgia de Não-Ferrosos. São
Paulo Brasil: Editora Blucher; 2017. p. 368–78.
[128] Escobar DP, Castro CSB, Borba EC, Oliveira AP, Camei K,
Taiss EJM, et al. Efeito do caminho de solidificação na
microestrutura e na precipitação de (Ti,Nb)(C,N) em um aço
HTP com baixo Mn – alto Nb. In: 72o Congresso Anual da
ABM. São Paulo, SP: Editora Blucher; 2017. p. 2598–609.
[129] Shibata H, Yin H, Yoshinaga S, Emi T, Suzuki M. ln-situ
observation of inclusions in steel melt engulfment and
pushing by advancing melt/solid of nonmetallic interface.
ISIJ Int 1998;38(2):149–56.
[130] Stefanescu DM, Catalina AV. Calculation of the critical
velocity for the pushing/engulfment transition of
nonmetallic inclusions in steel. ISIJ Int 1998;38(5):503–5.
[131] Juretzko FR, Stefanescu DM, Dhindaw BK, Sen S, Curreri PA.
Particle engulfment and pushing by solidifying interfaces:
Part 1. Ground experiments. Metall Mater Trans A
1998;29(6):1691–6.
[132] Stefanescu DM, Juretzko FR, Catalina A, Dhindaw BK, Sen S,
Curreri PA. Particle engulfment and pushing by solidifying
interfaces: Part II. Microgravity experiments and theoretical
analysis. Metall Mater Trans A 1998;29(6):1697–706.
[133] Kaptay G. Interfacial criterion of spontaneous and forced
engulfment of reinforcing particles by an advancing
solid/liquid interface. Metall Mater Trans A
2001;32(4):993–1005.
[134] Bigelow LK, Flemings MC. Sulfide inclusions in steel. Metall
Trans B 1975;6B:275–83.
[135] Oikawa K, Ishida K, Nishizawa T. Effect of titanium addition
on the formation and distribution of MnS inclusions in
steel during solidification. ISIJ Int 1997;37(4):332–8.
[136] Oikawa K, Ohtani H, Ishida K, Nishizawa T. The control of
the morphology of MnS inclusions in steel during
solidification. ISIJ Int 1995;35(4):402–8.
[137] Mitsui H, Oikawa K, Ohnuma I, Kainuma R, Ishida K.
Morphology of sulphide formed in the Fe–Cr–S ternary
alloys. ISIJ Int 2002;42(11):1297–302.
[138] Ban-ya S. Preface – special issue on physics and chemistry
of molten slags. ISIJ Int 1993;33(1):1.
[139] Shiwaku K, Yamada Y, Koarai J, Kawaguchi Y. Improvement
of fatigue life of valve spring wire by reducing non-metallic
inclusions, SAE technical paper 850364. SAE Trans
1985;94–85, 3.32–3.43.
[140] Oshiro T, Ikeda T, Matsuyama H, Okushima S, Oki Y, Ibaraki
N. Verbesserung der Dauerhaltbarkeit von Ventilfederdraht.
Stahl Eisen 1989;10(21):1011–5.
[141] Murakami Y. Metal fatigue: effects of small defects and
nonmetallic inclusions. Kidlington, England: Elsevier
Science; 2002.
[142] Onoe T, Ito S, Ogawa K, Mimura T, Matsumoto H, Maeda S.
Shape control of inclusions for steel tire cord (development
in ladle arc refining). Trans ISIJ 1987;27:B249.
[143] Maeda S, Soejima T, Saito T, Matsumoto H, Fujimoto H,
Mimura T. Shape control of inclusions in wire rods for high
tensile tire cord by refining with synthetic slag. In:
Steelmaking conference proceedings. 1989. p. 379–85.
298 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
[144] Suito H, Inoue R. Thermodynamics on control of inclusions
composition in ultra-clean steels. ISIJ Int 1996;36(5):528–36.
[145] Jung I-H, Kang Y-B, Decterov SA, Pelton AD. Thermodynamic
evaluation and optimization of the MnO–Al2 O3 and
MnO–Al2 O3 –SiO2 systems and applications to inclusion
engineering. Metall Mater Trans B 2004;35(2):259–68.
[146] Zhang L. State of the art in the control of inclusions in tire
cord steels – a review. Steel Res Int 2006;77(3):158–69.
[147] Chen S, Jiang M, He X, Wang X. Top slag refining for
inclusion composition transform control in tire cord steel.
Int J Miner Metall Mater 2012;19(6):490–8.
[148] Li SX. Effects of inclusions on very high cycle fatigue
properties of high strength steels. Int Mater Rev
2012;57(2):92–114.
[149] Hoo JJC, Green WB, editors. American Society for Testing
and Materials. STP1327: bearing steels: into the 21st
century. West Conshohocken, PA: ASTM; 1998. p. 524.
[150] Bhadeshia HKDH. Steels for bearings. Prog Mater Sci
2012;57(2):268–435.
[151] Monnot J, Heritier B, Cogne JY. Relationship of melting
practice, inclusion type, and size with fatigue resistance of
bearing steels. In: ASTM STP 987: effect of steel
manufacturing processes on the quality of bearing steels.
1988. p. 149.
[152] Gatellier C, Zbaczyniak Y, Duplomb G, Damie PH, Saleil J,
Cecconi R. Control of steelmaking practices for bearing
steels in Ascometal. Cah Inf Tech/Rev Metall
1993;90(4):479–86.
[153] Ma W-J, Bao Y-P, Wang M, Zhao DW. Influence of slag
composition on bearing steel cleanness. Ironmak Steelmak
2014;41(1):26–30.
[154] Haug-Warberg T, Chen L-Q, Kattner U, Hallstedt B, Costa e
Silva A, Lee J-H, et al. Thermodynamics. In: Handbook of
software solutions for ICME. Weinheim, Germany:
Wiley-VCH Verlag; 2016.
[155] Luyckx L, Bell JR, McLean A, Korchynsky M. Sulfide shape
control in high strength low alloy steels. Metall Trans
1970;1(12):3341–50.
[156] Tipnis VA, Joseph RA, Doubrava JH. Calcium deoxidized
improved machining steels for automotive gears. SAE
technical paper 730115. In: 1973 International Automotive
Engineering Congress and Exposition. SAE – Society of
Automotive Engineers; 1973.
[157] Mercier M [Master of Science dissertation] Adição Dinamica
de CaSi. Belo Horizonte, Brazil: UFMG; 2014.
[158] Lu D [Doctor of Philosophy thesis] Kinetics mechanisms
and modelling of calcium treatment of steel. Canada:
McMaster University; 1992.
[159] Abraham S, Bodnar R, Raines J. Inclusion engineering and
the metallurgy of calcium treatment. Iron Steel Technol
2014;(July):57–69.
[160] Basak S, Kumar Dhal R, Roy GG. Efficacy and recovery of
calcium during CaSi cored wire injection in steel melts.
Ironmak Steelmak 2010;37(3):161–8.
[161] Verma N, Pistorius PC, Fruehan RJ, Potter M, Lind M, Story S.
Transient inclusion evolution during modification of
alumina inclusions by calcium in liquid steel: Part I.
Background, experimental techniques and analysis
methods. Metall Mater Trans B 2011;42(4):711–9.
[162] Verma N, Pistorius PC, Fruehan RJ, Potter M, Lind M, Story
SR. Transient inclusion evolution during modification of
alumina inclusions by calcium in liquid steel: Part II. Results
and discussion. Metall Mater Trans B 2011;42(4):720–9.
[163] Holappa L, Hamalainen M, Liukkonen M, Lind M.
Thermodynamic examination of inclusion modification
and precipitation from calcium treatment to solidified
steel. Ironmak Steelmak 2003;30(2):111–5.
[164] Bielefeldt WV, Vilela ACF, Moraes CAM, Fernandes PC.
Computational thermodynamics application on the
calcium inclusion treatment of SAE 8620 steel. Steel Res Int
2007;78(12):857–62.
[165] Guo J, Cheng S, Cheng Z, Xin L. Thermodynamics for
precipitation of CaS bearing inclusion and their
deformation during rolling process for Al-killed Ca-treated
steel. Steel Res Int 2013;84(6):545–53.
[166] Lino RE, Marins AMF, Marchi LA, Mendes JA, Penna LV, Neto
JGC, et al. Influence of the chemical composition on steel
casting performance. J Mater Res Technol 2017;6(1):50–6.
[167] Pielet H, Bhattacharya D. Thermodynamics of nozzle
blockage in continuous casting of calcium-containing
steels. Metall Mater Trans B 1984;15(3):547–62.
[168] Holappa L. Inclusion control for castability of resulphurized
steels. In: ISS-AIME ironmaking conference, ironmaking
conference proceedings. ISS-AIME; 2001. p.
1023–35.
[169] Botelho T, Medeiros G, Ramos GL, Costa e Silva A. The
application of computational thermodynamics in the
understanding and control of clogging and scum in
continuous casting of steel. J Phase Equilibria Diffus
2017;38(3):201–7.
[170] Ramsay CW, Matlock DK, Olson DL. The influence of
inclusions on the microstructures and properties of a high
strength steel weld metal. In: David SA, Vitek JM, editors.
Recent trends in welding science and technology TWR89.
Gatlinburg, TN/Materials Park, OH: ASM International; 1989.
p. 763–7.
[171] Stlaurent S, Lesperance G. Effects of chemistry, density and
size distribution of inclusions on the nucleation of acicular
ferrite of C–Mn steel shielded-metal-arc-welding
weldments. Mater Sci Eng A 1992;149(2):203–16.
[172] Koseki T, Thewlis G. Overview Inclusion assisted
microstructure control in C–Mn and low alloy steel welds.
Mater Sci Technol 2005;21(8):867–79.
[173] Ogibayashi S. Advances in technology of oxide metallurgy.
Nippon Steel Tech Rep 1994;(61):70–6.
[174] Sarma DS, Karasev AV, Jönsson PG. On the role of
non-metallic inclusions in the nucleation of acicular ferrite
in steels. ISIJ Int 2009;49(7):1063–74.
[175] Yan P [Doctor in Engineering dissertation] Formation and
behavior of Mn containing oxysulphide inclusions during
desulphurization, deoxidation and alloying. Belgium:
Katholieke Universiteit Leuven; 2014.
[176] Byun JS, Shim JH, Cho YW, Lee DN. Non-metallic inclusion
and intragranular nucleation of ferrite in Ti-killed C–Mn
steel. Acta Mater 2003;51(6):1593–606.
[177] Kikuchi N, Nabeshima S, Kishimoto Y, Matsushita T, Sridhar
S. Effect of Ti de-oxidation on solidification and
post-solidification microstructure in low carbon high
manganese steel. ISIJ Int 2007;47(9):1255–64.
[178] Ito A, Suito H, Inoue R. Size distribution of multi-phase
deoxidation particles for heterogeneous crystallization of
TiN and solidification structure in Ti-added ferritic stainless
steel. ISIJ Int 2012;52(7):1196–205.
[179] Inoue H, Koseki T. Clarification of solidification behaviors in
austenitic stainless steels based on welding process.
Nippon Steel Tech Rep 2007;95:62–70.
[180] Koseki T, Inoue H. TiN equiaxed solidification of steel
nucleating on titanium nitride. J Jpn Inst Met
2001;65(7):644–51.
[181] Inoue H, Koseki T. Solidification mechanism of austenitic
stainless steels solidified with primary ferrite. Acta Mater
2017;124:430–6.
[182] Park J-H, Kang Y. Inclusions in stainless steels – a review.
Steel Res Int 2017;88(12):1700130.
 j m a t e r r e s t e c h n o l . 2 0 1 8;7(3):283–299
[183] Ruby-Meyer F, Lehmann J, Gaye H. Thermodynamic analysis
of inclusions in Ti-deoxidised steels. Scand J Metall
2000;29(5):206–12.
[184] Costa e Silva A. An overview of the use of CALPHAD
methods in steelmaking. J Min Metall 1999;35B(1):
85–112.
[185] Jung I-H, Decterov SA, Pelton AD. Computer applications of
thermodynamic databases to inclusion engineering. ISIJ Int
2004;44(3):527–36.
[186] Liu Z, Gu K, Cai K. Mathematical model of sulfide
precipitation on oxides during solidification of Fe–Si alloy.
ISIJ Int 2002;42(9):950–7.
[187] Bartosiaki BG, Pereira JAM, Bielefeldt WV, Vilela ACF.
Assessment of inclusion analysis via manual and
automated SEM and total oxygen content of steel. J Mater
Res Technol 2015;4(3):235–40.
[188] Atkinson HV, Shi G. Characterization of inclusions in clean
steels: a review including the statistics of extremes
methods. Prog Mater Sci 2003;48(5):457–520.
299
[189] Kaushik P, Lowry M, Yin H, Pielet H. Inclusion
characterisation for clean steelmaking and quality control.
Ironmak Steelmak 2012;39(4):284–300.
[190] Jacobi H, Wünnenberg K. Methods for determination of
oxide cleanness in steel. In: IISI study on clean steel.
Brussels, Belgium: International Iron and Steel Institute;
2004.
[191] Kaushik P, Lehmann J, Nadif M. State of the art in control of
inclusions, their characterization, and future requirements.
Metall Mater Trans B 2012;43(4):710–25.
[192] Cicutti C, Capurro C. Steel cleanliness evaluation
techniques review. Application to different industrial cases.
In: 21a Conferencia del Acero IAS. 2016.
[193] Murakami Y, Toriyama T, Coudert EM. Instructions for a
new method of inclusion rating and correlations with the
fatigue limit. J Test Eval 1994;22(4):318–26.