First Demonstration of in Situ Electrochemical Control of a Base Metal Catalyst: Spectroscopic and Kinetic Study of the CO + NO Reaction over Na-Promoted Cu

Article abstract of DOI:10.1021/jp992270d

Controlled, reversible electrochemical promotion (EP) of a base metal catalyst has been demonstrated for the first time. Electropumping of Na from a ?2a?3 alumina solid electrolyte to a Cu film catalyst results in large improvements in both activity and selectivity of the latter. In the catalytic reduction of NO by CO, the reactive behavior, surface composition, and response to EP are a strong function of the composition of the reactant gas. Electron spectroscopic data show that these effects are due to pumping of Na to the catalyst where, under reaction conditions, it is present as NaNO₃ on an oxidized Cu surface. Taken together, the spectroscopic and reactor results show that Cu⁰ sites are not of significance and that the catalytically active surface is dominated by Cu⁺ and Cu²⁺ sites. They also suggest that Cu⁺ is of principal importance for the dissociative adsorption of NO and that EP is due to Na-induced enhancement of the adsorption and dissociation of NO at these sites.

Full text of DOI:10.1021/jp992270d

9960
J. Phys. Chem. B 1999, 103, 9960-9966
First Demonstration of in Situ Electrochemical Control of a Base Metal Catalyst:
Spectroscopic and Kinetic Study of the CO + NO Reaction over Na-Promoted Cu
Federico J. Williams, Alejandra Palermo, Mintcho S. Tikhov, and Richard M. Lambert*
Department of Chemistry, UniVersity of Cambridge, Cambridge CB2 1EW, U.K.
ReceiVed: July 1, 1999; In Final Form: September 20, 1999
Controlled, reversible electrochemical promotion (EP) of a base metal catalyst has been demonstrated for the
first time. Electropumping of Na from a â′′ alumina solid electrolyte to a Cu film catalyst results in large
improvements in both activity and selectivity of the latter. In the catalytic reduction of NO by CO, the reactive
behavior, surface composition, and response to EP are a strong function of the composition of the reactant
gas. Electron spectroscopic data show that these effects are due to pumping of Na to the catalyst where,
under reaction conditions, it is present as NaNO3 on an oxidized Cu surface. Taken together, the spectroscopic
and reactor results show that Cu0 sites are not of significance and that the catalytically active surface is
dominated by Cu+ and Cu2+ sites. They also suggest that Cu+ is of principal importance for the dissociative
adsorption of NO and that EP is due to Na-induced enhancement of the adsorption and dissociation of NO
at these sites.
Introduction
of the metal. This suggests that at much higher reactant pressures
it is unlikely that Cu0 metal sites survive in a net oxidizing
environment. Key information regarding the oxidation state of
alumina-supported Cu catalysts under reaction conditions at
atmospheric pressure comes from an in situ X-ray absorption
near-edge structure (XANES) study carried out by Fernandez-
Garcia et al.19 Their results indicate that both Cu+ and Cu2+
sites participate in the overall process, with Cu+ being involved
in the rate-limiting step.
Here we report on the electrochemical promotion (EP) by
Na of the CO + NO reaction over a copper film catalyst
supported on Na â′′ alumina solid electrolyte, which acts as
the source of electropumped spillover Na. This is the first time
that EP has been observed with a base metal catalyst. The results
shed light on the chemical composition of the catalytically active
surface and on the way in which this composition varies as the
CO/NO ratio varies from net reducing to net oxidizing condi-
tions. In addition, we address certain aspects of the reaction
mechanism and identify the chemical state of the Na promoter.
In agreement with London and Bell12 it is found that Cu nitrate
is not present on the active surface. In agreement with Fernadez-
Garcia et al.,19 we show that the latter consists of Cu(I) and
Cu(II) oxides, at least within the sampling depth of the X-ray
photoelectron (XP) spectra (∼1 nm). Comparison with our
earlier EP study of the Pt-catalyzed CO + NO reaction20 is
instructive and allows conclusions to be drawn about active sites
and the mechanism of promotion in the Cu system.
Copper-based catalysts for NO reduction have been studied
for many years because their low cost offers large potential
economic benefits in the field of emission control catalysis.1
Extensive research on the use of such catalysts for NO reduction
carried out in the 1960s and early 1970s2,3 has been reviewed
by Shelef.4 In this connection, the CO + NO reaction has
received the most attention because of its simplicity and relative
ease of investigation; a recent comprehensive review is avail-
able.5 There is no generally agreed reaction mechanism, and
there are divergent opinions about the nature of the catalytically
active site or sites. An important issue concerns the oxidation
state of the catalytically active surface. It seems likely that, at
least in part, disagreements in the literature reflect the sensitivity
of the copper/oxygen system to changes in oxidation potential
of the gas phase. With oxygen as oxidant, small changes in O2
partial pressure can result in abrupt phase transitions between
Cu metal, Cu2O, and CuO, accompanied by pronounced changes
in catalytic behavior.6-11 Furthermore, depending on the rates
of oxygen transfer to and from the solid, the catalyst composition
may not be uniform as one proceeds from the surface to bulk.
With NO as oxidant, similar effects are to be expected, although
less pronounced than in the case of O2. In this connection, it is
noteworthy that the steady-state reaction kinetics are essentially
the same, regardless of whether the starting material is Cu metal
or Cu2O.5 A variety of reaction mechanisms has been proposed,
often containing a large number of postulated elementary steps
and adsorbed species. Some authors have proposed that Cu0,
Cu+, and Cu2+ sites are all involved in catalytic turnover.12
Others have suggested13 that adsorbed nitrogen oxyanions
participate in the reaction, although infrared spectra12 indicate
that such species are not present on the active catalyst. Single
crystal data obtained under ultrahigh vacuum conditions provide
valuable basic information. They show that NO dissociates
completely on both smooth and stepped Cu surfaces14-18
accompanied by facile desorption of N2 and incipient oxidation
Experimental Methods
The Cu catalyst (working electrode, W) consisted of a porous,
continuous thin film deposited on one face of a 10 mm × 10
mm × 1.5 mm wafer of Na â′′ alumina solid electrolyte. Au
reference (R) and counter (C) electrodes were deposited on the
other face, all three electrodes being deposited by Cu or Au
sputtering in argon.21 XPS analysis of the as-deposited Cu
electrode showed a measurable C 1s signal (binding energy of
284 eV) corresponding to a submonolayer quantity of carbon.
After exposure to reaction gas, the C 1s intensity was reduced
* To whom correspondence should be addressed.. Phone: 44 1223
336467. Fax: 44 1223 336362. E-mail: RML1@cam.ac.uk.
10.1021/jp992270d CCC: $18.00 © 1999 American Chemical Society
Published on Web 10/26/1999

CO + NO Reaction
J. Phys. Chem. B, Vol. 103, No. 45, 1999 9961
to an undetectable level. The surface area of the Cu catalyst
working electrode (W) was estimated by measuring galvano-
static transients in a He atmosphere, as described in detail
elsewhere.21 The sample was suspended in a quartz, atmospheric
pressure well-mixed reactor (50 cm3) with all electrodes exposed
to the reactant gas mixture. The system behaved as a single
pellet, continuous stirred tank reactor (CSTR) as described and
discussed elsewhere.22
Inlet and exit gas analysis was carried out by a combination
of on-line gas chromatography (Shimadzu-14B; molecular sieve
5A and Porapak-N columns) and on-line mass spectrometry
(Balzers QMG 064). N2, N2O, CO, and CO2 were measured by
gas chromatography, and NO was monitored continuously by
mass spectrometry after performing the required calibrations.
Reactants were pure NO (Distillers MG) and CO (Distillers MG)
diluted in ultrapure He (99.996%) and fed to the reactor by mass-
flow controllers (Brooks 5850 TR). The total flow rate was kept
constant in all experiments at 6.8 × 10-5 mol s-1 (100 cm3
(STP)/min), with partial pressures PNO and PCO of 0.5-6 and
1.5-2.5 kPa, respectively, and PHe giving a total pressure of 1
atm in every case. Conversion of the reactants was restricted to
<15% in order to avoid mass transfer limitations. Control
experiments were carried out in which the total flow was varied
by a factor of 2 in order to verify that the observed changes in
activity were due to actual changes in surface reaction rates
and not due to mass transfer limitations. Nitrogen and carbon
mass balances always closed to within 2%.
Figure 1. Effect of catalyst potential (VWR) on the CO2, N2, N2O
formation rates and the selectivity of NO reduction to nitrogen.
Conditions are T ) 648 K, CO/NO ) 5:1.
A galvanostat-potentiostat (Ionic Systems) was used to
maintain a potential difference between the working and
reference electrodes (potentiostatic mode). All experiments were
carried out in potentiostatic mode by following the effect of
catalyst potential (VWR, measured with respect to the reference
electrode) on the reaction rates.
Postreaction XPS measurements were carried out in a VG
ADES 400 ultrahigh vacuum spectrometer system equipped with
a reaction cell. The EP sample was mounted on a manipulator
that allowed translation between the reaction cell and the
spectrometer chamber. The sample holder was made from a
machinable ceramic containing embedded Nichrome filaments
for resistive heating, and all electrical connections were made
with gold wire. XP spectra were acquired with Mg KR radiation,
constant pass energy of 20 eV, and with the Cu working
electrode held at ground potential. Quoted binding energies are
in reference to the Au 4f7/2 line at 83.8 eV, emitted by the Au
wire used to make the electrical connection to the working
electrode.
Figure 2. Effect of catalyst potential (VWR) on the CO2, N2, N2O
formation rates and the selectivity of NO reduction to nitrogen.
Conditions are T ) 648 K, CO/NO ) 2:1.
Results
Effect of Catalyst Potential (VWR) on Reaction Rates.
Figures 1-3 show steady-state (potentiostatic) rate data for CO2,
N2, and N2O production as a function of catalyst potential (VWR
)
at 648 K for a range of reactant gas compositions determined
by (fixed) inlet pressures (P0NO, P0CO) as follows: (Figure 1)
P0 ) 2.5 kPa, P0 ) 0.5 kPa (CO/NO ) 5:1); (Figure 2)
Figure 3. Effect of catalyst potential (VWR) on the CO2, N2, N2O
formation rates and the selectivity of NO reduction to nitrogen.
Conditions are T ) 648 K, CO/NO ) 1:4.
CO
NO
P0CO ) 2 kPa, P0NO ) 1 kPa (CO/NO ) 2:1); (Figure 3) P0
CO
) 1.5 kPa, P0NO ) 6 kPa (CO/NO ) 1:4). Turnover frequencies
(TOF) are expressed as molecules of product per Cu surface
atom per second. Also shown in Figures 1-3 is the dependence
of N2 selectivity on catalyst potential, where selectivity is
defined as
These data show that all rates (CO2, N2, and N2O) increase with
decreasing VWR, which corresponds to progressively increasing
the sodium loading of the Cu catalyst surface. Two features
are evident: (i) for values of VWR between 100 and -100 mV
the reaction rates are almost constant (unpromoted regime) and
(ii) there is a sharp increase in activity as VWR is reduced below
-100 mV (promoted regime). This behavior was reproducible
and fully reversible. The rate enhancement ratios (F) defined
r
N2
SN
)
(I)
2
r
+ r
N2O
N2

9962 J. Phys. Chem. B, Vol. 103, No. 45, 1999
Williams et al.
Figure 5. Na 1s XP spectra of the catalyst film after exposure to
different reaction mixtures. The results refer to unpromoted and
promoted conditions for different CO/NO ratios: (i) CO/NO ) 5:1,
(ii) CO/NO ) 2:1, and (iii) CO/NO ) 1:4.
exposing the catalyst film to the same conditions of temperature
and reactant partial pressures as those used in the CST reactor.
During exposure of the catalyst to the reaction mixture, the VWR
conditions were such that the Cu film was either (i) unpromoted
or (ii) Na-promoted. The spectra were acquired by the following
procedure. (i) After exposure to reaction gas at the appropriate
temperature, the sample was cooled to 330 K in reaction gas.
(ii) Open-circuit conditions were imposed when the sample
temperature was below 350 K, i.e., when Na mobility was low.
The object of this procedure was to “freeze out” the surface
conditions pertaining to the catalytically active surface.
These postreaction XP/AE experiments were performed in
order to shed light on several important issues. (i) What evidence
is there for the reversible transport of Na between electrolyte
and catalyst under reaction conditions? (ii) What differences in
catalyst surface composition are there between the unpromoted
and promoted regimes? (iii) What surface species are formed
by Na pumping to the Cu catalyst at reaction temperature in
the presence of the reactant gases? (iv) What is the oxidation
state of the Cu catalyst? We therefore exposed the sample to
the same conditions as those used in the reactor, all at 648 K,
1 bar total pressure, as follows: (i) CO/NO ) 5:1, (ii) CO/NO
Figure 4. (a) Cu 2p XPS and (b) Cu LMM XAES for metallic Cu,
Cu2O, and CuO.
as the ratio of the maximum promoted rates r to the unpromoted
rates r0 are as follows for the three cases shown: (1) F(CO2) )
20, F(N2) ) 26, F(N2O) ) 11, (2) F(CO2) ) 26.4, F(N2) ) 66,
F(N2O) ) 8.8, and (3) F(CO2) ) 2.5, F(N2) ) 4.4, F(N2O) )
1.2. It is apparent that decreasing VWR not only caused
substantial increases in all reaction rates but also caused
appreciable enhancement of the nitrogen selectivity in every
case: (a) from 50% to 75%, (b) from 29% to 80%, and (c)
from 20% to 50%. (Note that the behavior observed at the
stoichiometric composition CO/NO ) 1:1 was very similar to
that found for CO/NO ) 2:1, namely, F(CO2) ) 16, F(N2) )
34, F(N2O) ) 3.2; S(N2) increased from 25% to 75%.)
Postreaction XPS Studies of the EP Catalyst. XPS and
Auger spectra were acquired after sample transfer to the electron
spectrometer chamber following exposure to appropriate reaction
gas environments in the reactor cell. For comparison purposes,
we acquired reference Cu 2p XP and Cu LMM Auger electron
(AE) reference spectra for Cu, Cu2O, and CuO using the
procedure described by Poulston et al.23 These are illustrated
in Figure 4. Postreaction XP/AE spectra were obtained after
) 2:1, and (iii) CO/NO ) 1:4 under unpromoted (VWR
)
+100mV) and promoted (VWR ) -400mV) conditions. After
these gas exposures the sample was transferred to the spec-
trometer chamber, as described above, and the XP spectra
recorded.
A combination of Cu XP and X-ray excited Auger electron
(XAE) spectra is required in order to distinguish between Cu
metal, Cu2O, and CuO.11,23 Thus, CuO is characterized by high-
intensity shake-up satellites at ∼9 eV higher binding energy
than the main 2p3/2 and 2p1/2 peaks. Additionally, the CuO main
lines are shifted 1.2 eV toward higher binding energy and
broadened compared to Cu2O and Cu metal.24 Cu metal and
Cu2O are best distinguished by their X-ray excited Cu LMM
Auger spectra, since their Cu 2p transitions differ by only ∼0.2
eV. We therefore recorded appropriate reference spectra, and
these data are shown in Figure 4.
Figure 5 shows Na 1s XP spectra of the catalyst film taken
after exposure to three different reaction mixtures under both

CO + NO Reaction
J. Phys. Chem. B, Vol. 103, No. 45, 1999 9963
TABLE 1: Determination of Surface Oxidation State Using
Cu 2p XP and Cu LMM XAE Spectraa
unpromoted (+100 mV)
promoted (-400 mV)
CO/NO ) 5:1
CO/NO ) 2:1
CO/NO ) 1:4
Cu0 + Cu2O
(48% + 52%)
Cu2O
(100%)
Cu2O + CuO
(19% + 81%)
Cu2O + CuO
(67% +33%)
Cu2O + CuO
(28% + 72%)
Cu2O + CuO
(13% + 87%)
a Spectra acquired after exposure of the catalyst to the stated CO/
NO ratios for values of VWR corresponding to the promoted and
unpromoted catalyst.
consists of metallic Cu alone; (2) for a given catalyst potential,
as the CO/NO ratio decreases (i.e., as the gaseous environment
becomes more oxidizing), the oxidation state of the copper
increases; (3) for a given CO/NO ratio, as the sodium coverage
increases, the oxidation state of the copper is increased. Thus,
the bottom spectra of Figures 6 and 7 show the Cu 2p and Cu
LMM XAES spectra of the catalyst “as received”. The sharp
peak present at ∼918.5 eV kinetic energy in the Cu LMM
spectra and the shape and binding energy positions of the Cu
2p3/2 and Cu 2p1/2 demonstrate that the initial oxidation state of
the Cu in the catalyst is zero, as comparison with the reference
spectra in Figure 4 confirms. Exposing this “as-received” film
to the same conditions of pressure and temperature as those
used in the CST reactor with CO/NO ) 5:1 under unpromoted
conditions causes an increase in oxidation state of the Cu
catalyst. The catalyst consists of Cu metal and Cu2O as
determined from the relevant Cu 2p XP and Cu LMM XAES
spectra. Exposing the Cu film under the same initial conditions
as before to the same gas environment and temperature but under
promoted conditions causes a further increase in the oxidation
state of the Cu film. In this case, the catalyst surface consists
of Cu2O and CuO. The same general trends were observed when
the procedure was repeated with gas composition CO/NO )
2:1. In this case, the unpromoted catalyst consists of Cu2O,
whereas under promoted conditions the catalyst surface consists
of Cu2O and CuO, no Cu0 being detectable. With CO/NO )
1:4 the unpromoted surface consists of CuO and a small amount
of Cu2O; the promoted surface consists almost entirely of CuO.
These findings are summarized in Table 1. Quantification was
achieved by estimating the integrated intensity of the Cu0, Cu+,
and Cu2+ components and by assuming that (i) the photoion-
ization cross section of Cu 2p is independent of oxidation state
and (ii) the sampled volume is unaffected by the state of
oxidation of the Cu film. Although these are not gross
approximations, the estimated compositions should be regarded
as no more than indicative.
Figure 6. Cu 2p postreaction XP spectra of the catalyst film after
exposure to the following CO/NO ratios: (i) CO/NO ) 5:1, (ii) CO/
NO ) 2:1, and (iii) CO/NO ) 1:4 for unpromoted and promoted
conditions.
Figures 8 and 9 show the O 1s XP spectra and N 1s XP
spectra, respectively, taken after exposing the film to exactly
the same range of conditions as those pertaining to Figures 6
and 7. It is evident that (i) in all the unpromoted cases the O 1s
signal exhibits only one feature at 530 eV binding energy and
(ii) when sodium is present on the surface, a second feature
appears at 532.5 eV binding energy. The 530 eV may be
attributed with confidence to the oxides of copper25 present on
the surface under both unpromoted and promoted conditions.
No carbonate species were present as judged by the complete
absence of C 1s emission. Since the O 1s peak at 532.5 eV
appeared only when sodium was present on the surface, we
assign it to a sodium compound. We now show that this
compound is NaNO3.
Figure 7. Cu LMM postreaction XAE spectra of the catalyst film after
exposure to the following CO/NO ratios: (i) CO/NO ) 5:1, (ii) CO/
NO ) 2:1, and (iii) CO/NO ) 1:4 for unpromoted and promoted
conditions.
unpromoted and promoted conditions. These results clearly
demonstrate that the value of VWR is directly correlated with
the loading of Na on the catalyst surface under reaction
conditions: decreasing the catalyst potential (VWR ) -400mV)
causes an increase in the amount of Na. It is important to note
that these spectral changes were fully reversible as a function
of catalyst potential.
Analogous Cu 2p XPS and Cu LMM Auger spectra are
shown in Figures 6 and 7, respectively. Several features are
apparent from Figures 6 and 7 and may be summarized as
follows: (1) initially, before use, the surface of the catalyst film
Figure 9 shows relevant N 1s XP spectra; this signal was
detectable at all CO/NO ratios but only after the catalyst had

9964 J. Phys. Chem. B, Vol. 103, No. 45, 1999
Williams et al.
Discussion
Given the lack of sensitivity of the CO + NO reaction toward
the details of surface structure,14-18 it is reasonable to take the
view that our data and those obtained with conventional
dispersed catalysts should be interpretable on the same general
basis.
It is useful to summarize the overall pattern of behavior
illustrated in Figures 1-3. First, regardless of reaction gas
composition and Na loading, sodium promotion always leads
to enhancement of both activity and N2 selectivity. Second, the
Na-induced enhancement in activity is very large under reducing
conditions (F ≈ 20 for CO/NO ) 5:1 and 2:1) and an order of
magnitude smaller under oxidizing conditions (F ) 2.5, CO/
NO ) 1:4). Third, on the clean surface (VWR ) +100 mV) the
nitrogen selectivity decreases with decreasing CO/NO ratio.
Fourth, under the most reducing conditions, there is some
retardation of all rates at the most negative potential (highest
Na loading) for CO/NO ) 5:1 but none when the reaction gas
is more NO-rich.
Origin of Na Promotion and Nature of the Active Sites.
The spectroscopic data show clearly that promotion is associated
with the pumping of Na to the catalysts surface (Figure 5). They
also establish that under reaction conditions (i) this sodium is
present as the nitrate (Figures 8 and 9), (ii) no alkali carbonate
is formed, (iii) no copper nitrate is present, and (iv) the promoted
catalyst surface consists of Cu oxides. In view of (iii) we rule
out participation of Cu nitrate in the overall catalytic process,
in agreement with London and Bell12 and at variance with
Dekker et al.13
Figure 8. O 1s postreaction XP spectra of the catalyst film after
exposure to the following CO/NO ratios at reaction temperature for
unpromoted and promoted conditions: (i) CO/NO ) 5:1, (ii) CO/NO
) 2:1, and (iii) CO/NO ) 1:4.
In our earlier work on NO reduction by CO20 or propene29
over platinum, we proposed that EP results from acceleration
of the key reaction-initiating step: dissociation of adsorbed NO.
In the case of a pure metal surface, we argued as follows. As
the catalyst potential is decreased (pumping Na to the surface),
the electronic effect of the sodium promoter on adsorbed NO
strengthens the metal-N bond (increasing NO coverage) and
weakens the N-O bond (facilitating NO dissociation). These
effects are well understood and have been discussed in detail
by Lang et al.30 They are caused by the electric field of the
alkali cation, which lowers the energy of the π* orbital in the
adjacent diatomic molecule. Increased electron density in the
π* orbital (i) strengthens the metal-N bond, thus increasing
NO coverage, and (ii) facilitates dissociation of the N-O bond.
These effects operate together to produce large increases in both
activity and selectivity as a result of accelerating reaction 1 and
inhibiting reaction 4:
Figure 9. N 1s postreaction XP spectra of the catalyst film after
exposure to the following CO/NO ratios at reaction temperature for
unpromoted and promoted conditions: (i) CO/NO ) 5:1, (ii) CO/NO
) 2:1, and (iii) CO/NO ) 1:4.
NOa f Na + Oa
Na + Na f N2
COa + Oa f CO2
NOa + Na f N2O inhibited by Na
promoted by Na
(1)
(2)
(3)
(4)
been run under promoted conditions. The binding energy (407
eV) corresponds to sodium nitrate.26 Moreover, in any given
case, the ratio of the intensity of the O 1s higher binding energy
component to that of the corresponding N 1s emission was 2.86
(after normalization to account for photoionization cross sec-
tion27). This provides strong quantitative support for the view
that under promoted conditions sodium is present on the surface
as NaNO3. The binding energies observed in the Cu 2p spectra
rule out the presence of copper nitrate28 under all conditions.
Heating the promoted catalyst in vacuum to 483 K causes the
disappearance of both the higher binding energy O 1s signal
and the N 1s peak, commensurate with the thermal stability of
sodium nitrate.
It is worth noting that guided by these earlier EP results,20,29
we synthesized conventional dispersed Pt/alumina catalysts that,
when optimally promoted by Na, exhibited very large enhance-
ments of both activity and selectivity for NO reduction by
propene31 and under simulated exhaust gas conditions.32
What happens in our case, which is complicated by the effects
of Cu metal oxidation? There is little agreement in the literature
regarding the question of which types of Cu sites are involved
in catalytic turnover and which types of site adsorb NO, CO,
or other possible adsorbates under reaction conditions. For

CO + NO Reaction
J. Phys. Chem. B, Vol. 103, No. 45, 1999 9965
example, on the basis of their transient isotope reactor measure-
ments, Dekker et al.13 concluded that strongly bound NO is
associated with Cu+ sites. This is at variance with the proposal
by Fu et al.,33 based on low-temperature IR data, that CO and
NO adsorb on both Cu+ and Cu2+, with NO adsorbed somewhat
more strongly on Cu2+ and CO adsorbed somewhat more
strongly on Cu+. London and Bell12 invoke Cu0, Cu+, and Cu2+
in the process of catalytic turnover; most other authors consider
that only Cu+ and Cu2+ sites are involved. Our results cannot
resolve all these issues, but they do provide answers to some
of them. As argued below, they demonstrate that Na promotion
of NO dissociation is pivotal and accounts for the observed
effects of EP on activity and selectivity in a natural way. In
addition, they clearly demonstrate that the chemical composition
of the active surface is strongly dependent on the composition
of the reaction gas mixture, an issue that previously has not
been investigated systematically. We therefore suggest that in
the present case the reaction mechanism and mode of promotion
are analogous to those proposed to explain the behavior of Pt
catalysts undergoing EP of NO reduction.20,29
that this quantity should be lowest on Cu+, consistent with our
model. Thus, the observation38 of a substantially lower N-O
stretching frequency (1765 cm-1) for Cu+-NO than for Cu2+
-
NO (1880 cm-1) points to a reduced activation energy for NO
dissociation in the former case. However, we cannot lose sight
of the fact that the measured34 and calculated35,37 adsorption
energies of NO on Cu2+ differ by a factor of 4. Clearly, further
experimental and theoretical work is necessary in order to
reconcile this apparent divergence.
Note that our results do not permit any conclusions to be
drawn about the adsorption sites for CO. In passing, we note
that an additional role of Na is that it catalyzes the oxidation of
Cu by NO, thus accelerating formation of the active oxidized
surface. Therefore, on balance, we agree with Fernandez-Garcia
et al.19 that the process of catalytic turnover may be written as
follows:
The postreaction XP and Auger spectra (Figures 6 and 7)
indicate that the promoted surface does not contain detectable
Cu0. It consists of a mixture of Cu2+ and Cu+ sites; the number
density of the latter is relatively small under strongly oxidizing
conditions, being detectable only in the Auger spectra. This is
in line with the observations of Fu et al.33 whose IR/XPS data
obtained with CuO/alumina catalysts indicated the simultaneous
presence of Cu2+ and Cu+ at the surface. We therefore propose
that the reaction is triggered by the Na-induced dissociation of
NO adsorbed on Cu+. Cu+ rather than Cu2+ is suggested as the
principal site for NO activation on the grounds of its finite 3d
electron density. This d electron density should favor NO
dissociation on Cu+ (relative to Cu2+) and thus should enhance
the effect of Na promotion. This view is supported by the
observation that the effects of Na promotion on both selectivity
and activity are most pronounced when the Cu+ content of the
clean (unpromoted) surface are highest (Figures 1-3, 6, 7 and
Table 1). It is also in accord with the findings of Fernandez-
Garcia et al.,19 whose XANES data led them to conclude that
NO adsorption and dissociation at Cu+ sites is the rate-limiting
step for the unpromoted reaction. Formally, we may write
Figures 1-3 show that the nitrogen selectivity of the
unpromoted (VWR ) +100 mV) catalyst decreases while
decreasing the CO/NO ratio. This may be rationalized as
follows. The mode of NO chemisorption on Cu0, Cu+, and Cu2+
is linear bonding through the nitrogen atom. The conventional
view of this process is the donation/back model involving the
π* level of NO and d orbitals of the metal. Net electron transfer
may thus vary from NOδ+-Mδ- to NOδ--Mδ+. On this basis,
and in agreement with observation,14-17 one expects the efficacy
of Cu species for NO dissociation to lie in the order Cu0 >
Cu+ > Cu2+. The spectroscopic data show that decreasing the
CO/NO ratio causes an increase in overall oxidation state of
the copper. This in turn should result in a decrease in NO
dissociation and hence a decrease in nitrogen selectivity.
Why Is There No Strong Poisoning at Negative VWR? The
behavior observed here bears interesting similarities to and
differences from that observed for Pt20 and Rh39 surfaces
undergoing EP by Na. In all three cases, there are large
increments in both activity and selectivity under the influence
of electropumped Na. Also, in all three cases, there is no
tendency toward strong poisoning at the most negative values
of catalyst potential that could be achieved. In this latter respect,
the CO + NO reaction differs markedly from the propene +
NO reaction on Pt29 and Rh40 and from propene combustion on
Pt41 under EP by Na. In these three cases, a region of strong
promotion is followed by a regime of strong poisoning as the
catalyst potential is decreased; small amounts of Na cause
promotion and sufficiently large amounts of Na result in
poisoning. Our XP spectra for CO + NO/Cu provide a basis
for understanding this previously unexplained difference in
behavior. The strongly poisoned systems result from overloading
the catalyst with Na while propene is undergoing oxidation either
by NO or by O2. XP and XANES data show very clearly that,
under such conditions, oxidation of this carbon-rich molecule
leads to formation of thick films of Na carbonate that cover the
surface and are stable at reaction temperature. This is consistent
with the thermal properties of Na carbonate (melting point 1124
K). On the other hand, the present results clearly show that when
Cu+-NO + Cu+ f Cu+-N + Cu2+ + O-
or
Cu+-NO + 2Cu+ f Cu+-N + 2Cu2+ + O2-
The postreaction O 1s XP spectra favor the second possibility
in that they show the presence of only O2-. However, given
the limitations of the method, O- cannot be entirely ruled out.
The identity of the NO activating site(s) deserves further
comment. Shelef and Gandhi have shown34 that NO adsorbs
faster at Cu2+ sites than it does at Cu+ sites and that the
adsorption enthalpy in the former case is a strongly decreasing
function of coverage. Preferential adsorption of NO at Cu2+
sites may be rationalized in terms of the unpaired spins on
adsorbate and adsorbent.34 Our suggestion is based on adsorbed
NO acting as a σ donor and π acceptor. Recent calculations35,36
show that NO does indeed bond to Cu+ by a σ/π mechanism
and that the strength of the bond (∼200 kJ) is not greatly
different from that of the Cu2+-NO adsorption bond (∼260
kJ).37 With respect to mechanism, a key issue is the activation
enery to dissociation of NO on the two types of Cu site. The
weight of theoretical30 and experimental evidence33,38 suggests

9966 J. Phys. Chem. B, Vol. 103, No. 45, 1999
Williams et al.
(4) Shelef, M. Catal. ReV. Sci. Eng. 1975, 11, 1.
CO is oxidized by NO (a more facile reaction than propene
oxidation), there is no formation of Na carbonate. Instead, the
promoter phase consists exclusively of NaNO3. Thick films of
NaNO3 (i.e., possessing the thermal properties of bulk sodium
nitrate) would become volatile at ∼580 K, decomposing
altogether at ∼650 K, which is the reaction temperature used
in the present work. Submonolayer films of NaNO3, however,
would be expected to be more stable than the bulk material
because of their strong interaction with the metal’s surface. The
very strong stabilization of two-dimensional films of ionic
compounds adsorbed on metal surfaces relative to their three-
dimensional analogues has been demonstrated42 by experiment.
Thus, in the present case, submonolayer quantities of NaNO3
survive at reaction temperature, promoting the system. Thick
films cannot be built up because of their volatility. As a result,
strong poisoning does not occur. Note that there is some
retardation (∼10%) of the reaction rate at the highest Na
loadings in CO-rich gas (CO/NO ) 5:1). This may reflect onset
of the formation of a more stable surface compound under these
conditions. A possible cause for this retardation could be
formation of a stable alkali-CO surface complex of the type
detected on a number of different surfaces.43
(5) Parvulescu, V. I.; Grange P.; Delmon, B. Catal. Today 1998, 46,
233.
(6) Hildebrand, H. H.; Lintz, H.-G. Appl. Catal. 1990, 65, 241.
(7) Hildebrand, H. H.; Lintz, H.-G. Catal. Today 1991, 9, 153.
(8) Hildebrand, H. H.; Lintz, H.-G. Ber. Bunsen-Ges. Phys. Chem.
1991, 95, 1191.
(9) Bruck, J.; Brust, M.; Lintz, H.-G. Ber. Bunsen-Ges. Phys. Chem.
1995, 99, 1509.
(10) Estenfelder, M.; Lintz, H.-G. Stud. Surf. Sci. Catal. 1997, 110, 981.
(11) Jernigan, G. G.; Somorjai, G. A. J. Catal. 1994, 147, 567.
(12) London, J. W.; Bell, A. T. J. Catal. 1973, 31, 96.
(13) Dekker, F. H. M.; Kraneveld, S.; Bliek, A.; Kapteijn, F.; Molelijn,
J. A. J. Catal. 1997, 170, 168.
(14) Balkenende, A. R.; Gijzeman, O. L. J.; Geus, J. W. Appl. Surf.
Sci. 1989, 37, 189.
(15) Balkenende, A. R.; den Daas, H.; Huisman, M.; Gijzeman, O. L.
J.; Geus, J. W. Appl. Surf. Sci. 1991, 47, 341.
(16) Balkenende, A. R.; Hoogendam, R.; de Beer, T.; Gijzeman, O. L.
J.; Geus, J. W. Appl. Surf. Sci. 1992, 55, 1.
(17) Balkenende, A. R.; van der Gift, C. J. G.; Meulenkamp, E. A.;
Geus J. W. Appl. Surf. Sci. 1993, 68, 161.
(18) Wee, A. T. S.; Lin, J.; Huan, A. C. H.; Loh, F. C.; Tan, K. L. Surf.
Sci. 1994, 304, 145.
(19) Fernandez-Garcia, M.; Marquez Alvarez, C.; Rodriguez-Ramos, I.;
Guerrero-Ruiz, A.; Haller, G. L. J. Phys. Chem. 1995, 99, 16380.
(20) Palermo, A.; Lambert, R. M.; Harkness, I. R.; Yentekakis, I. V.;
Marina, O.; Vayenas, C. G. J. Catal. 1996, 161, 471.
(21) Tracey, S.; Palermo, A.; Holgado Vazquez, J. P.; Lambert, R. M.
J. Catal. 1998, 179, 231.
(22) Yentekakis, I. V.; Bebelis, S. J. Catal. 1992, 137, 278.
(23) Poulston, S.; Parlett, P. M.; Stone, P.; Bowker, M. Surf. Interface
Anal. 1996, 24, 811.
(24) Leiro, J. A.; Heinonen, M. H.; Werfel, F.; Nordstrom, E. G.;
Karlsson, K. H. Philos. Mag. Lett. 1993, 68, 153.
(25) Parmigiani, F.; Pachioni, G.; Illas, F.; Bagus, P. S. J. Electron
Spectrosc. Relat. Phenom. 1992, 59, 255.
Conclusions
(1) In the catalytic reduction of NO by CO, the reactive
behavior, surface composition, and response to electrochemical
promotion of Cu film catalysts are a strong function of the
composition of the reactant gas, consistent with a redox
mechanism.
(26) Allen, H. C.; Laux, J. M.; Vogt, R.; Finlayson-Pitts, B. J.;
Hemminger, J. C. J. Phys. Chem. 1996, 100, 6371.
(27) Yeh, J. J.; Lindau, I. At. Data Nucl. Data Tables 1985, 32, 1.
(28) Wagner, C. D. In Practical Surface Analysis by Auger and X-ray
Photoelectron Spectroscopy; Briggs, D., Seah, M. P., Eds.; J.Wiley and
Sons: New York, 1983.
(29) Yentekakis, I. V.; Palermo, A.; Filkin, N. C.; Tikhov, M. S.;
Lambert, R. M. J. Phys. Chem B 1997, 101, 3759.
(30) Lang, N. D.; Holloway, S.; Norskov, J. K. Surf. Sci. 1985, 150,
24.
(2) Under reducing or fairly oxidizing conditions the Cu-
catalyzed CO + NO reaction exhibits strong electrochemical
promotion of both activity and selectivity. Under strongly
oxidizing conditions, promotion is still observed, but the effects
are much less pronounced.
(3) The spectroscopic data indicate that these effects are due
to pumping of Na to the catalyst where, under reaction
conditions, it is present as NaNO3 on an oxidized Cu surface.
Na2CO3 and Cu nitrate are not detectable.
(31) Konsolakis, M.; Nalbantian, L.; McLeod, N.; Yentekakis, I. V.;
Lambert, R. M. Appl. Catal. B, in press.
(4) The spectroscopic and reactor data indicate that Cu0 sites
are not of catalytic significance. In this regard, the weight of
evidence suggests that Cu+ is the principal site for NO
adsorption and dissociation. EP is due to Na-induced enhance-
ment of the adsorption and dissociation of NO at these sites.
(32) Macleod, N.; Isaac, J.; Konsolakis, M.; Yentekakis, I. V.; Lambert,
R. M. Manuscript in preparation.
(33) Fu, Y.; Tian, Y.; Lin, P. J. Catal. 1991, 132, 85.
(34) Gandhi, H. S.; Shelef, M. J. Catal. 1973, 28, 1.
(35) Casarin, M.; Maccato, C.; Vittadini, A. Chem. Phys. Lett. 1997,
280, 53.
(36) Casarin, M.; Vittadini, A. Surf. Sci. 1997, 387, L1079.
(37) Duan, Y. H.; Zhang, K. M.; Xie, X. D. Surf. Sci. 1994, 321, L249.
(38) Sepu´lveda, A.; Ma´rquez, C.; Rodriguez-Ramos, I.; Guerrero-Ruiz,
A.; Fierro, J. L. Surf. Interface Anal. 1993, 20, 1067.
(39) Williams, F. J.; Palermo, A.; Tikhov, M. S.; Lambert R. M.
Manuscript in preparation.
Acknowledgment. F.J.W. acknowledges the award of a
scholarship by Fundacio´n YPF. Financial support from the U.K.
Engineering and Physical Sciences Research Council and from
the European Union is gratefully acknowledged under Grants
GR/M76706 and BRPR-CT97-0460, respectively.
(40) Williams, F. J.; Palermo, A.; Tikhov, M. S.; Lambert, R. M.
Manuscript in preparation.
(41) Filkin, N. C.; Tikhov, M. S.; Palermo, A.; Lambert, R. M. J. Phys.
Chem. A 1999, 103, 2680.
References and Notes
(42) Lambert, R. M.; Kitson, M. Surf. Sci. 1981, 110, 205.
(43) Bertolini, J. C.; Delichere, P.; Massardier, J. Surf. Sci. 1985, 160,
331.
(1) Taylor, K. Catal. ReV. Sci. Eng. 1993, 35, 457.
(2) Shelef, M.; Otto, K. J. Catal. 1968, 10, 408.
(3) Shelef, M.; Otto, K.; Gandhi, H. J. Catal. 1968, 12, 361.