Electro-reduction of hematite using water as the redox mediator

Article abstract of DOI:10.1039/c8gc02756k

The primary production of iron from its oxides, about 1.6 billion tons per annum, alone is several times greater than the production capacity of all other metals together. This massive production is conventionally achieved by carbothermal reduction causing several million tons of greenhouse CO₂ emissions every year, with catastrophic environmental consequences. Here, we present a green, efficient and economical molten salt route for the electrolytic production of iron from iron oxide (hematite, Fe₂O₃). The dissolution of water in molten LiCl causes the formation of hydrogen cations in the melt, which can be discharged on cathodically polarized solid iron oxide pellets immersed in the melt, leading to the reduction of Fe₂O₃ of metallic iron and the regeneration of water. The Fe₂O₃ pellet was thoroughly reduced to iron under 1.4 V within 5 h, considerably faster and more energy efficient than the current electro-deoxidation methods available. The feasibility of the water induced electrochemical reduction of metal oxides is confirmed using Fe₂O₃, but this novel methodology should be applicable to a broad range of metal oxides.

Full text of DOI:10.1039/c8gc02756k

Green Chemistry
View Article Online
View Journal
COMMUNICATION
Electro-reduction of hematite using water as the
redox mediator†
Cite this: DOI: 10.1039/c8gc02756k
a
a,b
Kaiyu Xie
and Ali Reza Kamali
*
Received 30th August 2018,
Accepted 27th November 2018
DOI: 10.1039/c8gc02756k
rsc.li/greenchem
The primary production of iron from its oxides, about 1.6 billion
So far, a number of electrometallurgical routes have been
tons per annum, alone is several times greater than the production proposed for direct production of iron, which can be divided
2
–5
capacity of all other metals together. This massive production is into four groups as follows: (1) molten oxide electrolysis, (2)
6
–8
conventionally achieved by carbothermal reduction causing electrowinning of solid iron from aqueous solutions,
several million tons of greenhouse CO emissions every year, with electrodeposition of iron in high temperature molten salts,
catastrophic environmental consequences. Here, we present a and (4) direct electro-deoxidation of solid iron oxides in
(3)
9
–12
2
1
3–17
green, efficient and economical molten salt route for the electro- molten salts.
Molten oxide electrolysis (MOE) for iron
lytic production of iron from iron oxide (hematite, Fe ). The dis- extraction suggests the idea of using an electrochemical cell
2 3
O
solution of water in molten LiCl causes the formation of hydrogen similar to that used in the Hall–Héroult cell for the production
cations in the melt, which can be discharged on cathodically of aluminum. In this technique, a supporting molten oxide
polarized solid iron oxide pellets immersed in the melt, leading to electrolyte with satisfactory ionic conductivity, composed of
the reduction of Fe
water. The Fe pellet was thoroughly reduced to iron under amounts of iron oxide can be dissolved in the molten oxide
.4 V within 5 h, considerably faster and more energy efficient than electrolyte to form iron cations Fe which can be electrolysed
2 3
2 2 3
O of metallic iron and the regeneration of multiple oxides (SiO , CaO, MgO, and Al O ), is used. Small
2 3
O
3
+
1
2
3
4
the current electro-deoxidation methods available. The feasibility using Cr90Fe10 alloy, iridium, and graphite anode materials
of the water induced electrochemical reduction of metal oxides is to produce iron at the cathode. An optimal cathodic current
2 3
confirmed using Fe O , but this novel methodology should be efficiency of around 46% was achieved during the galvanostatic
4
applicable to a broad range of metal oxides.
electrolysis with a current of 0.85 A for 2 hours. The cathodic
electrochemical reduction was found to proceed in two steps,
The development of energy efficient, carbon-free and green
routes for the production of iron and steel is of great techno-
logical and environmental importance, particularly in
response to the urgent requirement for the reduction of either
high greenhouse gas CO2 emission, or the extensive energy
consumption involved in the traditional carbothermal
methods of iron and steel making. In comparison with the
non-sustainable carbon reductant used in the preparation of
metallic iron in blast furnace plants, electricity can be con-
sidered as a renewable and green reductant for the future iron-
5
i.e., from iron(+III) to iron(+II) and then to zero-valence Fe.
Although the continuous operation of this process might be
possible, its complexity, low energy efficiency, and high operat-
ing temperature of up to 1600 °C as well as the highly corrosive
nature of the molten oxide electrolytes used provide critical
challenges for the large-scale operation of this approach.
Electrowinning of iron in caustic alkaline solutions is an
alternative way with a greater level of energy efficiency. In this
approach, the deposition of iron can be achieved using a rotat-
ing graphite disk cathode immersed in a 50 wt% NaOH solu-
1
making industry.
2 3
tion containing 33 wt% Fe O at 114 °C, providing a current
7
efficiency in the range 90–95%. To achieve this, a cell voltage
from 1.4 to 1.9 V should be applied between a graphite or
copper cathode and a nickel or platinum anode for a period of
over 3 hours. The insufficient mass transportation behavior in
aqueous solutions, the corrosive nature of the electrolytes and
also the high oxygen content of the iron product (up to
a
2
Energy and Environmental Materials Research Centre (E MC), School of Metallurgy,
Northeastern University, Shenyang 110819, China
b
Department of Materials Science and Metallurgy, University of Cambridge,
Cambridge, CB3 0FS, UK. E-mail: ali@smm.neu.edu.cn
7
,8
2
wt% ), however, may suppress the viability of this approach
†
Electronic supplementary information (ESI) available: Experimental infor-
mation and supplementary figures. See DOI: 10.1039/c8gc02756k
for large-scale operations.
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Communication
Green Chemistry
Solid iron can also be electrodeposited from Fe
2
O
3
or Fe
3
O
4
the system, can enhance the economic viability of the electro-
9
,10
11
containing molten salts such as Li CO ,
CaCl -KF and chemical extraction of iron.
2
2
3
1
2
CaCl
2
-CaF
2
at temperatures above 800 °C by galvanostatic
We have recently reported that water can be dissolved in
1
8,19
electrolysis methods. The limited solubility of iron oxides in molten LiCl to produce hydrogen cations in the melt.
the above-mentioned electrolytes and the difficulty in separ- Here, we propose that the protons formed in molten LiCl can
ating the solid dendritic iron product from the solidified salt be discharged on cathodically polarised metal oxides
are the predominant disadvantages of this process. As an excit- immersed in the melt to form atomic and molecular hydrogen
ing alternative method, the direct electro-deoxidation of solid leading to the in situ reduction of the metal oxides into the
iron oxides to solid iron in molten salts has received an exten- corresponding metals under a relatively low cell potential in a
sive attention. In this approach, the electro-deoxidation of short time. Given that hydrogen is considered as a clean redu-
Fe O pellets employed as the cathode in molten CaCl occurred cing agent, our approach provides a green and economical way
2
3
2
2
0–22
at 800 °C under an applied cell voltage of around 2.2 V with a for the solid-state carbon-free reduction of metal oxides.
current efficiency of higher than 90% and an energy consump- Amongst all metals, iron possesses the largest production
tion around 3.0 kW h per kilogram of Fe. Iron was also elec- capacity (>1.6 billion tons of crude steel in 2017) with the most
trochemically extracted at 750 °C in Na
melts using a Fe pellet cathode under an applied cell voltage this investigation, hematite (Fe
of 1.9–2.3 V with an optimal cathodic current efficiency of 95% to explore the feasibility of this novel process and also assess
and an energy consumption of 2.87 kW h per kilogram of Fe. the achievable production yield. For this, Fe pellets were
The carbon content of the iron product was found to be in the reduced by cathodically generated hydrogen in molten LiCl. To
1
3
2
3
2
CO
-K
3 2
CO
3
eutectic extensive technological contribution to modern society. In
O
2 3
2 3
O
) was preferentially selected
2 3
O
14
range of 0.035–0.76 wt%. The electro-deoxidation of Fe O
pellets could also be operated in molten sodium hydroxide
this end, a constant cell voltage of 1–1.4 V was applied
between an assembled iron oxide cathode and a graphite cruci-
2
3
1
5,16
at above 500 °C under a cell voltage of 1.7 V, and in molten ble anode over a period of 1–5 h at 660 °C in molten LiCl
LiF-NaF-1 wt% Li O at 750 °C, producing a carbon-free iron under a humid argon atmosphere. The experimental setup
2
17
product with a coral-like morphology. It should be mentioned used in this study is shown in Fig. 1 and described in the ESI.†
that the large-scale production of electrolytic iron, despite the The XRD patterns of the cathode materials after being used
worldwide attention, is limited by some critical challenges in the molten salt electrochemical process under different con-
mostly related to the economic barriers and/or the complexity ditions are shown in Fig. 2. As can be observed, after 1 h of
of the equipment/electrolyte required. Therefore, reducing the electrolysis at a cell voltage of 1 V, all the diffraction peaks of
value of essential processing parameters, especially the applied iron oxide disappeared, indicating a complete phase conver-
2 3 5
cell voltage and the processing time as well as the complexity of sion to Li Fe O . At the same cell voltage of 1 V but a pro-
2 3
Fig. 1 Schematic of the experimental setup and the mechanism involved in the electrochemical process. As shown in the left panel, a Fe O pellet
immersed in molten LiCl was used as the cathode in the electrolysis process under a humid argon atmosphere. Water is carried into the reactor by an
argon gas flow, and dissolved in the molten salt to form hydrogen cations (protons) and oxygen anions in the melt. As shown in the right panel, the
reduction of iron oxide occurs in three stages: stage I involves the formation of Li
reduction of the cathodically polarised Fe . Stage II involves the reduction of Li Fe
process. Stage III involves the complete metallisation of the cathode, and the subsequent evolution of hydrogen on reduced iron particles.
2
Fe
3
O
5
intermediate phase through the electrochemical hydrogen
O
2 3
2
3 5
O to metallic iron and the inward propagation of the metallisation
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Communication
Fig. 2 XRD patterns of the cathodic products obtained under the indi-
cated conditions.
longed electrolysis time of 5 h, the product mainly consisted of
Li Fe O , but two small new peaks at 2θ = 44.7° and 82.4° were
2 3 5
also detected which can be assigned to the (110) and (211)
diffraction reflections of metallic iron, respectively. We should
Fig. 3 Morphological and chemical analysis of the cathodic products.
(
a) Optical photographs of the Fe
the electrolysis process at 1.4 V for 5 h; (b) SEM micrograph of the sin-
tered Fe pellet; (c–e) SEM images of the cathodic products produced
by conducting electrolysis at: (c) 1 V for 1 h, (d) surface layer at 1.4 V for
h, (e) the core section at 1.4 V for 5 h; (f) EDS spectrum from the
corresponding region marked in (e).
2 3
O pellet before (left) and after (right)
2 3
O
consider that the theoretical decomposition potential of Fe
2 3
O
1
at 660 °C under standard conditions can be calculated to be
1
.003 V (FactSage).
Fundamentally, however, all electrochemical reactions
require an actual potential higher than the thermodynamic
value to proceed, due to factors such as polarisation and LiCl at 660 °C for 1 h under a humid argon atmosphere
ohmic losses. The results obtained here, therefore, are interest- without supplying any electric charge. After cooling down the
ing, demonstrating that the Fe O precursors could be reduced reactor and retrieving the pellet, XRD analysis was conducted
2
3
to Fe under such a low applied voltage of 1.0 V, opening a to detect any possible phase change, which could occur during
promising window for low potential electrochemical reduction the heat treatment. The XRD pattern of the heat-treated
of iron oxides.
The product obtained by performing electrolysis for only pellet (Fig. S1†) and no detectable phase transition was found.
h at a cell voltage of 1.4 V was found to possess distinct The formation of Li Fe observed can, therefore, be attribu-
material obtained was compared with that of the initial Fe O
2 3
1
2
3 5
O
surface and core layers. The dense external surface layer with a ted to the electrochemical reaction occurring during the elec-
thickness of approximately 0.3 mm was retrieved from the trolysis. Owing to the multivalent nature of iron oxide and to
cathode pellet and found to be predominantly constituted of maintain the electroneutrality of the electrolyte during the
+
metallic iron. The internal core part, however, was identified electrolysis, Li cations from the LiCl melt can migrate towards
as Li
2
Fe
3
O
5
, as can be seen in Fig. 2. After conducting electro- the negatively charged Fe , leading to the electrochemical
2
O
3
lysis for 5 h at 1.4 V, whilst the external geometric dimension formation of the Li
2
3 5
Fe O intermediate phase. The same
of the retrieved pellets didn’t show a considerable change in phenomenon was also observed during the electro-deoxidation
2
4,25
comparison with the initial pellets, the grey color metallic of solid TiO
appearance of the material was clear, as evidenced in Fig. 3a. eutectic melt and TiO
2
in molten CaCl
2
,
Fe
2
O
3
in the Na
2 3 2 3
CO -K CO
2
6
27
2
in LiCl, resulting in the formation
The XRD analysis demonstrated that the oxide pellet was of Ca–Ti–O, Na–Fe–O, and Li–Ti–O intermediate phases,
thoroughly reduced to pure iron. From the XRD results, it is respectively.
evident that an enhanced electrochemical reduction rate could
be achieved at larger applied potentials, which can contribute after the electrolysis process at 1.4 V for 5 h, and the SEM mor-
to the promoted discharge of protons on the cathode, provid- phology of the sintered Fe pellet and those of the products
ing a stronger driving force for the in situ generation of obtained under different indicated conditions. As can be
hydrogen. depicted, the particle size of the sintered pellets was around
Fig. 3 shows the optical images of the pellet before and
2 3
O
In order to comprehensively study the changes occurring in 200–500 nm. The SEM morphology of the product produced
the pellet during the electrolysis process and therefore the under 1 V for 1 h (Fig. 3c) demonstrates the presence of uni-
reduction pathway, the Fe O pellet was immersed in molten formly formed octahedral crystals with a size of around
2
3
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Communication
Green Chemistry
1
–5 μm, considerably larger than the Fe
2 3
O particles in the sin-
2
5,28
tered pellet. Based on the previous studies
on the electro-
deoxidation of solid oxide precursors and also the common
knowledge of the solid-state electrochemical reduction, it can
be assumed that, in our case, the reduction process has firstly
occurred on the surface of the pellet and then gradually devel-
oped into the pellet core. The SEM image obtained from the
surface layer of the sample produced under 1.4 V for 1 h
(Fig. 3d) displays a cluster-shaped porous structure with par-
tially interconnected nodular reduced iron particles, with
respect to the XRD results of Fig. 2. Moreover, some distin-
guished octahedral particles could also be observed around
the iron particles. Fig. 3e shows the SEM morphology of the
product produced by conducting the electrolysis process at 1.4
V for 5 h. The EDX analysis results such as those shown in
2 3
Fig. 3e revealed that the Fe O precursor was completely
reduced to clusters of iron particles with irregular shapes and
sizes in the range of a few to more than 10 micrometers. By
comparing the morphology of the product produced at 1.4 V
for 1 h and 5 h (Fig. 3d and e), one can realise that prolonging
the reduction time will eventually result in the agglomeration
of interconnected metal particles and the enhancement of the
sintering effect of the iron particles. It is noteworthy that the
electrolyte would penetrate into the generated pores and inter-
connecting channels between the particles, and hence the
reaction boundary was continuously propagated and more
electrochemically active sites were formed. As a result, the
charge transfer kinetics and mass diffusion phenomena were
kinetically enhanced, thereby promoting the bulk reduction of
Fig. 4 Electrochemical measurements. (a) The changes of cell current
and the hydrogen concentration in the outlet gas flow versus the elec-
trolysis time achieved by conducting electrolysis at 1.4 V. (b) The poten-
tial differences between the cathode/anode and the Mo quasi-reference
electrode.
2 3
Fe O .
The typical current–time curve which could be obtained
during the electrolysis is plotted in Fig. 4a. It is acknowledged
that the electric current obtained is in fact the response of
the electrode reactions occurred driven by the defined con-
stant voltage applied between the cathode and the anode.
With this in mind, based on the different regimes observed of the interior part of the pellet are much more sluggish than
in the current–time plot, the electrolysis process can be con- the surface layer.
sidered to proceed in five different steps as shown in Fig. 4a.
Fig. 4a also shows the concentration of hydrogen in the cell
Step I, corresponding to the beginning period of the electroly- outlet gas stream against the electrolysis time. As can be
sis process, can be characterised by a high electric current depicted, during the first hour of the electrolysis, the curve
which even further increased to a maximum value of 1.385 A exhibits a plateau at a hydrogen concentration value below
within the initial 2 min upon the application of the electric 0.05%, which then gradually increases up to around 0.6% in
potential, indicating that the reduction process occurred pre- the next two hours. Subsequently, the hydrogen concentration
ferentially in the area of the pellet nearby or connected to the reaches a second plateau lasting over an hour. The latter,
surrounding Ni wire. Afterwards, in step II, the current which corresponds to steps III and IV of the current–time
decreased down to 0.13 A in the next 33 min. It may corres- curve, can be associated with the completion of the reduction
pond to the complete metallisation of the cathode pellet at process. From Fig. 4a, it can be further noted that the concen-
the surface. Then, the current first increased to around 0.75 A tration of hydrogen increased almost linearly with time from
within 35–186 min, and then decreased to 0.23 A within the fourth to fifth hours of electrolysis, attributable to the pro-
2
8 min, corresponding to steps III and IV of the process as gress of the reduction and metallisation process. It is because,
presented in Fig. 4a, respectively. These two stages can be by converting a higher fraction of the iron oxide cathode to
attributed to the development of the reduction process metallic iron, a higher quantity of electro-generated hydrogen
towards the interior area of the pellet. Therefore, the can be released to the atmosphere, instead of being consumed
reduction can be assumed to be complete at the end of stage in the reduction process. It should also be mentioned that the
IV. Furthermore, the derived peak-profile observed in these actual cathode surface area becomes larger and therefore a
two stages demonstrates that the reduction and metallisation greater amount of hydrogen can be generated whilst the
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Communication
reduction process proceeds, as can clearly be observed in
Fig. 4a (step III). Moreover, this figure suggests that the
reduction of the cathode is complete after 3.5 h of the electro- cathode and dissolved Li
lysis, indicated by a sharp decrease in the current–time curve.
(3) Electrochemical reduction of the Fe
In the presence of an applied negative potential on the
O in the melt, the electrochemical
reduction of the cathodically charged Fe proceeded
As a conclusion, the reduction and metallisation of the through the formation of the intermediate phase Li Fe O .
2 3
O cathode
2
2 3
O
2
3 5
2 3
Fe O pellet cathodes in molten LiCl melt under a humid This reaction can be proposed to occur via the two-step mecha-
argon atmosphere can be speculated to occur in three stages, nism shown below:
as shown in Fig. 1 and described as follows:
þ
H þ e ¼ Ha
ð5:1Þ
ð5:2Þ
ð5Þ
(1) The dissolution of water in molten salt
Water carried into the reactor by the argon gas flow inter-
acts with the molten LiCl to form hydrogen cations (protons)
and oxygen anions in the melt. The mechanism involved can
be described by the occurrence of the hydrolysis of molten
lithium chloride under the influence of water present in the
atmosphere to form lithium hydroxide and hydrogen chloride.
The latter dissolves in molten LiCl to form hydrogen cations.
þ
3
Fe2O3 þ 2Li2O þ Ha þ H þ e ¼ 2Li2Fe3O5 þ H2O
3Fe O þ 2Li O þ H ¼ 2Li Fe O þ H O
2
3
2
2
2
3
5
2
ΔG° = −210.6 kJ mol− at 660 °C.
First, a hydrogen cation is discharged on the Fe
to form H (5.1). The cathodic formation of H from molten
LiCl containing protons has been discussed elsewhere. The
cathodic discharge of the following H could lead to the for-
mation of Li Fe O at the electrode/melt interface where a
sufficient quantity of dissolved Li
1
2
O
3
cathode
a
a
1
9
+
The diffusion coefficient of hydrogen cations (H ) arising from
+
the dissolution of HCl is known to be an order of magnitude
greater than most other ions in high temperature ionic melts.
On the other hand, the generated lithium hydroxide can spon-
taneously decompose to form lithium oxide with a high solubi-
lity of about 12 mol% in the LiCl melt, forming oxygen
2
3 5
2
O is available (5.2). Reaction
(
5) shows that this cathodic phase transition is energetically
favorable. At a prolonged electrolysis time, the removal of
oxygen from the cathode, and therefore a complete metallisa-
tion of the cathode could be achieved. Reactions (6) and (6.1)
represent the possible electrochemical and chemical reactions
responsible for the formation of metallic iron.
2
9
anions. More details of this interaction mechanism can be
found in ref. 18 and the references in it.
(2) The cathodic discharge of cations
Under the influence of the cathodic potential applied,
+
+
cations present in the melt (H or Li ) may be discharged on
the cathode. In order to investigate this, the four possible
electrochemical decomposition reactions that might occur
during the electrolysis at 660 °C, together with the thermody-
namical and electrochemical driving forces required are
shown as below. The first two reactions correspond to the
þ
þ
2ꢀ
Li2Fe3O5 þ 4Ha þ 4H þ 4e ¼ 2Li þ 3Fe þ 4H2O þ O
Li2Fe3O5 þ 4H2 ¼ 3Fe þ ½Li2Oꢁin LiCl þ 4½H2Oꢁin LiCl
ΔG° = −1.1 × 10− kJ mol−1 at 660 °C.
Although reaction (6.1) is energetically close to the equili-
brium, the presence of a cathodic potential and the dis-
ð6Þ
ð6:1Þ
16
2
electrochemical formation of Li and the others to that of H .
2 2
solution of the reaction products (Li O and H O) into the melt
1
provide a further driving force for the reaction to proceed. In
this process, water is consumed to reduce iron oxide to iron,
and then reproduced. Hence, water acts as the redox mediator.
The overall cathodic reaction on the Fe
explained as follows:
2
LiCl ¼ 2Li þ Cl2ðgÞ ΔG° ¼ 666:3 kJ molꢀ E° ¼ ꢀ3:4 V
ð1Þ
2
Li2O ¼ 4Li þ O2ðgÞ ΔG° ¼ 948:7 kJ molꢀ E° ¼ ꢀ2:5 V
1
O cathode can be
2 3
ð2Þ
þ
1
6H þ Fe O þ 6e ¼ 2Fe þ 3½H Oꢁ
2
HCl ¼ H2ðgÞ þ Cl2ðgÞ ΔG° ¼ 200:8 kJ molꢀ E° ¼ ꢀ1:0 V
2
3
2
in LiCl
ð7Þ
ð3Þ
Iron produced by reaction (7) can then function as the
cathode to produce H2:
2
H2O ¼ 2H2ðgÞ þ O2ðgÞ ΔG° ¼ 303:9 kJ molꢀ E° ¼ ꢀ0:79 V
1
þ
ð4Þ
2H þ 2e ¼ H2ðgÞ ðon newly formed Fe cathodeÞ ð8Þ
From these reactions, it can be concluded that despite the
As can be observed from Fig. 4b, the absolute value of the
+
presence of Li in the melt, these cations might not be dis- potential difference between the cathode and the Mo quasi-
charged on the cathode, since the electrochemical decompo- reference electrode gradually decreased in the electrolysis time
sition potentials required are considerably larger than the cell 3.5–5 h, corresponding to step V presented in Fig. 4a. These
voltage of 1–1.4 V applied in this study. The cathodic for- trends are in agreement with the proposed mechanism. The
mation of hydrogen, on the other hand, is favorable. From reduction of hydrogen cations on the electrochemically pro-
these reactions it can also be shown that the evolution of duced iron, reaction (8), would require less voltage in compari-
oxygen at the anode is more favorable than that of chlorine. son with that required for the reduction of iron oxide, reaction
The protons formed in the melt, therefore, can solely be dis- (7). Upon progression of the electrochemical process, hence
charged on the cathode to generate atomic (H
H ) hydrogen.
a
) and molecular metallisation of the cathode, the cathodic reaction becomes
preferentially the discharge of protons to form hydrogen gas
(
2
This journal is © The Royal Society of Chemistry 2018
Green Chem.

View Article Online
Communication
Green Chemistry
on the electrochemically reduced Fe. Therefore, the cathodic the retrieved salt after electrolysis under 1.4 V for 5 h
potential is correspondingly decreased, while the concen- (Fig. S2†) shows that the predominant phase is LiCl, although
tration of hydrogen in the outlet gas is linearly increased. several other small peaks associated with LiCl·H O can also
2
Since electrolysis was conducted at a constant cell voltage, the be detected. LiCl can be dehydrated at temperatures less than
decrease in the cathodic potential led to an increase of the 200 °C (ref. 34) and the residual water remaining in the salt
anodic potential.
at higher temperatures might not be a problem since water is
It should be mentioned that carbon, which has historically the reaction mediator in the present process. This process
been used as the reducing agent for the reduction of iron which uses only electricity to reduce iron oxides to iron is sus-
oxides, is currently a main source of CO greenhouse gas emis- tainable and considerably more economical in comparison
2
sions. The production of one ton of crude steel by the car- with the available alternative technologies. The production
bothermal method is accompanied by the formation of about yield was found to be high at about 98.4% (1.4 V, 5 h)
1
.9 tons of CO
2
which contribute to about 5% of the annual achieved in our lab-scale setup. The current method may also
CO increase globally. It is, in fact, a fundamental challenge be considered as an efficient method of producing hydrogen
2
2
3
facing the steel industry. To address this challenge, hydrogen gas. The process presented here deserves further studies
can be considered as a green alternative to carbon since its focusing on the modification of the assembled cathode con-
reduction product is water. Currently, hydrogen is mainly gen- figuration, the cell design and also the optimisation of para-
erated through the decomposition of methane and photoelec- meters such as the applied voltage, temperature, the chemical
3
0–32
trolysis as well as photocatalysis and biocatalysis of water.
composition of the electrolyte, the porosity of the oxide
However, since hydrogen is in the gas phase at ordinary temp- pellets used and the reduction time to achieve the optimal
eratures and pressures, a main challenge associated with its reduction efficiency and energy consumption. In addition,
use as the reducing agent is related to the storage and trans- the solubility of hydrogen cations in LiCl melt and their
portation of the hydrogen produced.
electrochemical behavior should be studied to understand
Considering these, therefore, it is highly beneficial to the mechanism(s) involved in the process in more details. In
combine the production of hydrogen and its utilisation as the this work, a constant voltage below those used in direct
reducing agent to produce metals from their oxides in a single electro-deoxidation methods was used to highlight the effect
one-step process. This concept was successfully demonstrated in of water as the redox mediator. A constant current approach
the current study by presenting a molten salt method in which can also be used in future studies. These issues are currently
water is dissolved in the melt to generate hydrogen cations. investigated in our lab (ESI).
Hydrogen is subsequently electro-generated on the metal oxide
cathode leading to the in situ reduction of oxides into the corres-
ponding metals. Here, water is employed as the source of hydro-
gen. However, an interesting aspect of the proposed method is
Conclusions
that water is continuously consumed to produce hydrogen, and In summary, the feasibility of the electro-reduction of metal
regenerated by the hydrogen reduction of iron oxide. Hence, the oxides in molten salts in the presence of water as the redox
process is green and only uses electricity to reduce metal oxide. mediator was demonstrated using Fe O as the metal oxide.
2
3
In fact, water is used as the redox mediator and not practically The results obtained showed that Fe
consumed during the process. thoroughly be reduced to pure iron of sponge-like structure in
By integrating the area under the current–time curve of molten LiCl in the presence of water. Furthermore, a uniform
Fig. 4a (1.4 V, 3.5 h), the electric charge consumed to produce octahedral Li Fe crystal was identified as the intermediate
2
O
3
pellets could
2
3 5
O
about 1.12 g metallic Fe could be calculated to be around 6031 product during the electrolysis process. The electro-reduction
C. From this, the current efficiency of the process could be cal- of iron oxide to iron using the water redox mediator presented
culated to be about 96%. The energy efficiency of the molten in this study was specified to progress in three steps, i.e.,
salt process described here is considerably high in comparison hydrolysis of LiCl melt under a humid argon atmosphere to
with that of alternative molten salt approaches (see Table S1†). form protons and oxygen anions in the melt, the electro-trans-
2 3 2 3 2 3 5
It has been found that the electro-deoxidation of Fe O in formation of Fe O into Li Fe O by the in situ generated
molten salts without the involvement of the water redox hydrogen, and the reduction of the latter to metallic iron. The
mediator occurs at cell voltages greater than 1.8 V requiring 5 to proposed process may provide a breakthrough towards the
1
3,14,26
1
5 h for the completion of the reduction process.
It was green production of metals from their oxides, and particularly
demonstrated here that the water redox mediator provides a introduces an alternative electrometallurgical technology for
considerable saving in terms of both the reduction voltage and environmentally friendly production of iron and steel. This
time. The energy consumption achieved was calculated to be technique may be applied to a broad range of metal oxides.
about 2500 kW h per ton of Fe, which is very attractive in com-
parison with that of the traditional carbothermal technique in
blast furnaces (2900–3300 kw h per ton of Fe (ref. 33)).
Conflicts of interest
It should be further pointed out that the salt can be
retrieved after the process and reused. The XRD pattern of There are no conflicts to declare.
Green Chem.
This journal is © The Royal Society of Chemistry 2018

View Article Online
Green Chemistry
Communication
1
5 A. Cox and D. J. Fray, J. Appl. Electrochem., 2008, 38, 1401–
Acknowledgements
1407.
The authors would like to acknowledge the financial support 16 S. Wang, J. Ge, Y. Hu, H. Zhu and S. Jiao, Electrochim. Acta,
provided by the National Natural Science Foundation of China 2013, 87, 148–152.
No. 51750110513), and the Fundamental Research Funds for 17 M. Gibilaro, J. Pivato, L. Cassayre, L. Massot, P. Chamelot
(
the Central Universities (No. N172505001).
and P. Taxil, Electrochim. Acta, 2011, 56, 5410–5415.
18 A. R. Kamali and D. J. Fray, Nanoscale, 2015, 7, 11310–
11320.
19 A. R. Kamali, Green Chem., 2016, 18, 1952–1964.
20 J. Oh and D. Noh, Fuel, 2017, 196, 144–153.
21 J. Zieliński, I. Zglinicka, L. Znak and Z. kaszkur, Appl.
Catal., A, 2010, 381, 191–196.
Notes and references
1
2
A. Allanore, J. Electrochem. Soc., 2014, 162, E13–E22.
A. Allanore, L. Yin and D. R. Sadoway, Nature, 2013, 497,
3
53–356.
2
2
2
2 J. Dang, K. Chou, X. Hu and G. Zhang, Steel Res. Int., 2010,
3
4
5
6
D. Wang, A. J. Gmitter and D. R. Sadoway, J. Electrochem.
Soc., 2011, 158, E51–E54.
Z. Zhou, H. Jiao, J. Tu, J. Zhu and S. Jiao, J. Electrochem.
Soc., 2017, 164, E113–E116.
K. Zhang, H. Jiao, Z. Zhou, S. Jiao and H. Zhu,
J. Electrochem. Soc., 2016, 163, D710–D714.
84, 526–533.
3 https://www.worldsteel.org/steel-by
topic/sustainability/
Sustainable-Steel-Indicators-2017-and-the-future.html.
4 R. Bhagat, D. Dye, S. L. Raghunathan, R. J. Talling,
D. Inman, B. K. Jackson, K. K. Rao and R. J. Dashwood,
Acta Mater., 2010, 58, 5057–5062.
A. Allanore, H. Lavelaine, G. Valentin, J. P. Birat,
P. Delcroix and F. Lapicque, Electrochim. Acta, 2010, 55,
2
2
2
2
5 D. J. Fray, G. Z. Chen and T. W. Farthing, Nature, 2000, 407,
361–364.
4
007–4013.
6 D. Tang, H. Yin, W. Xiao, H. Zhu, X. Mao and D. Wang,
J. Electroanal. Chem., 2013, 689, 109–116.
7 K. Jiang, X. Hu, H. Sun, D. Wang, X. Jin, Y. Ren and
G. Z. Chen, Chem. Mater., 2004, 16, 4324–4329.
8 A. M. Abdelkader, K. T. Kilby, A. Cox and D. J. Fray, Chem.
Rev., 2013, 113, 2863–2886.
7
8
9
B. Yuan, O. E. Kongstein and G. M. Haarberg,
J. Electrochem. Soc., 2009, 156, D64–D69.
X. Zou, S. Gu, H. Cheng, X. Lu, Z. Zhou, C. Li and W. Ding,
J. Electrochem. Soc., 2015, 162, D49–D55.
S. Licht and B. Wang, Chem. Commun., 2010, 46, 7004–
7
006.
2
3
9 Y. Sakamura, J. Electrochem. Soc., 2010, 157, E135–E139.
0 M. S. Dresselhaus and I. L. Thomas, Nature, 2001, 414,
3
1
1
0 B. Cui and S. Licht, Green Chem., 2013, 15, 881–884.
1 S. Wang, G. M. Haarberg and E. Kvalheim, J. Iron Steel Res.
Int., 2008, 15, 48–51.
2 G. M. Haarberg, E. Kvalheim, S. Rolseth, T. Murakami,
S. Pietrzyk and S. Wang, ECS Trans., 2007, 3, 341–345.
3 G. Li, D. Wang and Z. Chen, J. Mater. Sci. Technol., 2009,
32–337.
1 Z. Zou, J. Ye, K. Sayama and H. Arakawa, Nature, 2001, 414,
25–627.
3
1
1
1
6
3
3
2 M. Gratzel, Nature, 2001, 414, 338–344.
3 L. D. D. Harvey, Energy efficiency and the demand for energy
services, Earthscan, London, 2010.
2
5, 767–771.
4 H. Yin, D. Tang, H. Zhu, Y. Zhang and D. Wang,
34 A. R. Kamali, D. J. Fray and C. Schwandt, J. Therm. Anal.
Electrochem. Commun., 2011, 13, 1521–1524.
Calorim., 2011, 104, 619–626.
This journal is © The Royal Society of Chemistry 2018
Green Chem.