Electrocatalytic reduction of H2O2 at P2Mo18O626- modified glassy carbon

Article abstract of DOI:10.1016/S0013-4686(99)00399-0

Catalysis of the electroreduction of H₂O₂ by the polyoxometalate P₂Mo₁₈O₆₂^6- is described. The presence of polyoxometalate in solution leads to electrocatalytic currents at potentials as positive as +0.35 V versus Ag/AgCl. Modification of the surface of a glassy carbon electrode by adsorption of P₂Mo₁₈O₆₂^6- results in the formation of an anionic monolayer showing catalytic redox activity at the same unusual potential. The currents of H₂O₂ electroreduction can be greatly enhanced by coating the electrode with multiple layers of polyoxometalate and a charge balancing cation. A possible catalytic mechanism is proposed.

Full text of DOI:10.1016/S0013-4686(99)00399-0

Electrochimica Acta 45 (2000) 1829–1836
www.elsevier.nl/locate/electacta
6−
62
Electrocatalytic reduction of H2O2 at P2Mo18O
glassy carbon
modified
D. Martel, A. Kuhn *
Laboratoire d’Analyse Chimique par Reconnaissance Mole´culaire, Ecole Nationale Supe´rieure de Chimie et de Physique Bordeaux,
B.P. 108, 33402 Talence, France
Received 26 July 1999; received in revised form 15 September 1999
Abstract
6−
62
Catalysis of the electroreduction of H2O2 by the polyoxometalate P2Mo18O
is described. The presence of
polyoxometalate in solution leads to electrocatalytic currents at potentials as positive as +0.35 V versus Ag/AgCl.
6−
62
Modification of the surface of a glassy carbon electrode by adsorption of P2Mo18O
results in the formation of an
anionic monolayer showing catalytic redox activity at the same unusual potential. The currents of H2O2 electroreduc-
tion can be greatly enhanced by coating the electrode with multiple layers of polyoxometalate and a charge balancing
cation. A possible catalytic mechanism is proposed. © 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Modified electrodes; Electrocatalysis; Polyoxometalate; Monolayer; Multilayer
POM have been used for the activation of both,
electrooxidation and electroreduction reactions [5]. Ex-
amples concerning the latter case are catalysis of the
hydrogen evolution reaction [2], the reduction of bro-
mate [6,7], nitrite [8] and hydrogen peroxide [9]. The
reactive site for H2O2 and NO−2 reduction seemed to be
an iron center present in the catalyst molecule, but
recently the electrocatalytic reduction of NO−2 has also
been reported for very simple, unsubstituted Keggin
and Dawson ions [7,10–12]. Because the improvement
of H2O2 electroactivity is of general interest for applica-
tions such as biosensors and fuel cells we became
interested in the question whether the electroreduction
of H2O2 could eventually also be activated by POM
containing no transition metal substituents. The answer
is not obvious because it is known from organic synthe-
sis that POM, used to catalyze the oxidation of organic
molecules by hydrogen peroxide are preferentially
mixed addenda species [13–16]. However a few exam-
ples exist where unsubstituted POM have been success-
fully used instead [17,18], and this made us believe that
they should also be able to activate the electroreduction
of H2O2.
1. Introduction
Polyoxometalates (POM) are ideal candidates for
catalyst design because their chemical properties can be
tuned by choosing the constituent elements. This is one
reason for their popularity in recent years. Especially
transition metal substituted derivatives have been ap-
plied frequently in homogeneous and heterogeneous
catalysis [1]. In the context of electrocatalytic reactions,
POM can be used either homogeneously dissolved in
the electrolyte solution or attached to the electrode
surface. Generally three different methods can be ap-
plied to immobilize POM on the electrode surface.
Keita and Nadjo have reported the electrodeposition of
POM on electrodes [2], Bidan et al. entrapped POM in
various conducting polymer films [3] and Dong’s group
described the spontaneous adsorption after soaking the
electrode in an aqueous POM solution [4].
* Corresponding author. Tel.: +33-5-56846573; fax: +33-
5-56842717.
E-mail address: kuhn@enscpb.u-bordeaux.fr (A. Kuhn)
0013-4686/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0013-4686(99)00399-0

1830
D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
lytic rate constants. The catalytic currents can be sig-
Despite the positive standard potential of the H2O2/
nificantly enhanced when a simplified strategy, recently
developed to bind multiple layers of POM to the elec-
trode surface [24–26], is employed to modify the elec-
trode surface.
H2O couple (E°= +1.76 V versus NHE), hydrogen
peroxide is not reduced at significant rates at most
electrode surfaces. Electrode modification with transi-
tion metal catalysts like iron porphyrins [19] or copper
and ruthenium complexes [20,21] enhances the reactiv-
ity of H2O2, however the reported potentials are all
negative versus SSCE.
2. Experimental
In contrast, phosphopolyoxomolybdates are readily
reduced at quite positive potentials and therefore
seemed to be good candidates to catalyse the electrore-
duction of hydrogen peroxide in an interesting range of
potential. In a recent paper Sun et al proposed a
Keggin polyoxomolybdate for the catalysis of H2O2
reduction [22]. Multiple layers of POM were fixed at
the electrode surface by a technique proposed earlier by
Ingersoll et al. [23] and catalytic activity was observed
qualitatively for a gold electrode modified with six
layers of anions, however the influence of the number
of layers on the catalytic efficiency has not been
investigated.
2.1. Materials
The polyoxometalate (NH4)6P2Mo18O62 was avail-
able from a previous study [24]. All other chemicals
were reagent grade and were used as received. Solutions
were prepared from ultra pure water that had been
passed through a purification train (Milli-Q Plus 185,
Millipore). Commercial glassy carbon electrodes with 3
mm diameter were used (Bioanalytical Systems, BAS).
2.2. Apparatus and procedures
This report summarises the results obtained using
Cyclic voltammetry (CV) experiments were carried
out in a three compartment cell (10 ml) with a BAS
Model 100B electrochemical analyzer at ambient tem-
perature (2092°C). The electrolyte had been bubbled
with nitrogen for at least 10 min prior to use. All
potentials are given with respect to a commercial Ag/
AgCl reference electrode (BAS) and a platinum wire
was used as counter electrode in a separate compart-
ment in order to avoid contamination of the electrolyte
by traces of platinum. If not otherwise mentioned,
scans were started at the positive end of the potential
range. The glassy carbon electrodes were polished prior
to each experiment with 0.05 mm alumina powder (Mi-
cropolish, Buehler) and then sonicated for 5 min in
ultrapure water.
6−
62
P2Mo18O
as a catalyst dissolved in the electrolyte
and adsorbed as a monolayer on the electrode surface.
The activity is quantified by measurements of the cata-
The general procedure of polyoxometalate mono-
and multilayer preparation used in this study has been
described in detail in previous publications [24,25].
3. Results
6−
62
3.1. Catalytic reduction of H2O2 with P2Mo18O
in
solution
The direct reduction of H2O2 is severely inhibited at
most electrode surfaces as can be seen for glassy carbon
in Fig. 1A. Successive addition of H2O2 to the support-
ing electrolyte results in a slight increase of the general
background current but no defined reduction process
can be distinguished. This contrasts with the behaviour
observed when POM is present in the supporting elec-
trolyte (Fig. 1B). The three waves present in the first
CV (curve a) correspond to the following reactions [27]:
Fig. 1. (A) Cyclic voltammogram at 10 mV s−1 in a solution
containing 0.5 M H2SO4 (curve a), 0.5 M H2SO4 and 150 mM
H2O2 (curve b), 0.5 M H2SO4 and 300 mM H2O2 (curve c). (B)
6−
62
Same as in (A) but 0.5 mM P2Mo18
O
is present in the
electrolyte.

D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
1831
The addition of 150 mM H2O2 dramatically changes
the CV. The increase of the reduction current situated
at the potential of the second redox wave (+350 mV
versus Ag/AgCl) is accompanied by the vanishing of
the corresponding oxidation peak. This clearly indicates
that hydrogen peroxide is reduced at this quite positive
potential via an electrocatalytic process involving the
6−
62
polyoxometalate P2Mo18O . Noting that the catalytic
current increases linearly with the concentration of
6−
62
H2O2 for a fixed concentration of P2Mo18O , it seems
that the reaction rate is first order in the substrate. For
the sake of comparison with the experiments described
further down, the catalytic efficiency (CE) is calculated
as the ratio of the catalytic current over the current of
the catalyst alone at a potential of +350 mV. For a
POM concentration of 0.5 mM and [H2O2]=300 mM a
value of CE=2.2 is obtained. Taking into account the
high concentration of hydrogen peroxide, this value is
quite low, indicating that the kinetics of the catalytic
process are rather slow.
3.2. Catalytic reduction of H2O2 at electrodes coated
6−
62
with a monolayer of P2Mo18O
As mentioned above polyoxometalates can be at-
tached to various electrode materials by spontaneous
adsorption [5]. The resulting monolayer seems to be at
least in some cases well organized [28] however the
mechanism of adsorption is still not completely under-
stood. Partial charge transfer from the substrate to the
POM has been suggested [24] but also other mecha-
nisms like a condensation with surface hydroxyl groups
can be imagined. In the first case it is obvious that the
adsorption should change the electronic structure of the
POM, and eventually also its catalytic activity. We were
Fig. 2. (A) Cyclic voltammogram (6=100 mV s−1) of glassy
carbon in 0.5 M H2SO4 (curve a), after the modification of the
6−
62
electrode with a monolayer of P2Mo18
O
(curve b). (B)
Cyclic voltammogram (6=10 mV s−1) of the modified elec-
trode in 0.5 M H2SO4 (a), after addition of 150 mM H2O2 (b),
300 mM H2O2 (c) and 450 mM H2O2 (d). (C) The electrode is
rinsed after experiment (B) and a cyclic voltammogram is
recorded in 0.5 M H2SO4 immediately (c) and after 15 min of
cycling (b). For comparison the cyclic voltammogram of the
monolayer before catalysis is also shown (a).
therefore interested in comparing the catalytic be-
6−
62
haviour of a P2Mo18O
monolayer with the results
obtained in bulk solution.
Fig. 2A shows the cyclic voltammogram for the first
6−
62
three redox waves of a P2Mo18O
monolayer ad-
sorbed on glassy carbon. Instead of spontaneous ad-
sorption at open circuit this monolayer has been
obtained by cycling the electrode in 0.5 mM
(NH4)6P2Mo18O62+0.5 M H2SO4 between +600 and
0 mV for several minutes. This procedure leads to more
stable layers and is therefore used for the study of
catalysis. Knowing from the above equations that the
waves are two-electron waves we can use the charge
under the peaks to calculate an apparent surface cover-
age with respect to the geometric surface of the elec-
trode. The obtained value of 4.4×10−11 mol cm−2 is
in good agreement with what would be expected for a
monolayer on a smooth electrode surface.
P2Mo18O +2 e+2 H+UH2P2Mo18O
(1)
(2)
(3)
6−
62
6−
62
H2P2Mo18O +2 e+2 H+UH4P2Mo18O
H4P2Mo18O +2 e+2 H+UH6P2Mo18O
6−
62
6−
62
6−
62
6−
62
For the discussion of the catalytic mechanism further
down it is important to know that all three waves move
to more negative potentials with a slope of approxi-
mately −60 mV/pH at low pH. The most positive
redox wave becomes pH independent above pH 3 indi-
cating that the reduced species is no longer protonated
at pH\3. The second wave is pH dependent up to pH
6 and the third wave seems to remain pH dependent up
to pH values where this POM is known to undergo
rapid decomposition [27].
The addition of increasing amounts of H2O2 leads,
like in the case of homogeneously dissolved catalyst to
an enhanced current at the potential of the second

1832
D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
redox wave (Fig. 2B), however the catalytic efficiency
seems to be better, compared to the experiment with
dissolved catalyst. For [H2O2]=300 mM a value of
CE=6.1 is obtained at +350 mV, which is three times
higher than the corresponding value for the catalyst in
solution. This kind of enhancement of catalytic activity
by supporting POM on carbon has been observed also
in the context of liquid phase oxidation of organic
molecules [29].
Despite this activation of the catalyst due to adsorp-
tion, the overall rate constant of the reaction is still
quite small. Attempts to record plateau currents as a
function of rotation rate in a typical rotating disk
electrode experiment lead to no increase of the current,
indicating that the rate of the catalytic reduction of
H2O2 is too low to produce a significant depletion of its
concentration at the electrode surface. This lack of
variation means that the general equation describing
the variation of the limiting current Ilim with rotation
rate
The catalytically active monolayer is fairly stable. We
have performed steady state voltammetric measure-
ments with a rotating disk electrode and even after
three hours of continuously holding the potential at
+300 mV no decrease in catalytic current was ob-
served. Furthermore, if the electrode is rinsed after
catalysis and transferred back into pure supporting
electrolyte the initial CV of the monolayer is recovered.
Curve c in Fig. 2C shows the first scan in pure support-
ing electrolyte after the catalysis experiment. It is
slightly different compared to the scan of the modified
electrode before the catalysis experiment (Fig. 2C curve
a). This seems to be due to a reversible transformation
of the catalyst, because already after a few cycles in
pure supporting electrolyte, a CV almost indistinguish-
able from the original monolayer is observed (Fig. 2C
curve b). This is an important result, as some publica-
tions suggest that the generation of the active form of
the catalyst implies decomposition of the POM [31]. In
our case an irreversible modification or degradation of
the catalyst by H2O2 seems to be neglectable, probably
because the surface confined molecules are more resis-
tant compared to the catalyst in solution.
1/Ilim=1/Ilev+1/Ikin
(4)
can be simplified to
Furthermore the monolayer is sufficiently well at-
tached in order to be not removed from the surface by
the hydrodynamic action of the rotating disk electrode.
This allows us to record the steady-state current re-
sponse of the modified electrode for long periods of
time and upon injection of increasing amounts of H2O2
(Fig. 3A). As can be seen from Fig. 3B the electrode
response is linear over at least one order of magnitude.
The upper limit of linearity was not explored because
we were more interested in the detection of small
concentrations. The monolayer modified electrode al-
lows detection of H2O2 concentrations down to 10 mM.
As our final goal is the detection of hydrogen peroxide
released from biochemical processes it is clear that this
detection limit is not satisfying so far.
I
lim=Ikin=n F A k C G
(5)
where C is the concentration of H2O2 in the bulk, G is
6−
62
the surface coverage with P2Mo18O
and the other
constants have their usual meaning [30]. The mean
value obtained in our experiment is k=4.8 M−1 s−1
.
3.3. Catalytic reduction of H2O2 at electrodes coated
6−
62
with multiple layers of P2Mo18O
In order to improve the sensitivity of the modified
electrode we thought to increase the total amount of
POM on the electrode surface. Several strategies exist
like electrodeposition of POM [32], incorporation of
POM into polymer films attached to the electrode
surface [3,10,33] or the recently developed multilayer
deposition techniques [22–26].
The latter one allows, in contrast to the electrodepo-
sition and incorporation approach, the deposition of
POM layers, separated by layers of counter cations
with a structural control on the molecular level. This
offers the possibility to combine in one multilayer
different POM and cations in a well controlled way.
With the help of such a molecular architecture on the
Fig. 3. (A) GC electrode modified with a monolayer of
6−
62
P2Mo18
O
is held at +300 mV in an RDE experiment
(ꢀ=1000 rpm). Addition of aliquots of 30% H2O2 to the 0.5
M H2SO4 supporting electrolyte. (B) Calibration curve ob-
tained from experiment (A).

D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
1833
formation of a cationic adlayer which allows the subse-
quent adsorption of a second layer of POM when
dipping the electrode again in POM solution (Fig. 4A
curve c). As can be seen from the CV the amount of
adsorbed POM is now approximately twice as high.
Repeated alternate dipping results in a multilayer with
interesting redox properties. If the total number of
layers is not too high, all POM molecules seem to keep
their electroactivity and as H2O2 is a small and neutral
molecule it should be able to penetrate the layers and
react electrocatalytically. In a very recent study [12] the
authors used this multilayer approach for the electro-
catalytic reduction of NO−2 but they stated that the
catalytic currents were essentially independent of the
number of layers in the electrode coating.
With the bilayer modified electrode shown in Fig. 4A
we performed for comparison an experiment analogue
to the one of Fig. 2B. It can be easily seen from Fig. 4B
that not only the absolute currents are higher in the
bilayer experiment, but also the catalytic efficiency for
[H2O2]=300 mM yields CE=8 at +350 mV. It seems
therefore advantageous to use the bilayer instead of a
simple monolayer. However a plot of the catalytic
current against hydrogen peroxide concentration re-
vealed a non-linear dependence. This led us to look
closer into the time dependence of the catalytic current
for a given H2O2 concentration. The result is shown in
Fig. 4C. In contrast to the monolayer experiment the
current is a function of time and goes through a
maximum after a couple of cycles (curve c). For the
maximum value of the catalytic current, which is ob-
tained after 20 min of cycling, a catalytic efficiency of
CE=19 can be calculated. This corresponds to a rate
constant of k=34 M−1 s−1 if G=9.6×10−11 mol
cm−2 is used as the new surface coverage of the bilayer
in Eq. (5). This higher rate constant allows the detec-
tion of smaller H2O2 concentrations (around 1 mM)
compared to the monolayer modified electrode.
Fig. 4. Background of glassy carbon in 0.5 M H2SO4 (a).
6−
62
Deposition of a second P2Mo18
O
layer by dipping the
monolayer modified electrode (b) for 10 min in a saturated
solution of DADD in 0.5 M H2SO4. After rinsing, the elec-
trode is dipped for 10 min in 0.5 mM POM solution. After
rinsing a CV is recorded in 0.5 M H2SO4 with 6=10 mV s−1
(c). (B) Cyclic voltammogram of the bilayer modified electrode
in 0.5 M H2SO4 (a), after addition of 150 mM H2O2 (b), 250
mM H2O2 (c), 300 mM H2O2 (d) and 400 mM H2O2 (e). (C)
Increase of catalytic current as a function of the number of
cycles for a bilayer modified electrode in 0.5 M H2SO4+300
mM H2O2. First cyclic voltammogram (6=10 mV s−1) after
the addition of H2O2 (a), after 10 min of cycling (b), after 20
min of cycling (c), after 30 min of cycling (d).
However, due to the variation of current with time it
is difficult to assign a specific current to a certain H2O2
concentration and hence no calibration curve was
recorded.
Apparently the species with higher catalytic activity
needs some time to be formed. This could be explained
by a reaction of hydrogen peroxide with POM and/or
DADD. One argument supporting this assumption is
that some catalytic POM/H2O2 systems from the litera-
ture for the oxidation of organic molecules seem to
work best in the presence of ammonium cations [1].
Ishii et al. reported enhanced epoxidation of olefins by
POM and H2O2 in the presence of ammonium salts
containing at least one long alkyl chain [17]. A second
argument is that the system goes through a maximum
of catalytic activity. This means that once the forma-
tion of the species with higher catalytic activity by
reaction of H2O2 with POM/DADD is achieved, the
electrode surface it might be possible to obtain synergis-
tic effects between different catalyst molecules, but this
type of study is out of the scope of this paper.
A recent report from our laboratory [25] described
the growth of stable POM multilayers using a series of
divalent cations (NH+3 (CH2)xNH3+) as spacer. In the
following we concentrate on NH+3 (CH2)12NH3+
(DADD) because it gave the best results.
Dipping of the POM monolayer modified electrode
in an acidic solution of DADD will result in the

1834
D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
catalyst on the surface, and a layer thin enough to
ensure good electron transfer and substrate diffusion.
A similar observation is made when methylviologen
is used instead of DADD as spacer cation. Fig. 5A
illustrates the enhancement of the H2O2 reduction cur-
rent when going from a monolayer to a bilayer. Again,
the bilayer is much more efficient in terms of catalysis
compared to the monolayer.
The main difference between using DADD or
methylviologen for the construction is the stability of
the obtained bilayer. As illustrated in Fig. 4C the
POM/DADD/POM bilayer shows a quite important
degradation during catalysis, whereas the POM/
methylviologen/POM bilayer seems to be rather stable
after reaching the maximum of catalytic activity (Fig.
5B). This increase in stability might be ascribed to the
permanent nature of the positive charges in methylvio-
logen. In the case of DADD the positive charges are
due to a protonation of the nitrogen and therefore a
local change in pH that might occur during catalysis
leads to the destruction of the bilayer.
bilayer is no longer stable. If DADD is involved in the
formation of this species it seems reasonable that it
cannot act any longer as an electrostatic anchor for the
second POM layer. Indeed we observe in an experiment
analogue to the one of Fig. 2C, that after the catalysis
experiment the equivalent of only one monolayer of
POM stays on the electrode surface.
In an additional experiment we checked the role of
the DADD adlayer with respect to catalysis. Deposition
of the first POM layer and one layer of DADD leads to
a decrease of the catalytic activity compared to the
POM monolayer alone. The hydrophobic character of
the long alkylchain creates a hostile environment for
the diffusion of H2O2 to the catalyst layer and it seems
therefore important for the outermost layer to be POM.
Transport problems due to this hydrophobic environ-
ment also prevent the efficient use of all catalytic layers
6−
62
when more than a bilayer of P2Mo18O
is adsorbed
on the electrode surface. Deposition of a triple layer for
example leads to no further significant increase in cata-
lytic currents. Therefore a bilayer of POM seems to be
an optimisation between the need of a high amount of
4. Discussion
POM are good examples for bifunctional catalysts,
that means their acid/base and redox properties work
cooperatively in the catalytic process. In the present
case of H2O2 reduction this bifunctionality is quite
helpful because we are dealing with a two electron, two
proton process
H2O2+2 e+2 H+U2 H2O
The intermediate formation of a peroxo complex
between POM and hydrogen peroxide is a conceivable
path for the catalysis [17] and should lead to a weaken-
ing of the oxygen–oxygen bond. The electron transfer
from the POM and cleavage of the oxygen–oxygen
bond has to be accompanied by a protonation of the
resulting fragments. This is possible at high rates if the
reduced POM is sufficiently protonated.
It is clear from our study and also from the results
obtained by other groups [7,12] that the catalytically
active form of this POM is in general the twice, two
6−
62
electron reduced species H4P2Mo18O . Increasing the
pH, all three redox waves shift with 60 mV/pH to more
negative redox potentials. However the most positive
redox wave becomes pH independent at around pH 3
as mentioned above. This means that at pH 4, like in
Fig. 6, the predominating species at the second wave is
Fig. 5. (A) Comparison of the catalytic current in 0.5 M
H2SO4+150 mM H2O2 obtained for an electrode modified
6−
62
with
a
monolayer of P2Mo18
O
(a) and an electrode
8−
6−
H2P2Mo18O
instead of H4P2Mo18O . Consequently
62
62
modified with a POM/methylviologen/POM layer (b). (B)
Evolution of catalytic current as a function of the number of
direct proton transfer from the POM to the H2O2
fragments becomes more difficult and catalysis of the
H2O2 reduction is suppressed on the second redox
cycles for
a POM bilayer modified electrode in 0.5 M
H2SO4+150 mM H2O2. CV (6=10 mV s−1) after the addi-
tion of H2O2 (a), after 10 min of cycling (b), after 20 min of
cycling (d), after 30 min of cycling (c).
wave. An increase in current is now visible on the third
8−
62
redox wave corresponding to the species H4P2Mo18O

D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
1835
5. Conclusions
It has been demonstrated that the simple Dawson
6−
62
polyoxometalate P2Mo18O
is able to catalyse elec-
troreduction of H2O2 at an unusual positive potential
of +0.35 V versus Ag/AgCl, comparable to the po-
tential observed on platinum electrodes. The immobil-
isation of this anion as a monolayer on glassy carbon
enhances its catalytic efficiency.
In contrast to the observation made by Kloster and
Anson [12] a significant improvement of catalytic ac-
tivity can be obtained by controlled deposition of
Fig. 6. The cyclic voltammogram of a monolayer modified
electrode is recorded in 0.5 M Na2SO4 adjusted with H2SO4 to
pH 4 (a). After addition of 50 mM H2O2 (b) and 400 mM
H2O2 (c) to the electrolyte.
6−
62
multilayers consisting of P2Mo18O
and a large
cation. This discrepancy might be explained by the
different substrate used in both studies. H2O2 is a
neutral molecule and can therefore penetrate the elec-
trostatic barrier of the outermost anionic POM layer
in order to access inner layers. NO−2 which was used
in the study by Kloster and Anson can probably not
penetrate the multilayer because of the electrostatic
repulsion with the last POM layer and therefore no
increase in catalytic activity is observed when piling
up multiple layers. This illustrates a limitation of the
multilayer modification technique, but the architecture
being very versatile it can be adapted to a specific
substrate.
Another critical point of the multilayer approach is
its stability under catalytic conditions. From the re-
sults presented in this paper it seems to be favourable
to use ‘permanent’ cations like methylviologen as a
spacer between the POM layers instead of temporary
cations like the protonated DADD.
. The increase in driving force and the higher proton
concentration on this species compared to
8−
62
H2P2Mo18O
seems to facilitate catalysis but the ef-
fect is very small.
With this information we can summarize the mech-
anism as follows:
P2Mo18O +2 e+2 H+UH2P2Mo18O
(6)
(7)
(8)
(9)
6−
62
6−
62
H2P2Mo18O +2 e+2 H+UH4P2Mo18O
6−
62
6−
62
6−
62
H4P2Mo18O +H2O2Uperoxo complex
6−
62
peroxo complexUH2P2Mo18O +2 H2O
The formation of the peroxo complex in Eq. (8)
seems to be necessary for an efficient inner sphere
catalysis as can be seen from the time dependence of
the catalytic current in Fig. 4C. The dramatic varia-
tion of catalytic currents with pH indicates that the
proton transfer from the POM to the substrate (Eq.
(9)) is the overall rate limiting step.
A maximum of catalytic current is observed for a
6−
62
bilayer of P2Mo18O
separated by one layer of
NH+3 (CH2)12NH3+ or methyviologen. The rate con-
stant for this optimized structure is 34 M−1 −1. This
s
shows that electrode surfaces with catalytic potency
can be fashioned by appropriate assembling of multi-
ple layers of this polyoxometalate. However the rate
constant is small and therefore our initial goal to
measure hydrogen peroxide produced in enzyme reac-
tions can not be achieved. In addition the catalytic
activity decreases with increasing pH and at the work-
ing conditions for enzymes (around pH 7) no more
catalysis is observed. This is due to the insufficient
protonation of the polyoxometalate and its limited
stability. One way around this problem might be the
use of polyoxometalates with higher pKa and in-
creased stability.
The increase of catalytic efficiency observed when
the monolayer is replaced by a bilayer separated by
DADD is difficult to explain. On the one hand
DADD does not seem to be directly responsible for
the enhanced current because the addition of
a
DADD adlayer on top of the POM monolayer does
not effect the current in a positive sense, on the other
hand there is a slow conversion of the POM/DADD
system into a form with higher catalytic activity,
which is not seen for POM alone. At present we only
can speculate that DADD activates indirectly the re-
duction due to the different mode of attachment of
the additional POM layer. The first POM layer is
tightly bound by chemisorption to the glassy carbon
surface, whereas the second POM layer is only bound
by electrostatic interaction, leading to a higher activ-
ity. This explanation is consistent with our observa-
tion that a similar enhancement of catalytic activity is
also observed when methylviologen is used as a cation
instead of DADD.
Acknowledgements
This work was supported by the Region Aquitaine.
Helpful discussion with the other members of our re-
search group are acknowledged with gratitude.

1836
D. Martel, A. Kuhn / Electrochimica Acta 45 (2000) 1829–1836
[17] Y. Ishii, K. Yamawaki, T. Ura, H. Yamada, T. Yoshida,
References
M. Ogawa, J. Org. Chem. 53 (1988) 3587.
[18] M. Shimizu, H. Orita, T. Hayakawa, Y. Watanabe, K.
Takehira, Bull. Chem. Soc. Jpn. 63 (1835) 1990.
[19] K. Shigehara, F.C. Anson, J. Phys. Chem. 86 (1982) 2776.
[20] K. Shigehara, F.C. Anson, J. Electroanal. Chem. 132
(1982) 107.
[21] N. Oyama, F.C. Anson, J. Electroanal. Chem. 199 (1986)
467.
[22] C. Sun, J. Zhao, H. Xu, Y. Sun, X. Zhang, J. Shen, J.
Electroanal. Chem. 435 (1997) 63.
[1] I.V. Kozhevnikov, Chem. Rev. 98 (1998) 171.
[2] B. Ke¨ıta, L. Nadjo, J. Eletroanal. Chem. 191 (1985) 441.
[3] G. Bidan, M. Laprowski, J.P. Travers, Synth. Met. 28
(1989) C113.
[4] B. Wang, S. Dong, J. Electroanal. Chem. 328 (1992) 245.
[5] M. Sadakane, E. Steckhan, Chem. Rev. 98 (1998) 219.
[6] X. Xi, S. Dong, Electrochim. Acta 40 (1995) 2785.
[7] A. Papadakis, A. Souliotis, E. Papaconstantinou, J. Elec-
troanal. Chem. 435 (1997) 17.
[23] D. Ingersoll, P.J. Kulesza, L.R. Faulkner, J. Electrochem.
Soc. 141 (1994) 140.
[8] J.E. Toth, F.C. Anson, J. Am. Chem. Soc. 111 (1989)
2444.
[24] A. Kuhn, F.C. Anson, Langmuir 12 (1996) 5481.
[25] A. Kuhn, N. Mano, C. Vidal, J. Electroanal. Chem. 462
(1999) 187.
[26] C. Sun, J. Zhang, Electrochim. Acta 43 (1998) 943.
[27] E. Papaconstantinou, M.T. Pope, Inorg. Chem. 6 (1967)
1152.
[9] J.E. Toth, J.D. Melton, D. Cabeli, B.H.J. Bielski, F.C.
Anson, Inorg. Chem. 29 (1952) 1990.
[10] B. Ke¨ıta, A. Belhouari, L. Nadjo, R. Contant, J. Elec-
troanal. Chem. 381 (1995) 243.
[11] S. Dong, X. Xi, M. Tian, J. Electroanal. Chem. 385
(1995) 227.
[28] W.G. Klemperer, C.G. Wall, Chem. Rev. 98 (1998) 297.
[29] S. Fujibayashi, K. Nakayama, M. Hamamoto, S.S.Y.
Nishiyama, Y. Ishii, J. Mol. Catal. 110 (1996) 105.
[30] N. Oyama, F.C. Anson, Anal.Chem. 52 (1980) 1192.
[31] C. Aubry, G. Chottard, N. Platzer, J.M. Bre´geault, R.
Thouvenot, F. Chauveau, C. Huet, H. Ledon, Inorg.
Chem. 30 (1991) 4409.
[32] B. Ke¨ıta, L. Nadjo, J. Electroanal. Chem. 243 (1988) 87.
[33] H. Sung, H. So, W.-K. Paik, Electrochim. Acta 39 (1994)
645.
[12] G.M. Kloster, F.C. Anson, Electrochim. Acta 44 (1999)
2271.
[13] H. Furukawa, T. Nakamura, H. Inagaki, E. Nishikawa,
C. Imai, M. Misono, Chem. Lett. 877 (1988).
[14] A.M. Khenkin, C.L. Hill, J. Am. Chem. Soc. 115 (1993)
8178.
[15] R. Neumann, in: M.T. Pope, A. Mu¨ller (Eds.), Polyox-
ometalates, Kluwer, 1994, pp. 307.
[16] R.D. Gall, C.L. Hill, J.E. Walker, Chem. Mater. 8 (1996)
2523.
.