Non-faradaic electrochemical modification of catalytic activity: 11. Ethane oxidation on Pt

Article abstract of DOI:10.1006/jcat.1997.1775

The rate of ethane oxidation on polycrystalline Pt catalyst films deposited on 8 mol% Y₂O₃-stabilized ZrO₂, an O^2- conductor, can be reversibly enhanced by up to a factor of 20 by varying the potential of the Pt film. The rate increase is typically a factor of 200 higher than the electrochemically controlled rate of supply of promoting oxide ions. Electrochemical oxygen removal from the Pt catalyst surface also enhances the catalytic rate by up to a factor of 7. In this case the rate increase is typically a factor of 20 higher than the electrochemically controlled rate of oxygen removal. The ethane oxidation activation energy varies linearly by a factor of 3 with varying catalyst potential, leading to the appearance of the compensation effect. The observed pronounced promotional phenomena are discussed on the basis of the reaction mechanism, previous electrochemical promotion studies and recent TPD and XPS investigations.

Full text of DOI:10.1006/jcat.1997.1775

JOURNAL OF CATALYSIS 171, 148–159 (1997)
ARTICLE NO. CA971775
Non-Faradaic Electrochemical Modification of Catalytic Activity
11. Ethane Oxidation on Pt
Anthony Kaloyannis and Constantinos G. Vayenas
Department of Chemical Engineering, University of Patras, GR-26500 Patras, Greece
Received January 31, 1997; revised May 19, 1997; accepted June 2, 1997
electrolyte the induced increase in catalytic rate is up to
3
105 times larger than the rate, I/2F, of O2 supply to
The rate of ethane oxidation on polycrystalline Pt catalyst films
deposited on 8 mol% Y2O3-stabilized ZrO2, an O2 conductor, can
be reversibly enhanced by up to a factor of 20 by varying the po-
tential of the Pt film. The rate increase is typically a factor of 200
higher than the electrochemically controlled rate of supply of pro-
moting oxide ions. Electrochemical oxygen removal from the Pt
catalyst surface also enhances the catalytic rate by up to a factor
of 7. In this case the rate increase is typically a factor of 20 higher
than the electrochemically controlled rate of oxygen removal. The
ethane oxidation activation energy varies linearly by a factor of 3
with varying catalyst potential, leading to the appearance of the
compensation effect. The observed pronounced promotional phe-
nomena are discussed on the basis of the reaction mechanism, pre-
vious electrochemical promotion studies and recent TPD and XPS
c
the catalyst (2, 15–17) (where I is the applied current and
F is Faraday’s constant) and up to 100 times larger than the
catalytic rate when no current is applied (18).
Three parameters are commonly used for the macro-
scopic description of electrochemical promotion:
1. The enhancement factor or Faradaic efficiency, 3, de-
fined by
3
1r/(I/2F),
[1]
where 1r isthe induced change in catalytic rate expressed in
mol O/s. The current, I, is defined positive when anions are
supplied to the catalyst. A reaction exhibits the NEMCA
effect when |3| > 1. When 3 > 1 the reaction is termed elec-
trophobic, while when 3 < 1 the reaction is termed elec-
trophilic. As shown both experimentally and theoretically
(15–17) the order of magnitude of the absolute value, |3|, of
the Faradaic efficiency 3 can be estimated for any reaction,
catalyst, or solid electrolyte from
investigations.
1997 Academic Press
INTRODUCTION
During the past 8 years the effect of non-Faradaic electro-
chemical modification of catalytic activity (NEMCA) (1–3),
or electrochemical promotion (4) or in situ controlled pro-
motion (5), has been described for over 40 catalytic reac-
tions on Pt, Pd, Rh, Au, Ag, Ni, and IrO2 catalysts using
O2 -conducting solid electrolytes, such as yttria-stabilized
zirconia (YSZ) (1–3); Na+-conducting solid electrolytes,
such as 00-Al2O3 (5, 6); H+ conducting solid electrolytes,
such as CsHSO4 (7), CaZr0.9In0.1O3 a (8) and Nafion (9); F
conducting solid electrolytes, such as CaF2 (10); and mixed
ionic–electronic conductors, such as TiO2 (11) as the active
catalyst support. The effect has also been recently studied
in aqueous electrolyte systems (12–14). Work in this area
has been reviewed extensively (15–17).
|3| 2Fro/Io,
[2]
where ro is the open-circuit (unpromoted) catalytic rate and
Io is the exchange current of the electrocatalytic (charge-
transfer) reaction at the catalyst–solid electrolyte interface.
Equation [2] dictates that in order to obtain a non-Faradaic
rate enhancement (|3| > 1) it is necessary to have a cata-
lytic reaction intrinscically faster than the corresponding
electrocatalytic reaction, i.e., ro > (Io/2F).
2. The rate enhancement ratio, , defined by
= r/ro.
[3]
The porous metal or metallic oxide (18) catalyst film
serves simultaneously as an electrode in the galvanic cell
3. The promotion index PIi of the promoting ion, i, de-
fined by
gaseousreactants, metal catalyst|solid electrolyte|metal,O2
1r/ro
PIi =
,
[4]
,
1
i
and the NEMCA effect is induced by applying currents or
potentials (typically 2 to +2 V) between the catalyst and where i is the coverage of the promoting species (e.g., O
the metal counter electrode. When using YSZ as the solid Na +) on the catalyst surface. When the promoting species
148
0021-9517/97 $25.00
c
Copyright
1997 by Academic Press
All rights of reproduction in any form reserved.

ETHANE OXIDATION ON Pt
149
does not react appreciably with any of the reactants (e.g., ied in detail by McMaster and Madix (37), who found that
Na +) then its coverage, i, can be measured accurately on the initial step for ethane dissociation involves C H bond
–
the catalyst surface via coulometry (5, 6), and thus PIi can cleavage. Molecular ethane adsorption is weak (37, 38) and
be easily computed (5, 6); PINa values up to 3000 have been thusdissociative ethane chemisorption, which isa structure-
measured for the NO reduction by H2 on Pt (20). When the sensitive process (37, 38), is likely to be the first step for
promoting species (e.g., O ) is also partially consumed by ethane oxidation on Pt (37).
one of the reactants, as in the present case (where C2H6 re-
In this work we study the effect of electrochemical
actswith the promotingspeciesO at a rate 3timessmaller promotion for ethane oxidation on polycrystalline Pt
(19) than with normally chemisorbed oxygen), then the deposited on YSZ. Similarly with the cases of C2H4 (2) and
measurement of PIi is less accurate. In such cases a conser- C3H6 (39) oxidation on Pt/YSZ, ethane oxidation is found
vative estimate of PIi can be obtained from PIi = (1r/ro)max
,
to exhibit a pronounced electrophobic NEMCA enhance-
i.e., by assuming that the maximum rate enhancement is ment with positive currents, i.e., O supply to the catalyst.
obtained for 1
for measuring
work.
= 1 (16, 21). A more precise method Interestingly it is found that the reaction also exhibits a
, and thus PIO , is given in the present strong electrophilic enhancement with negative currents,
i.e., with removal of atomic oxygen from the catalyst
O
O
Recent XPS (21–23), SERS (24, 25), cyclic voltammetric surface.
(26), TPD (26), and STM (27) investigations have shown
that, as originally proposed on the basis of rate transients
(2, 15), the NEMCA effect is due to an electrochemi-
EXPERIMENTAL
cally controlled migration (backspillover) of ionic species
from the solid electrolyte onto the gas-exposed catalyst-
electrode surface. These backspillover ionic species alter
the catalyst surface work function, e8, by
The atmospheric pressure continuous flow apparatus has
been described previously (6, 11, 15, 19, 39). Reactants were
Messer Griesheim certified standards of C2H6 in He and O2
in He. They could be further diluted in ultrapure (99.999% )
He. Reactants and products were analyzed by on-line gas
chromatography (Perkin–Elmer Sigma 300) with a TC de-
tector and Perkin–Elmer LCI-100 computing integrator. A
1(e8) = e1VWR
,
[5]
where VWR is the catalyst (working electrode, W) potential Porapak N column was used to separate O2, C2H6, CO2, and
with respect to the reference (R) electrode (3, 28, 29). Elec- H2O. No other products were detected. The CO2 concentra-
trochemical promotion is thus due to the promoting action tion in the reactor effluent was also continuously monitored
+
of these backspillover species (O in the case of YSZ, Na
using an Anarad Infrared CO2 analyzer.
in the case of 00-Al2O3) which affect the binding strength of
The well-mixed quartz reactor of volume 30 cm3 was of
chemisorbed reactants and intermediates via through-the- the single-pellet type (40, 41); i.e., the YSZ disc (19 mm
metal and through-the-vacuum interactions (15–17). Re- diameter, 2 mm thickness) was suspended in the quartz
cent ab initio quantum mechanical calculations using metal tube (Fig. 1a) with three Au wires attached to the catalyst,
clusters and oxygen atoms coadsorbed with ions or point counter, and reference electrodes.
charges have confirmed this picture and have provided a
The Pt catalyst film (mass 3 mg, superficial surface area
firm theoretical basis for a fundamental understanding of 1.05 cm2) was deposited on one side of the YSZ disc us-
electrochemical promotion and more generally of promo- ing a thin coating of unfluxed Engelhard A1121 Pt paste
tion on metal catalysts (30).
followed by drying and calcination in air, first for 2 h at
The nature of the promoting O species on Pt interfaced 400 C, then for 1 h at 800 C (Fig. 1b). The true surface
with YSZ has been investigated with XPS (21), SERS (24), area of the Pt film was determined by measuring its re-
cyclic voltammetry (26), and TPD (26). The XPS investiga- active oxygen uptake at 380 C via the isothermal titration
tion has shown that its O1s binding energy is at 528.8 eV technique: The Pt film is first exposed to O2 for several
(21). The TPD investigation has shown that its peak des- min and then the reactor is purged with ultrapure He for
orption temperature is at Tp 750–780 K vs Tp 675– a time tHe, several times longer than the reactor residence
685 K for weakly bonded atomic oxygen (O1s binding en- time to remove gaseous O2. Subsequently the reactor is
ergy at 530 eV (21)) which coexists on the Pt surface and purged with CO and the amount of oxygen, N, remaining
is displaced to this weakly bonded state from its normal on the surface is obtained by integrating the area of the CO2
chemisorptive state (Tp 720–730 K) by the coadsorbed peak in the reactor effluent. By varying tHe one can study
strongly bonded O species (26).
the kinetics of oxygen desorption (15–17). By extrapolat-
The complete (31–34) and partial (34, 35) oxidation of ing N to tHe = 0, one obtains the reactive oxygen catalyst
ethane has been investigated on several metal oxide (31, 32, uptake NO. Figure 2 shows the results of the surface titra-
34, 35) and metal (33, 36) surfaces including Pt (36). The tion measurements. The reactive oxygen uptake is NO =
ethane dissociation dynamics on Pt(111) have been stud- 1.1 10 6 mol O. Assuming a maximum oxygen coverage

150
KALOYANNIS AND VAYENAS
FIG. 1. (a) Single pellet catalytic reactor and three-electrode configuration. (b) Scanning electron micrographs of the Pt catalyst electrode.
of 0.25 (27) this corresponds to a true Pt surface area of for 10 min. Au was chosen as the auxiliary electrode ma-
1800 cm2.
terial because of its inertness for C2H6 oxidation (checked
Gold counter and reference electrodes were deposited via blank experiments) which makes it an adequate pseu-
on the other side of the disc using thin coatings of a Au doreference electrode when using the single pellet design,
paste prepared from a slurry of Au powder (Aldrich Chem- as discussed recently (42). Thus to a good approximation
icals) in ethyl acetate followed by calcination in air at 810 C ( 0.1 V) (42) the measured open-circuit potentials, VWo R
,

ETHANE OXIDATION ON Pt
151
FIG. 2. Isothermal titration of surface oxygen with CO. Effect of oxy-
gen desorption time, tHe, on the mass of reactive oxygen on the Pt catalyst
surface.
FIG. 3. Effect of PO and temperature on the rate and turnover fre-
quency of ethane oxidation.
2
ing PC H between 0 and 4 kPa at fixed PO (0.6 kPa) was
2
6
2
but also the imposed potentials VWR can be referenced to
the equilibrium potential of oxygen corresponding to the
oxygen partial pressure in the reactive gas mixture. The
found to decrease VWo R by less than 0.1 V in contrast with
previous studies of CO and C2H4 oxidation on Pt/YSZ (16)
where PCO and PC H valuesup to 2kPa cause a pronounced,
open-circuit potential, VWo R , in presence of PO = 21 kPa
2
4
2
up to 0.4V, decrease in VWo R under similar conditions. There-
fore the VWo R values also suggest weak adsorption of ethane
relative to oxygen.
was very near the theoretical value of zero ( 0.05 V) over
the temperature range of the present investigation (400–
500 C). The origin of such small offset potentials at these
relatively low temperatures has been discussed previously
(15, 16).
Promotional Transients
Figures 5a and 5b show typical galvanostatic transients;
i.e., they depict the catalytic rate and catalyst potential re-
sponse upon application of a positive and of a negative
current, respectively.
RESULTS
Open-Circuit Kinetics
Figures 3 and 4 show the dependence of the rate, r, of
ethane oxidation on the partial pressures, PO and PC H ,
2
2
6
of oxygen and ethane, respectively, under open-circuit, i.e.,
unpromoted, conditions. In these and subsequent figures
the rate of ethane oxidation is expressed in mol O/s and also
as turnover frequency (TOF), i.e., oxygen atoms reacting
per surface site per second, by using the reactive oxygen
uptake, NO = 1.1 10 6 mol O, of the Pt catalyst film.
As shown in Fig. 3 the rate exhibits a pronounced maxi-
mum with respect to oxygen pressure at a PO value, here-
2
after denoted PO , which is weakly dependent on tempera-
ture. Figure 4 sho2ws that the rate is positive order in ethane
over the entire range of gaseous composition and temper-
ature investigated in this work. The kinetics depicted in
Figs. 3 and 4 suggest competitive adsorption of dissocia-
tively chemisorbed oxygen and ethane (37) with stronger
binding of oxygen than ethane on the catalyst surface, which
leads to the appearance of the rate maximum at PO . This is
2
also corroborated by the fact that the open-circuit potential,
FIG. 4. Effect of PC
and temperature on the rate and turnover
6
H
2
VWo R , was found to be rather insensitive to PC H . Thus vary- frequency of ethane oxidation.
2
6

152
KALOYANNIS AND VAYENAS
FIG. 5. Electrochemical promotion. Ethane oxidation rate (continuous line) and catalyst potential (dashed line) response to step changes in
applied positive (a) and negative (b) current. The experimental ( ) and computed (2FN/I) rate relaxation time constants are also indicated. See text
for discussion. ro = 1.25 10 8 mol O/s, PO = 10.5 kPa, PC
= 1.62 kPa, N = 1.1 10 6 mol O.
H
6
2
2
The effect of positive current, I = 220 A, correspond- kinetics of the reaction of the promoting oxide species with
9
ing to the supply of I/2F = 1.14 10 mol O/s is shown C2H6 and can be predicted by
in Fig. 5a. At t 0 the circuit is open (I = 0) and the
1
steady-state regular (unpromoted) catalytic rate is 1.25
= 3 (TOFp)
,
[9]
S
8
10 mol O/s. Upon positive current application the rate
increases by a factor of 15.5. The rate increase 1r =
19.4 10 8 mol O/s is 170 times larger than I/2F. This im-
plies that each O2 supplied to the catalyst causes, on the av-
erage, 170 chemisorbed oxygen atoms to react with ethane
and form CO2 and H2O. Consequently for the galvanostatic
transient of Fig. 5a the enhancement factor, or Faradaic ef-
ficiency, 3, and rate enhancement ratio, , equal 170 and
16.5, respectively:
where TOFp isthe turnover frequencyofthe catalyticrate in
the promoted state. Equation [9] reflects the fact that, for
oxidation reactions, 3 expresses the ratio of the average
lifetimes of the promoting oxide species and of normally
chemisorbed oxygen on the catalyst surface (19). Thus for
the transient of Fig. 5b it is (TOFp) = 0.19 s 1, 3 = 170, and
consequently from Eq. [9] one obtains S = 15 min, which is
in very good qualitative agreement with the experimental
S value (Fig. 5a).
Figure 5b shows the effect of negative current applica-
tion under the same temperature and gaseous composition
as in Fig. 5a, thus starting from practically the same open-
circuit (unpromoted) rate. Negative current application
3 = 1r/(I/2F) = 170
= r/ro = 16.5.
[6]
[7]
It is worth noting the very good agreement between the
rate relaxation time constant (defined as the time required
for the rate increase to reach 63% of its final steady-state
value) and the parameter 2FNO/I, which isthe time required
to form a monolayer of promoting oxide ions (supplied at
a rate I/2F) on the catalyst surface with NO adsorption sites
(Fig. 5a):
9
(I = 450 A corresponding to the removal of 2.33 10
mol O/s) causes a nearly fivefold enhancement in the rate
( = 5.8). The rate increase 1r = 5.74 10 8 mol O/s is 25
times larger than the rate ( I/2F) of oxygen removal. Con-
sequently the Faradaic efficiency 3 equals 1r/(I/2F) = 25.
The negative sign shows that the reaction exhibits elec-
trophilic behavior (15, 17) upon negative current applica-
tion, i.e.,
2FNO /I.
[8]
Upon current interruption both the catalytic rate and
the catalyst potential return to their initial, unpromoted
values showing the reversibility of in situ controlled pro-
2
3 < 0, ∂r/∂VWR < 0, PIO < 0
[10]
motion. It is worth noting that according to the analysis of in contrast with the electrophobic behavior (15–17) ob-
electrochemical promotion transients outlined recently (19, served upon positive current application (Fig. 5a), i.e.,
42), and presented under Discussion, the time constant,
,
S
of the rate transient upon current interruption reflects the
2
3 > 0, ∂r/∂VWR > 0, PIO > 0.
[11]

ETHANE OXIDATION ON Pt
153
TABLE 1
Electrokinetic Parameters and Open-Circuit Catalytic Rate
(PO = 10.7 kPa, PC H = 1.65 kPa)
2
2
6
T
( C)
Io,a
A)
Io,c
A)
(
(
ro/(10 8 mol O/s) 2Fro/Io,a 2Fro/Io,c
a
c
420
460
500
0.11
0.95 0.81 0.29
0.68
1.27
2.04
11900
6600
2870
1380
230
97
0.37 10.63 0.87 0.25
1.37 40.64 0.87 0.25
hancement factor, or Faradaic efficiency, 3, takes values
between 100 and 500 for positive currents and between 10
and 100 for negative currents. The measured |3| values
are typically an order of magnitude smaller than those pre-
dicted by the approximate relationship
FIG. 6. Effect of catalyst overpotential (= VWR VWo R ) on the cell
current.
|3| 2Fro/Io
[2]
Electrocatalytic Kinetics
asalso shown in Table 1. It isworth noting, however, that de-
spite this significant deviation, the measured 3 values show
the same qualitative trends with the parameter 2Fro/Io; i.e.,
|3| is larger in the anodic region where Io is significantly
smaller (Table 1).
Figure 6 shows the dependence of current, I, on catalyst
overpotential, . The latter is defined from
= VWR VWo R
,
[12]
where VWo R is the open-circuit (I = 0) catalyst potential. It
is clear that the current-overpotential behavior is not sym-
metric for anodic (I > 0) and cathodic (I < 0) currents and
thus cannot be described by a single value of the exchange
current Io in the Butler–Volmer equation
Effect of Catalyst Potential
Figure 8 is based on the data of Figs. 6 and 7 and shows the
dependence of the catalytic rate on catalyst potential VWR
or, equivalently (3, 28, 29), change 1(e8) in catalyst work
function e8. The rate exhibits “inverted-volcano”-type
behavior (16) with the minimum near to the open-circuit
conditions. This type of behavior has also been observed
in several previous electrochemical promotion studies
ln(I/Io) = a F /RT
c F /RT.
[13]
This is because a single Butler–Volmer equation is satis-
fied only as long as the coverages of adsorbed species do not
change appreciable with catalyst potential which is clearly
not the case in most NEMCA studies (15–17). However, as
shown in Fig. 6, the I vs data can be adequately described
for intermediate | | values by the Tafel equations
ln(I/Io,a) = a F /RT
ln( I/Io,c) = c F /RT
[14]
[15]
for anodic and cathodic operation, respectively. In these
equations Io,a, Io,c and c are the apparent anodic and ca-
thodic exchange currents and transfer coefficients, respec-
tively. Table 1 lists the extracted Io,a, Io,c a, and c values.
,
a
,
Both Io,a and Io,c increase exponentially with temperature
with apparent activation energies of the order of 30 and
50 kcal/mol respectively, while a and c remain practically
constant.
Steady-State Effect of Current on the Catalytic Rate
FIG. 7. Effect of the rate, I/2F, of electrochemical oxygen ion supply
(I > 0) or removal (I < 0) on the induced increase in the rate of C2H6
oxidation.
Figure 7 shows the steady-state effect of applied current,
I, on the increase in the rate of ethane oxidation. The en-

154
KALOYANNIS AND VAYENAS
in this region. This exponential-type rate dependence on
catalyst potential and work function has also been observed
in numerous previous NEMCA studies (15–17).
For negative potentials the reaction exhibits electrophilic
behavior and Eq. [16] is again approximately satisfied with
values near 0.1.
By comparing Figs. 6 and 8, one can observe that the
potential—or overpotential—ranges where the exponen-
tial rate vs potential relationship holds (Eq. [16]) coincide
with the ranges where the exponential current vs potential
dependence is valid.
Figure 9a shows the effect of catalyst potential VWR on
the reaction kinetics with respect to PO . Two observations
2
can be made: (a) the rate is enhanced with both positive and
negative potentials and (b) the PO value at the rate maxi-
2
mum, P , increases significantly with increasing VWR and
O2
decreases with decreasing VWR (Fig. 9b). These observa-
tions indicate that increasing VWR, and catalyst work func-
tion e8, weakens significantly the chemisorptive bond of
oxygen as in all previous electrochemical promotion stud-
FIG. 8. Effect of catalyst potential and work function change on the
rate enhancement ratio (= r/ro) at fixed gaseous composition. The open-
circuit rate, ro, equals 0.68 10 8, 1.27 10 8, and 2.04 10 8 mol O/s at
T = 420, 460, and 500 C, respectively.
ies of catalytic oxidations (15–17). As shown in Fig. 9b, PO
2
increases by a factor of eight upon increasing VWR by 1.5 V.
Interestingly the variation is near linear.
of catalytic oxidations on Pt/YSZ (15–17) and Ag/YSZ
(15–17).
For positive potentials the reaction exhibits electropho-
bic behavior and in the range 0.15 V < VWR < 0.5 V the rate
increases exponentially with catalyst potential and work
function according to
Figure 10showsthe effect ofcatalyst potentialVWR on the
reaction kinetics with respect to ethane. It is worth noting
that: (a) positive potentials significantly enhance the rate
and decrease the apparent order ofthe reaction with respect
to PC H , the latter indicating enhanced binding of ethane
2
6
on the surface at high potentials where the rate dependence
on PC H parallels that of a Langmuir adsorption isotherm,
and (b) negative potentials enhance the rate only for low
ln(r/ro) = (e8 e8 )/kb T,
[16]
2
6
where kb is Boltzmann’s constant and the parameter
,
PC H values. Under such low PC H /PO values the rate is
2
6
2
6
2
termed NEMCA coefficient (15–17), is of the order of 0.35 negative order in oxygen (Fig. 3).
FIG. 9. (a) Effect of PO on the rate of C2H6 oxidation under open-circuit (o.c.) conditions and at various imposed catalyst potentials. (b) Effect
2
of imposed catalyst potential on the PO value, PO , which maximizes the rate.
2
2

ETHANE OXIDATION ON Pt
155
FIG. 10. Effect of PC
circuit (o.c.) conditions and at various imposed catalyst potentials.
on the rate of C2H6 oxidation under open-
6
H
2
FIG. 12. Effect of catalyst potential and work function on the appar-
ent activation energy, E, and on the logarithm of the preexponential factor
ro; roo is the open-circuit preexponential factor and T2, T20 are the isokinetic
points for positive and negative potentials, respectively.
Effect of Catalyst Potential on the Catalytic
Activation Energy
It is worth noting that the open-circuit Arrhenius line
does not pass from any of the two isokinetic points. This
must be because in open-circuit operation the catalyst po-
tential is not fixed to a specific value but varies somewhat
with temperature (between 0.1 and +0.05 V).
Figures 11a and 11b show Arrhenius plots obtained at
various fixed values of catalyst potential and compare them
with the open-circuit (o.c.) Arrhenius plot. Figure 11a refers
to positive potentials and Fig. 11b to negative ones. In both
cases compensation effect behavior is obtained (43, 44).
In the former case all Arrhenius lines corresponding to
a fixed potential converge to a single isokinetic point,
T2 = 230 C. In the latter case again all the fixed potential
Arrhenius lines converge to a second isokinetic point T20 =
294 C.
Figure 12 shows the dependence of the apparent acti-
vation energy values, E, extracted from the Arrhenius plots
of Figs. 11a and 11b, on the catalyst potential VWR. In-
terestingly the E vs VWR dependence parallels the r vs
VWR dependence (Fig. 7), i.e., E exhibits a minimum near
the open-circuit conditions. In the electrophobic region
FIG. 11. Promotionally induced compensation effect. Arrhenius plots at various fixed positive (a) and negative (b) catalyst potentials; comparison
with the open-circuit (o.c.) Arrhenius plot.

156
KALOYANNIS AND VAYENAS
an active catalyst support and promoter donor. Similarly
to the case of C2H4 oxidation on Pt/YSZ (2), Pt/ 00-Al2O3
(4) and Rh/YSZ (19) the reaction is strongly electrophobic
at high potentials; i.e., the rate increases significantly with
O2 supply to the catalyst. The maximum observed catalytic
rate enhancement in this region ( = r/ro 20) is a factor
of three lower than the maximum value ( 60) measured
with C2H4 oxidation on Pt/YSZ.
All the observed features of electrochemical promotion
in this region, i.e., the order of magnitude of 3( 2Fro/Io)
and of ( 2FN/I), are consistent with the electrochemi-
cally controlled oxide ion backspillover promoting mecha-
nism (1, 15–17) now strongly supported by XPS (21–23),
SERS (24, 25), TPD (26), cyclic voltammetry (26), and
STM (27).
Interestingly in the present case ethane oxidation is also
found to exhibit a rather strong electrophilic behavior
FIG. 13. Effect of the apparent activation energy, E, on the apparent
preexponential factor, ro, at various fixed positive (squares) and negative
(
7) at low potentials, i.e., with electrochemical removal
of oxygen from the Pt surface. Clearly this type of elec-
trophilic electrochemical promotion, also observed in the
cases of CO oxidation on Pt/YSZ (15, 16) and Ag/YSZ
(15, 16) and of C3H6 oxidation on Pt/YSZ (39), is not due
to oxide ion backspillover, since oxygen is in this case being
removed electrochemically from the Pt surface. That the
promoting mechanism in this region, of negative currents
and low potentials, is different from that in the region of
positive currents and high potentials is also manifest by the
pronounced minima in reaction rate and activation energy
near the open-circuit conditions (Figs. 8 and 12) and also by
the appearance of two different isokinetic points (Figs. 11a
and 11b).
(circles) potentials; open symbol corresponds to open-circuit conditions.
(∂r/∂VWR > 0) the activation energy E increases linearly
with VWR (and e8) with a slope H = 0.8:
1E = H 1(e8).
[17]
In the electrophilic region (∂r/∂VWR < 0) the activation
energy E conforms again to Eq. [14] with 0.4. It is
=
H
worth noting that E in the electrophobic region is signifi-
cantly higher than in the electrophilic one.
The apparent preexponential factor ro, defined from
r = ro exp( E/kb T ),
[18]
The observed compensation effect (42, 43) shows, in con-
junction with other recent similar results (15–17), that the
parallels the activation energy behavior (Fig. 12). Its loga- promotion-induced compensation effect is a rather com-
rithm varies linearly with VWR and in fact when plotted as mon phenomenon resulting from the linear variation of ac-
kbT2ln(ro/roo), where T2 (or T20 ) is the isokinetic point, and tivation energy and logarithm of preexponential factor with
roo is the open-circuit preexponential factor, one obtains the catalyst potential and work function (Fig. 12).
same functional dependence on VWR, as the E dependence
Origin of the Promoting Action
on VWR, diplaced vertically by the open-circuit activation
energy Eo (Fig. 12).
As previously noted it is now well established that when
using YSZ and positive currents, the NEMCA effect is due
to the promotional action of backspillover oxide ions O
(O1s binding energy at 528.8 eV (21) peak desorption tem-
perature Tp 750–780 K (26)) which spread over the entire
catalyst surface, increase its work function by
This is because, as shown previously (19), when an isoki-
netic point exists, then E and ln ro are related via
E
Eo = kb T2 ln ro/roo
.
[19]
Equation [19] is satisfied for every VWR value and conse-
quently when ln(ro/roo) is plotted vs E two straight lines are
obtained with slopes kbT2 and kb T20 , respectively as shown
in Fig. 13.
1(e8) = e1VWR
,
[5]
and act as sacrificial promoters by reacting with ethane at
a rate 3 times smaller than the displaced (26) more weakly
bonded atomic oxygen (Tp 675–685 K).
DISCUSSION
The present work shows that the catalytic activity of
It is worth emphasizing that Eq. [5] is also valid when
Pt for ethane oxidation can be markedly affected via the applying negative currents, i.e., in the low-catalyst potential
NEMCA effect by using Y2O3-stabilized zirconia-ZrO2 as region as well. Also in both regions the Helmholz equation

ETHANE OXIDATION ON Pt
157
is valid (15–17),
e1VWR = 1(e8) =
Activation Energy Dependence on Work Function
and Compensation Effect
X
eNm
1(
j ),
[20]
j
The above simple promotional rule can also explain the
observed peculiar dependence of the apparent activation
εo
j
where Nm is the surface metal atom concentration (e.g., energy on catalyst potential (Fig. 12).
19
1.53 1019 atom/m2 for Pt(111)) e = 1.6 10
C/atom,
We start the analysis by noting that the apparent activa-
12
εo = 8.85 10
C2/J m and
stands for the (average) tion energy of a bimolecular catalytic reaction
j
dipole moment ofallspeciesjpresent on the catalyst surface
at a coverage j. When applying positive currents and thus
increasing VWR in the Pt/YSZ system, the increase in e8 is
accommodated primarily by the increase in the coverage of
the backspillover oxide ions O whose dipole moment is
expected to be higher than that of other covalently bonded
species (e.g., coadsorbed atomic oxygen and ethane). How-
ever, changes in the coverages and dipole moments of these
coadsorbed species are also quite possible, as actually man-
ifest in TPD studies of oxygen on Pt/YSZ (26) where sig-
nificant changes in the chemisorptive state of coadsorbed
atomic oxygen was observed (27).
When usingnegative currentsin the Pt/YSZ system, there
can be no backspillover oxide migration to the catalyst
surface, only chemisorbed oxygen removal, and thus the
Helmholz equation [20] has to be accommodated only by
changes in the coverages and dipole moments of dissocia-
tively chemisorbed oxygen and C2H6. A decrease in oxygen
coverage will in general cause a decrease in e8 (15–17). A
decrease in dipole moment due to tighter binding of oxy-
gen on the Pt surface will also cause the same effect. As
manifest in Fig. 9 negative potentials cause a decrease in
A + B → C + D
following Langmuir–Hinshelwood kinetics; i.e.,
r = kR kA kB PA PB/(1 + kA PA + kB PB)2
is given by the expression (5)
[21]
[22]
kA PA QA + kB PB QB
E = ER QA QB + 2
,
[23]
1 + kA PA + kB PB
where ER is the true activation energy, kA and kB are the
adsorption equilibrium constants of A and B, and QA, QB
are the corresponding heats of adsorption of A and B. In
the case of ethane oxidation and in view of the fact that the
stoichiometry of the dissociative ethane chemisorption is
not well known (37) one can write to a first approximation
Eq. [23] as
kO P1/2 QO + k
P
Q
C2H6 C2H6 C2H6
O2
E = ER QO QC H + 2
.
2
6
1 + kO P1/2 + kC H PC H
2
6
2
6
O2
[24]
==
PO , thus a strengthening of the Pt O bond, as expected for
ele2ctron donor adsorbates, such as atomic oxygen, upon de-
creasing the catalyst work function (15–17). Consequently
one can attribute the electrochemical promotional action
in the region of negative currents (and low potentials) to
Eq. [5] in conjunction with the Helmholz equation [20],
i.e., to the fact that a decrease in potential forces a tighter
binding of dissociatively chemisorbed oxygen, and weaker
binding of dissociatively chemisorbed C2H6 on the Pt sur-
face (15–17) with a concomitant readjustment in surface
coverages of oxygen and C2H6 in order to satisfy Eq. [5].
One can thus rationalize all the observed promotional
In view of the observed kinetic behavior (Figs. 9 and 10) it
is reasonable to assume that kO P1/2 > 1 and kC H PC H > 1,
2
6
2
6
O2
in which case Eq. [24] can be written as
kO P1/2
k
P
C2H6 C2H6
O2
E = ER +
[QO QC H ].
[25]
2
6
kO P1/2 + k
P
C2H6 C2H6
O2
We then note that, as previously discussed, QO is ex-
pected to decrease with increasing e8, while QC H is ex-
pected to increase. Assuming linear variations (30), i.e.,
2
6
O2
QO = QoO
11(e8)
21(e8),
[26]
[27]
features on the reaction kinetics with respect to P (Fig. 9)
and PC H (Fig. 10) in termsof the followingsimple rule used
2
6
QC H = QoC H
+
2
6
2
6
to explain all previous NEMCA studies. Increasing cata-
lyst work function, either via the action of backspillover
ions or via readjustment in surface coverages and dipole
moments due to the applied potential, causes a weaken-
ing in the chemisorptive bond strength of dissociatively
chemisorbed oxygen (electron acceptor) and a strength-
ening in the chemisorptive bond strength of dissociatively
chemisorbed C2H6 (electron donor). This causes both the
where
, 2 > 0, one can rewrite Eq. [25] as
1
kO P1/2
k
P
C2H6 C2H6
O2
E = ER +
kO P1/2 + k
P
C2H6 C2H6
O2
QoO QoC H
(
+
2)1(e8) . [28]
1
2
6
increase in PO (Fig. 9) and the decrease in apparent re-
This equation can be used in the attempt to rationalize
2
action order with respect to C2H4 (Fig. 10) with increasing the observed complex dependence of E on 1(e8) in a qual-
VWR and e8. itative fashion. The sign of ∂E/∂(e8) depends on the sign

158
KALOYANNIS AND VAYENAS
of the parameter A = kO P1/2 kC H PC H . For high 1(e8)
tained from the initial rate slope upon current interruption.
From the transient upon current interruption of Fig. 5a one
obtains = 5.6, in reasonable agreement with the value es-
timated upon current imposition.
2
6
2
6
O2
values oxygen adsorption is weakened and ethane adsorp-
tion is strengthened; A is thus negative and, according to
Eq. [28], E increases linearly with 1(e8) as experimentally
observed (Fig. 12). For negative 1(e8) values oxygen ad-
sorption is strong and ethane adsorption is weak; thus A
is positive and, according to Eq. [28], E increases linearly
with decreasing 1(e8), again in good qualitative agreement
with experiment (Fig. 12).
Once has been estimated, one can then compute
O
from Eq. [34] for any time t from the corresponding r/ro
value, i.e.,
1
ln(r/ro)
=
ln(r/ro) =
(I/2F NO ).
[35]
O
d ln(r/ro)/dt
Transient Analysis and Promotion Index
The maximum
sient of Fig. 5a to be
the observed value (
order of 33 in good qualitative agreement with PIO values
value is thus computed for the tran-
0.5. This, in view of Eq. [4] and
O
As previously noted, determination of the promotion in-
dex PIO of the promoting oxide ions O requires knowl-
edge of the coverage of O on the catalyst surface. This
can be done by using the rate transient analysis upon cur-
rent interruption. We first consider the mass balance of O
O
16.5) gives a PIO value of the
for other Pt catalyzed oxidations [15–17].
CONCLUSIONS
during current application (0
t
to)
d
Ethane oxidation on Pt films deposited on YSZ exhibits
a rather strong electrochemical promotion effect at tem-
peratures 400 to 500 C. Both positive and negative applied
potentials enhance the rate of oxidation by up to a factor
of 20 and 7, respectively. The rate increase is typically a
factor of 200 to 20 higher than the rate of electrochemical
supply or removal of oxygen, respectively. The reaction ki-
netics and apparent activation energy are also significantly
modified by the applied potential leading to the appearance
of the compensation effect. These modifications as well as
the pronounced rate enhancement are consistent with the
notion that increasing catalyst potential and work function
weakens the chemisorptive bond strength of oxygen and
enhances dissociative ethane chemisorption on the Pt sur-
face while decreasing catalyst potential and work function
has exactly the opposite effect.
The observed transient behavior with positive current
application is in good agreement with the oxide-ion back-
spillover mechanism for electrochemical promotion and al-
lows for the estimation of the coverage of the promoting
oxidic backspillover species. The picture is still less clear re-
garding the promoting mechanism with negative currents,
i.e., electrochemical oxygen removal from the catalyst sur-
face. The use of surface spectroscopic techniques under
such conditions could provide useful additional insight.
O
NO
= (I/2F) NO f (
),
[29]
O
dt
where f (
) is the rate of consumption of O on the
O
catalyst surface, due to reaction with C2H6. We then note
that in the region of positive potentials the catalytic rate is
exponentially dependent on 1(e8) (Eq. [16], Fig. 8). The
change in work function e8, 1(e8), is linearly related to
, provided the dipole moment of O is constant with
coverage. On the basis of this assumption one can write
O
r = ro exp(
),
[30]
O
where is a constant. It follows then from Eq. [29] and the
initial condition (t = 0,
= 0) that
d ln(r/ro)
O
d
O
NO
= NO
= (I/2F) .
[31]
dt
dt
t=0
t=0
Consequently can be obtained from the initial slope of
ln(r/ro) vs t. For the rate transient upon current imposition
(Fig. 5a) one can estimate a value of 6.5.
We then note that as steady state it is
(I/2F) = NO f (
)
[32]
O
and that, in view of Eq. [31],
ACKNOWLEDGMENT
d
O
NO
=
NO f (
) = (I/2F),
[33]
O
We thank the PENED Programme of the Hellenic Secretariat of
Research and Technology for partial financial support.
dt
t=to
where the last equality in [33] holds because of Eq. [32]. In
view of Eq. [33] one has
REFERENCES
1. Vayenas, C. G., Bebelis, S., and Neophytides, S., J. Phys. Chem. 92,
5083 (1988).
2. Bebelis, S., and Vayenas, C. G., J. Catal. 118, 125 (1989).
3. Vayenas, C. G., Bebelis, S., and Ladas, S., Nature (London) 343, 625
(1990).
d ln(r/ro)
NO
=
(I/2F),
[34]
dt
t=to
which implies that a second estimate of can also be ob-

ETHANE OXIDATION ON Pt
159
4. Pritchard, J., Nature (London) 343, 592 (1990).
5. Yentekakis, I. V., Moggridge, G., Vayenas, C. G., and Lambert, R. M.,
J. Catal. 146, 293 (1994).
6. Vayenas, C. G., Bebelis, S., and Despotopoulou, M., J. Catal. 128, 415
(1991).
7. Politova, T. I., Sobyanin, V. A., and Belyaev, V. D., React. Kinet. Catal.
Lett. 41, 321 (1990).
8. Makri, M., Buekenhoudt, A., Luyten, J., and Vayenas, C. G., Ionics 2,
(1996).
22. Palermo, A., Tikhov, M. S., Filkin, N., Lambert, R. M., Yentekakis,
I. V., and Vayenas, C. G., Stud. Surf. Sci. Catal. 101, 513 (1996).
23. Arakawa, T., Saito, A., and Shiokawa, J., Chem. Phys. Lett. 94, 250
(1983); Appl. Surf. Sci. 16, 365 (1983).
24. Basini, L., Cavalca, C. A., and Haller, G. L., J. Phys. Chem. 88, 10853
(1994).
25. Kondarides, D. I., Papatheodorou, G. N., Vayenas, C. G., and Verykios,
X. E., Ber. Bunsenges. Phys. Chem. 97, 709 (1993).
26. Neophytides, S., and Vayenas, C. G., J. Phys. Chem. 99, 17063 (1995).
27. Makri, M., Vayenas, C. G., Bebelis, S., Besocke, K., and Cavalca, C. A.,
Surf. Sci. 369, 351 (1996).
9. Tsiplakides, D., Neophytides, S., Enea, O., Jaksic, M. M., and Vayenas,
C. G., J. Electrochem. Soc. [In press]
10. Yentekakis, I. V., and Vayenas, C. G., J. Catal. 149, 238 (1994).
11. Pliangos, C., Yentekakis, I. V., Ladas, S., and Vayenas, C. G., J. Catal.
159, 189 (1996).
12. Neophytides, S., Tsiplakides, D., Stonehart, P., Jaksic, M. M., and
Vayenas, C. G., Nature (London) 370, 45 (1994).
13. Neophytides, S. G., Tsiplakides, D., Stonehart, P., Jaksic, M. M., and
Vayenas, C. G., J. Phys. Chem. 100, 14803 (1996).
14. Anastasievic, N., and Werner, D., in “Proceedings, GDCh Angewandte
Elektrochemie,” Monheim, October 1996.
28. Ladas, S., Bebelis, S., and Vayenas, C. G., Surf. Sci. 251/252, 1062
(1991).
29. Zipprich, W., Wiemho¨fer, H.-D., Vohrer, U., and Go¨pel, W., Ber. Bun-
senges. Phys. Chem. 99, 1406 (1995).
30. Pacchioni, G., Ilas, F., Neophytides, S., and Vayenas, C. G., J. Phys.
Chem. 100, 16653 (1996).
31. Yao, Y.-F. Y., and Kummer, J. T., J. Catal. 28, 124 (1973).
32. Fairbridge, C., and Ross, R. A., Can. J. Chem. 60, 893 (1982).
33. Wachs, I. E., and Kelemen, S. R., J. Catal. 68, 213 (1981).
15. Vayenas, C. G., Bebelis, S., Yentekakis, I. V., and Lintz, H.-G., Catal. 34. Erdo¨helyi, A., and Solymosi, F., J. Catal. 123, 31 (1990).
Today 11, 303 (1992).
16. Vayenas, C. G., Jaksic, M. M., Bebelis, S., and Neophytides, S. G.,
35. Oyama, S. T., Middlebrook, A. M., and Somorjai, G. A., J. Phys. Chem.
94, 5029 (1990).
in “Modern Aspects of Electrochemistry” (J. O’M. Bockris, B. E. 36. Harold, M. P., and Luss, D., Ind. Eng. Chem. Res. 26, 2092 (1987). [And
Conway, and R. E. White, Eds.), Vol. 29, p. 57. Plenum Press, New
York, 1995. [And references therein]
references therein]
37. McMaster, M. C., and Madix, R. J., Surf. Sci. 275, 265 (1992).
17. Vayenas, C. G., and Neophytides, S. G., in “Catalysis,” Vol. 12, Chap. 6, 38. Liu, A. C., and Friend, C. M., Surf. Sci. 216, 33 (1989).
p. 199. The Royal Society of Chemistry, Cambridge, UK, 1996.
18. Varkaraki, E., Nicole, E., Plattner, E., Comninellis, Ch., and Vayenas,
C. G., J. Appl. Electrochem. 25, 978 (1995).
39. Kaloyannis, A. C., Pliangos, C. A., Yentekakis, I. V., and Vayenas,
C. G., Ionics 1, 159 (1995).
40. Yentekakis, I. V., and Bebelis, S., J. Catal. 137, 278 (1992).
19. Pliangos, C., Yentekakis, I. V., Verykios, X. E., and Vayenas, C. G., 41. Cavalca, C. A., Larsen, G., Vayenas, C. G., and Haller, G. L., J. Phys.
J. Catal. 154, 124 (1995).
Chem. 97, 6115 (1993).
20. Marina, O., Yentekakis, I. V., Vayenas, C. G., Palermo, A., and
Lambert, R. M., J. Catal. 166, 218 (1997).
42. Karavassilis, Ch., Bebelis, S., and Vayenas, C. G., J. Catal. 160, 190, 205
(1996).
21. Ladas, S., Kennou, S., Bebelis, S., and Vayenas, C. G., J. Phys. Chem. 43. Cremer, E., Adv. Catal. 7, 75 (1955).
97, 8845 (1993). 44. Schwab, G.-M., J. Catal. 84, 1 (1983).