Engineering, 2010, 2, 496-501
doi:10.4236/eng.2010.27065 Published Online July 2010 (http://www.SciRP.org/journal/eng)
The Effects of an Applied Voltage on the Corrosion
Characteristics of Dense MgO
Brian Joseph Monaghan1, Sharon April Nightingale1, Qiang Dong2, Michael Funcik2
1PYROmetallurgy Research Group, University of Wollongong, Wollongong, Australia
2Former graduate students, University of Wollongong, Faculty of Engineering,
Wollongong, Australia
Received March 10, 2010; revised March 10, 2010; accepted May 2, 2010
Abstract
In a recent investigation Mills and Riaz [1] showed that industrial oxide refractory corrosion by liquid oxides
could be changed by the application of a small voltage across the liquid oxide-refractory interface. They ex-
plained their result in terms of penetration of the refractory pores with liquid oxide. Their analysis of the
corrosion effect was to some degree limited by the use of an industrial refractory material in the study. Fur-
ther, it was not clear whether their findings were limited to solely industrial refractories or had wider ranging
application to more dense ceramic type solid oxide systems. In this study, a simpler and more easily charac-
terized solid oxide (dense MgO) has been used to examine the effects of an applied voltage on the solid ox-
ide in a liquid oxide melt. The dissolution rate of an MgO ceramic in a CaO-SiO2-Al2O3 and CaO-SiO2-
Fe2O3-FeO-MgO liquid oxide composition at various applied voltages has been measured at 1540°C. It was
found that the MgO corrosion in the CaO-SiO2-Al2O3 system was insensitive to an applied voltage over the
voltage range –0.5 to 0.3 V. In the CaO-SiO2-Fe2O3-FeO-MgO liquid oxide system the MgO corrosion rate
showed a maximum at –0.45 V. This effect has been explained by considering the consequences of an ap-
plied voltage on the rate of Marangoni flow at the liquid oxide-refractory-gas interface and in turn, the flow
effect on the rate of the mass transfer controlled MgO dissolution reaction.
Keywords: Slags, Liquid Oxide, Refractory, MgO, Electrode Potential
1. Introduction
MgO based refractories are widely used in the metallur-
gical industries and their service lifetimes are often lim-
ited by the rate the refractory dissolves in the liquid ox-
ides generated in the processes [2]. The dissolution rate
is a function of contact area between the liquid oxide and
refractory which in turn is a function of the area calcu-
lated using the macro dimensions of the refractory and
the pore structure within the refractory. Recently there
have been two separate studies that have shown that the
rate of liquid oxide penetration into a refractory and
hence the rate of dissolution of a refractory into the liq-
uid may be affected by the application of a small applied
voltage across a liquid oxide-refractory interface [1,3].
The results of these studies were explained using an in-
tegrated form of Poiseulle’s law [3], shown in Equation
(1), for liquid penetration, l, into pores,

2cos 2
r
l








t


(1)
where r is the pore radius,
 is the interfacial tension, 
the contact angle between the liquid oxide and refractory,
η is the dynamic viscosity of the liquid oxide and t is
time. In these studies it was argued that by application of
an electrode potential across the refractory-liquid oxide
interface the
 and  change thereby reducing the depth
of penetration for a given time. A schematic showing
liquid penetration in a wetting (  < 90°) and non-wetting
(
 > 90°) system are given in Figure 1.
The findings presented in these studies [1,3] require
further investigation. Why should the interfacial tension
and contact angle of a liquid oxide-refractory system
change with changing applied voltage? Does this change
in interfacial tension and contact angle of the liquid ox-
ide-refractory system change the rates of dissolution of

B. J. MONAGHAN ET AL.497
θ < 90°
Wetting & penetrating
θ > 90°
Non-wetting
Non-
p
enetratin
g
Figure 1. A sche matic showing the effects of liquid penetr a-
tion in a wetting and non-wetting system, where L = liquid
oxide, R = refractory and G = Gas.
refractories with extremely low porosity where increased
contact area generated by liquid oxide penetration into
pores is not an issue? If their findings are to be fully un-
derstood with respect to refractory and ceramic corrosion
then these question need to be addressed.
An answer to the first question may be found by con-
sidering the application of a number of existing theories
that have attempted to explain the composition stratifica-
tion in other systems such as metal-aqueous [4,5] solu-
tions, metal-polymers [6] and molten salts-metal [5,7,8]
systems under an applied voltage and the effects of this
change on the interfacial tension of the system. In these
systems it has been found that ions/molecules at or near
the phase interface order themselves to minimize the
voltage affects. This has the effect of changing the inter-
facial properties of the phase interface. A schematic of
this ordering process is shown in Figure 2, for a molten
salt system under an applied voltage. This ordering has
the effect of changing the interfacial properties of the
phase interface, shown as
 in Figure 2.
Highly basic liquid oxides are ionic in nature [9] and
to all intents and purposes can be considered as a molten
salt. Acidic liquid oxides are more covalent [9] and may
not be easily represented as molten salts. It is likely that
models, which describe the behavior of molten salts, will
have some applicability to liquid oxides in general and
basic liquid oxides in particular. Unfortunately, Kazakov
et al. [3] and Riaz et al. [1] investigations were on pene-
tration of liquid oxide into pores. It is difficult to define
the contact area between a porous refractory and the liq-
uid oxide in these experiments. This precludes direct
application of molten salt theories and hence an answer
to the question why should the interfacial tension and
contact angle of a liquid oxide-refractory system change
with changing applied voltage? The study of a simpler
refractory-liquid oxide system under an applied voltage
with a more dense refractory, (minimal porosity) and
therefore a better defined contact area may enable an
answer to the question. It may be expected, though it has
Solid
–
–
–
–+
+
+
+
+
+
+
+
+
+
Molten Salt
–
–
–
–
–
–
+
+
+
+
+
+
+
–
–
–
–
–
+
+
+
++
–
––
–
–
Ordered
Layer
Diffuse
Layer
Random
Layer
Solid
––
––
––
–– ++
++
+
+
+
+
+
++
++
++
Molten Salt
––
––
––
––
––
––
++
++
++
++
++
++
++
––
––
––
––
––
++
++
++
++ ++
––
–– ––
––
––
Ordered
Layer
Diffuse
Layer
Random
Layer

Voltage
0 -V +V

Voltage
0 -V +V

Voltage
0 -V +V

Voltage
0 -V +V
Solid
––
––
––
–– ++
++
+
+
+
+
+
++
++
++
Molten Salt
––
––
––
––
––
––
++
++
++
++
++
++
++
––
––
––
––
––
++
++
++
++ ++
––
–– ––
––
––
Ordered
Layer
Diffuse
Layer
Random
Layer
Solid
––
––
––
–– ++
++
+
+
+
+
+
++
++
++
Molten Salt
––
––
––
––
––
––
++
++
++
++
++
++
++
––
––
––
––
––
++
++
++
++ ++
––
–– ––
––
––
Ordered
Layer
Diffuse
Layer
Random
Layer
––
Ordered
Layer
Diffuse
Layer
Random
Layer

Voltage
0 -V +V

Voltage
0 -V +V

Voltage
0 -V +V

Voltage
0 -V +V
Figure 2. A sche matic of the orde ri ng in a molten salt under
an applied voltage and the resulting change in interfacial
tension . Circles with + and – denote an ion with the re-
spective polarity.
to be confirmed, that changes in interfacial tension and/or
wetting of the liquid oxide-refractory system due to an
applied voltage would also affect the dissolution rate of
more dense refractories. It is known that changing the
interfacial tension and/or wetting will also change the
interfacial (Marangoni) flow conditions [10] at the liquid
oxide-refractory-gas interface. Marangoni flows are in-
terfacial tension driven flows and are known to affect
corrosion at the liquid oxide-refractory-gas interface
[2,10]. Such a change in local flow conditions would
have a significant effect on refractory-liquid oxide dis-
solution systems that are at least in part controlled by
mass transport in the liquid oxide phase. The confirma-
tion that an applied voltage affects the dissolution be-
havior of the simpler dense refractory-liquid oxide sys-
tem is a necessary pre-curser to addressing more funda-
mental issues of why the interfacial tension and/or wet-
ting change with applied voltage. In this study the disso-
lution of a dense MgO refractory in contact with a CaO-
SiO2-Fe2O3-FeO-MgO liquid oxide under an applied
voltage has been studied and reported.
MgO was chosen as the refractory to be studied as it is
a widely used refractory in the liquid metal processing
industry and it is known that the dissolution of this re-
fractory in a CaO-SiO2-Fe2O3-FeO-MgO system is at
least in part controlled by mass transfer in the liquid ox-
ide phase [11,12] and therefore will be significantly af-
fected by Marangoni flows.
2. Experimental
The effect of an applied voltage across a solid dense
MgO-liquid oxide interface on the dissolution of rate of
MgO was studied. The composition of liquid oxides used
are given in Tabl e 1. A cylindrical piece of high density
MgO was dipped in liquid oxide for 90 or 120 minutes,
removed and water quenched. The resulting corroded
MgO sample was then sectioned, ground and polished to a
1 μm finish for microscopy using standard ceramographic

B. J. MONAGHAN ET AL. 498
Table 1. Liquid oxide compositions and dense MgO/liquid
oxide contact times used in this study .
Mass % CaO SiO2 Al2O3 MgO Fe2O3 FeO Time
(min)
Liquid
oxide 1 45.8 45.8 \ 3.0 4.6 0.990
Liquid
oxide 2 49.5 40.5 10 \ \ \ 120
methods. The experiments were carried out in a dried air
atmosphere and heated to 1540°C in a molybdenum dis-
ilicide vertical tube furnace. A schematic of the furnace
set-up is given in Figure 3. A schematic of the electrical
circuit used to apply the voltage is given in Figure 4. It
should be noted that the electrical cell in Figure 4 is
analogous to a capacitance cell and that there is no cur-
rent flowing.
The MgO dissolution was measured by measuring the
minimum diameter, Dmin, of the sample as defined in
Figure 5. This change in diameter is a result of the re-
fractory dissolution process, therefore the smaller Dmin,
the greater the amount of refractory dissolution/corrosion.
The cylindrical MgO test samples were supplied by Ro-
jan Advanced Ceramics, Henderson, Western Australia.
They are greater than 97% pure and had a density of in
excess of 96% theoretical.
The liquid oxide was prepared by pre-melting appro-
priate mixtures of CaO, SiO2, Fe2O3 and MgO in air,
quenching the liquid oxide and then crushing the resul-
tant glass. This process was repeated to obtain a ho-
mogenous liquid oxide. The crushed oxide was then
melted in-situ in the platinum crucible prior to running
an experiment. The reported oxide compositions in Ta-
ble 1 are based on ICP measurements for CaO, SiO2,
Water cooled end
MgO refractory
Slag
Heating
elements
Air in
Air ou
t
Water cooled end
Pt crucible
Gas tight
alumina
work tube
Water cooled end
MgO refractory
Slag
Heating
elements
Air in
Air ou
t
Water cooled end
Pt crucible
Gas tight
alumina
work tube
Figure 3. A schematic of the furnace set-up.
MgO refractory
Slag
Pt crucible
0.3mm
diameter Pt
wire
Power
supply
V
+-
Figure 4. A schematic of the electrical circuit used to apply
a voltage.
Before After
20mm diameter
25-30 mm
Dmin
Before After
20mm diameter
25-30 mm
Dmin
D
min
Figure 5. A schematic showing the MgO sample before and
after the dissolution experiment showing liquid oxide line
corrosion attack. Dmin is the minimum diameter of the sam-
ple after corrosion.
MgO, Al2O3 and total Fe. A back titration method was
used to establish the FeO content. The rest of the iron
was assumed to be Fe2O3.
The solid MgO was chosen to be studied as it is a
widely used refractory in the liquid metal processing
industry and it is known that the dissolution of this re-
fractory in a CaO-SiO2-Fe2O3-FeO-MgO system (Liquid
oxide 1, Table 1) is at least in part controlled by mass
transfer in the liquid oxide phase [11,12] and therefore
will be significantly affected by Marangoni flows. The
CaO-SiO2-Al2O3 liquid oxide (Liquid oxide 2, Table 1)
was chosen as there are published electrical capacitance
data [13] available for this system. These data would aid
future studies that assess models describing the ion be-
havior in liquid oxide systems under an applied voltage.
It was planned that the experimental immersion times
for all experiments in this study would be constant. Un-
fortunately after 90 minutes using liquid oxide 2 there
was minimal corrosion of the dense MgO. It was there-
fore the decided to increase the experimental time for
these experiments.

B. J. MONAGHAN ET AL.499
3. Results/Discussion
Experimental results for the effects of an applied voltage
on the dissolution/corrosion of dense MgO in liquid ox-
ides 1 and 2 are given in Figures 6 and 7 respectively.
From inspection of Figures 6 and 7 it can be seen that
an applied voltage has a significant effect on the meas-
ured Dmin when using liquid oxide 1 but not liquid oxide
2. At –0.5 V using liquid oxide 2 there is a slight lower-
ing of Dmin. It is not clear whether this lowering of Dmin
using liquid oxide 2 is significant given the uncertainty
in the data. The majority of the corrosion observed in this
study takes place at the MgO refractory-liquid oxide-gas
interface. At approximately –0.45 V the amount of MgO
corrosion using liquid oxide 1 and hence dissolution rate
is at a maximum.
Why should an applied voltage across the refractory-
liquid oxide interface affect the dissolution rate of the
MgO refractory? It is known that MgO dissolution in the
liquid oxide used in this study is at least in part con-
trolled by mass transfer in the liquid oxide phase [11,12].
6
8
10
12
14
16
18
20
-0.8 -0.6 -0.4 -0.20
Volta
g
e/V
Dmin/mm
D
min
/mm
Vo lta ge /V
Figure 6. The effects of an applied voltage on the dissolution
of dense MgO as characterized by Dmin for liquid oxide 1 at
1540°C for 90 minutes.
6
8
10
12
14
16
18
20
-0.6 -0.4 -0.200.20.4
Vol ta
g
e/V
Dmin/mm
Vo lta ge /V
D
min
/mm
Figure 7. The effects of an applied voltage on the dissolution
of dense MgO as characterized by Dmin for liquid oxide 2 at
1540°C for 120 minutes.
To discuss the effects of an applied voltage on a mass
transfer controlled process it is instructive to consider a
simple mass transfer model for steady state refractory
corrosion, as shown in Figure 8. In this figure A and B
represent the MgO in the refractory and liquid oxide
phase respectively, C is the molar concentration and i
represents interface. On the assumption that the dissolu-
tion process, as describe by Equation (2), is controlled
purely by mass transfer in the liquid oxide


MgOsolid MgOliquid oxide

(2)
then the MgO dissolution flux, J, into the liquid oxide
can be obtained from an application of Fick’s law (Equa-
tion (3)) to the refractory-liquid oxide system

21
..
i
MgO MgO
D
J
CC molecms


  (3)
where D is the interdiffusion coefficient, δ is the stagnant
(boundary) layer as defined in Figure 8, CMgO is the
MgO concentration in the bulk of the liquid oxide and
Ci
MgO is the MgO concentration in the liquid oxide at
reaction interface.
Factors that effect the velocity of the liquid oxide ad-
jacent the stagnant layer δ will effect the concentration
profile across the stagnant layer and hence the value of δ.
Any changes in δ will change the MgO dissolution rate
(flux J in Equation (3)). At the refractory-liquid oxide-gas
interface it is known that Marangoni flows are active and
play a significant part in the wear/corrosion profiles ob-
tained in the refractory at that interface [14]. These flows
are a result of shear stresses, δ, caused by interfacial ten-
sion gradients and can be described by Equation (4) [14]
ddTdCd
dxTdxC dxdx

 

 
 


(4)
Figure 8. A schematic of solid MgO dissolving in a liquid
oxide showing concentration profiles on either side of the
reaction interface.

B. J. MONAGHAN ET AL. 500
where x is distance, T is temperature, C is concentration
and ψ voltage. Other symbols are as defined previously.
For all other things being equal, a change in the voltage
term in Equation (4), as a result the applied voltage to the
liquid oxide-refractory system will be reflected by a
change in the Marangoni flow velocity. This in turn will
affect the MgO dissolution flux, as the thickness of the
stagnant layer δ in Equation (3) will change. The MgO
dissolution data as a function of applied voltage obtained
in this study are consistent with, and therefore could be
explained by such an argument. The argument that the
applied voltage in the dense refractory-liquid oxide sys-
tem studied in this investigation affects the interfacial
phenomena, and hence the dissolution rate is consistent
with Kazacov et al. [3] and Riaz et al. [1] explanation of
the liquid oxide penetration observations of their porous
refractory under an applied voltage. Essentially they ar-
gued that applying a voltage to their porous refrac-
tory-liquid oxide system changed the interfacial charac-
teristics of the liquid oxide penetration.
The argument for the voltage effect on MgO for dis-
solution using liquid oxide 1 is based on the assumption
that the MgO dissolution process is at least in part mass
transfer controlled. Therefore providing MgO dissolution
in a liquid oxide 2 is also at least in part mass transfer
controlled a similar voltage effect should be observed.
The results shown in Figure 7 do not bear this out. Why
there is this discrepancy is not clear and requires further
study. It may be that the expected voltage effect takes
place outside the range tested in this study, or that the
particular liquid oxide composition used and ions formed
are not as susceptible to and applied voltage effect, or
that the voltage effect is not as significant in liquid oxide
2 as liquid oxide 1 and therefore masked by the uncer-
tainty in the results, or that the dissolution process of
MgO in liquid oxide 2 is not controlled by mass transfer
in the liquid oxide phase.
It is unlikely that the dissolution process is not at least
in part mass transfer controlled in the liquid oxide phase
given previous work in this area [2,11,12], and more
specifically [12] on the same solid liquid system, showed
mass transfer control.
From the results of using liquid oxide 1 in this study
and the work of Kazacov et al. [3] and Riaz et al. [1] it
can be seen that the corrosion of a refractory may be al-
tered by applying a voltage across the refractory-liquid
oxide interface. This phenomenon offers the potential to
be able to control, among other things, refractory corro-
sion in liquid metal production, a significant cost to
many processes.
At present, it is not possible to predict with any cer-
tainty whether the applied voltage to the liquid ox-
ide-refractory system will lessen or increase the dissolu-
tion rate of the refractory. Nor is it known with any cer-
tainty if any of the current models that have been devel-
oped to describe the behavior of a solid-liquid interface
when a voltage is applied across the interface [4-8] are
applicable to liquid oxide-refractory systems. These
models describe how the ions stratify at the solid-liquid
interface under an applied voltage and that the stratifica-
tion is a function of temperature, liquid oxide composi-
tion (principally ion size) and liquid oxide capacitance.
Our understanding of ion sizes in steelmaking liquid
oxides is primarily limited to basic (depolymerised) liq-
uid oxides [9]. The ions present in acid liquid oxides, for
example liquid oxides rich in silica, are not well de-
scribed [9]. The lack of knowledge of acid liquid oxides
is not necessarily a major problem as most liquid oxides
used in liquid steel processing are basic.
Obtaining and utilizing liquid oxide capacitance data
are key issues that must be addressed if a solution to the
problems of direct application of the solid-liquid inter-
face models to refractory dissolution that account for
applied voltage effects are to be sought. There are very
little capacitance data for liquid oxides [13,15,16]. This
makes direct application of the solid-liquid interface
models difficult. Also, if these data are to be used to con-
firm which model is most appropriate for use in liquid
oxide-refractory systems then an accurate description of
the contact area between the liquid oxide and refractory
is required. This contact area is more easily described for
dense refractories. The confirmation that an applied vol-
tage across a refractory-liquid oxide interface for dense
MgO affects the MgO corrosion rate is an essential first
step in developing and understanding of this important
phenomenon.
4. Conclusions
The effect of an applied voltage across the solid-liquid
oxide interface on the dissolution rate of dense MgO
refractory has been studied for two liquid oxide compo-
sitions in the CaO-SiO2-Al2O3 and CaO-SiO2-Fe2O3-
FeO-MgO liquid oxide systems respectively. It was
found that the MgO corrosion in the CaO-SiO2-Al2O3
based liquid oxide proved insensitive to an applied volt-
age over the voltage range –0.5 to 0.3 V. The corrosion
rate of the dense MgO in the CaO-SiO2-Fe2O3-FeO-MgO
proved to be effected by an applied voltage, showing a
maximum corrosion rate at –0.45 V. This effect has been
explained by considering the consequences of an applied
voltage on the rate of Marangoni flow at the liquid ox-
ide-refractory-gas interface and in turn, the flow effect
on the rate of the mass transfer controlled MgO dissolu-
tion reaction.
The fact that the dissolution rate of a refractory can
be altered by the application of a voltage across the re-
fractory-liquid oxide interface offers a potential new tool
to control refractory losses in liquid metal processing.

B. J. MONAGHAN ET AL. 501
5. References
[1] S. Riaz, K. C. Mills and K. Bain, “Experimental Exami-
nation of Liquid Oxide/Refractory Interface,” Ironmaking
and Steelmaking, Vol. 29, No. 2, 2002, pp. 107-113.
[2] W. E. Lee and S. Zhang, “Melt Corrosion of Oxide and
Oxide-Carbon Refractories,” International Materials Re-
views, Vol. 44, No. 3, 1999, pp. 77-104.
[3] A. A. Kazakov, “Influence of the Electrical Potential on
Converter Smelting,” Russian Metallurgy, Vol. 6, 1997,
pp. 25-29.
[4] R. R. Dogonadze and Y. A. Chizmadzhev, “Structure and
Capacitance of the Metal-Fused Salt,” Interface Pro-
ceedings Academy of Sciences USSR Physical Chemical
Section, Vol. 157, No. 944, 1964, pp.778-781.
[5] E. A. Ukshe, N. G. Bukun, D. I. Leikis and A. N. Frum-
kin, “Investigation of the Electric Double Layer in Salt
Melts,” Electrochimica Acta, Vol. 9, No. 4, 1964, pp.
431-439.
[6] M. Vallet and B. Berge, “Electrowetting of Water and
Aqueous Solutions on Poly(Ethylene Terephthalate) In-
sulating Films,” Polymer, Vol. 37, No. 12, 1996, pp.
2465-2470.
[7] F. D. Richardson, “Physical Chemistry of Melts in Metal-
lurgy,” Academic Press, London, UK, 1974.
[8] A. D. Graves, G. J. Hill and D. Inman, “Electrode Proc-
esses in Molten Salts,” Advances in Electrochemistry and
Electrochemical Engineering, Vol. 4, No. 117, 1966, pp.
159-159.
[9] K. C. Mills, “Liquid Oxide Atlas,” 2nd Edition, Verlag
Stalheisen GmBH, Dusseldorf, Germany, 1995.
[10] N. Eustathopoulos, M. G. Nicholas and B. Drevet, “Wet-
tability at High Temperatures,” 12th Edition, Pergamon
Press, London, UK, 1999.
[11] B. J. Monaghan, S. A. Nightingale, L. Chen and G. A.
Brooks, “The Dissolution Behavior of Selected Oxides in
CaO-SiO2-Al2O3 Liquid Oxides,” VII International Con-
ferance on Molten Liquid oxides and Fluxes and Salts,
SAIMM, Johannesburg, South Africa, 2004, pp. 585-594.
[12] S. A. Nightingale, G. A. Brooks and B. J. Monaghan,
“Degradation of MgO Refractory in CaO-SiO2-MgO-
FeOx and CaO-SiO2-Al2O3-MgO-FeOx Liquid Oxides
under Forced Convection,” Metallurgical and Materials
Transactions B, Vol. 36B, No. 4, 2005, pp. 453-461.
[13] M. Ashizuka and K. Oe, “Measurement of the Electrical
Capacitance of the Double Layer at the Interface between
CaO-SiO2, CaO-SiO2-Al2O3 and CaO-Al2O3 Oxide
Melts,” Journal of the Japanese Institute of Metals, Vol.
39, No. 12, 1975, pp. 1243-1249.
[14] K. Mukai, “Wetting and Marangoni Effect in Iron and
Steelmaking Processes,” ISIJ International, Vol. 32, No.
1, 1992, pp. 19-25.
[15] M. Ashizuka and K. Oe, “On the Capacitance of Platinum
Electrode in Na2O-SiO2 Melts,” Journal of the Japanese
Institute of Metals, Vol. 39, No. 4, 1975, pp. 388-393.
[16] T. Emi, T. Sakuraya and K Sanbongi, “Electrochemical
Studies of the Interfacial Phenomena at Metal/Molten
Liquid Oxide Boundaries,” The 4th Japan–USSR Joint
Symposium on the Physical Chemistry of Metallurgical
Processes, The Iron and Steel Institute, Tokyo, Japan,
1973, pp. 72-86.