Preparation and electrochemical studies of metal-carbon composite catalysts for small-scale electrosynthesis of H2O2

Article abstract of DOI:10.1016/j.electacta.2011.06.043

Numerous transition metal-carbon composite catalysts (M = V, Zn, Ni, Sn, Ce, Ba, Fe, Cu) have been synthesized and tested for electroreduction of O ₂ to H₂O₂, The activity and selectivity of all synthesized catalysts for electrosynthesis of H₂O₂ were determined by the rotating ring-disk electrode method in acidic and neutral electrolytes. The Co-based catalysts in general showed the highest activity towards H₂O₂ formation. Experiments with different loading contents of Co showed that the activation overpotential losses of oxygen reduction to H₂O₂ reduces as loading increases to about 4 wt% Co. Addition of Co beyond this level did not seem to impact the overpotential losses. The cobalt-based catalysts, were spray-coated onto 120 μm thick Toray graphite substrates, and were studied in bulk electrolysis cells for up to 100 h at potentiostatic conditions (0.25 V vs. RHE) in pH 0, 3, and 7 electrolytes. At (25 °C and 1 bar) with a catalysts loading of about 1mgcmgeometric-2 and using dissolved O₂ in 0.5 M H₂SO₄, typical H₂O₂ electrosynthesis rates of about 5μmolh^-1cmgeometric-2 were reached with current efficiencies of about 85 ± 5% at 0.25 V (vs. RHE).

Full text of DOI:10.1016/j.electacta.2011.06.043

Electrochimica Acta 56 (2011) 9074–9081
Contents lists available at ScienceDirect
Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta
Preparation and electrochemical studies of metal–carbon composite catalysts for
small-scale electrosynthesis of H2O2
Arman Bonakdarpour∗, Daniel Esau, Hillary Cheng, Andrew Wang, Elöd Gyenge, David P. Wilkinson∗∗
Department of Chemical and Biological Engineering and Clean Energy Research Center, 2360 East Mall, University of British Columbia, Vancouver, BC V6T 1Z3, Canada
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 13 April 2011
Received in revised form 10 June 2011
Accepted 12 June 2011
Available online 30 June 2011
Numerous transition metal–carbon composite catalysts (M = V, Zn, Ni, Sn, Ce, Ba, Fe, Cu) have been syn-
thesized and tested for electroreduction of O2 to H2O2, The activity and selectivity of all synthesized
catalysts for electrosynthesis of H2O2 were determined by the rotating ring-disk electrode method in
acidic and neutral electrolytes. The Co-based catalysts in general showed the highest activity towards
H2O2 formation. Experiments with different loading contents of Co showed that the activation overpo-
tential losses of oxygen reduction to H2O2 reduces as loading increases to about 4 wt% Co. Addition of
Co beyond this level did not seem to impact the overpotential losses. The cobalt-based catalysts, were
spray-coated onto 120 ␮m thick Toray® graphite substrates, and were studied in bulk electrolysis cells
Keywords:
Electrosynthesis of H2O2
O2 reduction
Carbon composite catalysts
Acidic and basic media
for up to 100 h at potentiostatic conditions (0.25 V vs. RHE) in pH 0, 3, and 7 electrolytes. At (25 ◦C and
−2
1 bar) with a catalysts loading of about 1 mg cm
and using dissolved O2 in 0.5 M H2SO4, typical
geometric
H2O2 electrosynthesis rates of about 5 ␮mol h−1 cm−2
were reached with current efficiencies of
geometric
about 85 5% at 0.25 V (vs. RHE).
© 2011 Elsevier Ltd. All rights reserved.
1. Introduction
basic solution (2 M NaOH) and H2O2 (1 M) are used in bleaching and
brightening of pulp and paper thus making the on-site electrosyn-
thesis of alkaline H2O2 very attractive for this application. Oloman
discusses different electrochemical reactors, including Dow/H-D
Tech, UBC, and E-TEK designs for producing alkaline H2O2 [3]. These
designs typically use carbon material for the cathode, Ni for the
anode, and either a diaphragm or a membrane as a separator [4–7].
NaOH and O2 are introduced co-currently as catholyte and anolyte,
the carbon electrode is cathodically polarized by an external power
source and the resulting mixed NaOH/H2O2 trickles out of the reac-
tor [3]. Electrode and separator materials, temperature, and reactor
design are all important factors in producing the desired concen-
trations of H2O2, enhancing the reactor lifetime, and improving the
overall economics.
Electroreduction of O2 to H2O2 occurs readily in alkaline
solutions through the two electron reduction mechanism on
carbon-based electrodes [8]. In acidic media, however, carbon
shows high oxygen reduction overpotentials and thus poor activity
as an ORR catalyst. Research has shown that addition of transition
metals like Co or Fe, and nitrogen, followed by a pyrolysis treatment
lead to formation of catalytic sites which are highly active for ORR
to either H2O or H2O2 [9,10]. Fe-activated carbons are especially
active for production of H2O and as such have been intensively stud-
ied for the cathodic reactions of fuel cells in the past few decades
[11–13]. Co-activated carbons, on the other hand are known for the
reduction of O2 to H2O2. Marcotte et al. have performed detailed
studies on the impact of carbon precursors and Co concentration on
Commercial manufacturing of H2O2, with a current global
production rate of 2.2 × 106 tonnes yr−1, involves sequential hydro-
genation and oxidation of alkylanthraquinone (Riedl-Pfleiderer
process), organic solvents, high temperatures (∼500 ◦C), Pd cata-
lysts and liquid–liquid extraction [1]. This thermochemical process
is neither energy efficient nor environmentally benign; however,
when coupled to other chemical processes in petrochemical plants,
where excess heat is available, the process becomes economical
[2]. There has also been significant research into alternative meth-
ods which rely on direct reactions of H2 and O2 to form H2O2 [1].
Direct routes, however, face challenging issues like handling poten-
tially explosive mixtures of O2/H2, simultaneous side reactions and
requirements of highly selective catalysts. Samanta and Campos-
Martin et al. review alternative methods of H2O2 production in
recent publications [1,2].
Hydrogen peroxide can also be directly synthesized by elec-
troreduction of O2 in either acidic or alkaline media. Electrochem-
ical synthesis of H2O2 in alkaline media is both facile and highly
relevant, for example, in the pulp and paper industry. Concentrated
∗
Corresponding author. Tel.: +1 604 827 3181.
Corresponding author. Tel.: +1 604 822 4888.
E-mail addresses: arman@chbe.ubc.ca (A. Bonakdarpour),
∗∗
dwilkinson@chbe.ubc.ca (D.P. Wilkinson).
0013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2011.06.043

A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
9075
the activity and selectivity of catalysts for H2O2 formation [14]. The
authors hypothesize that of the two possible CoN4/C and CoN2/C
active sites, the former could lead to more H2O2 generation. Funda-
mental understanding of the nature of active sites for the reduction
of O2 on these types of catalysts is still under investigation. Very
recently, Olson et al. proposed a model for reduction of O2 on Co
catalysts synthesized from pyrolyzing Co porphyrins where Co–Nx
centers are active for reduction of O2 to H2O2 and any surface
oxides (CoxOy) would be effective in a further reduction of step
of H2O2 to H2O [15]. Different carbons (multi-walled nanotubes,
XC 72 or other high surface area carbons), various Co precursors
(Co porphyrins, polypyrroles, acetates), the impact of Co concentra-
tion, and ORR properties have also been reported in the literature
[16–18]. Yang et al. observed high selectivities for H2O2 generation
on sputter deposited Co–N–C catalysts spanning very wide compo-
sition ranges [19]. In more recent publications, Olson et al. discuss
Co polypyrrole derived catalysts in alkaline media and provide evi-
dence for a dual site functionality by which O2 is reduced to HO−
first and then to either OH− or H2O2 [15,20]. Furthermore, organic
redox catalysts, such as anthraquinone derivatives, loaded onto gas
diffusion electrodes, are shown to be effective in electrogeneration
of H2O2 [21].
environmentally friendly and decomposes to O2 and H2O. For these
reasons the on-site production H2O2 by the electrosynthesis route
remains an ideal option [39–41].
This paper presents results for catalysts and electrodes pre-
pared for the electrosynthesis of H2O2 in acidic and near neutral
media. Rotating ring-disk electrode (RRDE) measurements are used
to screen metal–C–N catalysts synthesized in-house for their activ-
ity and selectivity towards H2O2 generation from O2 reduction in
different pH electrolytes. Batch electrolysis cells, using electrodes
made from gas diffusion layers coated with catalyst, were used
to measure the production rate of H2O2 from dissolved O2 in the
electrolyte and to make preliminary studies of electrode durabil-
ity. The importance of electrode preparation and stability of H2O2
in the electrolyte are highlighted and discussed. Optimization of
electrode structures and their evaluation in continuous or batch
reactors (e.g., trickle bed reactors, fuel cell-type reactors, or large
volume batch electrolysis) are currently underway and will be
reported in subsequent publications. The larger scope of our on-
going project is to develop an electrochemical reactor for on-site
generation of H2O2 and system integration with an H2O2/UV AOP
water treatment facility.
Besides trickle-bed reactors, both alkaline and acidic fuel cell
type reactors for generation of H2O2 have also been investigated.
Brillas et al. describe an alkaline fuel cell with KOH electrolyte, a
Pt/C anode and an uncatalyzed carbon electrode for oxidation of
2. Experimental methods
H
2 and reduction of O2 (to HO2−), respectively [22]. Yamanaka and
2.1. Catalyst, ink and electrode preparations
co-workers have in the recent years investigated the possibility of
an acidic H2 fuel cell for producing H2O2 [18,23]. In this config-
uration, H2 fuel and O2 oxidant are fed to the anode and cathode
electrodes of a fuel cell reactor and H2O2 is generated at the cathode
when appropriate catalysts like Co-porphyrins are used. The elec-
trodes are separated by a proton conducting polymer membrane
and in addition to H2O2, electrical power is also generated. In sum-
mary, for production of H2O2 in acidic environments, development
of catalysts which are active, i.e., exhibit a low ORR overpotential
drop for H2O2, have high selectivity for H2O2 production, and are
stable in acidic pH and in the presence of H2O2, are highly desired.
Major applications of H2O2 are in chemical synthesis, the tex-
tile industry and paper bleaching [2,24,25]. A very important and
rapidly growing usage of H2O2 is in advanced oxidation processes
(AOP), which are increasingly being applied for the disinfection of
waste and drinking water, following significant research and devel-
opment over the past two decades [26–35]. The more commonly
used chlorination treatment, still in wide spread use, leads to for-
mation of disinfection by products (DBPs) like trihalomethanes,
which are potentially toxic or carcinogenic compounds and whose
contents in drinking water are regulated by health and environ-
mental agencies [36,37]. There are many different routes to AOPs;
however, most common approaches rely on efficient production
and usage of hydroxyl radicals (OH•). Hydroxyl radicals, are highly
reactive, short-lived, electrophilic and non-selective species, and
can disintegrate organic waste matter by initiating kinetically fast
oxidative reactions, which ideally proceed to complete mineraliza-
tion, i.e., terminate with the formation of CO2, H2O and mineral
acids. These radicals can be generated from oxygen species like O3
or H2O2, chemically by Fenton reagents like Fe2+, photochemically
by ultra violet (UV) radiation, or photocatalytically by a combina-
tion of UV radiation and TiO2 [31,35,38]. As such, H2O2 remains a
very important chemical in the advanced oxidation treatment of
water. For small systems, especially in remote locations, it is highly
desired to have on-site H2O2 generation because potential hazards
associated with the transport and handling of concentrated H2O2
will be avoided and the water treatment plant will be self-sufficient
in this respect. In general, very dilute concentrations (∼10 ppm) of
H2O2 are required in the UV AOP water treatment reactors. H2O2 is
Appropriate amounts of carbon Vulcan XC 72 (Cabot Corpora-
tion) and metal nitrate salts (Alfa Aesar) were mixed and sonicated
in de-ionized (DI) water and dried overnight at 60 ◦C in an oven. The
cobalt precursor consisted of Co(NO3)2·6H2O with a 99.99% purity
(Alfa Aesar). Similar nitrate salts were used for V, Zn, Ni, Sn, Ce, Ba,
Fe, and Cu as well. The mixture was heat treated in a ceramic tube
furnace (Fischer Scientific) at 900 ◦C under flowing N2 for 2 h. From
an economical point of view, nitrate salts of Co (98% purity) are
generally about 100× less expensive per gram of Co than porphyrin
compounds.
After heat treatment the powder was ground and inks for the
RRDE and electrodes used in the electrolysis cell were prepared. For
the RRDE experiments a mixture of 8 mg powder, 100 ␮L Nafion
(Aldrich, 5 wt% Nafion in ethanol), 75 ␮L ethanol, and 350 ␮L DI
water were mixed and sonicated for about 0.5 h in order to produce
very uniform slurries. Careful studies were performed to opti-
mize the ink formulation and loadings on the tip of the RRDE. An
8 ␮L drop was placed on the tip of the glassy carbon RRDE by a
micropipette to produce a total composite catalyst (i.e., transition
metal–carbon–nitrogen) loading of 0.6 mg cm−2 on the tip of the
RRDE. For a 4 wt% Co catalysts, this corresponds to a metal loading
of about 0.025 mgCobalt cm−2 on the tip of the RRDE. The catalyst
ink, deposited on the glassy carbon (GC) tip of the RRDE, was dried
under flowing N2. The impact of catalyst loading on the measured
activity and selectivity of oxygen reduction electrocatalysts has
been reported in the literature [42,43].
For the batch electrolysis experiments the catalyst inks for the
electrodes were made by mixing 600 mg catalyst powder (4 wt%
Co), 10 mL Nafion, 30 mL DI water and 10 mL ethanol and sonicating
the mixture for 20 min. The slurries were coated onto Teflon treated
Toray carbon papers (TGP-H-120 Fuel Cell Earth) by using either
brush painting or an air-powered, gravity-feed spray gun (Master-
craft, Canadian Tire) while heating them on a hotplate at about
70 ◦C. The catalyst loadings on the Toray papers varied from 1 to
5 mg cm−2. During the spray-coating the electrodes were held tight
in a home-made aluminum jig. The spray-painted electrodes have
much higher quality in terms of uniformity and longevity. However,
the deposition time is several hours compared to 10–20 min for the

9076
A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
brush painted electrodes. All the electrolysis results reported in this
paper were made using air-sprayed electrodes.
2.2. X-ray diffraction and electron microscopy
High resolution transmission electron imaging was performed
by a JEOL JEM 2010F using a 200 keV beam energy. Elemental anal-
ysis was performed by an Oxford INCA Pentafet energy dispersive
spectrometer (EDS) connected to the JEOL microscope. The speci-
men samples were prepared by placing a droplet of a thoroughly
sonicated mixture of the sample and methanol onto TEM copper
grids covered with lacy carbon films.
The X-ray diffraction (XRD) data were collected using a Bruker
D8 Advance diffractometer equipped with a position-sensitive Van-
˚
tec detector, Cu K␣ radiation (1.51 A) source, and appropriate
detector and anti-scatter slits. Microscope glass slides were used
as sample holders.
Fig. 1. Schematics of the electrolysis cell used for the batch H2O2 electrosynthesis
experiments.
2.3. Electroanalytical measurements
2.4. Batch electrolysis measurements
The experimental setup for these measurements consisted of a
Pine RRDE system, a Pt counter electrode (Radiometer Analytical),
a Cl−-free Hg/Hg2SO4 (Radiometer Analytical) reference electrode,
and 0.1 M double distilled HClO4 (GFS Chemicals). The Pine instru-
mentation consisted of a bipotentiostat (AFCBP1), a Pine rotator
(AFMSRCE) and a Pine RRDE electrode (AFE6R2GCPT) with a GC disk
(5 mm dia.) and a Pt ring (55 mm inner and 65 mm outer dia.). To
keep the cell O2-saturated, O2 (99.998% purity) was introduced in
the electrolyte through a 5 ␮m porous dispersion tube (Ace Glass)
during the measurements. All measurements were performed at
room temperature (21 0.5 ◦C). Extreme care was taken to ensure
the cleanliness of the glassware, cell parts, and the samples during
handling and setup. Electrochemical potentials mentioned in this
article are referenced to the reversible hydrogen electrode (RHE)
and currents are all normalized by the geometric area of the elec-
trodes.
For batch electrolysis experiments, a two compartment glass
cell, shown in Fig. 1, was used. In this configuration, the anode
compartment was separated from the cathode by a Nafion® 112
(Dupont) membrane. The membrane was pre-treated by a three
step boiling treatment in 3 wt% H2O2, DI water, and 0.5 M H2SO4
prior to use. The membrane prevents diffusion of H2O2 to the anode
compartment, where it could oxidize at the counter electrode. The
cell was equipped with a Ag/AgCl reference electrode (BASi), a
Pt counter electrode (Radiometer Analytical) and gas dispersion
tube (Ace Glass) for purging O2 in the main compartment. Double-
distilled H2SO4 (HFS Chemicals) at different concentrations (pH 0,
3) was used for the electrolyte. For pH 7 measurements a buffer
solution (KH2PO4, Fischer Scientific) was used. The cathode (see
Fig. 1) was held in place by a flat tip copper alligator clip which also
provided the electrical contact. The solution in the cathode com-
partment was continuously mixed using a magnetic stirrer bar. A
side port on the cell allowed for regular extraction of electrolyte
for UV spectrophotometric analysis. For each UV measurement a
3 mL sample was removed using a micropipette and was immedi-
ately replaced by fresh electrolyte to maintain the solution level
constant in the cell. All the electrosynthesis experiments were per-
formed potentiostatically at a constant potential of 0.25 vs. RHE
using the Pine bipotentiostat system.
The RRDE electrode used (Pine AFE6R2GCPT) had a nominal col-
lection efficiency of about 35%. Selectivity of the catalysts towards
H2O2 formation (or selectivity) was calculated by:
200 × Iring
(N × |Idisk|) + Iring
SelectivityH
=
(1)
O
2
2
where N, Iring and Idisk refer to the collection efficiency of the ring,
ring current, and disk current, respectively [44]. SelectivityH
The theoretical concentration of H2O2, which is produced by a
2 electron (n = 2) reduction of O2 molecules according to Eq. (2), at
time T, was determined by Faraday’s law of electrolysis as shown
in Eq. (3).
O
2
here is the percentage of the reduced O2 molecules that are con2-
verted to H2O2. The ring electrode was held at 1.2 V vs. RHE during
the oxygen reduction reaction (ORR) measurements to selectively
oxidize the H2O2. For electrochemical cleaning, the Pt ring elec-
trode was initially cycled between 0 and 1.2 V vs. RHE for 100 (or
more) cycles until steady state CVs were obtained. The first CV was
always done in N2-purged cell to obtain the background CV. Simi-
lar measurements were also performed after the oxygen reduction
reaction (ORR) experiments to determine the adhesion of the cat-
alyst to the RDE. Consistent double layer capacitance (before and
after ORR measurements) would indicate that the catalyst layer is
adhering well to the GC tip of the RRDE.
O2 + 2H+ + 2e− → H2O2, E2098 = 0.70VSHE
(2)
ꢀ
T
CE
2FV
[H2O2]T
=
I dt
(3)
0
where CE, F, I, t, V and refer to current efficiency for peroxide for-
mation, Faraday’s constant, electric current, time, and electrolyte
volume, respectively.
ORR measurements consisted of CVs (3 cycles) between the
open circuit voltage (OCV) (0.7 vs. RHE) and 0.1 V with a sweep rate
of 5 mV s−1. The capacitive currents obtained by CV measurements
in the N2-saturated electrolyte were subtracted from the ORR data
to extract the Faradaic currents. The RHE scale for the reference
electrode was established by purging H2 in the cell and recording
the OCV of the Pt ring electrode.
2.5. H2O2 analysis by UV–Vis spectrophotometry
For analytical detection of H2O2, the UV spectrophotometric
method described by Klassen and co-workers was used [45]. This
sensitive method uses the 351 nm wavelength absorption of UV
−
radiation by the I3 species which are formed when H2O2 reacts
with KI in presence of ammonium molybdate catalyst in buffered

A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
9077
solutions. For blank measurements, it was important to use the cor-
−
responding electrolyte without any H2O2. The I3 method is well
suited for detecting very low concentrations (∼␮M) of H2O2. When
H2O2 concentration was higher, samples were first diluted down
to a range of 10–50 ppm. The UV absorption measurements were
performed with a Varian Cary 100 UV–Vis spectrophotometer. At
least two measurements were performed to check the repeatability
of the results.
To study the pH stability of H2O2, commercial grade H2O2 (Fis-
cher Scientific, 3%, w/w) was used to prepare 350 ppm (mg L−1
)
stock solutions using several different pH (0, 0.7, 4, 6.5, 8, and 10)
solutions with either double distilled H2SO4 (GFS Chemicals) or
NaOH (Fischer Scientific) and were kept at 295 K. The concentra-
tion of H2O2 in these samples was periodically determined using
the UV spectrophotometric method.
3. Results and discussion
3.1. Physical characterizations
Fig. 2 shows the X-ray diffraction (XRD) patterns of 0, 2, and
4 wt% Co loading catalysts. Heat treatment at 900 ◦C leads to for-
mation of crystalline face centered cubic cobalt grains. The impact
of aging is also illustrated in Fig. 2. Aging (exposure to air for >1
year) leads to significant oxidation of Co grains and formation of
face centered cubic Co3O4 phase.
TEM image of the catalysts and the composition of selected
regions are shown in Fig. 3. The energy dispersive spectra show
both high and low Co content regions. No evidence for presence of
nitrogen is observed. Further work is in progress to detect presence
of nitrogen in the samples. These results also show that Co is not
very uniformly distributed as, for instance, there is no indication of
significant Co in spot marked ‘C’ in the image. Further research will
aim to improve the distribution of Co during the synthesis.
Fig. 3. TEM image and energy dispersive spectra of a 4 wt% Co–C catalyst.
3.2. Electroanalytical and electrolysis measurements
3.2.1. Co-based catalysts
composite catalyst deposited on the RRDE tip. Loss of catalyst from
the RRDE tip, due to poor adhesion, would decrease the double
layer capacitance current seen here. The difference between CVs
taken before and after the oxygen reduction measurements is neg-
ligible showing that the high surface area catalyst layer retains
a good adhesion to the GC RRDE tip during the entire time of
The adhesion of the electrocatalyst to the GC disk was exam-
ined by comparing the cyclic voltammograms recorded in N2
purged electrolyte before and after performing three oxygen reduc-
tion experiments. Fig. 4 shows two voltammograms, which were
recorded before and after the ORR measurements. The double layer
capacitance observed in Fig. 4 predominantly originates from the
Fig. 4. Cyclic voltammograms of Co-catalyst loaded on GC disk electrode in N2
purged electrolyte before and after ORR measurements. The experiments were
performed using a catalyst loading of 0.6 mg cm−2 on the RRDE tip, 0.1 M HClO4
electrolyte, at 295 K and 1 bar pressure.
Fig. 2. X-ray diffraction pattern of the samples shown for (a) blank glass slide, (b)
heat treated carbon Vulcan XC-72, (c) 2 wt% Co, (d) 4 wt% Co, and (e) 4 wt% Co after
exposure to air for about one year. The patterns are shifted for clarity.

9078
A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
100
80
c)
60
40
ring potential = 1.2 V vs. RHE
0.1 M HClO 4, 293 K, 5 mV s-1
20
cathodic sweeps, 900 rpm
0.2
b)
0.15
0.1
Fig. 6. Tafel analysis of O2 reduction for several composite Co–C catalysts. The exper-
iments were performed using a catalyst loading of 0.6 mg cm−2 on the RRDE tip,
0.1 M HClO4 electrolyte, at 295 K and 1 bar pressure.
From disk and ring current data, the selectivity of electrocata-
lysts towards H2O2 formation was calculated using (1) and results
are shown in Fig. 5a. Overall for potentials below 0.4 V a selectiv-
ity of 80–90% is observed. The competing secondary reaction is the
further reduction of H2O2 to H2O:
0.05
0
a)
H2O2 + 2H+ + 2e− → 2H2O, E0 = 0.53VSHE
(5)
-
0.5
-1
Guillet et al. also report a similar trend, i.e., increased H2O2 genera-
tion as a function of Co loading. However, they note that increasing
the Co levels above 0.2 wt% reduces the H2O2% released into the
electrolyte [16]. A subsequent reduction to H2O is then most likely
operative. Thus, from a combined optimization of overpotential
and H2O2 production, the authors conclude that, the 0.2 wt% Co
catalyst is the best among the Co-porphyrin derived catalysts. In
the catalysts reported here, the overpotential for the O2 reduction
decreases as the Co content is increased up to about 4 wt%, how-
ever, the H2O2 levels remain the same for all catalysts at potentials
below 0.3 V. Thus an optimized catalyst of 4 wt% Co shows both the
lowest overpotentials and still high H2O2 selectivity levels of above
80%. Furthermore, the H2O2 level of about 80% is nearly constant
and does not drop as the electrode potential is further decreased,
in contrast to what has been reported for the Co-porphyrin derived
catalysts [16]. Therefore, it appears that catalysts prepared here are
predominantly H2O2 producing catalysts and are deficient in H2O
generating active sites.
0 wt% Co
0.5 wt% Co
1 wt% Co
1.5 wt% Co
2 wt% Co
3 wt% Co
3.5 wt% Co
4 wt% Co
-
1.5
-2
0
0.2
0.4
0.6
Disk Potential vs. RHE [V]
Fig. 5. Rotating-ring disk electrode results for oxygen electroreduction on Co–C
composite catalysts with different Co loadings. (a) O2 reduction current densities
obtained from the disk electrode, (b) oxidation current densities on the Pt ring, and
(c) calculated selectivity of the catalysts for reduction of O2 to H2O2. The experiments
were performed using a catalyst loading of 0.6 mg cm−2 on the RRDE tip, 0.1 M HClO4
electrolyte, at 295 K and 1 bar pressure.
Fig. 6 shows a Tafel analysis of the kinetic regions of sev-
eral selected electrocatalysts whose ORR activity and selectivity
are shown in Fig. 5. The kinetic currents were obtained from the
Koutecky–Levich relation:
measurements. Furthermore, the three ORR cycles recorded for
each catalyst were stable.
Fig. 5 shows the oxygen reduction currents of several differ-
ent Co-based electrocatalysts. In general as Co wt% increases an
improvement in the catalytic activity is observed. The current
shown on the y-axis is first corrected for the non-Faradic current
arising from the charging/discharging of the double layer capaci-
tance and then is normalized by the electrode geometric surface
area (SA = 0.2 cm2) to obtain the net Faradaic current due to oxygen
electroreduction:
1
1
1
=
+
(6)
jFaradaic,geometric
jkinetic
jdiffusion-limited
where jFaradaic,geometric, jkinetic, and jdiffusion-limited are the Faradaic
current density measured, the kinetic, and the diffusion-limited
current densities, respectively. The diffusion-limited current den-
sity for a perfect smooth disk electrode rotating with an angular
velocity of ω, is expressed by the Levich equation:
jdisk-ORR − jdisk-N
2
jFaradaic,geometric
=
(4)
SA geometric
jdiffusion-limited = 0.62nFD2/3
ꢀ
−1/6ω1/2CO
(7)
,b
O
2
2
The highest electroreduction currents were obtained for catalysts
with a Co loading of 4–5 wt%. With these catalysts an open circuit
potential (OCP) of about 0.7 vs. RHE could be reached. Increasing
the Co content beyond these levels did not seem to improve the
oxygen electroreduction. The composite catalyst (Co–C) loading on
all disks was similar (600 ␮g cm−2).
where C, F, D, n, ꢀ and ω refer to concentration of the ana-
lyte, Faraday’s constant, diffusivity of the analyte, number of
electrons transferred, viscosity of the electrolyte, and angular
velocity of the rotating electrode. Using the literature reported
values of CO = 1.26 × 10−3 mol L−1, DO = 1.93 × 10−5 cm2 s−1
,
2
2,b

A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
9079
and ꢀH = 1.009 × 10−2 cm2 s−1, and the experimental value of
100
O
2
n = 2.4 (equivalent of 80% selectivity for H2O2 production), we
estimate the jdiffusion-limited to be about 2.65 mA cm−2. Actual cur-
rent plateaus indicating mass transfer limitations are not always
observed for these types of electrocatalysts (Fig. 4). Linear regres-
sion of all the Tafel data shown by Fig. 6, generates a slope of about
45 2 mV dec−1, suggesting that a similar electroreduction mech-
anism is operative for all the electrocatalysts shown. For the best
electrocatalyst, i.e., 4 wt% Co, a kinetic gain of about 300 mV (equiv-
alent to a seventy fold increase in the current at 0.45 V) is observed
over the baseline XC-72 carbon for the operating conditions used.
The activity gain is a result of the enhanced exchange current den-
c)
80
60
40
20
sity, which for the 4 wt% Co sample reaches 5.8 × 10−5 mA cm−2
,
0.2
indicating that the number of catalytic sites increases markedly by
addition of Co and performing a heat treatment.
ring potential = 1.2 V vs. RHE
0.1 M HClO4 , 293 K, 5 mV s-1
cathodic sweeps, 900 rpm
b)
0.15
0.1
3.2.2. Other transition metal-based catalysts
Other metals like Mn-based porphyrins have also been reported
for the reduction of O2 to H2O2 [18]. For this reason, a few transition
(Cu, Zn, Ni, Ba, Fe, Sn) and one rare earth (Ce) metals were used to
prepare transition metal–carbon composite catalysts. All these cat-
alysts were prepared in a similar manner to the Co catalysts using
nitrate salts. Over 25 catalysts with metal loadings of 1, 3, 5 and
10 wt% metal were synthesized and studied for the ORR behavior.
For brevity, only data with 5 wt% loading are shown in Fig. 7. It is
evident that Co activated samples have the lowest overpotentials
for the electroreduction of O2 to H2O2. The Co-catalysts also show
very high selectivity for H2O2 generation. Thus Co-based catalysts
were chosen for the next step which consisted of durability and
batch analysis in bulk electrolysis cells. For a comparison, ORR mea-
surements were also done with NaOH (pH = 13) electrolyte using
several different high surface area carbon powders. Reduction of
O2 to H2O2 occurs readily on untreated carbons in alkaline media
0.05
0
-0.5
-1
a)
V
Zn
Ni
Sn
Ce
Ba
Fe
Co
Cu
−
through formation of HO2 which forms a chemical equilibrium
with H2O2 [8].
-1.5
-2
O2 + H2O + 2e− → 2HO2− + OH−, E0 = −0.065VSHE
HO2− + H2O ↔ H2O2 + OH−
(8)
(9)
0
0.2
0.4
0.6
Disk Potential vs. RHE [V]
Fig. 8 shows electroanalytical measurements of ORR using sev-
eral different carbon electrodes consisting of the GC tip of the
RRDE, and inks prepared form high surface area Black Pearls
2000 (1450 m2 g−1, Cabot), Vulcan XC 72 (232 m2 g−1 Cabot), and
acetylene black (75 m2 g−1, Chevron Philips). The carbon pow-
der depositions on the RRDE tip amounted to a loading of about
225 mg cm−2. All the carbon powders show low overpotentials (i.e.,
high activity) and excellent selectivity (∼60–80%) for generation of
HO2−. The plain glassy carbon (0.2 cm−2) shows the highest selec-
tivity for H2O2 generation.
Fig. 7. Oxygen electroreduction on different metal catalysts with 5 wt% metal load-
ing. (a) O2 reduction measurements obtained from the rotating disk electrode, (b)
ring current densities, and (c) selectivity of the catalysts for reduction of O2 to H2O2.
The experiments were performed using a catalyst loading of 0.6 mg cm−2 on the
RRDE tip, 0.1 M double distilled HClO4 electrolyte, at 295 K and 1 bar pressure.
of RRDE as a quick and reliable method of electrocatalyst screening
for H2O2 studies. The currents measured were stable (≤10% drop)
for up to 24 h, which also reflects in the nearly linear accumulation
of H2O2 concentration over time. Increasing the catalyst loading
leads to higher production rates with very good current efficien-
cies; indicating that increasing the catalyst loading (at least up to
5 mg cm−2) does not lead to a further reduction of H2O2 to H2O. Fur-
ther optimization of the catalyst loading could even lead to higher
H2O2 yields. The pH of the electrolyte also shows a marked impact
on the H2O2 production. At higher pH values the rates decrease
partly due to lack of H+ and conductivity of the supporting elec-
trolyte.
In these experiments the rate of H2O2 production is limited
by the concentration and transport of O2 to the electrode all of
which occurs in the solution phase. In solution phase, in which our
electrolysis experiments were conducted, the dissolved O2 concen-
tration is about one hundred times smaller than that of gaseous
O2 at atmospheric pressure. In these experiments the typical cur-
rent density was about 1 mA cm−2 which is similar to the RDE
3.2.3. Batch electrolysis
Fig. 9 shows the accumulated H2O2 content in the electrolysis
cell, normalized by the electrode geometric surface area, for differ-
ent catalyst loadings and pH conditions as a function of time. At
pH = 0 and with composite catalyst loadings of 1 and 5 mg cm−2
H2O2 production rates of about 5 and 10 ␮mol h−1 cm−2 were
obtained. Lines and symbols indicate the expected and measured
H2O2 concentrations, respectively. The expected concentrations
are based on the current densities and were obtained using Eq. (2)
with an overall current efficiency of 0.8. The measured H2O2 quan-
tities by UV spectrophotometry are in good agreement with the
expected content at pH = 0. The small discrepancies may stem from
parasitic corrosion current (e.g., the alligator tip) which could be
avoided in a modified cell configuration. Furthermore, the results
at pH = 0 generally confirm the current efficiency ∼0.8 obtained
from the RRDE measurements, hence highlighting the importance

9080
A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
100
80
c)
60
ring potential = 1.2 V vs. RHE
0.1 M NaOH, 293 K, 5 mV s-1
cathodic sweeps, 900 rpm
40
20
0.25
0.2
0.15
0.1
0.05
Glassy Carbon
Acetylene
BP 2000
b)
Vulcan XC72
0
-0.5
-1
a)
-1.5
-2
Fig. 9. H2O2 content in the electrolysis cell normalized by the electrode surface area
as a function of time are shown for (a) pH = 1 (0.05 M H2SO4), (b) pH = 3 (0.5 × 10−3
M
0
0.2
0.4
0.6
0.8
H2SO4), and (c) pH = 7 (using KH2PO4 buffer) electrolytes and for a range of different
catalyst (4 wt% Co) loadings. The experiments were performed using a wide range
of catalyst loadings at 0.25 V vs. RHE, 295 K and 1 bar pressure.
Disk Potential vs. RHE [V]
Fig. 8. Rotating-ring disk electrode measurements of various high surface area car-
bons in alkaline electrolyte. (a) O2 reduction current densities obtained on the
disk electrode, (b) oxidation current densities on the Pt ring current densities, and
(c) selectivity of the catalysts for reduction of O2 to H2O2. The experiments were
performed using a catalyst loading of 0.6 mg cm−2 on the RRDE tip, 0.1 M NaOH
electrolyte, at 295 K and 1 bar pressure.
H2O2 at different pH levels for up to 7 days. For pH ≤ 8 no reduc-
tion in the H2O2 concentration is observed over 7 days. However,
at a pH = 10, a dramatic reduction is observed in the H2O2 concen-
tration. The decomposition of H2O2 here can be modeled by a 1st
experiments with a 400 rpm rotation speed. Our current work is
focused on examining the catalysts in a membrane reactor where
gaseous O2 is used on the cathode electrode and H2 is used at the
anode electrode. A catalyst coated membrane such as Nafion would
provide a highly acidic environment, compact reactor design and
more importantly allows for neutral H2O2 electrogeneration. Very
recently Yamanaka and co-workers have demonstrated production
of neutral H2O2 using O2 reduction and H2 oxidation in a mem-
brane fuel cell reactor [46]. Further research on the catalysts could
include examination of higher surface area carbon (e.g., BP-2000)
and pyrolysis conditions. Non-precious metal ORR catalysts for fuel
cells with BP-2000, for instance, have generally shown higher activ-
ity for reduction of oxygen [12,47]. Furthermore, impact of reactive
atmosphere (e.g., NH3) for incorporation of nitrogen could be exam-
ined as well.
3.3. H2O2 stability studies
H2O2 stability was examined over a range of different pH at
295 K. Fig. 10 shows stability results for commercially available
Fig. 10. Chemical stability of H2O2 in different pH media as a function of time at
295 K.

A. Bonakdarpour et al. / Electrochimica Acta 56 (2011) 9074–9081
9081
order rate law:
d[H2O2]
[9] F. Jaouen, J.P. Dodelet, Electrochimica Acta 52 (2007) 5975.
[10] J.P. Dodelet, Oxygen reduction in PEM fuel cell conditions: heat-treated
non-precious metal-N4 macrocycles and beyond, in: N4-Macrocyclic Metal
Complexes, Springer, New York, 2006.
[11] E.B. Easton, A. Bonakdarpour, R. Yang, D.A. Stevens, J.R. Dahn, Journal of the
Electrochemical Society 155 (2008) B547.
= −k[H2O2]n
(10)
dt
where the
k
and
n
parameters were found to be to be
0.027 day−1 ppm−0.76 and 1.76, respectively. These studies are very
important in choosing the right medium and designing a reactor
for the electrosynthesis of H2O2. For example in trickle-bed reac-
tors, high alkalinity, high temperatures and exposure of H2O2 to the
anode all lead to loss of H2O2 [3]. Several loss factors can be averted
by using an acidic H+ conducting Nafion membrane in conjunction
with an appropriate catalyst like the Co-catalysts discussed here or
elsewhere. This allows operation in a more stable range of pH and
prevents contact of the H2O2 with the anode.
[12] M. Lefevre, E. Proietti, F. Jaouen, J.P. Dodelet, Science 324 (2009) 71.
[13] P.H. Matter, U.S. Ozkan, Catalysis Letters 109 (2006) 115.
[14] S. Marcotte, D. Villers, N. Guillet, L. Roue, J.P. Dodelet, Electrochimica Acta 50
(2004) 179.
[15] T.S. Olson, S. Pylypenko, J.E. Fulghum, P. Atanassov, Journal of the Electrochem-
ical Society 157 (2010) B54.
[16] N. Guillet, L. Roue, S. Marcotte, D. Villers, J.P. Dodelet, N. Chhim, S. Trevin, Journal
of Applied Electrochemistry 36 (2006) 863.
[17] W. Zhang, A.U. Shaikh, E.Y. Tsui, T.M. Swager, Chemistry of Materials 21 (2009)
3234.
[18] I. Yamanaka, T. Onizawa, H. Suzuki, N. Hanaizumi, K. Otsuka, Chemistry Letters
35 (2006) 1330.
[19] R.Z. Yang, K. Stevens, A. Bonakdarpour, J.R. Dahn, Journal of the Electrochemical
Society 154 (2007) B893.
[20] T.S. Olson, S. Pylypenko, P. Atanassov, K. Asazawa, K. Yamada, H. Tanaka, Journal
of Physical Chemistry C 114 (2010) 5049.
4. Conclusions
Highly active Co-based catalysts for electroreduction of O2 to
H2O2 in acidic media were synthesized using a simple procedure
and inexpensive Co nitrate salts. Catalysts with a nominal 4 wt%
Co content showed the highest activity and selectivity for elec-
trogeneration of H2O2. Batch electrolysis cells using dissolved O2
in the solution phase showed no degradation for up to 24 h and
current efficiencies of up to 85% were achieved. A study of other
metals (Ni, Cu, Zn, Ba, Ce) with various loadings was performed
but none reached the electrocatalytic performance of Co-activated
catalysts. H2O2 examined in acidic media showed excellent sta-
bility for up to seven days at ambient conditions. Further work
is now focused on implementing a membrane-based electrolytic
cell using the catalysts synthesized here specifically for integration
with water disinfecting UV reactors for remote communities.
[21] J.C. Forti, C.E. Venancio, M.R.V. Lanza, R. Bertazzoli, Journal of the Brazilian
Chemical Society 19 (2008) 643.
[22] E. Brillas, F. Alcaide, P.L. Cabot, Electrochimica Acta 48 (2002) 331.
[23] I. Yamanaka, Catalysis Surveys from Asia 12 (2008) 78.
[24] C.W. Jones, Applications of Hydrogen Peroxide and Derivatives, Royal Society
of Chemistry, London, 1999.
[25] G. Strukul, Catalytic Oxidations with Hydrogen Peroxide as Oxidant, Kluwer
Academic Publisher, London, 1992.
[26] M.N. Chong, B. Jin, C.W.K. Chow, C. Saint, Water Research 44 (2010) 2997.
[27] B. Kasprzyk-Hordern, M. Ziolek, J. Nawrocki, Applied Catalysis B-Environmental
46 (2003) 639.
[28] O. Legrini, E. Oliveros, A.M. Braun, Chemical Reviews 93 (1993) 671.
[29] B.L. Loeb, Ozone-Science & Engineering 31 (2009) 379.
[30] N.N. Mahamuni, Y.G. Adewuyi, Ultrasonics Sonochemistry 17 (2010) 990.
[31] A. Matilainen, M. Sillanpaa, Chemosphere 80 (2010) 351.
[32] T. Oppenlander, Photochemical Purification of Water and Air, Wiley-VCH Ver-
lag, Weinheim, 2003.
[33] S. Parsons, Advanced Oxidation Processes for Water and Wastewater Treat-
ment, IWA Publishing, London, 2004.
[34] P.C. Vandevivere, R. Bianchi, W. Verstraete, Journal of Chemical Technology and
Biotechnology 72 (1998) 289.
Acknowledgements
[35] R. Venkatadri, R.W. Peters, Hazardous Waste & Hazardous Materials 10 (1993)
107.
[36] S.D. Richardson, Journal of Environmental Monitoring 4 (2002) 1.
[37] Y.T. Woo, D. Lai, J.L. McLain, M.K. Manibusan, V. Dellarco, Environmental Health
Perspectives 110 (2002) 75.
[38] E. Neyens, J. Baeyens, Journal of Hazardous Materials 98 (2003) 33.
[39] A. Alvarez-Gallegos, D. Pletcher, Electrochimica Acta 44 (1998) 853.
[40] A. Alverez-Gallegos, D. Pletcher, Electrochimica Acta 44 (1999) 2483.
[41] A. Da Pozzo, E. Petrucci, C. Merli, Journal of Applied Electrochemistry 38 (2008)
997.
The authors would like to acknowledge the financial support of
NSERC through the Res’eau WaterNET program. Arman Bonakdar-
pour likes to acknowledge partial funding from the NSERC PDF
program. Daniel Esau and Hillary Cheng would like to acknowledge
financial assistance through NSERC USRA undergraduate award
program.
[42] A. Bonakdarpour, M. Lefevre, R.Z. Yang, F. Jaouen, T. Dahn, J.P. Dodelet, J.R. Dahn,
Electrochemical and Solid State Letters 11 (2008) B105.
[43] A. Bonakdarpour, T.R. Dahn, R.T. Atanasoski, M.K. Debe, J.R. Dahn, Electrochem-
ical and Solid State Letters 11 (2008) B208.
[44] T.J. Schmidt, H.A. Gasteiger, Rotating thin-film method for supported catalysts,
in: W. Vielstich, A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells: Fun-
damentals, Technology and Applications, vol. 2, John Wiley & Sons, Chichester,
2003, p. 316.
[45] N.V. Klassen, D. Marchington, H.C.E. McGowan, Analytical Chemistry 66 (1994)
2921.
[46] I. Yamanaka, T. Murayama, Angewandte Chemie-International Edition 47
(2008) 1900.
References
[1] C. Samanta, Applied Catalysis A-General 350 (2008) 133.
[2] J.M. Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Angewandte Chemie-
International Edition 45 (2006) 6962.
[3] C.W. Oloman, Electrochemical Processing for the Pulp & Paper Industry, The
Electrochemical Consultancy, New York, 1996.
[4] C. Oloman, A.P. Watkinson, Journal of Applied Electrochemistry 9 (1979) 117.
[5] N. Gupta, C.W. Oloman, Journal of Applied Electrochemistry 36 (2006) 255.
[6] N. Gupta, C.W. Oloman, Journal of Applied Electrochemistry 36 (2006) 1133.
[7] E.L. Gyenge, C.W. Oloman, Journal of the Electrochemical Society 152 (2005)
D42.
[47] A.D. Pauric, A.W. Pedersen, T. Andrusiak, E.B. Easton, Journal of the Electro-
chemical Society 157 (2010) B370.
[8] K. Kinoshita, Oxygen Electrochemical Technology, 1st ed., Wiley-Interscience,
1992.