Tailoring the selectivity and stability of chemically modified platinum nanocatalysts to design highly durable anodes for PEM fuel cells

Article abstract of DOI:10.1002/anie.201100744

Big and small: Chemically modifying platinum with calix[4]arene yields a highly stable anode catalyst that effectively suppresses the oxidation reduction reaction without altering the maximum activity for the hydrogen oxidation reaction (see picture, Ptblue, Cgray, Ored, Syellow). This behavior extends from long-range-ordered stepped single-crystal surfaces to nanocatalysts. Copyright

Full text of DOI:10.1002/anie.201100744

Communications
DOI: 10.1002/anie.201100744
Heterogeneous Catalysis
Tailoring the Selectivity and Stability of Chemically
Modified Platinum Nanocatalysts To Design Highly
Durable Anodes for PEM Fuel Cells**
Bostjan Genorio, Ram Subbaraman, Dusan Strmcnik, Dusan Tripkovic,
Vojislav R. Stamenkovic, and Nenad M. Markovic*
Angewandte
Chemie
5468
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5468 –5472

An ever-changing energy landscape and the global drive
toward greener energy technologies have made fuel cells a
focal point of numerous research initiatives. In their current
form, proton-exchange-membrane fuel cells (PEMFCs) have
been shown to perform well under operating conditions for
both automotive and stationary applications. In order for
PEMFCs to reach commercial implementation, issues such as
durability under both normal and startup and shutdown
conditions have to be tackled effectively. Reactivity, selectiv-
ity, and stability are the quintessential properties that need to
be tailored to develop catalysts that can tackle the durability
issues arising during startup and shutdown.[1] One approach to
accomplish this task is to design an anode catalyst that can
efficiently suppress the undesired oxygen reduction reaction
(ORR; imparting selectivity) and preserve the platinum-like
hydrogen oxidation reaction (HOR) activity (imparting
reactivity), while remaining stable under operating conditions
(imparting stability). Such an approach not only reduces the
overpotential on the cathode side owing to negligible ORR
currents on the anode but also prevents formation of
detrimental products such as hydrogen peroxide, which can
be formed under “normal” anode startup and shutdown
conditions. We have shown that chemically modified electro-
des (CME) consisting of self-assembled monolayers (SAMs)
of calix[4]arene molecules on extended platinum single-
crystal surfaces can selectively block the ORR without
affecting the HOR activities and kinetics.[2] Usually, the
lessons learned from such extended surfaces have helped in
the understanding of nanocatalysts that mimic the reactivity
and catalytic behavior of the extended surfaces.[3] Seldom,
however, can the behavior of extended surfaces be completely
translated down to the nanocatalysts.
surfaces. Then, we developed a methodology to form highly
selective and stable SAMs of calix molecules on commercial
nanocatalysts (3M nanostructured thin film (NSTF)[4] and
Tanaka 5 nm Pt/C (TKK) catalysts). We find that if the
synthesis is precisely controlled, the selectivity of nano-
particles for the ORR in the presence of hydrogen under
conditions relevant to PEMFC operations is almost 100%.
We start with the electrochemical characteristics of
calix[4]arene-decorated Pt(110) and Pt(1099). As summar-
ized in Figure 1, both stepped surfaces show characteristic
cyclic voltammograms; the under-potentially deposited (Hupd
)
hydrogen (0–0.4 V) is followed first by a double-layer region
and then at E > 0.6 V by reversible (OHad) and irreversible
oxide formation.[5] On the calix-covered surfaces, however,
both the Hupd and OHad regions are significantly suppressed.
In line with Ref. [2], on the highly covered surfaces the
number of “free” Pt sites (determined from the Hupd charge) is
extremely low (ca. 2–3%). However, on the same surface the
HOR is similar to calix-free Pt, thus confirming that the
turnover frequency (TOF) of the hydrogen reaction is
extremely high[6] and that the Pt–H2 energetics are not
affected by the adsorbed calix molecules.
More importantly, we find that at E > 0.6 V, while the
ORR is almost completely inhibited (these modifications are
not accompanied by undesired peroxide production),[2] the
HOR is under pure diffusion control. This unique selectivity is
attributed to very strong ensemble effects in which the
number of bare Pt sites available for adsorption of O2 is much
smaller than that for the adsorption of H2 and the subsequent
HOR.[2] We conclude therefore that the selectivity achieved
using calix-modified electrodes is not affected by the presence
of steps. This result is very important because it provides
evidence that such behavior can be successfully translated to
the most commonly used forms of nanocatalysts, which are
known to contain a vast majority of such sites (steps and
short-range terraces).
Having established the behavior of well-defined surfaces,
we move on to the most relevant electrocatalyst systems:
nanocatalysts. To encompass a wide range of electrocatalyst
designs and properties, we provide an analysis for the two
most commonly used commercial electrocatalysts. The TKK
catalyst and the 3M NSTF catalyst were both studied
(Figure 2). The TKK catalyst represents supported nano-
catalysts, where platinum nanoparticles 2–10 nm in diameter
are supported on amorphous carbon black. NSTF catalysts,
comprised of a unique catalyst structure which is free of
carbon support, are usually applied directly to the membrane
to provide a compact membrane electrode assembly structure
(Figure 2). Aqueous electrochemical experiments conducted
using the RDE/RRDE (RDE = rotating disk electrode,
RRDE = rotating ring disk electrode) methods for these
nanocatalysts are well-established[7] and have been shown to
correlate very well with operating fuel-cell systems.
Herein, we show that the platinum modified with
calix[4]arene (calix) is, in fact, one of these rare examples in
which the modified nanocatalyst system behaves in line with
the corresponding extended-surface system. First, we dem-
onstrate high selectivity of the HOR on calix-modified
Pt(1099){10(111) ꢀ (100)}and Pt(110){2(111) ꢀ (100)} step
[*] Dr. B. Genorio, Dr. D. Strmcnik, Dr. D. Tripkovic,
Dr. V. R. Stamenkovic, Dr. N. M. Markovic
Materials Science Division, Argonne National Laboratory
Argonne, IL 60439 (USA)
E-mail: nmmarkovic@anl.gov
Dr. B. Genorio
Faculty of Chemistry and Chemical Technology
University of Ljubljana (Slovenia)
Dr. R. Subbaraman
Nuclear Engineering Division
Argonne National Laboratory (USA)
[**] This work was supported by the Director, Office of Science, Office of
Basic Energy Sciences, Division of Materials Sciences, US Depart-
ment of Energy under Contract No. DE-AC03-76SF00098 and the
Center of Excellence Low Carbon Technologies Slovenia (CO NOT),
Center of Excellence Advanced Materials and Technologies for the
Future Slovenia (CO NAMASTE). R.S. is grateful for financial
support from an Argonne postdoctoral fellowship. PEM=Proton-
exchange membrane.
We present herein results obtained from the RDE study
that should be relevant for operating fuel-cell systems.
Various modifications of the calix molecules were studied,
including the thiolated derivatives of calix[6]arenes and
calix[8]arenes (see the Supporting Information for the syn-
thesis), but only the derivatives of the calix[4]arene family
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201100744.
Angew. Chem. Int. Ed. 2011, 50, 5468 –5472
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 5469

Communications
Figure 1. Electrochemical characteristics of calix-modified stepped surfaces: a) Pt(1099), b) Pt(110). Potentials of interest during startup and
shutdown are shown in the shaded region.[1,2]
limiting currents for the HOR are observed at potentials
above 0.1 V and the activities below 0.1 V are, within the
experimental limits, almost identical.
Furthermore, the ORR polarization curves show limited
or insignificant currents in the potential region of interest for
the anode-side catalyst. As was shown in the earlier study with
Pt(111),[2] the peroxide yield on all extended and nanoparticle
Pt–calix systems is negligible above 0.6 V, and the overall
ORR behavior of these surfaces mimic ORR on uncovered or
partially covered patches.[6] All of these observations suggest
that SAMs of calix molecule can be used to tailor the
selectivity of the nanocatalyst toward ORR while preserving
the HOR activity, the goal for an ideal anode catalyst. It is
also important that the established selectivity was possible
only because the required number of active sites for maximal
rates of the HOR is, in fact, extremely small but is sufficient to
provide enough sites for the diffusion-limiting currents.[2]
In addition to selectivity of CME, both thermal and
electrochemical stability of these electrodes are important
properties that need to be addressed to evaluate the anode
catalystꢁs applicability to PEMFC. In order to study the
stability of calix-modified electrodes, we tested a Pt–calix
system in an oxygen-rich environment at 0.8 V for approx-
imately 14 h in solution at 608C. These conditions are
expected to be harsher than those experienced by the
electrode in a real fuel-cell system. The exposure of the
anode catalyst to high potentials (E < 0.8 V for anode) in an
air (oxygen)-rich atmosphere during startup and shutdown is
expected to last between tens of seconds and a few minutes a
day. The temperatures are expected to be similar to those
used in our test conditions.
Figure 2. a) SEM image of an unmodified Pt nanowhisker. b) TEM
morphology of typical TKK nanocatalysts. Calix molecules are not
visible by electron microscopy. Based on our previous STM study,[2] we
present a schematic representation of calix-modified nanocatalysts:
c) Model morphology of calix[4]arene (yellow-gray-red)-modified Pt
NSTF nanowhisker (blue). d) Model for TKK nanocatalyst chemically
modified with calix[4]arene.
proved to be effective in achieving the selectivity, so only the
results pertaining to the latter are presented. As can be seen
in Figure 3, the calix[4]arene molecules are found to suppress
the Hupd region (0.05–0.4 V) for both NSTF and TKK
catalysts. The relative coverages for similar methods of
preparation are slightly different, but the net results appear
to be the same: an exceptional selectivity for the HOR versus
ORR. As for stepped surfaces discussed above, the diffusion-
Figure 4 shows the current–time relationship for the CME
held at 0.8 V in an oxygen-rich atmosphere. The ORR current
5470
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5468 –5472

Figure 3. Electrochemical characteristics of calix-modified Pt nanocatalysts: a) NSTF, b) TKK (50% Pt loading). Catalyst loadings were
approximately 14–16 mgcmÀ2; lower loadings were also used to mimic anode catalyst performance, yielding similar qualitative results. All current
densities are given with respect to the disk geometric area.
actually shows a small decay, thus suggesting that there is no
loss of the calix molecules from the surface owing to
oxidation. (Removal of the molecules by oxidation or
desorption would increase the reduction current.) A similar
experiment was also performed for the nanocatalysts (TKK)
modified with calix[4]arene molecules, which show qualita-
tively similar results. This finding suggests that the calix[4]-
arene-modified electrodes are stable under these operating
conditions. Moreover, during the long-term experiments, the
HOR (results not shown) is not affected at all.
tune the selectivity of anode catalysts for ORR without
altering the maximum activity of the HOR. This behavior is
highly transformational, extending from long-range-ordered
stepped single-crystal surfaces to nanocatalysts. The CME
approach is not restricted to a Pt–calix system, and we
envision it to provide many applications in analytical,
synthetic, and materials chemistry as well as in chemical
energy conversion and storage.
In conclusion, our CMEs prepared by modifying Pt with
calix[4]arene molecules are highly stable and can effectively
Experimental Section
Synthesis of the thiolated derivative of calix[4]arene: Adsorption of
organic groups on noble-metal surfaces has been well-established for
various groups (-S, -CN, -A, where A denotes the anchoring group).[8]
The driving force for ordering of such large molecules is presumable
governed by a synergy between the strong chemical bond between the
anchoring groups and Pt surface atoms and the local steric interaction
between adsorbed molecules. The calix[4]arene moleculesꢁ anchoring
groups are usually of the form S(R) where (R) is used to cap the thiol
group on the quadrupole anchoring groups. A detailed description of
the synthesis procedure as well as molecular designs considered are
presented in the Supporting Information.
Preparation of Pt(1099), Pt(110), and Pt(polycrystalline) surfa-
ces and self-assembly: Pt electrodes were prepared by inductive
heating for 10 min at approximately 1100 K in an argon–hydrogen
flow (3% hydrogen). The annealed specimen was cooled slowly to
room temperature in this flow stream and immediately covered by a
droplet of water. The electrode was then immersed in a THF solution
of calix[4]arene for 24 h, allowing the formation of a calix[4]arene
SAM. The concentration of calix[4]arene in THF was 600 mm to
obtain samples with very high coverages of calix on Pt surface.
Coverages were estimated from the Hupd measurements. The effect of
coverage on ORR and HOR was previously presented.[3] The
coverages can be modified by either varying the concentrations of
the calix/THF solution or the exposure time to the high-concentration
solution. After SAM preparation, the crystals were washed thor-
Figure 4. Stability of the Pt–calix system at 608C in O2-saturated 0.1m
HClO4 at 0.8 V. Inset: ORR curves for unmodified surface and Pt–calix
surface both before and after the stability test. Note: the HOR remains
unchanged for the duration of the experiment.
Angew. Chem. Int. Ed. 2011, 50, 5468 –5472
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5471

Communications
oughly with deionized water before assembly and immersion in the
electrochemical cell.
the electrode was rotated at 1600 rpm. Electrode potentials are given
versus the reversible hydrogen electrode (RHE).
Preparation of NSTF and TKK catalyst electrodes and their self-
assembly: The catalysts were mixed with water at a concentration of
1 mgmLÀ1. This dispersion was then ultrasonically mixed for one
hour, after which a stable suspension was obtained. A glassy carbon
disk (6 mm diameter) was then mechanically polished. Known
volumes of the suspensions were then added using a micropipette
onto the glassy carbon disk electrode. The electrode was dried at 608C
in an inert atmosphere. The suspension was applied so that it coated
the surface of the electrode very uniformly. Once dry, these electrodes
were washed with water to verify the good adhesion of particles to the
glassy carbon substrate. Subsequently, the electrodes were immersed
in 1000 mm solution of calix in THF. We chose to use a high
concentration of calix owing to the larger surface area of Pt compared
to the disk electrodes. The systems were equilibrated for 24 h.
Another method involved assembly of the disk electrode in a hanging
meniscus arrangement with subsequent immersion of the electrode in
the calix solution with rotation (600 rpm) for 4 h. Both of these
methods yielded similar coverages. After equilibration, the samples
were washed thoroughly with water before being immersed in the
electrochemical cell. For a discussion on the relative coverages
obtained for the same conditions for TKK and NSTF, please refer to
the Supporting Information.
RDE method, electrolytes, and electrochemical setup: After
extensive rinsing, the electrode was embedded into the rotating-disk
electrode (RDE) and transferred into a standard three-compartment
electrochemical cell containing 0.1m HClO4 (Sigma–Aldrich). In each
experiment, the electrode was immersed at 0.07 V in solution
saturated with Ar. After obtaining a stable voltammogram between
0.07 and 0.7 V the polarization curve for the ORR was recorded on
the disk on the disk electrode. Subsequently, oxygen was purged out
of the solution and replaced with hydrogen, and HOR polarization
curves were measured. Finally, the voltammetric response was again
recorded in argon-purged solution to confirm that calix coverages had
not changed significantly.
Received: January 29, 2011
Published online: May 12, 2011
Keywords: calixarenes · electrochemistry ·
.
heterogeneous catalysis · oxygen reduction reaction · platinum
[1] R. Borup, J. Meyers, B. Pivovar, Y. S. Kim, R. Mukundan, N.
Garland, D. Myers, M. Wilson, F. Garzon, D. Wood, P. Zelenay, K.
More, K. Stroh, T. Zawodzinski, J. Boncella, J. E. McGrath, M.
Inaba, K. Miyatake, M. Hori, K. Ota, Z. Ogumi, S. Miyata, A.
Nishikata, Z. Siroma, Y. Uchimoto, K. Yasuda, K. I. Kimijima, N.
Iwashita, Chem. Rev. 2007, 107, 3904 – 3951; C. A. Reiser, L.
Bregoli, T. W. Patterson, J. S. Yi, J. D. L. Yang, M. L. Perry, T. D.
Jarvi, Electrochem. Solid-State Lett. 2005, 8, A273 – A276; J. P.
Meyers, R. M. Darling, J. Electrochem. Soc. 2006, 153, A1432 –
A1442; R. A. Sidik, J. Solid State Electrochem. 2009, 13, 1123 –
1126.
[2] B. Genorio, D. Strmcnik, R. Subbaraman, D. Tripkovic, G.
Karapetrov, V. R. Stamenkovic, S. Pejovnik, N. M. Markovic,
Nat. Mater. 2010, 9, 998 – 1003.
[3] H. A. Gasteiger, N. M. Markovic, Science 2009, 324, 48 – 49; V. R.
Stamenkovic, B. S. Mun, M. Arenz, K. J. J. Mayrhofer, C. A.
Lucas, G. F. Wang, P. N. Ross, N. M. Markovic, Nat. Mater. 2007, 6,
241 – 247.
[4] M. K. Debe, A. K. Schmoeckel, G. D. Vernstrorn, R. Atanasoski,
J. Power Sources 2006, 161, 1002 – 1011.
[5] N. M. Markovic, P. N. Ross, Jr., Surf. Sci. Rep. 2002, 45, 117 – 229.
[6] D. S. Strmcnik, P. Rebec, M. Gaberscek, D. Tripkovic, V.
Stamenkovic, C. Lucas, N. M. Markovic, J. Phys. Chem. C 2007,
111, 18672 – 18678.
[7] K. J. J. Mayrhofer, D. Strmcnik, B. B. Blizanac, V. Stamenkovic,
M. Arenz, N. M. Markovic, Electrochim. Acta 2008, 53, 3181 –
3188.
All gases were 5N5 quality purchased from Airgas Inc. The sweep
[8] J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, G. M.
Whitesides, Chem. Rev. 2005, 105, 1103 – 1169.
rate for all measurements was 50 mVsÀ1; for the ORR measurements,
5472
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 5468 –5472