Noncovalent modification of carbon nanotubes with pyrene-functionalized nickel complexes: Carbon monoxide tolerant catalysts for hydrogen evolution and uptake

Article abstract of DOI:10.1002/anie.201005427

Go with the CO: The functionalization of multiwalled carbon nanotubes with molecular complexes through π-π stacking produces robust, noble-metal-free electrocatalytic nanomaterials for H₂ evolution and uptake. The catalysts are compatible with the conditions encountered in classical proton-exchange membrane devices and are tolerant of the common pollutant CO, thus offering significant advantages over traditional Pt-based catalysts.

Full text of DOI:10.1002/anie.201005427

DOI: 10.1002/anie.201005427
Bioinspired Nanocatalysts
Noncovalent Modification of Carbon Nanotubes with Pyrene-
Functionalized Nickel Complexes: Carbon Monoxide Tolerant
Catalysts for Hydrogen Evolution and Uptake**
Phong D. Tran, Alan Le Goff, Jonathan Heidkamp, Bruno Jousselme,* Nicolas Guillet,
Serge Palacin, Holger Dau, Marc Fontecave, and Vincent Artero*
Hydrogen production through the reduction of water appears
to be a very attractive solution for the long-term storage of
renewable energy. However, economically viable processes
require platinum-free catalysts, since this expensive and
scarce metal is not a sustainable resource.[1] We recently
showed that the combination of a bioinspired molecular
approach with nanochemical tools, through the covalent
attachment of mimics[2,3] of the active site of hydrogenase
enzymes onto carbon nanotubes (CNTs), results in a noble-
metal-free electrocatalytic nanomaterial with low overpoten-
tial and exceptional stability for H2 evolution or uptake.[4,5] In
this initial study, we used the electroreduction of a diazonium
salt to decorate multiwalled carbon nanotubes (MWCNTs)
deposited on the electrode support with a polyphenylene
layer bearing amino groups.[6] These amino groups were then
used to attach an activated ester derivative [Ni(PPh2NAr2)2]2+
of the nickel bisdiphosphine bioinspired catalyst developed
by DuBois and co-workers through the formation of an amide
linkage. However, as current manufacturing techniques for
active layers for fuel cells or electrolyzers rely on standard
deposition or printing of an ink containing the electroactive
material, this three-step procedure has obvious practical and
technical drawbacks. We thus turned toward a more direct
and smoother method involving noncovalent p–p stacking
interactions for the functionalization of MWCNTs, since this
methodology[7] has recently been shown to result into efficient
electronic communication between CNTs and immobilized
metal coordination complexes.[8] This approach allows
straightforward and highly convenient preparation of very
stable electrocatalytic materials for H2 evolution and uptake
with tunable catalyst loading. In addition, the catalytic activity
for H2 uptake displayed by this novel molecular engineered
electrocatalytic material is sustained in the presence of carbon
monoxide (CO), a major impurity in H2 fuels derived from
reformed hydrocarbons or biomass. This improvement con-
stitutes a major breakthrough for nafion-based proton-
exchange membrane (PEM) fuel cells technology, since CO
poisoning limits the commercialization of devices based on Pt
electrocatalysts.[9]
[*] Dr. P. D. Tran, Prof. M. Fontecave, Dr. V. Artero
Laboratoire de Chimie et Biologie des Mꢀtaux
Universitꢀ Joseph Fourier, CNRS UMR 5249, CEA DSV/iRTSV
17 rue des Martyrs, 38054 Grenoble cedex 9 (France)
E-mail: vincent.artero@cea.fr
Condensation of pyren-1-yl methylamine with formalde-
hyde in the presence of phenyl- or cyclohexylphosphine yields
1,5-di(pyren-1-ylmethyl)-3,7-diphenyl-1,5-diaza-3,7-diphos-
Dr. A. Le Goff, Dr. B. Jousselme, Dr. S. Palacin
CEA, IRAMIS, SPCSI, Chemistry of Surfaces and Interfaces Group
91191 Gif sur Yvette Cedex (France)
À
Á
phacyclooctane P2PhNC2 H2Pyrene and 1,5-di(pyren-1-ylmethyl)-
3,7-dicyclohexyl-1,5-diaza-3,7-diphosphacyclooctane
E-mail: bruno.jousselme@cea.fr
À
Á
PC2 yNC2 H Pyrene ligands, respectively.[10] The corresponding
2
J. Heidkamp, Prof. H. Dau
FB Physik, Free University Berlin
Arnimallee 14, 14195 Berlin (Germany)
Â
À
Á Ã
orange-to-red nickel(II) complexes Ni PP2hNC2 H2Pyrene
-
2
Â
À
Á Ã
(BF4)2 (1(BF4)2) and Ni PCyNC2 H Pyrene (BF4)2 (2(BF4)2) are
2
2
2
Dr. N. Guillet
Institut LITEN, CEA LITEN/DTH/LCPEM
17 rue des Martyrs, 38054 Grenoble cedex 9 (France)
air- and moisture-stable in the solid state, and their acetoni-
trile solutions can be handled in air for more than 24 h
without degradation, as evidenced by 31P NMR measure-
ments.
Prof. M. Fontecave
Collꢁge de France
11 place Marcellin-Berthelot, 75005 Paris (France)
[**] This work was supported by the CEA (Nanosciences program, grant
Grafthydro) and the ANR (PAN-H 2008, EnzHyd). The X-ray
absorption experiments were supported by Dr. F Schꢂfers at
Helmholtz Zentrum Berlin (HZB/BESSY) and by the Berlin cluster
of excellence on Unifying Concepts in Catalysis (UniCat). We thank
A. Fihri and J. Fize for experimental contributions and P. Jegou for
XPS measurements.
Supporting information for this article (experimental details
including synthetic and catalytic assay procedures, cyclic voltam-
mograms of 1(BF4)2, 2(BF4)2, and [Ni(CH3CN)6](BF4)2, XPS, XANES,
and EXAFS of functionalized MWCNTs/GDL, and electrocatalytic
behavior of bulk catalysts and the derived materials) is available on
the WWW under http://dx.doi.org/10.1002/anie.201005427.
Angew. Chem. Int. Ed. 2011, 50, 1371 –1374
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1371

Communications
Both 1(BF4)2 and 2(BF4)2 are electroactive for catalytic H2
evolution from protonated N,N-dimethylformamide
whereas (35 ꢁ 15)% are octahedrally coordinated by light
atoms. In the latter species, nickel may be coordinated by
water molecules, carboxylate, or hydroxo defects present at
the surface of MWCNTs or by an oxidized diphosphine
ligand, either through the phosphine oxide function or
through the amine function, as shown recently in a similar
system.[15] The extended X-ray absorption fine-structure
(EXAFS)[10] of grafted 1(BF4)2 confirms our conclusion
derived from XANES, namely, the prevalence of the NiIIP4
coordination of 1(BF4)2 and presence of NiII(O/C/N)6 coor-
dination for about one third of the Ni ions. A mixed-ligand
environment of light atoms and phosphorous both coordi-
nated to the same NiII ion is unlikely.[10]
([DMFH]OTf) in CH3CN with similar catalytic rates,[10] and
the overpotential of 0.1 V is slightly lower than that displayed
by the previously reported [Ni(PPh2NPh2)2](BF4)2 complex,[11]
probably because of the distinct nature of the nitrogen
substituent. As both catalysts are unstable in CH3CN solution
in the presence of Et3N, we were unable to measure their
activity for H2 oxidation using the electrochemical assay
procedure previously described by DuBois and co-work-
ers.[11,12] We then turned to a homogeneous assay, adapted
from the conventional procedure used to determine the
specific activity of native hydrogenase enzymes.[13] Both
1(BF4)2 and 2(BF4)2 proved to be active for catalytic H2
(105 Pa) oxidation in CH3CN in the presence of 2,6-lutidine
(2,6-Lut) [Eq. (1)] when methyl viologen hexafluorophos-
phate ([MV](PF6)2) was used as the electron acceptor.[10]
The cyclic voltammogram recorded in pure electrolyte
(CH3CN, 0.1 molLꢂ1 nBu4NBF4) at the MWCNTs/GDL
(MWCNTs loading of 0.05 mgcmꢂ2) electrode modified
with 2(BF4)2 displays two one-electron quasi-reversible
systems at ꢂ0.25 (DEp = 72 mV for a scan rate of 50 mVsꢂ1
)
ð1Þ
and ꢂ0.60 V versus NHE (DEp = 68 mV; 50 mVsꢂ1).[10] The
intensities of both anodic and cathodic peaks are directly
proportional to the scan rate, thus confirming the immobili-
zation of the nickel complexes onto the electrode surface. The
cyclic voltammogram recorded at an electrode modified with
1(BF4)2 shows only one reversible wave at ꢂ0.58 V versus
NHE (DEp = 60 mV; 50 mVsꢂ1), which is likely to be a
combination of the two one-electron waves observed in
solution.[16,17] Integration of the waves allows us to determine
a surface concentration for both catalysts of (2 ꢁ 0.5) ꢀ
10ꢂ9 molcmꢂ2. Such electrochemical responses were found
to be highly reproducible with distinct electrodes and do not
evolve with time nor depend on storage conditions. No
electrochemical signal could be attributed to the Ni species
coordinated to six light atoms. Control experiments with bulk
[Ni(CH3CN)6]2+ show that such species display irreversible
response at potentials distinct from that observed for the
modified electrodes.[10] XAS studies are underway to inves-
tigate whether the degraded Ni species can be electrochemi-
cally converted back into NiP4 under cycling conditions.
We then investigated H2 production and uptake catalyzed
by the new electrode materials with 0.5 molLꢂ1 aqueous
sulfuric acid as the electrolyte. The Ni-functionalized
MWCNTs/GDL electrodes were assembled with a nafion
membrane to protect the catalyst from the acidic solution
while allowing protons to reach or escape the catalytic layer.
These membrane electrode assemblies (MEAs) containing
either 1(BF4)2 or 2(BF4)2 display electrocatalytic activities for
H2 evolution as well as for H2 oxidation (Figure 1). Remark-
ably, both processes occur at vanishingly small overpotentials
as shown by the fact that the traces steeply cut through the
potential axis at the equilibrium potential. Chronoampero-
metric measurements carried out at ꢂ0.3 V versus NHE
under the same conditions did not show any loss of activity
after a 6 h experiment corresponding to 8.5 ꢀ 104 turnovers,[10]
clearly indicating the remarkable robustness of this catalytic
material and consistent with the lack of any leaching of the
p-stacked catalysts. The material obtained from 2(BF4)2
proves to be slightly more efficient for H2 oxidation than
the material obtained from 1(BF4)2, which is as anticipated
from the solution study.[10]
þ
þC
H2 þ 2 ½MVꢀ2þ þ 2 ½2,6-Lutꢀ ! 2 ½MVꢀ þ 2 ½2,6-LutHꢀ
Catalysts 1(BF4)2 and 2(BF4)2 can be physisorbed on
MWCNTs deposited on an electrode substrate through the
establishment of p–p stacking interactions between the
pyrene moieties and graphene motifs. In a first step,
MWCNTs were deposited by filtration onto commercial gas
diffusion layers (GDL), developed for proton exchange
membrane (PEM) applications and consisting of a carbon
fiber cloth coated with a microporous teflon layer embedding
carbon black so as to retain electronic conductivity properties.
Scanning electron micrographs show the high specific surface
displayed by the resulting electrode thanks to the formation
of bundles of MWCNTs with extensive branching.[10] Sec-
ondly, a millimolar solution of catalyst 1(BF4)2 or 2(BF4)2 in
CH2Cl2 was slowly filtered through these MWCNTs/GDL
electrodes. The electrode was then washed with CH3CN, so as
to eliminate any unbound nickel complexes, and air-dried.
X-ray photoelectron spectroscopy (XPS) analysis of the
Ni-functionalized MWCNTs/GDL electrodes shows on the
survey spectrum the presence of Ni, P, N, B, and F constitutive
elements of complex 1(BF4)2 and of oxygen from alcohol or
carboxylic defects of pristine MWCNTs.[10] The decomposi-
tion of the expanded P 2p region shows four peaks: the first
two peaks centered at 132.8 and 133.6 eV correspond to the
P2p3/2 and P2p1/2 peaks, respectively, of the metal-bound
phosphorous atoms[14] and the other peaks at 131.6 and
132.4 eV are attributed to P2p peaks of uncoordinated
phosphine ligands adsorbed on MWCNTs. Fitting and
integration of these peaks gave a ratio of 4:1. The Ni 2p3/2
region is centered at 856.76 eV, which is in good agreement
with the presence of a NiII ion.
We detected a Ni K-edge position indicative of NiII in the
X-ray absorption spectrum. However, grafting 1(BF4)2 on
MWCNTs modifies the X-ray absorption near-edge structure
(XANES) pronouncedly.[10] The XANES spectrum is well
reproduced by a weighted addition of spectra collected for
1(BF4)2 before grafting and for a NiII coordinated to six light
atoms (O, N, C), as found in [Ni(H2O)6]2+ used as a model.[10]
The weighting coefficients suggest that (65 ꢁ 15)% of the Ni
ions are bound to the unmodified ligand system of 1(BF4),
1372
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 1371 –1374

Figure 1. Evolution of current density as a function of potential for
both H2 production and uptake from a 0.5m H2SO4 aqueous solution
under an atmosphere of H2 (105 Pa, H2 flux rate=20 sccm), recorded
at an MEA consisting of a gas diffusion layer (GDL) assembled with a
nafion membrane (2 mVsꢂ1): c unfunctionalized MWCNT/GDL;
c MWCNT/GDL functionalized through p-stacking of 1(BF4)2 (sur-
face catalyst concentration (2ꢁ0.5)ꢃ10ꢂ9 molNi cmꢂ2); c MWCNT/
GDL functionalized through p-stacking of 2(BF4)2 (surface catalyst
concentration (2ꢁ0.5)ꢃ10ꢂ9 molNi cmꢂ2).
Figure 2. Evolution of the current densities for H2 evolution (ꢂ0.3 V
vs. NHE) and uptake (0.2 V vs. NHE) as a function of the surface
catalyst concentration in MEA obtained from 2(BF4)2.
complexes is either ineffective or reversible without inhibiting
H2-oxidation catalysis.[11,18] Nicely, this property is retained in
the materials after grafting. For comparison, commercial
MEAs containing highly dispersed platinum (0.5 mgPtcmꢂ2)
are rapidly and completely deactivated over minutes under
the same conditions (Figure 3).
The p–p stacking functionalization of MWCNTs with
bioinspired molecular complexes thus appears to be a
straightforward methodology to prepare highly robust, CO-
tolerant, noble-metal-free, and bidirectional electrocatalytic
nanomaterials for H2 evolution and uptake, which are
compatible with the conditions encountered in classical
proton-exchange membrane devices. The materials reported
herein can be prepared in one step from MWCNTs and
pyrene-functionalized complexes and then deposited onto
any electrode support. A major challenge will now consist of
As a major drawback, the previously used electrografting
procedure imposes some limitation on the amount of grafted
catalysts, since the growth of the polyphenylene layer
partially fills the pores of the MWCNTs deposit, thus leading
to degradation of the electron- and mass-transport perform-
ances. As a consequence, we were not able to increase the
surface concentration above (1.5 ꢁ 0.5) ꢀ 10ꢂ9 molNi cmꢂ2,
regardless of the amount of CNTs deposited on the GDL,
and the current densities above a few mAcmꢂ2 [4]
. By contrast,
the p-stacking methodology reported herein allows decora-
tion of the MWCNTs with a monolayer of catalysts so that
catalyst loading on MWCNTs/GDL electrodes can be easily
tuned by controlling the amount of CNTs initially deposited
on the GDL electrode. Thus, surface catalyst concentrations
can be linearly increased up to (1.1 ꢁ 1) ꢀ 10ꢂ8 molNi cmꢂ2 for
CNT loading of 0.3 mgcmꢂ2.[10] Catalytic current density for
H2 evolution linearly increases with catalyst loading to reach
almost 20 mAcmꢂ2 (Figure 2). Current densities for H2
oxidation increased much less. Moreover, the two new
materials display catalytic current densities for H2 oxidation
comparable to those generated by the previously reported
one,[4] in which the nickel catalyst was covalently grafted.[10]
Given the contrasting catalytic performances of bulk
(ungrafted) 1(BF4)2, 2(BF4)2, and [Ni(PPh2NAr2)2]2+ (Ar=
aryl) for H2 evolution,[10] these observations indicate that
mass transport of H2 gas within a film of functionalized CNTs
limits the catalytic rates and thus the anodic current density.
A very important limitation of the use of platinum
nanoparticles as electrocatalysts for H2 oxidation in fuel
cells comes from the poisoning effect of CO.[9] Hence, we
investigated the catalytic performances of Ni-functionalized
materials for H2 (105 Pa) oxidation in the presence of CO
(50 ppm) and found no inhibiting effect of this pollutant
(Figure 3 and the Supporting Information). Solution studies
have shown that binding of CO to nickel(II) bisdiphosphine
Figure 3. Evolution upon repeated cycling (2 mVsꢂ1) of the H2 oxida-
tion current density per mg of deposited catalyst recorded at 0.25 V
versus NHE for a nickel-functionalized MWCNT/GDL electrode (pre-
pared from 2(BF4)2; surface catalyst concentration:
(2ꢁ0.5)ꢃ10ꢂ9 molcmꢂ2) and a commercial Pt-based active layer
(Pt nanoparticles deposited on carbon black, 0.5 mgPt cmꢂ2) under an
atmosphere of H2 polluted with 50 ppm CO (flux rate: 20 sccm).
Angew. Chem. Int. Ed. 2011, 50, 1371 –1374
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1373

Communications
[6] A. Le Goff, F. Moggia, N. Debou, P. Jegou, V. Artero, M.
Fontecave, B. Jousselme, S. Palacin, J. Electroanal. Chem. 2010,
641, 57 – 63.
[7] Y. L. Zhao, J. F. Stoddart, Acc. Chem. Res. 2009, 42, 1161 – 1171.
[8] E. W. McQueen, J. I. Goldsmith, J. Am. Chem. Soc. 2009, 131,
17554 – 17556.
significantly improving the catalytic loading so as to reach the
power performances displayed by commercial Pt-based active
layers. To this end, again the p–p stacking methodology
allows such tuning of the catalytic loading by increasing the
amount of supporting MWCNTs. Since specific H2 oxidation
current densities (i.e. relative to the mass of immobilized
catalyst including metal, ligands, and counteranions, if
[9] J. J. Baschuk, X. G. Li, Int. J. Energy Res. 2001, 25, 695 – 713.
[10] See the Supporting Information.
[11] A. D. Wilson, R. H. Newell, M. J. McNevin, J. T. Muckerman,
M. R. DuBois, D. L. DuBois, J. Am. Chem. Soc. 2006, 128, 358 –
366.
[12] J. Y. Yang, R. M. Bullock, W. J. Shaw, B. Twamley, K. Fraze,
M. R. DuBois, D. L. DuBois, J. Am. Chem. Soc. 2009, 131, 5935 –
5945.
applicable, as shown in Figure 3) obtained with the Ni-based
materials (250 mAcmꢂ2 mgcatalyst
for 2(BF4)2) and with
ꢂ1
commercial active layers containing highly dispersed plati-
num (200 mAcmꢂ2 mgPtꢂ1; 0.5 mgPt cmꢂ2) are comparable, this
methodology paves the way for the implementation of noble-
metal-free electrocatalytic materials in operative PEM devi-
ces.
[13] V. Artero, M. Fontecave, Coord. Chem. Rev. 2005, 249, 1518 –
1535.
[14] The signal of oxidized phosphine ligand, either coordinated or
not, would appear at the same energy.
Received: August 30, 2010
[15] J. Y. Yang, R. M. Bullock, W. G. Dougherty, W. S. Kassel, B.
Twamley, D. L. DuBois, M. R. DuBois, Dalton Trans. 2010, 39,
3001 – 3010.
Published online: January 5, 2011
Keywords: carbon · heterogeneous catalysis · hydrogen ·
.
[16] Actually, the two NiII/I and NiI/0 redox features are quite narrow
in 1(BF4)2 and [Ni(PPh2NPh2)2](BF4)2 relative to that of 2(BF4)2.
Such behavior has been traced to structural distortions.[19]
[17] G. M. Jacobsen, J. Y. Yang, B. Twamley, A. D. Wilson, R. M.
Bullock, M. R. DuBois, D. L. DuBois, Energy Environ. Sci. 2008,
1, 167 – 174.
nanotubes · nickel
[1] R. B. Gordon, M. Bertram, T. E. Graedel, Proc. Natl. Acad. Sci.
USA 2006, 103, 1209 – 1214.
[2] S. Canaguier, V. Artero, M. Fontecave, Dalton Trans. 2008, 315 –
325.
[3] C. Tard, C. J. Pickett, Chem. Rev. 2009, 109, 2245 – 2274.
[4] A. Le Goff, V. Artero, B. Jousselme, P. D. Tran, N. Guillet, R.
Metaye, A. Fihri, S. Palacin, M. Fontecave, Science 2009, 326,
1384 – 1387.
[18] A. D. Wilson, K. Fraze, B. Twamley, S. M. Miller, D. L. DuBois,
M. R. DuBois, J. Am. Chem. Soc. 2008, 130, 1061 – 1068.
[19] J. W. Raebiger, A. Miedaner, C. J. Curtis, S. M. Miller, O. P.
Anderson, D. L. DuBois, J. Am. Chem. Soc. 2004, 126, 5502 –
5514.
[5] P. D. Tran, V. Artero, M. Fontecave, Energy Environ. Sci. 2010, 3,
727 – 747.
1374
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 1371 –1374