This is the pre-peer reviewed version of the following article: Angew. Chem. 2017, 129, 1801 –1805, which has been published in final form at
https://onlinelibrary.wiley.com/doi/pdf/10.1002/ange.201610552. This article may be used for non-commercial purposes in accordance with
Wiley Terms and Conditions for Use of Self-Archived Versions.

Zuschriften

Heterogeneous Catalysis Hot Paper

Angewandte

Chemie

Deutsche Ausgabe:
DOI: 10.1002/ange.201610552
Internationale Ausgabe: DOI: 10.1002/anie.201610552

Hybrid Palladium Nanoparticles for Direct Hydrogen Peroxide
Synthesis: The Key Role of the Ligand
Giacomo M. Lari+, BegoÇa Puÿrtolas+, Masoud Shahrokhi, NÇria LÛpez, and Javier PÿrezRamÌrez*
Abstract: Ligand-modified palladium nanoparticles deposited
on a carbon carrier efficiently catalyze the direct synthesis of
H2O2 and the unique performance is due to their hybrid
nanostructure. Catalytic testing demonstrated that the selectivity increases with the HHDMA ligand content from 10 % for
naked nanoparticles up to 80 %, rivalling that obtained with
state-of-the-art bimetallic catalysts (HHDMA = C20H46NO5P).
Furthermore, it remains stable over five consecutive reaction
runs owing to the high resistance towards leaching of the
organic moiety, arising from its bond with the metal surface. As
rationalized by density functional theory, this behavior is
attributed to the adsorption mode of the reaction intermediates
on the metal surface. Whereas they lie flat in the absence of the
organic shell, their electrostatic interaction with the ligand
result in a unique vertical configuration which prevents further
dissociation and over-hydrogenation. These findings demonstrate the importance of understanding substrate–ligand interactions in capped nanoparticles to develop smart catalysts for
the sustainable manufacture of hydrogen peroxide.

organic substrates, 2) use of green solvents such as water or
methanol, and 3) simplified purification.[1, 4] Its applicability
has been widely investigated as such or in tandem with
selective oxidation reactions using heterogeneous or, seldom,
homogeneous catalysts.[5] Albeit the number of advantages
over the traditional route, its industrial implementation has
been hindered by the limited selectivity of the most active
catalyst identified, that is, supported Pd nanoparticles
(NPs).[6] Indeed, besides facilitating the reaction between
molecularly adsorbed O2 and H atoms, Pd NPs favor O2
dissociation leading to undesired water formation
(Scheme 1).[7] To overcome this limitation, the addition of
a second metal, chiefly Au or Sn, which was shown to generate
active sites that prevent OˇO bond cleavage, resulted in the
most selective catalysts reported to date.[8] Still, the added
loading of costly and/or toxic metals hampers their widespread use.

Hydrogen peroxide, H O , attracts growing attention as
2

2

a green alternative to traditional stoichiometric oxidants in
a wide range of applications within the textile, pulp bleaching,
waste water treatment, metallurgy, cosmetic, and pharmaceutical industries. Owing to the advantages of its use, that is,
the high atom economy and the generation of water as the
only byproduct,[1] an increase of its global market value from
3.7 to 6.0 billion USD between 2014 and 2023 is forecasted.[2]
Nowadays, the production of H2O2 relies exclusively on the
anthraquinone process, which, despite being safe and suitable
for continuous operation, involves the use of quinones and
solvents resulting in energy-intensive purification steps and
the generation of high amounts of waste, thus being competitive only at large scale.[1, 3] The direct synthesis of H2O2 is
an appealing alternative that has the potential to be exploited
in decentralized plants at any scale due to the 1) absence of
[*] G. M. Lari,[+] Dr. B. Puÿrtolas,[+] Prof. Dr. J. Pÿrez-RamÌrez
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. Shahrokhi, Prof. Dr. N. LÛpez
Institute of Chemical Research of Catalonia (ICIQ) and
Barcelona Institute of Science and Technology
Av. PaÔsos Catalans 16, 43007 Tarragona (Spain)
[+] These authors contributed equally to this work.
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under http://dx.doi.org/10.
1002/anie.201610552.
Angew. Chem. 2017, 129, 1801 –1805

Scheme 1. Reactants, products, and relevant adsorbed intermediates
in the a) direct synthesis of H2O2 and the side reactions leading to
water formation by b) unselective O2 hydrogenation, c) H2O2 hydrogenation, and d) H2O2 decomposition over supported Pd catalysts.

In our quest to develop a superior catalyst, we were
attracted by nanostructured hybrid materials, incorporating
active metal NPs covered by an adsorbed layer of large
organic molecules, such as polymers or surfactants.[9] The
latter not only prevent the aggregation of the active phase, but
were also demonstrated to boost the selectivity in the
stereospecific hydrogenation of carbonyl groups, the partial
hydrogenation of alkynes, nitroaromatics, and unsaturated
epoxides, and the alcoholysis of silanes owing to the blockage
of unselective facets, the modification of the adsorption
energies of reactants, intermediates, and products, the precise
control of surface structures, or the interfacial electronic
effect.[10] Despite these encouraging results, they remain
overlooked for direct H2O2 synthesis, where only poly(vinyl
alcohol) (PVA)-capped Pd particles were recently suggested

⌫ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

1801

Angewandte

Zuschriften
to limit O2 dissociation.[11] Although the initial H2O2 selectivity was remarkable (ca. 95 %), it rapidly decreased due to
leaching of the organic moiety, resulting in particle agglomeration. Herein, we show for the first time the suitability of
carbon-supported, hexadecyl-2-hydroxyethyl-dimethyl ammonium dihydrogen phosphate (C20H46NO5P, HHDMA)capped Pd NPs as catalyst in the direct synthesis of H2O2.
This commercially available material benefits from the nontoxic nature of the ligand and its hybrid structure has been
proved stable in alkyne hydrogenation.[10b] Furthermore, we
demonstrate through the combination of experimental and
computational evidences based on density functional theory
(DFT) the crucial role of the ligand in attaining an efficient
catalyst. To this end, a series of Pd catalysts with 0.5–0.8 wt %
metal loading labelled as Pd-HHDMAn/C, where n = 1–5
denotes increasing HHDMA/Pd molar ratios,[12] and a catalyst
(1.0 wt % Pd loading) featuring naked NPs prepared by dry
impregnation (Pd/C) were evaluated. Additionally, a titanium
silicate-supported Pd material (Pd-HHDMA5/TiS) was confronted to its C-supported counterpart to unravel the role of
the support. Compositional analysis of the solids (Table 1)
Table 1: Characterization data of the Pd catalysts.
Catalyst

Pd[a]
[wt %]

HHDMA/Pd[a]
[mol molˇ1]

N[b]
[wt %]

DCO[c]
[%]

Pd/C
Pd-HHDMA1/C
Pd-HHDMA2/C
Pd-HHDMA3/C
Pd-HHDMA4/C
Pd-HHDMA5/C
Pd-HHDMA5/TiS

1.0
0.7
0.8
0.6
0.5
0.6
0.5

0.0
0.7
1.0
4.3
8.2
11.1
8.3

0.0
0.1
0.2
0.3
0.4
0.7
0.5

5
12
13
12
13
10
14

[a] Inductively coupled plasma-optical emission spectroscopy (ICPOES). [b] Elemental analysis. [c] Pd dispersion by CO pulse chemisorption.

revealed that both the N content and HHDMA/Pd molar
ratio increased from Pd-HHDMA1/C to Pd-HHDMA5/C in
line with the higher amount of ligand used in the preparation.
Transmission electron microscopy (TEM) (Figure 1) showed
that the Pd NPs possessed a uneven spherical shape with an
average diameter of around 4 nm. Despite they were

Chemie

unevenly distributed over the carbon support, no evidence
of particle agglomeration was observed, which is related to
the function of the ligand moiety in separating the metal
phase during the synthesis. Indeed, examination of the Pd
dispersion (Table 1) revealed similar values for all the
HHDMA-modified samples,[13] which was higher than that
attained via conventional impregnation (Pd/C).
Assessment of the catalytic response of the Pd/C catalyst
in the direct synthesis of H2O2 displayed comparable results
to those previously reported for these materials.[6] Upon an
increase of the HHDMA ligand content (Figure 2 a), augmented H2O2 productivity as well as selectivity to H2O2 was
observed. In line with its highest ligand content, PdHHDMA5/C
exhibited
the
highest
productivity
(8.4 molH2 O2 hˇ1 gPdˇ1) and selectivity (80 %), surpassing most
of the state-of-the-art materials (Table 2). The H2O2 selectivTable 2: Comparison of the performance of Pd-HHDMA5/C with state-ofthe-art catalysts.
Catalyst

Pd
[wt %]

Ligand

SH2 O2
[%]

rH2 O2 formation
[molH2 O2 hˇ1 gPdˇ1]

Ref.

Pd-HHDMA5/C
Pd/C fibers
Pd/N-doped C
Pd/C
AuPd/C
SnPd/TiO2

0.6
1.0
0.9
5.0
2.5
3.0

HHDMA
PVA
PVA
none
none
none

80
n.a.
45
n.a.
80
96

8.4
0.6
14.2[a]
2.7
4.4
2.0

see [b]
[13]
[11a]
[6]
[8b]
[8c]

[a] H2SO4 was added to the reaction medium as promoter. [b] This work.

ity was comparable to that attained for the bimetallic AuPd
system but lower than that recently reported for SnPd alloys
supported on TiO2. Furthermore, the H2O2 hydrogenation
activity was hindered upon increased amount of HHDMA
and was negligible for the best-performing Pd-HHDMA5/C
catalyst. Additional H2O2 decomposition experiments
showed that the rate of this reaction did not depend on the
ligand content neither on the carbon support, despite a different C support was presumably used for the commercial
preparation of Pd-HHDMA5/C. Accordingly, similar results
were attained upon evaluation of the sole activated carbon
(see Table S1 in the Supporting Information). The key
influence of the support on this transformation was confirmed

Figure 1. TEM of a) Pd-HHDMA3/C, b) fresh and c) used Pd-HHDMA5/C. The inset shows the particle size distributions and the scale bar applies
to all the micrographs.

1802

www.angewandte.de

⌫ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. 2017, 129, 1801 –1805

Zuschriften

Angewandte

Chemie

Figure 2. a) Effect of the HHDMA/Pd molar ratio on the rates of H2O2 formation, decomposition, and hydrogenation and on the H2O2 selectivity.
b) Fraction of the initial activity and H2O2 selectivity over Pd-HHDMA5/C upon five consecutive runs. Conditions: T = 273 K, P = 40 bar, 3.75 vol %
H2 ; 7.50 vol % O2 ; 88.75 vol % N2, mcat = 10 mg in 5 cm3 of 67 vol % methanol in water.

by the high decomposition rates observed over the bare TiS,
which limited the productivity of Pd-HHDMA5/TiS to
0.3 molH2 O2 hˇ1 gPdˇ1. Finally, the stability of the PdHHDMA5/C catalyst was evaluated by monitoring the
performance over five consecutive runs (Figure 2 b), revealing
that both the productivity and selectivity to H2O2 were
maintained. Additional characterization of the used catalyst
showed that, in contrasts to previous reports,[11b] the morphology of the Pd NPs was fully preserved (Figure 1 c) and no
leaching of Pd or ligand was observed upon ICP-OES analysis
of the reaction mixtures.
DFT calculations performed over Pd(111) and Pd(111)HHDMA surfaces enabled to ascertain the key role of the
ligand at the molecular scale. The complex structure of the
hybrid surface has been disclosed previously, and was shown
to comprise H2PO4ˇ groups strongly adsorbed on the metal
surface, whose charge is compensated by the ammonium
headgroups of the organic surfactant.[10b] The direct synthesis
of H2O2 proceeds via O2 adsorption followed by two
subsequent hydrogenation steps leading to an hydroperoxy
(OOH) and to H2O2. Overall, the activation energy for this
path is lower on the naked (0.74 eV, 1 eV = 96 kJ molˇ1) than
on the hybrid Pd surface (1.18 eV, Table S2). Nevertheless,
the different selectivity observed experimentally can be
explained considering the activation barriers for the selective

and unselective transformations of the OOH intermediate
(Figure 3 and Table S2). The over-hydrogenation and decomposition pathways are favored over Pd(111) compared to the
formation of H2O2 (Figure S1). In contrast, on the Pd(111)HHDMA surface, not only the hydrogenation of the adsorbed
OOH intermediate to H2O2 is preferred versus the overhydrogenation to water, but also the desorption of the
product is energetically advantageous over its further hydrogenation and/or decomposition. Notably, the presence of the
ligand increases the activation energy of the two side
reactions. Furthermore, water formation attributed to O2
dissociation is only feasible on the naked Pd surface, while
it is thermodynamically impeded on the Pd(111)-HHDMA
surface (Table S3). Other hydrogenation pathways, involving
the protonation of adsorbed species due to interfacial acidity,[14] are likely not taking place if the low energy barriers for
H2 dissociation and OOH formation are considered.
The differences in the energy profiles between the naked
and the hybrid Pd NPs arises from the adsorption configuration of O2 and the hydroperoxy radical. These intermediates flatten on the Pd(111) surface (Figures 3 and S2), thus
favoring the cleavage of the OˇO bond. In contrast, on
Pd(111)-HHDMA, the electrostatic interaction of the adsorbed intermediates with the H2PO4ˇ group and the quaternary ammonium headgroups of the HHDMA molecule

Figure 3. Central panel: activation energies for the direct synthesis of H2O2 and the side reactions leading to water formation by H2O2
hydrogenation and decomposition. The drawings represent the adsorption configuration of the hydroperoxy (OOH) radical on Pd(111) and
Pd(111)-HHDMA surfaces. Color code: H in white, C in grey, O in red, P in yellow, and Pd in blue.
Angew. Chem. 2017, 129, 1801 –1805

⌫ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.de

1803

Zuschriften
enables the peculiar vertical configuration of the intermediate, impeding the cleavage of the OˇO bond and its overhydrogenation. This is reflected by the decrease of the
average atomic O–N distance from 6.64 to 5.95 ä between
the horizontal and the vertical adsorption (Figure S3). Such
a ligand-imposed adsorption geometry resembles that
observed at the active site of metalloenzymes, where the
allowed reaction paths are determined by the directionality of
the interaction between the substrate, the metal center, and
the amino acid residues in the surroundings. Whether this
behavior is unique to the HHDMA ligand clearly merits
further exploration, especially as the interest in the role of
electric fields in boosting selectivity is growing.[15] Finally, it
should be noted that, apart from the molecular reorientation
on ligand-decorated structures, previously proposed as selectivity descriptors,[10e] steric hindrance caused by the presence
of the ligand can also contribute to the observed selectivity
patterns as observed on earlier studies with bulkier molecules.[10b] However, the small size of the molecules involved in
this case results in a minor impact of the ensemble confinement on the selective performance.
In conclusion, we have shown the outstanding activity,
selectivity, and stability of carbon-supported HHDMAcapped Pd NPs in the direct synthesis of H2O2. Their
performance was correlated with the amount of HHDMA
ligand on the surface of the metal. DFT attributed this
behavior to the adsorption mode of the reaction intermediates, deriving from their electrostatic interactions with the
surfactant. The attained results not only demonstrate the
applicability of hybrid materials for the direct synthesis of
H2O2 but also highlight the importance of understanding the
molecular role of the organic ligand for the design of novel
nanocatalysts.

Experimental Section
Pd-HHDMAn/C (n = 1–4) catalysts were prepared by immobilizing onto activated carbon (NORIT) of colloids obtained using
Na2PdCl4 (Sigma–Aldrich) as Pd precursor and varying amounts of
HHDMA (Sigma–Aldrich) as stabilizing and reducing agent following the experimental protocol described elsewere.[12] Pd-HHDMA5/C
(NanoSelect LF 100TM) and Pd-HHDMA5/TiS (NanoSelect LF
200TM), prepared industrially over unspecified carbon and titanium
silicate supports following a similar procedure, were used as received.
Pd/C was prepared by dry impregnation of activated carbon using
Na2PdCl4 as the metal precursor. The compositional and structural
properties were studied by ICP-OES, elemental analysis, CO pulse
chemisorption, and TEM measurements. Direct synthesis of H2O2
was conducted in a batch reactor at 273 K and a total pressure of
40 bar (3.75 vol % H2 ; 7.50 vol % O2 ; 88.75 vol % N2) using a 67 vol %
methanol/water mixture as solvent and 10 mg of catalyst. H2O2
(60 mm in 67 vol % methanol/water) hydrogenation and decomposition experiments were performed at 40 bar under 3.75 vol % H2/N2 or
N2 atmosphere, respectively. The reaction mixture was stirred for
30 minutes and the H2O2 formed was quantified by titration with
KMnO4. The stability of Pd-HHDMA5/C was assessed upon 5
consecutive catalytic runs. DFT calculations were conducted with
the Vienna ab initio simulation package (VASP),[16] using the RPBE
functional and a kinetic cut-off energy of 450 eV. The Pd(111) slabs
were constructed with five atomic layers which were periodically
repeated and separated by a vacuum gap of 15 ä. The surface
contained four adsorbed HHDMA molecules, whose alkyl chain has

1804

www.angewandte.de

Angewandte

Chemie

been shortened to lighten the calculations. The optimization thresholds were 10ˇ5 eV and 0.001 eV äˇ1 for electronic and ionic relaxations, respectively. Further details on catalyst preparation, characterization, testing, and DFT simulations are given in the Supporting
Information.

Acknowledgements
This work was funded by the Swiss National Science
Foundation (grant number 200020-159760), the MINECO
(grant number CTQ2015-68770-R) and Marie Curie–
COFUND (grant number 291787-ICIQ-IPMP, M.S.). The
Scientific Center for Optical and Electron Microscopy at
ETH Zurich, ScopeM, is acknowledged for the use of its
facilities. Dr. S. Mitchell is thanked for the TEM analyses.

Conflict of interest
The authors declare no conflict of interest.
Keywords: colloidal synthesis · density functional calculations ·
hybrid catalysts · palladium nanoparticles · peroxides
How to cite: Angew. Chem. Int. Ed. 2017, 56, 1775 – 1779
Angew. Chem. 2017, 129, 1801 – 1805
[1] Y. Yi, L. Wang, G. Li, H. Guo, Catal. Sci. Technol. 2016, 6, 1593 –
1610.
[2] “Hydrogen Peroxide Market—Global Industry Analysis, Size,
Share, Growth, Trends and Forecast, 2015 – 2023”, can be found
under https://globenewswire.com/.
[3] J. GarcÌa-Serna, T. Moreno, P. Biasi, M. J. Cocero, J.-P. Mikkola,
T. O. Salmi, Green Chem. 2014, 16, 2320 – 2343.
[4] J. M. Campos-Martin, G. Blanco-Brieva, J. L. G. Fierro, Angew.
Chem. Int. Ed. 2006, 45, 6962 – 6984; Angew. Chem. 2006, 118,
7116 – 7139.
[5] a) Q. Chen, E. J. Beckman, Green Chem. 2008, 10, 934 – 938;
b) S. Shibata, T. Suenobu, S. Fukuzumi, Angew. Chem. Int. Ed.
2013, 52, 12327 – 12331; Angew. Chem. 2013, 125, 12553 – 12557;
c) S. Fukuzumi, Biochim. Biophys. Acta Bioenerg. 2016, 1857,
604 – 611.
[6] E. N. Ntainjua, J. K. Edwards, A. F. Carley, J. A. LÛpez-Sµnchez,
J. A. Moulijn, A. A. Herzing, C. J. Kiely, G. J. Hutchings, Green
Chem. 2008, 10, 1162 – 1169.
[7] A. Plauck, E. E. Stangland, J. A. Dumesic, M. Mavrikakis, Proc.
Natl. Acad. Sci. USA 2016, 113, E1973 – E1982.
[8] a) J. K. Edwards, J. Pritchard, L. Lu, M. Piccinini, G. Shaw, A. F.
Carley, D. J. Morgan, C. J. Kiely, G. J. Hutchings, Angew. Chem.
Int. Ed. 2014, 53, 2381 – 2384; Angew. Chem. 2014, 126, 2413 –
2416; b) J. K. Edwards, B. Solsona, E. Ntainjua, A. F. Carley,
A. A. Herzing, C. J. Kiely, G. J. Hutchings, Science 2009, 323,
1037 – 1041; c) S. J. Freakley, Q. He, J. H. Harrhy, L. Lu, D. A.
Crole, D. J. Morgan, E. N. Ntainjua, J. K. Edwards, A. F. Carley,
A. Y. Borisevich, C. J. Kiely, G. J. Hutchings, Science 2016, 351,
965 – 968; d) M. GarcÌa-Mota, N. LÛpez, Phys. Chem. Chem.
Phys. 2011, 13, 5790 – 5797; e) J. Li, T. Ishihara, K. Yoshizawa, J.
Phys. Chem. C 2011, 115, 25359 – 25367; f) J. Li, K. Yoshizawa,
Catal. Today 2015, 248, 142 – 148.
[9] C.-J. Jia, F. Sch¸th, Phys. Chem. Chem. Phys. 2011, 13, 2457 –
2487.
[10] a) T. B¸rgi, A. Baiker, Acc. Chem. Res. 2004, 37, 909 – 917; b) G.
Vilÿ, N. Almora-Barrios, S. Mitchell, N. LÛpez, J. Pÿrez-RamÌrez,

⌫ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Angew. Chem. 2017, 129, 1801 –1805

Zuschriften
Chem. Eur. J. 2014, 20, 5926 – 5937; c) T. Taguchi, K. Isozaki, K.
Miki, Adv. Mater. 2012, 24, 6462 – 6467; d) Z. Niu, Y. Li, Chem.
Mater. 2014, 26, 72 – 83; e) S. T. Marshall, M. O Brien, B. Oetter,
A. Corpuz, R. M. Richards, D. K. Schwartz, J. W. Medlin, Nat.
Mater. 2010, 9, 853 – 858; f) G. Chen, C. Xu, X. Huang, J. Ye, L.
Gu, G. Li, Z. Tang, B. Wu, H. Yang, Z. Zhao, Z. Zhou, G. Fu, N.
Zheng, Nat. Mater. 2016, 15, 564 – 569.
[11] a) S. Abate, R. Arrigo, M. E. Schuster, S. Perathoner, G. Centi,
A. Villa, D. Sub, R. Schlçgl, Catal. Today 2010, 157, 280 – 285;
b) S. Abate, M. Freni, R. Arrigo, M. E. Schuster, S. Perathoner,
G. Centi, ChemCatChem 2013, 5, 1899 – 1905.
[12] a) P. T. Witte, P. H. Berben, S. Boland, E. H. Boymans, D. Vogt,
J. W. Geus, J. G. Donkervoort, Top. Catal. 2012, 55, 505 – 511;
b) P. T. Witte, WO2009096783 A1, 2009.

Angew. Chem. 2017, 129, 1801 –1805

Angewandte

Chemie

[13] A. Villa, S. J. Freakley, M. Schiavoni, J. K. Edwards, C.
Hammond, G. M. Veith, W. Wang, D. Wang, L. Prati, N.
Dimitratos, G. J. Hutchings, Catal. Sci. Technol. 2016, 6, 694 –
697.
[14] D. Albani, Q. Li, G. Vilÿ, S. Mitchell, N. Almora-Barrios, P. T.
Witte, N. LÛpez, J. Pÿrez-RamÌrez, Green Chem. 2016, DOI:
10.1039/C6GC02586B.
[15] S. Shaik, D. Mandal, R. Ramanan, Nat. Chem. 2016, 8, 1091 –
1098.
[16] G. Kresse, J. Furthm¸ller, Phys. Rev. B 1996, 54, 11169 – 11186.
Manuscript received: October 28, 2016
Revised: November 27, 2016
Final Article published: December 16, 2016

⌫ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.angewandte.de

1805