Electrochemical promotion of NO reduction by C3H6 on Rh catalyst-electrode films supported on YSZ and on dispersed Rh/YSZ catalysts

Article abstract of DOI:10.1016/j.jcat.2007.07.034

The selective catalytic reduction of NO by C₃H₆ in the presence of O₂ was investigated on porous polycrystalline Rh catalyst-electrode films deposited on YSZ (Y₂O₃-stabilized ZrO₂) and dispersed Rh/YSZ (0.5 wt%) catalysts. In the former case, it was found that the application of current or potential between the Rh catalyst-electrode and an Au counterelectrode enhanced the rate of NO reduction and CO₂ formation by factors of up to 55 and 200, respectively. These rate increases were strongly non-Faradaic with apparent Faradaic efficiencies, Λ, of up to 7000 for the formation of CO₂, demonstrating a significant effect of electrochemical promotion or non-Faradaic electrochemical modification of catalytic activity. A comparison of the activity between the electropromoted Rh paste catalyst and the dispersed Rh/YSZ catalyst shows that the TOFs on the dispersed Rh/YSZ catalyst fall between the corresponding TOFs on the unpromoted and electropromoted Rh catalyst-electrode film.

Full text of DOI:10.1016/j.jcat.2007.07.034

Journal of Catalysis 251 (2007) 400–409
www.elsevier.com/locate/jcat
Electrochemical promotion of NO reduction by C H on Rh
3
6
catalyst-electrode films supported on YSZ and on dispersed Rh/YSZ
catalysts
a
a
a
b
c
a,∗
I. Constantinou , D. Archonta , S. Brosda , M. Lepage , Y. Sakamoto , C.G. Vayenas
a
Department of Chemical Engineering, Caratheodory 1 St., University of Patras, GR-26504 Patras, Greece
b
Toyota Motor Europe NV/SA, Technical Centre, Hoge Wei 33, B1930 Zaventem, Belgium
c
Toyota Central R&D Laboratory Inc. Catalysis Div., Nagakute, Aichi 480-1192, Japan
Received 13 April 2007; revised 19 June 2007; accepted 30 July 2007
Abstract
The selective catalytic reduction of NO by C H in the presence of O was investigated on porous polycrystalline Rh catalyst-electrode films
3
6
2
deposited on YSZ (Y O -stabilized ZrO ) and dispersed Rh/YSZ (0.5 wt%) catalysts. In the former case, it was found that the application of
2
3
2
current or potential between the Rh catalyst-electrode and an Au counterelectrode enhanced the rate of NO reduction and CO formation by
2
factors of up to 55 and 200, respectively. These rate increases were strongly non-Faradaic with apparent Faradaic efficiencies, Λ, of up to 7000 for
the formation of CO , demonstrating a significant effect of electrochemical promotion or non-Faradaic electrochemical modification of catalytic
2
activity. A comparison of the activity between the electropromoted Rh paste catalyst and the dispersed Rh/YSZ catalyst shows that the TOFs on
the dispersed Rh/YSZ catalyst fall between the corresponding TOFs on the unpromoted and electropromoted Rh catalyst-electrode film.
©
2007 Elsevier Inc. All rights reserved.
Keywords: Electrochemical promotion; Non-Faradaic electrochemical modification of catalytic activity, NO reduction; Rh catalyst-electrode; NO reduction
by C H
3
6
1
. Introduction
metal/gas exposed catalytic interface [9,13,14] and that classi-
cal promotion, electrochemical promotion, and metal–support
interactions with O2 conducting and mixed ionic–electronic
supports are functionally identical and different only opera-
tionally [14–16].
−
The application of small currents or potentials on porous cat-
alyst films deposited on solid electrolytes leads to pronounced
strongly non-Faradaic and reversible changes in catalytic activ-
ity and selectivity. This phenomenon is described as the electro-
chemical promotion of catalysis (EPOC) or electropromotion,
or as non-Faradaic electrochemical modification of catalytic ac-
tivity (NEMCA) [1–15].
Electrochemical promotion allows for the in situ control of
catalyst activity and selectivity by controlling in situ the pro-
moter coverage via potential application [1–17]. Work over the
last 20 years has demonstrated that the phenomenon results
from electrochemically controlled migration of promoting ionic
The main parameters commonly used to describe the magni-
tude of electrochemical promotion are (i) the apparent Faradaic
efficiency, Λ, defined as
ꢀr
catalytic
Λ =
,
(1)
I/2F
where ꢀrcatalytic is the current or potential induced change in
catalytic rate, I is the applied current, and F is the Faraday con-
stant (96,480 C/mol), and (ii) the rate enhancement factor, ρ,
defined as
δ−
δ+
species (O , Na ) from the solid electrolyte support to the
ρ = r/r = (r + ꢀr
)/r ,
0
(2)
0
0
catalytic
*
where r is the electropromoted catalytic rate and r is the un-
Corresponding author. Fax: +30 2610 997269
0
E-mail address: cat@chemeng.upatras.gr (C.G. Vayenas).
promoted (open circuit) catalytic rate.
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2007.07.034

I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
401
A reaction exhibits electrochemical promotion when
unit, the reactor, and the analysis unit. The feed unit uses cer-
tified gases of C H (3000 ppm in He), NO (3000 ppm and
|
Λ| > 0, whereas electrocatalysis is limited to |Λ| ꢀ 0. When
3
6
Λ > 1, the reaction is termed electrophobic; when Λ < −1, it is
electrophilic. In the former case, the rate increases with catalyst
potential, U, whereas in the latter case, the rate increases with
2% diluted in He), O (1, 5, and 20% diluted in He), and ultra-
2
pure He (99.999%) for further dilution of the reaction mixture.
The C H and NO were provided by Messer Griesheim, and
3
6
5
decreasing catalyst potential. To date, Λ values up to 3 × 10
the O and He were provided by Air Liquide. Gas flow rates
2
[
1,13,14] and ρ values up to 1000 [14,18,19] have been reported
were regulated by Brooks mass flow controllers connected to
for several catalytic systems.
An important question regarding the practical usefulness of
electrochemical promotion is whether the turnover frequency
a four-channel Brose control box (model 5878). The total gas
flow rate was 200 cm (STP)/min.
3
The electrochemical promotion experiments were carried
3
(
TOF) of a catalytic reaction on an electropromoted catalyst
out in a “single pellet” reactor with a volume of ∼30 cm
film (e.g., Pt or Rh on yttria-stabilized zirconia [YSZ]) can ex-
ceed the corresponding TOF for the same reaction on the same
metal obtained with a fully dispersed nanocrystalline catalyst
supported on the same support (e.g., YSZ). This is an important
question from a practical standpoint as well as for understand-
ing the extent to which thermal migration of promoting species
[
14] (Fig. 1a). Under the experimental conditions used in this
study, this reactor behaved as a continuously stirred tank reactor
CSTR), as described in detail previously [13,14,31,32]. The re-
(
action temperature was monitored via a thermocouple (type K)
in contact with the sample.
The dispersed catalyst experiments were conducted in the
reactor shown in Fig. 1b. This reactor behaved as a CSTR and
δ−
from the support (e.g., O ) can match the electrochemically
controlled promoting species migration, which is responsible
for the NEMCA effect.
Reduction of pollutant emissions present in the exhaust of
lean-burn and diesel engines has attracted great scientific in-
terest in recent years [20–22]. One possible way to achieve
high efficiency in this process is the selective catalytic reduction
3
had a volume of approximately 10 cm .
Gas analysis was performed using an online gas chromato-
graph (Hewlett Packard 5890 series II), a chemiluminescence
analyzer (Eco Physics CLD 700 EL ht) for NO and NO2,
and an infrared analyzer (Rosemount Binos 100) for CO2 for
continuous analysis of the reactor feed and products. The gas
chromatograph used a thermal conductivity detector (TCD)
equipped with two columns, a Molecular Sieve 5A (for the de-
(
SCR) of nitrogen oxides by hydrocarbons over noble metal cat-
alysts. Rhodium is a quite selective catalyst for NO reduction
but becomes unreactive under highly oxidizing (i.e., lean-burn)
conditions due to surface Rh2O3 formation [23–26].
The main overall stoichiometric reactions occurring on the
Rh catalyst-electrode are
tection of O , N , and CO) and a Porapak Q column (for the
2
2
detection of C3H6, CO2, and N2O). Both TCD signals were
integrated and recorded online by a Hewlett Packard integra-
tor (model 3395). The chemiluminescence analyzer continu-
ously detected the concentrations of NO and NO2, and the
infrared analyzer monitored the CO2 concentration in the ef-
fluent stream. The signals of the two analyzer’s were recorded
on a three-channel pen recorder (Yokogawa Hokushin Electric,
type LR 4120E). Currents or potentials were applied using an
Amel Instruments model 533 potentiostat/galvanostat.
C3H6 + 9NO → 3CO2 + (9/2)N2 + 3H2O,
C3H6 + 18NO → 3CO2 + 9N2O + 3H2O,
and
(3)
(4)
C3H6 + (9/2)O2 → 3CO2 + 3H2O.
(5)
In addition to these reactions and in high excess of oxygen
in the gas phase, the oxidation of NO to NO2 also occurs. The
electrochemical promotion of Pt and Rh catalyst films for the
reduction of NO by hydrocarbons, CO, and H2 in absence of O2
or with low O2 concentrations has been studied in detail [4,10–
2
2
.2. Catalyst preparation and characterization
.2.1. Preparation of electropromoted films
The electropromoted Rh film catalysts were deposited on a
1
2,14].
In the present study, to address the foregoing question re-
disk of YSZ solid electrolyte (YSZ, 8 wt% Y O ) and served as
2
3
the working electrode of the electrochemical cell. The other two
electrodes (reference and counter) were deposited on the other
side of the disk (Fig. 1a). The YSZ disk was 1.9 cm in diameter
and 2 mm thick. The geometric area of the working electrode
garding electrochemical promotion and metal–support interac-
tion, the selective catalytic reduction of NO by propylene in
the presence of a significant amount of oxygen was examined
over Rh paste catalyst-electrodes supported on YSZ and com-
pared with the performance of Rh catalysts finely dispersed on
porous YSZ.
2
was ∼2.1 cm , and those of the reference and counterelectrodes
2
were ∼0.5 and ∼2.1 cm , respectively.
Gold counter and reference electrodes were deposited on the
one side of the solid electrolyte plate via deposition of a thin
paste layer (A1113 gold resinate; METALOR), followed by cal-
2
. Experimental
◦
2
.1. Apparatus
cination first at 400 C for 2 h (at a heating rate of 10 K/min)
◦
and then at 800 C for 90 min (at the same heating rate). Blank
The experimental apparatus and reactors used are presented
schematically in Fig. 1 and have been described in detail previ-
ously [13,27–30]. The apparatus comprises three parts: the feed
experiments showed that the catalytic activity of the Au elec-
trode was negligible compared with that of the Rh catalyst for
the reactions under study.

402
I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
Fig. 1. Schematic of the experimental apparatus and of the two reactors used (a) for the electrochemical promotion of the Rh film catalysts and (b) for the dispersed
catalyst experiments.
The Rh polycrystalline catalyst film (8826 rhodium resinate;
METALOR) was deposited on the other side of the solid elec-
trolyte, exactly opposite to the counterelectrode. This catalyst-
electrode was calcined at 550 C for 3 h with a heating rate
of 3 K/min. The sample was reduced in situ in the reactor
catalyst surface area, NG, expressed in mol via [13,14]
NG = Iτ/2F.
(6)
◦
From Eq. (6), and taking the average of 10 galvanostatic and
potentiostatic transient experiments at constant temperature, the
active catalyst surface area, NG, of the Rh film was estimated to
using 2% H2 in He, with the temperature raised from 150 to
◦
4
50 C at a rate of 5 K/min, then remaining in the same re-
−7
be 4 × 10 mol Rh. Four such galvanostatic and potentiostatic
◦
ducing 2% H2 mixture at 450 C for 20 min. After reduction,
transients are shown in Appendix A. The above NG value was
used to compute turnover frequencies (TOFs). Table 1 summa-
rizes the properties of the Rh film catalyst-electrode.
the sample was cooled to the desired operating temperature
◦
(
∼380 C) at which the reactive NO, O2, and C3H6 mixture
was fed.
The surface area of the Rh catalyst films was estimated using
2.2.2. Preparation of dispersed catalysts
the galvanostatic transient technique [13,14]. This technique in-
volves first measuring the time, τ, required for the rate increase
to reach 63% of its maximum steady-state value during anodic
polarization, and then, assuming a 1:1 surface Rh:O ratio, esti-
mating the reactive oxygen uptake of the Rh film and the active
The dispersed catalysts were prepared using a wet-impreg-
nation method. Commercial YSZ (Tosoh 8Y-SZ) was used as
the carrier, and a solution of Rh(NO ) ·2H O (Alpha Products)
3
3
2
was used as the precursor. The YSZ support was impregnated
with an aqueous solution of the metal precursor salt. Weighed

I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
403
Table 1
Properties of the catalytic samples
Rh film deposited on YSZ disc
Dispersed Rh/YSZ catalyst
Metal dispersion
Specific surface
77%
∼2.1 cm2
∼2.1 cm2
∼0.5 cm2
4 ×
345 m /g
Rh metal
2
Catalyst-electrode
superficial area
Counterelectrode
superficial area
Reference electrode
superficial area
Actual surface Rh
catalyst area
Mean size of
metal particles
Total catalyst
surface area
1.39 nm
2
12 m /gsupport
Total catalyst
40 mg
−
7
10 mol Rh mass per catalytic
testing
expressed as surface
metal mols, NG
amounts of Rh(NO3)3·2H2O for the desired composition were
◦
dissolved in distilled water at 25 C, followed by the addition
of the desired amount of the carrier. After impregnation, the
◦
water was slowly evaporated at 80 C under continuous stir-
◦
ring. The residue was dried at 120 C for 24 h, and the resulting
solid was ground and sieved. The powder catalyst was treated
◦
at 200 C for 30 min under a flow of N , after which the flow
2
was switched to 100% H2 and the temperature was increased
◦
slowly to 300 C, where the sample was held for 2 h. Finally,
the sample was cooled to room temperature under a flow of N2.
The total metal loading was 0.5%, based on the total weight of
the catalyst.
The total surface area of the support was estimated by the
BET method, and the metal dispersion of the catalyst was
obtained with H2 chemisorption (Table 1). The adsorption
isotherms in H2 were determined by the volumetric technique
using a modified Fisons Instruments Sorptomatic 1900 appara-
Fig. 2. Galvanostatic transient NEMCA experiment (application of +70 µA).
Experimental conditions: P
= P
= 0.1 kPa, P
= 0.43 kPa,
NO
C H
3
O
2
6
◦
3
T = 370 C, F = 200 cm (STP)/min; dotted line refers to the oxygen partial
v
pressure and dashed line refers to the working electrode potential, U, during the
transient. (Top) Conversion of propylene and rate of CO formation. (Bottom)
2
Conversion of NO and rate of NO consumption.
tus. The mass of the catalyst sample used for chemisorption was
◦
1
.0 mg. Measurements were performed at 25 C in the pressure
3
. Results and discussion
range of 0–75 Torr. Before each measurement, the catalyst was
◦
(
(
3
i) pretreated by dynamic vacuum at 250 C for 1 h, followed by
◦
3.1. Transient experiments
ii) reduction with 1 bar H2 at 250 C, and (iii) evacuation for
0 min at 250 C, and (iv) finally cooled to room temperature
◦
Fig. 2 shows a typical galvanostatic NEMCA experiment,
demonstrating the transient effect of constant positive current
application on the rates of CO2 formation and NO reduction
and on the catalyst potential UWR. The experiment was car-
under vacuum. Afterwards, the weakly adsorbed hydrogen was
removed from the catalyst by evacuating the adsorption cell at
−
6
room temperature for 10 min to a pressure of 10 Torr [33].
The total hydrogen uptake at monolayer coverage was obtained
by extrapolation to the adsorption isotherm at zero pressure.
The results are given in Table 1.
ried out at fixed feed gas compositions of PO = 0.43 kPa and
2
◦
PC H = PNO = 0.1 kPa at 370 C. Initially, for t < 0, the cir-
3
6
cuit was open and the steady-state rates of C3H6 oxidation to
Before each experiment, an specific amount of fresh catalyst
−7
CO2 and of NO reduction to N2 were equal to 0.02 × 10
(
ca. 40 mg, particle size of 0.18 < d < 0.25 mm) was placed
−7
and 0.02 × 10 mol O/s, respectively. At t = 0, a constant
◦
in the reactor and reduced at 300 C for 30 min. Then the re-
anodic current, (I = 70 µA) was applied between the cata-
actor was purged with He and the temperature was raised to
2−
lyst and the counterelectrode. Oxygen ions, O , were trans-
◦
5
50 C under He flow. The catalyst was maintained at that tem-
ferred from the YSZ support to the Rh catalyst-electrode at
perature for 1 h and then cooled under a flow of He to room
temperature for measuring the metal dispersion (Table 1) or to
the desired temperature for conducting the kinetic experiments.
In the latter case, the feed was switched directly after the He
purge to the reaction mixture, and light-off curves or steady-
state measurements were obtained at given gas compositions in
−9
a rate, I/2F, equal to 0.36 × 10 mol O/s, where F is the
Faraday constant. At the same time, the rate of C3H6 oxida-
tion to CO2 increased, and after 120 min, it approached a new
−
7
steady state (r = 4.4 × 10 mol O/s). The increase in the cat-
−
7
alytic rate ꢀr (=4.38 × 10 mol O/s) was 1200 times greater
than that in the rate of ion transport, I/2F. This implies that
◦
2−
the temperature range of 100–550 C.
each O supplied to the rhodium catalyst surface caused on av-

404
I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
erage 1200 chemisorbed O atoms to react with propylene and
form CO2 and H2O. A 220-fold rate enhancement (ρ = 220)
was observed, one of the highest ρ values ever observed in
electrochemical promotion studies. The C3H6 conversion in-
creased from 0.15 to 33%. At the same time, the rate of NO
−
7
−7
reduction increased from 0.02 × 10 to 0.74 × 10 mol O/s
−7
(
ꢀr = 0.72 × 10 mol O/s). The latter corresponds to a 35-
fold rate enhancement and a Λ value of 199. After current inter-
ruption, the catalytic rates of CO2 formation and NO reduction
returned to their initial open circuit value, demonstrating the re-
versibility of the phenomenon.
It also is interesting to examine the corresponding catalyst
potential with respect to the reference electrode, UWR, response
in Fig. 2 (dashed line). When the current was first applied to
the sample, the catalyst potential increased abruptly from 50 to
4
50 mV, indicating a change in the oxidation state of the cata-
(a)
lyst. This behavior is quite common when the catalyst surface
is initially oxidized (i.e., has a very high oxygen coverage, very
likely forming an ordered monolayer surface Rh2O3 phase) [34]
and the oxygen partial pressure in the feed is near the stoi-
chiometrically required pressure for complete fuel combustion.
Along with the potential UWR, Fig. 2 also shows the actual oxy-
gen partial pressure obtained from the oxygen mass balance
(
dotted line). It can be seen that after the application of current,
the PO dropped monotonously due to the increased propylene
2
oxidation rate, and the potential exhibited a more complex tran-
sient, indicative of the reduction of the surface Rh2O3 phase.
Note that the open-circuit potential was 50 mV before cur-
rent application but dropped to −180 mV immediately after
current interruption, indicating that the catalyst surface was in-
deed reduced to metallic Rh. It can be seen that under this new
open-circuit potential value, the rates of CO2 formation and
NO consumption were very high. On current interruption, the
gaseous oxygen oxidized the catalyst surface, as manifested by
the gradual increase in catalyst potential back to its initial value
(
b)
Fig. 3. Effect of catalyst potential on the rate and TOF of CO produc-
tion and on C H conversion (a) and on the rates and TOFs of NO conver-
2
3
6
(
Fig. 2). This is in agreement with former studies of electropro-
sion to N2 (b) at various PO2 feed values. Experimental conditions: PNO =
◦
3
moted Rh catalyst films [35,36] demonstrating that application
of currents or potentials can have a significant (and reversible)
effect on the stability limits of surface Rh oxide, thus signifi-
cantly affecting the catalytic activity [14,37,38].
PC H = 0.1 kPa, T = 370 C, Fv = 200 cm (STP)/min. (2) PO =
3
6
2
0
.5 kPa, (1) P = 0.67 kPa, (") P = 0.88 kPa, (!) P = 1.26 kPa,
O2
O2
O2
(F) P = 1.77 kPa, (E) P = 2.5 kPa, (ꢀ, £) P = 4.8 kPa for the rate
O
O
O
2
2
2
of N and NO formation, respectively.
2
2
polarization. This most likely is due to the weakening of the
3
.2. Effect of oxygen partial pressure
Rh=O bond via repulsive lateral interactions with co-adsorbed
δ−
spillover O supplied electrochemically to the catalyst surface
Fig. 3 shows the steady-state effect of oxygen inlet par-
and the concomitant hindrance of the O poisoning of the cat-
alyst surface [14]. At low oxygen inlet partial pressures, the
tial pressure and of catalyst potential on the rate and TOF of
CO2 formation, the conversion of C3H6 (Fig. 3a), and the rates
and TOFs of N2 and NO2 formation and conversion of NO
C H conversion increased from 0.2 to 40%. Under lean burn
3
6
conditions (4.8% O2), the effect was smaller, and rate increases
on the order of 25% were observed. Under cathodic polarization
(i.e., application of negative potentials), the rate increase most
likely was due to the lower work function of the catalyst sur-
face [14], which enhances NO adsorption and dissociation [34].
Fig. 3b presents the corresponding dependence of the NO
consumption rate on the oxygen inlet partial pressure. As shown
in the figure, the pronounced inverted-volcano behavior for NO
(
Fig. 3b). The experiments were carried out at constant inlet
partial pressures of NO and C3H6 equal to 0.1 kPa (1000 ppm)
◦
3
at 370 C and at a constant flow rate of 200 cm (STP)/min.
The catalyst potential was varied from −1500 mV (cathodic
polarization) to +1500 mV (anodic polarization). It can be
seen that both reactions exhibited a pronounced inverted vol-
cano behavior [14]; that is, the rates of C3H6 oxidation and
NO reduction increased dramatically with both positive and
negative potential. At higher oxygen inlet partial pressures,
the electrochemical promotion was more pronounced at anodic
reduction to N2 was present for PO values up to 2.5 kPa. In this
2
range, the selectivity to N2 was near 100%. When the oxygen

I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
405
inlet partial pressure increased to 4.8 kPa, the effect diminished
and; only under these conditions did significant NO2 formation
occur. During anodic and cathodic polarization in this high PO2
range, the selectivity for N2 formation increased slightly, but the
rate of N2 formation remained an order of magnitude smaller
than the rate of NO2 formation.
At low oxygen inlet partial pressures (and consequently low
coverages), anodic polarization reduced the binding strength of
oxygen on the catalyst surface, leading to a more active ad-
sorbed oxygen species and enhancing the rate of propylene
oxidation. The resulting decreased oxygen coverage liberated
surface sites for NO chemisorption and reduction. The ad-
sorbed hydrocarbon species generally reduced the coverage of
adsorbed oxygen, also leading to higher NO reduction rates. At
the highest inlet oxygen partial pressures, the oxygen coverage
was high, the Rh active sites were occupied almost completely
by oxygen, and surface Rh2O3 was formed [36]. This is the
reason why at such high inlet oxygen pressures the conversions
were low and the electrochemical promotion effect was lower
as well, because anodic polarization of the Rh film was not suf-
ficient to decompose the surface Rh2O3 [39].
(a)
3
.3. Effect of hydrocarbon partial pressure
The effect of hydrocarbon partial pressure, PC3H6 , on the
rates of formation of CO2 and N2 is shown in Figs. 4a and 4b.
The inlet partial pressure of NO was kept constant at 0.1 kPa
and that of oxygen was kept constant at 1.25 kPa. The experi-
◦
ments were performed at 370 C and at a constant flow rate of
2
3
00 cm (STP)/min. The C3H6 inlet partial pressure was var-
ied from 150 ppm to 2000 ppm, and the applied potential was
between −1500 and +1500 mV.
In general, increasing the propylene inlet partial pressure
was found to increase significantly both the unpromoted and
electropromoted rates of CO2 formation (Fig. 4a) and the elec-
tropromoted rate of NO reduction to N2 (Fig. 4b). At the highest
propylene partial pressure (PC H = 2000 ppm), the open cir-
(
b)
Fig. 4. Effect of catalyst potential and P
of CO formation (a) and NO reduction (b). Experimental conditions:
on the rates and TOFs
6
C H
3
3
6
2
cuit rate of NO consumption is significantly increased as well.
As expected in view of Fig. 3, both anodic and cathodic po-
larization cause a significant increase in both catalytic rates
◦
3
P
= 0.1 kPa, P = 1.25 kPa, T = 370 C, Fv = 200 cm (STP)/min.
NO
O
2
(P) P
= 150 ppm, (!) P
= 500 ppm, (2) P
= 750 ppm,
C H
3
C H
C H
3 6
6
3
6
(") P
= 1000 ppm, (F) P
= 2000 ppm.
C H
3
C H
3 6
6
(inverted volcano behavior [14]).
It is clear from Figs. 3 and 4 that the PC H /PO ratio of the
3
6
2
values of up to 400, but the electropromotion effect gradually
diminishes as the PC H /PO ratio is decreased below 0.15 due
reaction mixture has a significant effect on the open-circuit cat-
alytic and electropromoted rates. It is known [23] that NO dis-
sociates only at reduced active Rh sites, producing adsorbed N
and O atoms, and that a combination of two adsorbed N atoms
leads to dinitrogen formation. Thus, high hydrocarbon partial
pressures maintain the surface in a reduced state and decrease
the surface coverage of O, accelerating the rate of NO dissoci-
ation and N2 desorption, as has been observed experimentally.
These observations are summarized and quantified in the
three-dimensional Figs. 5–7 which show the effect of catalyst
potential and PC H /PO ratio on the rates of CO2 and N2
3
6
2
to the formation of surface Rh2O3 [14].
The Faradaic efficiencies ΛCO and ΛNO, increased signif-
2
icantly with increasing PC H /PO . At the highest PC H /PO
3
6
2
3
6
2
ratio used, and for small overpotentials (<500 mV), the mea-
sured Faradaic efficiency values, ΛCO2 , were as high as 7000.
These Faradaic efficiencies are by an order of magnitude higher
than those obtained at the lowest PC H /PO ratio. Similar be-
3
6
2
havior was observed for NO reduction. The Faradaic efficiency,
ΛNO, for small overpotentials (<500 mV) varied from 50 at the
lowest PC H /PO ratio to 550 at the highest PC H /PO ratio.
3
6
2
formation (Fig. 5), the rate enhancement ratios ρCO and ρN2
2
3
6
2
3
6
2
(
Fig. 6), and the corresponding Faradaic efficiency values ΛCO2
The measured Faradaic efficiency and rate enhancement ra-
tio values were of comparable magnitude with those obtained
recently on Ir [40] and Pt [41] catalysts for the SCR of NO.
and ΛN (Fig. 7). As already noted, ρN reaches values on the
order of 60 for high PC H /PO ratios with corresponding ΛN
2
2
2
3
6
2

406
I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
(a)
(a)
(b)
(b)
Fig. 6. Effect of catalyst potential and P
/P feed ratio on the rate en-
C H
3
O
6
2
hancement ratio of CO (a) and N (b) formation rate as a function of propylene
Fig. 5. Effect of catalyst potential and P
/P
feed ratio on the rates
2
2
C H
O
3
6
2
to oxygen inlet partial pressures. Experimental conditions: P
= 0.1 kPa,
of (a) CO formation and (b) N formation as a function of catalyst poten-
NO
2
2
◦
3
T = 370 C, Fv = 200 cm (STP)/min. Open symbols refer to constant oxy-
gen inlet pressure of 1.25 kPa and filled symbols refer to constant propylene
inlet partial pressure of 0.1 kPa.
tial for different propylene inlet partial pressures. Experimental conditions:
◦
3
P
= 0.1 kPa, PO = 1.25 kPa, T = 370 C, Fv = 200 cm (STP)/min.
NO
2
Open symbols refer to constant oxygen inlet pressure of 1.25 kPa and filled
symbols refer to constant propylene inlet partial pressure of 0.1 kPa.
This implies that the dispersed catalysts are, to some extent,
partially promoted due to thermal migration of the promoting
3
.4. Comparison of electropromoted and dispersed catalysts
2−
O
species from YSZ to the dispersed catalyst surface [14]. It
As discussed earlier, an important question regarding the
should be noted that the TOFs compared in Figs. 8 and 9 were
based on the number of surface Rh atoms, which did not change
when surface Rh2O3 oxide phases were formed, because the lat-
ter most likely did not exceed a monolayer [14].
Fig. 8 presents the TOFs of the CO2 production rate as a
function of the oxygen inlet partial pressure. Under open cir-
cuit conditions, the TOF of CO2 formation of the film catalyst
was quite low (open circles). The rate of CO2 formation was
positive order with respect to the partial pressure of O2 under
open circuit conditions and became near-zero order under elec-
tropromoted conditions. Similarly, the rate was near-zero order
practical usefulness of the electrochemical promotion effect for
this and other catalytic systems is to what extent the TOFs
on the electropromoted electrode surface can exceed the corre-
sponding TOFs of fully dispersed commercial catalysts where
2−
the O spillover mechanism is also operative [14]. Figs. 8
TOFs for CO2 formation) and 9 (TOFs for NO formation)
(
present direct comparisons between electropromoted and dis-
persed Rh on YSZ catalysts. It can be seen that, in general, the
TOF values of the dispersed catalysts fall between the TOFs
on the promoted and unpromoted Rh paste catalyst-electrodes.

I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
407
(a)
Fig. 8. Effect of P on the TOF of CO formation on the unpromoted (!) and
O
2
2
electropromoted (") Rh film and on the dispersed Rh/YSZ catalyst (×).
(b)
Fig. 7. Effect of catalyst potential and P
/P feed ratio on the Faradaic
C H
3
O
6
2
Fig. 9. Effect of P
moted (!) and electropromoted (") Rh film and on the dispersed Rh/YSZ
on the TOF of NO conversion to N on the unpro-
2
efficiency of CO formation (a) and N formation (b). Conditions as in Fig. 6.
O2
2
2
catalyst (×). Encircled points correspond to significant formation of NO2.
with respect to the oxygen partial pressure for the dispersed
catalyst. The TOFs for CO2 formation of the dispersed catalyst
were between the corresponding TOFs for CO2 of the unpro-
moted (open circuit) and the electropromoted (closed circuit)
catalyst-electrode film.
The TOFs for NO consumption depicted in Fig. 9 show again
that the TOFNO of the dispersed Rh/YSZ catalyst was between
the corresponding TOFs of the unpromoted and electropro-
moted Rh paste catalyst-electrodes. The rate of NO consump-
The results shown in Figs. 8 and 9 underscore the func-
tional similarities and only operational differences between the
effect of electrochemical promotion and metal–support inter-
actions with ionic and mixed ionic–electronic conducting sup-
ports [14–16].
4
. Conclusion
tion was near-zero order with respect to the PO for the (highly
2
inactive) unpromoted catalyst and was strongly negative order
in PO2 for both the highly active electropromoted Rh paste
catalyst-electrode and the dispersed Rh/YSZ catalyst. This is
due to the blocking of surface Rh sites by adsorbed O and con-
The present study has demonstrated that electrochemical
promotion can significantly affect the SCR of NO by propy-
lene, even in the presence of a significant excess of oxygen. The
rates of CO2 and N2 production are enhanced both with positive
(electrophobic behavior) and with negative (electrophilic be-
havior) potentials or currents. The rate enhancement ratios, ρ,
comitant formation of surface Rh2O3 for values PO > 2.5 kPa.
2
At PO > 4 kPa, significant NO2 formation can be seen for both
2
the electropromoted and the dispersed catalysts.

408
I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
(a)
(b)
Fig. 10. Galvanostatic (a) and potentiostatic (b) transients used for the measurement of the Rh film surface area, N . Conditions for all transients:
G
◦
3
−7
mol O/s,
P
= P
= 0.1 kPa, P = 0.4 kPa, T = 370 C, Fv = 200 cm (STP)/min; galvanostatic transient for 40 µA: r = 0.02 × 10
NO
C H
3
O
2
0
6
−
7
−7
−7
ꢀ
r = 0.53×10 mol O/s, ρ = 27, Λ = 256; galvanostatic transient for 50 µA: r = 0.02×10 mol O/s, ꢀr = 1.04×10 mol O/s, ρ = 53, Λ = 400; potentio-
0
−
7
−7
−7
static transient for 500 mV: r = 0.02×10 mol O/s, ꢀr = 0.3×10 mol O/s, ρ = 11, Λ = 58; potentiostatic transient for 500 mV: r = 0.05×10 mol O/s,
0
0
−
7
ꢀ
r = 1.12 × 10 mol O/s, ρ = 23, Λ = 76.
for the two main products, CO2 and N2, can take very high val-
ues. Application of positive current (or potential) leads to ρ val-
ues up to 200 and 55 for CO2 and N2 production, with apparent
Faradaic efficiencies on the order of 7000 and 200, respectively.
Comparing the two phenomena (metal–support interaction
Rh mols, NG, via Eq. (6), that is, NG = Iτ/2F. In the case
of the potentiostatic transients, in Eq. (6) we used the time-
averaged current value during the rate transient, which differs
little (<10%) from the near-steady-state current value obtained
after the initial sharp current transients.
2−
with O -conducting supports and electrochemical promotion)
may lead to the conclusion that they are functionally identical
and only operationally different and that the thermal migra-
As noted earlier, more complex rate versus time transients
can be obtained under some conditions (e.g., Fig. 2), appar-
ently due to the abrupt decomposition of the surface monolayer
Rh O phase [14]. Such transients were not used in computing
2−
tion (back-spillover) of promoting O ions from the support
to the metal–gas interface is not sufficient to fully promote a
dispersed catalyst (i.e., to establish sufficiently high coverage
2
3
−
7
the average NG value (∼4 × 10 mol Rh), but appear to con-
sist of two successive transients each giving a N value close
G
2−
of O on the dispersed catalyst surface), so that the TOFs on
the dispersed catalyst do not reach the corresponding TOFs on
the fully electropromoted catalyst. This is an important result
that shows promise for the practical usefulness of electrochem-
ical promotion.
to the above average value.
References
[
1] C.G. Vayenas, S. Bebelis, S. Ladas, Nature 343 (1990) 625.
[2] R.M. Lambert, F. Williams, A. Palermo, M.S. Tikhov, Top. Catal. 13
2000) 91.
Similar strong promotional effects are obtained when the
support (e.g., Y zeolite) can donate monovalent or divalent
cations to the supported metal catalyst [42,43]. This is analo-
gous to the electropromotion of metal films deposited on mono-
valent or divalent ion-conducting solid electrolytes [3,14,15].
(
[
[
[
3] G. Foti, S. Wodiunig, C. Comninellis, Curr. Top. Electrochem. 7 (2000) 1.
4] C.A. Cavalca, G.L. Haller, J. Catal. 177 (1998) 389.
5] L. Ploense, M. Salazar, B. Gurau, E.S. Smotkin, J. Am. Chem. Soc. 119
(1997) 11550.
[6] P. Vernoux, F. Gaillard, L. Bultel, E. Siebert, M. Primet, J. Catal. 208
Acknowledgments
(2002) 412.
7] I. Metcalfe, J. Catal. 199 (2001) 247.
[
[
8] C. Sanchez, E. Leiva, in: W. Vielstich, H. Gasteiger, A. Lamm (Eds.),
Handbook of Fuel Cells: Fundamentals, Technology and Applications,
Wiley, England, 2003.
The authors thank Toyota for financing this research and
Professor Verykios and his laboratory for providing the hydro-
gen chemisorption measurements on the Rh-dispersed catalyst.
D. Archonta and I. Constantinou received graduate student fel-
lowships from the PENED 2001 program of the General Secre-
tariat of Research and Technology (GSRT) Hellas.
[9] G. Fóti, O. Lavanchy, Ch. Comninellis, J. Appl. Electrochem. 30 (12)
(2000) 1223.
[10] C. Pliangos, C. Raptis, T. Badas, C.G. Vayenas, Ionics 6 (2000) 119–126.
[11] C. Pliangos, C. Raptis, T. Badas, C.G. Vayenas, Solid State Ionics 136/137
(2000) 767–773.
[12] C. Pliangos, C. Raptis, T. Badas, D. Tsiplakides, C.G. Vayenas, Elec-
trochim. Acta 46 (2000) 331–339.
Appendix A
[
13] C.G. Vayenas, M.M. Jaksic, S. Bebelis, S.G. Neophytides, in: J.O.M.
Bockris, B.E. Conway, R.E. White (Eds.), Modern Aspects of Electro-
chemistry, Kluwer Academic/Plenum, New York, 1996, p. 57.
Fig. 10 presents two galvanostatic (a) and two potentio-
static (b) transients used to estimate the number of surface

I. Constantinou et al. / Journal of Catalysis 251 (2007) 400–409
409
[
14] C.G. Vayenas, S. Bebelis, C. Pliangos, S. Brosda, D. Tsiplakides, Electro-
chemical Activation of Catalysis: Promotion, Electrochemical Promotion
and Metal–Support Interactions, Kluwer Academic/Plenum, New York,
[29] C. Cavalca, G. Larsen, C.G. Vayenas, G. Haller, J. Phys. Chem. 97 (1993)
6115.
[30] I.V. Yentekakis, S. Bebelis, J. Catal. 137 (1992) 278.
[31] I. Riess, C.G. Vayenas, Solid State Ionics 159 (3–4) (2003) 313.
[32] D. Tsiplakides, C.G. Vayenas, J. Catal. 185 (1999) 237.
[33] A. Kotsifa, T.I. Halkides, D.I. Kondarides, X.E. Verykios, Catal.
Lett. 79 (1–4) (2002) 113.
2001.
[
15] C.G. Vayenas, C. Pliangos, S. Brosda, D. Tsiplakides, in: A. Wieck-
owski, E. Savinova, C.G. Vayenas (Eds.), Catalysis and Electrocatalysis
at Nanoparticles, Dekker, New York, 2003.
[16] J. Nicole, D. Tsiplakides, C. Pliangos, X.E. Verykios, Ch. Comninellis,
C.G. Vayenas, J. Catal. 204 (2001) 23.
[34] C. Pliangos, C. Raptis, T. Badas, C.G. Vayenas, Solid State Ionics 136/137
(2000) 767.
[
17] X.E. Verykios, in: A. Wieckowski, E.R. Savinova, C.G. Vayenas (Eds.),
Catalysis and Electrocatalysis at Nanoparticles Surfaces, Marcel Dekker,
New York, 2003.
[35] C. Raptis, Ph.D. thesis, Department of Chemical Engineering, University
of Patras, 2003.
[36] C. Pliangos, Ph.D. thesis, Department of Chemical Engineering, Univer-
sity of Patras, 1995.
[37] C.G. Vayenas, S. Brosda, C. Pliangos, J. Catal. 203 (2001) 329.
[38] C. Pliangos, I.V. Yentekakis, X.E. Verykios, C.G. Vayenas, J. Catal. 154
(1995) 124.
[39] I. Constantinou, Ph.D. thesis, Chemical Engineering Department, Univer-
sity of Patras, 2005.
[
[
18] S. Bebelis, N. Kotsionopoulos, Solid State Ionics 177 (2006) 2205.
19] C. Kokkofitis, G. Karagiannakis, S. Zisekas, M. Stoukides, J. Catal.
234 (2) (2005) 476.
[
[
[
[
20] M.D. Amiridis, T. Zhang, R.J. Farrauto, Appl. Catal. B 10 (1996) 203.
21] A. Fritz, V. Pitchon, Appl. Catal. B 13 (1997) 1.
22] V.I. Parvulescu, P. Grange, B. Delmon, Catal. Today 46 (1998) 233.
23] T.I. Halkides, D.I. Kondarides, X.E. Verykios, Catal. Today 73 (2002)
[40] P. Vernoux, F. Gaillard, R. Karoum, A. Billard, Appl. Catal. B 73 (2007)
73.
213–221.
[
[
24] T. Chafik, D.I. Kondarides, X.E. Verykios, J. Catal. 190 (2000) 446.
25] T. Chafik, A.M. Efstathiou, X.E. Verykios, J. Phys. Chem. B 101 (1997)
[41] F. Dorado, A. de Lucas-Consuegra, P. Vernoux, J.L. Valverde, Appl. Catal.
B 73 (2007) 85.
7
968.
[42] T. Visser, T. Alexander Nijhuis, AdM.J. van der Eerden, K. Jenken, Y. Ji,
W. Bras, S. Nikitenko, Y. Ikeda, M. Lepage, B.M. Weckhuysen, J. Phys.
Chem. B 109 (2005) 3822.
[
[
26] D.I. Kondarides, T. Chafik, X.E. Verykios, J. Catal. 191 (2000) 147.
27] C.G. Vayenas, S. Bebelis, I.V. Yentekakis, H.-G. Lintz, Catal. Today 11 (3)
(
1992) 303.
28] C.G. Vayenas, I.V. Yentekakis, in: G. Ertl, H. Knötzinger, J. Weitcamp
Eds.), Handbook of Catalysis, VCH, Weinheim, 1997, p. 1310.
[43] AdM.J. van der Eerden, T. Visser, T. Alexander Nijhuis, Y. Ikeda, M.
Lepage, D.C. Koningsberger, B.M. Weckhuysen, J. Am. Chem. Soc. 127
(2005) 3272.
[
(