the effect of ladle treatment on steel cleanness in tool...
TRANSCRIPT
The Effect of Ladle Treatment on
Steel Cleanness in Tool Steels
Karin Steneholm
Doctoral Thesis
Stockholm 2016
Department of Materials Science and Engineering
School of Industrial Engineering and Management
KTH, Royal Institute of Technology
SE-100 44 Stockholm, Sweden
Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm,
framlägges för offentlig granskning för avläggande av Teknologie Doktorsexamen, fredag den
10 juni 2016, kl 10:00 i B3, Brinellvägen 23, Kungliga Tekniska Högskolan, KTH,
Stockholm.
ISBN 978-91-7595-951-1
i
Karin Steneholm The Effect of Ladle Treatment on Steel Cleanness in Tool Steels
Department of Materials Science and Engineering
School of Industrial Engineering and Management
KTH, Royal Institute of Technology
SE-100 44 Stockholm
Sweden
ISBN: 978-91-7595-951-1
Printed by: Universitetsservice US-AB, Stockholm 2016
© Karin Steneholm
ii
First say to yourself what you would be; and then do what you have to do.
- Epictetus
iii
TABLE OF CONTENTS
ABSTRACT ............................................................................................................................................ vi SAMMANFATTNING ......................................................................................................................... viii ACKNOWLEDGEMENTS ..................................................................................................................... x PREFACE ............................................................................................................................................... xi AUTHOR CONTRIBUTION ................................................................................................................ xii 1. INTRODUCTION ......................................................................................................................... 13
1.1 Change of inclusion characteristics during ladle refining and casting .................................. 13
1.2 Refractory .............................................................................................................................. 14
1.3 Top slag ................................................................................................................................. 14
1.4 Focus of this thesis ................................................................................................................ 15
2. THEORY....................................................................................................................................... 17 2.1 Inclusion Size Distribution .................................................................................................... 17
2.2 Thermodynamic examinations .............................................................................................. 17
2.3 Removal rates of impurities................................................................................................... 19
2.4 Steel cleanliness in the industry ............................................................................................ 21
3. EXPERIMENTAL ........................................................................................................................ 23 3.1 Plant description .................................................................................................................... 23
3.2 Plant trials .............................................................................................................................. 23
3.3 Sample evaluations ................................................................................................................ 23
3.4 Thermodynamic calculations................................................................................................. 23
4. RESULTS AND DISCUSSION ................................................................................................... 25 4.1 Change of top slag (supplement 1) ........................................................................................ 25
4.1.1 Inclusion content ........................................................................................................... 25
4.1.2 Inclusion composition .................................................................................................... 26
4.1.3 Thermodynamic calculations ......................................................................................... 27
4.2 Evolution of inclusion characteristics during vacuum degassing (supplement 2) ................. 29
4.2.1 Total oxygen content...................................................................................................... 29
4.2.2 Inclusion content ........................................................................................................... 30
4.2.3 Inclusion composition .................................................................................................... 30
4.2.4 Desulphurization rate .................................................................................................... 31
4.3 Removal of Hydrogen, Nitrogen and Sulphur (supplement 3) .............................................. 33
iv
4.3.1 Hydrogen removal ......................................................................................................... 33
4.3.2 Sulphur removal ............................................................................................................ 33
4.3.3 Nitrogen removal ........................................................................................................... 35
4.4 The role of process control (supplement 4) ........................................................................... 37
4.4.1 Steel properties .............................................................................................................. 37
4.4.2 Slag properties .............................................................................................................. 38
4.4.3 Deoxidation ................................................................................................................... 39
4.4.4 Oxygen activity calculations .......................................................................................... 40
5. CONCLUDING DISCUSSION .................................................................................................... 43 6. CONCLUSIONS ........................................................................................................................... 45 7. FUTURE WORK .......................................................................................................................... 47 8. REFERENCES .............................................................................................................................. 48
v
vi
ABSTRACT
The aim of the present work was to get an overview of the steel cleanness in tool steel.
Studies of three steel grades were done with different focuses.
Firstly the change of the inclusion characteristics during the vacuum degassing in the ladle
was looked upon. The results are based on plant trials as well as on thermodynamic
calculations. Trials were made on two different tool steel grades. During the trials, the top
slag composition was altered and steel and slag samples were collected before and after
vacuum degassing. The steel samples were analysed to determine the chemical composition of
steel and slag as well as the inclusion numbers. With the aid of thermodynamic calculations,
the oxygen activity for the equilibria steel/inclusion and steel/top slag was calculated. The
data was compared for different calculations as well as to measured oxygen activities from the
steel melts. The results showed that the top slag composition has an effect on the inclusion
composition. When the CaO-content in the top slag was increased, the CaO-content in the
inclusion was increased as well. The thermodynamic calculations indicated that the oxygen
activity for the steel/inclusion was twice as large as that of the steel/top slag before vacuum
degassing. However, after vacuum degassing the oxygen activity values were close for the
steel/inclusion and steel/top slag equilibriums.
In order to see how the inclusion characteristics evolved during vacuum degassing the
vacuum treatment process was interrupted at five pre-determined time points. At each
interruption of the vacuum degassing, steel and slag samples were obtained. Thereafter, the
total oxygen, inclusion content and composition were determined. Before vacuum treatment
the inclusion content was high in number. Thereafter, during vacuum the number of the
smaller inclusions was decreased. However, for the larger inclusions (>11.2µm) the number
of inclusions had increased. It was also noticed that the inclusion composition varied in
samples taken before vacuum degassing. However, throughout the degassing process the
inclusion composition was found to approach the top slag composition. Finally, at the end of
the degassing operation only one type of inclusion composition was found in the steel melt.
Also, after around 10 minutes of degassing, the inclusions reached their minimum (or plateau)
values.
The removal of hydrogen, nitrogen and sulphur is also of great importance during a vacuum
degassing treatment in order to obtain a clean steel. The three elements were studied by using
samples taken before, during (at five predetermined time points) and after vacuum degassing.
The chemical steel and slag compositions were used to calculate the removal of hydrogen,
nitrogen and sulphur. It was found that the removal kinetics of hydrogen and nitrogen can be
described with first-order reactions. This is valid for the studied steel grade, as long as the
sulphur content is below 0.003 wt-%. When 10 minutes of vacuum degassing have passed, the
removal of hydrogen and nitrogen is more or less finished for the studied steel grades. In the
case of sulphur it is not possible to apply a first-order reaction. Instead the sulphur content
seems to follow the equilibrium sulphur content at all stages during the degassing operation.
vii
Controlling the process is of great importance when reaching a good cleanliness of the steel.
Two steel grades were studied, steel grades with similar process route and with only a few
differences in steel composition. Steel and slag samples were collected and analysed. The sum
of FeO and MnO was found to be a clear indicator for when reoxidation had taken place.
However, the amount of carry-over slag from the electric arc furnace could not be predicted
by a conclusive indicator. Also the oxygen activity was calculated and the calculations were
compared with the measurements made in the liquid steel. The results indicate that
calculations of oxygen activities with multivalence slag species such as Fe and Cr requires
measurements for validations to ensure that reliable results are obtained.
Keywords: oxygen activity, inclusion characteristics, vacuum degassing, desulphurization,
slag, ladle, hydrogen, nitrogen, deslagging
viii
SAMMANFATTNING
Syftet med denna studie var att få en överblick av verktygsståls renhet. Studier på tre olika
stålsorter utfördes.
Inneslutningsbilden under vakuumbehandling undersöktes. Resultaten är baserade på
industriförsök likväl som termodynamiska beräkningar. Försöken gjordes på två olika
stålsorter där toppslaggens sammansättning varierades och stål- och slaggprover tog ut före
och efter vakuumbehandling. Proverna analyserades för att bestämma den kemiska
sammansättningen på stål och slagg, men inneslutningsbilden undersöktes också. Med hjälp
av termodynamiska beräkningar beräknades syreaktiviteten för jämvikterna stål/inneslutning
och stål/toppslag. Beräkningarna jämfördes sinsemellan samt med uppmätt syreaktivitet från
stålsmältan. Resultaten visar att toppslaggens sammansättning har en påverkan på
inneslutningarnas sammansättning. Specifikt så fanns att när CaO-halten i toppslaggen ökar
syns en ökning av CaO i inneslutningssammansättningen. De termodynamiska beräkningarna
visar att syreaktiviten för jämvikten stål/inneslutning är dubbelt så stor som den för jämvikten
stål/topplagg före vakuumbehandling. Dock är syreaktiviteterna för de olika jämvikterna lika
efter vakuumbehandlingen.
För att studera hur inneslutningsbilden utvecklas under vakuumbehandling avbröts
vakuumbehandlingen vid fem förutbestämda tidpunkter. Vid varje avbrott av
vakuumbehandlingen togs stål- och slaggprover som analyserades för att bestämma totalsyre,
inneslutningsmängd och inneslutningssammansättning. Före vakuumbehandlingen var antalet
inneslutningar stort, medan antalet små inneslutningar minskade under vakuumbehandlingen.
Däremot konstaterades att antalet stora inneslutningarna (>11.2µm) ökade under avgasningen.
Vad beträffar inneslutningssammansättningen syntes en variation i analys före
vakuumbehandlingen som sedan förändrades till att likna toppslaggens sammansättning under
vakuumbehandlingstiden. I slutet av processen märktes endast en typ av
inneslutningssammansättning i stålsmältan. Vad som också noterades var att efter ungefär 10
minuter av vakuumbehandling så hade inneslutningsantalet nått sitt minimumvärde.
För att erhålla rent stål är väte-, kväve- och svavelreningen också av stor vikt under
vakuumbehandlingen. Dessa tre element studerades genom att prover togs vid olika tillfällen i
processen; före, under (vid fem olika förutbestämda tidpunkter) och efter vakuumbehandling.
De kemiska sammansättningarna för stål- och slaggproverna användes för att beräkna väte-,
kväve- och svavelreningarna. Beräkningarna visade att för väte och kväve så kan kinetiken
beskrivas med en förstagradsekvation. Specifikt så gäller detta för stålsorten i studien, samt
under förutsättningen att svavelhalten är lägre än 0,003 vikts-%. Dessutom så visade
resultaten att väte- och kvävereningen är i stort sett klar för den studerade stålsorten efter 10
minuters vakuumbehandling. När det gäller svavel så kan en förstagradsekvation inte
användas vid beräkningarna Istället uppvisar svavel en tendens att följa jämviktshalten av
svavel genom hela avgasningsprocessen.
Genom att kontrollera processen finns stora möjligheter att erhålla en bra renhet på stålet. Två
stålsorter studerades i denna specifika studie, där de hade en liknande processväg samt endast
ix
några små skillnader i stålanalysen. Stål- och slaggprover samlades in och analyserades.
Summan av FeO och MnO visade sig vara en klar indikation på när reoxidation har skett, men
mängden ”carry-over” slagg från ljusbågsugnen kunde inte predikteras med hjälp av någon
specifik indikator. Dessutom så beräknades syreaktiviten och beräkningarna jämfördes sedan
med uppmätt syre i stålet. Resultaten indikerar att beräkningar av syreaktivitet med
toppslagger som innehåller multivalenselement, såsom Fe och Cr, kräver valideringar med
mätningar för att trovärdiga predikteringar ska erhållas.
x
ACKNOWLEDGEMENTS
First of all I would like express my sincere gratitude and appreciation to my supervisor
Professor Pär Jönsson. You have given me inspiration, support and encouragement throughout
this journey.
I am also thankful for the support and advices I have been given by my co-supervisor, Anders
Tilliander.
I especially would like to thank Uddeholms AB and Bruksägare C J Yngström’s scholarship
for financing the writing of this thesis.
Many thanks go to my colleagues at Uddeholms AB for helping me and for keeping my spirit
up and giving me energy.
My beloved parents, thank you.
Karin Steneholm
Stockholm, 2016
xi
PREFACE
The present thesis is based on the following papers:
Supplement 1: “Effect of top slag composition on inclusion characteristics during vacuum
degassing of tool steel”
K. Steneholm, M. Andersson and M. Nzotta and P Jönsson
Steel Research International, Vol. 78, issue 7, 2007, pp. 522-530
Supplement 2: “Change of Inclusion Characteristics During Vacuum Degassing of Tool
Steel”
K.M. Steneholm, M.A.T. Andersson and P.G. Jönsson
Steel Research International, Vol. 77, issue 6, 2006, pp. 392-400
Supplement 3: “Removal of Hydrogen, Nitrogen and Sulphur from Tool Steel during
Vacuum Degassing”
K. Steneholm, M. Andersson, P.G. Jönsson, and A. Tilliander
Ironmaking & Steelmaking, Vol. 40, issue 3, 2013, pp.199-205
This paper received the Adrian Normanton Award from the Institute of Metals in London,
which is a price for the best paper published in Ironmaking and Steelmaking during 2013.
Supplement 4: “The role of process control on steel cleanliness”
K. Steneholm, Nils Å. I. Andersson, A. Tilliander and P.G. Jönsson.
Manuscript.
Parts of this work have been presented at the following conferences:
I: “Influence of Process Step on Inclusion Characteristics”
K. Steneholm
Nordic Symposium for young Scientists, Production Metallurgy for Iron,
Steel and Ferroalloys, June 11-13, 2003, Pohto/Oulu, Finland
II: Influence of Process Parameters during Vacuum Degassing on the Non-Metallic Inclusion
Content in Tool Steel
K. Steneholm, M. Andersson
STÅL2004, 5-7 May, 2004, Borlänge, Sweden
xii
AUTHOR CONTRIBUTION
Supplement 1: Project planning, literature survey, plant trials, microscopic analysis, analysis
of thermodynamic calculations and major part of writing.
Supplement 2: Project planning, literature survey, plant trials, microscopic analysis and
major part of writing.
Supplement 3: Project planning, literature survey, plant trials, analysis of thermodynamic
calculations and major part of writing.
Supplement 4: Project planning, literature survey, plant trials, analysis of thermodynamic
calculations and major part of writing.
13
1. INTRODUCTION
Today’s steel production is more and more specialized, due to higher demands from the
customers as well as due to an increased competition. It is therefore important to obtain a
deeper insight to how the process can be controlled, in order to make clean steel that meets
the demands on material properties from customers. Here, the making of clean steel includes a
control of the inclusion characteristics, which includes the size, composition and morphology
of inclusions. Furthermore, a control of sulphur, hydrogen and nitrogen is of importance. The
present thesis is focused on tool steelmaking and specifically for the conditions that are valid
for Uddeholms AB. The research at this plant has increasingly focused on an increased
understanding of especially how the inclusion characteristics change during different parts of
the process from the steelmaking to the final product.
At Uddeholms AB as well as at other tool steel plants it is known that especially the
inclusions have a negative influence on material properties that are essential for tool steel
grades. Specifically, the chemical compositions of the inclusions are important to determine,
since they will affect how the inclusions behave during a deformation of the steel. Thus, these
inclusions will also influence the material properties of the finished steel product. One of the
most important material properties in tool steels is the polishability. Previous research has
shown that the polishability increases with a decreased inclusion content [1]. It should also be
mentioned that the important relationship between inclusions and material properties has, for
example, been discussed by Wijk [2]. In 1995, he reported the importance to control the
amount, size, size distribution and composition of non-metallic inclusions in order to reach
optimum steel properties to obtain a steel performance within tight specifications.
1.1 Change of inclusion characteristics during ladle refining and casting
In order to make a change of how an improved cleanness of the steel can be obtained it is
important to initially map how especially the inclusion characteristics changes in different
parts of the process. For example, Cheng et al [3] has studied how the inclusion characteristics
changes during the ladle refining and casting operations. The focus was on macro inclusions
present in steel during different parts of the ladle treatment and casting process, since these
are most detrimental to the material properties. The results showed that clusters were
observed before vacuum degassing, so it was concluded that macro inclusions were formed
due to collisions. In addition, the results showed that the major sources of macro inclusions
during casting were mold flux entrainments and refractory breakdowns. Several other studies
have also been carried out to at KTH to determine how the inclusion characteristics vary
during ladle treatment, based on plant trials [4-7].
When discussing studies of how the inclusion characteristics vary with process steps, it is also
of importance to discuss how reliable the samples are when determining the inclusion
characteristics. Typically, the inclusion sizes are determined using light optical microscopy
and the inclusion characteristics is determined using scanning electron microscopy in
14
combination with energy dispersive spectroscopy. The knowledge from previously published
research on the influence of sampling on determining the inclusion characteristics in process
samples has been carefully considered when carrying out the current studies in this thesis [8-
10]. When discussing sampling it is also of importance to consider when and where it is
suitable to take a sample. Therefore, Söder et al [11] studied the concentration gradients of
inclusions during both gas and induction stirring. They found that the number of inclusions
varied both in relation to sampling position and time.
It also need to mentioned that studies have been made to determine the effect of deoxidation
praxis and ferroalloy additions on the inclusion characteristics [12-15], - desulphurisation and
inclusion characteristics related to the top slag [16-23], - stirring conditions related to
inclusion characteristics [24-30]. All these aspects will, of course, influence how the inclusion
characteristics changes during different parts of the process,
1.2 Refractory
The effect of the refractory on the inclusion content in the steel during vacuum degassing has
been studied by researchers from Ovako Steel. More specifically, in two separate studies
[31,32] they found that MgO and carbon in the magnesite lining react as follows during
vacuum degassing:
MgO(s) + C(s) => Mg(g) + CO(g) (1)
This magnesium gas will dissolve into the steel and react with alumina inclusions due to that
magnesium has a higher affinity for oxygen than for aluminium. The inclusion sizes and
compositions during vacuum degassing can also be changed due to reactions with slag present
on the refractory wall from the previous heat, defined as ladle glaze [33].
1.3 Top slag
The top slag represents another source of oxygen that can be transferred to steel during ladle
refining. This is especially true during sulphur refining during vacuum degassing. For
example, oxygen is transferred from the slag to the steel according to the following reaction:
CaO(s) + S => CaS(s) + O (2)
where O is the new oxygen dissolved into the steel. This oxygen will react with strong
deoxidizers in the steel such as Al, Si, Ti etc under the formation of new inclusions.
Furthermore, the oxygen can also participate in reactions which will modify already existing
inclusions. The intense stirring during a vacuum degassing operation also give rise to
reactions between strong deoxidizers such as Al and oxides in the slag such as FeO, MnO and
SiO2. These leads to the formation of new inclusions exemplified by the following reaction
involving FeO:
15
3FeO(l) + 2Al => 3Fe(l) + Al2O3 (3)
The influence of the top slag composition on the inclusion composition was also studied by
Almcrantz et al [22]. The focus was the effect of three different slag compositions with
different viscosities on the inclusion composition during vacuum degassing. The major
conclusion was that the magnesium content in the alumina-based inclusions increased during
vacuum degassing using high viscosity top slags. This was due to that the magnesium gas that
was released through equation (1) reacted with alumina inclusions present in the steel.
Almcrantz et al [22] also found that the calcium content in the alumina-based inclusions
increased when using low-viscosity slags. This was believed to be due to a more intense
mixing between slag and steel for low-viscosity slags compared to high-viscosity slags, which
in turn promoted reactions between slag and steel. Here, it should be noted that Almcrantz et
al’s [22] investigation focused on bearing steels, while the present focus is tool steels. These
tool steels have different steel compositions and thus different thermodynamic conditions
compared to bearing steels.
1.4 Focus of this thesis
Based on the results from
the above mentioned
studies as well as the
experience from
Uddeholms AB it was
decided to focus most of
the work on the vacuum
degassing part of the ladle
refining process. During
this vacuum degassing, the
content of dissolved gases
in the steel melt such as
hydrogen and nitrogen can
be decreased. Furthermore,
carbon and sulphur can be
removed to the gas phase
(as CO) and the slag phase,
respectively. In addition,
vacuum degassing is an
important unit process for
removal and control of non-metallic inclusions.
This thesis is based on four supplements. The focus of the supplements is ladle refining and
especially the vacuum degassing step, as mentioned above. More specifically, the studies are
16
focused on the influence of the operation praxis on the inclusion characteristics as well as on
unwanted elements. The first supplement concerns the effect of a changed composition of the
top slag on the inclusion characteristics. Specifically, the chemical compositions of steel,
slags and inclusions as well as the size distribution of inclusions are determined in samples
taken during ladle refining. Supplement 1 also discusses thermodynamic calculations of
steel/slag and steel/inclusion equilibria using both Wagner’s theory and Thermo-Calc
simulations. Supplement 2 focuses on the change of the inclusion characteristics during
vacuum degassing. The purpose is to provide new information on how inclusion
characteristics vary for different conditions during the degassing operation. Note, that in order
to obtain this information the normal experimental praxis has been altered. More specifically,
the vacuum degassing operation has been interrupted at different times for different heats. To
the authors’ knowledge, this type of plant trial has not been reported previously related to
experiments focused on inclusion characteristics. However, it should be noted that Hallberg et
al [34] have reported on trials where the vacuum degassing operation was terminated in order
to obtain information on hydrogen and sulphur contents in steel.
The third supplement focuses on the refining of hydrogen, nitrogen and sulphur during the
vacuum degassing part of the ladle refining process. Note, that the samples used in
supplement 3 have been collected from supplement 2. The objective of the research in
Supplement 3 study was not to obtain a model, but to be able to use the latest theoretical
results to implement them in the production process. Finally, the fourth supplement concerns
an evaluation of the ladle process aiming at optimizing the process in order to reduce the
oxygen content. Specifically, the focus is on the oxygen activity and aluminium content based
on plant trials.
It should also be mentioned that one important motivation for carrying out this thesis is to
provide new information that would assist in the implementation of a process model that is
under construction at Uddeholms AB. It is the belief that a good process model could be used
to make predictions of what will happen if certain actions are taken during the process.
However, such a model needs to be carefully validated with experimental data such as those
provided in this thesis.
17
2. THEORY
2.1 Inclusion Size Distribution
Two kinds of steel samples were used to determine the inclusion size distributions, namely
Liquid Sampling Hot Rolling (LSHR) samples [8] and Rapid Solidification (RS) samples
[35]. The LSHR sample provides a larger area to be investigated in a microscope as well as a
possibility to study the sample after hot rolling compared to a RS sample. However, small
inclusions might grow or be separated in the LSHR sample due to a large sample volume and
a longer solidification time. At the same time, the chance of obtaining better results with
respect to the macro inclusions is better with the LSHR sample than with the RS sample due
to the larger sample volume of the former.
The inclusions in the samples taken from the liquid steel were classified according to SS 11
11 16 standard [36] and following the Jernkontoret´s chart II. Based on this chart, the
inclusions can be classified into the following four inclusion types: ductile (A), brittle (B),
brittle-ductile (C) and undeformable (D) inclusions. In addition, these four inclusion types can
be classified then into the following four sub-groups based on the length/width ratios: thin
(T), medium (M), heavy (H) and particular (P). It should also be mentioned that inclusions
found in liquid steel samples not deformed. Therefore, they are commonly only classified as
D type inclusions. Specifically, the following size ranges are commonly used to classify
inclusions detected in samples taken from liquid steel:
28 57. . DT m
57 113. . DM m
113 22 6. . DH m
DP m 22 6.
During the determination of inclusions sizes in the light optical microscope, the PC-MIC
software [37] was used to calculate the number of inclusions per mm2 as well as the area
percentage of inclusions in each analysed sample.
2.2 Thermodynamic examinations
It is clear that several thermodynamic models can be used to calculate slag/metal equilibriums
and inclusion/metal equilibriums. For tool steels, the Wagner equation is less suitable to use
since the alloy content is so high that the liquid metal could not be treated as a dilute solution.
At the same time, the dilute solution assumption is often used to determine the dissolved
element activities in high alloyed steels due to that this approach is so simple to use.
However, in order to make a compare the simplified calculation based on a dilute solution
model it is common to calculate the slag/metal equilibrium by using thermodynamic
softwares such as, for example, Thermo-Calc [38] which has been used in this work.
18
When aluminium is added as a deoxidizer to the liquid steel, the following reaction takes
place if it is assumed that the dissolved oxygen and aluminium in the molten steel and the
alumina in the oxide phase are in equilibrium:
)(32 32 sOAlOAl (4)
where the oxygen activity (aO) and the aluminium activity (aAl) in the molten steel as well as
the alumina activity (aAl2O3) in the oxide phase (inclusion or top slag) can be calculated in the
following manner:
32
32
Al
OAl
OaK
aa
(5)
where K is the equilibrium constant, which can be calculated in the following way based on
the Gibbs free energy:
)exp(RT
GK
(6)
where the Gibbs free energy can be expressed as follows according to Hayes [39]:
TG 714.3861205115 (J/mol) (7)
Henry’s law is used to calculate the aluminium activity in equation (5):
jfajj
% (8)
and Wagner’s equation is used to calculate the activity coefficient fj:
19
ief i
jj %log (9)
where jf is the activity coefficient for element j in the molten steel, %i is the dissolved
elements in the molten steel, and i
je is the interaction parameter for element j .
The thermodynamic software Thermo-Calc [38] is made up of different thermodynamic
models and databases, which may be used for calculations of component activities in oxide
phases. In the present work, calculations were done to determine the alumina activities of the
different oxide inclusion phases and the top slag samples.
2.3 Removal rates of impurities
The removal rates of impurities are often found to obey a first order reaction kinetics. For
example, the thermodynamics and kinetics for refining of hydrogen and sulphur have been
reported by several researchers [21, 34]. The nitrogen removal rate is more difficult to
calculate than the sulphur or hydrogen removal rates. For nitrogen it is necessary to decide
whether the reaction is a first order or a second order reaction.
Typically, a first-order differential equation to describe the refining rate for impurity elements
in steels may be written as follows:
eqXX
M
Ak
dt
Xd%%
%
(10)
where X is the concentration of the impurity element, for example H, N or S. Furthermore, the
parameter k is the mass transfer coefficient, ρ is the steel density, A is the interface area
(effective reaction area) between the steel phase and the refining phase, M is steel mass and
[%X] is the concentration of the impurity element. Also, the parameter [%X]eq is the
hypothetical concentration of the refined impurity element in the steel phase in equilibrium
with the current concentration of the refining phase.
In general, an overall rate constant, β, is often used to replace the effective reaction area A.
This is due to that it is hard to estimate reasonable values of the reaction area A. Specifically,
the rate constant β may be defined as follows:
20
M
Ak
(11)
The equilibrium content of element X in equation (10) may be calculated by using Sieverts’
law, which can be expressed as follows:
2
% X
X
X pf
KX (12)
where KX is expressed for each element of interest. Therefore, equation (12) can be written as
follows for the elements H and N:
RT
G
f
pX x
X
X0
exp% 2 (13)
where 2Xp is the partial pressure of element X in the gas phase, fX is the activity coefficient
for element X, R is the gas constant, and T is the temperature given in Kelvin. In the present
work it was assumed that the partial pressure of element X was equal to the total pressure in
each case. This simplified the calculations.
In order to perform the equilibrium calculations, the interaction parameters, i
He in equation (9)
were taken from Engh [40]. In addition, the standard Gibbs free energy values were taken
from Hayes [39]. During the calculations, equation (10) was solved for each of the studied
elements, namely H, N, S. As a result, the following equation was obtained:
tXX
XXH
eq
eqt
%%
%%ln
0
(14)
where %Xt, %X0 and %Xeq are the element contents in the steel at time t = t, t = 0 and at
equilibrium, respectively. Here, the parameter X is used to represent H, S or N.
21
As mentioned above, some reactions may take place according to second-order kinetics rather
than first-order kinetics. For example, Cho and Rao [41] reported that the nitrogen desorption
is a second order reaction when the sulphur contents in the steel is high. Similar results have
also been reported by Harada and Janke [42]. Specifically, they showed that both first-order
and second-order reactions will take place for steels with high sulphur and oxygen contents.
Based on these results, it was decided to also consider a second order reaction in this work as
well. Specifically, a second-order reaction may be expressed as follows:
22
%%´%
eqNNM
Ak
dt
Nd
(15)
where k´ is the second order mass transfer coefficient.
Deo and Boom [43] have reported that the two models that can be used to predict the mass
transfer control (the first order model) and chemical reaction control (the second order model)
can be used for the same theory as mentioned earlier. Specifically, a first order reaction can be
used for low surfactant concentrations and a second order reaction can be used for high
concentrations. These reactions can be expressed as follows:
1%%
ieq
m
CR NNk
k ; a first-order reaction (16)
1%%
ieq
m
CR NNk
k ; a second-order reaction (17)
where kCR is the chemical reaction coefficient, km is the mass transfer coefficient, iN% is the
interfacial concentration, and eqN% is the nitrogen concentration in the steel in equilibrium
with a nitrogen gas.
2.4 Steel cleanliness in the industry
As mentioned in the introduction, a clean steel grade can be indicated by low contents of
hydrogen, nitrogen, sulphur and oxygen. Also, typically low total oxygen contents in the steel
22
correspond to low oxygen inclusion contents [44]. Here, low total oxygen contents are
achieved by adding elements which have high affinities to oxygen, according to the
Ellingham diagram. This procedure is named a deoxidation process. Typically, a deoxidation
process involves a reaction between a deoxidizer and dissolved oxygen under the formation of
an oxide. In general, this may be expressed by the following reaction:
xMe + yO = MexOy (18)
where Me is the deoxidiser, O is the dissolved oxygen, and MexOy is the oxide. Typically,
deoxidizers such as Si, Al, Ti and C are used in industry. However, in some cases elements
act mot only as a deoxidizer but also as an alloying element, such as in the case of Mn.
Several previous studies have reported the best practice to obtain a low sulphur content during
a refining operation [45-47]. Here, the sulphur removal is often discussed in combination of a
lowering of the oxygen content in the steel. In this study, the focus is on the relation between
the aluminium content in the steel and the removal of sulphur. Here, the sulphur removal
reaction can be expressed as follows:
𝑆 + (𝑂2−)𝑠𝑙𝑎𝑔 = 𝑂 + (𝑆2−)𝑠𝑙𝑎𝑔 (19)
2𝐴𝑙 + 3𝑂 = 𝐴𝑙2𝑂3(𝑠) (20)
These reactions result on that sulphur is removed to the slag and that the dissolved aluminium
reacts with oxygen under the formation of alumina inclusions. Reaction (19) results in that the
dissolved aluminium content in the steel is decreased. Also, Andersson et al. [48] has reported
that the aluminium content can decrease during vacuum treatment due to reactions with FeO,
MnO and SiO2 which are present in the slag. Overall, this decrease of the dissolved
aluminium content in the steel is often taken as a measure of an overall reoxidation that has
taken place during ladle refining.
23
3. EXPERIMENTAL
3.1 Plant description
Uddeholms AB, Hagfors, Sweden is a scrap-based steelmaking company. The scrap is melted
in an electric arc furnace and then transferred to the ladle station in a 65-tonne ladle. At the
ladle station, the steel melt is deoxidized and alloyed. Thereafter, the ladle is transferred to the
vacuum degassing station, where it will undergo vacuum treatment to remove H, N and S.
Finally, the steel is cast using uphill teeming.
3.2 Plant trials
Three tool steel grades were evaluated; i) a modified 420 steel grade (0.25%C, 0.35%Si,
0.55%Mn, 13.3 %Cr, 0.35%Mo, 1.35%Ni, 0.017%Al and 0.12%N), ii) a H13-steel grade
(0.39% C, 1.0% Si, 0.4% Mn, 5.3% Cr, 1.3% Mo and 0.9% V) and iii) a Mn-alloyed
chromium steel. In the first supplement, samples were taken before and after the vacuum
degassing treatment. Furthermore, in the second and third supplements sampling were done
before vacuum degassing and at different times during the vacuum degassing process. Note,
that the normal vacuum treatment was interrupted in order to obtain the samples. The
sampling in the fourth supplement was made before a deslagging (removal of EAF-slag) was
carried out at the ladle station as well as before and after the vacuum degassing operation. The
samples obtained were steel, slag and temperature samples. Also, two kinds of steel samplers
were used, namely the LSHR [8] and the RS [35] samplers. The temperature measurements
were combined with oxygen activity measurements.
3.3 Sample evaluations
In the case of RS-samples, two samples were obtained at each steel sampling. One sample
was used to determine the chemical composition of steel and later on for light optical
microscopy. The other steel sample was used for determining the oxygen, carbon and sulphur
contents. Also, the LSHR-samples were constructed to be sufficient for all evaluation. These
steel samples were analysed by using a light optical microscope (LOM) to establish the
inclusion content. The classification was made using the JK’s chart SS 11 11 16 [36].
Thereafter the chemical composition of the inclusions was determined by using a scanning
electron microscope (SEM) equipped with energy dispersive spectroscopy (EDS).
3.4 Thermodynamic calculations
Thermodynamic calculations were made for supplement 1, 3 and 4. A prediction of the
oxygen activity for the equilibria steel/inclusion (supplement 1) and steel/top slag
(supplement 1,4) as well as a calculation of the sulphur, hydrogen and nitrogen activity
(supplement 3) was made with the use of the data of the steel composition, temperature, bulk
24
and top slag weight. Also, the activities were calculated by using both the Wagner’s equation
(supplement 1,3) and by using a thermodynamic software. The latter was the thermodynamic
software Thermo-Calc [38] which was used together with appropriate databases as explained
in detail in the supplements.
25
4. RESULTS AND DISCUSSION
A short review from the supplements referring to their main topics is shown below. The topics
are the following: the effects of changing the top slag, the evolution during vacuum treatment
time, the removal of H, N, S and process control.
4.1 Change of top slag (supplement 1)
4.1.1 Inclusion content
In supplement 1, the main focus was to study the effect of a changed top slag composition on
the inclusion composition. Samples were taken before and after vacuum degassing for three
heats (A, B, C) of a 420 modified steel. The top slag composition for the two latter heats were
altered in the way that heats B and C had a higher CaO-content and a lower MgO-content than
that of heat A. The compositions of the slag samples were determined and the steel samples
were analysed both with respect to the steel composition and inclusion characteristics. Also, a
comparison of the steel/slag and steel/inclusion equilibria was made by using Wagner’s
theory as well as by using Thermo-Calc simulations.
When looking at the total number of inclusions per mm2 in Figure 1, it can be seen that the
number of inclusions decreases during the vacuum degassing operation. Furthermore, after
vacuum treatment the total amount of inclusions are similar in number for the different top
slag compositions. It can also be seen that the samples taken before vacuum degassing show a
different amount of inclusions. Hence, in this case it does not seem like the different top slag
compositions do affect the total amount of inclusions in the steel.
Figure 1. Total amount of inclusions per mm2 as a function of the process step
0
0,5
1
1,5
2
2,5
3
Before vacuum After vacuum
To
tal
nu
mb
er
of
inc
lus
ion
s/m
m2
A
B
C
26
On the contrary, a correlation between a changed top slag composition and the number of
larger inclusion has been found. Since the larger inclusions are of greater concern for the
manufacturer it is therefore of importance to be aware of eventual risks of creating large
inclusions. In Figure 2 it can be seen that the number of large (>11.2μm) inclusions have
different values before vacuum degassing. For heat A there is a large decrease in the inclusion
number, a 70% loss. However, in the case of heats B and C, there is a slight increase in the
inclusion number. More specifically the gain is 20-30%, which could be due to their similar
top slag compositions.
Figure 2. Number of inclusions (>11.2μm) per mm2 as a function of the process step
4.1.2 Inclusion composition
The inclusion compositions were also examined, which is shown in Table 1. Before vacuum
degassing the inclusion composition varies and it is not similar to the top slag composition.
However, after vacuum treatment the inclusion composition has been altered and resembles
more the top slag composition. Also, there are no variations, since only one type of inclusions
was found. This applies to all three heats. The most common type before vacuum treatment is
type a, which is a single phase Al2O3•MgO spinel inclusion. After vacuum treatment, only the
phases d, f and k were found in heats A, B and C, respectively. Also, the top slag of heat A
has a higher MgO content than the top slags in heat B and C. However, heats B and C have a
higher CaO content than heat A. This is reflected in the inclusion composition after vacuum
degassing.
0
0,01
0,02
0,03
0,04
0,05
0,06
0,07
Before vacuum After vacuum
Nu
mb
er
of
inc
lus
ion
s/m
m2
A
B
C
27
Table 1. Phases of micro inclusions – determined by using SEM-EDS.
Phase Al2O3 SiO2 MnO CaO MgO Cr2O3
Found in
sample*
a 75-78 - - - 22-25 - A1, B1, C1
b 78 - - 7 15 - A1
c 62 2 - 36 - - A1
d 45 6 - 46 4 - A2
e 40 8 - 52 - - B1
f 33 6 - 61 - - B2
g 33 34 - 33 - - C1
h 25 39 23 - - 13 C1
i 63 - - 31 6 - C1
j 49 20 14 - - 17 C1
k 30 12 - 58 - - C2
(*A1, B1 and C1 are samples taken before vacuum treatment, while A2, B2 and C2 are
samples taken after vacuum treatment.)
4.1.3 Thermodynamic calculations
The second part of supplement 1 deals with thermodynamic calculations. Three cases of
equilibria were considered, I: steel melt/inclusion (before and after vacuum degassing), II:
steel melt/top slag (before and after vacuum degassing) and III: steel melt/top slag (after
vacuum degassing). In the first two cases (I and II) the oxygen activity was calculated using
the equilibrium reaction (4). Wagner’s equation was applied for the calculations of the
dissolved element activities together with the Thermo-Calc software to calculate the alumina
activity (aAl2O3). In case III, only the Thermo-Calc software was used to calculate the
metal/top slag equilibrium. Overall, the values obtained in all three cases were not close to
those of the measured values. This could be due to uncertainties during sampling or errors in
the thermodynamic models used. It is seen that the values of the steel/inclusion oxygen
activity is closer to that of the steel/top slag after vacuum degassing compared to the oxygen
activity before vacuum degassing. This can be related to the fact that the inclusion
composition approaches the top slag composition during the vacuum treatment.
28
In Figure 3 the difference between the two cases II and III is shown. The difference between
the two cases is greater for heat A than for heats B and C. One reason could be that heats B
and C are closer in top slag composition compared to heat A.
Figure 3. A comparison of calculated oxygen activity values for the slag/metal
equilibriums in cases II and III (after vacuum degassing). The reference state for the
activity of oxygen is a 1% hypothetical solution.
0
0,2
0,4
0,6
0,8
1
1,2
A2 B2 C2
Ac
tivit
y o
f o
xyg
en
(*1
0^
4)
Sample
Case II Case III
29
4.2 Evolution of inclusion characteristics during vacuum degassing (supplement
2)
Supplement 2 deals with the evolution of inclusion characteristics during vacuum degassing
in H-13 steels. Ten interruptions during vacuum treatment were made during the same amount
of heats. Five different interruptions were made, at 3, 6, 9, 15 and at 22 minutes of vacuum
degassing. The experiment was made in two series, named the a-serie and the b-serie. The aim
was to obtain a first understanding on how the inclusion characteristics change during vacuum
treatment. For each heat, steel, slag and oxygen activity samples were collected before and
during the interruption of the vacuum degassing operation. The steel and slag samples were
examined with respect to their composition. Furthermore, the steel samples were also cut to
enable a determination of the inclusion characteristics inside the steel samples.
4.2.1 Total oxygen content
In Figure 4, the relative decrease of the total oxygen content in the steel is shown. During the
first ten to fifteen minutes of the vacuum treatment the total oxygen is decreased with more
than 50%. However, after fifteen minutes, the total oxygen content seems to increase again.
The analysis of the total oxygen is somewhat uncertain, since a high total oxygen value could
be explained by a large inclusion being present in the piece of steel that is analysed. Hence,
this would result in a large value of the total oxygen content. It should also be noted that each
point in the figure represents one single heat.
Figure 4. Decrease of the total oxygen content in the steel during vacuum treatment
0
0,5
1
1,5
2
2,5
0 5 10 15 20 25
Oto
t(t)
/Oto
t(t=
0)
Time of vacuum treatment (minutes)
( )
30
4.2.2 Inclusion content
Figure 4 can be related to Figure 5 due to that the total oxygen content represents an
approximate measurement of the amount of inclusions. Thus, in Figure 5 the same trend is
noted as was seen in Figure 4. At first, there is a decrease in the number of inclusions. Then,
after some fifteen minutes the decrease has diminished or even turned into a small increase.
The results for the smaller inclusion size classes is the same as those found in Figure 5.
However, not enough data is available for the larger inclusions (>11.2µm) due to a low
number of inclusions, so the results for these inclusions are not statistically reliable.
Figure 5. Decrease of total inclusion content in steel (#/mm2) during vacuum treatment
4.2.3 Inclusion composition
The compositions of the inclusions are shown in Table 2. Before and during the first six
minutes of vacuum treatment, several types of inclusions, (types m-v), were found in the steel
samples. Thereafter, the inclusion compositions seemed to approach the top slag composition,
as represented by the type x inclusions.
0
0,2
0,4
0,6
0,8
1
1,2
0 5 10 15 20 25
#In
clu
sio
n(t
)/#
Inc
lus
ion
(t=
0)
Time of vacuum treatment (minutes)
( )
31
Table 2. Phases of micro inclusions – determined by using SEM-EDS
Type Al2O3 SiO2 CaO MgO Cr2O3 MnO
Found in
sample*
m 69 <1 2 29 a and b (BV)
n 68 32 a (BV, V3)
o 59 <1 32 4 4 a and b (BV)
p 48-50 1.5-2 38-41 7-7.5 0-4
b (BV), a
(BV)
r 25 42 43 8 a (V3)
s 63 <1 1.5 32 3 <1 a (V3)
t 47 1.5 43 9 a (V6)
u 25 22 41 12 a (V6)
v 55 22 1,7 <1 3 16 a (V6)
x 36.5-40 1.5-4 49.5-55 5-7
a and b (V9,
V15, V22)
y 33 18.5 39 10 b (V15)
z 25 30 31.5 14.5 b (V15)
(*BV: before vacuum degassing, V3: stop after 3min of vacuum degassing, V6: stop after
6min, V9: stop after 9min, V15: stop after 15 min, V22: stop after 22min.)
4.2.4 Desulphurization rate
In Figure 6, the desulphurization rate is shown as a function of the vacuum treatment time. It
can be seen that the rate decreases rather fast up to approximately a time of ten minutes of
vacuum treatment. After that, the rate remains constant at a low desulphurization rate.
However, there is one heat which has a very high desulphurization rate. This heat also has a
higher total oxygen content (Figure 4) and a total inclusion content (Figure 5). Furthermore, it
has a different inclusion composition. The other five heats at nine, fifteen and twenty-two
minutes of vacuum degassing contain mainly one singular type of inclusions. However, this
heat (sampled after fifteen minutes of vacuum degassing) contains several different inclusion
types, types x-z. In general, this heat contains inclusions with a higher amount of SiO2. One
possible reason for that may be that the higher desulphurization rate causes a local depletion
of aluminium in the steel. Since silicon is another element with a high oxygen affinity it will
also react with the dissolved oxygen that is supplied during the desulphurization. This will
result in a formation of silica inclusions.
32
Figure 6. Rate of desulphurization as a function of the vacuum treatment time
0
0,00005
0,0001
0,00015
0,0002
0,00025
0 5 10 15 20 25
Time of vacuum treatment (minutes)
ΔS
/Δt
(wt%
/min
)
( )
33
4.3 Removal of Hydrogen, Nitrogen and Sulphur (supplement 3)
This supplement focuses on the removal of hydrogen, nitrogen and sulphur during the vacuum
degassing process. The main goal was to be able to implement the result into the production
process. The sampling series used are described in section 3.2.
4.3.1 Hydrogen removal
In Figure 7 the relative decrease of hydrogen can be seen. During the first ten minutes, almost
70% of the initial hydrogen content is removed. Hence, the dehydrogenation process is very
rapid and the required hydrogen limit for the product is reached well before the vacuum
treatment time of thirty minutes has passed.
Figure 7. Relative decrease of the hydrogen in steel during the vacuum degassing
operation. The two different series representing the two sampling series described in
section 3.2.
4.3.2 Sulphur removal
The relationship between calculated and experimentally determined equilibrium sulphur
concentrations can be seen in Figure 8. Three different slag weights (740 kg, 1000 kg and
1500 kg) were used in the calculations. This was done to cover the normal range of the
amounts of slag used in the process. The data indicates that the steel’s sulphur concentrations
are close to the equilibrium values for the plant trials.
34
Figure 8. Equlibrium and actual sulphur contents
Figure 9 illustrates the desulphurization kinetics for three different amounts of slag mentioned
above. As can be seen, the slopes are very similar, (0.0505; 0.0506 and 0.0423 respectively).
Hence it is very likely that for fairly similar sulphur contents, the amount of slag does not
affect the desulphurization kinetics to a large degree.
0
0,001
0,002
0,003
0,004
0,005
0,006
0,007
0 0,001 0,002 0,003 0,004 0,005 0,006 0,007
[%S]plant trials
[%S]e
quil
ibri
um
740 kg slag
1000 kg slag
1500 kg slag
35
Figure 9. Desulphurization kinetics calculated by using the first order reaction model
for three different slag weights, namely 740kg, 1000kg and 1500kg.
4.3.3 Nitrogen removal
In figure 10, the nitrogen removal kinetics as a function of vacuum treatment time is seen.
Both first-order and second-order equations are presented. The second-order rate equation
was examined as to evaluate if the rate calculations would be improved. However, no
improvement was seen and the R2-value also indicates that there is no difference between the
two rate equations.
y = 0,0506x + 0,407
R2 = 0,5328
y = 0,0423x + 0,3268
R2 = 0,6953
y = 0,0505x + 0,7459
R2 = 0,3872
0
0,2
0,4
0,6
0,8
1
1,2
1,4
1,6
1,8
2
0 5 10 15 20 25
Deg as s ing tim e (m inutes )
740kg
1000kg
1500kg
36
Figure 10. Denitrogenation kinetics calculated using the first and second order reaction
models
It is known that oxygen and sulphur have a large influence on the kinetics of nitrogen
removal. In order to state that the first-order reaction is valid, the oxygen content and sulphur
content should be below 80ppm. For the studied steel grade, the values are well below the
given limits.
37
4.4 The role of process control (supplement 4)
Supplement 4 discusses the relation between some process parameters to the the properties of
steel and slag. The effect of using the same process route for two steel grades having different
properties was studied.
4.4.1 Steel properties
The analyses of the two steels grades (A and B) have three major differences taken into
account in this study, namely Si-, C- and Mn-content. All three elements represent
deoxidizers of different magnitudes. In figure 11 the difference in analysis is seen.
Figure 11. Si-, C-, Mn-contents in the studied steel grades (A and B)
The process route is similar for the studied steel grade, both in the electric arc furnace (EAF)
as in ladle (LF) and vacuum treatment (VD). More specifically, with respect to the melting of
return scrap of the specific steel grade and the process time in the LF and VD reactors.
However, the temperatures are not equal due to the higher liquidus temperature of steel grade
B compared to steel grade A.
0,00
0,50
1,00
1,50
2,00
2,50
3,00
Si C Mn
wt-
%
Si-, C-, Mn-content i steel grade A and B
A
B
38
4.4.2 Slag properties
As mentioned above, the steel grades have a similar process route. This also applies to the
addition of new synthetic slag in the LF. After the steel is tapped into the ladle, the EAF-slag
is raked of and new slag added. Both the amounts and materials are similar for the studied
steel grades. However, as seen in figure 12, the slags of the different steel grades do not
behave in similar ways.
Figure 12. FeO+MnO in slag vs S-content in steel at different process steps after
deslagging and new synthetic slag has been added. Points representing heat A has been
marked to differentiate from the B-heats.
The sulphur content in steel seems to follow the sum of the FeO and MnO contents in the
slag, where a lower content in the slag corresponds to a lower content of sulphur in the steel.
What also is noticeable is that the FeO+MnO contents and sulphur contents are lower for the
A-heat than the B heats at all process steps. One of the B-heats has a higher content of
FeO+MnO, which may be caused by a late addition of scrap in the EAF. The late addition
caused a larger amount of slag as well as an extra addition of oxygen compared to a normal
procedure. The EAF-slag of steel grade B has properties that make it harder to rake off.
Hence, some EAF-slag may be left in the ladle which will result in a higher content of
FeO+MnO in the new synthetic slag.
0
1
2
3
4
5
6
0 0,001 0,002 0,003 0,004 0,005 0,006 0,007 0,008 0,009
FeO
+Mn
O (
wt-
%)
S (wt-%)
Fe+MnO in slag vs S-content in steel
DH02
BV
AV
A A
A
39
4.4.3 Deoxidation
The steel grades have similar deoxidation processes, i.e. the same aluminium additions are
being used. However, the steel grades react differently to this addition since their oxygen
contents in steel and slag differs.
When looking at the aluminium content it can be seen in figure 13 that the aluminium content
is higher at all process steps for steel grade A compared to steel grade B. The loss of
aluminium in steel may be explained by the desulphurization process, which is described in
equation (19).
The dissolved oxygen forms alumina inclusions according to equation (4). However, the
dissolved aluminium may also react with oxygen dissolved when FeO and MnO are being
reduced during vacuum treatment. This can be described as seen in equation (3).
It has been stated earlier that the slags of the B heats have higher FeO+MnO contents than the
A heats. This may be a reason why the aluminium content is lower in the B-heats than in the
A-heats, as seen in figure 13.
Figure 13. Aluminium content at different process steps for all heats
The connection between aluminium, sulphur and oxygen is also visible in figure 14.
0
0,01
0,02
0,03
0,04
0,05
0,06
0,07
0,08
0,09
DH02 BV AV
(wt-
%)
Al-content at different process steps
A1
B1
B2
B3
B4
B5
40
Figure 14. S- and Al-content in steel vs measured oxygen activity at process step AV for
the different steel grades (A, B).
The aluminium contents for the B-heats are lower than for the A-heat and the sulphur content
as well as the oxygen activity are higher in the B-heats compared to the A-heat. This may
indicate that the aluminium content is not high enough to support a deoxidation and
desulphurization, as in the case of steel grade A.
4.4.4 Oxygen activity calculations
Oxygen activity calculations were made to compare with measured oxygen activities. In
figure 15 the calculated oxygen activity at different process steps for steel grade A is shown.
The relation of the calculated activities for process step BD to the measured activity is
approximately 60 to 90% and for the following to steps, the relation is even larger, namely
approximately 15% and 30%, respectively. Between the process steps BD and BV, a
deslagging operation is performed. During this period it is possible for the steel to have an
oxygen pick-up. However, this reoxidation has not been taken into account when making the
calculations.
41
Figure 15. Calculated oxygen activities at different process steps for steel grade A
If the measured values of oxygen activities are assumed to have small uncertainties and hence
being close to the actual oxygen activity, the values may be set as a condition for the oxygen
activity. Then it is possible to compare the activity of FeO in liquid slag with the original
calculations to those using the measured oxygen activity. In figure 16 the comparison of the
two calculated activities is seen. Here, the relative error for activities at process step BD is
small but for process steps BV and AV the error is larger. What is noticeable is that the probes
used for sampling should not be used for temperatures below 1570°C and samplings made at
BD and AV are close to the lower limit of the temperature range. An interpretation of the
result may be that the available oxygen in the system is poorly estimated due to the use of the
assumption that multivalent species, such as Fe and Cr, only comes as a single valence state,
e.g. Fe2+
and not as a mixture of Fe2+
and Fe3+
. Therefore, the Fe2+
/Fe3+
ratio should ideally be
measured to make a clear conclusion with respect to the oxygen activity.
0
5
10
15
20
25
BD BV AV
(pp
m)
Oxygen activities at different process steps for steel grade A
TCFE8+SLAG2 (ppm)
TCOX5 (ppm)
Measured oxygen activity (ppm)
42
Figure 16. Activity of FeO in liquid slag with calculated and measured aO set as the
specific conditions.
0,0001
0,001
0,01
0,1
1
0,0001 0,001 0,01 0,1 1
a(Fe
O),
cal
cula
ted
aO
a(FeO), measured aO
a(FeO), calculated vs measured aO
BD
BV
AV
43
5. CONCLUDING DISCUSSION
This thesis is based on four supplements. The focus of the supplements is ladle refining and
vacuum degassing. More specifically, the influences of the operation praxis in these process
parts on the inclusion characteristics and on the content of unwanted elements. The first
supplement concerns the effect of changing the composition of the top slag and the second
deals with the evolution of the inclusion characteristics during vacuum degassing. The third
supplement focuses on the evolution of the hydrogen, nitrogen and sulphur contents during
vacuum treatment. Finally, the fourth supplement concerns an evaluation of the ladle process
aiming at optimizing the process in order to reduce the oxygen content.
It has been seen that the inclusion composition changes with a changed top slag composition.
For the experiments with an increased CaO-content in the top slag, the CaO-content in the
inclusions was increased. After vacuum treatment, the inclusion composition was found to
approach the composition of the top slag. It is also seen that the inclusions have a more
homogenous composition after vacuum treatment, with only one inclusion type, compared to
before vacuum treatment. The thermodynamic calculations showed that the oxygen activity
for the steel/ inclusion (aO,eq ) was 2-3 times higher than that of the steel/top slag before
vacuum treatment. It was also concluded that the theoretical equilibrium oxygen activity for
the steel/ inclusion (aO,eq ) will become more similar to the oxygen activity for the equilibrium
steel/top slag during the course of vacuum degassing.
The study also showed that the number of inclusions was reduced during the course of the
vacuum degassing operation. However, for the larger inclusion type, DP, there is a trend that
the number of inclusions increases at the end of the treatment. Then, after around 10 minutes
of degassing the inclusion characteristic data have reached their minimum or plateau values. It
was also observed that when the desulphurization rate is high, the total oxygen in the steel
will remain high. This, in turn, will lead to a high number of inclusions.
During vacuum degassing treatment, the removal of hydrogen and nitrogen is more or less
finished after 10 minutes, due to that the target values have been reached. This may be
described by a first-order reaction. However, the sulphur removal in the studied steel grade
seems to follow the equilibrium sulphur content at all times throughout the process. Thus, the
sulphur removal can continue for longer times than that of the hydrogen and nitrogen removal
in order to reach even lower contents.
44
The process control before vacuum degassing will set the conditions for how the outcome
may be. The oxygen activity is dependent on the aluminium content. Therefore, it is necessary
to have control of the addition of aluminium to the steel (the deoxidation process). Also, the
slag from the electric arc furnace may have properties that cause problems to make a complete
deslagging operation. In this case, the remaining electric arc furnace slag will force an
increase in aluminium addition as to keep the oxygen content low.
The results given in this thesis are valuable to use in a process model made to ensure a
stability in the process of steel making during ladle refining. The experiments carried out
using an interrupted vacuum degassing can be used as a way of verifying the process model.
However, since the argon flow is not a true value, it is difficult to know how reliable the exact
numbers are. During the investigations it has also been noted that the argon flow is different
for each heat. However, the stirring seems to be similar when looking at the open eye itself.
Overall, it is necessary to have a stirring that is vigorous for the top slag to have effect on the
inclusion characteristics. With no stirring the top slag has no or little effect on the inclusion
composition.
This study has also shown that the initial conditions before vacuum treatment, such as the
sulphur content, are very important to know. The initial conditions will have a great influence
on the behaviour of oxide inclusions in the steel melt during the degassing. Therefore, a future
process model for vacuum treatment must consider these initial conditions carefully.
45
6. CONCLUSIONS
The focus of this thesis is the effect of the ladle treatment on the steel cleanness in tool steels
based on plant trials. Steel samples have been evaluated with OES to determine the steel
composition, LOM to determine the inclusion size, and SEM in combination with EDS to
determine the inclusion composition. In addition, thermodynamic calculations have been
carried out to determine the oxygen activity in steel. Based on the results in this study the
most important specific conclusions may be summarized as follows:
Supplement 1
When changing the top slag composition, increasing CaO content and decreasing
MgO content it is seen in the change of inclusion composition. Hence, it is clearly
seen that the top slag composition has an influence on the inclusion composition.
The number of inclusions larger than 11.2µm increased slightly during vacuum
degassing.
Supplement 2
During vacuum degassing, after approximately 10 minutes the large inclusions reach a
plateau in minimum number.
When desulphurization rate is high, the total oxygen content will remain high, hence
rendering a high number of inclusions in the steel.
Supplement 3
Sulphur content for the studied steel seems to follow the equilibrium sulphur content
at all stages during vacuum degassing and therefore it is not possible to apply a first
order reaction to calculate the sulphur removal.
Hydrogen and nitrogen removal kinetics can be described by first order reactions for
the studied steel grade, as long as the sulphur content is below 0.003%.
The hydrogen and nitrogen removal is more or less finished after 10 minutes of
vacuum degassing.
Supplement 4
A careful deslagging is of great importance to reach stable aluminium content and
small reoxidation. The latter is indicated by the amount of FeO+MnO in the slag.
46
Results from calculations of oxygen activity show similar values independent of
database used. However, the calculations differ largely from the measured values
which may be due to the assumption of the true oxidation state being disregarded.
47
7. FUTURE WORK
Though some work has been done to investigate the behaviour of the inclusion characteristics
during the vacuum degassing treatment still many unanswered questions remain. In order to
get an even better insight to the area and a deeper understanding of the reasons for the change
of the inclusion composition during vacuum treatment, more studies should be performed.
Based on the results of this thesis, the following studies are suggested as further work:
Make detailed investigations on the mechanism of the change of the inclusion
composition during vacuum treatment.
Make further plant trials to investigate the evolution of the inclusion characteristics
during vacuum degassing, when other process parameters are varied (i.e. initial
sulphur content, ladle age, temperature, previous steel grade in the ladle, etc.)
Study if it is possible to relate the inclusion content to the oxygen activity and -
dissolved oxygen during the ladle treatment process in tool steels.
Study how it is possible to accurately measure and control the argon flow during the
ladle treatment.
Implement the degree of deslagging as a parameter in a future process control model
48
8. REFERENCES
1. K. Bengtsson: Inclusions and polishability, Swedish Symposium on Non-Metallic
Inclusions in Steel, 1981, pp. 450-462.
2. O. Wijk: Proc. Scaninject VII, MEFOS, Luleå, Sweden, June 1995, 35.
3. J. Cheng, R. Eriksson and P. Jönsson, “Determination of Macro Inclusions during
Clean Steel Production”, Ironmaking and Steelmaking, 30, 2003, pp. 66-72.
4. P. Sjödin, P.G. Jönsson, M. Andreasson, and A. Winqvist, ”Oxidic steel cleanness in
high-carbon chromium bearing steel”, Scand. J. Metallurgy, Vol. 26, No. 1,1997, p.41.
5. E. Fuchs and P. Jönsson, "Inclusion characteristics in bearing steel before and after
ingot casting", High Temp. Mat. and Proc., Vol. 19, No. 5, 2000, pp. 333-344.
6. R. Eriksson, P. Jönsson and A. Gustafsson, “Determination of Inclusion
Characteristics in Low-Carbon Steel during Up-Hill Teeming”, Scandinavian J.
Metallurgy, 33, 2004, pp. 160-171.
7. K.M. Steneholm, A.M.T. Andersson, P.G. Jönsson, “Change of Inclusion
Characteristics during Vacuum Degassing of Tool Steel”, Steel Research International,
No.6, 2006, pp. 392-399.
8. T. Hansen, P.G. Jönsson, S.E. Lundberg and K. Törresvoll, ”The Concept of the
Liquid Sampling and Hot Rolling Method for Determination of Macro Inclusion
Characteristics in Steel, Steel Research International, No. 3, 2006, pp. 177-185.
9. O.T. Ericsson, A.V. Karasev and P.G. Jönsson, “Effect of Slag Protection System and
Sample Geometry on Homogeneity of Total oxygen Content in Samples from Liquid
Steel”, Steel Research International, Vol. 82, No. 3, 2011, pp. 222-229.
10. O.T. Ericsson, M. Lionet, A.V. Karasev, R. Inoue and P.G. Jönsson, ”Changes in
inclusion characteristics during sampling of liquid steel”, Ironmaking and
Steelmaking, Vol. 39, No. 1, 2012, pp. 67-75.
11. Mats Söder, Pär Jönsson, Lage Jonsson, Mselly Nzotta, ”An Experimental Study of
Number Concentration Gradients of Inclusions during Deoxidation using Inductive
and Gas Stirring”, Steel Research International, No. 7, 2005, pp. 481-490.
12. L. Huet, P. Jönsson and F. Reinholdsson, "The effect of deoxidation practise on
inclusion characteristics in bearing steel production", Steel Times International, Nov.
1997, pp. 47-50.
49
13. J. Björklund, M. Andersson, M. Nzotta and P. Jönsson, “The Effect of Ladle
Treatment on Inclusion Composition in Tool Steel Production”, Steel Research
International, Vol. 79, No. 4, 2008, pp. 261-270.
14. T. Sjöqvist, M. Göransson, P. Jönsson and P. Cowx,”Influence of Ferromanganese on
Micro-Alloyed Engineering Steel”, Ironmaking and Steelmaking, 30, 2003, pp. 73-80.
15. Andersson M, Jonsson L and Jönsson P, “A Thermodynamic and Kinetic Model of
Reoxidation and Desulphurisation in the Ladle Furnace” ISIJ International, Vol.40,
No.11, 2000, pp. 1080-1088.
16. T. Sjöqvist, S. Jung, P. Jönsson and M. Andreasson, “Influence of Calcium Carbide
Slag Additions on Inclusion Characteristics in Steel”, Ironmaking and Steelmaking,
Vol. 27, No. 5, 2000, pp. 373-380.
17. M. Andersson, M. Hallberg, L. Jonsson and P. Jönsson,”Slag-metal Reactions during
Ladle Treatment with Focus on Desulphurisation”, Ironmaking and Steelmaking, Vol.
29, No. 3, 2002, p. 224.
18. J. Björklund, M. Andersson and P. Jönsson, ” Equilibrium between slag, steel and
inclusions during ladle treatment: comparison with production data”, Ironmaking and
Steelmaking”, Vol. 34, No. 4, 2007, pp. 312-324.
19. K. Steneholm, M. Andersson, M. Nzotta and P. Jönsson, “Effect of Top Slag
Composition on Inclusion Characteristics during Vacuum Degassing of Tool Steel”,
Steel Research International. No.7, 2007, pp. 522-530.
20. Andersson M, Berlin D, Jönsson P, Löwnerts M, “The Influence of Different Calcium-
Based Additions on Desulphurisation and Inclusion Characteristics”, Scandinavian
Journal of Metallurgy, Vol 30, No 3, 2001, pp.127-135.
21. Nzotta M M, Sichen Du, Seetharaman S, “A Study of the Sulfide Capacities of Iron-
Oxide Containing Slags” Metallurgical Trans. B, Vol. 30B, 1999, pp. 909-920.
22. Almcrantz M, Andersson MAT, Jönsson PG, “Determination of Inclusion
Characteristics in the Asea-SKF process using the Modified Spark-Induces OES
Technique as a Complement in Studying the Influence of Top Slag Composition”,
Steel Research International, 76, No9, 2005, pp. 624-634.
23. Dayal P, Beskow K, Björkvall J, Sichen D, “Study of Slag/Metal Interface in Ladle
Treatment”, Ironmaking and Steelmaking, Vol 33, No 6, 2006 , pp. 454-464.
24. K. Malmberg, M. Nzotta, A. Karasev and P.G. Jönsson, ”Optimisation of stirring
conditions during vacuum degassing in order to lower the inclusion content in steel”,
Ironmaking and Steelmaking, Vol. 40, No. 3, 2013, pp. 231-237.
50
25. K. Malmberg; A. Karasev; M. Nzotta; J. Alexis; P. G. Jönsson, “Influence of final
stirring treatment on inclusion number in tool steel”, Ironmaking and Steelmaking,
Vol. 40, No. 6, 2013, pp. 407-412.
26. C. Medioni, P. Jönsson, D. Sichen, "Effect of stirring practice on sulphur and nitrogen
refining as well as inclusion removal in ladle treatment", Steel Research International,
No. 12, 2015, pp. 1498-1507.
27. Jönsson P, Karlsson L and Reinholdsson F, “The Effect of Final Induction Stirring
Practice on Inclusion Characteristics Bearing Steel Production”, Proc. 5th Clean Steel
Conf., Balatonfured, Hungary, Vol. 1, Jun. 1997, pp. 218-226.
28. Sheng D-Y, Söder M, Jönsson P, Jonsson L, “Modeling Micro-Inclusion Growth and
Separation in Gas-Stirred Ladles”, Vol 31, No.2, 2002, pp. 134-147.
29. Zheng S, Zhu M, “Physical Modelling of Inclusion Behaviour in Secondary Refining
with Argon Blowing”, Steel Research International, Vol. 79, No.9, 2008, pp. 685-690.
30. Thunman M, Eckert S, Hennig O, Björkvall J, Sichen D, “Study on the Formation of
Open-Eye and Slag Entrainment in Gas Stirred Ladle”, Steel Research International,
78, No 12, 2007, pp. 849-856.
31. Hallberg B. Licentiate Thesis, Dept. of Metallurgy, KTH, SE-100 44 Stockholm,
Sweden, TRITA-TPM-35, 1996.
32. Pierard J, Sichen D, Jönsson P, Seetharaman S and Landin T., Ironmaking and
Steelmaking, 1998, pp. 374-381.
33. Tripathi N. N., Nzotta M., Sandberg A., Sichen Du, “Effect of ladle age on the
formation of non-metallic inclusions in ladle treatment”, Ironmaking and Steelmaking,
2004, pp. 235-240.
34. M. Hallberg, L.T.I Jonsson and J. Alexis: Proc. Scanmet I, MEFOS, Jun 1999, pp.119-
170.
35. Steneholm K. M.Sc. Thesis, LTU, Luleå, Sweden, 2002.
36. Johansson S. Proceedings of the Third Int. Conf. on Clean Steel, The Inst. of Metals,
Balatonfured, Hungary, 1986, pp. 60-67.
37. PCMIC, Jernkontoret, TO23, 111 87 Stockholm, Sweden.
38. J.O. Andersson, T. Helander, L. Höglund, P.F. Shi, and B. Sundman, “Thermo-Calc
and DICTRA, Computational tools for materials science”, Calphad, 26, 2002, pp. 273-
312.
51
39. Hayes P. Process Principles in Minerals & Materials Production, Brisbane, 1993, pp.
633-652.
40. T.A. Engh: Principles of Metal Refining, Oxford University Press, 1992.
41. W.D. Cho and Y.K. Rao, Ironmaking and Steelmaking, Vol. 22, No.3, 1995, p. 239.
42. T. Harada and D. Janke, Steel Research International, Vol. 60, No. 8, 1989, p. 337.
43. B. Deo and R. Boom, “Fundamentals of Steelmaking Metallurgy”, Prentice Hall
International, UK, 1993.
44. J. Wikström, K Nakajima, P. Jönsson, “Application of a Model for Liquid Inclusion
Separation at a Steel- Slag Interface for Laboratory and Industrial Situations”, Steel
Research International. Vol. 79, No. 11, 2008, pp. 826-834.
45. M.A.T. Andersson, P.G. Jönsson, M. Hallberg, “Optimisation of ladle slag
composition by application of sulphide capacity model”, Ironmaking and Steelmaking,
Vol. 27, No.4, 2000, pp. 286-293.
46. Margareta A.T. Andersson, Pär G. Jönsson and Mselly M. Nzotta, “Application of the
Sulphide Capacity Concept on High-Basicity Ladle Slags Used in Bearing-Steel
Production”, ISIJ International, Vol. 39, No. 11, 1999, pp. 1140-1149.
47. P.G. Jönsson, L Jonsson and D. Sichen, “Viscosities of LF Slags and their Impact on
Ladle Refining”, ISIJ International, Vol 37, No.5, 1991, pp. 484-491.
48. A.M.T. Andersson, L.T.I Jonsson, P.G. Jönsson, “A model of reoxidation from the top
slag and the effect on sulphur refining during vacuum degassing”, Scand. Journal of
Metallurgy, Vol. 32, 2003, pp.123-136.