Why Is MgO – C Refractory Material Applicable

in Steel Metallurgy?

J. Pötschke

Under the aspect of the high

solubilities of graphite in liquid
iron and periclase in slags arises
the question how these materi-
als can be used succesfully for
the production of liquid steel. In
addition the rate of dissolution is
remarcable high. From an esti-
mate by the Nernst equation
∆x ρ C − C0
= β ⋅ S ⋅ sat
∆t ρR 100

follows at about 1600 °C for

∆x
graphit/steel ––
∆t
≅ 14 cm/h and T [C]

for periclase/slag ∆x
–– ≅ 0,3 –
∆t Fig. 1 Reaction between periclase and graphite during the total pressures of 2 atm and
2 cm/h. This paper looks for ar- 0,2 atm [2]

guments why MgO – C refracto-

ry material is applicable in steel rities, especially at the determination of the ture, in which gaseous Mg and CO are gen-
components and phases, which falsify the erated in equal parts:
metallurgy. result [1]. Essential is
〈MgO 〉 + 〈C 〉 = {Mg } + {CO } (1.1.2)
1 Phase equilibria in MgO – C F=K+2-P-R-B (1.1.1)
materials Hence, it is essential:
F = Number of degrees of freedom, i.e. K = 4 (MgO, C, {Mg}, {CO})
1.1 MgO – C number of state variables. Degrees of P = 3 (MgO, C, Gas)
The “Gibbs phase rule” describes phase freedom (pressure, temperature, con- R = 1 (1.1.2)
equilibria. It is a thermodynamic justified centration), which are essential for B = 1 (PMg = PCO),
law, which is often just described as a rule, the determination of the equilibrium.
because its practical use can lead to insecu- K = Number of all chemical components i.e. F = 4 + 2 – 3 – 1 – 1 = 1 (temperature
P = Number of the coexisting phases dur- or gas pressure).
ing the equilibrium Fig. 1 shows that the reaction, at a total
Jürgen Pötschke R = Number of the linear independent re- pressure of 2 atm (PCO = PMg = 1 atm), hap-
45219 Essen-Kettwig actions pens at 1848 °C. If the total pressure drops
Germany B = Number of further constraining condi- to 0,2 atm (PCO = PMg = 0,1 atm), periclase
tions, e.g. stoichiometric composi- and graphite already react at 1598 °C. C is
E-mail: jpoet@t-online.de tions during the change from one completely transformed into carbon monox-
substantial initial state into the equi- ide, because of the stoichiometric excess of
Keywords: MgO – C refractory, steel librium. MgO.
metallurgy, slag, oxidation, applicability The temperature of reaction TMg, CO[°C], dur-
MgO bricks are consisting of about ing the given gas pressure and where the
Received: 21.07.2012 90 mass-% MgO and 10 % graphitated car- complete reaction into the gases {Mg} and
Accepted: 18.09.2012 bon. Both components (MgO and C) are re- {CO} happens, is fixed and diagramed in
acting with each other at a high tempera- Fig. 2. The red graph shows that the reaction

80 refractories WORLDFORUM 5 (2013) [2]

 T (Pges) Spinel and Co+Mg
(1.1.2), during a loss of pressure, proceeds
at an increasingly low temperature. As long P = 5 (MgO, C, Al4C3, MgAl2O4, gas) (1.2.4)
as both components 〈MgO〉 and 〈C〉 are ex- R = 1 (1.2.2)
isting this dependency is essential, even if i.e. F = 5 + 2 – 5 – 1 = 1 (temperature or
antioxidants like aluminium are in the mate- total pressure).
rial or iron infiltrates. Approximately is:
But, with an increase of temperature, this re-
TMg, CO = 1757 · P 0,056
[atm] [°C] (1.1.3) action leads in principle into the known re-
action

T [°C]
1.2 MgO – C – Al pressure
〈MgO 〉 + 〈C 〉P=={Mg
0,2atm
} + {COleads
}. to(1.1.2)
the reaction temperature ~1600°C and therefore
Fig. 3 graphs the case conditions which was chosen for the calculation of the all following examples.
show when the refractory material contains Al4C3 disapears (aAl4C3 < 1) and the previ-
3 mass-% metallic aluminium as antioxi- The built
ously conditions
spinel does during the presence
not take part. At of the metalic aluminium differ vastly from
dant. The oxygen affine metal shall bind the the two – component system
1329 °C the entire available gas pressure MgO – C. Given that aluminium melts at 660°C
aerial oxygen, which is infiltrating the struc- and at the latest reacts then
during the reaction (1.2.2) of carbide to quantitatively to aluminium carbide:
ture, as aluminium oxide, before it oxidates spinel is shown by {Mg}, namely P [atm]
the graphite into {CO}. Oxygen would oxi- PMg = 0,2 atm. The reaction (1.1.2) starts 4( Alal-) + 3 C = Al4C3 . (1.2.1)
date the graphitic bonding matrix of the most impalpably, already below the F ig. 2 Dependency of the reaction temperatures
TMg, CO [°C], the formation of {Mg} and {CO} (eq.
structure and lead to a decomposition of it. temperature
Here the Tmetalic Mg, CO. Because {CO} is formed
aluminium completely 1.1.3,disappears and parts
upper red graph), as wellofas the graphite
TK,S [°C], the for-
The total pressure P = 0,2 atm leads to the and the fixedbonded.
become propotionAfterwards
{CO}/{Mg} = 1 isthenottwo components Al4C3, and MgO
mation of spinel (eq. 1.2.3, lower blue graph)atofathe
react
reaction temperature ~ 1600 °C and there- further valid, one T
temperature now has F = 2:
K,S[°C], just dependenttotal on the gas Ppressure,
pressure with
[atm] of both the[2]
gases formation of
fore was chosen for the calculation of the all Tspinel,
K, S ≤ T ≤additional
TMg, CO graphite and gaseous magnesium:
following examples. Pges ≥ PMg ≥ 0,5 Pges (1.2.5) ature TMg,CO, namely at 1598 °C. The partial
The conditions during the presence of the 0 ≤ PCO2≤ 0,5 P
Al4C3 + 16 MgO = 4 MgAl2O4 of+the6 magnesium
pressure C + 12 decreases
Mg paral-
metalic aluminium differ vastly from the two
ges { }
lel from PMg = 0,2 atm to 0,1 atm although
(1.2.2)
– component system MgO – C. Given that The partial pressure of the carbon monoxide its mass increases. Thereby the trend of the
aluminium melts at 660 °C and at the latest inThisthe reaction
calculated at P = 0,2
example atm takes
amounts only place at 1330°C.
concentrations Thecomponents,
of all blue graph in Fig. 1.1.2.
during
reacts then quantitatively to aluminium car- shows the temperature dependency
PCO ≈ 10 atm at 1329 °C. The partial pres-
-5 T
identical initial conditions, is always the is valid:
K,S (P ges). The following approach
bide: sure reaches its stoichiometric required val- same, just shifted by the temperature
ue PCO = 0,1 atm only TKafter 1503
, S =the ⋅ P[0,07
complete atm ] [∞C ] 2) given
(Fig. (1.2.3)
by the pressure.
4 (Al ) + 3 〈C〉 = 〈Al4C3〉 (1.2.1) transformation of the minority component The following key sentences should be con-
graphite to carbon monoxide at the temper- sidered:
Here the metalic aluminium completely
disappears and parts of the graphite be-
come bonded. Afterwards the two com-
ponents Al4C3, and MgO react at a tem-
perature TK, S [°C], just dependent on the
gas pressure, with the formation of spinel,
additional graphite and gaseous magne-
sium:

2 〈Al4C3〉 + 16 〈MgO 〉 =
4 〈MgAl2O4〉 + 6 〈C 〉 +12 {Mg } (1.2.2)

This reaction at P = 0,2 atm takes place at

1330 °C. The blue graph in Fig. 2 shows the
temperature dependency TK, S (Pges). The fol-
lowing approach is valid:

TK, S = 1503 · P [0,07

atm] [°C] (1.2.3)

This reaction has just one degree of freedom

T [C]
as well:
Fig. 3 Reaction between periclase, graphite and aluminium at a total pressure of
K = 5 (MgO, C, Al4C3, MgAl2O4, {Mg}) P = 0,2 atm [2]
4

refractories WORLDFORUM 5 (2013) [2] 81

 It cannot be excluded that molten iron infiltrates into the porosities of the
structure, one talks about infiltration. To understand the chemical interaction of
the components in the steel metallurgy, temperatures above 1500°C should be
considered. Fig. 1.3.1 graphs the result of the calculated example /2/. The main
reaction (1.1.2) <MgO> + <C>= {Mg} + {CO} has one degree of freedom, as
always. Additionally the reactions of all components happen with the iron melt.
The concentrations of solute magnesium and oxygen during the equilibrium are
that marginal that their graphical depiction can be ignored.

ponents in the steel metallurgy, tempera-

tures above 1500 °C should be considered.
Fig. 4 graphs the result of the calculated ex-
ample [2]. The main reaction (1.1.2)
〈MgO〉 + 〈C〉 = {Mg} + {CO} has one degree
of freedom, as always. Additionally the reac-
tions of all components happen with the
iron melt. The concentrations of solute mag-
nesium and oxygen during the equilibrium
Periclase, graphite and solute aluminium start to react with each other to spinel, are that marginal that their graphical depic-
{Mg} and {CO} from a temperature of 1589°C i.e. it constitutes a gas phase justtion can be ignored.
below 1600°C: The solubility of the carbon amounts about
5 mass-%, so that the remaining 10 g of
[ ]
5 MgO + C + 2 Al = MgAl 2O 4 + 4 Mg + CO (1.4.4) { } { } graphite, during the reaching of the temper-
ature of reaction TMg, CO = 1598 °C, react
K = 7 (MgO, C, MgAl2O4, Fe, [Al], {Mg}, {CO}) spontaneously with periclase to {CO} and
P = 5 (MgO, C, MgAl2O4, melt, gas) (1.4.5) {Mg}. Subsequently the carbon, dissolved in
T [C]
R = 1 (1.4.4) the iron, reacts progressively with an in-
FFig.
= 7 F1.3.1: 5Reaction
+ig.24-Reaction between
- 1 =between periclase,
3 (temperature,
periclase, graphite
gas pressure
graphite and and
and
molten molten
at a totaliron
amounts
iron of at aoftotal
material).
pressure crease of temperature under reduction of the
(At TP>=1600°C pressure of P =is 0,2atm
solid carbon
0,2 atm [2] replaced /2/.by its dissolution and it remains F = 3.)remaining MgO. This reaction has two de-
grees of freedom:

The solubility of the carbon amounts about 5wt%, so that the remaining 10g of〈MgO〉 + [C]= {Mg} + {CO} (1.3.1)
graphite, during the reaching of the temperature of reaction TMg,CO = 1598°C,
react spontaneously with periclase to{CO} and {Mg}. Subsequently the carbon,K = 5 (MgO, Fe, [C], {Mg}, {CO})
dissolved in the iron, reacts progressively with an increase of temperature underP = 3 (MgO, melt, gas) (1.3.2)
reduction of the remaining <MgO>. This reaction has two degrees of freedom: R = 1 (1.3.1)
B = 1 (PMg = PCO)
6
F = 5 + 2 - 3 – 1 – 1 = 2 (temperature or
gas pressure and amounts of material).

1.4 MgO – C – Al – Fe
The risk of infiltration by molten iron in-
creases if the refractory material contains
metallic aluminium. The reasons are ex-
plained later (chapter 2.3.2).
The aluminium carbide (Al4C3), which is
T [C]
already stable at a low temperature,
Fig. 5 Reaction between periclase, graphite, aluminium and molten iron at a total reacts latest at its contact with the molten
Fig. 1.4.1: Reaction between periclase, graphite, aluminium and molten iron at a
pressure of P = 0,2 atm [2] iron to carbon and aluminium, dissolved
total pressure of P = 0,2atm /2/.
in iron:
• Temperature TMg, CO (Pges) determines the structure. This metalic magnesium func-
limit of application of the material tions as antioxidant. This is the first pro- 〈Al4C3〉 = 4 [Al ] + 3 [C ] (1.4.1)
MgO and the remaining graphite react when the reaction temperature TMg,CO ,
(eq. 1.1.2). The binder phase turns into tective mechanism.
here 1598°C, is reached and until the graphite is completely spent. This happens
gas when the limit is exceeded and the • The spinel hinders the diffusion of the gas- K = 4 (Al4C3, Fe, [C], [Al])
after the main reaction (Gl. 1.1.2) <MgO> + <C>= {Mg} + {CO}, which has
material dissolves. In the example this are es through the densification of the struc- P = 2 (Al4C3, melt) (1.4.2)
only one degree of freedom. The solute components and the remaining periclase
1600 °C at 0,2 atm. ture and consequently improves the R = 1 (1.4.1)
balance, parallel and with increasing temperature, into the thermochemical
• The spinel, built according to (eq. 1.2.2), (chemical) stability of the MgO – C mate- F = 4 + 2 – 2 – 1 = 3 (temperature, gas
equilibrium. However, this period of the reaction has no meaning for the
does not take part in any of the following rial. This is the second protective pressure and amounts of material)
resistance of the refractory material, because the bonding has already dissolved
reactions. mechanism.
and the material has lost its stability.
• From the thermodynamic point of view the (The instantaneous dissolution of graphite
aluminium is totally bonded as spinel and 1.3 MgO – C – Fe and the molten aluminium in molten iron
therefore is completely omitted as antioxi- It cannot be excluded that molten iron from 1500 °C is equal with the reaction
dant. infiltrates into the porosities of the structure,
8(1.4.1))
• As a result of the formation of spinel a one talks about infiltration. To under- Fig. 5 shows the further progress. Firstly, up
high amount of magnesium steam is in the stand the chemical interaction of the com- to 1589 °C, periclase and the solute alu-

82 refractories WORLDFORUM 5 (2013) [2]

minium react to spinel, without building a
gas phase:

Incorporation in steel mass-%

〈MgO〉 +2 [Al ] + 3 [O] = 〈MgAl2O4〉 (1.4.3)

Hereby the initially in iron dissolved oxygen

content attunes to equilibrium. Because of
the low content of just a few ppm and a
high thermochemical stability of both oxides
Fig: 2.2.1: Chronological sequence of the concentration of solute carbonate and
just very little spinel generates. (~ 10-2 g).
Periclase, graphite and solute aluminium aluminium in the iron melt /3/.
start to react with each other to spinel, {Mg}
and {CO} from a temperature of 1589The °C i.e.dashed graph describes the chronological increase of the solute carbon
it constitutes a gas phase just during below the use of the MgO-C material. Time
The[h] dissolution stopps after about 30
1600 °C: out to of MgO
besolute (compare Figin2.2.2), is
minutes. A white layer, which turns
Fig. 6 Chronological sequence of the concentration carbon and aluminium
the concentration of solute carbonate and
the Fig: [3] Chronological sequence of
2.2.1:
/3/. infiltrated just at a few places.
iron melt Iron
5 〈MgO 〉 + 〈C 〉 + 2 [Al ] = noticeable when the sample isinextracted.
aluminium the iron melt
〈MgAl2O4〉 + 4 {Mg } + {CO } (1.4.4)
The dashed graph describes the chronological increase of the solute carbon
K = 7 (MgO, C, MgAl2O4, Fe, [Al], {Mg}, {CO}) during the use of the MgO-C material. The dissolution stopps after about 30
P = 5 (MgO, C, MgAl2O4, melt, gas) (1.4.5) minutes. A white layer, which turns out to be MgO (compare Fig 2.2.2), is
R = 1 (1.4.4) noticeable when the sample is extracted. Iron infiltrated just at a few places.
F = 7 + 2 – 5 – 1 = 3 (temperature, gas
pressure and amounts of material).
At T > 1600 °C solid carbon is replaced by
its dissolution and it remains F = 3.

MgO and the remaining graphite react when

the reaction temperature TMg, CO, here
1598 °C, is reached and until the graphite is
completely spent. This happens after the
main reaction (eq. 1.1.2) 〈MgO〉 + 〈C〉 =
{Mg} + {CO}, which has only one degree of
freedom. The solute components and the re-
maining periclase balance, parallel and with
increasing temperature, into the thermo-
chemical equilibrium. However, this period of
the reaction has no meaning for the resist-
ance of the refractory material, because the
bonding has already dissolved and the ma-
terial has lost its stability.
The following key sentences are important:
• The main reaction (equ. 1.1.2) 〈MgO〉 +
〈C〉 = {Mg} + {CO} has always one de-
Fig. 7 MgO– C (at the top) and MgO-C-Al (at the bottom) after the reaction with iron at
gree of freedom, gas pressure or tempera-
ture..
1600 °C 10
10
• The interlinking between the reaction tem-
perature TMg, CO and the total pressure Pges component in a thermochemical consider- 2 Chemical wear of MgO – C
is always valid, even if iron infiltrates. Con- ation. The same is valid for solute carbon, materials in molten iron
versely this means that this equilibrium oxygen, magnesium etc.
pressure regulates in the structure at a • Infiltrated parts of the structure behave 2.1 Experiments
predetermined service temperature. At different than non – infiltrated. A mean Pure iron (2 kg) is inductively (10 kHz)
1600 °C this are 0,2 atm, hence it is pos- value in the behaviour of both will result molten in a crucible of sintered corundum
sible to calculate with PMg = PCO = 0,1. with some statistical spread, in case that and hold at 1600 °C. The vacuum flask con-
• Aluminium is solute as metal in the infil- the infiltration does not takes place com- tains purest argon (800 mbar). A test bar
trated iron and can be treated as own pletely. (20 x 20 mm2) is 40 mm deep dipped into

refractories WORLDFORUM 5 (2013) [2] 83

 Fig. 2.2.2: MgO-C (at the top) and MgO-C-Al (at the bottom) after the reaction
with iron at 1600°C

The MgO-C-Al material behaves completely different. The concentrations of

carbon and aluminium increase continuously. The surface stays black, i.e. it is
not covered by a MgO-layer. The structure is infiltrated by iron for about 4mm
deep. Fig. 2.2.3 depicts the distribution of solute carbon and aluminium in the
infiltrated iron. Both concentrations increase with a longer distance from the
surface.
crease with a longer distance from the sur-
face.
The solute oxygen concentration measured
in the infiltrate is, for a huge number of such
experiments, laid on the analyzed carbon
concentration and double logarithmic dia-
gramed in Fig. 9. The result is a relatively
dense distribution range, in which the MgO-
C values are by trend lower in carbon and
higher in oxygen than the MgO–C–Al pairs
of values. The envelope can be allocated to
CO partial pressures between 0,01 and
0,1 atm.
In future all calculations will be with
PCO = 0,05 atm, which is approximately that
calculated for 1600 °C.
Fig.1.2.4: Vacher – Hamilton equilibrium in molten iron (1600°C), measured in
dissolution experiments of different MgO-C and MgO-C-Al
Fig. 8 Solute concentrations of carbonate (black) and aluminium (white) in the infiltrat- 2.3 /4/. Discussion
materials
Fig. iron [4] Solute concentrations of carbonate (black) and aluminium (white) in
ed 2.2.3:
the infiltrated iron /4/. 2.3.1 Infiltration
Fig.1.2.4:2.3 Discussion
Vacher – Hamilton equilibrium in molten iron (1600°C), measured in
The solute oxygen concentration measured in the infiltrate is, for a huge dissolution
number The equilibrium
experiments between
of different the and
MgO-C gravitation
MgO-C-Al
2.3.1 Infiltration
materials /4/.
of such experiments, laid on the analyzed carbon concentration and double pressure and the capillary pressure are es-
logarithmic diagramed in Fig. 2.2.4. The result is a relatively dense distribution sential
The equilibrium for thetheinfiltration:
between gravitation pressure and the capillary pressure are
essential for the infiltration:
2.3 Discussion
range, in which the MgO-C values are by trend lower in carbon and higher in 4 σ cosθ  2
Pa + ρ ⋅ g ⋅ h =  dyn / cm  . (2.3.1.1) (2.3.1.1)
[O] [mass-%]

oxygen than the MgO-C-Al pairs of values. The envelope can be 2.3.1allocated
Infiltration to d
CO – partial pressures between 0,01 and 0,1 atm. In future we are going to
If the melt
The equilibrium wets the
the refractory
If the melt wets material
the (refractory
< 90°) thematerial
melt pressure
spontaneously
calculate with PCO = 0,05 atm, which is approximately that calculated for between
essential infiltrates
gravitation pressure and the capillary are
into porosities (slags). The gravitation pressure must be bigger than
for the infiltration:
1600°C. (θ < 90°) the melt spontaneously infiltrates
the capillary pressure
4 σ tocosreach
θ infiltration,
2 if the wetting angle is > 90°, i.e.
(slags).
/ cmThe . gravitation pres-
Pa +non
during ρ ⋅ gwetting
⋅ h = porosities
into (steel).
d  dyncritical
The diameter d [cm] is:(2.3.1.1)
sure must be bigger than the capillary pres-
If the melt wets the refractory material4σ (⋅ cos
< θ90°)[ifcm
the
sure to reach
d ≥ infiltration, ] .melt
the spontaneously
wetting angle (2.3.1.2)
infiltrates into porosities (slags). ThePa gravitation
+ ρ ⋅ g h pressure must be bigger than
the capillary pressureistoθreach
> 90°, i.e. during
infiltration, if thenon
wetting wetting
angle (steel).
is > 90°, i.e.
during non Bywetting Theinformation
critical
(steel).
the typical diameter
The critical d [cm]
Fediameter
= 1550mN/m,d [cm] is: is:Fe/R = 120°, d = 2 10-3cm, Fe=
[C] [mass-%] 7g/cm3 and g = 981cm/s2 and Pa= 9.81 105g/cm s2 one receives h = 83cm, i.e.
4σ ⋅ cosθ -3
the porositiesd11
with
≥ a diameter[of cmd] =. 2 10 cm are (2.3.1.2) just(2.3.1.2)
infiltrated by the iron in
a deeper metal level.
Pa + ρ ⋅ g h
Fig. 9 Vacher-Hamilton equilibrium in molten iron (1600 °C), measured in dissolution
Fig.1.2.4: Vacher
experiments – Hamilton
of different MgO– equilibrium
C and MgO– C–in Al molten
materialsiron
[4] (1600°C), measured
By the inBy 2 the
typical information typical information
Fe = 1550mN/m, Fe/R = 120°, d = σ =-3cm, Fe=
2 Fe10
7g/cm3 and g = 981cm/s 5 2 12
dissolution experiments of different MgO-C and MgO-C-Al 1550 mN/m, θFe/R = 120°, d = 2·10 cm, ρ=Fe 83cm, i.e.
and P a= 9.81 10 g/cm s one receives
-3 h
the porosities with a diameter of3 d = 2 10-3cm are just 2infiltrated by the iron in
the melt materials /4/. different materials served during the use of the MgO–C–Al
up to 5 h. Two a deeper metalma-level. = 7 g/cm and g = 9,81 cm/s and Pa =
are analyzed: MgO–C (14 % flake graphite, terial. Fig. 6 graphs the quantitative results 9,81 · 105 g/cm · s2 one receives h = 83 cm,
resin-bonded) and the same [3, 4]. i.e. the porosities with a diameter of
12
2.3 MgO–C
Discussion
with an addition of 3 mass-% The dashed graph describes the chronologi- d = 2 · 103 cm are just infiltrated by the iron
metallic aluminium as antioxidant. Both ma- cal increase of the solute carbon during the in a deeper metal level.
2.3.1terials
Infiltration
have been carbonized at 450 °C for use of the MgO–C material. The dissolution The infiltration depth of steel and slag is cal-
6 h. The concentration of oxygen is con- stopps after about 30 min. A white layer, culated by their flow rate in the porosities
The equilibrium between The infiltration depth of steel and slag is calculated by their flow rate in the
trolled during the the gravitation
entire experiment via a pressure
which turns and
outthe
to becapillary
MgO (compare Fig. 7),are (Hagen – Poiseuille [5]):
pressure
porosities (Hagen – Poiseuille /5/):
essential for theininfiltration:
permanently the crucible installed EMF – is noticeable when the sample is extracted.
probe. Samples are taken out 4 σof cos
the θmelt Iron infiltrated just at a The
few infiltration
places. depth of steel and  4 ⋅ σ ⋅ cosθ 
P + ρ ⋅g ⋅h=  2 d 2 slag
Pa +isρ calculated
⋅ g ⋅h + by their flow rate in the
during the aexperiment and testedd for their  dyn /The  .
cmMgO–C–Al material behaves (2.3.1.1)
complete-
dl
porosities (Hagen – Poiseuille= /5/):
 d  [cm / s ] (2.3.1.3)
dt 32 ⋅η ⋅ l
content of solute carbon and accordingly ly different. The concentrations of carbon
 4 ⋅ σ ⋅ cosθ 
If thealuminium
melt wets [3].the refractory material ( and aluminium increase continuously. The d 2  Pafactor/6,7/
+ ρ ⋅ g ⋅h +  (2.3.1.3)
< 90°) the melt spontaneous and ly
by the dltortuosity
 d  [cm / s ]
bydta= x (2.3.1.3)
infiltrates into porosities (slags). The gravitation surface stays black, must
pressure i.e. it isbenotbigger
covered than λ32=⋅η =⋅ l1,84 ⋅ ε − 0,3[−] . (2.3.1.4)
2.2 Results
the capillary pressure to reach infiltration, if the wetting angle is >It describes
MgO-layer. The structure is infiltrated by iron and by the l
tortuosity factor [6, 7]
90°,
and by the tortuosity
i.e.the relationship between the externally measured penetration depth x
duringGasnonbubbles, which(steel).
wetting bubble up
The to critical for about
the assaydiameter d 4[cm]
mm deep.
is: Fig. 8 depicts dis-factor/6,7/
theintricate
and the actual xpath l of the liquid in the inner part of the structure of the
surface, are observed after the immersion of tribution of solute carbon and aluminium porosity in . Equation = 1,84 ⋅is εinserted
λ = (2.3.1.4) − 0,3[−in ] . (2.3.1.3) and
(2.3.1.4) (2.3.1.4)after the
integrated
l
separation of the variables:
the MgO–C sample material.4This is
σ ⋅ cosθnot ob- the infiltrated iron. Both concentrations in-
It describes the relationship between the externally measured penetration depth x
2
d≥ [cm] .
and(2.3.1.2)
the intricate actual
x = λpath
d ⋅ σ ⋅ cosθ d in
⋅ l of the liquid+
(Pathe+ ρinner ) ⋅ of
⋅ g ⋅ hpart
t [cmthe
] structure of the
(2.3.1.5)
Pa + ρ ⋅ g h porosity . Equation (2.3.1.4)4 ⋅isη inserted in16(2.3.1.3) ⋅η and integrated after the
separation of the variables:
84 refractories WORLDFORUM 2
5 (2013) [2]
By the typical information Fe = 1550mN/m, Fe/R = 120°, d = 2 10 -3 result of
The
x cm,
= λ ⋅
d ⋅the
Fe
cosθ d (depth
σ=⋅ penetration
+
g ⋅ h )t - law. By the above used numbers, the
Pa + ρis⋅ the
⋅ t [ cm ] (2.3.1.5)
7g/cm3 and g = 981cm/s2 and Pa= 9.81 105g/cm s2 one receives h =viscosity 83cm,
Fe= 4
i.e.
0.02
⋅η , poise (g/(cm 16 ⋅ηs) and the tortuosity factor = 0,5 the result for
penetration depth of the iron in a metal level of h = 100cm is x[cm] = 5 ⋅ t [min] .
the porosities with a diameter of d = 2 10-3cm are just infiltrated bythethe iron in about
The resultThe
of cast penetrates
the penetration depth15cmis thein 10t -minutes
law. Byinto thetheabovestructure, if the cast the
used numbers, does not
The infiltration depth of steel and slag is calculated by their flow rate in the
porosities (Hagen – Poiseuille /5/):

 4 ⋅ σ ⋅ cosθ 
d 2  Pa + ρ ⋅ g ⋅ h + 
dl the relationship between
It describes d the
 [cmex-/ s] rial, (2.3.1.3)
which declines the melt and hinders PCO = 0,05 atm. This sixth protective
=
dt 32 ⋅ η ⋅
ternally measured penetration depth x andl their penetration into the structure. This ex- mechanism limits the disaggregation of
and by the the intricate
tortuosity actual path l of the liquid in the
factor/6,7/ plains the gas bubbles observed in the ex- the graphitic bonding phase by infiltrated
inner part of the structurex of the porosity ε. periment (see 2.2) and could be a fourth iron.
λ = = 1,84 ⋅ ε − 0,3[−] . (2.3.1.4)
Equation (2.3.1.4)l is inserted in (2.3.1.3) protective mechanism. The use of aluminium containing MgO – C
It describesandthe relationship
integrated between
after thetheseparation
externally measured
of the penetration depth x however, contains metallic
If the material, materials does not appear so nice. Spinel
and the intricate actual path l of the liquid in the inner part of the structure of the
porosity variables:
. Equation (2.3.1.4) is inserted in (2.3.1.3) and integrated aluminium afteras theantioxidant the concentration (MgAl2O4) generates via deoxidation of the
separation of the variables: of oxygen is reduced so far, by deoxidation infiltrate:
2
d ⋅ σ ⋅ cosθ d (Pa + ρ ⋅ g ⋅ h ) of the melt in the area of the phase bound-
x=λ⋅ + ⋅ t [cm] (2.3.1.5)
4 ⋅η 16 ⋅η ary that this protective mechanism cannot [Mg ] + 2 [Al ] + 4 [O ] =
(2.3.1.5) work anymore. MgAl
〈The use2Oof4〉,aluminium
∆G0 = – containing 692 kJ/molMgO(2.3.1.8) – C materials does not appear so
The result of the penetration depth is the t - law. By the above used numbers, the Spinel (MgAl2O4) generates via deoxidation of the infiltrate:
The result of the penetration depth is the The formation of the MgO-layer on the sur-
viscosity Fe= 0.02 , poise (g/(cm s) and the tortuosity factor = 0,5 the result for
√t-law.
the penetration depthByofthe theabove
iron inused numbers,
a metal level of the h = vis-
100cm isfacex[cmof] =the
5 ⋅ aluminium
t [min] . free MgO – C materi- [ Mg ]solubility
The + 2[ Al ] + 4[ product
O] = MgAl is O , ∆G 0 = −692kJ / mol (2.3.1.8)
2 4
cosity The use of aluminium containing MgO – C materials does not app
η
The cast penetrates about Fe = 0,0215cmPin(g/(cm·s)
10 minutes structure, if the cast doeseffect,
andintothethetortu- al is a further not connected with the hin-
Spinel (MgAl2O4) generates via deoxidation of the infiltrate:
solidify before
osityinfactor
the temperature
λ = 0,5 the gradient
resultoffor
the thelining.
pene-However, the non
dered wetting of(compare upper part of
infiltration [Mgsolubility
The ] · [Al ]2 ·product[O ]4 =is5 · 10-20 (1600 °C)
the refractory material is the third protective mechanism against corrosion. 0
tration depth of the iron in a metal level of Fig. 7). The evaluation of the Vacher-Hamil- [ Mg ] + 2[ Al ] + 4[O] = MgAl2O4 , ∆G (2.3.1.9)
= −692kJ / mol (2.3.1
cm isatxthe
[ Mg ] ⋅ [ Al ]2 ⋅ [O]4 = 5 ⋅ 10−20 (1600∞C ) (2.3.1.9)
A CO – bufferh = can100occur [cm]
two=phase
5 · √t[min].
boundaryThe cast
refractory/iron,ton ifequilibrium
the pure in Fig. 9 accounts a middle
MgO – C material
penetrates is used,
aboutcomparable
15 cm into 10 the min
“Leidenfrost
into theeffect”partial
for boiling
pressure of the carbon monoxide in The solubility product is
1/4 1/4
water. For the deoxidation with carbonate the following reaction is essential  5 ⋅ 10−20   5 ⋅ 10−20 
structure, if the cast does not solidify before the infiltrated structure of 0,05 atm. Accord- [ O] = =  −20 −3 = 3 ⋅ 10−4 wt.% (2.3.1.10)
[ Mg ]⋅ [[ Mg
Al ]]2⋅⋅[[Al O]]24 = 5 ⋅ 10 (1600 ∞ C ) 2 (2.3.
0    [1,2 ⋅ 10 ] ⋅ [0,1] 
[C] +in[O]the= CO{temperature
, G (1600°C) gradient
= -93ofkJ/mol.
the lining. ingly,(2.3.1.6)
the partial pressure of the magnesium
With: However, the non wetting of the refractory is PMg = 0,05 atm. The gases diffuse to the 3 · 102,4 -4 mass-%
0,05 1/ 4  5 ⋅ 10(2.3.1.10)
1/ 4
⋅ 10−3 −⋅ 20 −20
K = 4 ([C], [O], Fe, CO) [[CO]]==  5 ⋅ 10 = 0,4 = wt
 .%,(VacherHamilton  )
 = 3 ⋅ 10−4 wt.% (2.3.1.11)
(2.3.1
material
P = 2 (melt, gas phase) is the third protective mecha- surface of the material and dissolve into the 0,0003
[ Mg ] ⋅ [ Al ] 2 −3
[1,2 ⋅ 10 ] ⋅ [0,1] 2

nism
R = 1 (2.3.1.6), against corrosion. melt. The latter contains e.g. 3 · 10 mass-% -3
The average concentration of the aluminium, dissolved in the infiltrate, is [
F = 4 + 2 - A2 - 1CO-buffer
= 3 (pressure,can temperature,
occur one at concentration).
the two solute oxygen, so that magnesium immedi- 0,1wt% 2,4 ⋅ 10−3 ⋅ 0,05 -3wt%. According to Gl. (2.3.1.9) the concentrati
The concentration of oxygen is fixed, if the temperature, the pressure and the [C ] = next to [Mg] = =100,4 wt.%,(VacherHamilton) (2.3.
phase boundary oxygen accounts 0,0003[O] = 3 10-4wt%. Furthermore according to the Vacher –
carbon concentration is given.refractory/iron, if the pure ately reacts to MgO, which in heterogenous
The use of aluminium containing MgO carbon– C materials
Hamilton equilibrium, [C] = 0.4wt% is solubledoes at a COnot –appear
partial so
pre
MgO – C material is used, comparable to the nucleation precipitates on the surface of the Spinelof 0,4
The
(MgAl mass-%
PCOaverage
=2O
(Vacher-Hamilton)
concentration
4) generates
0,05atm ofThis
via deoxidation
(Gl. 2.3.1.7).
(2.3.1.11)
the aluminium, dissolved
of the infiltrate:
corresponds with in the infiltrat
the measuring. In
The solubility product (Vacher-Hamilton-equilibrium)
effect” for boiling water. Foris calculated by material.
equ. 0,1wt% next -3
“Leidenfrost refractory
The use of aluminium containing MgO – C comparison materials does to tothe[Mg] = 10 wt%. According
solubility
not appearofso5,4wt% -4 nice.
at 1600°C to this
Gl. effect
(2.3.1.9)leadsthe
to conce
a clea
(2.3.1.6) to: [ Mg ] oxygen
2[ Al ] accounts
4[ O ] [O] = O
MgAl 3 10 , wt%.
0 Furthermore
692 / according to the Va
(2.3.1.8)
the deoxidation with carbon the following Because the solubility
Spinel (MgAl 2 4 of the magnesium in
O ) generates via deoxidation + limitation
The
of the + of
average
infiltrate: the = disaggregation
concentration
2 4 ∆ of
G the
= bonding
− of kJ phase
the mol as well.
Hamilton equilibrium, [C] = 0.4wt% carbon is soluble at a CO – part
reaction
[C ]⋅ [O ] = 2.4 ⋅ 10−3 ⋅isPCO essential
, (1600°C) /8/. molten iron at 1600 °C and PMg = 1 atm
(2.3.1.7) aluminium,
of PCODissolution
2.3.3 = 0,05atm dissolved
(Gl. 2.3.1.7).
speed in the This infiltrate,
corresponds with the measuring.
[ Mg ] + 2[ Al ] + 4[O] = MgAl2O4 , ∆G 0The = −solubility
692 kJ / mol
comparison
product is
the (2.3.1.8)
is [Mg] = 0,023 mass-%, this are is [Al] =to 0,1 solubility
mass-% of 5,4wt% next at 1600°C to this effect leads to
13 limitation
A plate 4of
2 of graphite the disaggregation
dissolves
−20 in of the
molten ironbonding
(1600°C) phase
with asthewell.
speed of
[C] + [O] = {CO }, ∆G0 (1600 °C) = [Mg] The = solubility
0,05 · 0,023 = 1,15 (2.3.1.9)
is · 10 mass-% [ Mg ] ⋅ [Mg]
[ Al ] ⋅=[O10 ] = mass-%. According to eq.
-3 -3 5 ⋅ 10 (1600 ∞C )
product
– 93 kJ/mol (2.3.1.6) for the analyzed case. The solubility (2.3.1.9) ∆x theρ1/Stconcentration
2.3.3 Dissolution speed
C − C0 of 7oxygen 5,3 ac-
[ Mg ] ⋅ [ofAl ]2 ⋅ [the
O]4 = 5 ⋅MgO 10−20 (1600 ∞C ) = β ⋅ 4 ⋅ s(2.3.1.9)
20 = 0,03 ⋅ ⋅ 1/ 4= 4 ⋅ 10−3 [cm / s ] (2.3.3.1)
product accounts counts5 ⋅ 10∆−t[O] =ρ3C · 10 5 ⋅ 10−202,9
-4 mass-%.
100 Furthermore
100
[O] = A plate of graphite 2
= dissolves−3in molten  iron ⋅ 10−4 wt.%
= 3(1600°C) with(2.3.1.10)
the speed
With: [Mg] · [O] = 2 · 10-6 [mass-%] 2 [8] and is ex- according
[ Mg ] ⋅ [ Al ]to the [1,2 ⋅ 10 ] ⋅ [0,1]2equilibri-
Vacher-Hamilton
− 20 1/ 4 − 20 1/ 4
K = 4 ([C], [O], Fe, CO) ceeded  5current
⋅ 10   5 ⋅product
10  This are∆14cm/h /3,4/.C −=C 0,03cm/s is 7the mass
5,3at atransfer coefficient in a sti
[Oby] = the

concentration
2
= um,= 3[C]
2  2,4
melt. ⋅ 10 =x−40,4.%
−3 =wt
Without
mass-%
ρ
⋅ Sthigh
β the s carbon
⋅(2.3.1.10) 0
decrease = 0,03
is soluble
of the⋅ saturation
⋅ 4 ⋅ 10−3 [cm / sC]S =(5,3wt%
= concentration 2.3.3.
[ Mg ] ⋅ [ Al ] [1,2 ⋅ 10 −3
] ⋅ [0,1] ⋅ 10 ⋅ 0,05
P = 2 (melt, gas phase) 1,15 · 10 · 3 · 10 = 3,5 · 10 [mass-%] , CO-partial pressure 100 2,9 100
the useofofPthe CO = 0,05– Catm (eq. would be much
-3 -3 -6 [2C ] = ∆ t ρ
= 0,4 wt .%,( VacherHamilton ) (2.3.1.11)
less 0,0003
than 0,4wt% C MgO material more
R = 1 (2.3.1.6) which is condition for the exsolution of 2.3.1.7).
problematic. This corresponds with the measur-
2,4 ⋅ 10−3 ⋅ 0,05 This )are 14cm/h /3,4/. = 0,03cm/s is the mass transfer coefficient in
[C ] =
〈MgO〉. For this0,0003 = 0,4
reason the disaggregation ofwt .%,( VacherHamilton
The average
ing. concentration (2.3.1.11)
of theto aluminium, dissolved
of in the infiltrate, is [A
melt. In Without comparison
the high decrease the solubility
of the saturation concentration CS = 5
0,1wt% next to [Mg] = 10-3wt%. According to Gl. (2.3.1.9) the concentratio
F = 4 + 2 – 2 – 1 = 3 (pressure, tempera- this material stopps after 30 min: the MgO oxygen – 5,4
less mass-% at 1600 °C this effect leads to a
than 0,4wt% the use of the MgO
accounts [O] = 3 10-4wt%. Furthermore according to the Vacher –
– C material would be much mo
The average concentration of the aluminium, problematic. dissolved in the infiltrate, is [Al] =
ture, one concentration). layer is a fifth protective mechanism. Hamilton clearequilibrium,
limitation of [C]the disaggregation
= 0.4wt% carbon isofsoluble the at a CO – partial pres
0,1wt% next to [Mg] = 10-3wt%. According to Gl. (2.3.1.9) the concentration of
The concentration of oxygen is fixed, if the of PCObonding
= 0,05atm (Gl. 2.3.1.7).
phase This corresponds with the measuring. In
oxygen accounts [O] = 3 10-4wt%. Furthermore according toasthewell Vacher[9]. –
temperature, the pressure and the carbon 2.3.2Hamilton
Solubility of carbon
equilibrium, [C]in the carbon
= 0.4wt%
comparison to the solubility of 5,4wt% at 1600°C this effect leads to a clear
is soluble at a CO – partial pressure
limitation of the disaggregation of the bonding phase as well.
concentration is given. iron-infiltrate
of P CO = 0,05atm (Gl. 2.3.1.7). This corresponds 2.3.3 with Dissolution
the measuring. speed In
comparison to the solubility of 5,4wt% at 1600°C this effect leads to a clear
The solubility 2.3.3
is Dissolution speed
limitation of carbon
of the in the infiltrate
disaggregation of the bonding A phase
plate as of well.
graphite dissolves in molten iron
The solubility product (Vacher-Hamilton-equi- highly reduced by the vapour pressure of theA plate(1600 °C) with
of graphite the speed
dissolves of iron (1600°C) with the speed of
in molten
2.3.3 Dissolution speed
librium) is calculated by equ. (2.3.1.6) to: magnesium. The CO-partial pressure can be
calculated
A platevia Fig. 9 and via the Vacher-iron (1600°C) ∆x ρ St Cthe − C0 7 5,3
of graphite dissolves in molten = β ⋅ with ⋅ s speed = 0,03of ⋅ ⋅ = 4 ⋅ 10−3 [cm / s ] (2.3.3.1)
∆t ρC 100 2,9 100
[C ] · [O ] = 2,4 · 10 · PCO, (1600°C) [8]
-3 Hamilton equilibrium (2.3.1.7). As shown,
∆x ρ Cs − C0 7 5,3
(2.3.1.7) the partial pressure = β ⋅ inSt the ⋅ structure = 0,03 is always
⋅ ⋅ = 4 ⋅ 10−3 [cm / s ] (2.3.3.1) (2.3.3.1)
∆t ρC 100 2,9 This100 are 14cm/h /3,4/. = 0,03cm/s is the mass transfer coefficient in a stir
between 0,01 ≤ PCO ≤ 0,1 atm. The average melt. Without the high decrease of the saturation concentration CS = 5,3wt%
A CO-partial pressure of more than 1 atm valueThisPCO are≈ 0,05 atm/3,4/. fits the= 0,03cm/s
temperature less thanThis0,4wt%
are 14the cm/h use[3, of 4].theβMgO – C cm/s
material would be much more
14cm/h is the mass transfer coefficient in a = 0,03
stirred is the
problematic.
developes, if a steel melt contains more than rangemelt.
of 1550Without – 1600 the high °C. decrease
If the steel of the
melt saturation concentration C =
mass transfer coefficient in a stirred melt.
S 5,3wt% to
less than 0,4wt% the use of the MgO – C material would be much more
25 ppm solute oxygen and when the melt with problematic.
a solute oxygen concentration of Without the high decrease of the saturation
gets in contact with the graphite of the 30 ppm infiltrates the structure, there would concentration CS = 5,3 mass-% to less than
MgO–C material, whose activity is aC = 1. be just [C] = 2,4 · 10-3 · 0,05/3 · 0,4 mass-% the use of the MgO – C materi-
A gas cap developes on the refractory mate- 10-3 = 0,04 mass-% carbon soluble at al would be much more problematic.

refractories WORLDFORUM 5 (2013) [2] 85

15
MgO – C and pure graphite are carried out in a hot stage microscope in an argon
atmosphere. The pulverized ladle slag, which is subsequently pressed into cubes
(3 x 3 x 3mm3), is lying on the refractory plates (10 x 10mm2) and heated up
slowly (3°C/min). Fig. 3.1.1. shows the results. MgO – C is wetted, but the
infiltration is less than with MgO.

infiltration and disaggregation is observed

gravimetrically. The result is shown in
Fig. 11.
The complete infiltration happens during
one minute. The material dissolves temporal-
ly linear in the non-saturated slag for 16 min
and disaggregates afterwards. On the con-
trary, the weight stays unchanged in the
MgO-saturated slag for 90 min. Conse-
quently the sample disaggregates sponta-
neously.

3.2 Discussion

3.2.1 Wetting
Fig. 10 indicates that refractory material on
The slag is melted in a crucible of platinum, which is resting on a ceramic
base MgO is wetted by slags and therefore
weighing scale.ofThe
Fig. 10 Wetting MgO,preheated testgraphite
MgO – C and cylinder
by (40mm 18mmØ) is driven into the
ladle slagx[9]
slag and fixed. The chronological sequence of its infiltration and disaggregation gets infiltrated. The graphite is not wetted,
Fig. 3.1.1: Wetting
is observed of MgO, The
gravimetrically. MgO result– isCshown
and ingraphite
Fig. 3.1.2. by ladle slag i.e.
/9/no spontaneous infiltration occurs. A suc-
cessful infiltration can only occur when the
2,0
slag height is adequate.
MgO
an non
MgO saturated
ungesättigt
The experiments for the infiltration and dissolution MgO
an
of
non
MgO
decarbonised
saturated
gesättigt
MgO
The MgO – C –contains graphite in
– C material
ladle bricks are carried out in a batch furnace at 1600°C in an argon atmosphere.
1,5 the bonding phase, however it is wetted. The
graphitised porosities hinder the penetration
Two types of slags are used /10/: of the liquid slag. A slag of the density
change [%] [g]

∞ One, non-saturated
1,0
on MgO, with the following composition; ρS = 50 wt.%
Gewichtsveränderung

2,7 g/cm 3 and the surface tension

CaO, 35 wt.% Al2O3 and 15 wt.% SiO2. Its MgO – solubility is mN/m
σS = 570 9,6 forms
wt.% the wetting angle of
0,5
θS/R = 85° with the refractory material and
/11/, i.e. 37,2 g MgO dissolves in 350 g slag. enters its porosities of the diameter
mass

∞ The same, 0,0

but MgO-saturated slag /11/. d > 8 · 10-4 cm, if it is h = 1 m high
(eq. 2.3.1.2). The graphite reduces the wet-
ting and therefore functions as a seventh
-0,5
protective mechanism.
0 5 10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 100

time [min] Zeit [min] 1

3.2.2 and
graphite reduces the wetting Infiltration
therefore functions as a seventh protective
Fig. 3.1.2: Dissolution and disaggregation of MgO in two ladle mechanism.
slags at 1600°C
Fig. 11 Dissolution and disaggregation of MgO in two ladle slags at 1600 °C [10] The infiltration time of the MgO – materials
/10/
calculates from eq. (2.3.1.5). In the labora-
3.2.2 Infiltration tory test h ≈ 0 cm, hence, the following is es-
The
3 Wearcomplete
of MgO infiltration happensinduring one minute. The material dissolves
– C materials
temporally linear in the non-saturated – Cfor
slag is wetted,
16 but theand
minutes infiltration is less thantime of
disaggregates
The infiltration sential:
the MgO – materials calculates from Gl. (2.3.1.5). In the
liquid slag with MgO. Graphiteinisthe
not MgO
wetted.
laboratory test h
0 cm, hence, the following is essential:
afterwards. On the contrary, the weight stays unchanged – saturated
The experiments for thespontaneously.
infiltration and dis- 4 ⋅ η ⋅ x2 (3.2.2.1)
slag for 90 minutes. Consequently the sample disaggregates t= 2 [s] (3.2.2.1)
3.1 Experiments and results solution of decarbonised MgO–C-ladle
λ ⋅ d ⋅ σSchlacke ⋅ cos θ

Infiltration and dissolution of refractory ma- bricks are carried out in a batch furnace at
The porosity is 20 vol%, i.e. = 0,2, from what the tortuosity factor is
terials are intimately
3.2 Discussion connected with its wet- 1600 °C in an argon atmosphere.
calculatedTwowithtypesGl. (2.3.1.4)
The porosity
as =is0,20 52.vol.-%, i.e. ε =
The average pore0,2,diameter
from is d =
ting behaviour. Hence, experiments for these of slags are used [10]: 8 10-4 cm. Furthermore whatthethewetting angle factor
tortuosity has been identified as
is calculated with= 34° /10/.
3.2.1 Wetting areas are arranged, evalu- • One, non-saturated on MgO, The surface tension is S = 570 mN/m and the viscosity = 1, 7 g/cm s /10,
three application with the fol-
12/. The infiltration time eq. t(2.3.1.4) as λ = 0,52.
= 54s is calculated Thehalf
by the average pore
cylinder radius x = 0, 9
ated and discussed. lowing composition; 50 mass-% CaO, quite
cm, which coincides diameter is dthe
well with = observation.
8 · 10-4 cm. Furthermore
Fig.
The 3.1.1 indicates
experiments that
for the refractory
wetting material35onmass-%
behaviour base MgO is wetted
Al2O3 and 15 mass-% by slags
SiO2. Itsand the wetting angle has been identified as
therefore The tendency for infiltration is one of the most important characteristic for the
of materialsgets infiltrated.
on the The MgO
base of MgO, graphite
– C is not wetted, i.e.isno9,6
MgO-solubility spontaneous
mass-%
judgement of i.e.
[11], θ =material.
refractory 34° [10]. The surface
In addition, the tendencytension is highly, if
increases
infiltration occurs.
and pure graphite areAcarried
successful
out in infiltration
a hot cang only
37,2 MgO occur
dissolveswhen
in 350 the slag is
g slag.
the graphite
height = 570(cf.
σS5wt%
less than mN/m
Fig. 5.1),and becausetheon theviscosity
one hand the
is adequate. porosity
stage microscope in an argon atmosphere. • The same, but MgO-saturated slagand the tortuosity
[11]. η = 1,7factor
g/cmincrease
· s [10, rapidly
12]. and
Theoninfiltration
the other hand the
porosities are increasingly less blocked by enclosures of graphite.
The pulverized ladle slag, which is subse- The slag is melted in a crucible of pla- time t = 54 s is calculated by the half cylin-
The MgO – C material contains graphite in the bonding phase, however it is
quently pressed into cubes (3 x 3 x 3 mm3), tinum, which is resting on a ceramic weigh-
wetted. The graphitised porosities hinder the penetration of the liquid slag. A der
3.2.3 Disaggregation
radius x = 0,9 cm, which coincides quite
of MgO in slag
is of
slag lying on the refractory
the density plates
3 ing scale. The preheated test cylinder well with the observation. The tendency for
S = 2,7g/cm and the surface tension S = 570mN/m forms
(10wetting
the x 10 mmangle
2) and heated up slowly (40 mm x 18material
of S/R = 85° with the refractory mm Ø) isand driven into Non
3.2.3.1
enters the
its -slag
saturatedinfiltration
slags is one of the most important
(3 °C/min). of
porosities Fig.the
10diameter
shows the dresults.
> 8 MgO 10-4 cm,andif itfixed.
is hThe chronological
= 1m high (Gl.sequence
2.3.1.2). of The
its characteristic for the judgement of refractory
The disaggregation of refractory material in slags takes place in four phases:
Infiltration, disaggregation of the grain bonding, break down of the structure and
the dissolution of the left crystallites in the slag. The reaction on the free surface
86 refractories
is generally
2 WORLDFORUM
not constitutive for the whole event. This5 observation
(2013) [2]is completely
proven by the existing experiments/12/.

W. Gans /13/ has already theoretically dealt with these incidents in his
professorial dissertation “Disaggregation and dissolution of heaps of debris“ in
 Fig. (3.2.3.1) is calculated by equ. (3.2.3.4) via the data mentioned above.

phite reduces the wetting and therefore functions as a seventh protective Fig. (3.2.3.1) is calculated by equ. (3.2.3.4) via the data mentioned above.
chanism. material. In addition, the tendency increases
highly, if the graphite is less than 5 mass-%
.2 Infiltration (cf. Fig. 17), because on the one hand the
porosity and the tortuosity factor increase
e infiltration time of the MgO – materials calculates from Gl. (2.3.1.5). In the
rapidly and on the other hand the porosities
oratory test h
0 cm, hence, the following is essential: Dissolution of single
are 4increasingly less blocked by enclosures crystalsquite after
⋅ η ⋅ x2 disaggregation
t= 2 [s] (3.2.2.1)
λ of⋅ dgraphite.
⋅ σSchlacke ⋅ cos θ Disaggregation and dissolution
t (calculation for the experiment
3.2.3= 0,2,
Disaggregation τ' = ≤1 (3.2.3.1.2)
e porosity is 20 vol%, i.e. from what the of MgO in
tortuosity slagis
factor t max
culated with Gl. (2.3.1.4) as = 0, 52. The average pore diameter is d =
-4
10 cm. Furthermore the wetting angle has been identified as = 34° /10/.
e surface tension is S =3.2.3.1 Non-saturated
570 mN/m
isslags
and the viscosity
the =running time t in relation to time tmax , needed for the dissolution of the
1, 7 g/cm s /10,
. The infiltration time t The
= 54sdisaggregation
is calculated byofthe biggest grain.
half cylinder The
radius time for the non stationary diffusion of a spherical particle is
refractory material inx = 0, 9
which coincides quite well with the observation. calculated approximately /14/ as
slags takes place in four phases: infiltration, Fig. 3.2.3.1: Temporal process of the experimentally measured and calculated
disaggregation
e tendency for infiltration of the
is one of the most grain bonding,
important breakfor thedegree '
characteristic of conversion
1 R 2 (left axis of ordinates) and of the reduced
Dissolution saturation
without disaggregation, i.e.
gement of refractory material. In addition, the tendency increases highly, if tmax (right
quotient 2
⋅axis of[ sordinate),
]. (3.2.3.1.3)
(C0 = 0) /10,12/
dissolution exclusively over the surface
down of the structure and the dissolution of 2,1 4⋅ D of the sample
graphite is less than 5wt% (cf. Fig. 5.1), because on the one hand the
osity and the tortuosity the left crystallites
factor increase rapidly and on inthe the slag.the The lower graph describes the disaggregation of the material,
other hand Ø
exclusively over the
osities are increasingly The
less reaction The process is reduced curved
to the surface area ofofthe
dissolution the used cylinder (40mm
crystallites in the x 18mm
melt, in ), thus
the case without
blocked by on the free surface is generally disaggregation /15/.
enclosures of graphite.
of a rapid disaggregation of the material. Then the following is valid (upper graph
not constitutive for the whole event.
in Fig. This ob-
3.2.3.1) /13/: In the given case the cylinder would dissolve entirely, before achieving
.3 Disaggregation of MgO in slagis completely proven by the exist- 3 saturation. Hence,
servation 1 the following
1 1is− valid
x  forofthethe course of the lower graph in Fig.
FFig.
φZ = ⋅ (3.2.3.1):
−τ '− ⋅ τ '3 +Temporal
ig.'⋅ 1ln23.2.3.1:
τTemporal τ '2 − ⋅ of
process
process
τ '−the experimentally

experimentally
(3.2.3.1.4) measured and calculated
measured and calculated degree of con-
ing experiments [12]. x degree of6 conversion 2 (left3 axis  of ordinates) and of the reduced saturation
.3.1 Non - saturated slags version (left axis of ordinates) and of the reduced saturation quotient
W. Gans [13] has already theoretically dealt
phases: the smallest to the Dbiggest01grain
quotient (right ϕ ϕ
The relationship from (right axis of τaxis
ordinate) = 2(C⋅of =ordinate),arctan (C0 =. 0)
0) ⋅[10,12] /10,12/ (3.2.3.1.6)
with these
e disaggregation of refractory incidents
material in slagsintakes
his place
professorial
in four dis- − ϕ of the material1− ϕ is
ltration, disaggregationsertation
of the grain bonding, breakand
“Disaggregation down of the structure
dissolution of and
dissolution of the left crystallites in the slag. The reaction on the free surface Thea max lower
2− a mingraph describes the disaggregation of the material, exclusively over the
x ≡= R(t) = m(t) ≤ =1 .C(t) − C0 ≡ γ is(3.2.3.1.5) (3.2.3.1.5)
enerally not constitutiveheaps
for theofwhole
debris“ in This
1973. The numerical ration.
of theThe resultradius
of theR,relationship from the
Ø the
event. observation is completelyϕcurved R 2a surface area of the used cylinder (40mm x 18mm ), thus without
max m C − C
the relationship running
model, developed here, for the disaggrega-
ven by the existing experiments/12/. S
disaggregation
S S
/15/.
0 starting radius R0 = 0,9 cm to the radius of
solute mass
Fig. (3.2.3.1) is calculated m and
by equ. the concentration
(3.2.3.4) via the data Cmentioned
to the particular
above. (theoretical) value of
tion of refractory material in slags is material avail- contains
Refractory
In
Fig.the
material
given caseof contains
the coarse
cylinder grains
would saturation
dissolve isR R=above.
entirely, = 1,3 cm and
before therefore
achieving
Gans /13/ has already theoretically dealt with theseRefractory
incidents in his saturation. coarse
(3.2.3.1)The grains
isresult
calculatedagainst
theby wear
relationship
equ. and fine
from
(3.2.3.4) via bonding
the starting
the grains
radius
data mentioned 0 S 0.9cm to
able for all users
andon Maple(bonding
[12]. Thematrix).
relative against
Hence, x often
saturation. weartends
and
Hence,fine
towardsbonding
the (xgrains
1 following 1),(bond-
which is additionally
valid for
2
R the course
fessorial dissertation “Disaggregation dissolution of heaps
simplifies equ.
of debris“ in
the radius of saturation is RS = 1,3cm and thereforeϕϕE E== 2 =
The diameter
0
0,48.. For
= 0,48 the relationship graph
of
For the the lower of the in Fig.
73. The numerical model, conversion
developed here, for the disaggregation of (3.2.3.1.4).
ing
refractory matrix).
(3.2.3.1): Hence, xofoften the biggest
tends towardsgrain shall1 be amax R=S
0,5cm, the diameter ofrelationship
terial in slags is available for all users on Maple /12/. The relative (xthe
conversion →smallest
1), which amindegree
of the = 0,005cm.
additionally of conversionThe measured
simplifies φ toeq. diffusion
the reduced
degreesaturation quotient
of conversion φ to the reduced satu-
V0 − V(t) coefficient of the MgO in the slag is D = 1,3 10-7cm2/s /12/. Consequently, tmax
φ= ≤1 (3.2.3.1.1)
(3.2.3.1.1) (3.2.3.1.4). The diameter of the biggest ration quotient γ the following is valid
V0 = 454min (equ. 3.2.3.1.3) and x = 0,99 (equ. 3.2.3.1.5). The ϕ upper left graph in
ϕ
grain shall be amax = 0,5 cm, τD the = 2diameter
⋅ φ/γ = ϕE = 0,48. and therefore
of⋅ arctan 5 (3.2.3.1.6)
with eq.
of the observed refractory
he observed refractory starting volume V during starting volume V0
the time t [s] is the smallest amin = 0,005 cm. The measured
calculated. the
1 −
following
ϕ is φ γ
1=
−
ϕ
ϕ
(3.2.3.1.6)E τDE = 1,47. The entire time of
valid / = 0,48 and therefore with equ. (3.2.3.1.6)
0

during the time t [s] is calculated. diffusion coefficient of the MgO in the slag is τ = 1,47. The entire time of
dissolution is dissolution is
3 DE
2
D = 1,3 R(t) · 10-7=cm m(t)
2/s [12]. C(t)Consequently,
− C0
t ϕ ≡ = ≡ γ is the relationship τ ⋅V ⋅δ of the running radius R, the
τ' =
t max
≤1 (3.2.3.1.2)
(3.2.3.1.2) tmax = 454 2
R S min (eq. mS3.2.3.1.3) CS −and C0x = 0,99 t a = DEF ⋅ DS [s]. (3.2.3.1.7) (3.2.3.1.7)
0
(eq.
solute 3.2.3.1.5).
massThe m upper
and the left graph in Fig. 12 C to the particular (theoretical) value of
concentration
is the
he running time t in relation torunning
τ ' non
time
t tmax
= stationary ≤1
time t
, needed in relation
for theto time
dissolution
(3.2.3.1.2)
tmax ,
of theis calculated
saturation. by
The eq. (3.2.3.1.4)
result of 350via
the g the
of dataare in 350
slag
relationship thefromg of slag
crucible. theThe are in the crucible.
density
starting of the slag
radius The
Ris0 density
= 2,72 g/cm.
=S 0.9cm to
ggest grain. The time for the t diffusion of a spherical particle is Thus, the slag volume is VS = 128,7 cm. The thickness of the diffusion layer is
needed
culated approximately /14/ as
for
max the dissolution of the biggest mentioned above. of-3 the slag is ρS = 2,72 R g/cm
2 3. Thus, the
estimated = 1 10and cmtherefore
/12/. The curved=surface 0
=area of the sample
the (R = 0,9
grain. The time for the non stationary diffu- the lower
The radiusgraph of saturation
describes the is RhS==2,51,3cm volume is Vϕ=E 128,7 30,48
. The .isthick-
For
s the running time t in relation to time cm,disaggrega- cm) is Fslag0 = 14,1 cm. The Sdiffusion
cm
R S2coefficient D = 1,3 10-7 cm/s
2 tmax , needed for the dissolution of the
biggest grain. The time sion
for of
1 nona⋅ spherical
Rstationaryparticle is calculated ap- tionisprocess
of theofmaterial, exclusively /10/. Theover5insertion
the ness
of of the
all data in equ. diffusion
(3.2.3.1.7)layer ta = 7 104
is estimated
provides DE =
tmax 'the [ s ] . t diffusion Fig. of a(3.2.3.1.3)
spherical
3.2.3.1: particle
Temporal relationship the of experimentally
the degree measured
of and calculated
conversion φ to the reduced saturation quotient
2,1 2
4 ⋅ D 1,03 10 s
calculated approximately proximately [14]
/14/ as τ ' = as ≤ 1 degree of conversion curved
(left axis
Fig.
(3.2.3.1.2) surface
3.2.3.1: Temporalarea ofand
of ordinates) theofused
process of the cylinder
experimentally
the reduced δ = 1 · 10 cm [12]. The curved surface area
saturation measured
28,6 h as the necessary time of dissolution.
-3 and calculated
t max degree
2
quotient (right axis (40 mmofx conversion
of ordinate), 18(C mm 0 = Ø), (left
thusaxis
0) /10,12/ of ordinates)
without disaggre-and ofofthethe reduced
sample saturation
(R = 0,9 cm, h = 2,5 cm) is
e process is reduced to the dissolution 1 R of the crystallites in the melt, in the quotient case (right axis of ordinate), (C0 = 0) /10,12/
is the running
a rapid disaggregation t time
ofmax ' t2,1
the material. ⋅
in2 relation Then[ s ]
to . time
the t
following , (3.2.3.1.3)
needed
is (3.2.3.1.3)
valid for the
(upper gation
dissolution
graph [15].
of the The middle graph in Fig.
F0 =over 14,1thecm42describes
3.2.3.1 the actualcoefficient
. The diffusion process of is the first
4⋅ D max The lower graph describes the disaggregation of the material, exclusively
biggest
/13/: grain. The time for the non stationarycurved diffusion of aareaspherical particle is describes experiment.
Ø The graph was calculated via Maple /12/, but it is possible to 5
Fig. 3.2.3.1) surface Inofthe
The the given
lower
usedgraph case
cylinder the(40mm
cylinder thex would dissolve
disaggregation
18mm of the
), thus without D = 1,3 · 10 cm²/s [10]. The insertion of all
material, -7
exclusively over the
approximate the graph via the following relationship, because the behaviour of
calculated
3  approximately 1 3 /14/ as1 1to− the
x  disaggregation /15/. curved
case surface area of the used cylinder (40mm x 18mmØ), thus without
The process φZis=reduced The
⋅  −τ '⋅to τ '−process
ln the ⋅ τ ' + isτ '2reduced
dissolution −of ⋅the τ '−crystallites dissolution
in(3.2.3.1.4)
the melt, of entirely,
in the
 In the given case thedisaggregation
before achieving saturation.
dissolution Hence,of the datascatter
broad in is especially
eq. (3.2.3.1.7) influenced by provides
the big crystals.
x  6 material. Then 2 the23following cylinder would /15/.
dissolve entirely, before achieving
of a rapid disaggregation theof crystallites
the in the 1 melt,
R in saturation.
the case is valid a(upper
ofHence, the
In
the
graph
thefollowing
given
following case
is isvalid
valid
the for
cylinder
for the the For
course
would
course
(
of the
ofdissolve
the

(1 - t ))=the
lowerentirely,
grapha in ·following
7 Fig.
before · τDE is
104achieving = valid
1,03 /13/:
· 105 s ≈ 28,6 h as
n Fig. 3.2.3.1) /13/: tmax ' 2 ⋅ [s] . (3.2.3.1.3)
e relationship from3 the rapid disaggregation
 smallest 1to the biggest 2,11 grain of
4⋅ D the
1− ofxmaterial.
(3.2.3.1):
the
 material Thenis lower graphHence,
saturation. in Fig.the 12:following is valid for the course the necessary
of the lower 2time of3 dissolution.
graph in Fig. 2
φZ = ⋅  −τ '⋅ ln τ '− ⋅ τ '3 + τ '2 − ⋅ τ '− (3.2.3.1.4) (3.2.3.1): 1 −15α + 24α − 10α + (60 − 90α + 36α )τ + 
x  the following
6 is valid
2 (upper
3   graph in φ Exp = 3 The
⋅  middle graph 2
in Fig.
3
12 describes the ac-  ,(3.2.3.1.8)
ϕ case ϕ 20(1 − α) + (45α − 60)τ + 20τ
The process is a maxreduced
x = Fig.
− a min to the dissolution of the crystallites in theτ melt,
12) [13]: ≤ 1. (3.2.3.1.5) D = 2⋅
in the ⋅ arctan .ϕ (3.2.3.1.6)ϕ tual process of the first experiment. The 
(3.2.3.1.6)
of a rapid disaggregation
a of the material. Then the following is valid (upper 1 − graph
ϕ τ = 12− ⋅ ϕ ⋅ arctan . (3.2.3.1.6)
The relationship from the smallest to the biggest grain of the material is
1 − ϕ graph was calculated via tMaple [12], but it
max D
in Fig. 3.2.3.1) /13/: 1− ϕ t
3− a grains against 1 3 wear2 and 1 fine 1bonding
−2 x m(t)grains whereas the times τis ≡possible toand α ≡ athe graph via the (3.2.3.1.9)
approximate
efractory material contains a coarse R(t) C(t) − C t + t tz + ta
φ =
Z max ⋅ −τ '⋅ ln τ '− ⋅ τ '
1 . 6 1 (x 1),ϕ2which
 min ≤towards + τ ' − ≡ (3.2.3.1.5)
⋅ τ '−
3 = m = C −R(t)
(3.2.3.1.4) 2 ≡γ
0
is the relationship of the running radius R, z the a
onding matrix). Hence,x x=often axmax
tends R S2additionally m(t) C(t) − C0 following relationship, because the behav-
S ϕ ≡S C20 = = ≡ γ is the relationship of the running radius R, the
mplifies equ. (3.2.3.1.4). The diameter of the biggest grain solute shall be m amax = the concentration
RS mS C toCthe −C are defined. is the reduced running time andbroad thescatter
reducedis dissolution time of
cm, the diameter of the smallest
mass
(3.2.3.1.4) and S particular
0 (theoretical) value
iour of
of dissolution of the es-
The relationship from athe = 0,005cm.
min smallest to the Thebiggestmeasured graindiffusion
saturation. of
Thetheresult
material
solute
of the is
mass m and thefrom
relationship concentration the (biggest
the starting Cradius
to the cubic)
R 0 = crystal
particular (amax). tz is value
0.9cm(theoretical)
to the dissolution
of time of the complete cubic
Refractory material contains coarse grains against -7 2wear and fine bonding grains
is the relationship of the running radius R, pecially influenced by the big crystals.
efficient of the MgO in the slag is D = 1,3 10 cm /s /12/. Consequently, tsaturation. The result of the relationship sample,Rwhich 2
from has been observed
the starting radius R (A ) 0.9cm
(here 16min) and ta the dissolution time
0 0= to
max
(bonding matrix). Hence, x often tends towards 1 (x 1), which additionally 0
454min (equ. 3.2.3.1.3) The and xrelationship
= 0,99 a(equ. −from3.2.3.1.5).
a min the smallest the upper
The radius ofthe
toleft saturation
graphthe in is RS = mass 1,3cmmand andtherefore ofϕthe = 0,48
C . inFor
= crystallites thethe
simplifies equ. (3.2.3.1.4). The diameter
x = max of the≤biggest1. grain shall = solute
be amax(3.2.3.1.5) the concentration
E
R 2 For (α
slag τ ≤ (1
R≤02(here - α)) the
554min). Forfollowing 1, is istz valid
0, i.e. the process
the radius of saturation is R = 1,3cm andS therefore ϕ = = 0,48 . For the
0,5cm, the diameter of the biggest
smallest grain aminof=athe materialThe
0,005cm.
max isrelationship
measured diffusion to the particular (theoretical)
S is onlyof
value
of the degree of conversion φ to the reduced saturation quotient R S
intended
satu- for[13]:
dissolution:
E 2 equ. (3.2.3.1.4). (1 – ) is the reduced
coefficient of the MgO in the slag is D = 1,3 10-7cm2/s /12/. Consequently,relationship
tmax disaggregation time. Therefore the total process is ( + (1- ) = 1). (Symmetry
of the degree of conversion φ to the reduced saturation quotient
Refractory
= 454min (equ. material
3.2.3.1.3) and xcontains coarse
= 0,99 (equ. grains against
3.2.3.1.5). wear left
The upper andgraph
fine bonding
in grains is cognizable: A result for
0,5 is identical to one for (1 - ), if is  0,5.
(bonding matrix). Hence, x often tends towards 1 (x 1), which additionally Example: = 0,4 as a result is identical to (1 – ) for = 0,6.)
refractories
simplifies equ. WORLDFORUM
(3.2.3.1.4). The diameter 5 (2013)
of the biggest grain shall be[2]
amax = 5 87
0,5cm, the diameter of the smallest amin = 0,005cm. The measured diffusion 5 The whole process lasts
tZ = 16min and ta = 554min in the first experiment.
coefficient of the MgO in the slag is D = 1,3 10-7cm2/s /12/. Consequently, tmax 570min. Consequently = 0,966. Given that > 0,5, one calculates with
= 454min (equ. 3.2.3.1.3) and x = 0,99 (equ. 3.2.3.1.5). The upper left graph in (1 – ) = 0,034. For the current reduced time the following is valid: = t/tges =
t/570. For t = 285min is = 0,5 and consequently to equ.(3.2.3.1.8) φExp = 0,88,
 K / = (C )⋅ PCO . (3.2.4.3)

e solute carbon content is dependent on temperature, the slag

on and the CO – partial pressure: F = 3. MgO – C Part II b
.1 shows that up to (C) = 1wt% are soluble at a temperature of
nd PCO = 0,1 atm. The solubility diminishes with an increasing CO –
essure and thereby the risk of the dissolution of the binder phase.
to which extend an analyzed refractory ma-Saturated
3.2.3.3 to Ostwald-ripening
slag [17]. The fine grained
terial in an experiment tends to dissolve. The binder phase dissolves, by what bigger grains
By considering the second experiment, where the assay abruptly breaks down
observation approves the empiricalafterknowl- are even
90 minutes, growing. Thethe
though known
slag iscorrelation
saturated, is valid for explanation than
a different
edge whereupon a dense structuredissolution
with a needsthe to be found. In this
chronological case theofstructure
sequence coarsens due to Ostwald
the changing
– ripening /17/. The fine grained binder phase dissolves, by what bigger grains
chemically strong binder phase is anare
eighth medium radius of the grain [17]:
growing. The known correlation is valid for the chronological sequence of
protective mechanism. The MgO the – Cchanging
ma- medium radius of the grain /17/:
1/3
terial strongly belongs to this. 6 ⋅ σMgO−Slag ⋅ DSlag ⋅ VM2 ⋅ cSat MgO
thethe following
following is valid φ / φγ /=γϕ=E ϕ=E0,48
is valid = 0,48
andand therefore
therefore with
with equ.
equ. (3.2.3.1.6)
(3.2.3.1.6) r= − ⋅ t [cm] . (3.2.3.3.1)
The theoretical consideration, which is mentioned here highly abbreviated, R ⋅T
= 1,47.
τDE τ=DE1,47. TheThe entire
entire timetime of dissolution
of dissolution is is 3.2.3.2 Disaggregation time
allows to say to which extend an analyzed refractory material in an experiment
the following is valid φ / γ = ϕE = 0,48 and therefore with equ. (3.2.3.1.6)
-3
τ = 1,47. The entire time of dissolution tends to dissolve. The observation
The disaggregation time approves
can be just the
Forempirical knowledge
the available
roughly whereupon
experiment we receive r = 5, 9 10 t1/3 [cm], with the
(3.2.3.3.1)
DE
τDEτDE ⋅ VisS⋅ V
⋅ δS ⋅ δ
t a =t a = [s].[s]. a dense structure with a(3.2.3.1.7)
chemically strong binder
(3.2.3.1.7) phase data
following is an eighth = 430mN/m, DSlag = 1,3 10-7cm2/s, VM,MgO= 12cm3,
MgO-Slagprotective
τ ⋅ V ⋅ δ F0 ⋅FD
t a = DE S [s].
0 ⋅D mechanism. estimated,
(3.2.3.1.7) The MgO because the statistically
– C material collected
strongly belongs = 0.065mol/cm3, R = 8,31J/K mol and T = 1873K. A grain of
to this.
csat= 9,6wt%
F0 ⋅ D
characteristics MgO
of the structure 2mm– C Part
are not II
diameterForb the
may available
dissolve in one experiment we receive
hour. The saturation of the slag, pursued in
350350g ofg slag
of slagareare in the
in the crucible.
crucible. TheThe density
density of the
of the slagslag
is isS =S2,72
= 2,72 g/cm.
g/cm.
enough for anlayer
accurate the practical experience,
r = 5,9 · 10is-3the tenth protective following but like all the rest
mechanism,
350 g of
Thus,
Thus, slag
thethe are in
slagslagthe
volume crucible.
volume The density
is VisS V = 128,7
= S128,7 of the slag
cm. is
cm.
TheThe
S = 2,72 g/cm.
thickness
thickness of of the
the diffusion
diffusion is is calculation. The slag
layer · t1/3 [cm], with the
Thus, the slag volume is VS-3= 128,7
estimated -3 cm. The thickness of the diffusion layer is it is just of limited impact.
estimated
estimated = 1 =101-3=cm110/12/.
10cmThecm
/12/./12/.
curved The The
surface curved
curved
area of the surface
surface (R =area
area
3.2.3.2
sample of
0,9 of the
the sample
sample
Disaggregation
enters the (R (R = 0,9 in about 1 min.
= 0,9
time
porosities The data σ = 430 mN/m,
-7
= 1,310-710cm/s MgO-Slag
cm,cm,
cm, h2,5
hh==2,5 =cm)2,5
is Fcm)
cm) 0 is Fis0 cm.
= 14,1 F014,1
= =The
14,1
cm. cm.
The
diffusion The diffusion
diffusion
coefficient 1,3 coefficient
is D =coefficient is =D1,3
10-7 cm/sis D cm/s
DE = temporary law DEis valid
=be for the disaggregation DSlagthe
=to 1,3 · 10the
4 The result suggests choose -7 cm 2/s, V of the=binder
crystallites 12 cmphase
3, as coarse as
ta =ta7= 710410
4
/10/.
/10/.
/10/. The
The The insertion
insertion of all data
insertion of
of all all
in equ.
datadata in equ.
in equ.
(3.2.3.1.7) (3.2.3.1.7)
provides ta = 7 10
(3.2.3.1.7) provides
provides
The disaggregation time =
can
DE just roughly estimated, because statistically M, MgO
1,03
1,03
5
1010s 510 5 The theoretical consideration, which is mentioned 3.2.3.3 here highly abbreviated,
Saturated slagbecause then the Ostwald – ripening decelerates. The
possible, latter is easier to
1,03 s s ofcharacteristics
theanbinder phasethe into the infiltratedanslag c sataccurate
= 9,6 mass-% = 0,065 mol/cm , 3
 28,6 h as the necessary time of dissolution.
 28,6
 28,6 h ash theas the necessary
necessary time allows
time to say to collected
of dissolution.
of dissolution. which extend analyzed of refractory
structure are notinenough
material achieve
for an
experiment
with graphitic binding than with ceramic binding and equals the ninth
[16]: calculation. The slag enters the porosities in about one minute. The temporary
R= 8,31 J/K · where
mol and
.1: Solubility of experiment.
graphite
The middle graph in in basic
Fig. 3.2.3.1
tends
slag
describes
to dissolve.
atstructure
the actual 1600°C
process
The observation
of the first /18/
approves the empirical knowledge
By considering the whereupon
second
protective experiment,
mechanism. the Tassay
= 1873 K. A breaks
abruptly grain down
Fig. 13 Solubility
TheThe ofThegraphite
middle
middle graphgraph
graph was ininFig.
basic
calculated
in Fig. a dense
3.2.3.1
via
3.2.3.1 slag describes
/12/,
describes
Maple law
is the
but itthe withistoprocess
actual
possible
actual valid for
process of the
a chemically disaggregation
strong
of the
the first of theisbinder
first binder phase an phaseprotective
eighth into the infiltrated slag
own data.approximate the graph via the following relationship, because The [16]:
the behaviour
MgO of
– but
Cit material strongly
after 90 minutes, of
even though 2 mm diameter
the slag may
is saturated,dissolve in
a different 1 h. The than
explanation
at 1600°C experiment.
experiment.
[18] ( ofown TheThegraph
data) graph waswas mechanism.
calculated
calculated viaviaMaple Maple /12/,
/12/, but isitpossible to to belongs
is possible to this. needs to be found. In this case the structure coarsens due to Ostwald
dissolution
dissolution
Forapproximate
approximate
the broad
(

(1 -the)) the
scatter is especially influenced by the big crystals.
thegraph
graph viavia
the the following
following relationship,
relationship, because
because
The theoretical
2
dthe
the behaviour of c (tis) −
 8of which
behaviour
consideration, cs  /17/.
mentioned saturation of the slag, pursued in the practi-
following is valid /13/:
t = ⋅ ln − ln – ripening 3.2.4here
The highly
Solubility abbreviated,
fine grained binder
of carbon phase dissolves, by what bigger grains
in slag
(3.2.3.2.1)
dissolution
dissolution of the
of the broad
broad scatter
scatter is especially
is especially influenced
influenced
allows Zby
tobysay the
big2big
theto 
crystals.
crystals.  The known calin experience, is for
thethetenth protective
ForFor
( (
1

(1
π which
⋅ D extend
)τ +  to dissolve. The 
π 2 an analyzed
are cs refractory
c0 −growing.
 material an experiment
correlation is valid chronological sequence of
−-(1
15α-+))24the
))
α 2 the10αfollowing is+ valid
2 /13/:
−following
3
+ (60 −is valid
90α 36α/13/:
tends observation approves the empirical knowledge whereupon
φExp = 3
⋅ 2
3.2.3.2
3
Disaggregation time
 ,(3.2.3.1.8) the changingBesides
medium radius
MgO, of the
carbon
mechanism, grainbut
/17/:
is soluble inlike
lime
all–the
aluminate slags.
rest it is just The
of chemical reaction
20(1 − α) +(45α − 60)τ + 20τ
r solubility of graphite in 1slag is15the
−α
eleventh
242 α−210
a protective
103 α+3(60
dense

D(60
−=90 1,3
structure
90+α10 36 cm 2mechanism.
-7 2 2with
)/s
a chemically strong binder
τ+is the diffusion coefficient
phaseslag
inofbasic is an
the (MgO)
(3.2.3.2.1)
is eighth
/18/: protective
in the⋅impact.
limited slag, c⋅0V
,Mc2S ⋅ cSat
1/3
1 −15 +α24+α −α +mechanism.
− α 36+α Theα
)τ +
MgO –,(3.2.3.1.8)
C material strongly belongs to6 ⋅this.
σMgO 2 −−Slag
D2Slag
φExpφExp
= = 3 
⋅ 3t ⋅ 
− ατ−)≡α)+(45
the 20(1
whereas 20(1times (45
+and
The tdisaggregation
≡− 60)
τa2 +
α −ααcollected
60) τ220 3 3
(3.2.3.1.8)
τ20
and
+characteristics
τ
time
c(t) can
are
(3.2.3.1.9)
be
the
of during
just roughly
,(3.2.3.1.8)

concentrations
the structure

 saturation
estimated,
of the
are not enough
because
solute the
r 3
=
components, C
− +
accurate thatThe
for anEstimating
( )result
statistically
O
e.g. =
MgO, (R2 ⋅suggests
Cat−
)
the+
T of
{CO } ⋅ t [cm] . (3.2.3.3.1)
to choose the crystallites
(3.2.4.1)
MgO

t z + ta tz + ta beginning, the and in time. the moment

ction of graphite with slag calculation. Thedissolution
slag enters The activity of the pure graphite is a = 1 and the equilibrium constant
D =tthe
Z 1,3
porosities
· 10as-7 soon
occurs cmin2/sabout
asisthe one
the minute.ofThe
diffusion
saturation thetemporary
coeffi- of the
infiltrated slagbinder phaseupas<C>coarse
is reached as possible, be-
slag we receive raC=(2)
-3 1/3
are defined. is the reduced running t time t andis valid the reduced
a t adisaggregation
fordissolution
tthe time of For the intoavailable experiment 2 −5, 9 10 t [cm], with the
whereas the times τ(a≡ τ tz is thelaw
max).≡ and α ≡α of ≡the3.2.3.2 Disaggregation
cient of the
of the time
(MgO)
binder phase
in thefollowing
slag,
the infiltrated
d2
is tZ c=0,0,71
cdata
S and⋅ 2 c(t) cause thenKthe Ostwald-ripening decelerates.
whereas
whereas
the (biggest the
cubic)the
times times
crystal and andtime
dissolution complete cubic (3.2.3.1.9)
(3.2.3.1.9)
sample, which has been observed +[16]:
t z (A+t0z)ta(here
ta 16min) and ta the +tto
zt a+60%,
t z dissolution t a time the result of equ.(3.2.3.2.1) .
MgO-Slag = 430mN/m, =DSlag = 1,3⋅ PCO10-7cm2/s, 12cm3,
VM,MgO=(3.2.4.2)
ch contain (FeOofn)theand (MnO ) form carbon 1, is tz monoxide when a(O 2−mol
t ) −they cs beget in
2 the process are8 the concentrations ofcsatthe solute
= 9,6wt% compo- ⋅D
=π0.065mol/cm The3the,latter is easier
R statistically
= 8,31J/K to
) achieve
and T =with graphitic
1873K. A grain of
crystallites in the slag n(here 554min). For
is only intended for dissolution: equ. (3.2.3.1.4). (1 –t ) is=the reduced ⋅ ln
0, i.e.
The
d 
disaggregation
ln
(
ctime
−dissolution
can just roughly estimated, because
reduced  the 2mm are diameter may (3.2.3.2.1)
dissolve in onethanofhour. The saturation
ceramicis of the slag, pursued in
2 − = γ C ⋅ (C )
process is ( (3.2.3.1.9) nents, e.g. cMgO, at ofbeginning, during tZ the binding with binding
ith the graphite are
of arethedefined.
defined. refractoryis theis the
reducedreduced material running
running time time
(Fig.
and and the
3.2.5.1).
the reduced dissolution
For instance
time time
ofstructure The activity
for coefficient the solute carbon constant aCand
disaggregation
the (biggest
time. Therefore the total
cubic)
+Z (1- π) collected
=2The
1).
⋅D (Symmetry
result characteristics
πofof2the
the middle0 − cgrain
of the size d
practical
the 400 8m not enough
is
880s.
experience,
an accurate
This
is the are about
tenth 15
protective mechanism, but like (2) all the rest
the (biggest
is cognizable: Acubic)
result
crystal
forcrystal max(a).
0,5 is(aidentical maxtz).
toistone
ztheisforthe
(1 -dissolution
dissolution
), if is time
calculation.  0,5.time Theof the complete
complete
slag enters s cubic
cubic the porosities in about one minute. The temporary
on of the (FeO) sample,
is /2/:which
sample,
Example: = which
0,4 ashas has
been
a result been
is observed
observed
identical to (1 –(A0(A ))for0) (here
(here 16min)
= 0,6.) minutes
16min) andand saturation
in consideration
ta the
ta the and time
dissolution
dissolution in time
of time. Estimating
the time of
it is just of limited that theimpact.equals the ninth protective mechanism.
infiltration. 16 minutes are observed.
-7 law 2 is valid for the disaggregation of the binder phase into the infiltrated slag
ofare of defined.
the the crystallites
crystallites
0and τ inis the
inthe
in the
slag reduced
slag D (here
(here = 1,3 running
554min).
The10
554min). wholecm
For time
For /s is laststhe
1, is moment
1, tis 0,ofi.e.
z tz 0, i.e.
diffusion thedissolution
the
coefficient
process process oftthe
Z occurs as soon as
(MgO) in the slag, c0, cS 1
C = Fe + {CO}; ∆G (1600∞C ) = −131kJ / FC
tZ = 16min
is only
is570min.
only
ta = 554min
intended
intended forfor
the first
dissolution:
=dissolution:
experiment.
Givenand equ. equ.
c(t) (3.2.3.1.4).
(3.2.3.1.4).
are the
[16]:
process
(1(3.2.5.1)
Ostwald –(1 –) –is)ripening
concentrations is the
the theaffects
of2reduced
reduced solutethe structural
components, change if the times
e.g. suggests
MgO, are longer.
3.2.4 Then of
theSolubility of carbon inphase
slag as coarse as
and )α
Consequently
the For
(1 –disaggregation
disaggregation
= 0,034. reduced
thetime.
time.
0,966.
dissolution
Therefore
Therefore
current
that
thethe
reducedbeginning,
time
> 0,5,
total
the time
total
one calculates
of
process
process
following the
assuming
is valid: isthe
with
the
(is=(saturation
+ +=a(1-
t/tges(1- )saturation
=d)1).
saturation= 1). (Symmetry
oftime.
(Symmetry 99%8of Estimating
the
wecget
The result
(infiltrated
t )from
− cthat slag atisto
s equ.(3.2.3.2.1)
thechoose crystallites the binder
during t and in ln ln possible, the moment
because then of Ostwald
the – ripening decelerates. The latter is easier to
ibrium constantis(biggest
iscognizable:
t/570. cubic) crystal (amax
is cognizable:
For t = 285min A A
is result
result= 0,5forandfor ). tz is thetZ occurs
consequently



0,5 0,5
is
dissolution is
to identicaldisso- as soon
equ.(3.2.3.1.8)
identical to to
one φ one =
for for
0,88,
(1
=
Z -(1 -),2 if
reached π ⋅ Dup
as the),
⋅
if 
d toπ 60 of
is 2 
saturation is 
0,5. 2
−
0,5.%,the cthe − csresult
 of eq. Besides
(3.2.3.2.1)
MgO,withcarbon is soluble
binding in lime – alu-
0 infiltrated withslag is reached up than
Exp

Example:
which
Example: acceptably=0 = 0,4as aaswith
corresponds
0,4 a result
result is identical
Fig.is3.2.3.1. (The calculation
identical to (1to –(1with – ) Maple
for ) forist=Z 0,6.)= 4, 4⋅ 2
0,6.) = 5480 sachieve = 91min .
graphitic binding ceramic and equals the ninth
lution
suggested.)time of the complete sample, which
∆G π ⋅ D d2
protective mechanism. minate slags. The chemical reaction in basic
P /a = exp(− ) = 4521[atm] to 60%, the result of equ.(3.2.3.2.1) (3.2.3.2.1)
-7 2 isexperimentally.
t = 0,71 ⋅ 2
theZlasts
..
{CO} (FeO ) =tZ16min
tZhas =been
16min and and
observedta =ta554min
= 554min
(herein 16 in the
the firstfirst
min) experiment.
experiment.
and taDThis the= 1,3Thehas The bewhole
10
whole (3.2.52)
cm process
/s is lasts
process
observed diffusion π coefficient
⋅D of the (MgO)slag in the is slag,
[18]: c0, cS
570min.
570min. Consequently R ⋅T
Consequently = 0,966. = 0,966. Given Given thatthat and > 0,5,> 0,5,
c(t) areone
one 6calculates
thecalculates with
concentrations with of the solute components, e.g. MgO, at the
(1dissolution
–(1 –) =) 0,034. timeForofFor
= 0,034. thethe the crystallites
current current
Thereduced reduced
result
inof the
time time
the
slag
the the
beginning,
The
middle use The
following
following of during
the
grain
result
is valid:
is size
valid:the
graphite of
=400
d
saturationt/t=ges
bond the
t/t
8m=enables
ges
is
middle
=and
tZ3.2.4

inustime.
880s.
grain
Solubility
to use
This
sizecarbon
areofabout
Estimating
a coarser that15the
grain, inwhich
slag
moment of the
delays
t/570.t/570.
ity of an iron-II-oxide,
(here For For
554 t285min
=min). = 285min
t dissolved For is is α=→ in = 1,
0,5 0,5
minutes isand
aandtypical tin
z→
consequently
consequently LDAC
0,dissolution
dissolution
consideration
to equ.(3.2.3.1.8)
dof–
i.e.to equ.(3.2.3.1.8) ≤tand slag
400
occurs
Zthe µmisφas
illustrates
time of isφsoon
Exp aExp
= t0,88, 0,88,
aZ=ninth
≤as
infiltration.
(FeO) 880 s.16 This

protective
the saturation
minutes are
of about
the
mechanism.
are infiltrated
observed. 3 〈Cslag〉 +is(Oreached
-2) = (Cup 2-
2
) + {CO } (3.2.4.1)
which which acceptably
acceptably corresponds
corresponds with withFig.Fig. 3.2.3.1.
3.2.3.1. (The (The calculation
calculation with withMaple Maple is is Besides MgO, carbon 2 is soluble in lime – aluminate slags. The chemical reaction
.e. the processing the process
pressure
suggested.)
is only
is P intended
{CO}
for
2260atm. dissolution: to 60%, the15 min.
result in
of consideration
equ.(3.2.3.2.1) of is the
t = time
0,71 of
⋅
d infil-
.
suggested.) Ostwald – ripening affects the structural change if in thebasic Z
times slag
areisπlonger.
2/18/: Then
⋅D
eq. (3.2.3.1.4). (1 – α) is the reduced disag- tration. 16 min. are observed. 2 −The activity of the pure graphite is a =1
3C + O
assuming a saturation of 99% we get from equ.(3.2.3.2.1)
6 6 2 the structural
( )= (C )+ {CO} 2− 〈C〉
(3.2.4.1)

(α + (1 – α) = 1). Symmetry is cognizable:π ⋅ Dchange if the times are longer.

2
gregation time. Therefore the total process d 2 Ostwald
Theisresult -ripening
of the middle grain affects
size d
400 8m isoftZthe

880s. and the
This equilibrium
are isabout constant
a<C>=151 and the equilibrium constant
. timeThe activity pure graphite

A result for α ≤ 0,5 is iden

tZ = 4,minutes
4 ⋅ 2 in consideration
= 5480 s = 91min of the Then assum-
of infiltration. 16 minutes areaobserved.
2−

(1 – α), α
C
Thistical to observed
has be one for experimentally.
ing a saturation of 99 % one gets from eq. K = (2) ⋅ PCO (3.2.4.2)
(3.2.4.2)
Ostwald – ripening affects the structural change if the times are longer. a(O 2− ) Then

Example: α = 0,4 as a result

if is 0,5.
≥ assuming (3.2.3.2.1)
a saturation of 99% we get from equ.(3.2.3.2.1) 7
The use of the graphite bond enables us to use a coarser grain, which delays the
2 − = γ C ⋅ (C )

(1 – α) for α = 0,6.
is identical to d2
The activity coefficient ofThe
theactivity
solute carbon of the asolute
is constant
coefficient carbon
is constant aC = γ C · (C)
dissolution and illustrates a ninth protective mechanism. C(2)
tZ = 4, 4 ⋅ 2 = 5480 s = 91min .
π ⋅D 2-
(2)
tZ = 16 min and ta = 554 min in the This first has be observed experimentally. and the activity of the ions of dissolved oxy- 1
experiment, the whole process lasts The use ofThisthehas beenbond
graphite observed experimentally.
enables genwhich
us to use a coarser grain, as well. Both
delays the are drawn into the con-
570 min. Consequently α = 0,966. Given dissolution Theanduse of the graphite
illustrates bond enables
a ninth protective us to
mechanism. stant, therefore the equilibrium constant is
that α > 0,5, one calculates with use a coarser grain, which delays the disso-
α → (1 – α) = 0,034. For the current lution and illustrates a ninth protective Kl = 7(C ) · PCO (3.2.4.3)
reduced time the following is valid: mechanism.
τ = t/tges = t/570. For t = 285 min is τ = 0,5 Hence, the solute carbon content is depend-
and consequently to eq. (3.2.3.1.8) 3.2.3.3 Saturated slag ent on temperature, the slag composition
φEXP= 0,88, which acceptably corresponds By considering the second experiment, where and the CO-partial7pressure: F = 3.
with Fig. 12 (the calculation with Maple is the assay abruptly breaks down after 90 min., Fig. 13 shows that up to (C) = 1 mass-% are
suggested.) even though the slag is saturated, a different soluble at a temperature of 1600 °C and
The theoretical consideration, which is men- explanation than dissolution needs to be PCO = 0,1 atm. The solubility diminishes with
tioned here highly abbreviated, allows to say found. In this case the structure coarsens due an increasing CO-partial pressure and there-

88 refractories WORLDFORUM 5 (2013) [2]

composition and the CO – partial pressure: F = 3.
Fig. 3.2.4.1 shows that up to (C) = 1wt% are soluble at a temperature of
1600°C and PCO = 0,1 atm. The solubility diminishes with an increasing CO –
partial pressure and thereby the risk of the dissolution of the binder phase.

by the risk of the dissolution of the binder

phase.
The lesser solubility of graphite in slag is the
eleventh protective mechanism.

3.2.5 Reaction of graphite with slag

Fig. 3.2.4.1: Solubility of graphite in basic slag at 1600°C /18/
Slags which contain (FeOn) and (MnOn) form
own data.
carbon monoxide when they get in contact
with the graphite of the refractory material
The lesser solubility of graphite in slag is the eleventh protective mechanism.
(Fig. 14). For instance the reaction of the
(FeO)Reaction
3.2.5 is [2]: of graphite with slag

Slags
(FeO )which
+ 〈C 〉contain
= Fe + (FeO
{COn)};and
∆G(MnO n) form carbon monoxide when they get in
0 (1600 °C) =
contact with the graphite of the refractory material (Fig. 3.2.5.1). For instance
–131
the kJ/mol
reaction of the (FeO) is /2/: (3.2.5.1)
The
(FeOequilibrium
) + C = Fe constant
+ {CO}; ∆G is0 (1600∞C ) = −131kJ / FC (3.2.5.1)
The equilibrium constant is
∆G 0
K{CO} = P{CO} / a(FeO ) = exp(− ) = 4521[atm] (3.2.52)
R ⋅T
The activity of an iron-II-oxide, dissolved(3.2.52)
in a typical LDAC – slag is a (FeO)

0,5 /19/, i.e. the processing pressure is P{CO}
2260atm.

The activity of an iron-II-oxide, dissolvedTeil

in a III a 2
typical LDAC-slag is a(FeO) ≈ 0,5 [19], i.e. the
processing pressure is P{CO} ≈ 2260 atm.
The reaction with (MnO) leads to
PCO ≈ 10 atm. The infiltration of the slag 4.isWear of MgO – C materials through oxidation
highly disturbed, which corresponds addi-
tionally with the twelfth protecting 4.1 Experiments
mechanism and is comparable to the
fourth in iron, which is less pronounced.
Moreover, the solidification temperature Decarbonisat
of ion of carbon bonded refractory material during the preheating of
the slag is raised by several hundred de- ladles, via the ladle fire up to'about 1100°C in air, is a serious operational
grees, because of the reduction of (FeO), problem,
or because the surface layer decarbonises and consequently corrodes.
rather (MnO), whereby the infiltration Laboratory
is tests have been made for the quantitatively description of the process
highly disturbed. The porosities become at the DIFK/FGF in Bonn /20,21/. For this, amongst others, cylinders made of
blocked [19], which is the thirteenth industrially
pro- Fig. 14 produced
Reactions of MgO – C materials
MgO (green), (36FeO
MnO (blue) and mmØ x 100
(red) with mm)atwere
graphite oxidated on
different
tecting mechanism. air, during an increased
temperatures temperature
and CO-partial and
pressures in their loss of weight
a Richardson-Ellingham-diagram [1]:
Fig. 3.2.5.1: was continuously
Reactions of MgO (green), MnO (blue) and FeO (red) with graphite
Continuing, Fig. 14 indicates that 〈MnO〉 is
registered.
avoid Additionally,
low pressure (vacuum)
3 wt.%
reduced to metal by graphite at > 1400 °C at different temperatures and metallic
CO – partialaluminium
pressuresandin a/or boron carbide
Richardson – Ellingham
(B4C) were used as antioxidants. Fig.
diagram /1/: Avoid low Pressure (vacuum). 4.1. 1 shows the results by three examples.
and PCO ≈ 1 atm, whereas the partial pres-
sure of oxygen is PO2 ≈ 10-17 atm. Again, the
CO – partial pressure is equal to the equilib-
rium constant KP. But for the reduction of
MgO is PCO = √ KP because CO and Mg-gas-
es are produced in equal amounts.
temperature [°C]

4
weight loss [%]

4 Wear of MgO – C materials

through oxidation

4.1 Experiments
Decarbonisation of carbon bonded refracto-
ry material during the preheating of ladles,
via the ladle fire up to about 1100 °C in air,
is a serious operational problem, because time [h]
the surface layer decarbonises and conse-
Fig. 4.1.1: Decarbonisation of three MgO – C materials /20/.
quently corrodes. Laboratory tests have been Fig. 15 Decarbonisation of three MgO – C materials [20]

The samples were heated up by 5K/min. A significant reaction starts above

refractories WORLDFORUM600°C. Discontinuiti
5 (2013) [2] es 89
during the progress signalise the influence of the
antioxidants. The latter oxidate and cause an increase of weight, which evidently
delays the combustion reaction.
 2 C + {O
4.2 Evaluation 2 }discussion
and = 2 {CO}. (4.2.1)

The oxidation
The partial pressure of the reaction
oxygenofisthe graphite
fixed (F =is2), if the temperature and the CO
– partial pressure is given. The value of PO2= 10-18 atm at 1100°C and PCO = 1
2 C + {O2 }Ellingham
atm can be learned from the Richardson- = 2 {CO}. Diagram /1/, Fig.(4.2.1)
3.2.5.1.

so called
The4.2 The partial
pore
Evaluation andpressure
diffusion of the oxygen
coefficient
discussion is fixed (F = 2),for
is codetermined if the
thetemperature and the CO
chronological
– partial
procedure of the pressure is given. The value of PO2= 10-18 atm at 1100°C and PCO = 1
corrosion.
atm can be learned from the Richardson- Ellingham Diagram /1/, Fig. 3.2.5.1.
The oxidation reaction of the graphite is
D p ∫ D ⋅ λ ⋅ε [cm2/h]. (4.2.2)
The so called pore diffusion coefficient is codetermined for the chronological
With on
Tab. 1 Influence of antioxidants the the
porosity
procedure
de- andof2 the + {O2 } = factor
C tortuosity
the corrosion. 2 {CO}.. In a lot of cases is(4.2.1)
approximately valid /20, The oxidation
22/ of carbon during the preheat- sponds quite well with the practical results.
carbonisation (compare Fig. 15) [9] ing ofofxthea lined
The partial pressure D p ∫ DMgO
oxygen ⋅ λis⋅εfixed– C2/h].
[cm ladle
(F = inifthe
2), the steel
temperature It has been experimentally observed that the
(4.2.2)
λ = ' 1,84 ⋅ ε − 0,3[−] . (4.2.3) and the CO
4C
-18
MgO-C MgO-C-Al MgO-C-Al-B With theplant
– partial pressure l is The
isporosity
given. analysed
and value ofasPO2an
the tortuosity = 10example.
factor In aat
.atm The
lot1100°C kindPof
and
of cases is CO bonding
=1 phase, resin or carboplast,
atm can beapproximately
learned from valid
the /20, 22/
Richardson-
diffusion coefficient of oxygen in the Ellingham Diagram /1/, Fig. 3.2.5.1.
has no significant influence on the decar-
t0 [h] 14 19 The oxidation
19,5 of carbon during the preheating x
λ = 'is
of a lined MgO – C ladle in the
1,84 ⋅ ε − 0,3[−] . (4.2.3)
steelThe
plant
so is analysed
called pore O2as – anN2example.
diffusion– CO-gas l The[20]:
coefficient diffusion coefficient
is codetermined for of bonisation
theoxygen in the rate [23].
chronological
m0 [%] 15,5 12,7 O –N –CO
210,5
2 procedure – gas is /20/:
of the corrosion. The model offers the opportunity of an easy
The oxidation
D = 2,85 of carbon
⋅ 10−2during
⋅ TK1,79 the
[cm preheating
2
/ h] . of a lined MgO(4.2.4)
(4.2.4) – Candladlevery
in thedetailed examination of experi-
steel plant is 2
D analysed
D ⋅ λ ⋅asε an[cm
p ∫.4),
example.
/h]. The diffusion coefficient(4.2.2) of oxygen in the
Summing up equ.(4.2.2 to
O2–N2–CO – gas is /20/: we receive the pore diffusion coefficient ments for decarbonisation. For this, the
of:
With the porosity and the tortuosity factor . In a lot of cases is
made for the quantitatively description of Summing up D =eq. ⋅ 10−2 ⋅ Tto
2,85(4.2.2 1,79.4),
[cm2on / h]receives
. measured
(4.2.4) chronological progress of the
approximately valid −2 /20, 22/
1,79 K 2
the process at the DIFK/FGF in Bonn/DEDP =[20,2,85 ⋅ 10the
Summing Tequ.(4.2.2
up⋅pore
K xdiffusion 1,84
⋅ ε ⋅ to − 0,3 the
⋅weεreceive
.4),coefficient ([cmpore
of: / hdiffusion
] )
(4.2.5)
weightof:change has been simulated calcula-
coefficient
[cm2/h] resultsλfrom = 'an1,84 ε − 0,3[−of
⋅ porosity ] . = 0,37 after (4.2.3)
21]. For this, amongst others,Dcylinders
P = 3572 made l −2 open 1,79 (1100°C).
a complete
tionally, by what we receive the result t0
burnout and a temperature DP = 2,85 of T⋅ 10
K = 1373K
⋅ T ⋅ ε ⋅ 1,84( ⋅ ε − 0,3 [cm2 / h] ) (4.2.5)
of industrially produced MgO – C materials K [20, 21]. The dashed lines in Fig. 15 are the
The oxidation DP = of 3572 [cm2/h]
carbon during
resultsthe from preheating
an open porosity of a lined of =MgO 0,37 –after C ladle
a complete in the
(36 mm x 100 mm Ø) were The oxidated on air,
diffusion of oxygen to the reaction front is rate determined (4.2.5) /20/. result
The linearof such calculations. It is observable
steel plant burnout
is analysedand a as an example.
temperature of TKThe = 1373K diffusion
(1100°C). coefficient of oxygen in the
during an increased temperature rateOof2–Nand their gas is /20/:
oxidation
2–CO – follows the t − law /15,20/: that the matching succeeds the best, when
loss of weight was continuously registered. The diffusion
DP =Dof 3572 oxygen [cm to2 −the
2 reaction
1,79 front 2 is rate determined /20/. The linear
= 2,85 (1− )2/h]
φ⋅ 10 =τ⋅ ,Tresults [cmfrom / h] . an open (4.2.6) no antioxidants
(4.2.4) delay the process of decar-
rate of oxidation follows the t −Klaw /15,20/:
Additionally, 3 mass-% metallicSumming aluminium porosityto .4),
up equ.(4.2.2 of we ε =receive
0,37 after the pore a diffusion
completecoefficient bonisation. of:
and/or boron carbide (B4C) were at which
used isasessential:
an- burnout and a temperature (1− φ )2 =τ , of T = 1373 K (4.2.6)
The result is depicted in Tab. 1.
K
x − 2 1,79 2
tioxidants. Fig. 15 shows the results by three D =
Pat which i
φ = (1100
2,85 ⋅ 10 °C).K of the( reaction front/wall
⋅
is (position
essential:
T ⋅ ε ⋅ 1,84 ⋅ ε − 0,3 ) [ cm thickness),Two
/ h ] (4.2.5) facts become clear by the quantitative
(4.2.7)
x0
examples. DP = 3572 [cm2/h] The diffusion
results fromxi ofanoxygen open porosityto the reaction
of = 0,37 front afterexamination:
a complete
φ= (position of the reaction front/wall thickness), (4.2.7)
The samples were heated up byburnout 5 K/min. andAa temperature
is rate determined ofx0TK = 1373K [20]. The (1100°C).
linear rate of ox- • Aluminium delays the decarbonisation for
significant reaction starts above 600 °C. of oxygen
The diffusion idation followsto the reaction the √ t-law front[15, is rate20]:determined /20/.about The linear 25 %, which is not much in com-
Discontinuities during the progress rate signalise
of oxidation (1 – φ ) the
follows
2 = τ t − law /15,20/: (4.2.6) parison to its above mentioned disadvan-
the influence of the antioxidants. The latter (1− φ )2 =τ ,
tages.
(4.2.6) 2
oxidate and cause an increase of weight, at which is essential: Two facts become •clear The by addition 2 of about 3 % boron carbide
the quantitative examination:
- Aluminium delays the decarbonisation for about 25%, which is not much in
which evidently delays the combustion reac-
at which is essential: has no
comparison to its above mentioned disadvantages.
significant influence.
tion. x (position of the reaction To make the practical carbidecalculation easier, eq.
φ = i (position of the reaction- front/wall The addition of about
thickness), 3% boron
(4.2.7) has an insignificant influence.
A clear change in terms of colour of the x0 front/wall thickness) (4.2.7) (4.2.9) can be paraphrased into the so called
Two
To make factsthe becomepractical clear by the quantitative
calculation easier, equ. examination:
(4.2.9) can be paraphrased into the
bright, from the outside decarbonised part, -soAluminium
called oxidation delays oxidation coefficient
the decarbonisation
coefficient /21/:
[21]:
for about 25%, which is not much in
to the black core is visible in the saw – cut t (running time/total
t τ= (running time/total timecomparison duration), to with its above mentioned (4.2.8) disadvantages. x02 2
with the naked eye, if the experiment τis= t (running
t0 timetime/total duration), time with duration), with
- The addition(4.2.8) of about α(4.2.8)
Ox 3% = 2 ⋅boron
V ⋅ DPcarbide ⋅ n0 = has [an / h] .
cminsignificant influence. (4.2.11)
0
z ⋅ t0
2
stopped before the complete decarbonisa- x 0 α.Oxmake has the beenpractical
calculated by the (4.2.11)
t0 = x02
z ⋅ 2 ⋅VO ⋅ D p (n.a − niso
To
) called
Two facts become
calculation
(4.2.9) clear 2results
easier, ofequ.numerous
by the quantitative (4.2.9) experiments
examination:
can be paraphrased at different into the
tion of the samples. This indicates a trans- t0 = temperatures (4.2.9) (4.2.9)
oxidation - Aluminium
/21/: coefficient delays /21/:
the decarbonisation for about 25%, which is not much in
z ⋅ 2 ⋅VO ⋅ D p (na − ni ) comparison to its above
port controlled topochemical reaction. −5 mentioned x02bydisadvantages. 2
ααOxOxhas = 4,5 been ⋅ 10calculated⋅ TC + 0,015[ the cm2results
/ h] . of nu-
].
(4.2.12)
(4.2.11)
t0 is the time, which is required for the oxidation of the total wall Ox = 2of⋅ Vabout
- Theαaddition
thickness x
⋅ DP3%⋅ nboron
. 0 = carbide [cm has an / hinsignificant influence.
t is the time, which is trequired
0 is the time, for the which
oxidationis required
of the for
total thewall oxi-thickness merous x . experiments at 0different tempera-
o z ⋅ t
4.2 Evaluation and discussion 0 z stands for the dation geometry of the
of sample:
the total sample: 2 (panel), We 4has
(cylinder),
α Oxxreceive beenα Ox 6tures
To=make
calculated(ball).
0,064
o
during
by the T C = 1100°C
results ofeasier, and
numerous =0,37. canSummarised,
experiments at different forthe
2wall thickness o.
the[21]:
practical calculation equ. (4.2.9) be paraphrased into
z standsVfor = 5the cmgeometry
3
/mol is the of molar
the volume (panel), 4 (cylinder),
of the graphite. 6 (ball).
3 1173K < TK </21/:
temperatures so called and
1643K oxidation0,16coefficient


0,37 /21/: the following relation is valid:
The oxidation reaction of the V =graphite
5 cm /mol is is the molar z stands volume for ofthethegeometry
graphite. of the3 sample:
n0 = 1,23 ⋅ 10−4 ⋅ COx / TK [mol 3 / cm ] (4.2.10) −5 x02 2 / h2 ] .
n0 = 1,23 10−4 ⋅ C4Ox(cylinder),
2 ⋅(panel), / TK [mol 6/ cm (ball).
] (4.2.10) −α5Ox = 4,5 ⋅α10 = 2⋅ T⋅ VC⋅ +DP0,015[
⋅ n0 = cm 0,79[cm / h2] . (4.2.12)
(4.2.11)
is the concentration of oxygen in the gas. COx in theαair = 3,5 ⋅ 1021⋅vol.%,
Oxamounts ε ⋅ (1,84 ⋅ ε − 0,3)⋅ COx ⋅ Tz K
Ox
⋅ t0 [cm / h] . (4.2.13)
2 〈C 〉 + {O2} = 2 {CO } is the
i.e. n(4.2.1)
concentration
= 1,88
of
10V =[mol/cm
oxygen
-6 5 cm /mol
in3 the
3
]. n isatthe
gas. C Ox molar
the
in the
reaction volume
air amounts
front isofinsignificantly
the
21 vol.%,
small. (4.2.12)
0 -6 3 i We receive α Ox αOx= 0,064
has beenduring calculated TC by = 1100°C
the results and of numerous =0,37. Summarised,
experiments at differentfor
i.e. n0 =For 1,88TK =101373K [mol/cm and ].thenibrick
graphite. at thethickness
reaction xfront 0 = 12
is cminsignificantly small.
For TK one = 1373K and the brick thickness x = 12 cm
COx amounts
1173K
amounts
is <the
from
from
TKconcentration
<(4.2.9)
1643K (4.2.9)
temperatures
for and the
of
/21/:
for
0,16oxygen the
in0,37

the gas, e.g.21 Vol%.
the following relation It is isimmediately
valid:
The partial pressure of the oxygen issided fixedtotal oxidation t0= 1072 hours.
one sided total oxidation t0= 1072 hours.
0
possible to calculate One t0 withreceivestheαOxknown
α−5Ox⋅ T= +0,064
= 4,5 ⋅ 10geometry z and cm2the
during
/ h] size α Ox is
x0, if (4.2.12)
−T5 = 1100 °C C 0,015[ 0,79 2
.
(F = 2), if the temperature and the CO-par- n 0 = 1,23 ·10 -4 ·C Ox /T K [mol/cm
During industrial practice about 1 cm burns out during the preheating. So that
3] α(4.2.10)
measured Ox =or 3,5 ⋅ 10
calculated C ⋅ ε ⋅ (
out1,84 of ⋅ εand
(4.2.13). − 0,3ε =) ⋅
The 0,37.
C Ox ⋅ TSummarised,
decarbonisation
K [ cm / h ] for
.
time t(4.2.13)
[h] of the
tial pressure is given.During The industrial
value ofpractice about 1 cm burns out duringlength the preheating.
x [cm] results WeSo 1173
2 that
receivefrom αKOx< =TK
equ. 0,064 < during
1643and
(4.2.8) TKC =and(4.2.9).
1100°C0,16and≤ ε =0,37. ≤ 0,37 Summarised, for
4.2 Evaluation and discussion equ. (4.2.6) and (4.2.7) result for the reduced time τ = (1−11 2 / 12 ) = 0,008 . The
PO2 = 1018 atm at 1100 °Cequ. and(4.2.6)
Ptime
CO = for
1 atm
and (4.2.7)isresult
the concentration
for the reduced
the decarbonisation of 1cm of the MgO-C
of oxygen
time τ = inC(1Ox
the
−11 gas.
is 12C)concentration
/the
according
0,008
Ox=1173K the< .Tfollowing
to equ. (4.2.8)
The
K <of oxygen
1643K
is
relation
and in the
0,16
isgas,
valid:
0,37e.g.21 Vol%.relation
the following It is immediately
is valid:
can be learned from the Richardson-Elling-
time for the decarbonisation
t =τ ⋅ t0 = 0,008 ⋅1072 in the of
≈ 9 Std1cm air, which
of amounts
the MgO-C
corresponds 21 In vol.-%,
according
possible
quite
summary, toto
well i.e.with
equ.
calculate(4.2.8) t is
with
the 0practical
the following the known
possibilities geometry
of the oxidation z and the size x 0 if
, α Ox is
The oxidation reaction of the graphite is −5 0,79 protection
2 arise:
ham-diagram [1] (Fig. 14).t =τ ⋅ t0 results. = 0,008 ⋅1072 ≈ 9 Std , which corresponds
It has been n 0= 1,88 · 10 [mol/cm
experimentally -6 observed 3 quitemeasured
]. nithat well
at the ∞ with
A
the reaction
thecalculated
preferably
or practicalαdense
Ox
kind of bonding phase, resin
= 3,5
out ⋅ 10
structure
of ⋅ ε ⋅
(4.2.13). (1,84
is ⋅ ε −
generally
The 0,3 ) ⋅ C
an
decarbonisation
Ox ⋅ T [
advantage.
K cm / h ]
time . t (4.2.13)
[h] of the
2TheC +so{Ocalled } = 2 { CO } . results.or It carboplast,
has been(4.2.1) experimentally
has no significant observed that the
influence on kind
length
the of(compare
xbonding
[cm] results
decarbonisation phase,
protectivefrom
rate resin
/23/. mechanisms
equ. (4.2.8) and one and eight)
(4.2.9).
2 pore diffusion coefficient is front is insignificantly
or carboplast, has no significant influence on the decarbonisation
small. For T K = 1373 rate K /23/. (4.2.13)
∞ Antioxidants COx isare the aconcentration
fourteenth of protective
oxygen in the gas, e.g.21 Vol%.but
mechanism, It ishave different
immediately
codetermined for the chronological The proce-offersand
model the the brick thickness
opportunity of an x0 = 12
easy and cmvery amounts
disadvantages,
detailed examination e.g. at the
of reprocessing of used material.
The partial pressure of the oxygen is fixed (F =The 2), ifmodel
the temperature and the CO In summary, possible to calculate t with the known geometry z and the size x , if α is
dure ofThe thevalue
corrosion. offers thefor opportunity (4.2.9)offor anthe easy and very detailed usethe
examinationoffollowing of
COxburner possibilities
for 0preheating of the oxidation protection
gas, 0 arise: Ox of
and from one sided
the total ∞oxidation
The the is the concentration of
in oxygen
the in the
substoichiometric area the
– partial pressure is given. -18
atmexperiments
of PO2= 10experiments at 1100°C
for Pdecarbonisation.
CO = 1
decarbonisation. For this,
For
the
this,
measured
measured
∞ A
chronological
preferably
chronological measured dense
progress or
progress
calculated
structure out isof (4.2.13).
generally The an decarbonisation
advantage. time t [h] of the
of the weight change
t has been simulated calculationally, oxygen by concentration,
what we receive i. e.
the
1, can be the fifteenth protective
atm can be learned from the Richardson- Ellingham Diagram /1/, Fig. 3.2.5.1.
of the weighttchange has0= 1072 h.
been simulated calculationally, by(compare what welength
e.g. 21 results
x [cm]
protective
receive
vol.-%.
the mechanisms fromItequ. is immediately
(4.2.8)
one and and(4.2.9). possible to
eight)
mechanism.
DP = D · λ · ε [cm2/h] result t0result /20,21/.
0 /20,21/. The dashed lines in Fig. 4.1.1 are the result of such calculations.
(4.2.2) The Duringlines
dashed industrial
in Fig. practice
4.1.1 are about
the 1∞cmAntioxidants
result of
Theburns
such
avoidance calculate
calculations.
are of a the t0 with the
entrance
fourteenth of known geometry
external
protective air into
mechanism, the z and but have
inside of thedifferent
ladle
The so called pore diffusion coefficient is codetermined It is
forobservable
thethat that the matching succeeds the best, when no antioxidants delay
chronological
procedure of the corrosion.
It is observable
the process ofthe out during
matching
decarbonisation. the
succeeds preheating.
the best, when So that will equ.
result inthe
nodisadvantages,
antioxidants the size
e.g.
delay atx0the
sixteenth , if reprocessing
α is
protective
Ox measured of used
mechanism. or calculated
material.
∞ The use of In summary, the following possibilities of the oxidation protection arise:
With the porosity ε and the thetortuosity
process of decarbonisation.
factor
The result (4.2.6)inand
is depicted table(4.2.7) 4.2.1. result for the reduced ∞the out burner
of (4.2.13).
A preferably
for preheating in the
The decarbonisation substoichiometric
time t area of the
2 The result is depicted in table 4.2.1. oxygen concentration, i.dense
e. structure

1, can is generally
be the fifteenth an advantage. protective
Dλ. ε [cm
p ∫ InD ⋅aλ ⋅lot /h]. is approximately valid
of cases (4.2.2)time τ = (1 – 11/12)2 = 0,008. The time for
mechanism. [h](compare
of the lengthmechanisms
protective x [cm] results one and from eight) eq.
With the porosity and [20,the22]tortuosity factor . In a lot of cases is the decarbonisation of 1 cm of the ∞ The MgO-C avoidance ∞(4.2.8)
Antioxidants
of the and are
entrance a fourteenth
(4.2.9). of external protectiveair into mechanism,
the inside but have
of thedifferent
ladle
approximately valid /20, 22/ will result in the disadvantages,
sixteenthe.g. at the reprocessing
protective mechanism. of used material.
according to equ. (4.2.8) is ∞InThe summary,
use of the the burner following
for preheating possibilities of the
in the substoichiometric area of the
x
λ = ' 1,84 ⋅ ε − 0,3[−] . (4.2.3) (4.2.3) t = τ · t0 = 0,008 · 1072 ≈ 9 h, which corre- oxidation protectioni. e.arise:
oxygen concentration,
1, can be the fifteenth protective 4
l
mechanism.
∞ The avoidance of the entrance of external air into the inside of the ladle
The oxidation of carbon during the preheating of a lined MgO – C ladle in the will result in the sixteenth protective mechanism.
steel plant is analysed90
as an example. The diffusion coefficient of oxygen in the refractories WORLDFORUM 5 (2013) [2]
O2–N2–CO – gas is /20/: 4
D = 2,85 ⋅ 10−2 ⋅ TK1,79[cm 2 / h] . Table 4.2.1:(4.2.4)
Influence of antioxidants on the decarbonisation (compare Fig.
Table 4.1.1) Influence of antioxidants on the decarbonisation (compare Fig.
4.2.1: /9/.
Summing up equ.(4.2.2 to .4), we receive the pore diffusion
4.1.1) /9/. coefficient of:
 KH2O = 1/PH2O = 1,57 106 is the equilibrium constant of the reaction
room temperature, i.e. PH2O = 6,37 10-7 atm. The partial pressure of
at equilibrium with liquid water at 300K amounts P0 = 0,0354 atm /2/
KH2O = 1/PH2O = 1,57 106 is the equilibrium 13,3cm3of
– VMgO) =constant /mol
theisreaction
the change of the
(6.2.1) at volume and the crystallisa
10-7 atm.follows
room temperature, i.e. PH2O = 6,37 pressure The partial Go = Rof
with pressure 1 the steam
T ln(P 0/PH2O) :

• A preferably dense structure is generally R2 = 84,8 atmat· equilibrium

cm3/mol · Kwithare
3
liquid
the water
ideal at 300K amounts P0 = 0,0354 atm /2/. (VMg(OH)2
an advantage. (compare protective mech-
– VMgO3) = 13,3cm /mol is the change of the volume
gas constants,pressure
V [cm /mol] is the
R1 ⋅ Tand
⋅ ln the PH 2O 8,3 ⋅ 300 ⋅ ln(0,035 / 6.4 ⋅ 10−7 )
P0 / crystallisation
follows withmolar
Go =vol-
R1 T πln(P− 10=/PH2O) :
(VMg (OH )2 − VMgO ) ⋅ R1 / R2
=
13,3 ⋅ 0,1
=2
anisms one and eight) ume of the phases. The higher the change of
• Antioxidants are a fourteenth protec- the volume, theπ less R1 ⋅ T ⋅ lnincrease
the required P0 / PH 2O of 8,3 ⋅ 300 ⋅ ln(0,035 / 6.4 ⋅ 10−7 )
−1 = = = 20 ⋅ 103[atm]
tive mechanism, but have different dis- (VMg (OHfor
the crystallisation pressure )2 − VMgO
the ) ⋅ R1 of
break / R2 13,3 ⋅ 0,1
advantages, e.g. at the reprocessing of the structure. Hence, it is necessary to have a packing(5.3.2)for the storing of MgO – mat
which is impermeable for steam. However, (5.3.2) the graphite bonding prote
used material. MgO – C material for practice sufficiently. This is the eighteenth pro
• The use of the burner for preheating in the 5.2 Formation of graphite Hence,
Hence, it is necessary to have a packing for ittheisstoring
mechanism. necessary to have
of MgO a packing for
– material,
substoichiometric area of the oxygen con- which is impermeable
C-gas especially easily decomposesfortosteam.
CO2 However, the graphite
the storing of MgO bonding
material,protects
which istheimper-
MgO – C material for practice sufficiently. This is the eighteenth protective
centration, i. e. λ ≤ 1, can be the fif- and graphite at about 500 °C.
mechanism. meable for steam. However, the graphite
teenth protective mechanism. bonding protects the MgO – C material for
• The avoidance of the entrance of external 2{CO} = 〈C〉 + {CO2}, ∆GO (773 K) = practice sufficiently. This is the eighteenth
air into the inside of the ladle will result in – 35 522 J/mol (Boudouard) (5.2.1) protective mechanism.
the sixteenth protective mechanism.
Condition is a catalyser e.g. metallic iron. 6 Thermomechanical behaviour
5 Corrosion of MgO – C through This low temperature is just given after a of MgO – C material
new phase formation certain interval of the melting, but we can Comparably less thermal stresses develop in
Condition is a catalyser
still count on the evaporated iron in colder
e.g. metallic iron. This low temperature is
the structure of carbon bonded refractory
5.1 Fundamentals after
parts of the lining. For the crystallisation
a certain interval of the melting, but we can still count on the
materials, because of the, in comparison to
Condition is a catalyser e.g. metallic iron iron. This low
in colder temperature
partsMgO, is just
of the given
The thermodynamic stability of a phase is
after a certain pressure,
interval ofaccording
the melting, to buteq.we (5.1.5),
can stillfollows
count on the morelining.
evaporated
For the
than doubled heatcrystallisation
conductivity pressure, a
described by its free enthalpy of formation iron in colder [36]:
parts of the lining. For equ.
the (5.1.5),
crystallisation follows
pressure, /36/:
and the relatively low thermal expansion
according to
∆G (Gibbs energy) [1]:
5. Corrosion of MgO – C through new phase
equ. (5.1.5), follows /36/:
formation o
[24]. −∆ G ois−the
That ⋅ T ⋅ ln K C protective
R1nineteenth
Condition
π −1 =
−∆is R1 ⋅ T ⋅ lne.g.
G a −catalyser K Cmetallic iron. This
[atm] . (5.2.2) π − 1(5.2.2)
low =temperature
mechanism.
is just given [atm] . (5.2
after a certainVCinterval ⋅ R 1 / R of
2
the melting, but we can still count on V C R 1 / R2
the⋅evaporated
∆G = ∆G + RT ln K [J/FC ]
5.1 Fundamentals 0 (5.1.1) Fig.
iron in colder parts of the lining. For the crystallisation pressure, according to 17 demonstrates the influence of the
equ. (5.1.5), follows /36/: 2
The equilibrium Theconstant
equilibriumat 500°C constant
is KC = Pat CO2/(P 500 CO )°C graphite
= 2is(Fig 5.2.1) andconcentration
the in the structure on
The thermodynamic stability of a phase is described by its free
molar volume enthalpy
of the graphite of2is VC−∆ =The o cm
5,3
G − Requilibrium
3
/mol.
⋅ T ⋅ lnOut
K constant
of equ.(5.2.2) followsat 500°C is KC = PCO2/(PCO)2 = 2 (Fig 5.2
formation Out of theEnergy) chemical /1/: reaction (formula+35522 KC = PCO2/(PπCO)− 1== 2 (Fig. 16) and different characteristics of the material.
1 C
G (Gibbs [atm] . (5.2.2) 3
change (FC)), e.g. π − 1 =
− 8,31 ⋅ 773 ⋅ ln 2
the5,3molar = 58 molar
⋅ 10V3 ⋅ R volume
atm
volume of the graphite is
C 1, / R
which2
of
destroys the
the graphite
material.
Fig. 18 illustrates
is V
theC = 5,3
completely
cmdifferent
/mol. Out of equ.(5.2
o ⋅ 0,1
∆G = ∆G + RT ln K [ J / FC ] . Vthe (5.1.1) +is35522 −in8,31 ⋅ 773 ln 2 and
⋅5.2.1) 3 ceramic and
This effect inThe C =MgO-C
5,3 cm3material,
equilibrium /mol. Outhowever,
constant πofateq. 1(5.2.2)
−500°C =is not follows
Kobserved
C = PCO2/(P thermomechanical
)2 = 2 (Fig
practice,
CO because behaviour
= 58the⋅ 10of atm , which destroys the m
Out of the chemical reaction (formula change the (FC)), e.g.graphite 3
A + 2B = AB2, T = const. (5.1.2)
given molar volume
bonding of
does the graphite
not act is V
autocatalyticC = 5,3 cm
and 5,3
/mol. ⋅
Out 0,1of
graphitic
furthermore equ.(5.2.2)
bondsbonding.
the follows
35522 − 8,31 ⋅of
iron as carbide below+temperatures 773 ⋅ ln 2 This is3 the seventeenth
1200°C.
π − 1(5.1.2)
= = 58 ⋅ 10 atm , which It isdestroys
noticeable that the linear – elastic frac-
the material.
5,3 ⋅ 0,1 This effect in the MgO-C material, however, is not observed in pra
A + 2B = AB2, T = const protective mechanism.
the equilibrium constant follows This effect in the MgO-C the givenhowever,
material, graphite ture
is notbonding
mechanic
observed in does
is not applicable
practice,not
for refracto-
act autocatalytic
because and furtherm
the equilibrium constant follows
which
the givendestroys
graphite thebonding
material. does
iron as carbide not act autocatalytic ry materials.
and Their
furthermore heterogeneous
bonds the
below temperatures of 1200°C. This is the sevente structural
7
a
K AB2 = AB 22 . (5.1.3) iron
This as carbide
effect in the belowMgOtemperatures
– C material, of 1200°C.
howev- This is the seventeenth
a A ⋅ aB protective mechanism. protective mechanism.
er, is(5.1.3)
not observed in practice, because the
EquilibriumEquilibrium
is G = 0 (equ.is ∆G1.1.1),
= 0 (eq.i.e.1.1.1), i.e. given graphite bonding does not act auto-
catalytic and furthermore bonds the iron as
0
∆GG0 == -- RTRT· lnK lnK[J/FC]
[J/FC]. (5.1.4) carbide(5.1.4)
below temperatures of 1200 °C. This
The increasing activity aAB2 (supersaturation) often leads to an considerable
is the seventeenth protective mecha-
increase ofThe
theincreasing
crystallisation activity aAB2 (supersaturation)
pressure π, which is put on the nism.
environment by
, in theleads
phase AB2often case of to ananinhibition of growth
considerable of phase AB2 , e.g. by the
increase
nucleus surrounding crystal material. This pressure equals the
of the crystallisation pressure π, which is 5.3mechanical
Formation workof brucite
(π - 1) (VAB2 - VA - 2VB), because ∂G / ∂π ∆V /1/ and is added to the free
enthalpy ofput on the environment
formation G°. It followsby phase AB2, in MgO materials, which are kept on air dis-
the case of an inhibition of growth of phase solve, because brucite is formed:
G° + (π - 1) (VAB2 - VA - 2VB) R1/R2 + R1TlnKFig.5.2.1: Influence of the (PCO+ PCO2)-pressure in CO2/CO-compositions
AB2 = 0 (5.1.5)
AB2, e.g. by the nucleus surroundingoncrystal the reduction power of the carbon on iron oxides (shifting of the Boudouard-
We receivematerial. This pressure
the equilibrium for theequals graph
the mechani-
crystallisation with the
pressure total
π〈MgO〉
= 1 atm pressure)
+ {H /37/
2O} = 〈Mg(OH)
without the 2〉; ∆G (300 K) =
O

saturation cal
of ABwork
2 , (π
i.e. -
for 1)a · (V
AB2 AB2 - V1A - 2VB), because
= 1. R = 8,31 J/mol K –
and 35567 J/mol. (5.3.1)
R2 = 84,8 atm Fig.5.2.1:
3 Influence of the (PCO+ PCO2)-pressure in CO2/CO-compositions
∂G/∂π ≡cm
/mol K are the ideal
∆V [1] and is added to the 5.3 gas constants,
freeFormation
en- V[cm
onof
/mol]
thebrucite
is
reduction
the
power of the carbon on iron oxides (shifting of the Boudouard-
molar volume of the phases. The higher the change of the volume, the less the
thalpy ofofformation
required increase ∆G0. It follows
the crystallisation pressure for the break KHof
graph =with1/PHthe
2O the structure.
2O
=total
1,57 · 106 is /37/
pressure) the equilibrium
MgO materials, which are
constant of kept
the on air dissolve,
reaction (5.3.1)because at room brucite is formed:
∆G + (π - 1) (VAB2 - VA - 2VB) · MgO + H2O = Mg(OH)2; G° (300
O temperature, i.e. P H2OK)= =6,37 · 10 J/mol.
- 35567
-7 atm. The (5.3.1)
5.2 Formation of graphite 5.3 Formation of brucite
R1/R2 +R1TlnKAB2 = 0 (5.1.5) partial pressure of the steam at equilibrium
with liquid water at kept 300onK airamounts
COCO2
+ PCO
CO – gas especially easily decomposes to CO2 and graphiteMgO materials,
at about 500°C. which are
Fig.5.2.1: dissolve, because brucite is formed:
Influence of the 6 (P CO+(PP )-pressure in CO2/CO-c
2
Fig. 16 Influence of the )-pres-
One receives the equilibrium for the crystalli- P 0 = 0,0354
MgO + H
atm [2]. (VMg(OH)2 – VMgO) =
2O = Mg(OH)on 2; G° the
(300 reduction
K) = - 35567 sure inof
power
J/mol. CO2the
/CO-compositions
carbon
(5.3.1) on on theoxides (shifting of th
iron
2{CO} = <C>sation+ {CO2},
pressureG°π(773 = 1 K) atm= - without
35522 J/mol the (Boudouard).
13,3 cm3/mol (6.2.1)
is the change of the volume
supersaturation of AB2, i.e. for aAB2 = 1.
graph with the totalreduction
and the crystallisation pressure follows with
pressure) power of the carbon on iron
/37/
oxides (shifting of the Boudouard-graph
6
R1 = 8,31 J/mol · K and ∆GO = R1 · T · ln (P0/PH2O): with the total pressure) [37]
5

5.3 Formation of brucite

refractories WORLDFORUM 5 (2013) [2] 91

MgO materials, which are kept on air dissolve, because brucite is f

 Comparably less thermal stresses develop in the structure of carbon bonded
refractory materials, because of the, in comparison to MgO, more than doubled heat
conductivity and the relatively low thermal expansion /24/. That is the nineteenth
protective mechanism.

Fig. 5.1 demonstrates the influence of the graphite concentration in the structure on
during very high fracture stress to reach a
preferably high flexibility of structure. This is
against low amounts of graphite (compare
MgO–C in Fig. 17 and 18) and should be a
reason to choose RSt as thermal stress dam-
age resistance parameter [33]. Actually, the
so called characteristic length LC often
A low E-modulus has often worked well in the industrial experience of a steel-
proves of value, which is seen as measure for
plant during very high fracture stress to reach a preferably high flexibility of
the flexibility
structure. This is against low amounts of the
of graphite structure,
(compare thus the
MgO-C resid-
in Fig. 5.1
and 5. 2) and should be a reasonual resistance
to choose RSt asafter a previous
thermal mechanical
stress damage
resistance parameter /33/. Actually, the so called characteristic length LC often
pressure and which derives from R’’’’ (com-
proves of value, which is seen as measure for the flexibility of the structure,
thus the residual resistance afterpare 6.1) [31,
a previous 33]: pressure and which
mechanical
derives from R’’’’ (compare 5.1) /31,33/:
Gf ⋅E
Lc = [ m] . (5.3) (6.3)
σ2
f
The characteristic length LC should be preferably big.
For this H. Harmuth (priv. memo.) Thepoints
characteristic length LC should
to his experiences, whereby beLprefer-
C–

Fig. 17 Influence of the concentration of graphite in the structure values at the comparative sequence
on different ablyofbig.
product spectra (e.g. differently burned
Fig. 5.1: Influence of the concentration of graphite in the structure on different kinds of stones, but also the inclusion of carbon bonded kinds of stones)
characteristics of the MgO – C material [25] according to their characteristic For length, H. Harmuth
thiscomply (priv.with
very well memo.) points toof
that resistance
characteristics of the MgO – C material /25/ his experiences, whereby LThis
thermal shock, which is known from industrial observations. C-values at the
is especially
constitution is theessential, reason iffor
Fig. 5.2 comparison of the load – displacement behaviour of two materials with
thethis.
crackHence,
initiation iscomparative
authoritative.sequence of product spectra
Griffith [27]and
a ceramic bonding (picrochromite) and Irwine
a graphite [28]
bonding
The determination introduced
(MgO that–the /26/(e.g.
inC)equ. (5.1)differently burned kinds
no direct influence of the of temperature
stones, but is
Fig. 5.2 comparison of the load – displacement behaviour of two materials with
Fig. 5.2. illustrates the completely different thermomechanical behaviour of
aItceramic bonding
balance ofand
(picrochromite)
is noticeable that the linear – elastic fracture
energy
a
of thebonding
noticeable,
graphite
fracture
mechanicTis[K
coefficient
mechanic
is another
(MgO
not – C) /26/
-1 applicable for
] is in
also the
point. The
denominator
inclusion
square of the
position
of carbon
during a
bonded kinds
linear thermal
quasi-static
of
expansion
path of
process. Hasselman [29, 30]constitution
derived theis fol- stones) according to their characteristic
ceramic and graphitic bonding. refractory materials. Their
Itthis.
is noticeable that the
Hence, Griffith linear
/27/
heterogeneous
and –Irwine
lowing elastic
two /28/
structural
fracture
thermal
crack (Gl. 5.2).
mechanic
stressexperience.
introduced damage
the balance
the reason
of energyfor
resist-
for
should be as small as possible, which is proven by the
is not Tapplicable length,
of the comply very well with that resistance
practical
refractory materials.
fracture mechanic process.Their heterogeneous
Hasselman /29, structural
30/this: constitution
derived the following is thetworeason for
this. Hence, Griffith /27/ ance
and parameter
Irwine /28/ from
introduced the balance of energy of the
of thermal shock, which is known from in-
thermal stress damage resistance parameterGraphite from this: bonded materials, e.g. MgO-C, have a much better thermal shock
fracture mechanic process. Hasselman G f ⋅ E /29, resistance30/ derivedthan the ceramic
following dustrial observations.
two materials.
bonded Reason is mainly This theis especially
distinctive es-
thermal stress damage resistance R '''' = parameter2
[m] thermoplastic
from this: behaviour (5.1) (6.1)
of thesesential, if
materials, the crack
which is initiation
unaccounted is authoritative.
for no
G2f⋅ σ
⋅ Ef characteristic number. Fig.5.3 diagrams the comparison between MgO and
and R '''' = [ m ] (5.1) The determination that in eq. (6.1) no direct
2 ⋅ σ 2f MgO-C at an example of Hampel and Aneziris /33/. The plasticity follows the
and Gf increase influence of the temperature
of the cyclic loading, while the elastic behaviour hardly changes. The is noticeable, is
and Rst = ⋅ m1/2 ] behaviour
[ Kplastic . (5.2)
2 ⋅G 2
αT ⋅ E is a “mechanical another point.
reserve“ of The
the square
material of
and the linear
increases ther-
the
Rst = f
2
[ Kflexibility
⋅ m1/2 ] . of the structure, (5.2)(6.2)whichmal is important for its thermal shock
expansion coefficient αT [K-1] is in de-
resistance.
2 2 ⋅ α T ⋅E
Gf [J/m ] is the specific fracture energy, f [MPa] the initial strength of thenominator position during a quasi-static
material,2
E [GPa] the modulus of elasticity and T [K-1] the linear thermal
comparison of the Fload – displacement behaviour
G f [J/m ] iscoefficient.
expansion the Gf [J/m
two materials
of specific fracture
The
2]energy,
formation is of
with theincipient
f specific thefracture
[MPa]cracks initial energy,
strength
is made path of crack (eq. 6.2). αT should be as small
of thewith
difficult
ig. 18 Comparison of the load-displace-
material, E [GPa] the modulus of elasticity and [K -1
] the linear thermal
ic bonding (picrochromite) and a graphite an bonding
increasing (MgO
R’’’’, –
theirC)
σ /26/
[MPa]
growth
f the
with initial
an strength
increasing T R of
St . the material, as possible, which is proven by the practical
ment behaviour of two materials
expansion with a
coefficient. The formation of incipient cracks is made difficult with
an increasing
ceramic bonding (picrochromite) and R’’’’,
a their Egrowth
[GPa] with the anmodulus
increasing ofRStelasticity
. and experience.
By comparing the parameter for kinetic cracks growth at cracks initiation (equ.
iceable that the linear – elastic fracture mechanic is not applicable -1 for
graphite bonding (MgO –5.1) T [K ]propagation
[26] the quasi-staticαcrack
C) with the linear thermal
(equ. 5.2),expansion coeffi-is Graphite bonded materials, e.g. MgO–C, have
the following
ry materials. Their heterogeneous structural By constitution
comparing
conspicuous /31,32/:is the
the parameter reason forcracks growth
for kinetic
cient. The formation at cracks initiation (equ.
of incipient cracks is a much better thermal shock resistance than
nce, Griffith /27/ and Irwine /28/ introduced 5.1) withthe the quasi-static
balance crack propagation
of energy of the (equ. 5.2), the following is
conspicuous /31,32/: made difficult with an increasing R’’’’, their ceramic bonded materials. Reason is mainly
mechanic process. Hasselman /29, 30/ derived The modulus of elasticity two
the following is in numerator position at the kinetic crack process
and in denominator growth
position in with
the an increasing
quasi-static case. RHence,
St. the optimisationtheof distinctive1thermoplastic behaviour of 3
stress damage resistance parameter fromThe this:
modulusmaterial
of elasticity
a refractory Byis in numerator
is problematic.
comparing position at
The the
modulus of the
parameter kineticshall
elasticity
for crack process
kinetic these materials, which is unaccounted for no
Gf ⋅ E and in denominator
(theoretically) be asposition in the quasi-static
big, alternatively as small case. Hence,which
as possible, the optimisation of
is not possible
R '''' = [ m] aatrefractory (5.1)
material cracks
is problematic.growth
Theat cracks of
modulus initiation
elasticity (eq.
shall6.1) characteristic number. Fig. 19 diagrams the
2 ⋅σ f2 the same time. Therefore the required demands during the practical
(theoretically) be as big, with
experience is important. alternatively as small as
the quasi-static
The latter determines thepossible,
crack which is(eq.
propagation
choice. not possible
comparison between MgO and MgO–C at an
and at the same time. Therefore the required demands during the practical
6.2), the following
experience is important. The latter determines the choice.is conspicuous [31, 32]. example of Hampel and Aneziris [33]. The
Gf The modulus of elasticity is in numerator po- 2 follows the increase of the cyclic
plasticity
Rst = 2
[ K ⋅ m1/2 ] . (5.2)
2 ⋅ αT ⋅ E sition at the kinetic crack process and in de- loading, while the elastic behaviour hardly
2
nominator position in the quasi-static case. changes. The plastic behaviour is a “mechan-
2
] is the specific fracture energy, f [MPa] the initial strength Hence,
of the the optimisation of a refractory ma- ical reserve“ of the material and increases
-1
l, E [GPa] the modulus of elasticity and T [K ] the linear thermal terial is problematic. The modulus of elastic- the flexibility of the structure, which is impor-
on coefficient. The formation of incipient cracks is made difficult ity shallwith
(theoretically) be as big, alternative- tant for its thermal shock resistance.
asing R’’’’, their growth with an increasing RSt . ly as small as possible, which is not possible It becomes clear that the characteristic num-
at the same time. Therefore the required de- bers, calculated with eq. 6.1 and 6.2, allow
paring the parameter for kinetic cracks growth at cracks initiation (equ. the practical experience is im- no reliable statement. Rather are analyses
mands during
quasi-static
h theComparison
.5.3: crack
of the propagation
partly (equ.
elastic bending 5.2), the
behaviour of two materialsis
following
Fig. 19 Comparison
during an increasing loading /35/ of the partly elastic
portant. The latter determines the choice. with the consideration of plasticity and their
uous /31,32/:
bending behaviour of two materials A low E-modulus has often worked well in dependence of temperature and time need-
the industrial experience of a steel-plant ed to characterise this group of materials.
ecomesof
dulus elasticity
clear isduring an increasing loading
in numerator
that the characteristic position
numbers, at [35]
calculated kinetic
thewith equ. 5.1 and process
crack
denominator position
, allow no reliable in the
statement. quasi-static
Rather are analyses withHence,
case. the optimisation
the consideration of of
sticity and their dependence
tory material is problematic.of temperature and time
The modulus ofneeded to characterise
elasticity shall
group of materials. 92 refractories WORLDFORUM 5 (2013) [2]
ically) be as big, alternatively as small as possible, which is not possible
ame
e thermoplastic behaviourthe
time. Therefore the MgO –demands
of required during the practical
C material is a twentieth protective
nce
chanism The latter
against mechanical
is important. determines the choice.
wear.
The thermoplastic behaviour of the MgO – C Tab. 2 Evaluation of the characteristics, relevant for practice, of carbon containing mag-
material is a twentieth protective mech- nesia- and magnesia – carbon bricks during operation temperatures [38]
anism against mechanical wear. Magnesia Magesia – Carbon
Type of Stone Residual Carbon Fired Pitch
7 Concluding discussion Pitch Pitch Resin Bounded
[mass-%] Imbued Bounded Bounded + Antioxidants
Carbon bonded refractory materials have be- 2 5 14 14
come worldwide established in steel metal-
corrosian resistance 3 1–2 1 1–32)
lurgy. The reasons are multifaceted and di-
vided into four groups: oxidation and redox stability 1 2–3 3–4 2–4 2)
errosion strength 1–2 3 3 2
Structure tension load capability 3–4 2 1 1–32)
• High elasticity of the structure as a result
of the thermoplastic behaviour of the 1) relative evaluation
graphitic bonding matrix. 2) strongly dependantion
• Hindrance of the rapid oxidation, especial-
ly during the use of antioxidants (e.g. centrations of oxygen are more than • Keeping material dry during long-term
Almet), by the slow topochemical (depen- 30 ppm. storage.
dent on a clear position) reaction of the • The oxides mentioned above are reduced Finally the question arises, which percentage
well crystallised graphite. during the contact between slag and of graphite is at least required for the
• MA-spinel forms if metallic aluminium is graphite so that the eutectic setting tem- favoured function of the material. Material,
used as antioxidant, whereby the structure perature [19] of the infiltrated slag and its saturated with pitch or tar, contains about
is densifcated and less prone to infiltra- viscosity [19] highly increase, which hin- 2 – 6 mass-% carbon [24]. The structure be-
tion. The parallel generating {Mg} – gas ders the infiltration. comes poor in porosities and is better
getters the oxygen. protected against infiltration if the per-
• The graphite bonding offers the possible Solubility centage of graphite is higher. Well distrib-
application of specific (coarser) oxide – • The solubility of the graphite in iron is de- uted graphite, crystalline (flake graphite)
grain – size - distributions for the decrease pendent on the CO-partial pressure and or/and carbon black is existent in the struc-
of the dissolution of the structure by infil- the concentration of solute oxygen. Hence, ture and protects against infiltration and
trated slag. in iron infiltrate are at 1600 °C instead of corrosion, if the concentrations are higher
• No formation of graphite from the decom- 5,3 mass-% less than 0,4 mass-% carbon (10 – 20 mass-% carbon). Additionally the
position of CO takes place during working soluble, what stabilises the structure. thermal shock resistance and the mechani-
action, because of a missing catalyser, i.e. • The solubility of the graphite in slag cal loading are improved. Tab. 2 shows
no destruction of the structure via crys- decreases with an increase of the CO-par- the configuration of a relative evaluation
tallisation pressure. tial pressure and hardly exceeds [24, 38].
• Hindrance of the formation of brucite dur- 1 mass-%, which hinders the corrosion of The percentage of graphite of magnesia –
ing the storing of the stones on humid air, the structure as well. carbon – bricks is going to be decreased to
because of the cladding of the grains by • The dissolution of the refractory material, 3 mass-% for the protection of the environ-
graphite. as a result of the infiltration of slag, is af- ment (CO2 production), without worsen the
ter the achievement of saturation con- characteristics of the material [39]. It would
Infiltration trolled by Ostwald – ripening, thus hap- be possible to test, for instance, nano-crys-
• Neither steel nor slag wet graphite so that pens very slowly. talline graphite to achieve a consistent cov-
no spontaneous infiltration happens. Infil- er of all oxide surfaces already at the lowest
tration is just possible at an adequate Procedural actions percentage. However, the characteristics of
gravitation pressure. • Avoidance of low pressure (vacuum) saves this graphite differ completely from those of
• A layer of 〈MgO〉 can appear in the steel the MgO–C lining. the so far favoured flake graphite and have
at the boundary area, if the concentration • The cover of the lining via a slag coating to be tested first. It cannot be assumed that
of oxygen of the melt is high enough. The (Japan) reduces the corrosion by oxidation all discussed corrosion characteristics are
layer hinders the infiltration. This is also and chemical dissolution. achievable with 3 mass-% carbon or less,
observable on air. • The early input of (MgO) into the slag de- because the infiltration- and mechanical
• A {CO} – cap, which hinders the infiltra- lays the dissolution of the refractory mate- characteristics will worsen and the extreme-
tion, forms during the contact between rial by diffusion. ly small grain size will be very prone to oxi-
slag and graphite and the reduction of the • The adjustment of the ladle fire to a stoi- dation and dissolution of the graphite in
oxides (FeO) and (MnO). chiometric factor λ < 1 and the avoidance steel and slag (compare (3.2.3.1.3)). The
• A {CO} – cap, which hinders the infiltra- of unintentional air entrance delays the production of nano-crystalline graphite us-
tion, can also form during the contact burnout of the graphite and therefore pro- ing a gas [40] and its processing will pose
between steel and graphite, if the con- tects the lining. problems i.e. spontaneous ignition. For spe-

refractories WORLDFORUM 5 (2013) [2] 93

cial applications MgO – C materials based [13] Gans, W.: Zerfall und Auflösung von Haufwer- [27] Griffith, A. A.: Theory of Rupture. 1. Int. Congr.
on tar have been tested successfully ]24, 38] ken; gen. Habilitationsschrift RWTH-Aachen for Applied Mechanics, Delft, NL. (1924) 55–63
but its broad application should be difficult (1973) [28] Irwin, G. R.: Analysis of stress and strains near
to realise. [14] Darken, L. S.; Gurry, R. W.: Physical Chemistry of the end of a crack traversing a Plate, J. Appl.
Metals; McGraw-Hill Comp. N.Y. (1953) Mechanics 24 (1957) [3] 361–364
References [15] Friedrichs, H. A.; Knacke, O.: Verfahrenstechni- [29] Hasselman, D. P. H.: Unified theory of thermal
[1] Frohberg, M.G.: Thermodynamik für Werkstoff- sche Grundtypen isothermer Reaktionen in shock fracture initiation and crack propagation
ingenieure und Metallurgen. 2. Auflage, Deut- durch- und überströmten Schüttungen; For- in brittleceramics J. Am. Ceram. Soc. 52 (1969)
scher Verlag für Grundstoffindustrie, Leipzig schungsberichte des Landes NRW Nr. 2240, 600–604
(1994) Verl. Opladen (1973) [30] Hasselman, D. P. H.: Rolle der Bruchzähigkeit bei
2] Hack, K. u. M.: FACTSage, Thermfact & GTT [16] Carslaw, H. S.; Jaeger, J. C.: Conduction of heat in der Temperaturwechselbeständigkeit. Ber. Dt.
Technologies (2007), www.factsage.com solids, Oxford University Press, 2. Auflage, (1959) Keram. Ges. 54 (1977) [6] 195 – 201
[3] Ollig, M.: Reaktionen kohlenstoffhaltiger MgO [17] Kahlweit, M.: Grenzflächenerscheinungen; [31] Harmuth, H.; et al.: Bruchmechanische Charak-
– Werkstoffe mit flüssigem Eisen, Dr.- Ing. Diss. Grundzüge der Phys. Chemie Steinkopf, Darm- terisierung grobkeramischer feuerfester Werk-
RWTH – Aachen (2000) stadt (1981) stoffe; Veitsch-Radex Rundschau, 1–2 (1994) 4
[4] Pötschke, J.; Beimdiek, K.; Ollig, M.: Reaction [18] Schubert, H. G.; Schwerdtfeger, W.: Löslichkeit [32] Tschegg, E. K.: Prüfeinrichtung zur Ermittlung
between MgO-C Bricks and Steel Melts Proc. von Kohlenstoff in ESU – Schlacken; Arch. Ei- von bruchmechanischen Kennwerten, Austrian
6th. Bienn. Worldw. Congress of Refractories, senhüttenwesen 45 (1974) 437–9 Patent Specification AT 390328 (1986)
Berlin, September (1999) 166–169 [19] VDEh: Slag Atlas, 2. Auflage, Verlag Stahleisen [33] Schickle, B.; Telle, R.; Tonnesen, T.; Miyaji, D. Y.;
[5] Meschede, D.: Gerthsen Physik; Springer GmbH Düsseldorf (1995) de Anchieta Rodrigues, J.: Change of Mechani-
(2005) [20] Pötschke, J.; Routschka, G.; Simmat, R.: The eva- cal and Elastic Properties of Castables as a
[6] Jeschke, P.: Texture analysis of basic refractory luation of oxidation experiments on carbon- Function of Thermal Shock Cycles. Proc. 53th In-
brick, J. Am. Ceram. Soc. 49 (1966) [7] 360–363 containing refractories. Proc. 45th Int. Coll. on tern. Coll. on Refractories, Aachen (2010)
[7] Pötschke, J.; Routschka, G.; Simmat, R.: The eva- Refractories Aachen (2002) 36–39 86–89
luation of oxidation experiments on carbon- [21] Pötschke, J.; et. al.: Properties and Corrosion of [34] Buchebener, G.; Pirker, S.: Neue Hochleistungs-
containing refractories. Proc. 45th Int. Coll. on AMC-Refractories, Proc. UNITECR'03, 580–583 produkte für den Einsatz im Konverter. Veitsch–
Refractories, Aachen (2002) 36–39 [22] Jeschke, P.: Texture analysis of basic refractory Radex−Rundschau (1996) [2] 3–14
[8] Oeters, F.: Metallurgie der Stahlherstellung, Ver- brick, J. Am. Ceram. Soc. 49 (1966) [7] [35] M. Hampel, C. G. Anezieris: Thermal Shock Per-
lag Stahleisen/Springer (1989) 360–363 formance of Carbon Bonded Materials. Proc.
[9] Pötschke, J.: Feuerfestkorrosion durch Stahl und [23] Simmat, R.; Pötschke, J.: Oxidation Experiments 48th Int. Coll. on Refractories, Aachen (2005) 28
Schlacke; Seminarvortrag (2001) VDEh, Düssel- on different Types of Carbon Bond of Refracto- –32
dorf ries. Proc. 46th Int. Coll. on Refractories, Aachen [36] Baukloh, R.; et al.: Der Wachstumsdruck des
[10] Brüggmann, C.; Pötschke, J.: Disaggregation (2003) 11–12 Kohlenstoffs; Arch. Eisenhüttenwesen. 27.2
and Dissolution of Refractories using the [24] Routschka, G.; Wuthnow, H.: Praxishandbuch (1956) 95–99
Example of MgO/Lime-Aluminate Slags. Proc. Feuerfeste Werkstoffe, 5. Auflage (2011) Vulkan [37] Eisenhütte, 5. Auflage (1961) 94
54th Int. Coll. on Refractories, Aachen (2011) Verl. Essen [38] Barthel, H.; Kaltner, E.: The Effect of Carbon in
139–141 [25] Eschner, A.: Feuerfeste Werkstoffe in Theorie Carbon – Containing Magnesia Bricks on the
[11] Brüggmann, C.; Pötschke, J.: MgO – Saturation und Praxis; Seminar GHI RWTH-Aachen (2011) Wear in Basic Oxygen Furnaces; Reprint 1st Int.
in Secondary Metallurgical Lime-aluminate and [26] Vollenberg, I.; Mittler, G.; Brochen, E.; Brük- Conf. Refr. TOkyo (1983) 91–104
Lime-silicate Slags; steel research int. 82 kmann, C.; Pötschke, J.; Mudersbach, D.: Opti- [39] FIRE: German Research Foundation SPP 1418
(2011) [4] 422–427 mization of Refractories by Combining Labora- FIRE
[12] Brüggmann, C.: Ein Beitrag zur Verschlackung tory Investigations with Numerical Simulation [40] Walker, P. L.; et al.: Carbon Formation from
von MgO in sekundärmetallurgischen Schlak- using the Example of a MgO-Brick in Steel Lad- CO/H2 – Mixtures over Iron Catalysts. II. Rates
ken, Dr. Ing.-Diss. (2011) TU Bergakademie Frei- le and Converter. Proc. 53th Intern. Coll. on Re- of Carbon Formation; J. Phys. Chem. 63 (1958)
berg fractories, Aachen (2010) 152–154 140–149

94 refractories WORLDFORUM 5 (2013) [2]