Electrochemically assisted synthesis of fuels by CO2 hydrogenation over Fe in a bench scale solid electrolyte membrane reactor

Article abstract of DOI:10.1016/j.cattod.2016.02.025

The electrochemically assisted synthesis of fuels by CO₂ hydrogenation was studied over a cheap, widespread and non-precious Fe catalyst in a bench scale oxygen ion conducting membrane (YSZ) reactor. The studies were performed under conditions representative of potential practical application i.e., under atmospheric pressure, at relatively low temperatures and high gas flow rates, with varying H₂/CO₂ ratios and using gas compositions typical for postcombustion CO₂ capture exit streams and easily scalable catalyst-electrode configurations. The Fe-TiO₂ catalyst film was deposited by "dip-coating" and characterised both after pre-reduction and after testing. CO₂ conversion and selectivities to CH₃OH and C₂H₆O can be enhanced up to 4, 50 and 1.7 times, respectively, by electrochemically controlled migration of O^2- promoting ions to or from the catalyst surface. The optimum temperature for the process was 325°C. Lower gas flow rates favoured the synthesis of CH₃OH and C₂H₆O. CO₂ conversion and selectivities to CH₃OH and C₂H₆O showed a maximum for a stoichiometric H₂/CO₂ ratio of 3. Formation of C₃H₆ and CO is strongly favoured for a H₂/CO₂ ratio of 4 and 2, respectively, as a result of the increased and decreased hydrogen availability in the reaction system.

Full text of DOI:10.1016/j.cattod.2016.02.025

Catalysis Today 268 (2016) 46–59
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Electrochemically assisted synthesis of fuels by CO hydrogenation
2
over Fe in a bench scale solid electrolyte membrane reactor
∗
Esperanza Ruiz , Pedro J. Martínez, Ángel Morales, Gema San Vicente, Gonzalo de Diego,
José María Sánchez
Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas (CIEMAT), Av. Complutense, 40, 28040 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
The electrochemically assisted synthesis of fuels by CO2 hydrogenation was studied over a cheap,
widespread and non-precious Fe catalyst in a bench scale oxygen ion conducting membrane (YSZ) reac-
tor. The studies were performed under conditions representative of potential practical application i.e.,
under atmospheric pressure, at relatively low temperatures and high gas flow rates, with varying H2/CO2
ratios and using gas compositions typical for postcombustion CO2 capture exit streams and easily scalable
catalyst-electrode configurations.
Received 19 September 2015
Received in revised form 26 February 2016
Accepted 26 February 2016
Available online 11 March 2016
Keywords:
Solid electrolyte
Bench scale
CO2 hydrogenation
Fe-TiO2/YSZ
Electrochemical promotion
The Fe-TiO2 catalyst film was deposited by “dip-coating” and characterised both after pre-reduction
and after testing.
CO2 conversion and selectivities to CH3OH and C2H6O can be enhanced up to 4, 50 and 1.7 times,
2
−
respectively, by electrochemically controlled migration of O promoting ions to or from the catalyst
surface.
◦
The optimum temperature for the process was 325 C. Lower gas flow rates favoured the synthesis
of CH3OH and C2H6O. CO2 conversion and selectivities to CH3OH and C2H6O showed a maximum for a
stoichiometric H2/CO2 ratio of 3. Formation of C3H6 and CO is strongly favoured for a H2/CO2 ratio of 4 and
2
, respectively, as a result of the increased and decreased hydrogen availability in the reaction system.
2016 Elsevier B.V. All rights reserved.
©
1
. Introduction
Due to the increase in CO2 atmospheric levels and the dimin-
future. In this way, carbon dioxide can be chemically converted
from a harmful greenhouse gas causing global warning into a valu-
able, renewable, environmentally neutral and inexhaustible fuel
source for the future [1–4].
Two main reactions can occur on co-feeding CO2 and H2 over a
hydrogenation catalyst:
ishing fossil fuel resources arising from wide spread production of
energy by fossil fuels combustion, valorisation of CO2 emissions
to clean fuels is viewed as a complementary strategy to capture
and storage for an effective quantitative reduction of the CO emis-
2
x CO + (2x − z + y)H2 → CxHyOz + (2x − z)H O
(1)
(2)
2
2
sions, allowing their recycling and, therefore, a more sustainable
use of the energy resources. Chemical recycling of carbon dioxide
from combustion power plants, as an energy carrier, can be accom-
plished via its capture and subsequent hydrogenation to renewable,
useful and environmentally neutral fuels (methane, methanol,
dimethyl ether, etc.), provided that any available renewable energy
source (wind, solar or hydraulic) is used for both production of nec-
essary hydrogen (by water electrolysis) and chemical conversion
◦
−1
CO + H → CO + H OꢀG = 19.9 kJ mol
2
2
2
The former is the synthesis reaction resulting in the formation
of hydrocarbons and/or oxygenates (alcohols or ethers). The latter
is the reverse water gas shift (RWGS) reaction.
Most studies on conventional catalytic hydrogenation of CO2
to fuels, such as methanol and dimethyl ether, have been accom-
plished using fixed-bed reactor configurations [3] of metal or
metal oxide based catalysts supported on metal oxides [5] and
at high pressures [6] to favour the CO2 hydrogenation reaction.
Electrochemical promotion is reported to be a very powerful tool,
especially suitable for activation of very slow processes [7]. These
processes if realized with conventional catalytic technologies need
either extreme operating conditions (very high pressures and tem-
of CO . Moreover, it has been foreseen that increasing amounts of
cheap CO2 will be available from carbon sequestration in the near
2
∗ Corresponding author.
E-mail address: esperanza.ruiz@ciemat.es (E. Ruiz).
http://dx.doi.org/10.1016/j.cattod.2016.02.025
0
920-5861/© 2016 Elsevier B.V. All rights reserved.

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
47
peratures) or long retention times [7], as seems to be the case for
catalytic CO hydrogenation. In fact, CO adsorption and activation
Therefore, there were some scale-up facets, such as operation at
high flow rates, under realistic gas compositions and using catalyst-
electrode configurations easily adaptable to the existing devices
(conventional flow reactors), that need to be tackled in more depth
for the potential practical application of this technology [7,20,21].
Moreover, there are also several aspects regarding the practical
application of the technology, such as cost minimization, durabil-
ity, useful lifetime, etc. which have not yet been addressed in detail
[7,20,21].
2
2
have been suggested to be the rate limiting step in CO2 hydro-
genation [5,8,9]. Therefore, operation of conventional catalysts at
higher pressures results in an increased amount of adsorbed CO2,
and, hence, in an increase of reaction rate. However, electrochem-
ical promotion of catalysis allows increasing the surface coverage
of chemisorbed CO2 species, giving rise to a concomitant increase
in reaction rate and allowing the operation of the catalyst at lower
pressures, which may result in energy savings.
We have recently reported various bench-scale studies of elec-
Electrochemical promotion of catalysis (EPOC), by coupling
electrochemistry to catalysis, has been considered as an alter-
native approach to classical chemical promotion of catalyst by
electrochemically supplying and controlling the concentration of
a promoter on an active metal catalyst surface. The application of
small currents or potentials between a metal catalyst which is in
contact with a solid electrolyte, which acts as a source of promoter
species, and a counter electrode results in the migration of promot-
ing species to or from the catalyst surface, allowing increasing the
trochemically assisted CO hydrogenation to valuable hydrocarbon
2
and oxygenated fuels over Pt [22] and Cu [23] on K-␤Al O and over
2
3
Pt, Ni and Pd on YSZ [24]. These studies were carried out at high gas
flow rates, under atmospheric pressure, at relatively low tempera-
ture and using gas compositions representative of CO capture exit
2
streams and tubular configurations easily adaptable to the exist-
ing catalytic devices (conventional flow reactors) and prepared by
easily scalable procedures.
In the present work, we extend the study to other electrochem-
catalytic activity for the CO hydrogenation reaction and modifying
ical catalyst systems, more specifically to Fe-TiO2 on YSZ (an O2
−
2
the selectivity to the desired products, as well as to simultane-
ously monitor and control the reaction during the process [4,9].
However, there are few previous studies of electrochemically pro-
moted catalytic CO2 hydrogenation. It has been studied over Cu on
conductor). This system is based on a cheaper metal than Pt [22],
but was also prepared by “dip-coating”, which is thought to allow
obtaining a thinner metal film than the “electroless” method used
to deposit other non-noble metals, such as Cu [23] and Ni [24],
which cannot be deposited by “dip-coating”. Therefore, this study
may contribute to decrease material cost for the process and to the
development of the necessary catalyst manufacturing technique,
which may have an impact in the potential practical application of
the technology.
2−
SrZr0.9Y0.1O3-␣ (a proton conductor) [8], Pt on YSZ (an O conduc-
tor) [5,10], Pd on YSZ or Na-␤Al O3 (a Na+ conductor) [11], Rh on
2
YSZ [5,12], Cu on TiO -YSZ [5], Ru on YSZ, Na-␤Al O , K-␤Al O
2
2
3
2
3
(
a K+ conductor) or BZY (a proton conductor) [13–17], Ni or Ru
impregnated carbon nanofibers on YSZ [18] and Ni on K-␤Al O
2
3
[
[
19]. In addition, these studies have been carried out at lab scale
8,10–19] and using reactor configurations (discs) [8,10,13–19],
This work presents, to the extent of our knowledge for the first
time, a bench-scale study of electrochemically promoted synthesis
costly metals (Pt, Pd, Ru, Rd, etc.) [10–17], metal film preparation
of renewable fuels by CO hydrogenation over a cheap, widespread
2
methods [8,10–12,18] and simplified gas compositions (H and CO2
and non-precious Fe catalyst on −TiO /YSZ. This study was per-
2
2
diluted in He or N ) [5,8,10–19] which are not representative of the
formed using only one-tube solid oxide cell configuration, but it
could be envisaged to upscale the reactor design in a modular multi-
tubular solid oxide cell configuration [25] with only two external
connecting wires, which would be easily adaptable to available
conventional flow reactors (fixed bed) and may simplify electrical
supply to the electropromoted system.
2
potential practical application of the technology to real postcom-
bustion CO2 capture exit streams.
According to literature [7,20,21], to foster the ultimate uti-
lization of electrochemical promotion of catalysis in commercial
reactors, some challenges need to be addressed:
In this study, Au was used for the counter electrode and connect-
ing wires, given that it has been reported to be catalytically inactive
for the CO2 hydrogenation reaction [5,13], in order to ascribe the
observed electrocatalytic behaviour to the Fe catalyst only. How-
ever, with a view of potential application, the use of other cost
effective and easily available counter electrodes (stainless steel
mesh/sheet, composites, etc.) and connection wires should be con-
sidered.
The use of a wireless electrochemical promotion approach,
which is reported [26] to allow controlling the migration of oxide
anion species through the solid electrolyte by controlling the oxy-
gen chemical potential difference between both electrodes without
the need of using connection wires, could also result in a simple
configuration for upscaling the proposed catalyst system in a dou-
ble chamber configuration with the counter electrode exposed to
oxygen atmosphere.
a.) Material cost minimization. Moving from thick catalyst films
(
0.1–5 mm thick), mostly made via paste deposition of high cost
materials (noble metals) which lead to poor material utilization
metal dispersion below 0.01%) and low surface area, to thin (on
(
the order of few nm thick) catalyst electrodes or dispersed cata-
lysts that are sufficiently active, robust and inexpensive so as to
be deployed in practical reactors. The use of less expensive cata-
lysts (non-noble metals) and electrolyte systems is also needed.
Commercially available materials and semiproducts should be
used [20,21].
b.) Devising efficient and compact reactor design with easiness of
electrical connection [20,21].To simplify reactor design and to
make scale-up of electropromoted reactors easier, it is nec-
essary to use configurations (multitubular, monoltithic, etc.)
which would be susceptible of application in available devices
The present study was performed at high flow rates, under
atmospheric pressure, at relatively low temperatures and using
(
fixed bed conventional flow reactor), such as bipolar systems
concentrated CO streams, meant to represent exit streams of post-
2
in combination with a single chamber reactor.
c.) The research focus has to be shifted from lab-scale, fundamen-
tal research, to more applied research (pilot plant scale tests,
development of necessary catalyst manufacturing techniques
etc.) [7].
d.) To verify the electrochemical promotion of catalysis in larger
scale reactors at larger scale throughputs and in continuous
mode of operation [7].
combustion CO2 capture plants (commercial amine plants) and
H /CO ratios from 2 to 4 to consider the effect of a discontinuous H
flow, as the H2 needed for CO2 hydrogenation may be discontinu-
ously produced by water electrolysis only when electricity demand
is low and from intermittent renewable energy sources [2,3]. It
has been reported [3] that somewhat higher total CO2 conversions
2
2
2
(
especially to CH ) could be possible by using higher H /CO ratios,
4
2
2

4
8
E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
but the most efficient use of available H2 occurs at lower H /CO2
ratios.
The morphology of the catalyst film was investigated via SEM
using a HITACHI SU6600 FEG-SEM instrument with ZrO/W emis-
sion type of 1–30 kV (100 V steps) of accelerating voltage and 3 nm
at 1 kV/1.2 nm at 30 kV of resolution.
2
Reactions conditions of temperature, flow rate and H /CO2
2
ratios were varied to evaluate their influence on the extent of the
electrochemically assisted CO2 hydrogenation reaction and on the
selectivity to the different target oxygenated fuels.
XRD patterns of the catalyst-working electrode films were
recorded on a PHILIPS “Xpert-MPD” instrument using a Cu K␣ X-
◦
ray source (45 kV and 40 mA), a 2␪ range of 15–75 , a step size of
◦
2
␪ = 0.03 and a step time of 2 s.
2
. Experimental
The surface chemical composition of the catalyst electrode film
was examined by XPS using a Perkin-Elmer PHI 5400 System
equipped with a Mg K␣ (hꢀ = 1253.6 eV) excitation source running
at 15 kV and 20 mA and having a beam diameter of 1 mm. Base pres-
sure in the analysis chamber was maintained at about 10 Torr.
The pass energy was set at 89.5 eV for general spectra (0–1100 eV)
and at 35.75 eV for high resolution spectra. The energy scale was
referenced to the carbon 1s signal at 285.0 eV.
2
.1. Electrochemical catalyst
The electrochemical catalyst evaluated in the present work con-
−
9
sisted of a thin Fe film (catalyst-working electrode), coated on a
TiO intermediate layer (used to improve electronic conductivity of
the electrocatalyst), both deposited by dip-coating on the outer sur-
face of a 26-mm-i.d., 100-mm-long, 1–2-mm-thick YSZ tube open
on both sides.
2
A gold counter electrode was painted on the inner side of the
tube to perform polarizations.
2.3. Experimental set-up
Au was chosen as the auxiliary electrode material because it is
reported [5,13] to be inert for the CO2 hydrogenation reaction.
The Au counter electrode was prepared by painting the inner
side of the YSZ tube with a gold paste (HERAEUS-C5729). The
Electrochemically assisted CO hydrogenation over Fe-TiO cat-
2
2
alyst film was studied in a bench-scale plant, described in detail
3
−1
elsewhere [22], which is able to treat up to 20 m h (at 273 K and
◦
1
atm) of gas with temperatures ranging between 250 and 450 C,
◦
◦
deposited paste was dried at 150 C during 10 min, heated to 850 C
at about atmospheric pressure.
◦
at a controlled rate and, finally, annealed at 850 C during 10 min.
The different gas constituents of a post-combustion CO capture
2
Fe catalytic and intermediate TiO2 layers have been also coated
on the outer surface of the YSZ tube after deposition of the Au
counter electrode on the inner side of the tube. Both open ends of
the YSZ tube were plugged in order to avoid adding active catalyst
particles to the counter electrode side.
exit stream and hydrogen can be provided by mass flow controllers
from bottled gases. Steam can be supplied by vaporising water fed
into a boiler by a metering pump. The mixed wet gas is then pre-
heated and directed to a fixed-bed down-flow quartz reactor, with
5 mm of diameter and 900 mm of length, heated by a three-zone
electrical furnace. Polarization across the cell was measured and
controlled via a potentiostat-galvanostat.
3
Firstly, substrates were cleaned with absolute ethanol in an
ultrasonic bath, rinsed with absolute ethanol and dried in an oven
◦
at 90 C during 30 min.
Gaseous products from the reactor were delivered through a
heated transfer line to the gas analysis system in order to avoid
condensation of any volatile products. Gas composition was simul-
taneously determined using a gas micro-chromatograph, and an
NDIR CO /CO on line analyser [22], allowing the analysis of: H , N ,
Titanium dioxide has been deposited by sol-gel technique, using
a precursor solution that contains water, ethanol, hydrochloric
acid and tetrabutyl orthotitanate in a molar ratio 1/14/0.5/1. With-
drawal speed varied from 15 to 25 cm/min. Coatings have been
◦
2
2
2
thermally cured at 500 C during 30 min.
CO, CH , CO , C H , C H , C H , C H , C H , methanol, dimethyl
4
2
2
2
2
4
2
6
3
6
3
8
Iron oxide has been deposited by dip-coating, withdraw-
ether and ethanol.
ing substrate from
a precursor solution composed of 2-(2-
aminoethylamino) ethanol, ethanol and ferrous nitrate, according
to European patent EP1321539 [27], at speeds from 15 to
2
.4. Operating conditions and procedure
◦
2
3
5 cm/min. Coatings have been thermally cured at 500 C during
0 min.
The electrochemical catalysts were properly situated in the
The main role of TiO intermediate layer is to enhance electronic
2
reactor in order to lessen by-pass phenomena and to enhance
catalyst-reactive gas contact. The electrical connections in the reac-
tor were made from gold wires (HERAEUS), which is reported
to be catalytically inert in the process [5]; therefore, all the
potential-induced changes in catalytic activity and selectivity can
be exclusively attributed to the metal film.
conductivity of the Fe film which, as indicated above, was deposited
as iron oxide and, as will be explained later, after reduction in H
2
◦
at 400 C was mainly composed by metallic Fe and Fe O .
3
4
It has been reported, that under oxidizing conditions, TiO alone
2
2−
exhibits mixed electronic -ionic (O ) conductivity and can lead to
pronounced electrochemical promotion of oxidation reactions [28]
and, in some cases, may protect active metal (Rh) from oxidation
◦
The Fe film was reduced in a stream of H2 at 400 C during 1 h
[
22], before performing the electrochemically assisted CO2 hydro-
(
by lowering the stability of potentially formed surface Rh oxides)
genation experiments.
Open circuit potential was maintained during 30 min prior to
each test to define a reproducible reference state. Electrochemi-
under conditions (anodic polarization) where otherwise will be oxi-
dized [29]. However, under the strongly reducing conditions of the
present investigation, the conductivity of TiO is entirely electronic
2
cally assisted hydrogenation tests were performed under H /CO2
2
[
28] and the main role of TiO2 is only to enhance the electronic
binary mixtures, at different applied potentials (between −2 and
conductivity of the Fe film.
◦
2
V), H /CO ratios (from 2 to 4), temperatures (225–400 C) and
2
2
−
1
flow rates (90–216 N l h ), in order to determine the effect of
the utilized operating conditions on CO2 conversion, efficiency
of electrochemical promotion and selectivity to the different fuel
products.
2.2. Catalyst characterisation
A small fragment of the electrochemical catalyst was character-
ized, both after reduction and after testing, by Scanning Electron
Microscopy (SEM), X-Ray Diffraction (XRD) and X-ray Photoelec-
tron Spectroscopy (XPS) techniques.
CO2 hydrogenation tests were performed under H2 and CO2
binary mixtures; although a small amount of N2 (about 0.5%) was
added to the reaction gas mix as an internal standard. Accordingly

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
49
[
20], CO2 conversion (XCO2 ) and “CO2 free selectivity” is defined as
(
3) and (4), respectively:
ꢀ
ꢁ
1 − [
CO2] × [N2]
× 100
(3)
(4)
o
i
XCO2
=
[
CO2] × [N2]
i
o
ni × Mi
Si =
× 100
i=n
ꢂ
ni × Mi
i=1
where [CO ] and [CO ] are the corresponding CO2 molar quanti-
2
i
2
o
ties at the inlet and outlet of the reactor. As well, [N ] and [N ] are
2
i
2
o
N2 molar quantities at the inlet and outlet of the reactor, respec-
tively. Additionally,S is the selectivity to product i, n is the number
i
i
of carbon atoms of product i and M is moles of product i, respec-
i
tively.
The effect of polarization on catalyst performance for the CO2
hydrogenation reaction was gauged in terms of CO2 rate enhance-
ment ratio (Eq. (5)) [22].
CO2 rate enhancement ratio is defined as (Eq. (5)):
p
2
up
rCO
ꢁCO2
=
(5)
rCO2
p
2
up
2
where rCO and rCO are the CO catalytic rates with (under appli-
2
cation of a given potential V) and without (unpromoted reference
state, under open circuit conditions, i.e., no potential application
and measured open circuit potential almost equal to 0 V) electro-
chemical promotion, respectively.
According to literature [5,13,16], in the case of CO hydrogena-
2
2−
tion on metal catalysts deposited on YSZ (an O ion conductor),
O2 is not a reactant and, therefore, any positive potential-induced
catalytic rate change is due to electrochemical promotion indepen-
dently of the value of Faradaic efficiency. The same can be reasoned
in the case of negative potential, because even if CO is decom-
2
2−
posed via O removal from the catalyst surface, this does not lead
Fig. 1. SEM micrograph of the Fe catalyst-working electrode film: (a) after reduction,
to H O formation. In any case, the apparent Faradaic efficiency ꢂ
2
(
b) after testing.
[
5], which is also commonly used to quantify the magnitude of
electrochemical promotion, has been also calculated. The apparent
Faradaic efficiency ꢂ [5] is defined from (6):
The carbon mass balance (relative difference between carbon
mass at the reactor inlet and outlet) was also calculated for each
test and resulted to be below 0.5% for all cases.
r − ro
⁄
ꢄr
= I
ꢄr (mol/s) × E
i
i
ꢃ =
=
(6)
i
I
I
F
⁄F
⁄F
where r is the electropromoted rate and ro is the normal, open
3. Results and discussion
circuit catalytic rate (both in g-eq/s). ꢅr is the current- or potential-
induced change in the catalytic rate of each product formation
i
3.1. Catalyst characterisation studies
(
mol/s) and E is the g-eq factor for each product formation reaction.
i
In this study, ꢂ has been calculated for the main product obtained
at each H /CO ratios, i.e. ꢂ (EC2H6O = 6), ꢂC3H6 (EC3H6 = 6) and
ꢂCO (ECO = 2) for H /CO2 ratios of 3, 4 and 2, respectively. F is the
As reported in literature [6,30–35], the utilised preparation
technique determines also porous structure, surface morphology
and particle size of the metal thin film and therefore the elec-
trochemically assisted catalytic behaviour of the system. SEM
micrographs of Fe catalyst-working electrode film, both as pre-
pared and after reduction (a) and after testing (b), are shown in
Fig. 1a and b, respectively. As can be observed in SEM image of
the sample after reduction in Fig. 1(a), the obtained film seems to
resemble a typical foam structure [36], suggesting that they are
porous (allowing reactants and products diffusion), and continu-
ous, as verified by electrical conductivity measurements. The Fe
film thickness was also determined by SEM and resulted to be about
2
2
C2H6O
2
Faraday’s constant, I is the applied current. The term I/F corresponds
to the rate of ions (g-eq/s) supplied to catalyst according to the Fara-
day’s law. Thus, |ꢂ | = 1 refers to pure Faradaic rate enhancement,
and |ꢂ| > 1 implies electrochemical promotion.
The value of current efficiency (␩c) and energy cost (CE) were
calculated for each fuel product formation (methanol, ethanol and
dimethyl ether) under different applied potentials and operating
conditions, following the expressions (7) and (8), respectively:
ꢃ
ꢄ
ꢆc = (mp/Mp) × ꢀe/ꢀp × F /(I × t)
CE = V × F × (ꢀe/ꢀp)/(ꢆc × Mp)
(7)
(8)
2
00 nm.
XRD spectra of the fresh (as prepared and reduced) and used
after exposure to reaction conditions) samples of Fe film are com-
(
◦
◦
◦
where ␯e is the number of electrons transferred and ␯p is the
stoichiometric coefficient in each product formation reaction. Addi-
tionally, mp and Mp are the mass and molecular weight of product
formed, respectively.
pared in Fig. 2. The peaks at 2␪ = 44.8 , 65.1 and 82.6 , in the
XRD spectra of the as reduced sample, were identified as the typ-
ical diffraction peaks of Fe metal (JCPDS card no. 01-087-0722);
whereas the rest of peaks were assigned to the YSZ solid electrolyte

5
0
E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
Fig. 2. XRD analysis of the Fe catalyst-working electrode film: (a) after reduction,
b) after testing.
(
Fig. 3. Influence of temperature on (a) CO2 conversion under open circuit conditions
᭿), upon negative polarization (−2 V) (᭹) and upon positive polarization (2 V) (ꢀ)
and on (b) selectivities to CH OH (ꢁ, ꢂ) and C H O ( ) under open circuit
(
(
(
JCPDS card no. 98-008-9429) and to the TiO2 intermediate layer
JCPDS card no. 98-009-2363). No peaks of iron oxides, or other
3
2
6
,
conditions (filled symbols) and upon negative polarization (−2 V) (open symbols)
◦
−1
phases, were detected for the sample after reduction.
over Fe-TiO2/YSZ/Au. (225–400 C, 90 l h , H2/CO2 = 3).
The average crystallite size of the as reduced sample was eval-
uated from X-ray broadening of the main metallic Fe diffraction
peak at 2␪ = 44.815 by using the well-known Debey-Scherrer equa-
tion [37,38]. The average particle size resulted to be of about 38 nm.
The catalyst film metal dispersion has been also estimated from the
obtained particle diameter and resulted to be of about 3.2% [39].
The XPS analysis of the fresh (as prepared and reduced) sample
revealed the presence of several superficial compounds that could
not be identified by XRD, resulting in Fe being present mainly as
of the used sample. This fact seems to be confirmed also by XRD
which revealed a loss of crystallinity, as can be deduced from the
decrease of the intensity of XRD peaks associated with metallic Fe.
The XPS spectra of the used sample revealed also the coex-
istence of Fe oxides and oxides/hydroxides as resembled by the
appearance of peaks at binding energies of about 710.3 eV in the
Fe 2p3/2 spectra. The O1s spectra showed also that oxygen could
be present as OH groups (38.7%) and Fe oxides (61.4%) as indicated
by the appearance of peaks at binding energies of about 529.8 and
Fe O as shown by the appearance of peaks at a binding energy
3
4
of about 711 eV in the Fe 2p3/2 spectra and at a binding energy of
about 724.6 eV in the Fe 2p1/2 [40–43]. The appearance of a peak
at 529.77 eV in the O1s spectra seems to confirm the presence of
Fe O in the as reduced sample.
5
31.6 eV, respectively. These results suggest that Fe and or Fe O4
3
could be oxidized/oxohydroxized under reaction conditions.
3
4
Reflections of the rest of Fe O4 phase were undetectable in
3.2. Electrochemically assisted CO2 hydrogenation tests
3
the XRD pattern of the as reduced sample. This seems to indicate
that the as-deposited and reduced Fe film could be almost XRD-
amorphous or nanocrystalline.
3.2.1. Effect of temperature
Fig. 3(a) shows the effect of temperature on the conversion of
CO2 under open-circuit state and upon cathodic (–2 V) and anodic
According to literature [44], the reduction of iron catalyst by
hydrogen is known to be a two or three-staged process. That is
◦
(+2 V) polarization conditions, at temperatures from 225 to 400 C
◦
−1
first Fe O is reduced at 397 C to Fe O , which is then reduced
and using 90 l h of total gas flow rate and a H /CO2 ratio of 3.
2
3
3
4
2
◦
to metallic iron at 702 C. According to catalyst characterization,
initially deposited iron oxides were reduced to Fe O and partly
also to Fe during reduction in H2 at 400 C.
Dimethyl ether, methanol, ethanol, CO and in lesser extent
hydrocarbons (mainly C H6 and C H ), CH , formic acid and acetic
3
4
2
3
6
4
◦
acid were formed depending on polarization conditions and tem-
peratures.
As can be observed in Fig. 3(a), under open circuit and positive
polarization conditions, CO2 conversion monotonically decreases
as temperature increases for whatever value of applied potential.
Some agglomerates can be observed in SEM image of the sample
after testing (Fig. 1(b)), which are thought to be related with sin-
tering of iron particles upon exposure to testing gas environment
and not to the fact that the examined area is different in the case

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
51
This fact may be related to catalyst poisoning by carbon deposi-
tion over Fe and Fe O which progressively blocks the active sites
increases further and FT activity develops and is progressively stab-
lished on the catalyst surface. CO yield increases with temperature
via the RWGS reaction which is a reversible endothermic reaction.
However, the dependence of hydrocarbon and oxygenates yield is
under a compromise between higher driving force in activity and
larger tendency to inert carbon deposition [47,52].
3
4
[
(
(
44,45] or by oxidation of Fe O4 to inactive Fe oxides (Fe O ) [45]
3
2
3
during ongoing carbon deposition) or hydroxides (Fe(OH) ) [46]
3
−
2−
ions thermally and/or
by OH [13] species resulting from O
electrochemically supplied to Fe surface).
Upon negative polarization CO2 conversion decreases on
increasing temperature up to 300 C, also probably due to catalyst
It has been reported [13] that under reducing conditions, as
◦
in the present case (binary mixture of H2 and CO ), it is likely
2
−
poisoning by carbon deposition, and then slightly increases with
that hydroxyl group anionic species (OH ) are formed by reac-
tion of hydrogen with electrochemically/thermally migrated O
◦
2−
temperature, showing a maximum at about 350 C, and after that,
CO2 conversion decreases on increasing temperature.
species, which are reported [5] to have a short lifetime, on the
catalyst surface. Therefore, in spite of the strongly reducing envi-
ronment, Fe O might be oxidized to inactive Fe hydroxides [46] by
According to catalyst characterisation results, iron oxides were
reduced to Fe O and partly to metallic Fe during pre-reduction at
3
4
3
4
◦
4
00 C and subsequently oxidized to a less crystalline (amorphous)
the hydroxyl species, contributing also to magnetite deactivation.
The application of positive potentials (+ 2 V) results in O2 supply-
ing from solid electrolyte to the catalyst surface [5,13]. Therefore,
the decrease in conversion values obtained for positive polariza-
tion could be also because upon application of positive potentials
oxygen ions are transferred from the solid electrolyte to the Fe cata-
lyst surface resulting in an increased Fe O oxidation. In absence of
−
phase consisting of iron oxides and hydroxides upon exposure of
the catalyst to the testing gas environment and operating condi-
tions.
Iron oxides, which have been used as FT catalysts for many years,
are also active in reverse water gas sift reaction (RWGS) (Eq. (2))
[
47]. According to literature [48,49], active species for Fischer Trop-
3
4
sch (FT) reaction are generally obtained by the reduction of ␣-Fe O3
polarization, i.e., under unpromoted conditions, although no elec-
2
trical current passes across the cell, O2 ions could also thermally
migrate from the solid electrolyte to the catalyst surface as tem-
perature increases, due to the increased ionic conductivity of the
solid electrolyte at higher temperatures. In both cases CO2 conver-
−
via Fe O , i.e., the iron species derived from ␣-Fe O seems to be
3
4
2
3
the catalytic sites for FT-reaction. Moreover, during CO2 hydro-
genation, Fe O phase is reported to disappear slowly and a new
3
4
amorphous, probably oxidic phase is formed, which appears to be
also active for the RWGS reaction [47].
sion decreases as temperature increases, because migration of O2
−
In accordance with literature [47,50], CO2 hydrogenation over
Fe based conventional catalyst (unpromoted conditions) proceeds
to the catalyst surface increases with temperature as a result of the
increased ionic conductivity at higher temperatures.
2−
via a two-step process, with initial reduction of CO to CO via RWGS
On the contrary, under negative polarization (−2 V), O species
are pumped away from the catalyst surface to the solid electrolyte
YSZ, enhancing CO2 dissociative adsorption to CO (Eq. (10)) and
subsequent formed CO dissociation into surface carbon species (Eq.
(11)) [6,54,55], as well as direct CO2 reduction to carbon species
(Eq. (12)). Therefore, the initial decrease on CO2 conversion on
increasing temperature observed is thought to be due to Fe poi-
soning by carbon deposition, given that the amount of surface
carbon adsorbed on the catalyst surface is expected to increase
with time and temperature [2]. As commented above on increas-
ing temperature reverse water gas shift reaction (RWGS) activity
increases. Under negative polarization (−2 V), upon increasing tem-
2
reaction followed by conversion of CO via FT reaction. FT synthe-
sis is reported [51] to produce both hydrocarbon and oxygenated
products (oxygenates) by reducing CO with hydrogen on a suitable
catalyst, such as iron supported on TiO2 [49]. It has been reported
[
49] that on Fe-TiO2 catalysts, both metallic and oxidic iron phases
may play different roles in CO hydrogenation. Iron oxides are nec-
2
essary for the hydrogenation of CO2 to from CO. Metallic iron is
beneficial to form lower olefins in CO hydrogenation. Moreover,
iron oxides are also active to form oxygenates (CH OH, C H O, etc.)
3
2
6
from CO [51].
Another coexistent reaction is the CO2 reduction to carbon.
According to literature [44], CO2 can be at least partially dissoci-
perature ionic conductivity of the electrolyte, and therefore, O2
−
ated to CO on the Fe O4 species and in part decomposed further to
removal from the catalyst surface, increases also, facilitating further
3
carbon, which may result in catalyst deactivation by deposition of
carbonaceous material. Therefore, decomposition of adsorbed car-
bon dioxide (Eq. (9)) into carbon on magnetite species occurs in
two steps (Eq. (10) and (11)). Simultaneously, Fe O may capture
RWGS reaction by the electrochemically enhanced CO dissociative
adsorption to CO and subsequent CO hydrogenation to oxygenates
2
and hydrocarbons. Therefore, the increase in CO conversion levels
2
observed under negative polarization could be due to the improved
3
4
the oxygen released by CO2 and CO dissociation and converts itself
into Fe O (inactive for RWGS and carbon deposition), which may
CO2 dissociation, via O2 abstraction.
−
2
3
CO2 (ads.) + 4e → C (ads.) + 2O2−ꢀG◦ = 394.4 kJ mol−1 (12)
∗
-
∗
also result in catalyst deactivation. The amount of carbon on the
surface of the catalyst may be considered an indication of the dis-
sociative adsorption of CO . Propensity for CO2 decomposition is
reported to be higher over zirconium catalysts [44].
Given that there is no O in the gas phase, the reported [5] short
2
2
lifetime of the O2 species on the catalyst surface under reducing
−
conditions and the limited O2 capacity of the YSZ support, the sta-
−
◦
∗−
−1
CO + e → CO
(ads.)ꢀG = 183.3 kJ mol
(9)
(10)
(11)
2
2
bility of the solid electrolyte (YSZ) under working conditions can
◦
∗
−
∗
2−
−1
2−
CO2 (ads.) + e → CO (ads.) + O ꢀG = 257.2 kJ mol
CO (ads.) + e → C (ads.) + O ꢀG = 137.2 kJ mol
be questionable. However, the O species in the solid electrolyte
can be replaced by pumping of O2 species, resulting from elec-
−
◦
∗
∗
2−
−1
trochemical reduction of CO /adsorbed CO on the corresponding
2
CO conversion could decrease with temperature also as a result
electrode (gold or iron), to the solid electrolyte, allowing the solid
electrolyte to be stable under EPOC working conditions. In fact, it
has been previously demonstrated [13] that, upon positive polar-
2
of the concomitant increase in the extent of carbon deposition
and in the stability of the surface carbon species formed [52]. It
has been reported [47,53] that at the beginning of the reaction
ization, O2 ions could be generated at the gold counter electrode
−
(
at low temperature), carbon deposition on the catalyst surface
by electrochemical reduction of CO with the subsequent formation
2
takes place dominantly with a little formation of CO. On increas-
ing temperature, reverse water gas shift reaction yield increases
and products from RWGS dominate during ongoing carbon depo-
sition. Upon additional increase in temperature, the RWGS activity
of CO adsorbed on the surface and also possibly by subsequent elec-
trochemical reduction of the adsorbed CO species forming C surface
species, followed by removal of O2 from the gold film surface to
−
2−
the solid electrolyte. Upon negative polarization, O ions can be

5
2
E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
originated by same phenomena taking place at the iron catalyst
electrode.
Fig. 3(b) shows the effect of temperature on the selectivity to
the target oxygenated products (CH OH and C H O) under open-
3
2
6
circuit state and upon cathodic (–2 V) polarization conditions, at
◦
−1
temperatures from 225 to 400 C and using 90 l h of total gas flow
rate and a H /CO ratio of 3.
2
2
The selectivity to CH OH exhibited a maximum at a given
3
temperature and then decreased with the temperature incre-
ment, which agree with the fact that the equilibrium conversion
of CO2 to methanol decreases as temperature increases [49,50].
A H /CO ratio of 3 matched with the theoretical stoichiometry
2
2
in the methanol and dimethyl ether synthesis reactions by CO2
hydrogenation according to Eqs. (13)–(14), and thus they are ther-
modynamically favoured under these conditions. Moreover, C H O
2
6
can be also formed by dehydration (Eq. (15)) of CH OH.
3
In fact, as reported by literature [57], methanol converted to
◦
dimethyl ether at 250–300 C and both methanol and dimethyl
ether are reported to be transformed into hydrocarbons at temper-
◦
atures about 400 C [57], where RWGS (Eq. (2)) and methanation
(
Eq. (16)) reactions are also favoured [5].
◦
−
1
CO + 3H → CH OH + H OꢀG = −9.1 kJ mol
(13)
(14)
(15)
(16)
2
2
3
2
◦
−
1
2
CO + 6H → C H O + 3H OꢀG = −36.8 kJ mol
2
2
2
6
2
2
CH OH → C H O + H O
3
2
6
2
◦
−1
CO + 4H → CH + 2H OꢀG = −130.7 kJ mol
2
2
4
2
In the absence of surface oxygen ions (no polarization) and
under anodic (+2 V) polarization, CO2 conversion shows a maxi-
◦
mum (around 7% and 3%, respectively) at about 225 C. Cathodic
(
−2 V) polarization leads to a slight increase in CO2 conversion
◦
which reaches about 8% at an higher temperature of about 350 C,
where the effect of electrochemical promotion is more pronounced
due to the increased ionic conductivity of the solid electrolyte at
higher temperatures.
In the absence of surface oxygen ions (no polarization), the
selectivity to CH OH attained a maximum of about 19% around
3
◦
3
50 C, then decreased with the temperature increment, whereas
selectivity to C H O increased with temperature from about 4% at
2
6
◦
◦
2
25 C to around 20% at 400 C. Application of negative polariza-
tion (−2 V) leads to a slight increase in the maximum of CH OH
3
selectivity (20%), and it is obtained at lower temperature (around
◦
3
00–325 C). Selectivity to C H O is noticeably electrochemically
2
6
enhanced over the studied temperature range, showing the same
◦
tendency against temperature, increasing from about 10% at 225 C
to around 50% at 400 C.
Fig. 4. Influence of the applied potential on (a) CO2 conversion/TOF (᭿) and CO2
rate enhancement ratio (᭹), on (b) selectivities to CH3OH (᭿), C2H6O (᭹), CO (ꢀ)
and C2H5OH (ꢁ) and on (c) current (᭿) and apparent faradaic efficiency (᭹) over
◦
As commented above, negative polarization results in higher
levels of CO2 conversion and selectivity to CH OH and C H O with
◦
−1
Fe-TiO2/YSZ/Au. (325 C, 90 l h , H2/CO2 = 3).
3
2
6
respect to that under open circuit conditions, because CO2 disso-
ciative adsorption, which is reported to be the limiting step in CO2
hydrogenation [5,8,9,58], is facilitated by oxygen removal from the
catalyst surface.
ratios of 2 and 4, which correspond to a lack and an excess of H2,
respectively, in relation to the stoichiometry of the target products
(CH OH and C H O) formation reactions ((Eq. (10)) and (Eq. (11)),
3
2
6
Results obtained at different temperatures revealed that CO2
conversion and selectivity to the different target products were
respectively).
The steady state response of CO2 conversion, Turn Over Fre-
quency (TOF) and CO2 rate enhancement ratio to different applied
catalyst potentials (between −1.5 and 1.5 V) is displayed in Fig. 4a
◦
maximum at temperatures around 325–350 C. Therefore, a tem-
◦
perature value of 325 C was selected for subsequent experiments.
◦
−1
at 325 C and 90 l h and using a H /CO ratio of 3. Fig. 4b shows
2
2
3.2.2. Effect of potential
the effect of applied potential on the selectivity to CO, CH OH,
3
The steady state effect of polarization on the behaviour of Fe-
C H O, and C H5OH, respectively, whereas, Fig. 4c shows the vari-
2
6
2
TiO /YSZ/Au was investigated through potentiostatic experiments
ation of apparent faradaic efficiency (|ꢂ|) and current with applied
2
performed at different H /CO ratios (2–4) at the selected temper-
potential, under the same conditions.
2
−1
2
◦
ature of 325 C and 90 l h
.
As commented above, for H /CO2 ratios of 3, the hydrogenation
2
In order to consider the less favourable, although more realistic,
conditions associated with a discontinuous flow of renewable H2
of CO2 over Fe-TiO /YSZ/Au gives rise mainly to the formation of
methanol (Eq. (13)), dimethyl ether (Eq. (14)), ethanol (Eq. (17))
2
[
2], the behaviour of the catalyst was also studied using H /CO2
and CO (Eq. (2)), but (not shown) hydrocarbons (mainly C H6 and
2
2

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
53
C H ), CH , formic acid and acetic acid were also detected to be
which corresponds with a decrease in C H5OH formation of up to
3
6
4
2
formed.
approximately 50%.
Chemisorption of reactive molecules on a catalyst surface is
the previous step to any catalytic process. Chemisorption of an
adsorbate on a metal gives rise to a real chemical bond, thus imply-
ing electron donation from adsorbate to metal or from metal to
adsorbate. In the first case, the adsorbate is called electron donor
2
CO + 6H → C H5OH + 3H O
(17)
2
2
2
2
Moreover, C H5OH can be also formed by hydrogenation (Eq. (18))
and hydration (Eq. (19)) of CH OH.
2
3
(
electropositive), whereas in the second, is called electron accep-
2
2
CH OH + 3H → C H5OH + H O
(18)
(19)
3
2
2
2
tor (electronegative). There is a certain scale of electronegativity or
electron acceptor capacity, in which oxygen is one of the strongest
electron acceptors. The concept of electropositive or electronega-
tive is also valid for the promoter ions in charge of the phenomenon,
being these cations (electropositive promoters) and anions (elec-
tronegative promoters), respectively [9].
The origin of the electrochemical activation relies on the modifi-
cation of adsorption phenomena. In the electrochemical promotion,
promoter species are ions which electrochemically migrate in a
controlled manner from the support (solid electrolyte) to the metal.
The migration of these ions are accompanied by that of the corre-
sponding charge compensation ion, forming neutral surface dipoles
which are distributed along the metallic surface giving rise to the
ꢁeffective double layerꢂ. Formation of the effective double layer
gives rise to a change in the work function of the metal catalyst (ꢇ),
modifying its bond capacity which each reactive, and therefore its
catalytic behaviour [9,59].
CH OH + H O → C H5OH
3
2
2
In fact, a H /CO ratio of 3 is about the stoichiometric ratio
2
2
required for CH OH, C H5OH and C H O production and thus these
reactions are thermodynamically favoured and are in competition
over the catalyst surface under these conditions.
In this case, CO2 hydrogenation is considerably influenced by
the applied potential, with selectivities to C H O, C H5OH, CO and
CH OH up to about 47%, 10%, 12% and 50%, respectively. CO2 rate
enhancement ratio values are rather low (up to about 4) with
CO2 conversions up to 15%. TOF exhibited higher values (up to
5
alyst (Cu/TiO /YSZ/Au), which evidenced a better performance of
the present catalyst system.
As can be observed in Fig. 4c, |ꢂ| > 1 for whatever value of
applied potential. The maximum value of |ꢂ| is obtained for −0.5 V
and resulted to be 209. Such a high apparent Faradaic efficiency is
a distinguishing feature of electrochemical promotion vs. electro-
catalysis [9].
3
2
2
6
2
6
2
3
−
1
.5 s ) than those reported [5] for a similar electrochemical cat-
2
On the one hand, the increment in catalyst work function of the
metal catalyst by addition of an electronegative promoter (O2 for
YSZ), via application of positive potentials, favoured the adsorption
−
The catalyst showed an “inverted volcano type” electrochemical
behaviour, i.e., CO2 rate enhancement ratio exhibits a minimum at
certain potential (at around 0 V) [9,59] (Fig. 4a). This behaviour can
be attributed to the increase of hydrogen (electron donor) and CO2
of electron donor species (H ), because the electrode becomes pos-
itively charged (due to a defect of electrons), favouring the transfer
2
of electrons from electron donor molecules, like H , to the cata-
2
lyst electrode and thus the adsorption of the latter on the catalyst
surface, whereas at the same time hindered the adsorption of elec-
tron acceptor species (CO2 in this case), giving rise to an increase
in H2 (electron donor) coverage and to a decrease in the coverage
of CO2 (electron acceptor). The opposite occurs for application of
(
electron acceptor) adsorption strength and surface coverage upon
positive and negative polarization, respectively, as expected from
the rules of electrochemical promotion [5,13]. The difference in
the rate enhancement observed for negative and positive potential
application can be attributed to the difference in the chemisorption
propensity of CO2 and hydrogen on Fe. CO2 is less easily adsorbed
2−
negative potentials which results in the removal of O promoter
species from the catalyst electrode to the solid electrolyte, giving
rise to a decrease of the catalyst work function, i.e., the electrode
becomes negatively charged (due to an excess of electrons), which
favours the transfer of electrons from the catalyst to electron accep-
on Fe compared to H , thus enhancement of its chemisorption via
2
negative polarization is expected to have a more significant impact
on the rate [11].
However, selectivity to C H O diminished with potential
tor molecules, like CO , and, thus, the adsorption of the latter on
2
6
2
decrease (Fig. 4b), i.e., an electrophobic electrochemical behaviour
is observed, while selectivity to CO increases with decreasing cat-
alyst potential (Fig. 4b), showing an electrophilic electrochemical
behaviour [9,59].
the catalyst surface, whereas at the same time hindered the adsorp-
tion of electron donor species (H2 in this case), giving rise to an
increase in CO2 (electron acceptor) coverage and to a decrease in
the coverage of H2 (electron donor) [9–11,13,59].
As can be deduced also from Fig. 4b, the selectivity to C H5OH
In this way, depending on the relative electronegativity of the
different adsorbates involved in the reaction and on which of them
is in excess over the catalyst surface, the application of polariza-
tion will have a positive or negative effect on the overall kinetics of
the process [9,59]. Namely, the electrochemical promotion allows
modifying the adsorption of the different reactive molecules over
the catalyst-electrode surface. Thus, there is a given value of poten-
tial or promoter coverage which optimizes catalyst activity and
selectivity to the desired product, which depends on temperature
and gas composition.
2
follows opposite trends vs. applied potential with respect to that
of CO selectivity, whereas C H O production is strongly enhanced
2
6
at positive potentials, being a competition for the formation of the
different hydrogenation products over the catalyst surface. In fact,
as shown in Fig. 4b, with increasing catalyst potential by approxi-
mately 1 V (from 0 to 1 V), the selectivity to C H O increases by 1.4
2
6
times. As can be also observed in Fig. 4b, with decreasing catalyst
potential by about 1.5 V (from 1 to −0.5 V), CO selectivity increases
by about 100 times and C H5OH selectivity decreases by about 2
2
times.
Therefore, the observed electrochemically assisted catalytic
behaviour can be rationalized considering the effect of varying
applied potential on the chemisorptive bond strength of reac-
tants and intermediate surface species [9,59] and in accordance
As can be observed also in Fig. 4b, the optimum applied poten-
tial for CH OH formation is about −0.5 V, which corresponds with
3
an enhancement in their selectivity of up to about 4 times (with
respect to unpromoted open circuit conditions). Increasing or
decreasing applied potential from this value resulted in a decrease
in CH OH selectivity, i.e., selectivity to CH OH exhibited a maxi-
with the mechanisms proposed for the CO hydrogenation reaction
2
[44,47,50–52,60]. CO2 can be chemisorbed and at least partly dis-
sociated to CO on iron species (Fe and Fe O ) on a clean Fe catalyst
3
3
3
4
mum at this potential, showing a “volcano type” electrochemical
surface [44].
The pumping of O2 species away from the catalyst surface
−
behaviour. C H5OH exhibit a concomitant minimum at this poten-
2
tial, showing an “inverted volcano type” electrochemical behaviour,
to the solid electrolyte YSZ, enhances CO2 (electron acceptor)

5
4
E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
chemisorption (Eq. (9)) and dissociation to CO (Eq. (10). In fact,
under application of negative potentials, a strengthening of the
In fact, on decreasing applied potential, a strengthening of the
2−
Fe-(CO )* bond strength is achieved [13], since O
species are
2
2−
Fe-(CO )* bond strength is achieved [13], since O
species are
pumped away from the catalyst surface to the solid electrolyte
YSZ), enhancing CO2 adsorption and activation (Eq. (9)). Nega-
pumped away from the catalyst surface to the solid electrolyte
(YSZ), enhancing CO2 adsorption and activation (Eq. (9)), whereas
H2 dissociative adsorption (Eq. (20)) on Fe surface is hindered.
Lowering applied potentials also strengths Fe-C bond and weak-
ens C O bond enhancing CO2 dissociation to CO (10), resulting in
an enhancement of CO formation [10,11,13]. Moreover, CO disso-
ciative adsorption is less favoured at lower potentials (CO has a
2
(
tive polarization also strengths Fe-C bond and weakens C O bond
enhancing CO2 dissociation to CO (Eq. (10)).
The application of positive potentials give rise to a migration of
oxygen ions to the catalyst surface, favouring H2 (electron donor)
dissociative adsorption (Eq. (20)) on Fe surface, resulting in an
increase of the surface coverage of hydrogen and in a decrease in
CO2 coverage, being CO2 (electron acceptor) adsorption the reac-
tion limiting step.
lower electron acceptor capacity than CO ) [9,11,59,62] and it tends
2
to desorb (Eq. (22)). Both phenomena result in an increase in the
selectivity to CO.
Moreover, it can be observed that under application of
decreasing positive potentials formation of CH3OH is increasingly
enhanced reaching a maximum at about −0.5 V. As commented
above, around open circuit conditions (−0.5 V), at low negative
◦
H → 2H∗ (ads.) + 2eꢀG = 38.9 kJ mol
+
−1
(20)
2
−
At high O2 coverage or positive potentials, CO2 dissociative
adsorption on Fe is considered to be enhanced by the presence of
coadsorbed hydrogen (Eq. (21)) [60].
or positive potentials, selectivity to CH OH exhibited a maximum,
3
showing a “volcano type” electrochemical behaviour. In fact, forma-
tion of CH OH is reported to be enhanced over almost promoter free
According to literature [44,47,50–52,60], the mechanism of CO2
hydrogenation on Fe based catalyst implies firstly CO2 dissociative
adsorption to adsorbed CO (Eqs. (10) or (21)) on the catalyst surface.
CO adsorbed species resulting from CO2 dissociation may desorb
3
metal surfaces (around 0 V), where surface coverage of both reac-
tants (CO and H ) is expected to be very similar. Therefore, a strong
2
2
competition between CO2 and H2 adsorption could be expected
within this potential range, and, thus, according to this mechanism,
in the gas phase (Eq. (22)) or can react with H2 to form CH OH
3
it is expected that the formation of CH OH will be limited by the
(
Eq. (23)), C H O (Eq. (24)) or CH (Eq. (25)) adsorbed species.
3
2
6
2
dissociative adsorption of both H2 and CO2 [63,64].
CH2 adsorbed species are regarded as carbon–carbon propagation
species with chain growth taking place, presumably, by addition
of CH2 units, resulting in the formation of hydrocarbons (mainly
A subsequent decrease in potential below −0.5 V, gives rise to
an enhancement of CO2 dissociative adsorption on Fe surface, via
formation of adsorbed CO species, and to a decrease in H2 disso-
ciative adsorption, because H2 adsorption is retarded and tends to
be evolved from the catalyst surface [5,13]. Therefore, the amount
of hydrogen adsorbed on the surface is reduced and could limit
CO2 hydrogenation reactions. CO selectivity increases on decreas-
ing applied potential, because CO formation (via enhanced CO2
dissociative adsorption) and desorption (given that CO has a lower
electron acceptor capacity than CO2) are favoured. As commented
C H ) (Eq. (26)). Subsequent hydrogenation of CH species gives
3
6
2
rise to adsorbed CH species (Eq. (27)) which can be further hydro-
3
genated to CH4 (Eq. (28)) or can react with adsorbed CO to form
CH CO adsorbed species (Eq. (29)) which subsequent hydrogena-
3
tion results in ethanol formation (Eq. (30)) [60].
CO (ads.) + e + 2H+ → CO (ads.) + H O
∗
-
∗
(21)
(22)
(23)
(24)
(25)
(26)
(27)
(28)
(29)
(30)
2
2
∗
CO (ads.) → CO
above, selectivity to C H5OH selectivity exhibited a concomitant
2
∗
CO (ads.) + 4H(ads.) → CH OH
minimum at the same potential value (−0.5 V) as selectivity to
3
CH OH exhibited a maximum, resembling an “inverted volcano
3
∗
3
CO (ads.) + 6H(ads.) → CH OCH + CO
3
3
2
type” electrochemical behaviour. Therefore, selectivity to C H5OH
2
∗
∗
increases on decreasing applied potential below −0.5 V due to
CO (ads.) + 4H(ads.) → CH (ads.) + H O
2
2
the enhanced adsorbed CO, and thus CH CO*, species formation,
3
∗
2−
3
CH (ads.) → C H
as a result of the improved CO2 dissociation via O abstraction,
2
3
6
because, in spite of hydrogen evolution (which may limit the CO2
hydrogenation) is also favoured on decreasing applied potential,
∗
∗
CH (ads.) + H(ads.) → CH (ads.)
2
3
∗
under a stoichiometric H /CO ratio of 3, hydrogen coverage is still
CH (ads.) + H(ads.) → CH
2
2
3
4
enough [13] for CH CO* adsorbed species to be hydrogenated to
3
∗
∗
∗
CH (ads.) + CO (ads.) + H(ads.) → CH CO (ads.)
2
3
C2H5OH, given that CH3OH formation is inhibited (sharply decrease
to almost zero) for negative potentials [63]. As a consequence,
C2H5OH formation competitive reaction is increasingly enhanced,
∗
CH CO (ads.) + 3H(ads.) → C H5OH
3
2
At high positive potentials (hydrogen adsorption region), hydro-
gen coverage is supposed to be higher than that of CO adsorbed
species resulting from CO2 dissociation [13], and, thus, they can be
gradually hydrogenated to C H O and C H5OH.
It can be observed that under application of decreasing posi-
tive potentials to negative values (up to about −0.5 V), a migration
of oxygen ions from the catalyst surface to the solid electrolyte is
produced, decreasing oxygen surface coverage [9]. As a result, the
resulting in an increase in C H5OH selectivity, with the decrease
2
in potential. On the contrary, selectivity to C H O decreases for
2
6
potentials below −0.5, because, as commented above, the coverage
of both hydrogen and CO adsorbed species decreases on lowering
applied potential resulting in a decrease in selectivity to C H O
2
6
2
2
6
[2,12,13].
The steady state response of CO2 conversion, Turn Over Fre-
quency (TOF) and CO2 rate enhancement ratio to different applied
◦
−1
adsorption of electron acceptor species, as CO (Eq. (9)), is favoured
catalyst potentials (between −1.5 and 1.5 V), at 325 C and 90 l h
2
[
59] due to its higher electronegativity [61], increasing their surface
and using a H /CO ratio of 4, is displayed in Fig. 5a. Fig. 5b shows
2 2
coverage at the expense of electron donors (hydrogen) and giving
rise to a progressive increase in surface CO formation by dissocia-
the effect of applied potential on the selectivity to CO, CH OH,
3
C H O, C H5OH and C H , respectively, whereas Fig. 5c shows the
2
6
2
3
6
tive adsorption of CO , whereas H2 (electron donor) dissociative
variation of apparent faradaic efficiency (|ꢂ|) and current with
2
adsorption on Fe is hindered and therefore hydrogen evolution is
favoured [5,13]. As a result, CO selectivity increases and selectivity
to C H O and C H5OH decreases on decreasing catalyst potential
applied potential, under the same conditions.
As can be observed in Fig. 5b, for H /CO2 ratios of 4, the hydro-
2
◦
−1
genation of CO2 over Fe-TiO /YSZ/Au, at 325 C and 90 l h , gives
2
6
2
2
[
13].
rise also to the formation of CO (Eq. (2)), CH OH (Eq. (13)), C H O
3
2
6

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
55
C H is reported to be formed from CH radicals (Eq. (26)) result-
3
6
2
ing from hydrogenation of CO adsorbed species deposited by CO2
dissociation. The increase in the formation of C H6 is thought to
3
be due to the presence of an excess of H2 with respect to that sto-
chiometrically required for CH OH and C H O formation (H /CO
3
2
6
2
2
ratio of 3), i.e., due to an increased H2 availability in the system. In
fact according to literature [47,49], higher hydrocarbon selectivity
is attributable to a hydrogen excess in the reaction system [47,49].
Selectivity to C H diminished with potential decrease (Fig. 5b),
3
6
i.e., an electrophobic electrochemical behaviour is observed,
because, as commented above, on decreasing applied potential, H2
(electron donor) dissociative adsorption is hindered and hydrogen
evolution is favoured [5,13], limiting adsorbed CO species hydro-
genation to CH , and thus, C H formation. In addition, as more
2
3
6
water is formed, the amount of surface CO generated by CO2 dis-
sociation could decrease also trough the WGS (water gas shift) (Eq.
(
32)) [65], resulting in a concomitant decrease in C H6 yield.
3
◦
−
1
CO + H O → CO + H ꢀG = −19.9 kJ mol
(32)
2
2
2
Moreover, it can be observed that under application of decreas-
ing positive potentials within the positive range, formation of
CH OH and C H O is increasingly promoted reaching a maximum
3
2
6
at about 0.5 V. As indicated above, under application of high pos-
itive potentials, H2 (electron donor) dissociative adsorption on
Fe surface is favoured, being CO2 (electron acceptor) adsorption
the reaction limiting step. CO2 dissociation is considered to be
enhanced by the presence of adsorbed hydrogen. Hydrogenation
of CO adsorbed species resulting from CO2 dissociation leads to
the formation of these oxygenated products. Decreasing catalyst
potential results in the migration of oxygen species to the solid
electrolyte favouring CO2 (electron acceptor) dissociate adsorp-
tion on Fe surface, whereas H2 dissociative adsorption is hindered.
As a result, selectivity to CH OH and C H O increases up to a
3
2
6
maximum value, which corresponds with a maximum enhance-
ment in its selectivity of up to 50 and 1.7 times, respectively. In
this case, selectivities to CH OH and C H O showed similar trends
3
2
6
against potential, because in spite of hydrogen dissociative adsorp-
tion is hindered on decreasing applied potentials, for H /CO ratios
2
2
of 4 (higher than the stoichiometric ratio) hydrogen coverage is
always higher than that of CO adsorbed species resulting from
CO2 dissociation [56] and thus CO adsorbed species can be gradu-
ally hydrogenated to both compounds. Therefore, the competition
between hydrogenation reactions is less pronounced. A subsequent
Fig. 5. Influence of the applied potential on (a) CO2 conversion/TOF (᭿) and CO2
rate enhancement ratio (᭹), on (b) selectivities to CH3OH (᭿), C2H6O (᭹), CO (ꢀ),
C2H5OH (ꢁ) and C3H6 (ꢃ) and on (c) current (᭿) and apparent faradaic efficiency (᭹)
decrease in potential results in a decrease in selectivity to CH OH
◦
−1
3
over Fe-TiO2/YSZ/Au. (325 C, 90 l h , H2/CO2 = 4).
and C H O, because the hydrogen evolution reaction is increasingly
2
6
favoured, and the amount of hydrogen adsorbed on the surface is
reduced limiting hydrogenation reactions.
(
Eq. (14)), C H5OH (Eq. (17)) and C H (Eq. (31)), and it is, as well,
2 3 6
In general, the variation of C H5OH and CO selectivities vs.
affected by the applied potential, with selectivities to CO, CH OH,
C H O, C H5OH and C H up to about 10%, 5%, 25%, 15% and 50%,
respectively.
2
3
applied potential is very similar to that observed for a H /CO ratio
2
2
2
6
2
3
6
of 3.
In this case the maximum in C H and C H5OH formation corre-
3
6
2
◦
−
1
3
CO + 9H → C H + 6H OꢀG = −177 kJ mol
(31)
sponds with an enhancement in their selectivity of up to about 1.4
and 150 times, respectively, whereas the concomitant minimum in
CO formation corresponds with a decrease in its selectivity of up to
approximately 100 times.
2
2
3
6
2
CO2 rate enhancement ratio values are again rather low (up
to about 1.6) with CO conversions up to 6% and TOF up to
1
electrochemical behaviour, showing also a minimum in CO2 rate
enhancement ratio at about 0 V (Fig. 5a) [9,59].
As can be observed in Fig. 5c, |ꢂ| > 1 for whatever value of
applied potential. The maximum value of |ꢂ| is obtained for −0.5 V
and resulted to be 1570, demonstrating the non-Faradaic nature of
the process [9].
2
−
1
.7 s . The catalyst resembled, as well, an “inverted volcano type”
Fig. 6 shows the steady state effect of the applied potential on
CO conversion, Turn Over Frequency (TOF) and CO rate enhance-
2
2
ment ratio (Fig. 6a) and on the selectivity to CO, CH OH, C H O and
3
2
6
◦
−1
C H5OH (Fig. 6b), respectively, at 325 C and 90 l h and using a
2
H /CO2 ratio of 2, whereas Fig. 6c shows the variation of apparent
2
faradaic efficiency (|ꢂ|) and current with applied potential, under
the same conditions.
For this high H /CO2 ratio, the performance of the catalyst is
As can be observed in Fig. 6b, for H /CO2 ratios of 2, the hydro-
2
2
◦
−1
similar to that obtained for a H /CO ratio of 3, but the main product
genation of CO2 over Fe-TiO /YSZ/Au, at 325 C and 90 l h , gives
2
2
2
obtained is C H . Therefore, the selectivity to C H has been also
rise to the formation of CO (Eq. (2)), CH OH (Eq. (13)), C H O (Eq.
3
6
3
6
3
2
6
plotted consequently.
(14)) and C H5OH (Eq. (17)), being a competition for the formation
2

5
6
E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
For this lower H /CO2 ratio, the electrochemical behaviour is
2
similar to that obtained for an H /CO2 ratio of three, but CO is pro-
2
duced dominantly for all range of potentials, due to the defect of H2
[
49] with respect to that required by stoichiometry for methanol
and dimethyl ether formation (H /CO ratio of 3) and closer to that
2
2
required for reverse water gas shift reaction (H /CO2 ratio of 1).
2
Under high positive potentials H2 adsorption is favoured, being
CO adsorption the reaction limiting step. In this case, on decreasing
2
catalyst potential CO selectivity increases and selectivity to C H O
2
6
and C H5OH decreases due to the improved CO formation (via
2
CO2 dissociation) and desorption (CO has a lower electron donor
capacity than H ), and enhanced hydrogen evolution on decreas-
2
ing potential. For H /CO2 ratios of 2 (lower than the stoichiometric
2
ratio for C H O and C H5OH formation) and on hydrogen adsorp-
2
6
2
tion region (positive potential) coverage of H is still higher enough
2
for CO adsorbed species to be partially hydrogenated. On hydrogen
evolution region (negative potential), H coverage is lower, in addi-
2
tion CO formation and desorption (CO has a lower electron acceptor
capacity than CO ) are favoured as a result of the enhanced CO2
2
dissociative adsorption, being RWGS the dominant reaction. How-
ever, as more water is formed from CO2 hydrogenation reactions,
the amount of CO could decrease also through the WGS (water gas
shift) (Eq. (32)) reaction [65], resulting in a concomitant decrease
in CO yield, giving rise to the observed decrease in CO selectivity
and providing surface CO2 and H2 adsorbed species needed for the
different oxygenates production.
In this case, selectivity to CH OH, C H O and C H5OH showed
3
2
6
2
the same tendency vs. applied potential variation as for an
H /CO2 ratio of 3, showing a “volcano type” (maximum at −0.5 V),
2
“electrophobic” and “inverted volcano type” (minimum at 0 V)
electrochemical behaviour, respectively, whereas CO selectivity
exhibited also a “volcano type” electrochemical behaviour, i.e. it
attained also a maximum at a given potential value (0 V).
In fact, as shown in Fig. 6b, with increasing catalyst poten-
tial by approximately 0.5 V (from 0 to 0.5 V), the selectivity to
C H O increases by 1.5 times. As can be also observed in Fig. 6b,
2
6
with increasing catalyst potential by about 1.5 V (from 0 to 1.5 V),
C2H5OH selectivity increases by about 12.5 times and CO selectivity
decreases by about 1.3 times.
As can be deduced also from Fig. 6b, the optimum applied poten-
tial for CH OH formation is about −0.5 V, which corresponds with
3
an enhancement in their selectivity of up to about 2.5 times (with
respect to unpromoted open circuit conditions).
One of the main challenges for advancing CO hydrogenation to
2
fuels is increasing the energy efficiency of the process, i.e., decreas-
ing the energy requirements for producing the products. The trend
is to maximize products yield with minimal energy input. How-
ever, as stated in the introduction part of the manuscript, there are
other aspects which may justify the application of the electropro-
Fig. 6. Influence of the applied potential on (a) CO2 conversion/TOF (᭿) and CO2
rate enhancement ratio (᭹), on (b) selectivities to CH3OH (᭿), C2H6O (᭹), CO (ꢀ)
and C2H5OH (ꢁ) and on (c) current (᭿) and apparent faradaic efficiency (᭹) over
◦
−1
Fe-TiO2/YSZ/Au. (325 C, 90 l h , H2/CO2 = 2).
moted CO hydrogenation to fuels even with low energy efficiency,
2
because this technology offers a way for chemical storage of the
surplus of intermittent renewable energy that would otherwise be
wasted, turning the excess electricity, through electrolysis of water,
into renewable hydrogen which can be used to produce fuels.
From the above commented results, it seems that the control
of the yield of the CH OH, C H5OH and C H O can be carried out
of the different hydrogenation products over the catalyst surface,
and it is, as well, perceptibly influenced by the applied potential,
with selectivities to CO, CH OH, C H O and C H5OH up to about
3
2
6
2
8
0%, 6%, 5% and 5%, respectively.
3
2
2
6
CO2 rate enhancement ratio values are rather low (up to about
.5). CO2 conversion is also fairly low (up to 5%) with TOF values
by modification of the applied potential, i.e., there is an optimum
value of potential, at a given flow rate, temperature and H /CO and
1
2
2
up to 4.2. The catalyst showed also an “inverted volcano type” elec-
trochemical behaviour, i.e., CO2 rate enhancement ratio exhibits
a minimum at a certain potential (under open circuit conditions
approx.) (Fig. 6a) [8,59].
As can be observed in Fig. 6c, |ꢂ| > 1 for whatever value of
applied potential. The maximum value of |ꢂ| is obtained for −0.5 V
and resulted to be 44.5, demonstrating the non-Faradaic character
of the process [9].
ratio, which maximized CH OH, C H5OH or C H O selectivity.
3 2 2 6
The energy efficiency (energy balance) of the overall process
can be calculated as the ratio between useful energy that can be
produced by combustion of the synthesised fuel (output energy)
and the energy consumed in the formation of the product. The
energy produced can be calculated by multiplying the amount of
fuel produced by its calorific value, whereas the consumed energy
can be calculated as the sum of the thermal energy (needed to

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
57
increase temperature of reactives to that required for the reac-
tion) and electrical energy (needed to supply the current needed
for electropromotion of the process) necessary for obtaining the
fuel product. Therefore, the energy efficiency (energy balance) of
the overall process should be calculated for each target product
formation.
The electrical energy requirements for producing the different
products can be evaluated from the values of energy cost (kWh/Kg)
associated with each product formation calculated according to Eq.
(
8). In this work, the energy cost (kWh/Kg) has been calculated for
each product fuel (methanol, ethanol and dimethyl ether) forma-
tion under different operating conditions as an indication of the
energy efficiency of the process.
The minimum value of energy cost (kWh/Kg) associated with
each product formation was 0.004, 0.003 and 0.0005, for methanol,
ethanol and dimethyl ether, respectively. These values were
◦
obtained for −0.5 V at 325 C and using H /CO ratios of 3 (for
2
2
methanol and ethanol) and 4 (for C H O). In the case of methanol,
2
6
the minimum in energy cost coincides with the maximum in
CH OH selectivity, which makes methanol a candidate target prod-
3
uct, but also with a rather low CO conversion and high selectivity to
2
the rest of the products, which implies an additional cost associated
with product purification.
3
.2.3. Effect of H /CO2 ratio
2
For better understanding of the influence of H /CO2 ratio on
2
catalyst behaviour and selectivity to the different products, the
potentiostatic variation of CO2 conversion, CO2 rate enhancement
ratio and selectivity to CO, CH OH, C H5OH, and C H O are plotted
3
2
2
6
for different H /CO ratios in Fig. 7a–f, respectively.
2
2
Results (Fig. 7a) indicate that, in general, CO2 conversion is
slightly affected by H /CO2 ratio, except in the case of H /CO2 ratio
2
2
of three where the CO conversion versus potential curve is shifted
2
to the highest values of conversion, in special at highly negative
potentials.
As can be deduced from Fig. 7b, at positive potentials, in general,
CO2 rate enhancement ratios, i.e., promotion levels, decrease on
increasing H /CO ratio.
2
2
Selectivity to CO increases on decreasing H /CO ratio, as result
2
2
of the decreased hydrogen availability with respect to the stoichio-
metrically required for the synthesis of the different hydrogenated
products. In the case of H /CO2 ratio of two, the CO selectivity
2
vs. potential curve is shifted to the highest values (between 60
and 80%), presumably because an H /CO2 ratio of two is closer to
2
the value required by stoichiometry for RWGS reaction (about 1),
favouring CO production.
CH OH selectivity (Fig. 7d) exhibits a maximum around open
3
circuit conditions (+0.5–0.5 V), because the surface reaction could
be restricted by the coverage of both CO2 and H2 adsorbed on the
catalyst surface [63,64], and for H /CO ratios equal to three (of
2
2
about 50%), which corresponds with that required for stoichiomet-
ric synthesis of methanol by CO2 hydrogenation (Eq. (10)), and,
therefore, thermodynamically favours the formation of CH OH as
3
the expenses of other CO2 hydrogenation products. On the other
hand, for H /CO ratios distinct from three, selectivity to methanol
2
2
was almost unaffected by H /CO2 ratio [66].
2
Selectivity to ethanol (Fig. 7e) follows an opposite trend vs.
H /CO ratio with respect to methanol selectivity, showing a mini-
2
2
mum at about −0.5 V, which seems to indicate that they are formed
through competitive reactions. For other potential values, selec-
tivity to ethanol increases with H /CO ratio [67], as result of the
increased hydrogen availability with respect to the stoichiometri-
cally required for the ethanol synthesis reaction.
2
2
Fig. 7. Effect of H2/CO2 ratio on the potentiostatic variation of CO2 conversion (a),
CO2 rate enhancement ratio (b) and selectivities to CO (c), CH3OH (d), C2H5OH (e) and
C H O (f) on Fe-TiO /YSZ/Au. (325 ◦C, H /CO = 2 (᭿), H /CO = 3 (᭹) and H /CO2 = 4
2
6
2
2
2
2
2
2
−
1
As can be obtained from Fig. 7f, for an H /CO2 ratio of three, the
(ꢀ), 90 l h ).
2
C H O selectivity vs. potential curve is shifted to the highest values
2
6
(
between 25 and 50%), presumably because this H /CO ratio coin-
2 2

5
8
E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
As expected, the CO conversion values decreased upon increas-
2
ing gas flow rate as a result of the lower contact time available for
the reaction to take place, and, very interestingly, the methanol and
dimethyl ether selectivity also followed a negative trend [68,69].
This seems to suggest that both CH OH and C H O were produced,
3
2
6
to a large extent, from the CO hydrogenation, as a consecutive step
of the RWGS reaction, rather from the direct CO2 hydrogenation
reaction. On the contrary, selectivity to C H5OH and CO increases
2
on increasing gas flow rate. This could be indicative of the fact
that C H5OH and CO are produced in parallel paths through the
2
corresponding hydrogenation reactions, because a diminution in
selectivity to C H5OH with the increase in gas flow rate could
2
be expected if C H5OH was produced via consecutive hydration
2
or hydrogenation of the formed CH OH [70,71]. However, C H O
3
2
6
selectivity decreases with the increment in gas flowrate as CH OH
3
does, which seems to indicate that C H O is formed in any con-
2
6
secutive reaction from CH OH (via consecutive dehydration of
3
the formed CH OH) or from the same reaction intermediate (CO)
3
[
70,71].
The observed decrease in C H5OH selectivity at higher residence
2
times could resemble a possible inhibition of primary ethanol syn-
thesis, via hydrogenation of the adsorbed CO deposited by CO2
dissociation, by H O formed as side product of higher alcohols and
2
*
by hydrocarbons secondary reactions of adsorbed CO with CH2
radicals, which are favoured by long residence times [51,69].
4. Conclusions
In this study, a Fe-TiO /YSZ/Au tubular electrochemical catalyst
2
has been successfully prepared by dip-coating and characterised,
both after reduction and after testing, by Scanning Electron
Microscopy, X-Ray Photoelectron Spectroscopy and X-Ray Diffrac-
tion techniques.
The CO2 hydrogenation over Fe in an oxygen ion conduct-
ing membrane reactor can be electrochemically assisted under
atmospheric pressure, at relatively low temperatures and high gas
flow rates and under realistic postcombustion CO2 capture exit-
ing gas compositions. Selectivity to the different target fuels can be
modulated by modifying applied potential under given operating
conditions.
Fig. 8. Effect of gas flow rate on the potentiostatic variation of (a) CO2 conversion
᭿,ᮀ) and CO2 rate enhancement ratio (᭹,ꢃ), (b) selectivity to CH3OH (᭿,ᮀ) and
The hydrogenation of CO2 over the Fe catalyst leads to different
products depending on operating conditions. The main products
(
−
1
C2H6O (᭹,ꢃ) and (c) selectivity to CO (᭿,ᮀ) and C2H5OH (᭹,ꢃ) at 216 l h (filled
−1
◦
symbols) and 90 l h (open symbols) over Fe-TiO2/YSZ/Au. (325 C, H2/CO2 = 3).
were methanol, dimethyl ether, ethanol, CO and C H , but CH ,
3
6
4
C H , formic acid and acetic acid were also detected to be formed.
2
6
cides with that required for stoichiometric synthesis of C H O by
CO hydrogenation and selectivities to CH OH, C H O, C H and
2
6
2 3 2 6 3 6
CO2 hydrogenation (Eq. (14)), and, therefore, thermodynamically
favours the formation of C H O as the expenses of other CO hydro-
2
5
C H OH were electrochemically enhanced up to a maximum of 50,
2
6
2
1.7, 1.4 and 150 times, respectively.
◦
genation products. On the other hand, for H /CO2 ratios distinct
A temperature value of 325 C was selected to be the optimum
2
from three, selectivity to dimethyl ether decreased on increasing
for the electrochemically assisted CO2 hydrogenation process.
H /CO2 ratio.
Increasing gas flow rate, and correspondingly, decreasing con-
2
◦
As commented above, selectivity to C H6 reaches a maximum
tact time, for a given H /CO2 ratio of 3 and at 325 C, resulted in a
3
2
for a H /CO ratio of 4, because the amount of hydrogen is in excess
decrease in CO conversion and in CH OH and dimethyl ether selec-
2
2
2
3
with respect to that required by stoichiometry for the synthesis
reactions via CO2 hydrogenation [47].
2
5
tivities and in an increase in CO and C H OH selectivities, whereas
CO2 rate enhancement ratio is almost unaffected by the variation
of gas flow rate.
Results obtained at different H /CO ratios, at 90 l h 1 and
−
3.2.4. Effect of gas flow rate
2 2
◦
325 C, showed that CO conversion and selectivities to CH OH and
2 3
The influence of gas flow rate (the reciprocal of contact time)
◦
on CO2 hydrogenation behaviour, at 325 C and using a H /CO
C H O showed a maximum for a H /CO ratio of 3, whereas selec-
2 6 2 2
2
2
ratio of 3, is depicted in Fig. 8. Results indicate that increasing gas
flowrate, and correspondingly, decreasing contact time, resulted in
tivites to C H6 and CO strongly increase for a H /CO2 ratio of 4 and
3 2
2, respectively, as a result of the increased and decreased hydro-
a decrease in CO2 conversion (Fig. 8a) and in CH OH and dimethyl
gen availability in the reaction system, and selectivity to C H5OH
3
2
ether selectivities (Fig. 8b), as well as in an increase in CO and
increases with the increment in H /CO2 ratio, being the maximum
2
C H5OH selectivities (Fig. 8c), whereas CO rate enhancement ratio
values attained around 80%, 50%, 50%, 15% and 47% for CO, C H ,
2
2
3 6
is almost unaffected by the variation of gas flow rate (Fig. 8a).
CH OH, C H5OH and C H O, respectively.
3 2 2 6

E. Ruiz et al. / Catalysis Today 268 (2016) 46–59
[24] E. Ruiz, D. Cillero, P.J. Martínez, A. Morales, G. San Vicente, G. de Diego, J.M.
59
Therefore, this study addresses some scale-up aspects, such as
operation at high flow rates and atmospheric pressure, under real-
istic gas compositions and using catalyst-electrode configurations
easily adaptable to the existing catalytic devices (conventional flow
reactors), based on a cheap, widespread and non-precious Fe cat-
alyst, and prepared by commercial ready procedures, which may
have an impact on the potential practical application of the process
for CO2 recycling, contributing not only to controlling the global
Sánchez, J. CO2 Util. 8 (2014) 1–20.
25] P.R. Strutt, B.H. Kear, US Patent 8,173,327 B2.
26] D. Poulidi, A. Thursfield, I.S. Metcalfe, Top. Catal. 44 (2007) 435–449.
[
[
[27] A. Morales, EU EP 1 321 539 A2.
[
28] C. Pliangos, I.V. Yentekakis, S. Ladas, C.G. Vayenas, J. Catal. 159 (1996)
89–203.
1
[
29] E.A. Baranova, G. Foti, H. Jotterand, C. Comninellis, Top. Catal. 44 (2007)
355–360.
[30] W. Wang, J. Gong, Front. Chem. Sci. Eng. 5 (2011) 2–10.
31] C. Guizard, A. Princivalle, Catal. Today 146 (2009) 367–377.
32] E. Mutoro, C. Koutsodontis, B. Luerssen, S. Brosda, C.G. Vayenas, J. Janek, Appl.
Catal. B 100 (2010) 328–337.
[
[
“
Green-house Effect”, but also to the availability of fuel sources for
the future.
[
[
[
33] Y.-x Pan, C.-j. Ge, T.S. Liu, Q. Ge, Catal. Today 147 (2009) 68–76.
34] J.-S. Kim, S.-B. Lee, M.-J. Choi, K.-W. Lee, Catal. Today 115 (2006) 228–234.
35] M. Kisida, K. Umakoshi, J. -i. Ishiyama, H. Nagata, K. Wakabayashi, Catal. Today
Acknowledgements
29 (1996) 355–359.
The authors acknowledge support from Ministerio de Ciencia
e Innovación of Spain (Project ENE2010-15569). Pedro J. Martinez
is grateful to Ministerio de Ciencia e Innovación of Spain for the
research grant BES-2011-046902.
[36] A. de Lucas Consuegra, A. Caravaca, J. Gonzalez-Cobos, J.L. Valverde, F. Dorado,
Catal. Commun. 15 (2011) 6–9.
[
[
37] C.S. Chen, J.H. You, J.H. Lin, Y.Y. Chen, Catal. Commun. 9 (2008) 2381–2385.
38] X. Yao, Q. Yu, Z. Ji, Y. Lu, Y. Cao, C. Tang, F. Gao, L. Dong, Y. Chen, Appl. Catal. B
130–131 (2013) 293–304.
[
[
[
[
39] C.H. Bartholomew, R.J. Farrauto, Fundamentals of Industrial Catalytic
Processes, 2nd ed., John Wiley & Sons, Inc, New Jersey, 2006.
40] Z. Shu, S. Wang, J. Nanomater. 2009 (2009) 340217, http://dx.doi.org/10.1115/
Appendix A. Supplementary data
2009/340217.
41] K. Cheng, V.V. Ordomsky, M. Virginie, B. Legras, P.A. Chernavskii, V.O. Kazak, C.
Cordier, S. Paul, Y. Wang, A.Y. Khodakov, Appl. Catal. A 488 (2014) 66–77.
42] R.W. Dorner, D.R. Hardy, F.W. Williams, H.D. Willauer, Appl. Catal. A 373
(2010) 112–121.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2016.02.
025.
[
[
[
[
43] T. Yamashita, P. Hayes, Appl. Surf. Sci. 254 (2008) 2441–2449.
44] M. Niemelä, M. Nokosmaki, Catal. Today 100 (2005) 269–274.
45] J.S. Hwang, K.W. Jun, K.W. Lee, Appl. Catal. A 208 (2001) 217–222.
46] Z.L. Liu, H.B. Wang, Q.H. Lu, G.H. Du, L. Peng, Y.Q. Du, S.M. Zhang, K.L. Yao, J.
Magn. Magn. Mater. 283 (2004) 258–262.
References
[
[
1] G.A. Olah, A. Goeppert, G.K.S. Prakash, J. Org. Chem. 74 (2009) 487–498.
2] F. Ocampo, B. Louis, A. Kiennemann, A.C. Roger, Mater. Sci. Eng. 19 (2011)
[47] S. Seidi, N.A.S. Amin, M.R. Rahimpour, J. CO2 Util. 5 (2014) 66–81.
[48] T. Inui, M. Anpo, K. Izui, S. Yanegida, T. Yamaguchi, Advanced in Chemical
Conversions for Mitigating Carbon Dioxide, Elsevier Science B V, Amsterdam,
1998.
1
–11.
3] S.K. Hoekman, A. Broch, C. Robbins, R. Purcell, Int. J. Greenh. Gas Control 4
2010) 44–50.
[
(
[
[
4] D.T. Whipple, P.J. Kenis, J. Phys. Chem. Lett. (2010) 3451–3458.
5] E.I. Papaioannou, S. Souentie, A. Hammad, C.G. Vayenas, Catal. Today 146
[49] S. Zhanghuai, K. Yuan, J. Nat. Gas Chem. 9 (2000) 283–290.
[50] H.D. Willauer, R. Ananth, M.T. Olsen, D.M. Drab, D.R. Hardy, F.W. Williams, J.
CO2 Util. 3–4 (2013) 56–64.
(
2009) 336–344.
[
[
[
6] W. Wang, S. Wang, X. Ma, J. Gong, Chem. Soc. Rev. 40 (2011) 3703–3727.
7] A. Anastasijevic, Catal. Today 146 (2009) 308–311.
8] G. Karagiannakis, S. Zisekas, M. Stoukides, Solid State Ionics 162–163 (2003)
[51] D. Miller, M. Moskovits, J. Am. Chem. Soc. 111 (1989) 9250–9252.
[52] C.M. Finnerty, N.J. Coe, R.H. Cunningham, R.M. Ormerod, Catal. Today 46
(1998) 137–145.
3
13–318.
[53] H. Schulz, T. Riedel, G. Schaub, Top. Catal. 32 (2005) 117–124.
[54] S.J. Choe, D.H. Park, D.S. Huh, Bull. Korean Chem. Soc. 21 (2000) 779–784.
[55] C. Schild, A. Wokaun, R.A. Koeppel, A. Baiker, J. Phys. Chem. 95 (1991)
6341–6346.
[
9] C.G. Vayenas, S. Bebelis, C. Pliangos, S. Brosda, D. Tsiplakides, Electrochemical
Activation of Catalysis: Promotion Electrochemical Promotion and
Metal-Support Interactions, Kluwer Academic/Plenum Publishers, New York,
2
001.
[56] R. Sahki, O. Benlounes, O. Chérifi, R. Thouvenot, M.M. Bettahar, S. Hocine,
React. Kinet. Mech. Catal. 103 (2011) 391–403.
[
10] G. Pekridis, K. Kalimeri, N. Kaklidis, E. Vakouftsi, E.F. Iliopoulou, C. Athanasiou,
G.E. Marnellos, Catal. Today 127 (2007) 337–346.
11] S. Bebelis, H. Karisali, C.G. Vayenas, Solid State Ionic 179 (2008) 139–1395.
12] S. Bebelis, H. Karasali, C.G. Vayenas, J. Appl. Electrochem. 38 (2008)
[57] Y. Souma, H. Ando, M. Fujiwara, R. Kieffer, Energy Convers. Manag. 36 (1995)
593–596.
[
[
[58] S. Li, A. Li, S. Krishnamoorthy, E. Iglesia, Catal. Lett. 77 (2001) 197–205.
[59] S. Brosda, C.G. Vayenas, J. Wei, Appl. Catal. B 68 (2006) 109–124.
[60] H. Kusama, K. Okabe, K. Sayama, H. Arakawa, Energy 22 (1997) 343–348.
[61] F.J. Willians, M.S. Tikhov, A. Palermo, N. Macleod, R.M. Lambert, J. Phys. Chem.
B 105 (2001) 2800–2808.
[62] M. Kuriyama, H. Tanaka, S. Ito, T. Kubota, T. Miyao, S. Naito, K. Tomishige, K.
Kunimori, J. Catal. 252 (2007) 39–48.
[63] P. Dube, G.M. Brisard, J. Electroanal. Chem. 582 (2005) 230–240.
[64] G.M. Brisard, A.P.M. Camargo, F.C. Nart, T. Iwasita, Electrochem. Commun. 3
(2001) 603–607.
1
127–1133.
13] D. Theleritis, S. Souentie, A. Siokou, A. Katsaounis, C.G. Vayenas, ACS Catal. 2
2012) 770–780.
[
[
(
14] D. Theleritis, M. Makri, S. Souentie, A. Caravaca, A. Katsaounis, C.G. Vayenas,
ChemElectroChem 1 (2014) 254–262.
15] M. Makri, A. Katsaounis, C.G. Vayenas, Electrochim. Acta 179 (2015) 556–564.
16] I. Kalaitzidou, A. Katsaounis, T. Norby, C.G. Vayenas, J. Catal. 331 (2015)
[
[
9
8–109.
17] I. Kalaitzidou, M. Makri, D. Theleritis, A. Katsaounis, C.G. Vayenas, Surf. Sci.
46 (2016) 194–203.
[
[
6
[65] C.K. Kuo, N.D. Patel, A. Tan, P. Sarkar, P.S. Nicholson, Solid State Ionics 53–56
(1992) 564–570.
18] V. Jiménez, C. Jiménez-Borja, P. Sánchez, A. Romero, E.I. Papaioannou, D.
Theleritis, S. Souentie, S. Brosda, J.L. Valverde, Appl. Catal. B 107 (2011)
[66] J. Wu, M. Saito, M. Takeuchi, T. Watanabe, Appl. Catal. A 218 (2001) 235–240.
[67] W. Hua, W. Li, S. Xuelian, Coal Chem. Ind. 33 (2005) 30–35.
[68] Z. Huo, M. Hu, X. Zeng, J. Yun, F. Jin, Catal. Today 194 (2012) 25–29.
[69] M. Xu, M.J.L. Gines, A.M. Hilmen, B.L. Stephens, E. Iglesia, J. Catal. 171 (1997)
130–147.
2
10–220.
[
19] N. Gutiérrez-Guerra, J. González-Cobos, J.C. Serrano-Ruiz, J.L. Valverde, A. de
Lucas-Consuegra, Top. Catal. (2015), http://dx.doi.org/10.1007/s11244-015-
0
488-4 (published online).
[20] D. Tsiplakides, S. Balomenou, Chem. Ind. Chem. Eng. Q. 14 (2008) 97–105.
[21] D. Tsiplakides, S. Balomenou, Catal. Today 146 (2009) 312–318.
[22] E. Ruiz, D. Cillero, P.J. Martínez, A. Morales, G. San Vicente, G. de Diego, J.M.
Sánchez, Catal. Today 210 (2013) 55–66.
[70] A. Basonde, B. Tidona, P.R. von Rohr, A. Urakawa, Catal. Sci. Technol. 3 (2013)
767–778.
[71] M. Kasaie, M. Sohrabi, J. Mex. Chem. Soc. 53 (2009) 233–238.
[23] E. Ruiz, D. Cillero, P.J. Martínez, A. Morales, G. San Vicente, G. de Diego, J.M.
Sánchez, Catal. Today 236 (2014) 108–120.