SLAG – REFRACTORY ADHERENCE MECHANISMS
Jorge Madías Raúl Topolevsky
Sector Acería - I A S SIDERAR SAIC
madias.ias@cablenet.com.ar yaptky@siderar.com
Elena Brandaleze Silvia Camelli
Refractarios - I A S SIDERAR SAIC
brandaleze.ias@cablenet.com.ar apashc@siderar.com
Marcos Bentancour
Refractarios – IAS
bentancour.ias@intercom.com.ar
ABSTRACT
The wear process of refractories in service is produced by physical and
chemical causes linked to refractory quality, installation and steelmaking process. In order to
reduce the impact of the BOF operation on the refractory lining, the application of MgO
saturated coating slags is required. These slags form a protective layer on the MgO –C bricks
enhancing lining life.
The slag fluidity, the physics and chemistry of the refractory-slag interaction
are relevant to the design of these coating slags. The blowing practice, the end of blow
temperature, the ratio of solid to liquid phases in the slag, its morphology, its particle size and
chemistry are determining factors for the required properties.
The slag – refractory adherence mechanisms are analyzed and measurements of
slag fluidity at high temperature, up to 1700°C, are reported. These results are compared with
theoretical models.
INTRODUCTION
To extend BOF lining life and to maximize availability contributes to the
reduction of costs per ton of steel liquid. Different steel plants, depending on their lay out,
process conditions and operation, have developed and implemented different techniques of
protection of the MgO-C refractory lining: brick patching, gunning, manual splashing,
chemical splashing, slag coating and slag splashing.
The coating practice or build-up uses a MgO saturated slag to cover the bricks
with a slag layer. This adherent layer acts as a protective agent, extending lining life. Besides,
for the implementation of this technique, the control and design of the slags should be
optimized [1, 2]. Slags of appropriate characteristics allow a recovery of the lining, reducing
corrosion, oxidation and erosion of the refractory, as well as the impact of the charge, without
altering the metallurgical function of the BOF.
It is then relevant to study:
• Slag behavior at service temperature: chemistry, percentage of solid phase, liquidus
temperature, MgO saturation, fluidity and surface tension.
• Slag adherence on MgO-C bricks.
Experimental determination of physical properties as viscosity and surface
tension of multicomponent systems at high temperatures (higher than 1500°C) is difficult to
carry out, due to the complexity and time involved. Presence of impurities and possible
reactions between the crucible and the slag, give place to a transportation problem in the
interface that can decrease the values of the properties and occasionally can modify
wetability.
These factors took to the development of mathematical models to predict the
viscosity of a system as a function of slag chemistry [3, 4, 5, 6], being one of those more used
for the industrial slags, the Urbain model. [7]
In this piece of work, surface tension of end-of-blow slag was determined
experimentally and viscosity was estimated by means of theoretical models. The adherence of
these slags to MgO-C bricks was evaluated by dipping test. A post mortem study was done on
bricks taken out from the BOF bottom at the end of a campaign.
EXPERIMENTAL DEVELOPMENT
Three slag samples were taken from SIDERAR BOF#2 at the end of the blow by means of a
spoon. Slag chemistry is presented in table I. Melting range measured with Leco AF 400 is
presented in table II.
Table I. End of blow slag chemistry.
SiO2 MnO P2O5 S FeO MgO CaO Al2O3
Sample
(%) (%) (%) (%) (%) (%) (%) (%)
A 17.1 6.2 2.4 0.06 18.4 11.0 43.1 1.6
B 15.4 6.4 2.6 0.08 23.2 7.0 43.5 1.6
C 13.5 6.0 2.4 0.08 25.8 7.2 43.8 1.0
Table II. Melting range of end blow slag A.
Initial Softening Hemisphere Flow
Sample Temperature Temperature Temperature Temperature
(°C) (°C) (°C) (°C)
A 1398 1408 1412 1419
Surface tension and dipping test were carried out on these slags.
Surface tension
By using the modified Seybold method, tests were carried out in an Astro
furnace, at a temperature range of 1720 °C to 1750 °C under argon atmosphere. A graphite
crucible was designed in such a way that it could be used both for surface tension and
viscosity measurements, although viscosity measurements are not presented in this paper
(figure I). The crucible was fitted in an arrangement of graphite pipes, and placed into the
furnace chamber. The furnace was sealed and nitrogen was injected at a pressure of 1.8
kg/cm2.
Figure I. Experimental arrangement for surface tension and viscosity tests.
Stopper rod
Graphite crucible
Evacuante hole
Receiver
Balance
100 g of previously grinded slag, was introduced in the crucible at test
temperature. After 1 h holding time, the stopper rod was elevated permitting the free flow of
the liquid slag through the hole.
When hydrostatic pressure was equal to the pressure inside the chamber, the
flow of slag through the hole ended. Then the crucible was taken off the chamber and let cool
to environmental temperature. The bottom of the crucible was cut and included in resin for
measurement of the contact angle θ (figure II). The angle was measured by using a profile
projector.
Figure II. Contact angle between the refractory and the slag.
Graphite wall hole
Slag
θ
On the base of the measured contact angles, surface tension was
calculated by means of equation (1):
γ = ρ g h R / 2 cos θ [mN/m] (1)
Where:
γ = surface tension h = slag height after solidification (figure III)
ρ = slag density R = discharge hole radius
g = gravity θ = contact angle between the hole wall and the slag
Figure III. Scheme for definition of slag height after solidification.
D
h = slag height
Results of contact angle measurements and surface tension calculation are
summarized in table III.
Table III. Measured contact angle and calculated surface tension
Hole radio Contact Temperature γ
Sample Height (cm)
(cm) Angle (°C) [mN/m]
A 2.2 0.25 35 ° 1720 905
A 1.2 0.5 30.4 ° 1750 938
B 1.1 0.5 33 ° 1730 883
Dipping test
A rod of 2 cm diameter and 5 cm long was prepared by cutting and turning an
non-used MgO-C brick. 100 g of grinded slag (same as used for surface tension, see tables I
and II) were melted and heated to a given temperature in a graphite crucible. Then, the MgO-
C rod was submerged in the slag. After that, the rod was rotated 12 times during 5 minutes.
This time was selected in order to have the same tangential speed as the BOF when tilting. In
figure IV the experimental arrangement is shown.
Figure IV Experimental arrangement for dipping tests.
Holding sample rod
Sample
Graphite crucible with slag
Graphite tube
After test, the sample rod, covered with a slag layer (figure V, left), was
weighed. Samples were taken out for microscopy (figure V, right). Results of the
measurement of slag layer thickness and weight difference are detailed in table IV.
Figure V. MgO-C rod after dipping test in end-of-blow slag.
Table IV. Slag layer thickness and weight difference after dipping tests
Slag t (min) ∆W (g) Thickness (mm)
A 3 0.734 3
C 5 0.164 0.8
CALCULATION OF VISCOSITY
Most models for calculating viscosity are based on numeric adjustments
starting from experimental data and they are rarely based on the slag structure [8]. Urbain
model estimates viscosity of liquid silicates and alumino-silicates in function of chemical
composition and mineralogical nature.
This model describes temperature dependence of viscosity through the
Weymann equation [7]:
η = AT exp (103 B/T) (2)
where η: viscosity in poise; T: absolute temperature in K; A and B: parameters that depend
on slag analysis (A is expressed in poise K-1 and B in K).
- ln A = 0.2693 B + 11.6725 (3)
The oxides are classified into three groups:
Glass formers: XG = X SiO2 + X P2O5
Modifiers: XM = XCaO + XMgO + XNa2O + XK2O + 3 XCaF2 + XFeO +
+ XMnO + 2 XTiO2 + 2 XZrO2
Anfoters: XA = XAl2O3 + XFe2O3 + XB2O3
A group of industrial end-of-blow slags was selected and divided into three
groups, depending on the FeO content; viscosity was calculated at 1600 °C (table V).
Table V. Viscosity of selected end-of-blow slags. Calculation based on Urbain model.
SiO2 MnO P2O5 S FeO MgO CaO Al2O3 IB η Group
(Poise)
(
1 15.9 6.3 2.8 0.06 19.6 10.2 44.3 0.8 2.79 0.78
2 15.4 5.9 2.7 0.08 18.7 8.4 47 1.5 3.05 0.8 A
3 14.8 6.3 2.4 0.06 21.6 8.4 45.3 1.0 3.06 0.75
18<FeO<22
4 15.3 6.2 2.5 0.08 20.7 7.6 45.7 1.9 2.99 0.81
5 15.3 6.4 2.5 0.07 21 7.6 45.1 2.1 2.95 0.83
6 14.2 6.3 2.4 0.07 22.5 8.8 44.1 1.4 3.11 0.75
7 14.7 6.4 2.7 0.07 22.5 7.9 43.9 1.4 2.99 0.77 B
8 14.0 5.8 2.5 0.08 24.4 7.5 44.3 1.6 3.16 0.75
22<FeO<25
9 15.3 6.2 2.5 0.08 23.7 7.2 43.8 1.6 2.86 0.8
10 14.4 6.1 2.6 0.07 24.1 7.2 44.2 0.9 3.07 0.74
11 13.2 5.6 2.3 0.08 25.6 7.5 44.6 1.2 3.38 0.7
12 13.8 5.5 2.1 0.09 26.3 6.9 43.7 1.2 3.17 0.73 C
13 12.1 5.7 2.2 0.08 25.8 7.9 44.3 1.6 3.66 0.68
25<FeO<30
14 13.9 6.2 2.5 0.07 25.3 6.9 44 1.2 3.17 0.73
15 13.4 5.9 2.3 0.08 26 6.9 43.7 1.6 3.26 0.73
16 12.7 5.8 2.2 0.08 32.1 7.6 39.7 0.8 3.13 0.67
17 10.5 6.0 1.8 0.07 39.9 8.5 32.1 0.8 3.06 0.59 D
18 10.5 5.6 1.9 0.09 31.4 6.5 42.3 1.2 4.03 0.6
FeO>30
19 12.0 5.6 2.0 0.08 32.5 6.4 40 0.9 3.33 0.65
20 11.5 5.3 1.8 0.07 33.6 6.7 39.9 0.7 3.47 0.62
21 18.5 6.8 3.1 0.06 16.7 7.8 45.2 1.0 2.44 0.9
22 13.0 6.0 2.8 0.09 16.4 9.8 48.2 1.7 3.71 0.72 E
23 18.2 5.6 3.3 0.09 14.1 6.8 49.9 1.6 2.74 0.91
FeO<18
24 17.9 5.9 3.0 0.1 15.3 6.8 49.0 1.8 2.74 0.92
25 17.7 5.8 2.9 0.1 16.1 6.4 49.1 1.7 2.77 0.9
Figures VI to VIII show the influence of silica, iron oxide and basicity on viscosity at 1600°C.
Figure VI. Influence of silica content on viscosity at 1600 °C
0.95
0.90
0.85
Viscosity (Poise)
0.80
0.75
0.70
0.65
0.60
0.55
0.50
10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00
% de SiO2
Figure VII. Influence of iron oxide content on viscosity at 1600°C
0,95
0,90
0,85
Viscosity (Poise)
0,80
0,75
0,70
0,65
0,60
0,55
0,50
12,00 18,00 24,00 30,00 36,00
% FeO
Figure VIII. Influence of Basicity index on viscosity at 1600°C
0,95
0,90
0,85
Viscosity (Poise)
0,80
0,75
0,70
0,65
0,60
0,55
0,50
2,60 2,80 3,00 3,20 3,40 3,60 3,80
BI
Figure IX shows viscosity dependence with temperature for slags of group B (22<FeO<25%).
Figure IX. Temperature influence on viscosity (slags of group B, 22<FeO<25%).
0,80 Slag 6 Slag 7 Slag 8
Slag 9 Slag 10
0,75
Viscosity (poise)
0,70
Slag
0,65
0,60
0,55
1600 1610 1620 1630 1640 1650 1660 1670
Temperature (°C)
POST MORTEM STUDY OF MgO-C BRICKS
Several samples of MgO-C bricks were taken of the BOF bottom at the end of
a campaign. Figure X shows one of the samples, with layers of slag, steel and steel droplets,
adhered to the brick. The total thickness of these layers is 6 cm.
Figure X. MgO-C brick taken of BOF bottom at the end of a campaign. Layers of slag and
steel adhered to the brick are observed.
Steel layer
Slag layer
MgO – C brick
The microstructural study was based on the phase distribution in the slag and in
the slag - MgO-C brick interface. Two main phases were identified in the slag: one rich in
FeO, MnO and MgO and other corresponding to a calcium silicate (figure XI). EDS analyses
of these phases is presented in table VI.
Table VI. EDS analyses of different slags phases.
SiO2 MnO FeO MgO CaO
Phases
(%) (%) (%) (%) (%)
1 2.1 19 67 9.2 2.5
2 31 1 3.5 - 64.5
Figure XI Slag structure showing two phases: FeO-MnO-MgO rich phase (dark) and calcium
silicate (light). Back-scattered electron image and mapping of Ca, Si, Mg, Fe and Mn.
Back-scattered electron image Ca Si
Mg Fe Mn
Also, grains of magnesia removed from the brick were observed in the slag.
Inside these grains there was some Fe and Mn attack (Figure XI).
Figure XII. MgO grains in the slag, showing Fe and Mn attack. Back-scattered electron image
and mapping of Ca, Si, Mg, Fe and Mn.
Back-scattered electron image Ca Si
Fe Mg Mn
In the figure XIII the evolution of FeO, MnO, MgO and CaO from the center
of a periclase grain to the boundary is detailed. There is a decrease of MgO and an increase in
FeO and MnO.
Figure XIII. Evolution of FeO, MnO, MgO and CaO from the center of a periclasa grain to
the boundary.
100
90 MgO
80
70
EDS Analysis (%)
III 60
II 50
I 40
30
20
FeO MnO
10
CaO
0
0 20 40 60 80 100 120 140 160 180 200 220 240 260 280
µ m)
Distance from grain center (µ
ANALYSES OF RESULTS
Surface tension and viscosity determine the wetability behavior of the brick by
the slag and the adherence of the slag on the brick to form a protective layer and extend lining
life [9].
Slag A has higher surface tension than slag B, despite of slag A, presents
higher silica content. This result could be justified by differences in chemistry and
temperature. In general, surface tension decreases with the increase of glass formers,
indicating that the anionic groups are surface active. Surface tension also decreases because of
the formation of oxide nets. In particular, FeO surface tension is affected by the addition of
basic and acid oxides [10,11].
The application of the Urbain model to estimate viscosity confirms:
• When SiO2 is increased, viscosity also increases, because polymerization of silicates is
promoted.
• When FeO is increased, viscosity decrease, because this oxide acts as a modifier of the
system, breaking the silicate chains).
• When the basicity index increases, viscosity decreases because CaO also acts as a
modifier.
• The increase in temperature favors the fluidity, according to (η= f (Texp (1/T)).
For slag A, with 18% FeO, 3 mm of slag layer thickness were obtained in the
dipping test. For slag B, with 23% FeO, slag layer thickness was 0.8 mm. So, FeO content is a
major factor in the building of the coating layer.
Through microstructural analysis, the mechanism of attack of FeO and MnO to
the brick was identified.
CONCLUSION
The study of the behavior of the slag at temperatures higher than 1500°C
through experimental tests and application of theoretical models, allowed to establish the
influence of temperature and chemical composition of the slags on their physical properties as
surface tension, viscosity and of adherence on MgO-C bricks.
The post mortem study on the slag – MgO-C brick interface was useful to
identify the mechanism for the attack of FeO and MnO in the slag on the periclasa grains.
REFERENCES
1 J. Grosjean, P. Riboud, “Consistance des Laitiers de Convertisseurs et Tartinage”,
Revue de Métallurgie-CIT, 80, [7], 571-574 (1983).
2 D. Kim et all. “The Aplication Technique of the Converter Lining in Kwangyang
Steelworks”. 1st European Oxygen Steelmaking Congress, 226-231 (1984).
3 T. Iida et all. “An Equation for the Viscosity of Molten Multicomponent Fluxes and
Slags”, Molten Slags, Fluxes and Salts'97 Conference, 877-879. (1997).
4 T. Iida et all., “Accurate Prediction of the Viscosities of Various Industrial from
Chemical Composition”, International Conference on Steel and Society, Osaka, 265-268
(2000).
5 C. Oldani, A. Fernández Guillermet, “About the Mathematical Representation of the
Viscosity of Liquid Solutions of the CaO - Al2O3 - SiO2 System”, Bulletin of the National
Academy of Sciences, 58, 55-73. (1988) (in Spanish).
6 Du Sichen et all. “A Model for Estimation of Viscosities of Complex Metallic and
Ionic Melts”, Metallurgical and Materials Transactions B, 125B, 519-525 (1994).
7 G. Urbain et all. “Viscosity of Silicate Melts”, Trans. J. Br. Ceram. Soc., 80,139-141.
(1981).
8 K. Mills, “The Influence of Structure on the Physico-chemical Properties of Slags,
ISIJ International, 33, N° 1, 148-155, (1993).
9 P. Kozakévitch, “Surface Tension of Liquid Metals and Melts Oxidizes”, Slags and
Salts, 278-280, (1983).
10 E.T. Turkdogan, “Physicochemical properties of molten slags and glasses”, The
Metals Society, London (1983).
11 P. Riboud, M. Olette, “Phenomenes Superficiels Leur Rôle in Siderurgie”, Cours
Théorique of base pour l'élaboration of the fonte lácier et, IRSID-CESSID (1972).
