Electrocatalysis at conductive diamond modified by noble-metal oxides

Article abstract of DOI:10.1021/jp003216w

In the present work, a generalization of some of the interpretations of heterogeneous catalysis to the case of oxide film electrocatalysts has been attempted, using conductive boron-doped diamond supports. The very high stability of this material allowed the study of the catalytic properties of noble-metal oxide modifications from the level of submonolayer to a few monolayers, avoiding the serious interference of the reactivity of the generally adopted metal supports. A mechanism for the chlorine evolution reaction at submonolayers of oxide electrocatalyst has been proposed, basing on a chlorine radical spillover stage.

Full text of DOI:10.1021/jp003216w

© Copyright 2001 by the American Chemical Society
VOLUME 105, NUMBER 9, MARCH 8, 2001
LETTERS
Electrocatalysis at Conductive Diamond Modified by Noble-Metal Oxides
Achille De Battisti,* Sergio Ferro, and Maurizio Dal Colle
Department of Chemistry, UniVersity of Ferrara, Via L. Borsari 46, I-44100 Ferrara, Italy
ReceiVed: September 13, 2000; In Final Form: January 17, 2001
In the present work, a generalization of some of the interpretations of heterogeneous catalysis to the case of
oxide film electrocatalysts has been attempted, using conductive boron-doped diamond supports. The very
high stability of this material allowed the study of the catalytic properties of noble-metal oxide modifications
from the level of submonolayer to a few monolayers, avoiding the serious interference of the reactivity of the
generally adopted metal supports. A mechanism for the chlorine evolution reaction at submonolayers of oxide
electrocatalyst has been proposed, basing on a chlorine radical spillover stage.
The use of supported catalyst films is common practice in
heterogeneous catalysis, their properties essentially depending
on their chemical nature and their surface texture. Modified
electrodes, frequently used in electrochemical experiments,
represent a natural development of this conception in the field
of electrocatalysis.1 A wide variety of modifications have been
studied, including polymeric, metallic, and metal oxide films,
with amounts of deposits varying between submonolayers of
adatoms to multilayers, up to several hundreds of nanometer
thick films. Modifications have been inserted at the surface of
different supports, from carbon to different metal single crystals.
Although the thickness of the modifying film is generally
sufficiently high, still the interaction with the support may affect
its catalytic activity. The strong metal-support interaction
described in the particular case of supported metallic films is a
typical example of this class of phenomena.2
to take place.3 Under these conditions, the synthesis exhibits a
poor reproducibility4,5 and the metal support, generally a valve
metal, is partially oxidized.6 Typically, thin films allowing a
better understanding of charging mechanisms cannot be syn-
thesized at usual metal supports such as titanium or tantalum,
because of their massive involvement in the oxidation pro-
cesses: a limiting case of strong catalyst-support interaction.
The “model” modifying film in these cases is not available,
missing concrete possibilities of reaching better-defined systems.
The recent achievements in the preparation of highly boron-
doped diamond (BDD) films with minimal sp2 carbon contents
seem to supply an interesting solution to the above problems.
Supported BDD films exhibit very high stability upon different
chemical and electrochemical treatments.7 They can be heated
under oxygen up to about 500 °C, undergoing only partial
oxidation of surface carbon atoms. Very high stability has been
found also under far-positive electrochemical polarization.8
These features make BDD interesting not only as an electrode
film but also as an inert support for electrode films of a different
nature. In the present work, the electrochemical properties of
deposits of small amounts of RuO2 and IrO2 have been studied,
exploiting the absence of catalytic activity of Si/BDD besides
the above-mentioned high stability. Amounts of oxide electro-
catalyst were between 1.2E+13 and 2.65E+16 MO2 molecules/
cm2, corresponding to a nominal oxide loading between 0.01
The influence of the support material on the catalytic prop-
erties of the film, through changes of its surface composition,
is further enhanced when the deposit formation takes place under
drastic conditions, and the support itself is involved in the
precursor reaction. Oxide film electrodes for anodic electrosyn-
thesis, consisting of IrO2 or RuO2, are a typical example in this
sense. Their preparation involves several steps leading from
precursor salts to the respective oxides, and temperatures on
the order of 350-450 °C are required for the transformations
10.1021/jp003216w CCC: $20.00 © 2001 American Chemical Society
Published on Web 02/10/2001

1680 J. Phys. Chem. B, Vol. 105, No. 9, 2001
Letters
Figure 2. Three-dimensional array of distorted octahedra in the rutile
structure. View along the 001 direction.
at RuO2 and IrO2 thick-film electrodes exhibit poorly resolved
peaks, which convolute into broad flat maxima.3,11
The latter feature has possibly justified an overestimation of
the purely capacitive contribution to the charge-storage capacity
of oxide electrodes. Data in Figure 1 indicate that the comple-
mentary faradaic contribution could be the most important one.
Equilibria of the type:3,12
Figure 1. Cyclic voltammograms of two RuO2-modified BDD
electrodes, 1.2E+13 and 6.0E+14 molecules/cm2 (A), and of a BDD/
IrO2 electrode, 6.0E+14 molecules/cm2 (B), recorded in 1 N H2SO4 at
a sweep rate of 0.3 V s-1
.
MOx + H+ + e h MOx-1OH
and 10 monolayers. The properties of the BDD support and of
modified surfaces were studied in situ by cyclic voltammetry
in 1 N sulfuric acid. The ex situ characterization was carried
out by X-ray photoelectron spectroscopy (XPS), scanning
electron microscopy (SEM), and atomic force microscopy
(AFM). The effect of the modifications on the catalytic activity
was tested with the chlorine evolution reaction, making use of
quasi-steady potentiostatic curves.
BDD films of 1 µm thickness were grown on low-resistivity
(1 mΩ cm) Si wafers, according to an elsewhere described
procedure.9 RuO2 and IrO2 modifications were deposited by a
traditional sol-gel method,10 pyrolyzing the precursor salts at
400 °C for 30 h. Blank experiments with Si/BDD electrodes
showed that the above thermal treatment does not modify the
electrochemistry of the support.
In all cases, the preparation led to stable modifications,
unaffected by anodic polarization under chlorine evolution.
Cyclic voltammograms obtained at as-prepared electrodes
reached a stable shape after a limited number of cycles. Typical
CV data are shown in Figure 1. The main features of the
voltammogram (Figure 1A) are already present for the lowest
oxide deposit. They remain unchanged at electrodes with higher
oxide loading, whose only effect is the increase of voltammetric
current. Different solid-state redox couples between 0.00 and
1.25 V (vs SCE) for the RuO2-H2O and IrO2-H2O systems
are clearly identified. As shown in Figure 1, two well-resolved
peak pairs are observed, possibly related to the same oxidation
state transitions (MII-MIII and MIII-MIV). The voltammogram
in Figure 1B, relative to the IrO2 film, also exhibits weaker
signals due to higher oxidation state transitions. The more
positive one, partially superimposed with the oxygen evolution
onset, could be associated with the IrV-IrVI transition. The
existence of the above redox couples has been discussed in the
literature, in relation with the mechanism of oxygen evolution
reaction.3 The present results are a clear-cut in situ evidence of
distinct oxidation state changes, as the cyclic voltammograms
involving two or more oxidation state changes can better account
for the very high capacities measured, the proton exchange being
much faster than double-layer rearrangements caused by changes
of electrode potential.
The three-dimensional array of distorted RuO6 octaedra, on
which the rutile structure of RuO2 is based, presents void
channels normal to the 001 plane (Figure 2). As the inner walls
of these channels are oxygen-based, the above electrochemical
ion (proton)-exchange equilibria may extend several Ångstroms
below the oxide surface through, e.g., a Grotthus-type mecha-
nism.13 The intergrain-interparticle boundaries, often considered
the reason for the large capacities of the RuO2 films, would
rather magnify an intrinsic property of the structure of this oxide.
Voltammetric (anodic) charge relative to the lowest oxide
loading, 1.2E+13, indicates that 2.28 electrons are exchanged
per oxide molecule. Assuming an electronicity of 2 for the global
charging process, this is in agreement with the expectation that,
at very low coverage, most of the noble-metal sites can undergo
oxidation state changes. With increasing the oxide loading to
6.0E+14, which approximately represents one oxide monolayer,
the voltammetric charge increases up to only 6.77 µC, almost
twice the previous value (3.44 µC). The efficiency of the charge
storage is considerably decreased, possibly due to accumulation
of oxide in the deeper parts of the polycrystalline diamond
surface, which can incorporate even higher amounts of oxide
without exposing larger external surfaces. The coating containing
2.65E+16 molecules/cm2 exhibits an anodic voltammetric
charge of 50.62 µC, which is still lower than 224.34 µC, the
value calculated for a complete monolayer of RuO2 molecules
(7E+14) exchanging two electrons per molecule. This low
charging yield, at a relatively high loading, reflects both the
previously mentioned effect, related with the diamond film
surface, and the building up of the three-dimensional oxide phase
through clusters of molecules. The latter can be clearly seen in

Letters
J. Phys. Chem. B, Vol. 105, No. 9, 2001 1681
Figure 4. Current potential curves for chlorine evolution in 1 M NaCl/3
M NaClO4/0.01 M HClO4. Current values are normalized to the number
of surface sites.
Figure 3. AFM image of a BDD/RuO2 sample, with a nominal deposit
of 2.65E+16 molecules/cm2.
the AFM image in Figure 3. A simple calculation, based on the
average number of the approximately spherical particles and
their size, leads to an estimated charge of 24.90 µC, assuming
that only the first layer of RuO2 molecules can undergo oxidation
state changes at the clusters surface. This number is about one-
half the experimental one. In consideration of the presence of
channels in the rutile structure, this disagreement may be due
to the restriction of the charge storage to the surface sites only.
Within the same frame, the larger voltammetric area for the
IrO2 loading (Figure 1B) can be justified by the more acidic
character of this oxide, compared with RuO2, with higher proton
mobility expected in this case.
Figure 5. Tafel plots for chlorine evolution in 1 M NaCl/3 M NaClO4/
0.01 M HClO4. Current values are normalized to the number of surface
sites.
To ascertain the stoichiometry of the deposited oxide, XPS
measurements have been carried out for the ruthenium-based
samples. The signals relative to Ru3d5/2, Ru3p3/2, and O1s have
been used for this part of the investigation. The binding energy
of the first signal (280.6 eV) and the ratio between the other
two (2.0 ( 0.2) are evidence for the formation of a RuO2 phase
in all the investigated samples. Interestingly, the rate of increase
of the Ru signals with the amount of oxide deposited is quite
small and does not follow a linear law, supporting the hypothesis
formulated on the basis of the charge-storage results.
As above-mentioned, the electrochemical oxidation of chlo-
ride ions to chlorine was used as a model anodic reaction, to
trace the effect of RuO2 modifications of BDD films. A set of
polarization curves has been obtained in x M NaCl/(4 - x) M
NaClO4/0.01 M HClO4, with 0.1 e x e 4.
The results, in terms of polarization curves and Tafel plots,
for x ) 1, are shown in Figures 4 and 5, respectively. As the
extension of the working surface is different in the three cases
taken into consideration, current values have been normalized
to the number of surface atoms at which the reaction takes place.
In the case of BDD film, the number of C atoms/cm2 has been
used, assuming in a first approximation a unitary roughness
factor. The number of active sites for the two modified surfaces
depends on the oxide dispersion. Basing on previously discussed
CV data, for the smaller loading the nominal number of RuO2
molecules/cm2 of 1.2E+13 has been chosen. Making use of the
latter and of the ratio between the voltammetric charges found
for the two modified electrodes, the estimate for the higher
loading was 1.77E+14. The slope of the Tafel plot for the BDD
electrode is about 0.150 V, giving evidence for a rate-deter-
mining discharge of chloride ions.14,15 At the sample with the
highest oxide coating, the chlorine evolution reaction proceeds
with a much higher rate, as expected for a RuO2 electrode. The
Tafel slope is about 0.150 V, with a reaction order of 0.7 with
respect to Cl-. A Volmer-Heyrovsky mechanism14 with a rate-
determining electrochemical desorption can account for the
experimental evidence, provided a value of about 0.7 is assumed
for the coverage by the intermediate chlorine radicals.16 As a
further support to this mechanistic hypothesis, the rate of the
chlorine evolution reaction was independent from the solution
acidity.
The parameters of the chlorine evolution reaction at the
sample with the lowest catalyst loading are less easily interpret-
able. The catalytic activity of the modified electrode is quite
higher, compared with the unmodified one, which is in
agreement with the measurable presence of ruthenium oxide at
the electrode surface. However the Tafel slope is now 0.060 V,
and the acidity affects the chlorine evolution reaction, the
reaction order with respect to H+ being -0.6.
In this case, AFM does not show any particle, indicating that
oxide clusters should have a diameter not larger than ca. 9-10
nm. The very small size of the catalyst patches, at which the
chloride discharge takes place, may possibly favor the subse-
quent surface diffusion of the chlorine radical from the formation
site on the catalyst surface to a neighboring C(diamond) site.
The electrochemical desorption step can easily take place at
the latter, due to the poor interaction with the radical. As the
C(diamond) sites can be largely hydroxylated under the given
thermal and polarization conditions, the influence of the solution
acidity on the chlorine evolution rate could also be accounted
for along the above lines.

1682 J. Phys. Chem. B, Vol. 105, No. 9, 2001
Letters
effects and interpretations have been described20 in heteroge-
neous catalysis at supported RuO2.
The above-emphasized interest of conductive diamond sup-
ports in the fundamental investigation of oxide electrocatalysts
cannot overshadow the importance of stable supports in film
electrodes for industrial oxidative processes.
Acknowledgment. The authors gratefully acknowledge the
contribution of Prof. Ch. Comninellis for prompting the research
and for helpful discussion.
References and Notes
(1) For example: Gellings, P. J.; Bouwmeester, H. J. M. Catal. Today
2000, 58, 1.
(2) For example: Stakheev, A. Yu.; Kustov, L. M. Appl. Catal. A 1999,
188, 3.
(3) Trasatti, S.; Lodi, G. In Electrodes of ConductiVe Metallic Oxides,
Part A; Trasatti, S., Ed.; Elsevier: Amsterdam, 1980; pp 315-317.
(4) De Battisti, A.; Nanni, L.; Battaglin, G.; Comninellis, Ch. In New
Promising Electrochemical Systems for Rechargeable Batteries; Barsukov,
V., Beck, F., Eds.; NATO ASI Series; Kluwer: Dordrecht, 1996; pp 197-
211.
(5) Lodi, G.; De Battisti, A.; Bordin, G.; De Asmundis, C.; Benedetti,
A. J. Electroanal. Chem. 1990, 277, 139.
(6) Vercesi, G. P.; Rolewicz, J.; Comninellis, Ch.; Plattner, E.; Hinden,
J. Thermochim. Acta 1991, 176, 31.
(7) Carey, J. J.; Christ C. S., Jr.; Lowery, S. N.; U.S. Patent 5 399
247, 1995; Chem. Abstr. 1995, 123, 92375k.
(8) Michaud, P.-A.; Mahe´, E.; Haenni, W.; Perret, A.; Comninellis,
Ch. Electrochem. Solid-State Lett. 2000, 3, 77.
(9) Haenni, W.; Baumann, H.; Comninellis, Ch.; Gandini, D.; Nied-
ermann, P.; Perret, A.; Skinner, N. Diamond Relat. Mater. 1998, 7, 569.
(10) Guglielmi, M.; Colombo, P.; Rigato, V.; Battaglin, G.; Boscolo
Boscoletto, A.; De Battisti, A. J. Electrochem. Soc. 1993, 139, 1655.
(11) Benedetti, A.; Riello, P.; Battaglin, G.; De Battisti, A.; Barbieri,
A. J. Electroanal. Chem. 1994, 376, 195.
(12) Bock, C.; Birss, V. I. J. Electroanal. Chem. 1999, 475, 20 and
literature therein.
(13) Bockris, J. O’M.; Reddy, A. K. N. Modern Electrochemistry;
Plenum/Rosetta: New York, 1973; pp 474-476.
(14) Trasatti, S. Electrochim. Acta 1987, 32, 369.
(15) Ferro, S.; De Battisti, A.; Duo, I.; Comninellis, Ch.; Haenni, W.;
Perret, A. J. Electrochem. Soc. 2000, 147, 2614.
(16) Conway, B. E.; Tilak, B. V. Electrochim. Acta 1992, 37, 51.
(17) Weng, L. T.; Ruiz, P.; Delmon, B. In Studies in Surface Science
and Catalysis, New DeVelopment in SelectiVe Oxidation by Heterogeneous
Catalysis; Ruiz, P., Delmon, B., Eds.; Elsevier: Amsterdam, 1992; Vol.
72, pp 399-413.
(18) Aramata, A.; Toyoshima, I.; Enyo, M. Electrochim. Acta 1992, 37,
1317.
(19) Tseung, A. C. C.; Dhara, S. C. Electrochim. Acta 1974, 19, 845.
(20) Li, D.; Ichikuni, N.; Shimazu, S.; Uematsu, T. Appl. Catal. A 1999,
180, 227.
Figure 6. Schematic representation of the reaction pattern for the
chlorine evolution reaction at BDD/RuO2 electrodes. At low RuO2
coverage, radical spillover can take place.
The above reaction scheme could be seen as an example of
radical spillover, well-known in heterogeneous catalysis of gas-
phase reactions17 and also hypothesized in very specific cases
of electrochemical oxidation reactions.18,19 The pronounced
chemical and electrochemical stability of BDD films allows the
deposition of submonolayers of donor phases (e.g., RuO2, IrO2,
Pt) and obtainment of well-defined bifunctional surfaces where
the C(diamond) phase acts as an acceptor of adsorbed radicals
formed at the oxide surface.
Contrary to the expectations, the normalized values of current
at the two RuO2 coatings do not coincide. The decrease of
activity, with decreasing the oxide loading, can be understood
only assuming a strong diamond-ruthenium oxide interaction,
more important for thinner deposits. Apparently, the catalytic
activity at the oxide surface, requiring a rearrangement of
oxidation state of the metal ions in the oxide lattice, is enhanced
by the existence of underlying oxide layers, where such re-
arrangements can take place as well. On the basis of this
interpretation, the chlorine evolution reaction at the electrode
with the lowest oxide loading can be tentatively described by
Figure 6, which takes into consideration the two possible routes
of the chlorine radical consumption. The assumption of interac-
tion of adsorbed radicals with higher oxidation states of the
metal ion at the surface would necessarily imply the participation
of the whole near-surface region of the oxide film. Similar