Slag Foaming: Fundamentals, Experimental Evaluation and

Application in the Steelmaking Industry

A.P. Luz, T.A. vila, P. Bonadia, V.C. Pandolfelli

CO(g) formation is also related to the reac-

The slag foaming process is currently applied to some equipment for tion between C (from the metal) and oxy-
steel production (i.e., electric arc furnace (EAF), basic oxygen furnace gen, as shown by Eq. 3.

(BOF), etc.) aiming to save energy, improve productivity, enhance the C( l ) +

1
O2 CO( g ) (3)
refractory service life and inhibit steel re-oxidation. However, the cor- 2 (g)
rect control and optimization of the foam generation on the slag sur- Reaction 2 prevails at the slag-metal and
face are still limited and some factors such as basicity, FeO concen- slag-carbon interfaces, whereas reactions 1
and 3 will take place mainly at the contact
tration, surface tension, viscosity and how the suspended second area between oxygen and metal. Additional-
phase particles in the liquid act directly in the slag foaming behav- ly, the following transformations (Eq. 4 and
iour, need some further understanding. Therefore, this work address- 5) are expected to occur at the slag-gas and
es the main factors that induce foaming, experimental tests used for carbon-gas interfaces, respectively [7].

its evaluation and the use of thermodynamic data in order to control FeO(l ) + CO( g ) Fe(l ) + CO2( g ) (4)
the slag composition while the equipment is being operated. Con-
sidering the data presented in the scientific literature, there are vari-
ous conditions and situations to be explored and thermodynamic cal- CO2 ( g ) + C( s ) 2CO( g ) (5)

culations seem to be the most suitable tool for these investigations. All of these reactions (Eq. 25) are CO(g) and
CO2(g) formers and affect the mass transfer
1 Introduction There are basically two requirements for among metal, slag, carbon and gas phases,
For decades, slag has been considered as foaming: and are considered as a necessary condition
a residue of the steelmaking process. How- reactions or processes that generate small for slag foaming. The generated gas is re-
ever, this material plays an important role gas bubbles and
due to its capacity to withdraw impurities suitable slag properties to keep the bub- A.P. Luz, V.C. Pandolfelli
and inclusions from the steel, reduce bles as stable foam [10, 11]. Federal University of So Carlos (UFSCar)
energy losses and refractory wearing, and Usually, gases resulting from chemical reac- Materials Engineering Department
protect steel from re-oxidation [1]. The tions tend to foam the slag with smaller 13565-905 So Carlos, SP
slag composition (usually in the system of bubbles, whereas the injection of gas phas- Brazil
CaOMgOSiO2FeOAl2O3) directly affects es (oxygen, argon, etc.) results in larger bub-
its viscosity, thermal conductivity, density bles and less stable foams [12]. T.A. vila, P. Bonadia
and other properties, causing an impact on The foaming phenomenon can take place by Magnesita Refratrios S.A.
the ability of the slag to remove impurities injecting oxygen and carbon into the electric Research and Development Center (CPqD)
from the molten metal and to save energy. furnace bath (Fig. 1), leading to FeO genera- 32210-900 Contagem, MG
Slag foaming has been widely investigated in tion in the liquid (Eq. 1). This phase is one of Brazil
electric arc furnaces (EAF), basic oxygen fur- the major components in the slag, attaining
naces (BOF), etc. [27]. values higher than 20 mass-% [10]. Corresponding author: V.C. Pandolfelli
This practice protects the refractory materials E-mail: vicpando@power.ufscar.br
1
from the high heat intensity (radiation) gen- Fe( l ) + O2 FeO( l ) (1)
erated by the electric arc, improving produc- 2 (g) Keywords: foaming, slag, electric arc fur-
tivity and energy efficiency of the equipment. After that, carbon reacts with FeO from the nace (EAF), thermodynamic simulations
It has also been reported [8, 9] that the slag liquid, generating CO(g), which is mainly re-
foaming process can save 3 10 % and 25 sponsible for the foaming effect (Eq. 2) [7]. Received: 16.08.2010
63 % of energy and refractory consumption, Accepted: 02.01.2011
respectively. FeO( s ,l ) + C( s ,l ) Fe( l ) + CO( g ) (2)

refractories WORLDFORUM 3 (2011) [2] 1

 the slag height h [cm] and the superficial
O2 Gas bubbles
gas velocity Vgs [cm/s]. This parameter is di-
Carbon particles
rectly related to some of the slag physical
properties, such as viscosity, surface tension
and density. Fig. 3 and Eq. 6 and 7 are pre-
sented to better explain how to evaluate the
foaming index [14].
Slag
Qg
V gs = (6)
A
h
Molten steel
= (7)
V gs
Fig. 1 Sketch of the transformations that take place in EAF equipment during the slag where, Qg [cm3/s] is the gas flow rate, A is
foaming procedure [7]
the cross-section area of the crucible [cm2],
and h [cm] is the change of the slag height.
Resistor Moreover, the superficial gas velocity and
foam height [h] is related to the foam void
Mo wire fraction () (Eq. 8 and 9).
Voltmeter

V gs = V g (8)

Alumina
tube
h = L (9)

where, Vg [cm/s] is the actual gas velocity

Flowmeter and L [cm] the foam layer thickness. Consid-
Alumina ering that value is between 0.7 and 0.9 in
crucible
Ar any position that the foam has been meas-
ured [15], it is possible to assume that the
Foamed slag void fraction is constant. Therefore, the
Alumina foaming index can be expressed in terms of
tube Molten slag
the foam layer thickness and the actual gas
velocity (Eq. 10).
Fig. 2 Experimental apparatus used to measure the slag height when gas is injected
into the molten slag [19]
L
= (10)
V g
sponsible for the bubbles and foam forma- them only focused on the foam height analy-
tion, which should remain at the upper liq- ses when an inert gas or oxygen was inject-
uid surface, even after the reaction between ed into the liquid [3, 1220]. Fig. 2 shows a 2.2 Foam life time ()
FeO and C is completed. Thus, the CO(g) for- schematic drawing of a device (which was The time needed to reduce the foam height
mation is related to various process vari- adapted in an electric furnace) developed to from h0 (initial height) to a specific position
ables, such as size, type and mass flow rate evaluate the slag foamability at high tem- h, after halting the gas injection in the stud-
of carbonaceous material, flow rate of oxy- peratures. Nevertheless, this procedure only ied system, corresponds to the foam life
gen, carbon content in the bath, slag com- considers the liquid behaviour, not taking time. Equation 11 describes this parameter
position, slag-metal interfacial area, etc. [7]. into account the interaction between the and is related to the changes between the
However, despite the well known benefits of molten liquid and the refractory, as observed injected gas and the generation and
slag foaming, few steel producers have man- during the EAF operation in the steelmaking drainage of the foam [14].
aged to generate a consistent foamy prac- process. The foaming index and foam life t
tice due to the lack of slag chemistry and vis- time are the two parameters that can be at- = (11)
h
cosity control in the electric arc furnace op- tained in this simplified test. ln
eration [10]. h0
2.1 Foaming index () where, t is the time [s] measured during the
2 Foaming evaluation The foaming index or the average travelling foam height reduction. Therefore, for an ide-
Various authors experimentally evaluated time of the gas in the generated foam is at- al system (i.e., for which is constant) the
the slag foaming phenomenon, but most of tained by the ratio between the change of

2 refractories WORLDFORUM 2 (2010) [2]

foaming index will be equal to the average
foam life, as follows:

dh dL dV foam Q g (12)
= = =
dt dt A dt A
where, Vfoam is the foam volume. Simplifying
Eq. 12, the and correlation can be ob-
h L
tained: foam
h h
=
= = (13)
A Qg (Q g A)
Molten
slag
2.3 Mathematical models
Although various investigations are based Steady state
Qg = 0
on the analyses, the use of this parameter
is still controversial as the foaming phenom-
enon depends on several factors which are Fig. 3 Sketch of the foaming index measuring procedure [14]
not considered in Eq. 10 [16]. Some authors
state that this index does not correctly de- the properties of the chosen slag (data col- Experimental results obtained after evaluat-
scribe foaming (in a quantitative manner) lected by experimental tests or inferred from ing the slag with CaO / SiO2 = 1 and
when a chemical interaction between the the literature) and also with the use of the 5 15 mass-% of FeO at 1500 C, were
slag components and the metal takes place surface tension value, instead of the surface well correlated and adjusted to this model
at the contact area [13]. Another important tension depression one [16]. presented in Eq. 17 and 18. It should be
aspect is the fact that the linear relationship pointed out that the foaming behaviour of
between L and Vg only exists at high tem- = 115 (for basic slags) (15) basic slag is mainly affected by the visco-
peratures (> 1500 C), affecting the evalua-
sity of the liquid, whereas for the acid one
tion in other experimental conditions. it will depend strongly on the average
Thus, having the aim of optimizing and con- = 0.93 2
(for acid slags) (16) bubble size and surface tension of the
sidering the effect of the physical properties 3
liquid [16].
of slags on the value, various works fo- The surface tension depression is the slope When the bubbles are formed by gas injec-
cused on the design of mathematical models of the curve when the surface tension is tion, Db may depend on the liquid viscosity,
taking into account some of the main char- plotted against the molar concentration of surface tension, gas flow rate and nozzle
acteristics of the slag (density, viscosity, sur- the surface active element, (/ [moles geometry [12]. Nevertheless, it is always
face tension or surface tension depression, FexO]). Stadler et al. [16] stated that the in- possible to change the gas bubble size, just
bubble size of the gas generated in the liq- crease of FexO concentration commonly low- by changing the geometry of the blowing
uid, etc.). ers the slag viscosity, leading to a decrease nozzle, keeping all the other slag properties
Ito and Fruehan [14, 15] proposed an equa- in the foam index. Nevertheless, the surface constant.
tion based on dimensional analysis, not con- tension depression (which effect is more sig- The model proposed by Zhang and Fruehan
sidering the presence of solid particles and nificant in the case of acid slag) changes as [12] is the most accepted and used to de-
the bubble size effect in the slag for the a function of the FexO concentration and re- scribe the slag foamability (based on the
CaOSiO2FeOAl2O3 system (Eq. 14). sults in the increase, overriding the lower- measurement of ) [3, 11, 13, 14, 16, 18].
ing in the slag viscosity. However, the correlation presented by Frue-
= 5.7 x10 2 (14) Zhang and Fruehan [12] suggested expres- han and other authors [12, 14, 15] were
sions similar to the ones obtained by Jiang developed for a limited range of physico-
where, = viscosity [Pa.s], = surface ten- and Fruehan also considering the bubble chemical properties and service conditions.
sion [N.m-1] and = liquid density [kg.m-3]. size of the gas formed in the liquid (Db = av- Consequently, Eq. 17 and 18 do not predict,
In this expression, the viscosity is the most erage bubble diameter [m]). in a correct manner, the foam thickness gen-
important parameter to attain the most suit- 1.2 erated in molten slag presenting different
able value. = 115 properties from those specified in these
Jiang and Fruehan [3] adjusted the previous
0.2 Db0.9 works.
correlation quoted in the former work [14], (for basic slags) (17) Ghag et al. [21, 22] showed that the values
resulting in more accurate expressions derived from Eq. 17 have some discrepan-
(Eq. 15 and 16). However, this model still 12 cies when compared with the results at-
presented some limitations and did not = 10.3 x10 4 tained in practical tests. Based on their stud-
make the foaming behaviour for acid slag 0.4 11.7 Db23 ies, the authors concluded that the liquid vis-
clear. These observations are associated with (for acid slags) (18) cosity has a significant effect on the gener-

refractories WORLDFORUM 2 (2010) [2] 3

 and the gas bubbles cannot remain in the
liquid. The excessive amount of this oxide
Optimu m slag can also result in an acid slag, shortening
the refractory service life and wasting ex-
pensive chemical energy during the steel re-
index []
()

fining process [8].

Foaming index

3.3 Surface tension

Foaming

The bubble size increases by increasing the

liquid Over-saturated: surface tension of slag. Therefore, the film
Too much second between the bubbles becomes thin, because
phase particles
the void fraction in the foam scales with the
=
= 00 >>00 >>
>>00 bubble size. The drop of the foam height will
be induced by the pressure difference as a
Effective
Effective viscosity [
Viscosity ( ]e) result of the film surface curvature around
the bubble. Thus, the rate of rupture of the
Fig. 4 The relationship between foaming index and effective viscosity [28] bubbles at the upper surface of the foam in-
creases with the surface tension of the slag
ated foam height and the following correla- them and the slag foaming phenomenon is [18, 23].
tion was suggested: far from being completely understood.
3.4 Viscosity
E ff 3 Parameters that affect the slag
= 1x10 6 (19) An optimized slag chemical composition (for
( g ) 2 Db foaming practice
3
a chosen temperature) should be attained to
where, Eff is the effective elasticity of the liq- Slag properties may inhibit or promote the assure the formation and stability of the gas
uid film which forms the bubble and g is the generation of the foam, therefore some of bubbles. For a liquid with low viscosity, the
gravitational constant. Despite the advances these characteristics are presented and dis- bubbles will be easily eliminated and the
in the mathematical models, determining the cussed as follows: foam may be suppressed. Conversely, the
effective elasticity value is difficult and addi- bubbles will not be generated in a high vis-
tional parameters are still required. Further- 3.1 Basicity cous liquid [9, 23].
more, the comparison between the quoted The key issue for EAF slag engineering is to
model (Eq. 19) and the slag foaming data attain a balance between the refractory and 3.5 Suspended second phase
presented in the literature has not yet been fluxing oxides in its composition [28]. The re- particles
carried out. fractory oxides such as MgO and CaO can in- Their presence in a slag has a much greater
Scientific works commonly consider the crease the effective viscosity of the slag. On impact on foaming than the surface tension
injection of a gas phase in the liquid in the other hand, SiO2, Al2O3, FeO, MnO and and slag viscosity [28]. The solid particles
order to promote slag foaming. However, CaF2 will increase its fluidity and negatively (commonly Ca2SiO4 or MgO . FeO) act as gas
reactions involving oxygen and carbon affect the foam generation at the upper sur- nucleation sites, leading to a high amount of
(added to the liquid bath) in the electric face of the liquid. The slag basicity concept is small gas bubbles in the slag. Additionally,
arc furnace operation lead to the CO bub- an attempt to define a balance between the the suspended particles can change the ef-
bles generation, which will be responsible components of the liquid. Some expressions fective slag viscosity, as presented in Eq. 21.
for the foam formed at the top of the molten used to described this parameter are [28]: 5
slag. CaO
e = (1 1.35) 2
(21)
CaO
Aiming to understand the system changes in B2 = B3 = (20)
order to attain a condition similar to the
SiO2 SiO2 + Al2O3 where, e is the effective viscosity of the slag
practical tests, some authors [23] evaluated [Pa.s] and is the fraction of precipitated
and suggested a physical model considering 3.2 FeO concentration solid phases (0 < < 0,74). There is a close
a system containing a liquid material com- The iron oxide contained in the slag during relationship between the foaming index and
prised by slag and metal. This work attested the EAF operation is usually formed by the the effective viscosity (Fig. 4) and, the high-
that the main factors that affect the liquid iron oxidation when oxygen is injected into er e, the longer the residence time of the
foamability are the surface tension and vis- the molten metal. FeO is a major oxide com- gas bubbles in the slag will be, extending
cosity of the slag, interfacial tension be- ponent in slag and has a significant effect on the stability of the foam. However, there is a
tween slag-metal and surface tension of the viscosity, foamability, slag volume, productiv- maximum amount of second phase particles
metal phase. Other publications [2427] ity and power consumption. Low FeO con- that is beneficial for the foam stability.
also discuss the effectiveness of the existing tent (< 10 mass-%) results in a viscous slag Furthermore, other factors such as adding
models presented in the scientific literature, which is hard to foam. Conversely, a FeO carbon or MgO particles, the temperature of
but up to now there is no agreement among content > 40 mass-% increases the fluidity the process, etc., directly affect the foam for-

4 refractories WORLDFORUM 3 (2011) [2]

mation. When a high amount of carbon par-
ticles is added to the liquid, slag foaming
may be inhibited. It was postulated that
when a hydrophobic particle (or material
that presents limited wettability, such as car-
bon) comes into contact with the liquid film,
D

MgO [%]
the instantaneous contact angle at the
formed three-phase boundary (solid, liquid,
gas) is smaller than the equilibrium contact
one [29]. This difference results in a driving
force that makes the bubble move around
the particle, stretches the liquid film and de-
creases foaming. Thus, a control of the FeO
and C contents in the system is required, be-
cause the excess of carbon can inhibit foam- FeO [%]
ing instead of inducing the chemical reac-
tions responsible for keeping this process. Fig. 5 Isothermal solubility diagram (ISD) of the FeOMgOSiO2CaOAl2O3 system for a
basicity B3 of 2.0 and temperature = 1600 C [28]; MW = MgO . FeO, C2S = 2 CaO . SiO2,
The temperature effect in the foaming prac-
L = liquid
tice has also been investigated in some stud-
ies [19, 30] and the increase of this parame-
ter leads to the decrease in the foaming in- FeO and Al2O3. According to practical tests, it ture. Conversely, the Al2O3 effect is described
dex. The temperature also acts on the vis- is known that the suitable compositions are by the equation:
cosity and chemical saturation of the liquid, the MgO saturated ones, containing sus-
Al 2 O3 correction factor =
affecting the precipitation of second phases. pended second-phase particles (i.e.,
It must be pointed out that the saturated MgO . FeO (magnesio-wustite, MW)) [7]. The % Al 2 O3
MgO or CaO slag will not only foam better, MgO saturated liquids not will only positive- 0.615. .(% MgOref 6)
but will also decrease the refractory wearing ly affect the foaming, but also result in a de- SiO2 + Al 2 O3
of the electric arc furnaces [8]. crease of the refractory wearing. (23)
Therefore, based on the phase diagrams of Usually, the ISD diagrams (i.e., Fig. 5) are
4 Use of phase diagrams and the studied system, it is possible to define split into 4 regions: liquid slag (L), MgO . FeO
thermodynamic simulations to the influence of each slag component and (MW) saturated liquid, 2 CaO . SiO2 (C2S)
predict slag foaming behaviour which combination among them can result saturated liquid and the dual saturated
in EAF in liquid saturation and precipitation of the (L + C2S + MW) liquid. Point D is named
Optimization and control of the slag during MgO . FeO phase. Some works suggest the dual saturated point and the lines delimitat-
the EAF operation is very difficult, because use of the isothermal solubility diagrams ing the L + (MW + L) and L + (L + C2S) re-
not only molten metal but also scrap is (ISD) [1, 5, 7, 28], which are valid for a spe- gions (highlighted with bold lines in the
added to the furnace, affecting the liquid cific basicity value and express the relation- graph) are known as the MgO and CaO sat-
composition. Information about the scrap ship between the MgO and FeO in mass-%. urated ones, respectively. Moreover, the
particles size distribution, oxidation level, Si Fig. 5 shows an example of an ISD diagram dashed line surrounding the liquidus region
and Al content, etc., is very important to as- for the system FeOMgOSiO2CaOAl2O3 indicates the slag composition with peak
sure the efficiency of the equipment and at 1600 C. Considering the B3 basicity foaming performance, where the maximum
foaming practice at high temperatures. ratio, these graphs are prepared by analyz- slag foaming takes place (about 6 mass-%
Pretorius et al. [28] stated that the control of ing the isothermal sections of the ternary di- out of L region [5]).
the slag in conventional EAF operations can agrams MgOSiO2CaO, MgOSiO2FeO, Changing some parameters such as temper-
be compared to hitting a moving target, be- MgOCaOFeO and CaOSiO2FeO. In ad- ature, oxygen partial pressure and slag ba-
cause this is a dynamic system which con- dition, the effect of the Al2O3 on the MgO sicity, will move the saturation lines and the
stantly needs to be adjusted to keep a con- solubility can be calculated by the correction point of dual saturation (D) to a different lo-
tinue and stable foam generation, due to the factors: cation in the diagram and the MgO and FeO
changes in the slag throughout the steel- content of the slag should be re-adjusted to
making process.
% MgOsat = ensure the foaming effect.
The easiest way to adjust the slag chemistry % MgOref Al 2 O3 correction factor One of the most used procedures in an EAF
in this type of operation consists of injecting (22) to extend the foamy conditions consists of
carbon or adding limited amounts of MgO or injecting MgO fines into the slag towards
dolomite (CaMg(CO3)2), allowing the foam where, MgOref [%] corresponds to the MgO the end of the heating period. Fig. 6 shows
to form in the furnace. EAF slag typically content in the CaO-MgO saturated slag two examples where K and T are the orig-
contains five major oxides: CaO, MgO, SiO2, (double saturation) at a selected tempera- inal slag compositions.

refractories WORLDFORUM 3 (2011) [2] 5

 The ISD diagram analyzed by Pretorius et al.
[28] provides important information which
helps and indicates which adjustments can
MgO Addition be made to extend foaming in the furnace.
Nevertheless, some limitations can be de-
tected: 1) only applicable for the analysis

. of slag basicity (B3) between 1.2 3.5; 2)

MgO [%]

Slag K
suitable for FeO values < 55 mass-% and 3)

Slag T . K1 K3
K2
for temperatures between 1500 1750 C.
Due to these adjustments and correction
factors used in the diagrams, some impreci-
sion can still be found when using these
graphs.
A better way to obtain such diagrams and
also to define the effect of various parame-
ters on the slag characteristics is via thermo-
dynamic software such as FactSage, Thermo-
FeO [%] Calc, etc. [7, 8]. Thermodynamic calculations
Fig. 6 ISD diagram showing the effect of adding MgO to the path of the liquid composi-
can predict the position of the dual satura-
tion [28] tion point and the MgO saturated lines as a
function of the slag basicity (Fig. 7), temper-
As the FeO is generated and composition K MgO aims to attain two main purposes: 1) ature, oxygen partial pressure and other pa-
moves towards K2, the fluidity of the slag in- to reduce the slag temperature by incorpo- rameters. Compared to some mathematical
creases (Fig. 6). Initially the foaming proper- rating a material which was previously at models [7] and phase diagrams, these analy-
ties of the liquid increases moving from room temperature, and 2) to maintain the ses are very accurate and present an error
point K to K1, because the latter is located in MgO level due to the FeO increase during < 1 % when FeO content in the slag is less
the region where the foam generation is fa- the furnace operation. than 60 mass-%.
vored. Nevertheless, from K1 to K2, the slag However, the effectiveness of the MgO or In addition, simulations can be carried
becomes too liquid (low viscosity) to foam in doloma (CaO . MgO) injection into the EAF out for any chemical compositions and slag
a suitable way. The MgO fines injection into depends on the initial MgO content in the basicity values. Considering a practical
the slag (addition at point K1) will change slag. Considering composition T (Fig. 6), the condition, the liquid basicity is related to
the liquid composition and ensure the pres- magnesia level is too low and a further in- the EAF charging materials and directly de-
ence of second-phase particles, keeping the jection of this phase will not result in signif- pends on the cold charge content (scrap
foamy conditions [28]. In this case, adding icant changes. and/or solid pig iron) added during the op-
eration. The amount of cold charge is strate-
0.25 gic and based on the current price of these
25
materials. Consequently, the scrap and/or
solid pig iron additions reduce the slag ba-
0.2 MW + L sicity due to the increase in SiO2 content in
20 MW +
O X=1. 25 the system and, therefore, the steel plant
g may operate with B3 values lower than 1,2
M in the slag.
MgO [mass-%]

0.15
15
s X=1. 50 Thus, the use of thermodynamic simulations
s and the control of the slag compositions
a 0.1 X=1. 75 during the EAF operation can evaluate some
m 10
X=2. 00 conditions not explored by the scientific lit-
erature until today (B3 < 1).
X=2. 25
0.05
5 X=2. 50 For example, the ISD diagram for a slag
L L
basicity B3 = 0.9 (25.7 mass-% of CaO,
16.2 mass-% of SiO2 and 11.6 mass-% of
0 Al2O3) can be designed by the FactSage soft-
0 0.12 0.24 0.36 0.48 0.6
12 24
mass FeO
36 48 60 ware, using the Phase Diagram module.
FeO [mass-%]
In this case, MgO, FeO, O2 and (CaO)x(SiO2)y
Fig. 7 MgO saturated lines in the FeOMgOCaOSiO2 system as a function of the slag (Al2O3)z are considered, where x, y and z
basicity (CaO/SiO2 = X), at 1600 C and high oxygen partial pressure (pO2 = 10-4,72 atm) correspond to the molar stoichiometries of
[7] the oxides of the slag. Temperature and oxy-

6 refractories WORLDFORUM 3 (2011) [2]

gen partial pressure of the furnace environ-
25
ment can be easily defined at this calcula-
tion step. Fig. 8 shows a diagram prepared L+
using the FactSage software by the authors MgO .FeO (MW)
20 L + MgO .FeO (MW) + FeMg2O4
of this work, for the slag described above
(B3 = 0.9), temperature of 1600 C and
pO2 = 0.21.

MgO [mass-%]
15 L+
Significant changes in the diagram were ob-
FeMg2 O4
served when compared with those shown in
Fig. 5. For B3 < 1, the precipitation of C2S
10
and liquid saturation in this phase is not
possible, considering the chosen interval
Liquid (L)
(FeO and MgO content equal to
5
0 60 mass-% and 0 25 mass-%, respec-
tively) at 1600 C. Conversely, slag contain-
ing high amounts of FeO can still be saturat-
ed by the spinel phase (mainly FeMg2O4, but 0 12 24 36 48 60
it can also contain MgAl2O4 in its composi- FeO [mass-%]
tion Fig. 8). Therefore, slag with a basicity Fig. 8 Phase diagram in the FeOMgOSiO2CaOAl2O3 system for a basicity B3 = 0.9,
lower than 1 will show a distinct behaviour temperature = 1600 C and pO2 = 0.21 atm
at the thermodynamic equilibrium when
compared with the one with high basicity
values (B3). Acknowledgments in EAF and its influence on the steelmaking
It can be concluded that the data provided The authors are grateful to the Conselho shop productivity. ISIJ Inter. 35 (1995) [9]
by the simulations will help and indicate Nacional de Desenvolvimento Cientfico e 10541062
which adjustments may be carried out dur- Tecnolgico (CNPq) and Magnesita Re- [8] Kwong, K.S.; Bennet, J.P.; Krabbe, R.; Petty, A.;
ing the EAF operation in order to improve fratrios S.A. (Brazil) for supporting this re- Thomas, H.: Thermodynamic calculations pre-
and optimize suitable conditions for foam- search. dicting MgO saturated EAF for use in EAF steel
ing. There are various conditions and situa- production. In: Proc. TMS 2009 138th Annual
tions to be explored and the thermodynam- References Meeting and Exibition, San Francisco. EUA
ic calculations seem to be the most suitable [1] vila, T.A.; Freire, R.S.; Silva, G.F.B.L.; Borges, (2009) 6370
tool for these investigations. R.N.: Design of slags compatible with refracto- [9] Kwong, K.S.; Bennet, J.P.: Recycling practices of
ry systems. In: Proc. UNITECR 2009, Salvador. spent MgO-C refractories. J. Min. & Mat. Cha-
5 Final remarks Brazil, pp.14 racterization & Engineering 1 (2002) [2] 6978
Based on the various works in the literature [2] Guzzon, M.; Sahajwalla, V.; Memoli, F.; Pustori- [10] Pretorius, E.B.; Nunnington, R.C.: Stainless
and due to the complexity of slag foaming no, M.: The behaviour of the secondary metal- steel slag fundaments: from furnace to tundish.
phenomenon, it can be concluded that some lurgy slag into the EAF. How to create a good Ironmaking and Steelmaking 29 (2002) [2]
difficulties remain to define the best condi- foamy slag with the appropriate basicity using 133139
tions to attain the foam generation and its a mix of lime and recycled ladle slag as EAF [11] Matsuura, H.; Fruehan, R.J.: Slag foaming in an
stability at the upper surface of the molten slag former. In: Proc. 38th Seminrio de Aciaria, electric arc furnace. ISIJ Inter. 49 (2009) [10]
slag. Developed mathematical models are Belo Horizonte, Brazil, 2007, pp. 112 15301535
usually adjusted to a simple situation where [3] Jiang, R.; Fruehan, R.J.: Slag foaming in bath [12] Zhang, Y.; Fruehan, R.J.: Effect of the bubble
a constant gas flow (inert gas or oxygen) is smelting. Metallurg. & Mat. Trans. B. 22 (1991) size and chemical reactions on slag foaming.
injected into the liquid. However, the reac- 481489 Metallurg. & Mat. Trans. B. 26 (1995) 803-812
tions between oxygen and carbon with the [4] Kitamura, S.; Okohira, K.: Influence of slag [13] Wu, K.; Qian, W.; Chu, S.; Niu, Q.; Luo, H: Beha-
molten liquid components and the metal- composition and temperature on slag foaming. viour of slag foaming caused by blowing gas in
slag interactions result in changes of the ISIJ Inter. 32 (1992) [6] 741746 molten slags. ISIJ Inter. 40 (2000) [10] 954957
slag properties. Therefore, some adjustments [5] Novk, M.; Straka, J.; Pribyl, M.: Influence of [14] Ito, K.; Fruehan, R.J.: Study on the foaming of
in the liquid composition are frequently the slag foaming process applied in high alloy- CaO-SiO2-FeO slags, Part I: Foaming parameters
needed. Considering these aspects, the de- ed steel production on refractory wear of EAF and experimental results. Metallurg. & Mat.
velopment of isothermal solubility diagrams at Pilsen Steel melt shop. In: Proc. UNITECR Trans. B. 20 (1989) 509514
and the use of thermodynamic simulations 2009, Salvador, Brazil, pp. 14 [15] Ito, K.; Fruehan, R.J.: Study on the foaming of
seem to be a good option to evaluate the [6] Jung, S.M.; Fruehan, R.J.: Foaming characteristics CaO-SiO2-FeO slags, Part II: Dimensional analy-
best conditions to generate and maintain of BOF slags. ISIJ Inter. 40 (2000) [4] 348355 sis and foaming in iron and steelmaking pro-
the stability of the foam during the EAF op- [7] Morales, R.D.; Rubn Lule, G.; Lopez, F.; Cama- cesses. Metallurg. & Mat. Trans. B. 20 (1989)
eration. cho, J.; Romero, J.A.: The slag foaming practice 515521

refractories WORLDFORUM 3 (2011) [2] 7

 [16] Stadler, S.A. C.; Eksteen, J.J.; Aldrich, C.: An ex- lopment of slag foaming. ISIJ Inter. 38 (1998) del of slag foaming in oxygen steelmaking con-
perimental investigation of foaming in acidic, [11] 12081215 verters. ISIJ Inter. 38 (1998) [11] 12251232
high FexO slags. Min. Engng. 20 (2007) [22] Ghag, S.S.; Hayes, P.C.; Lee, H.G.: The prediction [27] Kim, H.S.; Min, D.J.; Park, J.H.: Foaming
11211138 of gas residence times in foaming CaO-SiO2- behaviour of CaO-SiO2-FeO-MgOsat-C (X =
[17] Ghag, S.S.; Hayes, P.C.; Lee, H.G.: Physical mo- FeO slags. ISIJ Inter. 38 (1998) [11] 12161224 Al2O3, MnO, P2O5 and CaF2) slags at high tem-
del studies on slag foaming. ISIJ Inter. 38 [23] Ogawa, Y.; Huin, D.; Gaye, H.; Tokumitsu, N.: peratures. ISIJ Inter. 41 (2001) [4] 317324
(1998) [11] 12011207 Physical model of slag foaming. ISIJ Inter. 33 [28] Pretorius, E.B.; Carlisle, R.C.: Foamy slag funda-
[18] Hara, S.; Kitamura, M.; Ogino, K.: The surface (1993) [1] 224232 mentals and their practical application to elec-
viscosities and the foamines of molten oxides. [24] Hong, L.; Hirasawa, M.; Sano, M.: Behaviour of tric furnace steelmaking. 56th Electric Furnace
ISIJ Inter. 30 (1990) [9] 714721 slag foaming with reduction of iron oxide in Conf. ISS-15-18, New Orleans/USA, 1998, pp.
[19] Ozturk, B.; Fruehan, R.J.; Effect of temperature molten slags by graphite. ISIJ Inter. 38 (1998) 275291
on slag foaming, Metallurg. & Mat. Trans. B. 26 [2] 13391345 [29] Zhang, Y.; Fruehan, R.J.: Effect of carbonace-
(1995) 10861091 [25] Lotun, D.; Pilon, L.: Physical modeling of slag ous particles on slag foaming. Metallurg. &
[20] Zhang, Y.; Fruehan, R.J.: Effect of gas type and foaming for various operating conditions and Mat. Trans. B. 26 (1995) 813819
pressure on slag foaming. Metallurg. & Mat. slag compositions. ISIJ Inter. 45 (2005) [6] [30] Sedivy, C.H.; Krump, R.: Tools for foaming slag
Trans. B. 26 (1995) 10881091 835840 operation at EAF steelmaking. Arch. Metallurgy
[21] Ghag, S.S.; Hayes, P.C.; Lee, H.G.; Model deve- [26] Misra, P.; DeO, B.; Chhabra, R.P.: Dynamic mo- & Mat. 53 (2008) [2] 409413

8 refractories WORLDFORUM 3 (2011) [2]

The author has requested enhancement of the downloaded file. All in-text references underlined in blue are linked to publications on ResearchGate.