Electrochemical promotion of the water-gas shift reaction on Pt/YSZ

Article abstract of DOI:10.1016/j.jcat.2011.07.009

The effect of electrochemical promotion of catalysis was investigated for the water-gas shift reaction over porous Pt catalyst electrodes interfaced with 8%mol Yttria-stabilized Zirconia. A fuel cell type electrochemical reactor was used at temperatures from 300 °C to 400 °C, under PH₂O/ ^PCO ratio values from 2.85 to 31. A negative order dependence of the catalytic reaction rate on P_CO and a positive one on PH₂O was found under open-circuit and polarization conditions. Positive potential application (+2.5 V), i.e., O^2- supply to the catalyst surface, causes a small decrease in the catalytic reaction rate, while negative potential application (-1.5 V) results in a pronounced rate increase, up to 200%, with apparent faradaic efficiency values up to 110. The rate increase obtained with negative polarization can be attributed to the weakening of the Pt-CO bond strength but also, to the increase in surface concentration of oxygen ion vacancies near the Pt-gas-support three-phase boundaries necessary for water dissociation.

Full text of DOI:10.1016/j.jcat.2011.07.009

Journal of Catalysis 283 (2011) 124–132
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Electrochemical promotion of the water–gas shift reaction on Pt/YSZ
a,
⇑
a
a
a
b
a
S. Souentie , L. Lizarraga , A. Kambolis , M. Alves-Fortunato , J.L. Valverde , P. Vernoux
a
Université de Lyon, Institut de Recherches sur la Catalyse et l’Environnement de Lyon, UMR 5256, CNRS, Université Claude Bernard Lyon 1, 2 Avenue A. Einstein,
9626 Villeurbanne, France
Departamento de Ingeniería Química, Facultad de Ciencias Químicas, Universidad de Castilla-La Mancha, Avenida Camilo José Cela 10, 13005 Ciudad Real, Spain
6
b
a r t i c l e i n f o
a b s t r a c t
Article history:
The effect of electrochemical promotion of catalysis was investigated for the water–gas shift reaction
over porous Pt catalyst electrodes interfaced with 8%mol Yttria-stabilized Zirconia. A fuel cell type elec-
Received 18 May 2011
Revised 27 July 2011
Accepted 28 July 2011
Available online 7 September 2011
trochemical reactor was used at temperatures from 300 °C to 400 °C, under P 2
2
H
O
=PCO ratio values from
.85 to 31. A negative order dependence of the catalytic reaction rate on PCO and a positive one on
H2O
was found under open-circuit and polarization conditions. Positive potential application (+2.5 V),
P
2ꢀ
i.e., O supply to the catalyst surface, causes a small decrease in the catalytic reaction rate, while neg-
ative potential application (ꢀ1.5 V) results in a pronounced rate increase, up to 200%, with apparent far-
adaic efficiency values up to 110. The rate increase obtained with negative polarization can be attributed
to the weakening of the Pt–CO bond strength but also, to the increase in surface concentration of oxygen
ion vacancies near the Pt-gas-support three-phase boundaries necessary for water dissociation.
Ó 2011 Elsevier Inc. All rights reserved.
Keywords:
Water–gas shift
Electrochemical promotion
EPOC
NEMCA effect
Pt electrode
YSZ
1
. Introduction
syn-gas feed at elevated temperatures (up to 500 °C) and pressures
(
19 bar).
The technologies of hydrogen production and purification will
The WGS pathway and the exact nature of any intermediates
be of great importance in the hydrogen-energy economy in the
near future. In the water–gas shift (WGS) reaction Eq. (1), hydro-
gen is produced from water in the presence of CO. This reaction
is a key technology in the hydrogen purification processes of
syn-gases obtained by steam reforming or partial oxidation of
hydrocarbons [1,2].
formed are still under discussion; however, it is generally agreed
that both the metal and the support play essential roles. Among
the two reactants, i.e., CO and H O, the latter is more difficult to
2
be activated due to its thermodynamic stability [4]. It has been re-
ported that Pt cannot activate water under WGS reaction condi-
tions [5]. Thus, in Pt-based WGS catalysts, a hydrophilic oxide
support is required to adsorb and activate water [4–9], while Pt
can adsorb and activate CO.
CO þ H
2
O ! CO
2
þ H
2
ð1Þ
The WGS state-of-the-art industrial process occurs in two steps
2
In the case of Pt/CeO catalysts, two main different reaction
to overcome the thermodynamic limitations at higher tempera-
pathways have been proposed in literature [6,10]: (i) the associa-
tive [11–14] and (ii) the regenerative red–ox mechanism
[5,7,15,16]. In the associative mechanism [11–14], terminal hydro-
xyl groups of the support react with CO on Pt to form surface for-
mate intermediate (reaction (2)). The decomposition of the
formate species is suggested to be facilitated by the presence of
water and is proposed to be rate determining (reaction (3)). No
oxygen is removed from the surface during the catalytic cycle,
and the oxide does not undergo any red–ox changes during the
WGS reaction sequence, since water activation occurs on the
formed formate intermediates.
ꢁ
tures (moderately exothermic reaction,
including a high-temperature (350–500 °C) shift (HTS) utilizing a
D
H ¼ ꢀ41:1 kJ=mol),
Fe low-temperature (200–
2
O
3
/Cr
2
O
3
catalyst, followed by
a
2
50 °C) one (LTS) over a Cu/ZnO/Al O catalyst [1]. However, a
2 3
single-step WGS process would be more desirable, avoiding the
technical complexities of a multistep process.
Supported noble metal (e.g., Pt, Pd and Au) catalysts are re-
ported as promising single-step WGS catalysts, due to their high
catalytic activity at low temperatures, tolerance in chemical poi-
soning (Cl and/or S), long-term stability at high temperatures and
regenerative potential after deactivation as a result of prolonged
use. Recently, a sulfur-resistant commercial WGS catalyst has been
reported by de la Osa et al. [3] using an industrial coal-derived
Pt ꢀ CO þ OHsup þ Hsup ! Pt ꢀ ½COOHꢂ ꢀ ssup þ Hsup
ð2Þ
2
þH O
Pt ꢀ ½COOHꢂ ꢀ ssup þ Hsup ! CO
2
ðgÞ þ H ðgÞ þ OHsup þ Hsup þ Pt
2
⇑
Corresponding author.
E-mail address: souentie@chemeng.upatras.gr (S. Souentie).
ð3Þ
0
021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.07.009

S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
125
In the regenerative red–ox mechanism [5,7,15,16], CO is adsorbed
on the metal particles and reacts at the metal-gas-support three-
phase boundaries (tpb) with the oxygen from the support, forming
When using YSZ as the solid electrolyte, the promoting ionic
dꢀ
+
species (O ꢀ d ) are generated in an electrochemical step at the
tpb (Eq. (9a)) which then spread, due the strong repulsive di-
pole–dipole interactions, over the entire metal–gas interface estab-
lishing there the neutral effective double layer,
CO
2
and creating a surface oxygen vacancy, V
O
(reaction (4)). The
support is then re-oxidized by H
2
O forming H (reaction (5)).
2
2
ꢀðYSZÞ ! ½Odꢀ ꢀ d ꢂPt þ 2e
þ
ꢀ
Pt ꢀ CO þ Osup ! CO ðgÞ þ Pt þ V
2
O
ð4Þ
ð5Þ
O
ð9aÞ
at a rate I/2F, where I is the current and F the Faraday’s constant.
In EPOC studies under reducing conditions, as is the present
case, it is likely that the promoting anionic species is a hydroxyl
groups formed via:
V
O
þ H
2
O ! Osup þ H ðgÞ
2
Recently, an ‘‘associative formate with red–ox regeneration’’ hy-
dride mechanism has been proposed for the case of Pt/ZrO
17,18]. In this mechanism, hydroxyl groups on the zirconia surface
react with CO to form intermediate formate. The latter can decom-
pose to H and CO , leaving oxygen vacancies that are filled by hy-
2
[
2ꢀðYSZÞ þ ð1=2ÞH
! ½OHdꢀ ꢀ d ꢂ þ 2e
þ
ꢀ
ð9bÞ
O
2
2
2
It must be clarified that the spillover species (commonly denoted
droxyl groups (reactions (6) and (7)), causing red–ox changes to the
support during the WGS reaction sequence.
dꢀ
+
O
ꢀ d in the EPOC literature) is overall neutral as the anionic oxy-
dꢀ
+
gen species O is accompanied by the image charge d in the metal.
dꢀ
+
H
sup
Consequently, the difference between the spillover O ꢀ d species,
which exists only at high coverages, and normally adsorbed oxygen
is only in the dipole moment (ꢃ2 vs. ꢃ1 Debye) and not in the total
charge which is zero in both cases [26].
Pt ꢀ CO þ OHsup ! Pt ꢀ ½COOHꢂ ꢀ ssup ! CO
2
ðgÞ þ H ðgÞ
2
þ Pt þ V
O
ð6Þ
ð7Þ
V
O
þ H
2
O ! Hsup þ OHsup
Three parameters are commonly used to quantify the magni-
tude of EPOC effect, the rate enhancement ratio,
26]:
q
, defined from
The exact role of Pt is still under discussion. Most groups [4–9,17]
propose that CO is adsorbed on Pt and reacts at the Pt-gas-support
tpb. However, Korhonen et al. [19] and Graf et al. [18] have recently
demonstrated that Pt is not necessary for the formation of the for-
[
q
¼ r=ro
ð10Þ
the open-cir-
where r is the electropromoted catalytic rate and r
cuit, i.e., unpromoted catalytic rate, the effective rate enhancement
ratio, , defined from [26]:
o
mate intermediate, but for its afterward decomposition to CO
, which is on the other hand, the rate determining step of the
associative mechanism.
2
and
H
2
q
c
The chemical promotion of the WGS catalysts by alkalis has
been investigated in several studies [20–24]. A parallel approach
to the classical chemical promotion is the use of electrochemical
promotion of catalysis (EPOC or non-faradaic electrochemical
modification of catalytic activity, NEMCA effect) to electrochemi-
cally promote metal catalyst electrodes deposited over solid elec-
qc
¼
q
=
qmax
ð11Þ
where
q
max expresses the maximum allowable
q
value due to com-
plete or, more generally, equilibrium conversion and the apparent
faradaic efficiency, , defined from [26]:
K
K
¼ ðr ꢀ r Þ=ðI=nFÞ
o
ð12Þ
2 2
trolyte supports, e.g., YSZ, TiO or CeO . Such an electrode can
where n is the charge of the ionic species (n = 2 for YSZ) and F the
Faraday’s constant. A reaction exhibits electrochemical promotion
when |K| > 1, while electrocatalysis is limited to |K| 6 1.
In a previous study on the electrochemical promotion of the
water–gas shift reaction utilizing a Pd catalyst electrode deposited
over a proton-conducting ceramic support (strontia–zirconia–
also be used as an electrochemical sensor. The idea of using a metal
electrode simultaneously as a catalyst to measure potentiometri-
cally the electrochemical potential, or thermodynamic activity, of
oxygen on metal catalysts, and thus to study catalytic mechanisms,
was originally proposed by Wagner [25] and led to the technique of
solid electrolyte potentiometry (SEP).
The phenomenon of electrochemical promotion of catalysis has
been extensively investigated in the last 30 years for more than 70
catalytic reaction systems [26,27]. In EPOC studies, the conductive
catalyst electrode is in contact with an ionic conductor ceramic
support, and the catalyst (e.g., noble metals and oxides) is electro-
chemically promoted by applying a current or overpotential be-
tween the catalyst film and a counter or reference electrode,
respectively. The cell overpotential is defined as:
yttria perovskite of the form SrZr0.95
Y
0.05
O
3-a), it was found that
H pumping upon positive polarization could lead to rate enhance-
ment, i.e., = 2 and = 8 at temperatures from 600 °C to 750 °C
29].
In the present study, the effect of electrochemical promotion of
catalysis was investigated for the first time in the water–gas shift
WGS) reaction over a porous Pt catalyst electrode interfaced with
+
q
K
[
(
O
YSZ at temperatures from 300 °C to 400 °C and PH2 =PCO ratio val-
ues from 2.85 to 31.
g
¼ U ꢀ Uoc
ð8Þ
where U is the applied potential difference and Uoc the open-circuit
potential difference.
2. Experimental
Numerous surface science and electrochemical techniques have
shown that EPOC is due to electrochemically controlled migration
2.1. Sample preparation
2ꢀ
(
reverse spillover or backspillover) of promoting ionic species (O
The solid electrolyte was a closed-end tube of 8 mol% Y
2 3
O -
+
+
00
in the case of YSZ, TiO
2
and CeO
2
, Na or K in the case of b -Al
) and BCN18 (Ba3-
9-a
Ca1.18Nb1.82O ), etc.) between the ionic or mixed ionic–electronic
2
O
3
,
stabilized ZrO (YSZ) of 13 mm outer diameter and 2 mm thick-
2
protons in the case of Nafion, CZI (CaZr0.9In0.1
O
3-
a
ness. A platinum electrode serving as the counter and/or reference
electrode was deposited on the inner side of the tube by applying a
thin coating of Pt organometallic paste (Engelhard-Clal 6926),
followed by calcination in air for 12 h at 800 °C. Pt was selected
for the counter electrode due to its well known good performance
conductor-support and the gas exposed catalyst surface, through
the catalyst-gas-electrolyte tpb. These backspillover species,
+
accompanied by their compensating (screening) charge d in the
metal, create an overall neutral effective double layer, modifying
the catalyst work function and thus affecting its chemisorption
properties and catalytic activity and selectivity [26–28].
2
in O dissociation and electrochemical reduction.
A similar Pt catalyst electrode was deposited on the outer side
of the tube, opposite to the counter electrode serving as the

1
26
S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
working electrode. The resulting metal loading was 0.83 mg Pt/
2
2
cm , while the geometrical area of the electrode was ꢃ3 cm . The
catalytically active surface area, N , as measured by isothermal
chemical titration with O mol Pt [26]. The
adsorption, was ꢃ1
G
2
l
latter implies for a very small metal dispersion (<1%). The surface
morphology of the solid electrolyte support and of the Pt catalyst
film was examined via scanning electron microscopy (SEM).
Fig. 1 presents a SEM micrograph of the Pt working-catalyst elec-
trode surface. The morphology of Pt layer is homogeneous with
significant porosity.
2.2. Catalytic activity measurements
Reactants and products analysis was performed by online gas
chromatography (Varian CP2003, equipped with two TC detectors
and a molecular sieve and Porapaq Q columns, for the CO, CO
and H analysis), and IR online analyzers (EMERSON NGA2000
for the CO and CO analysis). Reactants were certified standards
of CO in He, which could be further diluted in He. H O vapor was
2
2
2
2
introduced using an atmospheric pressure thermostated quartz
saturator at temperatures from 5 °C to 25 °C, which allowed for va-
por pressure variation between 0.8 kPa and 3.1 kPa. The atmo-
spheric pressure fuel cell type (or double chamber) quartz
3
reactor is shown in Fig. 2. It had a volume of ꢃ10 cm and has been
previously described in detail [30]. As shown in Fig. 2, in this reac-
tor type only the working-catalyst electrode is exposed to the reac-
2
tion mixture (CO, H O and He), while the counter electrode is
constantly exposed to ambient air. In this type of reactor, the work-
ing electrode is responsible for the observed catalytic activity,
while it can be also used as an electrochemical solid-state sensor.
Temperature was measured by a K-type thermocouple placed in
the outer side of the quartz tube in the proximity of the working
electrode. A second thermocouple placed in the inner side of the
bottom of the YSZ tube (exposed in air) revealed a negligible
Fig. 2. Schematic of the fuel cell type electrochemical reactor.
(
ꢃ3 °C) temperature difference. The gas flow and mixture compo-
ꢀ1
sition were regulated by four mass flow controllers (Brooks). The
to 5 s residence time and 1500 h HSV (hourly space velocity)
total gas flowrate was 120 cm3 (STP) min , which corresponds
ꢀ1
based on the reactor volume or 1300 h WHSV (weight hourly
ꢀ1
space velocity) based on the catalyst mass (ꢃ2.5 mg). These values
are comparable to the space velocities in commercial reactors un-
O
der similar PH2 =PCO ratio value and operation conditions. Constant
currents or potentials were applied using a VoltaLab (PGP201)
galvanostat–potentiostat.
3
. Results and discussion
Fig. 3 shows the transient effect of constant applied negative
2
and positive overpotential on the CO formation catalytic rate,
the conversion of CO and the current at 350 °C under
P
4
H2
O
=PCO ¼ 31. As shown in the figure, CO conversion was initially
2%, while negative overpotential application (
g
= ꢀ1.5 V) caused a
pronounced increase in the reaction rate, where CO conversion
reached 95%, i.e., 130% rate increase. The above value is near the
thermodynamic equilibrium conversion value (ꢃ97%) at 350 °C,
and thus, the effective rate enhancement ratio,
q
c
, is 0.98. This
value was
achieved upon negative polarization. The apparent faradaic effi-
ciency,
indicates that almost the maximum allowable
q
K
, was ꢀ30, which shows the strong non-faradaicity and
the high energy efficiency of the process.
After current interruption, the catalytic rate slowly returns to its
initial value, indicating the reversibility of the phenomenon. Worth
to note is that the relaxation time needed for rate recovery is much
higher than the corresponding to reach the electropromoted state
upon negative potential application. This is indicative of the high
thickness of the Pt layer and the low operation temperature, which
Fig. 1. SEM micrographs of the Pt catalyst electrode over the YSZ solid electrolyte
support.
dꢀ
inhibits the thermal migration of the O promoting species back

S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
127
Fig. 3. Transient effect of a constant applied negative (
current. T = 350 °C.
g
= ꢀ1.5 V) and positive (
g
= +2.5 V) overpotential on the CO
2
formation catalytic rate, the conversion of CO and the
to the catalyst surface to obtain its initial coverage. On the other
hand, positive potential application, i.e., O2 supply to the catalyst
surface, resulted in a decrease in the CO conversion from 42% to
The observed electrophilic type behavior, i.e., rate increase by
catalyst work function decrease, is in agreement with Zhai et al.
ꢀ
[20], where the addition of K on Pt was found to result in Pt–OH
x
3
5%, while
K
values were lower than unity.
species stabilization, which is suggested to be the active species
for the low-temperature water–gas shift reaction. Similar alkali
promotion has been reported by Pigos et al. [23] where Na addition
was found to facilitate the formate decomposition, i.e., weakening
of the C–H bond strength, which was assumed as the rate limiting.
The observed electrophilic type behavior, i.e., rate increase upon
negative polarization and rate decrease upon positive, indicates
strongly adsorbed electron donor species, i.e., CO on Pt surface un-
der open-circuit conditions. Under these conditions, there is a fi-
nite oxygen ion coverage on the Pt surface due to thermal
migration of ionic species from the oxide support to the catalyst
surface, which defines an open-circuit steady-state catalytic activ-
ity, by modifying the adsorbed species bond strength. Negative po-
tential application, i.e., O2 pumping from the catalyst surface,
causes a decrease in the catalyst work function and thus weaken-
ing of the Pt–CO bond strength in parallel with the stabilization of
the OH species (electron acceptor species) originating from water
dissociation in the vicinity of Pt particles [26,27,31]. Moreover,
negative potential application could also cause an increase in the
surface concentration of oxygen ion vacancies, by non-stoichiome-
Fig. 4 shows the effect of PH2
O
on the open-circuit potential
re-
(OCP) under constant PCOð0:28 kPaÞ at 350 °C. Increase in PH2
O
sulted in a shift of the OCP to less negative values, indicating
adsorption of electron acceptor species [25,26], i.e., hydroxyl
groups (OH ꢀ s) or atomic oxygen (O ꢀ s) species, formed by water
dissociation according to reactions (14) and (15), where s stands
for an active site either on Pt or on YSZ.
ꢀ
H
2
OðgÞ þ s ! OH ꢀ s þ 1=2H
2
ðgÞ
ð14Þ
H
2
OðgÞ þ s ! O ꢀ s þ H ðgÞ
2
ð15Þ
tric oxygen removal at steady state, due to the lack of O
phase, more likely near the tpb. The latter could probably increase
the WGS reaction rate due to the enhanced H O dissociation on
oxygen vacancies (reaction (6)) in agreement with the mechanism
proposed in the literature [5,7,15–18]. The enhanced H O dissoci-
2
in the gas
However, there is no evidence whether water activation occurs on
Pt or YSZ or both, since OCP can be affected by adsorption on either
metal particles or the support. Moreover, under open-circuit
2
2
ation and stabilization of the OH species could also facilitate the
decomposition of formate intermediate species, which according
to the associative mechanism [11–14,23] is rate limiting. Worth
to note is that under negative polarization, the cathodic electro-
chemical semi-reaction taking place at the cathode, i.e., the cata-
lyst-working electrode, is:
H
2
O þ 2eꢀ ! 2H þ O2ꢀ
ð13Þ
On the other hand, positive potential application, i.e., O2 ions sup-
ply to the catalyst surface, causes an increase in Pt work function
and thus strengthening of the Pt–CO bond strength [26,27,31,32],
ꢀ
which poisons the catalytic reaction rate. In this case,
K values were
found to be lower than unity. This can be explained on the basis of
dꢀ
small O lifetime on the CO-exposed Pt surface.
Furthermore, experiments using only CO in the feed revealed a
faradaic electrocatalytic oxidation of CO under positive polariza-
tion. Thus, under reaction conditions, a synergistic effect occurs be-
tween the electrocatalytic oxidation of CO and the strengthening of
the Pt–CO bond strength. The latter seems to be more significant
since a rate decrease is totally observed.
2
O
Fig. 4. Steady-state effect of the H O partial pressure, PH2 , on the open-circuit
potential, Uoc, under constant volumetric gas flowrate and CO partial pressure, PCO
.
T = 350 °C.

1
28
S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
conditions, i.e., no current or potential application, the support is
possible to undergo some surface reduction by CO and form oxygen
vacancies necessary for water activation [17,18]. However, the pos-
sibility for reduction of the oxide support under polarization is
rather limited, since all the applied cell overpotential values were
kept lower than the YSZ reduction potential (ꢃ2.3 V) [26]. Worth
to note is the ꢃ100 mV change in OCP upon introduction of 3 kPa
shift observed upon H O introduction under CO feed (80 mV). This
difference would perhaps indicate that water adsorption and acti-
2
vation occur mainly on surface oxygen vacancies of the YSZ near
the Pt particles. However, the effect of water dissociation on the
Pt–CO bond strength and thus on the OCP has also to be consid-
ered. Another indication for water adsorption mainly on the YSZ
surface (possibly on oxygen ion vacancies originating by O2
backspillover onto the Pt particles) rather than on Pt could also
ꢀ
H
2
O, which indicates strongly bonded CO on Pt.
The effect of the gas mixture on the OCP with respect to air at
be the higher N value that was calculated by isothermal chemical
G
3
50 °C is shown in Fig. 5. The addition of electron acceptor species,
i.e., H O and/or O in the gas phase, caused an increase in OCP,
from ꢀ145 mV to ꢀ120 and ꢀ100 mV in the presence of 3.1%
O and 2% O , respectively. It is worth noting that the OCP value
in the presence of O reflects the different PO2 values between the
titration [26] with H O adsorption (ꢃ20
lmol Pt) than that with O2
2
2
2
(ꢃ1
lmol Pt).
Fig. 6 presents a scheme of a possible electrochemical promo-
tion mechanism for the WGS reaction, based on the proposed in lit-
erature [17,18] ‘‘associative formate with red–ox regeneration’’
H
2
2
2
working (2%) and reference/counter electrode (20%). On the other
hand, introduction of CO resulted in a dramatic decrease in the
OCP to ꢀ1020 mV and to ꢀ940 mV in the presence of the reaction
mechanism for the case of Pt/ZrO . According to this mechanism,
hydroxyl groups on the zirconia surface react with CO, adsorbed
on Pt, to form intermediate formate. The latter decompose to H2
2
mixture (CO and H
2
O). Furthermore, under 2% O
2
/He where the
and CO , leaving oxygen vacancies that are filled by hydroxyl
2
surface concentration of oxygen vacancies near the tpb can be as-
sumed negligible, introduction of H
O caused ꢃ20 mV increase in
the OCP. The latter is significantly smaller than the corresponding
groups (reactions (6) and (7)). As mentioned before, surface oxygen
2
ꢀ
2
ion vacancies can be created due to O thermal migration and
backspillover onto the Pt surface, but also by partial reduction of
the oxide support in the presence of CO. As shown in the scheme,
negative polarization can lead to weakening of the Pt–CO bond
strength [26] but also to an increase in surface concentration of
oxygen vacancies near the tpb at steady state, which facilitates
the [OH] species adsorption. On the other hand, positive polariza-
tion (not shown here) causes an increase in the Pt–CO bond
strength, which poisons the catalytic activity and results in rate
decrease.
The steady-state effect of PCO on the conversion of CO and CO
2
formation rate and turnover frequency (TOF) under open-circuit
and positive and negative polarization conditions at 350 °C under
constant gas flowrate is presented in Fig. 7a. Under open-circuit
2
conditions, a negative effect of PCO on the CO formation rate was
ꢀ
7
ꢀ1
observed, where it decreased from 0.4 ꢄ 10
mol CO
2
s
at
ꢀ7
ꢀ1
0
.1 kPa CO to 0.3 ꢄ 10 mol CO
2
s
at 0.37 kPa CO. This negative
effect indicates the strong adsorption bond strength of CO on Pt.
The latter agrees with the electrophilic type behavior that observed
in Fig. 3 [26,31]. Positive potential application (g = +2.5 V) resulted
in a rate decrease in the examined PCO range with a similar nega-
tive effect of PCO. On the contrary, negative potential application
Fig. 5. The effect of gas composition on the open-circuit potential value, Uoc
T = 350 °C.
.
(g
= ꢀ1.5 V) caused a significant increase in the catalytic rate,
Fig. 6. Schematic of a possible electrochemical promotion mechanism for the WGS reaction.

S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
129
flowrate. Under open-circuit conditions, a positive effect of PH2
the CO
increased from 19% at 0.8 kPa H
tive effect of PH2 on the rate indicates the weak adsorption bond
O
on
2
formation rate was observed, where the conversion of CO
O to 29% at 3.1 kPa CO. The posi-
2
O
strength of electron acceptor species, i.e., hydroxyl groups. This re-
sult is in agreement with the electrophilic type behavior, which is
observed under strongly adsorbed electron donor species and
weakly adsorbed electron acceptor species [26,31]. A similar posi-
tive effect of PH2
potential application (
Fig. 8 shows the steady-state effect of PCO on the rate enhance-
ment ratio, , and the apparent faradaic efficiency, , values under
:1 kPa PH2 : As shown, and
O
was observed both under positive and negative
g
= +2.5 V).
q
K
3
O
q
K
values up to 3 and ꢀ50, respec-
tively, were obtained at low PCO values and under negative poten-
tial application, while the magnitude of EPOC effect decreased by
increase of PCO or decrease of the PH2
smaller values were obtained upon positive polarization, where
the effect was sub-faradaic, i.e., | | < 1 [34,35].
Fig. 9 shows the steady-state effect of the applied potential (U)
and overpotential ( ) on the rate enhancement ratio, , and the
apparent faradaic efficiency, , values at 3.1 kPa and
.28 kPa PCO at 350 °C. As shown, a maximum value was ob-
served at the optimum overpotential value
= ꢀ1.5 V, probably
O
=PCO ratio. On the other hand,
q
K
g
q
K
P
H2O
0
q
g
due to an optimal catalyst work function value and thus an optimal
reactants’ coverage (more likely CO) of the catalyst electrode, in
agreement with Fig. 7a. Similar optimal applied potential values
have been reported in previous EPOC studies under positive polar-
Fig. 7a. Steady-state effect of the CO partial pressure, PCO, on the CO
2
formation rate
O partial
and the conversion of CO, under constant volumetric gas flowrate and H
2
: T ¼ 350 ꢁC.
pressure, PH2
O
ization and were attributed to site blocking effects by backspillover
dꢀ
O
promoting ions on the positively polarized catalyst electrode
where the negative effect of PCO on the reaction rate was more pro-
nounced than that under open-circuit conditions. The observed
maximum in the reaction rate under negative potential application
could be attributed to an optimal CO coverage over the catalyst.
[34,35].
Arrhenius plots of the unpromoted and electropromoted reac-
tion rate of the WGS reaction are shown in Fig. 10. Negative poten-
tial application caused a pronounced increase in the apparent
ꢀ1
ꢀ1
ꢀ
Turnover frequency, TOF, values ꢃ0.03 s
were found under
open-circuit conditions, while negative potential application
activation energy, E , from 6 to 10.5 kcal mol , while positive
a
1
polarization led to a small decrease to 5 kcal mol . These values
are close to those reported for this reaction in literature, which fall
ꢀ1
caused a TOF increase up to 0.11 s . These values, which were cal-
ꢀ6
ꢀ1
culated using N
ature and fall between 0.5 and 1 s [9,21,33] at 350 °C depending
on the PH2 =PCO ratio.
The steady-state effect of PH2
G
= 10 mol Pt, are close to those reported in liter-
between 7.2 and 11 kcal mol [6,16,36,37]. The observed increase
ꢀ1
in E upon negative polarization and rate increase could be attrib-
a
O
uted either to a gradual change in the rate-limiting step under neg-
O
on the conversion of CO and CO
2
ative polarization conditions or to the fact that negative
polarization can be more effective at lower temperatures. The lat-
ter could be rationalized in the basis of weak CO adsorption on Pt
formation rate under open-circuit and positive and negative polar-
ization conditions is shown in Fig. 7b, at 350 °C under constant gas
2 2
H O
Fig. 7b. Steady-state effect of the H O partial pressure, P 2 , on the CO formation rate and the conversion of CO, under constant volumetric gas flowrate and CO partial
pressure, PCO. T = 350 °C.

1
30
S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
Fig. 10. Arrhenius plots under open-circuit (oc) state and negative (ꢀ1.5 V) and
positive (+2.5 V) overpotential application conditions. Inset: effect of overpotential
on the apparent activation energy, E
a
.
Eq. (16) can also be written in the form:
oc
E
a
¼ Ea þ a
H
eD
U
ð17Þ
since the work function change,
alyst potential, U, via [26,27]:
D
U
, is related to the change in cat-
Fig. 8. Steady-state effect of CO partial pressure, PCO, on the apparent faradaic
efficiency, , and the rate enhancement ratio, , under both positive and negative
potential application. T = 350 °C.
D
K
q
D
U
¼ eD
U
ð18Þ
where e is the electron charge.
3.1. Kinetic measurements
The effect of partial pressure of reactants (CO, H
2
O) on the ki-
netic rate of the WGS reaction was investigated at 350 °C.
Fig. 11a shows the dependence of the intrinsic reaction rate on
the partial pressure of CO, where H O partial pressure was kept
2
constant at 3.1 kPa. The total gas flowrate kept high in order to ob-
tain differential conditions, i.e., less than 10% conversion. As
shown, the increase in CO partial pressure led to a slight decrease
in the WGS reaction rate under either open-circuit state or nega-
tive polarization conditions, similar to Fig. 7a. On the other hand,
O
one observes a positive effect of PH2 on the intrinsic rate under
Fig. 9. Steady-state effect of the applied overpotential on the rate enhancement
ratio, and the apparent faradaic efficiency, under
¼ 3:1 kPa and
CO = 0.28 kPa, T = 350 °C.
q,
K,
P
H2 O
P
dꢀ
at high temperatures and thus smaller effect of O pumping under
negative polarization.
A linear dependence of the apparent activation energy on the
catalyst work function,
DU, has been observed and reported also
in previous EPOC studies [26] and conforms to the equation
[
9,21,33]:
oc
E
a
¼ Ea þ a
H
D
U
ð16Þ
oc
where Ea is the apparent activation energy in the unpromoted state
Fig. 11a. Steady-state effect of the CO partial pressure, PCO, on the kinetic CO
2
ꢀ1
(
6 kcal mol ) and a
H
is a constant, which is usually negative. In this
formation rate under constant H
2
O partial pressure, PH2
O
. (‘/2d) cos
x
= 1, F
T
= 100–
600 cm3 min , T = 350 °C.
ꢀ1
case, a
H
was calculated ꢃꢀ1.3.

S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
Table 1
131
both open-circuit state and negative potential application condi-
Effect of negative polarization on the parameters of the effective double-layer (EDL)
isotherm kinetic model.
tions (Fig. 11b). The reaction rate order value with respect to CO
partial pressure PCO, reported in literature of Pt-based catalysts,
varies between ꢀ1 and 1, while positive (near 1) is the order with
Model parameters
Open-circuit conditions
Negative polarization
OCP = ꢀ0.94 V
U = ꢀ2.44 V
respect to P
2
H O
[8,29,36,38–40].
g
= 0
g
= ꢀ1.5 V
= ꢀ46.56
The effective double-layer (EDL) isotherm Eq. (18), as has been
developed by Brosda and Vayenas [41] for adsorption in the pres-
ence of the ‘‘effective double layer’’, has been used to fit the exper-
imental intrinsic rate data shown in Figs. 11a and 11b, using
FORTRAN programming environment. The points shown in the fig-
ures correspond to the experimental data, while the solid curve to
the theoretically predicted rate according to the EDL model, i.e.:
P
= ꢀ17.52
P
2 ꢄ 10ꢀ
7
4.5 ꢄ 10
ꢀ7
k
k
k
R
A
D
3
10
1000
ꢀ0.12
0.001
500
k
k
A
ꢀ0.0008
D
0
0.0001
1.2 ꢄ 10ꢀ7
k
R
h
j
=ð1 ꢀ h
j
Þ ¼ k
j
P
j
expðꢀk
j
PÞ
ð19Þ
Langmuir-type adsorption isotherms that are mathematically similar
to Frumkin isotherms [41] and can account explicitly for the attrac-
tive or repulsive electrostatic interactions between the adsorbates
and the ‘‘effective double layer’’ present at the catalyst–gas interface
where h
j
is the coverage of the adsorbed j, k
j
is the adsorption coef-
ficient which quantifies the chemisorptive bond strength of j at the
potential of zero charge of the double layer, P is the partial pressure
of j in the gas phase, k is the partial charge transfer parameter of the
adsorbed j which is directly related to the dipole moment of the ad-
sorbed j (k > 0 for the electron donor species and k < 0 for the elec-
j
j
[
41]. Here, we make the assumption for competitive adsorption of
H
2
O and CO on the same active sites.
Using Eq. (19), the catalytic rate can be expressed as:
tron acceptor species), and
by Eq. (20a) [41].
P
is the dimensionless potential defined
k
A
k
D
P
A
P
D
exp½ðk
Þ þ k
is the surface reaction rate constant which can in general
A
þ k
D
Þ
P
ꢂ
r ¼ k
R
h
A
h
D
¼ k
R
ð21Þ
ꢀ
ꢁ
2
‘
½1 þ k
A
P
A
expðk
A
P
D
P
D
expðk
D
PÞꢂ
P
¼
D
U
cos
x
=k
b
T
ð20aÞ
2
d
where k
R
be expressed by:
where ‘ is the distance between the centers of the positive and neg-
ative charges in the adsorbed dipole, d is the thickness of the homo-
ꢀ
ꢁ
o
R
aF
RT
k
R
¼ k exp
U
ð22Þ
geneous double layer,
DU
(=U
ꢀ
U
0
) is the work function difference
between that of the actual surface and that of the surface at its po-
tential of zero charge [26,31,41–43], where the field strength in the
0
where k is the rate constant at zero potential conditions (U = 0 V)
R
double layer vanishes and
the adsorbate dipole and the field, which is vertical to the catalyst
electrode, and k the Boltzmann’s constant. In this analysis, we as-
sume adsorption in the direction of the field, thus
U
=
U
0
,
x
is the angle formed between
and
A D
a the transfer coefficient, which is related to k and k .
Table 1 summarizes the values of the parameters of the model
(k , k , k , k , k and k ) in open-circuit state and under negative
R A D A D
R
and positive polarization conditions at 350 °C. No results under po-
sitive potential application are presented here since the effect was
faradaic and no EPOC was observed. As shown, the surface reaction
0
b
o
x
= 0 and also
that ‘ ꢅ 2d. Hence, Eq. (20a) can also be written, using Eq. (18), in
the form:
rate constant, k
zation) in agreement with Eq. (22), since
behavior. Worth noting is the almost three orders of magnitude
difference between k and k , which indicates the strong CO
adsorption bond strength on Pt. Moreover, an increase in k and
an even more pronounced decrease in k were observed by nega-
R
, increases by potential decrease (negative polari-
P
ꢅ e
D
U=k
b
T
ð20bÞ
a
< 0 for electrophilic type
This model, which can be viewed as an extension of Langmuir–
Hinshelwood–Hougen–Watson kinetics, is based on electrochemical
A
D
A
D
tive polarization, indicating the weakening of the electron donor
species, i.e., CO, bond strength on the catalyst surface and the
strengthening of the electron acceptor species, i.e., OH, by negative
potential application. This agrees with the observed electrophilic
type behavior.
4
. Conclusions
The effect of electrochemical promotion of catalysis on the
water–gas shift reaction was investigated over porous Pt catalyst
electrodes interfaced with YSZ, in a fuel cell type electrochemical
reactor at temperatures from 300 °C to 400 °C, under PH2
values from 2.85 to 31. A negative-order dependence of the cata-
lytic reaction rate on PCO and a positive on PH2 was found under
open-circuit and under polarization conditions.
Positive overpotential application ð ¼ þ2:5 VÞ, i.e., O supply
to the catalyst surface, causes a small decrease in the catalytic
reaction rate, while negative potential application (
= ꢀ1.5 V) re-
sults in a pronounced rate increase, up to 200%, with apparent far-
adaic efficiency values up to 110 (under
O
=PCO ratio
O
2
ꢀ
g
g
g
= ꢀ0.5 V). The observed
rate increase upon negative polarization can be mainly attributed
to the weakening of the Pt–CO bond strength but also, to the
increase in surface concentration of oxygen ion vacancies, by
Fig. 11b. Steady-state effect of the H
2
O partial pressure, P 2
H
O
, on the intrinsic CO
ð‘=2dÞ cos ¼ 1,
2
formation rate under constant CO partial pressure,
P
H2
O
.
x
3
ꢀ1
F
T
¼ 100—600 cm min , T ¼ 350 ꢁC.

1
32
S. Souentie et al. / Journal of Catalysis 283 (2011) 124–132
removing non-stoichiometric oxygen from YSZ, near the Pt-gas-
support three-phase boundaries that are necessary for water
dissociation.
[21] P. Panagiotopoulou, D.I. Kondarides, J. Catal. 267 (2009) 57–66.
[
[
[
[
[
22] P. Panagiotopoulou, D.I. Kondarides, Appl. Catal. B 101 (2011) 738–746.
23] J.M. Pigos, C.J. Brooks, G. Jacobs, B.H. Davis, Appl. Catal. A 328 (2007) 14–26.
24] J.M. Pigos, C.J. Brooks, G. Jacobs, B.H. Davis, Appl. Catal. A 319 (2007) 47–57.
25] C. Wagner, Adv. Catal. 21 (1970) 323–381.
26] C.G. Vayenas, S. Bebelis, C. Pliangos, S. Brosda, D. Tsiplakides, Electrochemical
Activation of Catalysis: Promotion, Electrochemical Promotion, and Metal-
Support Interactions, Kluver Academic/Plenum Publishers, New York, 2001.
References
[
[
1] D.S. Newsome, Catal. Rev. Sci. Eng. 21 (1980) 275.
2] K.-O. Hinrichsen, K. Kochloefl, M. Muhler, in: G. Ertl, H. Knozinger, J. Weitkamp
574.
[
[
[
27] C.G. Vayenas, C.G. Koutsodontis, J. Chem. Phys. 128 (2008) 182506.
28] A. Katsaounis, J. Appl. Electrochem. 40 (2010) 885–902.
29] C. Kokkofitis, M. Ouzounidou, A. Skodra, M. Stoukides, Solid State Ionics 178
(
2
Eds.), Handbook of Heterogeneous Catalysis, second ed., VCH, Ludwigshafen,
008. pp. 2905–2920.
[
3] A.R. de la Osa, A. De Lucas, J.L. Valverde, A. Romero, I. Monteagudo, P. Sánchez,
Int. J. Hydrogen Energy 36 (2011) 44–51.
(
2007) 475–480.
[
[
[
30] X. Li, F. Gaillard, P. Vernoux, Top. Catal. 44 (2007) 391–398.
31] C.G. Vayenas, S. Brosda, C. Pliangos, J. Catal. 203 (2001) 329–350.
32] L. Lizarraga, S. Souentie, L. Mazri, A. Billard, P. Vernoux, Electrochem. Commun.
[
[
[
[
[
[
4] M.A. Henderson, Surf. Sci. Rep. 46 (2002) 1–308.
5] K. Takanabe, K.-i. Aika, K. Seshan, L. Lefferts, J. Catal. 227 (2004) 101–108.
6] T. Bunluesin, R.J. Gorte, G.W. Graham, Appl. Catal. B 15 (1998) 107–114.
7] R.J. Gorte, S. Zhao, Catal. Today 104 (2005) 18–24.
8] D.C. Grenoble, M.M. Estadt, D.F. Ollis, J. Catal. 67 (1981) 90–102.
9] P. Panagiotopoulou, D.I. Kondarides, Catal. Today 112 (2006) 49–52.
1
2 (2010) 1310–1313.
33] Q. Fu, W. Deng, H. Saltsburg, M. Flytzani-Stephanopoulos, Appl. Catal. B 56
2005) 57–68.
34] S. Souentie, A. Hammad, S. Brosda, G. Foti, C.G. Vayenas, J. Appl. Electrochem.
8 (2008) 1159–1170.
35] A. Hammad, S. Souentie, E.I. Papaioannou, S. Balomenou, D. Tsiplakides, J.C.
Figueroa, C. Cavalca, C.J. Pereira, Appl. Catal. B 103 (2011) 336–342.
36] C.M. Kalamaras, P. Panagiotopoulou, D.I. Kondarides, A.M. Efstathiou, J. Catal.
[
[
[
[
(
[
[
10] G. Jacobs, S. Ricote, B.H. Davis, Appl. Catal. A 302 (2006) 14–21.
11] G. Jacobs, U.M. Graham, E. Chenu, P.M. Patterson, A. Dozier, B.H. Davis, J. Catal.
3
2
29 (2005) 499–512.
[
12] A. Goguet, S.O. Shekhtman, R. Burch, C. Hardacre, F.C. Meunier, G.S. Yablonsky,
J. Catal. 237 (2006) 102–110.
264 (2009) 117–129.
[
[
13] S.Y. Choung, M. Ferrandon, T. Krause, Catal. Today 99 (2005) 257–262.
14] G. Jacobs, E. Chenu, P.M. Patterson, L. Williams, D. Sparks, G. Thomas, B.H.
Davis, Appl. Catal. A 258 (2004) 203–214.
[
[
[
37] K.G. Azzam, I.V. Babich, K. Seshan, L. Leffers, Appl. Catal. B 80 (2008) 129.
38] A.B. Mhadeshwar, D.G. Vlachos, Catal. Today 105 (2005) 162–172.
39] A.A. Phatak, N. Koryabkina, S. Rai, J.L. Ratts, W. Ruettinger, R.J. Farrauto, G.E.
Blau, W.N. Delgass, F.H. Ribeiro, Catal. Today 123 (2007) 224–234.
[
[
15] X. Wang, R.J. Gorte, Appl. Catal. A 247 (2003) 157–162.
16] S. Hilaire, X. Wang, T. Luo, R.J. Gorte, J. Wagner, Appl. Catal. A 215 (2001) 271–
[
40] R. Radhakrishnan, R.R. Willigan, Z. Dardas, T.H. Vanderspurt, Appl. Catal. B 66
2
78.
(
2006) 23–28.
[
[
[
17] K.G. Azzam, I.V. Babich, K. Seshan, L. Lefferts, J. Catal. 251 (2007) 153–162.
18] P.O. Graf, D.J.M. de Vlieger, B.L. Mojet, L. Lefferts, J. Catal. 262 (2009) 181–187.
19] S.T. Korhonen, M. Calatayud, A.O.I. Krause, J. Phys. Chem. C 112 (2008) 16096–
[
[
[
41] S. Brosda, C.G. Vayenas, J. Catal. 208 (2002) 38–53.
42] D. Tsiplakides, C.G. Vayenas, J. Electrochem. Soc. 148 (2001) E189–E202.
43] C.G. Vayenas, S. Brosda, C. Pliangos, J. Catal. 216 (2003) 487–504.
1
6102.
[
20] Y. Zhai, D. Pierre, R. Si, W. Deng, P. Ferrin, A.U. Nilekar, G. Peng, J.A. Herron, D.C.
Bell, H. Saltsburg, M. Mavrikakis, M. Flytzani-Stephanopoulos, Science 329
(
2010) 1633–1636.