You are on page 1of 7

Electrochimica Acta 56 (2011) 7659–7665

Contents lists available at ScienceDirect

Electrochimica Acta
journal homepage: www.elsevier.com/locate/electacta

Functionalized-carbon nanotube supported electrocatalysts and


buckypaper-based biocathodes for glucose fuel cell applications
L. Hussein a,b,∗ , Y.J. Feng c,1 , N. Alonso-Vante c,1 , G. Urban a,b , M. Krüger a,b,∗∗
a
Freiburg Materials Research Centre (FMF), Laboratory for Nanosciences, University of Freiburg, Freiburg, Germany
b
Institute for Microsystems Technology (IMTEK), Laboratory for Sensors, University of Freiburg, Freiburg, Germany
c
Laboratory of Electrocatalysis, UMR-CNRS 6503, University of Poitiers, Poitiers, France

a r t i c l e i n f o a b s t r a c t

Article history: The preparation and testing for electrocatalytic activity of functionalized carbon nanotube (f-CNT) sup-
Received 3 June 2011 ported Pt and Au–Pt nanoparticles (NPs), and bilirubin oxidase (BOD), are reported. These materials were
Accepted 21 June 2011 utilized as oxygen reduction reaction (ORR) cathode electrocatalysts in a phosphate buffer solution (0.2 M,
Available online 30 June 2011
pH 7.4) at 25 ◦ C, in the absence and presence of glucose. Carbon monoxide (CO) stripping voltammetry
was applied to determine the electrochemically active surface area (ESA). The ORR performance of the
Keywords:
Pt/f-CNTs catalyst was high (specific activity of 80.9 ␮A cmPt −2 at 0.8 V vs. RHE) with an open circuit
Biofuel cell
potential within ca. 10 mV of that delivered by state-of-the-art carbon supported platinum catalyst and
Oxygen reduction reaction (ORR)
Glucose
exhibited better glucose tolerance. The f-CNT support favors a higher electrocatalytic activity of BOD for
Functionalized carbon nanotubes (f-CNTs) the ORR than a commercially available carbon black (Vulcan XC-72R). These results demonstrate that
Nanocatalysts f-CNTs are a promising electrocatalyst supporting substrate for biofuel cell applications.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction the devices expensive and unsuitable for large scale applications.
Replacing or reducing the amount of Pt, while maintaining the cat-
Glucose is the most abundant monosaccharide. As a most alytic activity is an important goal for developing novel catalysts
important energy source (biofuel) for low-temperature, one- [4,5] and new supporting substrates [6]. In contrast to conven-
compartment biofuel cells, it has been the subject of many studies tional direct fuel cells, the challenge to establish an implantable
[1]. Particular attention has been paid to the development of elec- direct biofuel cell is to run the cell under physiological and mixed-
trocatalysts for mixed-reactant fuel cells, especially to the oxygen reactant conditions. The glucose fuel cell, as a mixed-reactant fuel
reduction reaction (ORR) process in relation to glucose tolerance cell, is considered to be an attractive power source for implantable
of the catalysts. Large overpotential losses, due to mixed poten- medical devices [7]. Although this design is simpler than a con-
tials resulting from mixed reactants and slow reaction kinetics, ventional one, the electrocatalysts need to be highly active and
usually occur. One possibility for overcoming these difficulties is resistant to the presence of other reactants (e.g. oxygen or glu-
to conceive new efficient and selective electrocatalysts for oxygen cose) under physiological conditions. The desired product of the
reduction in the presence of glucose. Several ORR electrocatalysts oxygen reduction at the cathode is water rather than hydrogen
based on Pt and oxophillic metals such as Pd, Co, Ni, Fe, Cu and Ag peroxide since the later results from low active ORR catalysts, and
have been investigated [2]. accelerates the corrosion of the electrode material and polymer
The overall electrocatalytic processes at electrodes are rather membrane [8]. With non-tolerant cathode catalysts, the glucose
complex and involve a number of adsorbed intermediates and would depolarize the cathode due to electrooxidation at the elec-
by-products [3]. Electrocatalysts based on Pt or its alloys make trode, resulting in direct electron transfer to oxygen. This results
in an internal current flow and a reduced voltage compared to the
thermodynamically expected potential of 1.23 V vs. RHE (Reversible
∗ Corresponding author at: Institute for Microsystems Technology (IMTEK), Labo- Hydrogen Electrode) for the ORR [9].
ratory for Sensors, University of Freiburg, Freiburg, Germany. Tel.: +49 7612 034781; Moreover, in order to minimize the additional ohmic drop, as
fax: +49 7612 037262. well as mass transport limitation and manufacturing problems
∗∗ Corresponding author at: Freiburg Materials Research Centre (FMF), Laboratory deriving from use of thick electrodes, catalysts for glucose fuel cells
for Nanosciences, University of Freiburg, Freiburg, Germany. Tel: +49 7612 034755; are usually based on non-supported active noble metals. It is well
fax: +49 7612 034701.
E-mail addresses: laith.hussein@fmf.uni-freiburg.de (L. Hussein),
known that the occurrence of catalyst agglomeration limits their
michael.krueger@fmf.uni-freiburg.de (M. Krüger). effectiveness and utilization in fuel cell systems [10]. In polymer
1
ISE member. electrolyte membrane fuel cells (PEMFCs), carbon black (Vulcan

0013-4686/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.electacta.2011.06.067
7660 L. Hussein et al. / Electrochimica Acta 56 (2011) 7659–7665

XC-72) is normally used as support material for Pt nanocatalyst


to enhance their utilization. However, the degradation of carbon
minimizes the long-term stability of the fuel cells [11]. Recently,
carbon nanotubes (CNTs) have been explored as promising sup-
port material for fuel cell catalysts due to their high stability and
unique mechanical properties [11,12]. The high amount of meso-
pores (2–50 nm in diameter) and the large specific surface area of
CNTs can significantly increase the dispersion of metal nanoparti-
cles and thus enhance their utilization and electrocatalytic activity.
The mesoporous network of CNTs additionally facilitates the immo-
bilization process and combined with their excellent electrical
conductivity, enables an improved charge transport as well as easy
accessibility of the reagent molecules to the catalytic sites. The
consequence is that CNT supported nanocatalysts show a better cat-
alytic performance than carbon black supported catalysts [12]. It is
also expected that CNT supported electrocatalysts possess a better Fig. 1. Scheme of the buckypaper fabrication process based on the vacuum filtration
long-term stability compared to carbon black supported ones. technique.
Novel buckypaper (BP) fabricated from commercially available
multi-walled carbon nanotubes (MWCNTs) is used in this work as MWCNTs were dispersed in 200 mL of an aqueous solution contain-
electrode support for glucose fuel cell applications. BP is a self- ing 1 wt.% Triton X-100 (Sigma–Aldrich) under mechanical stirring
supporting mat of entangled assemblies (ropes and bundles) of for 30 min followed by ultrasound treatment for 3 h. The resulting
MWCNTs which form membrane-type black films. It was origi- suspension was centrifuged for 15 min using a Sigma 2–5 cen-
nally developed to more easily handle carbon nanotubes [11]. BP trifuge at 2700 rpm to remove larger agglomerates of MWCNTs.
is a highly porous, mechanically stable and electrically conductive The supernatant, containing a stable CNT-suspension, was then fil-
material with various advantages over other carbon nanotube films tered through a nylon membrane filter (0.45 ␮m pore size) and
[13–15]. Some physical properties of BP have been investigated for compressed under vacuum by an oil-free diaphragm pump (KNF
potential applications in supercapacitors [16], electromechanical Neuberger, Germany). The obtained homogeneous black film was
actuators [17], artificial muscles [18], field emitters [19], transpar- washed repeatedly with an excess of deionised water, followed by
ent and conductive substrates [20], porous membranes [13,15,21], isopropyl alcohol and acetone. The prepared CNT-films were kept
and thin protective layers for electromagnetic shielding [22]. More- at room temperature for 30 min and then dried in a vacuum oven
over, its use in fuel cell [23,24], biofuel cell electrodes [25,26] and at 50 ◦ C overnight.
sensor/actuator devices [27], has just started.
Furthermore, it is reported that surface defects and functional- 2.3. Preparation of electrocatalysts and BP-based biocathodes
ization under strongly oxidizing conditions can be used effectively
as anchors for metal species [28]. The f-CNT supported Pt NPs and Au70 Pt30 NPs (nominal Au:Pt
Here we investigate the preparation and the electrochemical molar ratio of 70:30) were prepared via a water in oil (W/O)
activity of functionalized multi-walled carbon nanotube (f-CNT) microemulsion route with a metal mass loading of 40 wt.%, follow-
supported Pt and bimetallic Au–Pt nanoparticles (NPs) respectively ing the procedure reported by Habrioux et al. [29].
as potential electrode materials for the cathodic oxygen reduction The BP-supported BOD catalysts were prepared according the
reaction (ORR) in a glucose fuel cell. A direct comparison is made scheme illustrated in Fig. 2. BP pieces (1.0 cm × 2.0 cm) were cut
with conventional carbon black (Vulcan XC-72R) as substrate since out of a 15 ␮m thick BP film. Bilirubin oxidase (BOD, Amano-
it is widely used as a catalyst support material. Additionally, CNT-BP 3 [EC 1.3.3.5], activity 2.44 U mg−1 ) from Myrothecium verrucaria
has been utilized as electrode material. Both, the f-CNT and Vulcan was obtained from Amano Pharmaceutical Co. (Japan) and as
XC-72R supports are used to immobilize the system consisting of electron-transfer mediator, 2,2 -azinobis(3-ethylbenzothiazoline-
bilirubin oxidase and electron-transfer mediator in order to test the 6-sulfonate) diammonium salt (ABTS2− ) from Sigma–Aldrich was
electrocatalytic activity of the BP-based electrodes towards the ORR used.
for biofuel cell applications in the presence of glucose at neutral Typically, a BP biocathode based on f-CNTs was prepared using
medium (pH 7.4). the biocatalyst inks consisting of 3 mg f-CNTs, 3 mg of BOD, 12 vol.%
Nafion solution (Nafion® , 5 wt.%, Sigma–Aldrich) and 0.5 mM of
2. Materials and methods
ABTS2− , in 0.1 M phosphate buffer solution using Millipore-Q water
2.1. Functionalization of carbon nanotubes (18.2 M cm). A good dispersion of f-CNTs has been ensured with
the assistance of ultrasonic treatment for 15 min. Afterwards, the
Commercial MWCNTs (Baytubes C 150-HP, Bayer Material resulting ink was pipetted directly onto the BP to form BOD-
Science AG, Germany) were functionalised by harsh acid treatment ABTS/f-CNT-BP with a total enzyme loading of 175 ␮g cm−2 . For
(concentrated nitric acid, 65%) under reflux at 140 ◦ C for 3 h as direct comparison with f-CNTs, carbon black (Vulcan XC-72R, Cabot
previously reported [26]. The f-CNTs were collected by centrifuga- GmbH, Germany) was also used as BOD support to form BOD-
tion, re-dispersed in de-ionized water, washed thoroughly over a ABTS/C-BP catalysts, with the same enzyme loading. Afterwards the
nylon membrane filter with pore sizes of 0.45 ␮m (Whatman, UK) BP-based electrodes were cured and dried under nitrogen at room
during constant vacuum filtration, until the filtrate was neutral. temperature for 2 h. Finally, the electrodes were tested directly
Afterwards the f-CNT-cake was dried in a vacuum oven at 50 ◦ C after washing with a phosphate buffer solution (0.1 M, pH 7.4).
overnight.
2.4. Characterization of f-CNT supported catalysts and
2.2. BP fabrication buckypaper

The BP fabrication was performed as reported in a previous Transmission electron microscopy (TEM) images of f-CNT sup-
work [26] and is depicted in Fig. 1. In short, 100 mg of as-received ported and Vulcan XC-72R supported Pt and Au–Pt nanocatalysts
L. Hussein et al. / Electrochimica Acta 56 (2011) 7659–7665 7661

Fig. 2. Schematic representation of the preparation of BP-based biocathodes. The BOD-ABTS system is immobilized either on (a) carbon black (Vulcan XC-72R) or (b) f-CNTs.

were recorded on a Zeiss LEO 912 Omega microscope with an SCE against a commercial RHE (Hydroflex, Gaskatel GmbH, Kassel,
accelerating voltage of 120 kV. Nitrogen adsorption/desorption Germany).
isotherms were conducted at −196 ◦ C on an automatic analyzer All the linear-sweep voltammetries were conducted at a rotat-
(Sorptomatic 1990, Porotec, GmbH). Prior to the experiment, the ing speed of 2500 rpm using a potential scan rate of 5 mV s−1
samples were degassed by heating at 250 ◦ C for 5 h under high vac- in O2 -saturated phosphate buffer solution (0.2 M, pH 7.4) in the
uum. The texture properties were analyzed using advanced data absence and the presence of glucose (0.2 M). The obtained current-
processing software (ADP version 5.1, Thermo Electron Corpora- potential curves were recorded at a potential scan rate of 50 or
tion). 20 mV s−1 , after performing 20 or 5 potential cycles in the absence
The specific surface areas (given in m2 g−1 ) of carbon black and the presence of glucose respectively, in a nitrogen-purged elec-
and f-CNT powder samples were evaluated from the linear part trolyte which cleaned and activated the electrode surface. For the
of the BET adsorption plot of N2 , according to the standard testing of the BP-based biocathodes, the electrodes were mounted
Brunauer–Emmett–Teller (BET) method, at a relative pressure in a home-made single-cell test fixture, as described in literature
range of 0.05–0.3 p/p0 [30]. The mean diameter of mesopores [26].
(2–50 nm), the mesoporous surface area (given in m2 g−1 ) and the
volume (given cm3 g−1 ) were deduced from the desorption branch
2.5. Potentiodynamic CO stripping
of the isotherm relative pressure in the range of 0.4–0.999 p/p0
using a model developed by Barrett, Joyner and Halenda (BJH) [30].
The electrochemically active surface areas (ESA) of the elec-
The X-ray diffraction (XRD) measurements were carried out on
trocatalysts were obtained using the electrochemical carbon
a powder diffractometer (Siemens D-5000) with a Cu K␣ radiation
monoxide (CO)-stripping of an adsorbed monolayer [31]. Initially,
( = 1.5406 Å) at 40 kV and 30 mA using 2 scanning from 30◦ to 90◦
N2 gas was purged into a phosphate buffer solution (0.2 M, pH
with a step size of 0.015◦ per 1.5 s to ensue obtaining fine crystalline
7.4) for 30 min. Thereafter, 20 potential cycles were performed at
structure of metal nanoparticles.
50 mV s−1 between 0.1 and 1.2 V vs. RHE. The solution was then
The electrical conductivity measurements of BPs were carried
bubbled with CO (99.9% purity) for 5 min. It was then subjected to
out on a four-point probe instrument (QuadPro resistivity system,
a bias of 0.1 V vs. RHE for further 3 min. Finally, the dissolved CO was
Lucas Signatone) and the thicknesses of BPs were measured using
removed by bubbling N2 gas again into the solution for 30 min. The
a micrometer gauge (resolution 1 ␮m, Mitutoyo).
potential was cycled at 5 mV s−1 between 0.1 and 1.2 V vs. RHE for
All potentiodynamic measurements were performed in a con-
two complete oxidation and reduction cycles. The ESA was deter-
ventional thermostated three-electrode electrochemical cell at
mined by calculating the charge corresponding to the CO oxidation
25 ◦ C, using a potentiostat/galvanostat apparatus (Autolab PGSTAT-
peak. By assuming a nearly monolayer of adsorbed CO, a faradaic
30). The working electrode was a 3.0 mm diameter (0.07 cm2 )
charge of 420 ␮C cm−2 was used for the calculation [31].
glassy carbon disk, successively polished with Al2 O3 powder (5 Å)
using emery paper prior to the deposition of the catalyst ink.
The catalyst ink was prepared by dispersing 10 mg of the cat- 3. Results and discussion
alyst in a 200 ␮L Nafion solution and 1200 ␮L Millipore-Q water
(18.2 M cm) in an ultrasonic bath for 2 h. The catalyst ink (1.8 ␮L 3.1. Structure and morphology of nanocatalysts and buckypaper
in the case of 40 wt.% CNT supported Pt NPs or Au–Pt NPs cat-
alysts, and 3.6 ␮L in case of 20 wt.% Pt/C (E-TEK) catalyst), was Fig. 3 shows transmission electron microscopy (TEM) images
deposited onto the glassy carbon disk to obtain a metal mass of (A) f-CNT supported (40 wt.%) Pt-NPs, (B) Vulcan XC-72R sup-
loading of 72.7 ␮g cm−2 and then dried under nitrogen at room ported 20 wt.% Pt-NPs (E-TEK), and (C) f-CNT supported (40 wt.%)
temperature for 2 h. A KCl saturated calomel electrode (SCE) and Au–Pt NPs. In general, the nanoparticles were well-dispersed on
a glassy carbon plate (1 cm2 ) were used as reference and counter the f-CNT as for Vulcan XC-72R, with an average particles size of
electrodes, respectively. The reference electrode was separated 5.01 nm, 4.60 nm and 5.09 nm respectively and measured by TEM
from the electrochemical cell by a Luggin capillary. All the poten- image analysis using iTEM Desktop software (Olympus Soft Imag-
tials are quoted vs. RHE by directly measuring the potential of the ing Solutions GmbH, Münster, Germany).
7662 L. Hussein et al. / Electrochimica Acta 56 (2011) 7659–7665

The N2 -physisorption measurement showed that f-CNTs have


BET surface areas (ABET ) of ca. 279 m2 g−1 , median pore sizes of
54.50 nm and average pore sizes of 30.40 nm (calculated using
the equation: 4VBJH /ABET ) compared to ca. 232 m2 g−1 , 35.39 nm
and 9.53 nm respectively, for Vulcan XC-72R. Moreover, significant
increases in the BJH-mesopore area (ABJH ) of 240.4 m2 g−1 and BJH-
mesopore volume (VBJH ) of 2.2 cm3 g−1 have been found for f-CNTs,
in comparison to 75.77 m2 g−1 and 0.552 cm3 g−1 respectively, for
Vulcan XC-72R [26].
Additionally, a four-point probe measurement of the electrical
conductivity indicated that the BP based on as-received CNTs (Bay-
tubes) was highly conducting with conductivities of ca. 25 S cm−1 ,
in contrast to carbon black with a reported conductivity value of
ca. 5 S cm−1 [32].

3.2. Determination of the electrochemically active surface area

In order to normalize the current densities of our electro-


catalysts to their Pt content, the ESAs for the all Pt-based
electrocatalysts were investigated. The charge associated with the
adsorption of a monolayer of hydrogen atoms to the surfaces of
Pt atoms is 210 ␮C cm−2 [33]. Green and Kucernak [34] found that
hydrogen adsorption and stripping is not suitable for the assess-
ment of the active surface area of many alloy catalysts, since the
charges associated with H and other oxidation processes overlap
(i.e. occurrence of a flat shoulder peak). The hydrogen adsorption
process is more complex either due to the absorption of H into the
metal lattice or the redox behavior of surface-active groups, which
mask the platinum hydrogen adsorption–desorption characteris-
tics. Thus, the electro-oxidation of an adsorbed monolayer of CO is
used as the preferred method for the in situ measurement of the
surface area of electrocatalysts [31]. It is generally assumed that
CO is bonded linearly to a single metal surface atom, and there-
fore a 1:1 ratio between the number of adsorbed CO molecules
and metal surface atoms exists. Therefore, the active surface areas
for all catalysts were obtained by CO stripping experiments. It is
known that the shape of the CO-stripping peak depends on the
surface nature of the catalysts. In the case of Pt, the oxidation
wave of the adsorbed CO monolayer centers at 0.75 V vs. RHE. ESAs
for Au–Pt/f-CNT, Pt/C (E-TEK) and Pt/f-CNT catalysts were deter-
mined (Figs. S1, S2 and S3 in Supplementary data). The resulting
ESA values based on Pt loadings of the above mentioned catalysts
are: 48.60, 71.50 and 10.56 m2 g−1 , respectively. The important dif-
ference of these values between Pt/C (E-TEK) and Pt/f-CNT can be
due to differences in average particle size and/or due to agglom-
eration of nanoparticles [35] (see Fig. 3a and b). The oxidation
of adsorbed carbon monoxide usually takes place at more nega-
tive potentials (E ca. −0.1 V) on Pt/f-CNTs than on Pt/C (E-TEK)
(Figs. S2 and S3 in Supplementary data). This phenomenon can
be attributed to a substrate effect, which has been observed on
Pt deposited onto oxide sites [36]. The reason for this effect is
not clear. However, on the Au–Pt bimetallic nanoparticales, the
broad CO oxidation wave (from 0.68 to 1.2 V vs. RHE) reveals the
complex interplay of the Au and Pt atoms on these alloy nanopar-
ticles (Fig. S1 in Supplementary data). Moreover, the crystalline
structure of these nanoalloys was characterized by XRD measure-
ment (Fig. S4 in Supplementary data). From the XRD pattern, one
deduces that the peaks at 2-theta (38.7◦ , 44.8◦ , 65.4◦ , 78.4◦ ), can be
Fig. 3. Transmission electron microscopy (TEM) images of (a) f-CNTs supported Pt attributed to Au–Pt nanoalloy structure rather than the mixture of
NPs (40 wt.%), (b) carbon black Vulcan XC-72R supported Pt NPs (E-TEK, 20 wt.%) and individual nanoparticles [29,37,39].
(c) f-CNTs supported Au–Pt NPs (40 wt.%).
3.3. Surface electrochemistry of electrocatalysts

Fig. 4a shows cyclic voltammetry curves of commercially avail-


able 20 wt.% Pt/C (E-TEK), and of home-made 40 wt.% Pt/f-CNT
and 40 wt.% Au70 Pt30 /f-CNT catalysts after electrochemical stabi-
L. Hussein et al. / Electrochimica Acta 56 (2011) 7659–7665 7663

Fig. 5. ORR curves at the 10th potential cycle of (a) 20 wt.% Pt/C (E-TEK), 40 wt.%
Fig. 4. Cyclic voltammetry curves (CVs) of (a) 20 wt.% Pt/C (E-TEK), 40 wt.% Pt/f-CNTs
Pt/f-CNTs and 40 wt.% Au70 Pt30 /f-CNTs in O2 -saturated phosphate buffer solution
and 40 wt.% Au70 Pt30 /f-CNTs catalysts at the 20th potential cycle in a N2 -purged
(0.2 M, pH 7.4) in the absence (solid curves) and presence (dotted curves) of 0.2 M
phosphate buffer solution (0.2 M, pH 7.4), in the absence of 0.2 M glucose, (b) the
glucose, (b) the corresponding Tafel plots of the same ORR curves. A sweep rate of
CVs at the 5th potential cycle in the presence of 0.2 M glucose were performed. A
5 mV s−1 at a rotating speed of 2500 rpm, was applied at 25 ◦ C.
sweep rate of 5 mV s−1 was applied at 25 ◦ C.

lization. These curves were usually achieved, after 20 potential that glucose oxidation is still carried out on the Au-oxide/Au–OH
cycles at a sweep rate of 20 mV s−1 in N2 -purged phosphate buffer surface species which is in synergy with the Pt-oxide/Pt–OH species
solution (0.2 M, pH 7.4). Characteristic peaks for the hydrogen [29] and can be attributed to the alloying and “bifunctional” or
adsorption/desorption emerge between 0.1 and 0.35 V vs. RHE. As “ligand” effects [39]. Moreover the multiple anodic peaks can be
expected, these waves are strongly attenuated in case of the Au–Pt attributed to the oxidation of glucose and resulting intermediates,
alloy. The double layer region in the present electrolyte medium as observed during the positive-going scans on the Au–Pt/f-CNTs
is not as well defined as in KOH medium, which was observed by catalyst for the potential region between 0.15 and 1.25 V vs. RHE.
Hsueh et al. [38]. For all investigated catalysts, the reduction peak The activity of all catalysts is well visualized when it takes into
of the Pt-oxide species is centered at ca. 0.7 V vs. RHE. It can be seen consideration the current with respect to the ESA of Pt content,
that on Au–Pt/f-CNTs this process is also affected by the presence see Fig. 4b. Summing up, nanocatalysts (Pt/f-CNT and Pt/C (E-TEK))
of gold and the peak is shifted to 0.6 V vs. RHE. are less active towards glucose electro-oxidation as compared to
The curves have been recorded at 25 ◦ C after 5 initial poten- Au–Pt/f-CNTs catalyst.
tial cycles at a potential scan rate of 20 mV s−1 . Fig. 4 b shows the
faradaic oxidation currents for glucose (0.2 M) oxidation for the var-
ious catalysts. For the Pt/C (E-TEK) catalyst the oxidation wave on 3.4. Electrochemical oxygen reduction reaction – ORR
the positive-going potential is limited between the onset poten-
tial value (ca. 0.25 V) and 1.0 V vs. RHE, due to the formation of To test the degree of selectivity (tolerance) towards ORR and/or
Pt-oxide species. This oxidation process is recovered again in the the glucose oxidation reaction, carbon supported Pt and Au–Pt
negative-going scan. In contrast to this, glucose oxidation at the nanocatalysts were measured using the rotating disk electrode
Au–Pt catalyst occurs at 0.15 V indicating that either at Au sites (RDE) technique. Fig. 5a shows the results for Pt/C (E-TEK), Pt/f-CNT,
oxygenated species formed at low potentials, or a reduced overpo- and Au–Pt/f-CNT catalysts in O2 -saturated phosphate buffer solu-
tential occurs for the glucose oxidation process (as expected). The tion (0.2 M, pH 7.4) in the absence (solid curves) and the presence
oxidation potential is sustained up to 1.25 V vs. RHE due to the fact (dotted curves) of 0.2 M glucose at 25 ◦ C.
7664 L. Hussein et al. / Electrochimica Acta 56 (2011) 7659–7665

Table 1
Specific activity of the ORR electrocatalysts (total metal NPs loading 72.7 ␮g cm−2 ) in
a phosphate buffer solution (0.2 M, pH 7.4), containing 0.2 M glucose at 25 ◦ C using
rotation speed of 2500 rpm.

Nanocatalyst Onset potential ESA (m2 gPt −1 ) Specific activity


(V vs. RHE) at 0.8 V vs. RHE
(␮A cmPt −2 )

Pt/f-CNTs 0.89 10.56 80.9


Pt/C (E-TEK) 0.90 71.50 14.6
Au70 Pt30 /f-CNTs 0.54 48.60 NA

In the absence of glucose, both carbon supported Pt catalysts


(either on Vulcan XC-72R or f-CNTs) show similar kinetics, as
judged by the shape of the Tafel plots (see Fig. 5b). A similar
trend is also observed for Au–Pt/f-CNT in this activation region.
The differences between the Au–Pt bimetallic and Pt monometal-
lic based catalysts appear at higher overpotential. Indeed, the ORR
leading to water (4-electron charge transfer) is less pronounced
Fig. 6. ORR curves at the second potential cycle of (a) BOD-ABTS/C-BP and (b) BOD-
on the Au–Pt nanoalloy, since the current in the diffusion region
ABTS/f-CNT-BP, in O2 -saturated phosphate buffer solution (0.2 M, pH 7.4) in the
is ca. 88% of the diffusion current attained by Pt catalysts. This presence of 10 mM glucose without any rotating of the electrode. A sweep rate of
decrease can be attributed to the hydrogen peroxide formation. 3 mV s−1 was applied.
Furthermore, it clearly appears that in the activation region, the
substrate and the nature of the catalyst influence the overpoten-
formance may result from a better electrical conductivity of the
tial. The overpotential is decreasing according to following order:
support, lower impurities, a larger number of mesopores and a
Au–Pt/f-CNT > Pt/f-CNT > Pt/C (E-TEK). On the other hand, in the
higher total BJH-mesopore surface area of ca. 240 m2 g−1 for f-CNT
presence of 0.2 M glucose, both Pt monometallic based catalysts
which is three-fold larger than for Vulcan XC-72R (ca. 76 m2 g−1 ).
display a high tolerance towards glucose, since only a depolariza-
Even if the later possesses a similar value range for the overall spe-
tion of ca. −74 mV or −94 mV for Pt/f-CNT and Pt/C (E-TEK) can be
cific surface area as f-CNT, not all of the pores seem to be effective
observed, respectively (Fig. 5b). However, a limiting current den-
for mass transport and accessible for enzyme molecules because
sity on Pt catalysts is more sustained on f-CNTs than on Vulcan
the majority of its pores are micropores (less than 2 nm in diameter)
XC-72R. Furthermore, as expected, on Au–Pt/f-CNT the depolariza-
[26].
tion is evident (ca. −400 mV). The lower limiting current density
on Pt/C (E-TEK) can be attributed to a lower accessibility of oxygen
4. Conclusions
in the presence of glucose.
Furthermore, the results summarized in Table 1 clearly demon-
We have investigated the substrate effect of functionalized
strate that the Pt/f-CNTs electrocatalyst has a high electrocatalytic
multi-walled carbon nanotube (f-CNT) as supporting substrate for
activity (80.9 ␮A cmPt −2 at 0.8 V vs. RHE) for the ORR, exhibits
Pt and Au–Pt NPs and bilirubin oxidase (BOD) catalyst on the elec-
glucose-tolerance, and is stable during repetitive potential cycles
trocatalytic activity for the ORR in a phosphate buffer solution
as compared to the Pt/C (E-TEK) electrocatalyst in a phosphate
(0.2 M, pH 7.4), in the absence and presence of glucose at 25 ◦ C. In
buffer solution (0.2 M, pH 7.4) containing 0.2 M glucose. This can
comparison with carbon black (Vulcan XC-72R), the f-CNT substrate
be ascribed to the good dispersion and high utilization of Pt-NPs
shows an effective promotion for BOD towards ORR and it favors
on f-CNTs due to effective mass transport assisted by the pres-
the selective enhancement for the ORR in the presence of glucose.
ence of large numbers of mesoporous in f-CNTs, which enhance
The new Pt/f-CNTs catalyst is more active than the commercially
the catalytic activity and stability for ORR in comparisons to Vulcan
available Pt/C (E-TEK) catalyst. This can be rationalized based on
XC-72R which contains mainly micropores [26].
different factors: The lower charge-transfer resistance at the car-
bon/electrolyte interface, the larger pore sizes, higher dispersion of
3.5. ORR activity of f-CNT supported BOD
Pt NPs and the better affinity of the carbon support to water might
be responsible for the higher catalytic activity. Further investiga-
To investigate the substrate effect of f-CNT, two BOD-based bio-
tion of the device performance with respect to biocompatibility of
cathodes (geometric area 2.0 cm2 ) were prepared using f-CNT and
BP-based electrodes for various fuel cells and sensor applications
Vulcan XC-72R as the supporting substrate, and named BOD/f-CNT-
are currently in progress.
BP and BOD/C-BP, respectively. Fig. 6 shows the electrocatalytic
activity of BOD/C-BP and BOD/f-CNT-BP towards ORR under
steady-state conditions in O2 -saturated phosphate buffer solution Acknowledgements
(0.2 M, pH 7.4), containing 10 mM glucose. Here, we observe a
plateau-like current density at 0.6 V vs. RHE: −0.272 mA cm−2 for We are very grateful to the German Science Foundation (DFG)
BOD-ABTS/C-BP and −0.70 mA cm−2 for BOD-ABTS/f-CNT-BP. The within the graduate school Micro Energy Harvesting (GRK 1322)
current density of the later one is about 2.5 times higher. It is for the financial support. Thanks are due to Prof. B. Kokoh, Dr. Gre-
worth to mention here that the reported limiting current den- gory B. Stevens, Dr. Andreas Schreiber, Dr. Ralf Thomann, Dr. Ralf
sity of −0.3 mA cm−2 (at 0.95 V vs. RHE), using RDE at 100 rpm, in Sorgenfrei, and Dr. Nagham Mehaibes for their support and fruitful
air-saturated solution, was obtained by Habrioux et al. [40]. This discussions.
suggests that the f-CNT substrate enhances the reaction rate on
BOD/f-CNT-BP. To the best of our knowledge, this is the first time Appendix A. Supplementary data
that such an effective oxygen electro-reduction promotion of f-CNT
supported BOD towards ORR is reported in the presence of glucose Supplementary data associated with this article can be found, in
in direct comparison to a BOD/C-BP cathode. This enhanced per- the online version, at doi:10.1016/j.electacta.2011.06.067.
L. Hussein et al. / Electrochimica Acta 56 (2011) 7659–7665 7665

References [21] J. Tang, M. Itkis, C. Wang, X. Wang, Y. Yan, R. Haddon, Micro & Nano Letters 1
(2006) 62.
[1] P. Lens, P.N.L. Lens, Biofuels for Fuel Cells: Renewable Energy from Biomass [22] W.S. Kim, H.S. Song, B.O. Lee, K.H. Kwon, Y.S. Lim, M.S. Kim, Macromolecular
Fermentation , IWA Publishing, London, 2005, p. 169. Research 10 (2002) 253.
[2] P.B. Balbuena, S.R. Calvo, R. Callejas-Tovar, Z. Gu, G.E. Ramirez-Caballero, P. [23] J.M. Tang, K. Jensen, M. Waje, W. Li, P. Larsen, K. Pauley, Z. Chen, P. Ramesh,
Hirunsit, Y. Ma, Challenges in the design of active and durable alloy nanocat- M.E. Itkis, Y. Yan, The Journal of Physical Chemistry C 111 (2007) 17901.
alysts for fuel cells , in: P.B. Balbuena, V.R. Subramanian (Eds.), Theory and [24] W. Zhu, J.P. Zheng, R. Liang, B. Wang, C. Zhang, G. Au, E.J. Plichta, Electrochem-
Experiment in Electrocatalysis, vol. 50, Springer, New York, 2010, p. 351. istry Communications 12 (2010) 1654.
[3] Y.E. Seidel, A. Schneider, Z. Jusys, B. Wickman, B. Kasemo, R.J. Behm, Faraday [25] L. Hussein, S. Rubenwolf, F. von Stetten, G. Urban, R. Zengerle, M. Krueger, S.
Discussions 140 (2009) 167. Kerzenmacher, Biosensors and Bioelectronics 26 (2011) 4133.
[4] V. Selvaraj, M. Alagar, Electrochemistry Communications 9 (2007) 1145. [26] L. Hussein, G. Urban, M. Kruger, Physical Chemistry Chemical Physics 13 (2011)
[5] R. Bashyam, P. Zelenay, Nature 443 (2006) 63. 5831.
[6] J. Shi, D.-j. Guo, Z. Wang, H.-l. Li, Journal of Solid State Electrochemistry 9 (2005) [27] C. Li, E.T. Thostenson, T.W. Chou, Composites Science and Technology 68 (2008)
634. 1227.
[7] S. Kerzenmacher, J. Ducrée, R. Zengerle, F. von Stetten, Journal of Power Sources [28] Y. Fan, M. Burghard, K. Kern, Advanced Materials 14 (2002) 130.
182 (2008) 1. [29] A. Habrioux, E. Sibert, K. Servat, W. Vogel, K. Kokoh, N. Alonso-Vante, The Journal
[8] F.A. de Bruijn, V.A.T. Dam, G.J.M. Janssen, Fuel Cells 8 (2008) 3. of Physical Chemistry B 111 (2007) 10329.
[9] F. Harnisch, S. Wirth, U. Schröder, Electrochemistry Communications 11 (2009) [30] S. Lowell, J. Shields, M. Thomas, M. Thommes, Characterization of Porous Solids
2253. and Powders: Surface Area Pore Size and Density , Kluwer Academic Publishers,
[10] V. Baglio, A. Di Blasi, E. Modica, P. Cretì, V. Antonucci, A. Aricò, International Dordrecht, 2004.
Journal of Electrochemical Science 1 (2006) 71. [31] F.C. Nart, W. Vielstich, Normalisation of porous active surface areas , in: W.
[11] X. Wang, W. Li, Z. Chen, M. Waje, Y. Yan, Journal of Power Sources 158 (2006) Vielstich, A. Lamm, H.A. Gasteiger (Eds.), Handbook of Fuel Cells, vol. 2. Elec-
154. trocatalysis, John Wiley & Sons Ltd., New York, 2003, p. 302.
[12] W. Li, X. Wang, Z. Chen, M. Waje, Y. Yan, The Journal of Physical Chemistry B [32] J.S. Choi, W.S. Chung, H.Y. Ha, T.H. Lim, I.H. Oh, S.A. Hong, H.I. Lee, Journal of
110 (2006) 15353. Power Sources 156 (2006) 466.
[13] R. Smajda, Á. Kukovecz, Z. Kónya, I. Kiricsi, Carbon 45 (2007) 1176. [33] A. Essalik, K. Amouzegar, O. Savadogo, Journal of Applied Electrochemistry 25
[14] A. Kukovecz, R. Smajda, Z. Kónya, I. Kiricsi, Carbon 45 (2007) 1696. (1995) 404.
[15] S.M. Cooper, H.F. Chuang, M. Cinke, B.A. Cruden, M. Meyyappan, Nano Letters [34] C.L. Green, A. Kucernak, The Journal of Physical Chemistry B 106 (2002) 11446.
3 (2003) 189. [35] F. Maillard, O. Cherstiouk, S. Schreier, E. Savinova, U. Stimming, Faraday Dis-
[16] K.H. An, W.S. Kim, Y.S. Park, J.M. Moon, D.J. Bae, S.C. Lim, Y.S. Lee, Y.H. Lee, cussions 125 (2004) 357.
Advanced Functional Materials 11 (2001) 387. [36] W. Vogel, L. Timperman, N. Alonso-Vante, Applied Catalysis A: General 377
[17] R.H. Baughman, C. Cui, A.A. Zakhidov, Z. Iqbal, J.N. Barisci, G.M. Spinks, G.G. (2010) 167.
Wallace, A. Mazzoldi, D. De Rossi, A.G. Rinzler, O. Jaschinski, S. Roth, M. Kertesz, [37] J. Luo, M.M. Maye, V. Petkov, N.N. Kariuki, L. Wang, P. Njoki, D. Mott, Y. Lin, C.J.
Science 284 (1999) 1340. Zhong, Chemistry of Materials 17 (2005) 3086.
[18] U. Vohrer, I. Kolaric, M.H. Haque, S. Roth, U. Detlaff-Weglikowska, Carbon 42 [38] K. Hsueh, E. Gonzalez, S. Srinivasan, Electrochimica Acta 28 (1983) 691.
(2004) 1159. [39] A. Habrioux, W. Vogel, M. Guinel, L. Guetaz, K. Servat, B. Kokoh, N. Alonso-Vante,
[19] W. Knapp, D. Schleussner, Journal of Vacuum Science & Technology B 21 (2003) Physical Chemistry Chemical Physics 11 (2009) 3573.
557. [40] A. Habrioux, T. Napporn, K. Servat, S. Tingry, K. Kokoh, Electrochimica Acta 55
[20] D. Simien, J.A. Fagan, W. Luo, J.F. Douglas, K. Migler, J. Obrzut, ACS Nano 2 (2008) (2010) 7701.
1879.

You might also like