The adsorption of hematite particles on steel in strongly alkaline electrolyte

Article abstract of DOI:10.1149/1.3273198

The adsorption of hematite particles is an important but poorly understood elementary step in the electrolytic preparation of iron metal. In this work, a quantitative method of measuring the quantity of particles adsorbed onto a steel surface from a hematite suspension in concentrated sodium hydroxide electrolyte is proposed. The experimental procedure involves a preadsorption step in a hematite suspension and transfer to a particle-free solution, followed by a potentiostatic reduction of the film of adsorbed particles. The resulting current transient yields a quantitative analysis of the quantity of Fe ₂ O₃ adsorbed in the original solution. This work demonstrates that adsorption is only weakly dependent on the hematite particle concentration and pH (between 12 and 15). A simple phenomenological model for hematite reduction is proposed based on the form of the current-time transients.

Full text of DOI:10.1149/1.3273198

E24
Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
0013-4651/2010/157͑3͒/E24/7/$28.00 © The Electrochemical Society
The Adsorption of Hematite Particles on Steel in Strongly
Alkaline Electrolyte
,z
A. Allanore,a, J. Feng,b H. Lavelaine,a and K. Ogleb
*
aArcelorMittal R&D, F-57 283 Maizières-lès-Metz, France
bLaboratoire de Physico-Chimie des Surfaces, CNRS UMR7045, Ecole Nationale Supérieure de Chimie de
Paris, ParisTech, 75005 Paris, France
The adsorption of hematite particles is an important but poorly understood elementary step in the electrolytic preparation of iron
metal. In this work, a quantitative method of measuring the quantity of particles adsorbed onto a steel surface from a hematite
suspension in concentrated sodium hydroxide electrolyte is proposed. The experimental procedure involves a preadsorption step in
a hematite suspension and transfer to a particle-free solution, followed by a potentiostatic reduction of the film of adsorbed
particles. The resulting current transient yields a quantitative analysis of the quantity of Fe2O3 adsorbed in the original solution.
This work demonstrates that adsorption is only weakly dependent on the hematite particle concentration and pH ͑between 12 and
15͒. A simple phenomenological model for hematite reduction is proposed based on the form of the current–time transients.
© 2010 The Electrochemical Society. ͓DOI: 10.1149/1.3273198͔ All rights reserved.
Manuscript submitted September 16, 2009; revised manuscript received November 12, 2009. Published January 4, 2010.
Iron oxide electrolysis to produce iron metal has been known
since the pioneering work of Lloyd in 1929 and LeDuc et al. in
iron metal when pressed in contact with a polarized surface. The
reaction occurs via the local ͑microscopic scale͒ production of a
magnetite intermediate and the electrocrystallization of iron on the
particle surface. An investigation of iron solubility in alkaline
electrolyte7 showed that hematite bulk dissolution is low ͑2
ϫ 10−3 mol L−1͒ at 100°C and that the effect of the dissolved fer-
ric ions in the bulk electrolyte on the electrochemical reactions can
be neglected in the overall process. This latter study, however, dem-
onstrated that the final electrodeposition step could be obtained from
dissolved species in sufficiently reducing conditions, agreeing with
the physical theory of crystallization at low temperature and the
observation of deposits obtained in the suspension electrolysis
process.6
These results are important to establish an electrochemical
mechanism for the reaction, but the knowledge of the particle–
electrode interaction in the suspension process is still lacking. Two
limiting cases are possible for the contact mechanism of the hema-
tite reactant according to the residence time of adsorption, as pre-
sented in Fig. 1: ͑a͒ The particle collides with the cathode surface
such that only a fraction of oxide is converted to metal or ͑b͒ the
particle is adsorbed on the cathode with a residence time sufficient
for the entire particle to be reduced to metal.
Studies of hematite colloidal particle deposition on a stainless
steel substrate have been conducted in an electrolyte at a pH value
ranging from 2 to 12.9 The zeta potentials for hematite and stainless
steel at pH 12.8 ͑pH and ionic force being fixed by NaOH concen-
tration͒ have been evaluated at Ϫ48 and Ϫ40 mV, respectively, con-
firming the existence of an electrostatic repulsion between both sur-
faces in alkaline conditions. However, the use of an electrolyte at a
higher ionic strength enhances the deposition of colloids on surfaces
of a similar negative charge because of a reduction of the double-
layer thickness. Interactions governing the adhesion of the particle
to the substrate are therefore short-range forces, which may be
evaluated through the Delaguin–Landau–Verwey–Overbeek10
͑DLVO͒ theory. The deposition of particles can also be enhanced at
a higher temperature11 or in the presence of a magnetic field.12 The
deposition of particles with repulsive electrostatic interactions ͑simi-
lar charge͒ cannot be predicted accurately with existing theories
although models are under development, which, in addition, take
into account the effect of hydrodynamics.13 As a matter of fact, the
experimental results available do not concern the strongly alkaline
and high temperature range of the suspension electrolysis process
under evaluation and cannot be directly extrapolated to predict the
contact mode between the suspended particles and the electrode sur-
face.
1
1959.2 A process was patented by Offroy in 1974.3 However, as the
blast furnace is an efficient and mature technology, this work has
been very limited in scope and, in particular, the details of iron oxide
reduction have received very little attention. Recently, the race to
develop new steel-making processes with low greenhouse gas emis-
sion has led to a renewed interest in this technology,4,5 and small
scale trials have demonstrated a very high efficiency for direct re-
duction in alkaline solutions.6 Nevertheless, it is fair to say that a
fundamental understanding of the mechanism of iron ore reduction
is still limited. Indeed, to our knowledge, no similar process exists
involving the cathodic reduction of a suspension of oxide particles.
In general, the reduction of a suspended particle involves at least
two elementary kinetic steps: ͑i͒ the adsorption or adherence of the
hematite particle to the surface, followed by ͑ii͒ the electrochemical
reduction of the particle. Previous papers7,8 dealt with the second
phenomenon, the mechanism of reduction independent of particle
adsorption. In this paper, we make an attempt to investigate the
adsorption of the particles to the electrode surface.
The reduction may be carried out in a strongly alkaline aqueous
electrolyte ͑50 wt % NaOH–H2O at 100°C͒ in which finely ground
hematite ͑␣-Fe2O3, micrometer range͒ particles are suspended and
reduced to iron metal on the cathode.1-3 The overall net reduction
reaction may be written as
Fe2O3 + 3H2O + 6e− → 2Fe + 6 OH−
͓1͔
The suspension electrolysis process implies a reaction between a
polarized surface and an insulating oxide particle with the resulting
extraction of oxygen anions and the electrocrystallization of the
metal. Though the process efficiency is close to unity in an electro-
lyte highly concentrated in suspended hematite particles,2 the elec-
trochemical study of the reaction mechanism for the hematite con-
version to metal is hindered by the existence of the hydrogen
evolution reaction ͑HER͒ as a competing phenomenon on the cath-
ode
2H2O + 2e− → H2͑g͒ + 2 OH−
͓2͔
The evidence of the contact between the particles in suspension
and the cathode is difficult to establish because of the opacity of the
electrolyte due to the large amount of suspended matter ͑15 vol %͒.3
Two recent studies have been dedicated to the isolation of the
main steps of the reaction, namely, the role of dissolved species7 and
the reaction of a single particle on the cathode,8 where it was re-
vealed that a hematite particle can macroscopically be converted to
From these considerations, a simple experiment is proposed in
this work to assess the contact properties between the hematite par-
ticles and an iron cathode. The method involves an exposure of the
steel surface to the hematite particle suspension in a preliminary
*
Electrochemical Society Active Member.
z E-mail: allanore.antoine@free.fr
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
E25
(a)
(b)
Figure 1. ͑Color online͒ Possible particle/electrode interactions in the sus-
pension electrolysis process: ͑a͒ Collision and ͑b͒ adhesion of the particle on
the cathode.
Figure 3. ͑Color online͒ Steel surfaces used as cathode for ͑a͒ studying the
electrochemical potential effect and for ͑b͒ observations.
For the microscopic analysis, the mild steel substrate was em-
bedded in an epoxy resin holder to expose a square area of about
1 cm2 ͑Fig. 3b͒, polished with 1200 and 2400 grade emery papers
and thoroughly rinsed with distilled water before immersion. So-
dium hydroxide solutions were prepared by dissolving NaOH ͑Pro-
labo 97% purity, maximum 2.5% Na2CO3͒ in purified water ͑Milli-
pore system, resistivity 18.2 M⍀ m͒.
step, followed by an electrochemical characterization step where the
reduction of hematite to metal is obtained. This methodology is used
to characterize hematite particle adsorption as a function of suspen-
sion composition and exposure conditions.
Experimental
Materials.— Hematite particles from Prolabo ͑batch 0507825͒
with less than 1% impurities were used in this work. A particle size
distribution analysis ͑Malvern Mastersizer type S͒ revealed two size
distribution peaks at 0.4 and 10 ␮m, as illustrated in Fig. 2a and b.
A mild steel plate, typical of a packaging application, was used for
the adsorption substrate, as indicated in Fig. 3a.
Hematite adsorption.— The suspension was prepared from a so-
dium hydroxide and water solution at 110°C in a glass beaker ͑V
= 0.4 L͒, in which the desired amount of hematite particles had
been added. Concentrations in particles are given in the volume
fraction of hematite, which is readily calculated from the amount of
particles introduced ͑mp͒ and the hematite density ͑␳͒
mp 1
(a)
vol % =
100 = 47.5mp
7
␳ V
A magnetic stirring device, which also acted as a heating system,
was placed at the bottom of the vessel. Hematite adsorption was
carried out by immersing the electrode in the suspension for 3 min.
6
5
4
3
2
1
0
Electrochemical analysis step.— Particles adsorbed during the
adsorption step were detected and quantified by potentiostatic reduc-
tion in a separate particle-free electrolyte. Following the adsorption
step, the electrode was directly transferred to the analysis cell with-
out rinsing. The analysis electrolyte solution consisted of an un-
stirred, 50 wt % NaOH/water electrolyte at 110°C ͑pH25°C 15.3͒ in
a glass beaker ͑0.4 L͒. A new solution was used for each electrore-
duction experiment to guarantee replication. A large-area platinum
cylinder counter electrode and a HgO/Hg/KOH ͑10 M͒ reference
electrode were used, with the steel/adsorbed hematite electrode as
the cathode. A constant potential of Ϫ1.2 V vs Hg/HgO was applied
to the steel electrode by a Solartron 1286 potentiostat in chrono-
amperometric mode. Current transients were measured during the
cathodic polarization.
0.1
1
10
100
Particles Diameter / µm
(b)
The transfer of the electrode from the adsorption electrolyte to
the analysis electrolyte required approximately 3 s, and polarization
began approximately 5 s after the immersion in the analysis electro-
lyte. The surface was neither rinsed nor dried during the transfer;
therefore, it remained exposed to the alkaline solution throughout.
Nonadsorbed hematite particles, present in the electrolyte film after
the removal from the adsorption electrolyte, were in this way redis-
persed in the analysis electrolyte. Therefore, only the particles that
were strongly adsorbed on the surface were analyzed.
Scanning electron microscopy observation.— After the electro-
chemical reduction in the analysis step was achieved, the electrode
was removed from the analysis cell, rinsed with purified water, and
dried in an oven at 40°C under a N2 atmosphere. The electrode
surface was observed using a JEOL JSM-840 microscope. For each
20µm
Figure 2. ͑a͒ Particle size distribution and ͑b͒ SEM observation of the start-
ing hematite particles.
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E26
Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
0.10
0.08
0.06
0.04
0.02
4a.
4a. 15 vol.%
4b. no particles
4b.
2 µm
tP
tEND
30
Figure 5. Surface of the steel electrode after immersion in the suspension
and reduction at Ϫ1.2 V/HgHgO.
0
10
20
40
50
60
70
Time / s
Figure 4. ͑Color online͒ Current variation at Ecathodic = −1.2 V after immer-
sion ͑a͒ ᭺, in 15 vol % hematite suspension and ͑b͒ ᭝, in blank NaOH.
the current transient has returned to the background. A typical ex-
ample is shown in Fig. 5. The grains appearing on the surface have
been identified as pure iron metal by energy-dispersive X-ray analy-
sis ͑Ͼ90% Fe, Ͻ1% O͒. A detailed understanding of the current
transients in the first period is beyond the scope of this work, but
some comments are in order. The initial decrease during the first few
seconds cannot be attributed to double-layer charging, which has a
time constant of less than a second.14 This current might be attrib-
uted to the reduction of an oxidized layer on the steel plate. From
the determined charge ͑0.2 C for the blank and 0.4 C for hematite,
from raw current integrated between t = 0 and tp͒, the thickness of
the equivalent oxide layer ͑considered as hematite͒ is on the order of
100 nm, which agrees acceptably with the results obtained in such a
concentrated electrolyte and high temperature, where iron oxide
films form quickly.15 While this proposition requires additional stud-
ies for verification, it does set a lower limit to the thickness of oxide
films that can be measured with this technique.
The relatively slow increase in the current observed in the cur-
rent transient for the blank ͑Fig. 4b͒ after tp may be due to an
increase in the electrochemically active surface linked with the re-
moval of the insulating passivation layer. A larger concentration in
the reactant ͑water͒ provided by the fresh electrolyte might also be
responsible for such a behavior. This increase is much more pro-
nounced in the surface immersed in the suspension. The charge dif-
ference between tp and tend can be converted to the quantity of
Fe͑III͒ by Faraday’s law according to Reaction 1. The slow decay of
the current for both experiments after tend is commonly attributed to
the diffusion limitation for aqueous species ͑water or ions͒ in a
stagnant electrolysis configuration such as this.
observed sample, approximately 10 pictures were randomly re-
corded to represent the whole electrode surface from a statistical
viewpoint.
IFM characterization.— An interferometric microscope ͑IFM,
Microxam 100HR ADE Phase Shift͒ was used to obtain the topog-
raphy of the electrode surface.
Results and Discussion
The adsorption and reduction of hematite particles.— Figure 4
gives a typical example of the coulometric reduction of adsorbed
hematite particles ͓͑a͒, circles͔ after immersion in a 50 wt %
NaOH–H2O solution at 110°C for 3 min with 15 vol % of hematite.
A blank measurement ͓͑b͒, triangles͔ after immersion under identi-
cal conditions in a solution without hematite particles is also shown.
The analysis electrolyte was 50% NaOH at 110°C and was carried
out at an imposed potential of Ϫ1.2 V vs Hg/HgO. Following pre-
vious studies of hematite reduction,8 this potential is sufficient to
reduce the hematite particles while minimizing the hydrogen forma-
tion reaction.
At least three distinct kinetic regimes are observed in the current
transients whether or not hematite particle adsorption has occurred:
͑i͒ fast current decrease for approximately 1 s, ͑ii͒ transient current
regime where either a distinct peak or a slow decay with a shoulder
is observed, and ͑iii͒ steady-state current after 30 s until the end of
the experiment. As shown in Fig. 4, the effect of the adsorbed par-
ticles is essential in the second kinetic regime. Figure 4a and b
shows a start with a sharp decrease in the current density, which
ends after approximately 1 s ͑tp͒ although the initial current is in-
creased by a factor of 2 in the presence of adsorbed particles. Simi-
larly, the steady-state current obtained after approximately 30 s is
nearly identical for Fig. 4a and b. As the background current in the
blank electrolyte is no doubt due essentially to the HER reaction,
these results demonstrate that this reaction is not markedly affected
by the presence of the adsorbed hematite particles.
By contrast, the second kinetic period is markedly affected by the
presence of hematite particles. The blank ͑Fig. 4b͒ shows only a
slight increase in the current, obtaining a steady state after approxi-
mately 4 s. The current transient for the sample immersed in the
particle suspension shows a strong increase in current, passing
through the maximum value of 95 mA at about 7 s, forming a
well-defined peak, and then declines to a steady-state value, very
similar to that measured for the blank experiment at tend ͑28 s͒.
That the current maximum observed during the second period in
Fig. 4a is due to the reduction of the hematite particles has been
confirmed by scanning electron microscopy ͑SEM͒ performed after
Influence of hematite particle concentration.— The raw data of
current transients at Ϫ1.2 V vs Hg/HgO for various concentrations
of hematite particles in the suspension reactor are shown in Fig. 6a.
The current increases with the hematite concentration, though the
shape of the curves is not strictly identical throughout the trials. The
first seconds of current decrease noticed in the plate electrode stud-
ies ͑shown in Fig. 3a, current transient in Fig. 4͒ are not observed in
this configuration ͑electrode shown in Fig. 3b͒. To obtain an evalu-
ation of the overall reaction kinetics, the derivative of the current is
plotted as a function of time in Fig. 6b. Characteristic values are
taken at the inflexion of these curves for each concentration. The
corresponding reaction rate, presented in Fig. 7a, increases with the
hematite concentration and agrees with the assumption of a rate
limited by the amount of hematite particles. However, the reaction
rate reaches a plateau, suggesting the saturation of the surface by
iron oxide. To evaluate the charge related to the adsorbed hematite
particles, the steady-state background obtained for the blank has
been systematically subtracted. The overall charge obtained by inte-
grating the corresponding current–time transient is also presented in
Fig. 7a. It increases monotonically with the concentration of hema-
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
E27
(a)
0.20
0.15
0.10
0.05
0.00
0.3%
0.5%
1.2%
2%
2.9%
3.5%
4.3%
4.8%
Figure 8. Surface of the electrode after electrolysis for various volumic
fractions of hematite in suspension. Figures indicate the volumic fraction in
hematite. Scale bar: 10 ␮m.
0
20
40
60
80
100
120
particle concentration. The average size of the clusters is on the
order of a few micrometers, but their height can reach 10–30 ␮m.
The reason for such 3D growth requires further investigation, but
the heterogeneous aggregation of the colloidal hematite particles
seems a fair explanation. The movement of the particles on the
electrode surface occurs because of the H2 bubbles and other elec-
trohydrodynamic phenomena.16,17 As a matter of fact, such morphol-
ogy cannot be attributed to the conventional mechanisms of electro-
crystallization involving adions because the dimensions of the iron
clusters are much larger than the diffusion distance for adions.
The relationship between the concentration in particles and the
amount of metal deposit demonstrates that iron formation occurs by
a direct reduction of the hematite particles, suggesting the hematite
particles are adsorbed on the steel surface for a period of time suf-
ficient for total particle reduction to occur. The contact phenomena
involved are then of the type depicted in Fig. 1b. Such deposition of
negatively charged hematite particles on a negatively charged sur-
face is surprising in view of DLVO theory results, and hydrody-
namic aspects ͑kinetic energy of the particle͒ are invoked to justify
the overcoming of short-range repulsion forces.13
Time / s
(b)
0.000
-0.004
-0.008
-0.012
-0.016
-0.020
-0.024
-0.028
0.3%
0.5%
1.2%
2%
2.9%
3.5%
4.3%
4.8%
0
2
4
6
8
10
Time / s
Figure 6. Effect of the hematite volumic fraction in suspension on the ͑a͒
electrochemical signals and ͑b͒ corresponding derivative.
For the diluted suspensions, the amount of metal produced in-
creases with the hematite concentration, which agrees with a mass-
transfer limitation to the electrode. In concentrated suspensions
͑Ͼ5%͒, the saturation of the surface by the adsorbed particles is
observed, and the hematite concentration does not significantly af-
fect the amount of metal produced.
tite particles in the suspension reactor. A rough estimation of the
number of adsorbed particles as a function of hematite concentration
in the suspension may be obtained from the total charge by assum-
ing a particle diameter of 2 ␮m, as illustrated in Fig. 7b. Scanning
electron micrographs of the electrode surface after electrolysis in
suspensions with various hematite concentrations are presented in
Fig. 8. The amount of metal observed on the surface increases with
the concentration of hematite, but the surface coverage remains low.
The results suggest that the number of nucleation sites remains rela-
tively constant with hematite concentration, though the cluster size
increases markedly.
The topography of the electrode surface in such conditions is
presented in Fig. 9 where the increase in iron deposition is con-
firmed, especially up to a 6.5% volumic fraction of hematite. These
observations confirm the low surface coverage with metal and the
three-dimensional ͑3D͒ growth of the iron clusters with increasing
Influence of immersion time and hydroxide ion concentration.—
The effect of sodium hydroxide concentration in the suspension re-
actor has been studied in the 2.4 ϫ 10−2 to 18.6 mol L−1 range
͑corresponding to pH25°C 12.3–15.3͒ for 15 vol %. hematite. As
presented in Fig. 10a, no significant variation in the electrical charge
can be measured even if a tendency toward a larger extent of ad-
sorption at a higher pH can be observed. Further, no significant
effect of the immersion time in the suspension was observed, as
shown in Fig. 10b. These results imply that the stability of the sus-
pension against homogeneous aggregation is ensured even at pH
values as low as 12, which agrees with previous studies of hematite
suspensions in which the pH of zero charge was pHpzc ϳ 9.18 The
(a)
(b)
4.0x106
3.5x106
3.0x106
2.5x106
2.0x106
1.5x106
1.0x106
5.0x105
2.0
1.5
1.0
0.5
0.020
0.015
0.010
0.005
5
y= -33847+8.3226 10 x R²=0.997
x
Derivative
Figure 7. ͑Color online͒ ͑a͒ Variation in
the charge and initial derivative of current
density and ͑b͒ the number of 2 ␮m di-
ameter adsorbed particles with hematite
concentration.
Charge
0.1
1
10
0
1
2
3
4
5
Hematite Concentration / vol. %
Hematite Concentration / vol. %
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E28
Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
has been studied. The immersion step has been conducted for each
experiment in a concentrated suspension of hematite at 15 vol %.
Figure 11a presents the variation in the HER reaction for the
blank at different potentials, while Fig. 11b gives the variation ob-
tained in the presence of adsorbed particles. A strong variation in
both reactions is observed to occur with applied potential. The
steady-state HER reaction was measured as an average value near
the end of the experiment in Fig. 11. The initial rate of hematite
particle reduction is difficult to establish from the visual inspection
of the curves in Fig. 11; therefore, a simple model was developed,
which permits an interpretation of the entire data set after the sub-
traction of the background HER reaction. The model proposed for
the hematite particle reduction assumes an increase in the total cur-
rent due to particle reduction,8 linked to the hematite/electrode ac-
tive surface area increase under iron metal production. On the con-
trary, the background current is assumed constant throughout the
experiment, as exemplified in Fig. 4, where the very first current
values are the same as the final steady-state values.
Figure 9. ͑Color online͒ Observations of the electrode surface with optical
interferometry microscope. Numbers in bold indicate the volumic fraction in
hematite. ͑Observed surface of 410 ϫ 410 ␮m.͒
The total current is then the sum of two contributions
I = Ihem + Ibackground
the latest term being constant in the chronoamperometric experi-
ments, while the first one varies with hematite consumption.
Therefore, one can individually evaluate the hematite/iron metal
electrochemical reaction kinetic, thanks to the use of the derivative
of the current
steel surface is quickly saturated with hematite particles, and no
transient phenomena are detected on the time scale studied within
the experimental error.
The reduction of the adsorbed particles.— Although the focus
of this work is on the adsorption of the hematite particles during
immersion in the suspension reactor, it is of interest to investigate
the influence of the potential on the reduction reaction so as to better
understand the analytical process and perhaps shed light on the over-
all reduction mechanism. Using the steel plate as a cathode ͑Fig. 3a͒,
the effect of the polarization during the electrochemical analysis step
dI dIhem dIbackground
dShem
dt
=
+
= Khem͑V͒
dt
dt
dt
Evaluating the derivative of the current is then a simple method
to evaluate the variation in the reaction kinetic with the electrode
potential ͑Khem͒, assuming that the initial number of adsorbed he-
matite particles is constant between the experiments. The derivative
10
10
Figure 10. Influence of ͑a͒ sodium hy-
droxide concentration and ͑b͒ immersion
time ͓͑NaOH͔ = 20 mol L−1͒ in the sus-
pension on the electrochemical charge at-
tributed to iron production.
1
1
0.1
0.1
0
10
20
30
40
50
60
70
0.01
0.1
1
10
100
a - NaOH Concentration / mol.L-1
b- Immersion Time / s
-1.28V
-1.25V
-1.23V
-1.18V
-1.15V
-1.13V
-1.10 V
-1.28V
1
-1.25V
-1.23V
-1.18V
-1.15V
-1.13V
-1.10 V
1
no particles
with particles
0.1
0.1
Figure 11. Current transients for various
potentials at 110°C and 50% NaOH in the
͑a͒ absence of hematite particles and in
the ͑b͒ presence of adsorbed particles after
immersion in 15 vol % of hematite, as de-
scribed in the text.
0.01
0.01
0
20
40
60
80
100
0
20
40
60
80
100
Time / s
Time / s
(a)
(b)
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
E29
0.16
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
-1.28 V
log I - no particles
0.14
127 mV/dec
-1.23 V
0.12
0.10
0.08
0.06
0.04
0.02
0.00
log dI/dt - with particles
186 mV/dec
-1.18 V
-1.10 V
59 mV/dec
-1.15
0
20
40
60
80
100
120
-1.30
-1.25
-1.20
-1.10
-1.05
E / V
Time / s
Figure 12. ͑Color online͒ Tafel plot from current densities determined in
Fig. 11 for HER ͑current, ᮀ͒ and adsorbed particles ͑derivative of the cur-
rent, ᭺͒.
Figure 13. ͑Color online͒ Current variation with time at various potentials.
Experimental data are shown as points; the solid curve is a fit to Eq. 5.
M
r = ro − kFet
␳
has been evaluated at the inflection point of curves presented in Fig.
11b. The full derivation of such a model is given in the next section,
leading to a current density value solely linked with hematite par-
ticle reduction.
The fraction of the current for each particle is calculated from
ip = kFe4␲r2
In Fig. 12, a Tafel slope of 127 mV/dec is obtained for the HER
reaction, which may be compared with that of Ref. 19. By contrast,
two Tafel slopes of 59 and 186 mV/dec are obtained for particle
reduction at low and high current densities, respectively. The value
at a low current density could be related to one electron transfer
limiting step, but there may be synergistic effects between the HER
and the particle reduction because particle movement on the surface
may be associated with the hydrogen gas formation, as mentioned in
the previous section. Nevertheless, the fact that particle reduction
and hydrogen formation show different Tafel slopes demonstrates
that particle reduction is not controlled entirely by surface particle
movement. Also, by comparison, with the commonly assumed
mechanisms for high temperature formation and reduction of oxides,
it might seem reasonable to propose oxygen ion diffusion within the
solid as the rate limiting step. However, the low diffusivity of oxy-
gen ions at this temperature range makes this highly unlikely for the
rapid reaction rates observed here.
or
2
M
i = k 4␲ r − k t
͓4͔
͓5͔
ͩ
ͪ
p
Fe
o
Fe
␳
And, the overall total current is I = Npip
2
M
I = N k 4␲ r − k t
ͩ
ͪ
p
Fe
o
Fe
␳
The assumption of a homogeneous spherical geometry is obviously
not correct; however, it provides an ideal situation by which we can
compare the experimental data. The assumption that hematite reduc-
tion begins on the outer surface of the particle is consistent with
previous experiments.8
The adsorption step was the same for all the data in Fig. 11.
Therefore, within a statistical variation, we can assume that Np is
constant for these experiments. The data in Fig. 11, after the removal
of the contribution of background processes occurring during the
first few seconds and the steady-state HER reaction, have been ana-
lyzed using Eq. 5, assuming a constant number of particles Np
= 2 ϫ 106, with the radius ͑ro͒ and the electrochemical kinetic co-
efficient ͑kFe͒ as adjustable variables. The comparison between the
model and the experimental results is presented in Fig. 13, with
corresponding parameters presented in Table I. The model fairly
Phenomenological model for particle reduction.— In this sec-
tion, we propose a simple model to account for the contribution of
adsorbed hematite particle reduction to the total current transient. A
truly predictive model is, of course, beyond the scope of this work;
however, even a simple model makes a useful point of comparison
for the experimental data. We consider the surface to be partially
covered with Np hematite spherical particles of initial radius, ro, and
the particles are reduced from their outer surface toward their center
with an electrochemical kinetic coefficient kFe. The electroreduction
of hematite via Reaction 1 is conducted at the jFe current density
Table I. Experimental conditions and modeling results from the
current variation in Fig. 11.
jFe
kFe
=
͑mol m−2 s−1
͒
͓3͔
6F
E
tmax
͑s͒
r0
͑m͒
jFe
on the surface of the particle ͑Sp = 4␲r2͒. Faraday’s law predicts the
variation in the particle volume as follows
͑V/HgHgO͒
͑A m−2
͒
Ϫ1.1
169.6
75
55.33
37.5
33
1.9 ϫ 10−06
2.1 ϫ 10−06
2.3 ϫ 10−06
2.0 ϫ 10−06
2.2 ϫ 10−06
2.0 ϫ 10−06
1.9 ϫ 10−06
197
594
700
1000
1200
1450
1930
M
Ϫ1.15
Ϫ1.18
Ϫ1.2
Ϫ1.23
Ϫ1.25
Ϫ1.28
dV = − kFe4␲r2dt
␳
dV = 4␲r2dr
24.45
17
so that after integration
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).

E30
Journal of The Electrochemical Society, 157 ͑3͒ E24-E30 ͑2010͒
1.0x10-8
8.0x10-9
6.0x10-9
4.0x10-9
2.0x10-9
Lefevre ͑Laboratoire d’Electrochimie, Chimie des Interfaces et
Modélisation pour l’Energie, UMR 7575, CNRS, ENSCP͒ for the
useful discussions. This work was conducted within the Ulcos pro-
gram supported by grants from the European Sixth FP.
y = 5.5x x R²=0.99
10-11
ArcelorMittal Maizières Research assisted in meeting the publication
costs of this article.
List of Symbols
bC
F
Tafel slope, bC = 2.303RT/neF = 0.076 in V/decade for one elec-
tron transfer at 383 K
−1
Faraday’s constant, 96,485 C mol
e−
ip current resulting from the reduction of a single particle, A
total current attributed to adsorbed particles reduction, A
I
jFe current density for the hematite to metal reduction in Reaction 1,
A m−2
0
40
80
120
160
electrochemical kinetic coefficient, mol m−2 s−1
kFe
M
hematite molar mass, 0.1597 kg mol−1
t
/ s
max
Np number of particles adsorbed on the surface
ro radius of the adsorbed particles before reduction, m
Figure 14. ͑Color online͒ Variation in the ro/jFe fitted value from Fig. 11 as
a function of time tmax
.
Greek
hematite specific gravity,20 5260 kg m−3
␳
agrees with the experimental results. Note that from Eq. 5, the ro/jFe
ratio should be proportional to the time ͑tmax͒, at which the current
reaches zero, the proportionality constant being
References
1. S. J. Lloyd, Trans. Am. Electrochem. Soc., 55, 305 ͑1929͒.
2. J. A. M. LeDuc, R. E. Loftfield, and L. E. Vaaler, J. Electrochem. Soc., 106, 659
͑1959͒.
M 1
= 5.3 ϫ 10−11 m3 C−1
␳ 6F
3. J. Offroy, French Pat. FR 2 287 531 ͑1974͒.
4. J.-P. Birat, J. Borlée, B. Korthas, J. van der Stel, K. Meijer, C. Günther, M. Halin,
T. Bürgler, H. Lavelaine, and C. Treadgold, in The Third International Conference
on Process Development in Iron and Steelmaking, SCANMET III, Luleå, Sweden
͑2008͒.
This is also verified by experimental results, as shown in Fig. 14.
Conclusion
5. www.ulcos.org, last accessed December 2009.
6. B. Yuan, O. E. Kongstein, and G. M. Haarberg, J. Electrochem. Soc., 156, D64
͑2009͒.
The stability of the hematite particles in highly concentrated al-
kaline solution limits the aggregation of such iron oxides in suspen-
sion. Physical chemical interactions of such two-phase media en-
hance the wetting of a steel surface so that particles adhere to the
substrate. By the electroreduction of hematite to metal in a separate
analysis cell, the corresponding amount of adsorbed particles can be
evaluated. The trapped particles have a characteristic length of less
than 2 ␮m, and a 3D aggregation of iron metal on the steel surface
is observed. Compared with the hematite concentration, the influ-
ence of the base electrolyte concentration or immersion time on such
process indicates that bulk electrolyte/particle interactions are of
second order in determining the intensity of adhesion phenomena. A
preliminary analysis of the particle reduction rate reveals Tafel-like
kinetics for the electroreduction of iron oxide to iron.
7. A. Allanore, H. Lavelaine, G. Valentin, J. P. Birat, and F. Lapicque, J. Electrochem.
Soc., 154, E187 ͑2007͒.
8. A. Allanore, H. Lavelaine, G. Valentin, J. P. Birat, and F. Lapicque, J. Electrochem.
Soc., 155, E125 ͑2008͒.
9. R. J. Kuo and E. Matijevic, J. Colloid Interface Sci., 78, 407 ͑1980͒.
10. W. B. Russel, D. A. Saville, and W. R. Schowalter, Colloidal Dispersions, Cam-
bridge University Press, Cambridge ͑1989͒.
11. N. P. Ryde and E. Matijevic, Colloids Surf., A, 165, 59 ͑2000͒.
12. M. F. Haque, N. Kallay, V. Priman, and E. Matijevic, J. Colloid Interface Sci., 137,
36 ͑1990͒.
13. L. Čerović, G. Lefevre, A. Jaubertie, M. Fedorov, and S. Milonjic, J. Colloid
Interface Sci., 330, 284 ͑2009͒.
14. A. J. Bard and L. R. Faulkner, Electrochemical Methods, Fundamentals and Ap-
plications, John Wiley & Sons, New York ͑1980͒.
15. J. E. Castle and H. G. Masterson, Corros. Sci., 6, 93 ͑1966͒.
16. M. Trau, D. A. Saville, and I. A. Aksay, Langmuir, 13, 6375 ͑1997͒.
17. R. M. Penner, J. Phys. Chem. B, 106, 3339 ͑2002͒.
18. T. Hiemstra and W. H. Van Riemsdijk, Langmuir, 15, 8045 ͑1999͒.
19. M. R. G. de Chialvo, J. Electroanal. Chem., 372, 209 ͑1994͒.
20. R. M. Cornell and U. Schwetmann, The Iron Oxides: Structure, Properties, Occur-
rences and Uses, Wiley-VCH, Weinheim ͑2003͒.
Acknowledgments
The authors thank D. Bettinger and J. Mongin ͑ArcelorMittal͒
for the interferometry measurements. They are also grateful to G.
Downloaded on 2014-10-01 to IP 130.132.123.28 address. Redistribution subject to ECS terms of use (see ecsdl.org/site/terms_use) unless CC License in place (see abstract).