Angewandte

Chemie

This is the peer reviewed version of the following article: Angew. Chem. Int. Ed. 2015, 54, 11265 – 11269, which has
been published in final form at http://dx.doi.org/10.1002/anie.201505073. This article may Edition: DOI: 10.1002/anie.201505073
International be used for non-commercial
purposes in accordance with Wiley Terms and Conditions for Self-Archiving.
Heterogeneous Catalysis
German Edition:
DOI: 10.1002/ange.201505073

A Stable Single-Site Palladium Catalyst for Hydrogenations
Gianvito VilØ, Davide Albani, Maarten Nachtegaal, Zupeng Chen, Dariya Dontsova,
Markus Antonietti, Nfflria López,* and Javier PØrez-Ramírez*
Abstract: We report the preparation and hydrogenation
performance of a single-site palladium catalyst that was
obtained by the anchoring of Pd atoms into the cavities of
mesoporous polymeric graphitic carbon nitride. The characterization of the material confirmed the atomic dispersion of the
palladium phase throughout the sample. The catalyst was
applied for three-phase hydrogenations of alkynes and nitroarenes in a continuous-flow reactor, showing its high activity
and product selectivity in comparison with benchmark catalysts based on nanoparticles. Density functional theory calculations provided fundamental insights into the material structure and attributed the high catalyst activity and selectivity to
the facile hydrogen activation and hydrocarbon adsorption on
atomically dispersed Pd sites.

The selective hydrogenation of acetylenic and nitroaromatic

compounds is an important step in the synthesis of building
blocks for polymers, vitamins, fragrances, and agrochemicals.[1] Commercial catalysts for this family of reactions are
based on supported Pd or Pt nanoparticles modified with
harmful promoters such as lead (Lindlar) or, more recently,
with organic ligands (NanoSelect).[2] The modifiers not only
alter the adsorption energies of the reactants and products but
also reduce the size of the ensemble of surface atoms where
the reaction takes place. This prevents oligomer production
and hydride formation, the latter being responsible for the
undesired isomerization and over-hydrogenation pathways.[2]
In attempts to improve the performance of conventional
catalysts, enormous efforts have been devoted to the design of
materials with tailored metal-particle sizes.[3] In fact, engineering a minimal ensemble that is able to activate H2 and
displays interactions that are strong enough with the reactants
and weak with the products would offer unrivalled oppor[*] G. VilØ,[+] D. Albani,[+] Prof. J. PØrez-Ramírez
Institute for Chemical and Bioengineering
Department of Chemistry and Applied Biosciences, ETH Zurich
Vladimir-Prelog-Weg 1, 8093 Zurich (Switzerland)
E-mail: jpr@chem.ethz.ch
Dr. M. Nachtegaal
Paul Scherrer Institute, 5232 Villigen (Switzerland)
Z. Chen, Dr. D. Dontsova, Prof. M. Antonietti
Max-Planck Institute of Colloids and Interfaces
Department of Colloid Chemistry
Research Campus Golm, 14424 Potsdam (Germany)
Prof. N. López
Institute of Chemical Research of Catalonia (ICIQ)
Av. Països Catalans 16, 43007 Tarragona (Spain)
E-mail: nlopez@iciq.es
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201505073.
Angew. Chem. Int. Ed. 2015, 54, 11265 –11269

tunities for atom-efficient catalytic transformations. The
ultimate frontier in size reduction is a single-site heterogeneous catalyst, in which isolated metal atoms anchored on
a support catalyze hydrogenation reactions, mimicking the
activity of biological entities such as enzymes and antibodies.[4] Flytzani-Stephanopoulos and co-workers have
recently demonstrated that a single crystal containing a
Cu(111) surface doped with isolated Pd atoms can activate
hydrogen and display activity in the hydrogenation of
acetylene under ultra-high vacuum conditions.[5a] Nevertheless, the relatively low ethylene selectivity (30 %) at moderate
acetylene conversions (10–20 %) and the pressure and
material gaps of the model catalyst limit the broad applicability of this work. The synthesis of more realistic single-site
heterogeneous catalysts remains to be a challenging task as
isolated atoms are often thermodynamically unstable and
tend to leach and agglomerate during the first few hours of
a reaction.[5b–e] To design stable single-site catalysts, it is
crucial to select a high-surface-area support containing
“cages” or “cavities” that can entrap the catalytically active
atoms in an isolated form, similarly to what has been done in
the past for asymmetric organometallic catalysts and “ship-ina-bottle” nanostructures.[4]
Mesoporous polymeric graphitic carbon nitride (mpgC3N4) has attracted increasing interest in the field of material
science owing to its lattice–hole structure that partially
´
resembles a Sierpinsky sieve. The material, in fact, contains
characteristic N-coordinating cavities formed during the
polymerization of heptazine, which are called “six-fold
cavities” as they are delimited by six C3N4 rings (see
Figure 1 a; see also the Supporting Information, Figure S1).[6]
This structure gives rise to unique mechanical, electronic, and
conducting properties that are not encountered in bulk
carbon nitride samples.[7] Using this material, we were able
to strongly anchor isolated Pd species into host cavities,
designing the first stable single-site heterogeneous catalyst for
hydrogenations (hereafter referred to as [Pd]mpg-C3N4 ; for
details on its preparation, see the Supporting Information). A
combination of characterization studies, catalytic tests in
a continuous-flow reactor, and density functional theory
(DFT) calculations led to a molecular-level understanding of
the structure and reactivity of this novel material.
The [Pd]mpg-C3N4 sample contains 0.5 wt % Pd (Table 1).
The mpg-C3N4 carrier has a total surface area of 155 m2 gÀ1
and a total pore volume of 0.26 cm3 gÀ1. Microscopic examination (Figure 1 a) revealed that the palladium phase was
atomically distributed throughout the sample (metal dispersion: 100 %). This was confirmed by an analysis of the metal
distribution, displaying an average particle size of approximately 0.3–0.4 nm, which matches the van der Waals
diameter of a single palladium atom. X-ray absorption fine-

 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

11265

.
Angewandte
Communications

Figure 2. a) k2-weighted Fourier transform extended X-ray absorption
fine structure spectra (not phase-corrected). b) Pd 3d core level X-ray
photoelectron spectroscopy of the catalysts. The circles in (b) represent the experimental data, whereas the gray and violet lines show the
fitted Pd0 and Pd2+ components, respectively.

Figure 1. a–d) Structure of the materials (left), aberration-corrected
scanning transmission electron microscopy images (middle), and
particle size distribution (right) of [Pd]mpg-C3N4 (a), Pd-HHDMA/TiS
(b), Pd-Pb/CaCO3 (c), and Pd/Al2O3 (d). The structures depict the
increasing size of the active ensemble, from a single Pd atom (a) to
a bare Pd nanoparticle of approximately 800 atoms (d). The inset in
(a) shows a characteristic six-fold cavity in the carbon nitride structure,
which is crucial for confining palladium atoms in a stable manner. The
two-dimensional Gaussian function modeling of a selected region of
[Pd]mpg-C3N4 is shown in the inset of the micrograph in (a) to enable
visualization of the single Pd atoms (in red). C dark gray, H white,
N purple, O red, P light orange, Pb light gray, Pd blue.

Table 1: Sample characterization.
Sample
[Pd]mpg-C3N4
Pd-HHDMA/TiS
Pd-Pb/CaCO3
Pd/Al2O3

Pd[a]
[wt. %]
0.5
0.5
4.5
1.0

SBET[b]
[m2 gÀ1]
155
229
10
170

Vpore[c]
[cmÀ3 gÀ1]
0.26
0.16
0.03
0.64

DTEM[d]
[%]

DCO[e]
[%]

100
16
8
39

–
14
6
37

[a] ICP-OES. [b] BET method. [c] Volume of N2 at p/p0 = 0.98. [d] TEM.
[e] CO chemisorption.

structure spectroscopy (XAFS; Figure 2 a) corroborated the
atomic dispersion of palladium in [Pd]mpg-C3N4. No peaks
could be detected at a distance greater than 2 Š, which
indicates that no Pd–Pd bonds were formed, in contrast to
conventional nanoparticle-based catalysts.[8a]

11266

www.angewandte.org

This carbon nitride material was then studied by DFT
calculations for the first time to analyze the nature of the
isolated Pd sites incorporated into mpg-C3N4, their stability,
and reactivity. The unique features of mpg-C3N4 complicate
the calculations as a large rumpling is expected for this kind of
two-dimensional materials. In the modeling, special attention
was thus paid to the structural landscape, including the Pd
adsorption in “cages”, potential intercalation between the
graphitic layers, aggregation of Pd, and diffusion. The high
complexity observed has prompted us to present only the
most remarkable results. DFT calculations on a model surface
representing mpg-C3N4 revealed that the presence of an
atomic distribution of Pd can be attributed to the function of
the N species in the support, which electrostatically stabilize
the Pd species. In fact, the trapping of Pd in the six-fold
cavities is exothermic by 1.5 eV with respect to the gas-phase
atomic reference (Figure S1). This strong interaction between
the nitrogen atoms on the surface and the metal, coupled with
the large number of homogeneously distributed anchoring
centers, leads to the stabilization of palladium atoms and
appears to be fundamental for the design of a stable singlesite catalyst for hydrogenation reactions. A different support
with fully accessible sites (such as alumina) would suffer from
rapid agglomeration of the metal (see also Figure 3 c).
Considering that the six-fold cavities are present over the
entire mpg-C3N4 material with a density of approximately 1/
50 Š2, it is possible that during synthesis, parts of the metal
phase diffuse between the C3N4 layers until a stable, final
configuration is reached. This indicates that palladium can
also occupy subsurface and deeper positions in the form of an
intercalation compound (Figure S2). However, in the resting
state, Pd atoms placed in subsurface positions are slightly
more stable (see the Supporting Information for a detailed
analysis). Considering the specific surface area of mpg-C3N4,
the surface density of the six-fold cavities in a fully uniform
layer, the total palladium loading, and assuming that all
palladium is located at the surface, we could estimate that 8–
10 % of all surface six-fold cavities are occupied by palladium
(see the Supporting Information for a detailed analysis). This

 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2015, 54, 11265 –11269

Angewandte

Chemie

is only an estimate as Pd can also be distributed deep in the
graphitic layers, reducing the effective fraction of “surface
cage” occupation. Comparatively, in a 10 nm nanoparticle,
6 % of the Pd atoms would be located on the surface.[8b] From
a thermodynamic point of view, the aggregation of Pd atoms
to form dimers, trimers, or larger clusters is not favored
(Figure S3). Furthermore, considering the low palladium
loading and quite large number of nesting positions, entropic
contributions dominate, clarifying the stability of the atomic
sites. Only at high Pd contents, the six-fold cavities are
saturated with atomic species, and metal agglomeration starts.
This explains, for instance, the presence of nanoparticles of
approximately 3 nm when 20 times higher palladium loadings
were employed and the same preparation procedure was
followed.[6c]
The structure of [Pd]mpg-C3N4 can be compared with that
of three benchmark catalysts that are commercially available
and widely utilized in industry: ligand-modified PdHHDMA/TiS (HHDMA = hexadecyl-2-hydroxyethyl-dimethylammonium dihydrogen phosphate), Lindlar-type Pd-Pb/
CaCO3, and Pd/Al2O3 (Figure 1 and Table 1). Although
[Pd]mpg-C3N4 has the same metal loading as Pd-HHDMA/
TiS, the latter material contains particles with an average
diameter of approximately 8 nm deposited on titanium
silicate (SBET = 229 m2 gÀ1). Not all of the surface atoms in
the HHDMA-modified nanoparticles take part in the reaction: In fact, the ligand blocks about 75 % of the metal
surface, leading to ensembles with a size of approximately
0.8 nm.[2f] The reference Pd-Pb/CaCO3 catalyst contains
significantly more palladium as well as lead (5 wt % Pd,
3 wt % Pb, surface Pd/Pb ratio = 1.3). The catalyst exhibits
a total area of 10 m2 gÀ1 and a pore volume of 0.03 cm3 gÀ1. The
Pd-Pb nanoparticles are homogeneously distributed over
CaCO3 and have an average diameter of 14 nm (metal
dispersion: 6–8 %). Similar to the role of HHDMA, Pb
isolates the Pd sites, leading to ensembles with a size of
approximately 1 nm. Finally, Pd/A2O3 contains uniformly
sized bare palladium nanoparticles with average diameters of
about 3 nm. The particles are deposited on porous alumina
(Vpore = 0.64 cm3 gÀ1) with a total surface area of 170 m2 gÀ1. In
all these cases, XAS analyses detected only PdÀO (1.8–2.0 Š)
and PdÀPd (2.5–2.7 Š) bonds (Figure 2 a and Figure S4).
To confirm the various degrees of metal dispersion in the
four samples, CO chemisorption studies were conducted.
During pulse chemisorption, [Pd]mpg-C3N4 showed no
adsorption of the probe molecules, whereas the other
reference materials showed a quantifiable CO uptake, in
accordance with the degrees of dispersion determined by
microscopy (Table 1). The absence of CO uptake over
[Pd]mpg-C3N4 was confirmed by infrared spectroscopy in
transmission mode (Figure S5). This result indicates that
individual Pd atoms behave differently than those in a conventional nanoparticle, in line with the recent work by Schlçgl
and co-workers.[8c] As the Pd atoms lay below the plane
formed by the six-fold cavities, the cage is enriched in electron
density, and this prevents carbon monoxide from getting close
to the palladium atoms. This shield is very effective, and CO
cannot adsorb (the DFT-calculated adsorption energy is
largely endothermic, by 0.86 eV). Further insights into the
Angew. Chem. Int. Ed. 2015, 54, 11265 –11269

material structure were obtained by X-ray photoelectron
spectroscopy (Figure 2 b). A single Pd 3d5/2 binding energy of
335.5 eV, corresponding to Pd0, was observed for PdHHDMA/TiS and Pd/Al2O3. The signal is broader in the
case of Pd-Pb/CaCO3 owing to the presence of Pd2+ÀPb
species. In contrast, the XPS signal was not very intense for
[Pd]mpg-C3N4, likely because the detection of isolated Pd
species at low concentrations is below the detection limit of
the instrument. Sputtering with an Ar+ beam was conducted
to remove surface layers in a stepwise fashion to gain access to
the bulk composition of the material. In this case, an increase
in the palladium content with depth could be observed
(Figure S6), from a Pd content of 0.15 at. % in the surface to
0.5 at. % in the bulk. This finding confirms that part of the
palladium phase is homogeneously incorporated between the
graphitic layers of the support, in agreement with the DFT
calculations. Furthermore, during sputtering, the Pd 3d5/2 peak
is slightly shifted to higher binding energies (336.4 eV). This
could be due to quantum effects that appear for “particles”
smaller than 2 nm, although a Pd2+ contribution cannot be
excluded.
To illustrate the benefits of single-site catalysis, the
[Pd]mpg-C3N4 catalyst was used for the hydrogenation of 1hexyne (Figure 3 a), a reference compound that well represents typical alkynes employed for the manufacture of highadded-value compounds in the fine chemical and pharmaceutical industries. The rate of reaction at 303 K and 1 bar is
three orders of magnitude larger than that of other catalytic
systems (based on Ag, Au, CeO2).[9] This demonstrates that
the single-site catalyst can be applied in the manufacture of
fine chemicals and pharmaceuticals. Furthermore, below
363 K and 2 bar, the selectivity to 1-hexene is nearly 100 %,
pointing to the resistance of the material towards the
formation of b-hydrides. These results can be appreciated
when they are compared to the performance of benchmark
Pd-based hydrogenation catalysts (see also Figure S7). For
the hydrogenation of 1-hexyne at 343 K and 5 bar, [Pd]mpgC3N4 and Pd-HHDMA/TiS display similar activities (1.41 ”
103 and 1.38 ” 103 molproduct molPdÀ1 hÀ1, respectively) and high
olefin selectivities (90 %). In contrast, while the Lindlar
catalyst shows an excellent degree of selectivity to the
terminal
alkene
(90 %),
its
activity
(0.34 ”
103 molproduct molPdÀ1 hÀ1) is much lower in spite of the high
Pd content (5 wt %). The non-promoted Pd/Al2O3 catalyst is
fairly active (0.96 ” 103 molproduct molPdÀ1 hÀ1), but poorly selective to the corresponding olefin (69 %), yielding various
isomers and the alkane as side products. The resistance of
[Pd]mpg-C3N4 towards metal loss (a possible deactivation
pathway for single-site catalysts) was assessed in a catalytic
run at 343 K and 5 bar for 20 h (Figure 3 b). No decrease in
alkyne conversion and alkene selectivity could be observed,
pointing to the absence of any metal aggregates (Figure 3 c).
This catalytic performance was confirmed for the hydrogenation of 2-methyl-3-butyn-2-ol to 2-methyl-3-buten-2-ol
and 3-hexyne to cis-3-hexene, demonstrating the chemo- and
stereoselectivity of the catalyst in alkyne hydrogenation (cis/
trans ratio > 20), and finally applied to the hydrogenation of
nitrobenzene to aniline (Figure S8; see the Supporting
Information for the reaction conditions).

 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.org

11267

.
Angewandte
Communications

Figure 3. a) Reaction rates (in 103 molproduct molPdÀ1 hÀ1; left), selectivity
to 1-hexene (in %; right), and b) stability over the course of 20 h at
343 K and 5 bar during the hydrogenation of 1-hexyne over [Pd]mpgC3N4. The contour maps in (a) were obtained through spline interpolation of the experimental points shown as black dots. c) Representation showing isolated Pd species on carbon nitride (top) and alumina
(bottom). Whereas the atoms on the alumina support are unstable
and tend to aggregate, forming a Pd cluster, this does not occur with
mpg-C3N4. d) Energy profile for the hydrogenation of acetylene over
a single Pd atom. Al green, C dark gray, H white, N purple, O red,
Pd blue.

11268

www.angewandte.org

To be able to rationalize the outstanding hydrogenation
performance of [Pd]mpg-C3N4, we conducted DFT calculations using acetylene as a surrogate alkyne (Figure 3 d). The
reasons for the choice of a smaller alkyne are the following:
1) Several reaction paths were investigated in this work, with
and without considering the promoting role of the support,
complicating the calculations; 2) the large rumpling of the
carbon nitride, together with the conformational flexibility of
larger alkynes, renders the proper location of the transition
states difficult; and, most importantly, 3) the adsorption
energy of acetylene and 1-hexyne were found to be within
a few meV. Therefore, the results for acetylene can be easily
transferred to more complex compounds. The reaction closely
follows the pathway of homogeneous systems[10a] and takes
place on the isolated Pd species in the six-fold nests
(alternative configurations are described in the Supporting
Information). On the isolated Pd sites, the hydrogenation
occurs by the coordination of molecular hydrogen by
À0.30 eV (B), which undergoes heterolytic dissociation (C),
leaving one of the hydrogen atoms bound to an N atom in the
lattice and the second one on a Pd atom. Thus, there is an
active participation of the support, which fulfills the role that
ligands adopt in homogeneous catalysis. The N atom thus
becomes positively polarized and the aromaticity of the
heptazine core is partially lost. As a consequence, the Pd atom
is partially pulled out of the cavity. This step is almost barrierfree with respect to the gas-phase molecule when the zeropoint energy is considered and exothermic by À1.34 eV with
respect to the gas-phase reactants. The acetylenic compound
can be adsorbed in a highly activated form (H-C-C angle =
1538) with release of 0.5 eV (D). The H atom is transferred to
8
the organic moiety, leading to C2H3 (E) and C2H4 (F). The
energy barrier of this last step is 0.5 eV. Afterwards, the
alkene product desorbs (G). Therefore, the coordination of
the product without activation,[10b] preventing over-hydrogenation and oligomerization, and the absence of electronrich palladium centers as determined by Bader analysis[10c] are
key for the high selectivity of [Pd]mpg-C3N4.
In conclusion, we have shown that isolated Pd atoms can
be confined into the six-fold cavities of mpg-C3N4 in a stable
manner, enabling the design of a single-site Pd catalyst for the
flow hydrogenation of alkynes and nitroarenes. The material
surpasses the activity of conventional heterogeneous catalysts
based on nanoparticles, while maintaining an outstanding
degree of product selectivity (> 90 %) and circumventing
typical instabilities that occur with the miniaturization of the
ensemble size. The performance was rationalized at the
molecular level by DFT calculations, showing that the high
activity and selectivity is due to the facile hydrogen activation
and alkyne adsorption on the atomically dispersed Pd sites.
The support simulates enzymatic environments, such as those
in porphyrins, and simultaneously acts as a spacer, distributing
the active sites almost homogeneously, and as a ligand, both
inhibiting the adsorption of potential poisons (CO) and
promoting the activation of H2, which converts the metal into
the active form. We believe that the proof-of-concept study
based on the [Pd]mpg-C3N4 catalyst can be extrapolated for
the design of new heterogeneous single-site catalysts for

 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. Int. Ed. 2015, 54, 11265 –11269

Angewandte

Chemie

a variety of reactions, including continuous hydrogenations,
CÀC coupling reactions, and beyond.
[3]

Experimental Section

Details on the catalyst preparation, characterization, testing, and
theoretical calculations are given in the Supporting Information.
Mesoporous polymeric graphitic carbon nitride was prepared by
thermally induced self-condensation of cyanamide using colloidal
silica as the template.[6] The material was dispersed in deionized water
and mixed with an aqueous solution of PdCl2 and NaCl. The resulting
suspension was kept at room temperature for 1 h under vigorous
stirring in an ultrasound bath. After the addition of NaBH4, the slurry
was filtered and dried, yielding [Pd]mpg-C3N4. The three-phase
hydrogenation of 1-hexyne was studied in a continuous-flow threephase reactor[2f,g, 9] under the following conditions: catalyst mass
Wcat = 0.1 g (particle size: 0.2–0.4 mm), temperature T = 303–363 K,
total pressure P = 1–8 bar, hydrogen flow rate FG = 24 cm3 minÀ1, and
liquid flow rate FL = 1 cm3 minÀ1. Density functional theory calculations conducted with the Vienna ab initio simulation package[11]
enabled the modelling of the structure of the material and its
reactivity.

[4]

[5]

[6]

Acknowledgements
Dr. Frank Krumeich (ETH Zurich) and Dr. Roland Hauert
(EMPA Dübendorf) are acknowledged for the STEM and
XPS analyses, respectively. Delft Solids Solutions B.V. is
thanked for conducting CO chemisorption analyses. Financial
support from the ETH Zurich, the Spanish Ministerio de
Economía y Competitividad (CTQ2012-33826), BSC-CNS,
and the ICIQ Foundation is also acknowledged.

[7]
[8]

Keywords: carbon nitride · flow chemistry · palladium ·
selective hydrogenation · single-site catalysis

[9]

How to cite: Angew. Chem. Int. Ed. 2015, 54, 11265 – 11269
Angew. Chem. 2015, 127, 11417 – 11422

[10]

[1] R. A. Sheldon, H. van Bekkum, Fine Chemicals through
Heterogeneous Catalysis, Wiley-VCH, Weinheim, 2001, p. 351.
[2] a) J. A. Anderson, J. Mellor, R. P. K. Wells, J. Catal. 2008, 261,
208; b) D. Teschner, J. Borsodi, A. Wootsch, Z. RØvay, M.
Hävecker, A. Knop-Gericke, S. D. Jackson, R. Schlçgl, Science
2008, 320, 261; c) M. Armbrüster, K. Kovnir, M. Behrens, D.
Teschner, Y. Grin, R. Schlçgl, J. Am. Chem. Soc. 2010, 132,
14745; d) M. García-Mota, B. Bridier, J. PØrez-Ramírez, N.
López, J. Catal. 2010, 273, 92; e) P. T. Witte, P. H. Berben, D.

Angew. Chem. Int. Ed. 2015, 54, 11265 –11269

[11]

Vogt, J. W. Geus, Top. Catal. 2012, 55, 505; f) G. VilØ, N. AlmoraBarrios, S. Mitchell, N. López, J. PØrez-Ramírez, Chem. Eur. J.
2014, 20, 5926; g) G. VilØ, N. Almora-Barrios, N. López, J. PØrezRamírez, ACS Catal. 2015, 5, 3767.
´
a) A. Borodzinski, Catal. Lett. 2001, 71, 169; b) N. Semagina, L.
Kiwi-Minsker, Catal. Rev. Sci. Eng. 2009, 51, 147.
a) J. M. Thomas, R. Raya, D. W. Lewis, Angew. Chem. Int. Ed.
2005, 44, 6456; Angew. Chem. 2005, 117, 6614; b) J. M. Thomas,
Design and Applications of Single-Site Heterogeneous Catalysis,
Imperial College Press, London, 2012, p. 3; c) J. PØrez-Ramírez,
Angew. Chem. Int. Ed. 2013, 52, 9892; Angew. Chem. 2013, 125,
10076.
a) G. Kyriakou, M. B. Boucher, A. D. Jewell, E. A. Lewis, T. J.
Lawton, A. E. Baber, H. L. Tierney, M. Flytzani-Stephanopolous, E. C. H. Sykes, Science 2012, 335, 1209; b) S. F. J. Hackett,
R. M. Brydson, M. H. Gass, I. Harvey, A. D. Newman, K.
Wilson, A. F. Lee, Angew. Chem. Int. Ed. 2007, 46, 8593; Angew.
Chem. 2007, 119, 8747; c) X. Yang, A. Wang, B. Qiao, J. Li, J. Liu,
T. Zhang, Acc. Chem. Res. 2013, 46, 1740; d) E. J. Peterson, A. T.
DeLaRiva, S. Lin, R. S. Johnson, H. Guo, J. T. Miller, J. H.
Kwak, C. H. F. Peden, B. Kiefer, L. F. Allard, F. H. Ribeiro,
A. K. Datye, Nat. Commun. 2014, 5, 4885; e) H. Wei, X. Liu, A.
Wang, L. Zhang, B. Qiao, X. Yang, Y. Huang, J. Miao, T. Zhang,
Nat. Commun. 2014, 5, 5634.
a) F. Goettmann, A. Fischer, M. Antonietti, A. Thomas, Angew.
Chem. Int. Ed. 2006, 45, 4467; Angew. Chem. 2006, 118, 4579;
b) A. Thomas, A. Fisher, F. Gçttmann, M. Antonietti, J.-O.
Müller, R. Schlçgl, J. M. Carlsson, J. Mater. Chem. 2008, 18,
4893; c) Y. Li, X. Xu, P. Zhang, Y. Gong, H. Li, Y. Wang, RSC
Adv. 2013, 3, 10973.
Y. Wang, X. Wang, M. Antonietti, Angew. Chem. Int. Ed. 2012,
51, 68; Angew. Chem. 2012, 124, 70.
a) M. W. Tew, H. Emerich, J. A. van Bokhoven, J. Phys. Chem. C
2011, 115, 8457; b) M. Mavrikakis, P. Stoltze, J. K. Nørskov,
Catal. Lett. 2000, 64, 101; c) R. Arrigo, M. E. Schuster, Z. Xie, Y.
Yi, G. Wowsnick, L. L. Sun, K. E. Hermann, M. Friedrich, P.
Kast, M. Hävecker, A. Knop-Gericke, R. Schlçgl, ACS Catal.
2015, 5, 2740.
a) G. VilØ, J. PØrez-Ramírez, Nanoscale 2014, 6, 13476; b) G.
VilØ, S. Wrabetz, L. Floryan, M. E. Schuster, F. Girgsdies, D.
Teschner, J. PØrez-Ramírez, ChemCatChem 2014, 6, 1928.
a) J. López-Serrano, S. B. Duckett, A. Lledós, J. Am. Chem. Soc.
2006, 128, 9596; b) F. Studt, F. Abild-Pedersen, T. Bligaard, R. Z.
Sørensen, C. H. Christensen, J. K. Nørskov, Science 2008, 320,
1320; c) S. Abbet, A. Sanchez, U. Heiz, W.-D. Schneider, A. M.
Ferrari, G. Pacchioni, N. Rçsch, J. Am. Chem. Soc. 2000, 122,
3453.
G. Kresse, J. Hafner, Phys. Rev. B 1993, 47, 558.

Received: June 3, 2015
Published online: July 31, 2015

 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.org

11269