Direct synthesis of hydrogen peroxide over Pd/C catalyst prepared by selective adsorption deposition method

Article abstract of DOI:10.1016/j.jcat.2018.06.024

A new catalyst design based on selective adsorption deposition method was developed to achieve high reaction performance in the direct synthesis of hydrogen peroxide. The activity of the unprecedented Pd/C catalyst was superior to that of the conventionally prepared Pd/C catalysts, and the initial H₂O₂ productivity and H₂ selectivity reached as high as 8606 mmol H₂O₂/g Pd.h and 95.1%, respectively. This excellent activity may result from the intrinsic structural and electronic features of the active sites, i.e., the extremely small and monodispersed Pd nanoparticles with a high Pd²⁺/Pd⁰ ratio, which were realized by combining the selective adsorption of metal precursor cations on a negatively charged activated carbon surface and the subsequent homogeneous surface deposition of palladium hydroxide by the hydroxide ions that are slowly generated upon urea decomposition. The catalytic activity was significantly affected by the oxygen groups of the activated carbon support. The carboxyl groups do not efficiently suppress the unfavorable H-OOH dissociation but rather accelerate the H₂O₂ hydrolysis by forming hydrogen bonds with H₂O₂. Moreover, a sharp decrease in the reaction rates of H₂O₂ hydrogenation and direct synthesis of H₂O₂ was observed with the increase in the number of carboxyl groups on the activated carbon surface. This loss of activity, as confirmed by acid treatment and olefin hydrogenation experiments, implies that the carboxyl groups in close proximity to the active sites have a detrimental effect by hindering or poisoning the active sites.

Full text of DOI:10.1016/j.jcat.2018.06.024

Journal of Catalysis 365 (2018) 125–137
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Direct synthesis of hydrogen peroxide over Pd/C catalyst prepared
by selective adsorption deposition method
Seungsun Lee a,1, Hwiram Jeong b,1, Young-Min Chung a,
⇑
a
Department of Nano & Chemical Engineering, Kunsan National University, 558 Daehak-ro, Kunsan, Jeollabuk-Do 573-701, Republic of Korea
Department of Chemical Engineering, Hanyang University, Seoul 04763, Republic of Korea
b
a r t i c l e i n f o
a b s t r a c t
Article history:
A new catalyst design based on selective adsorption deposition method was developed to achieve high
reaction performance in the direct synthesis of hydrogen peroxide. The activity of the unprecedented
Received 27 April 2018
Revised 21 June 2018
Accepted 25 June 2018
Pd/C catalyst was superior to that of the conventionally prepared Pd/C catalysts, and the initial H
ductivity and H selectivity reached as high as 8606 mmol H /g Pd.h and 95.1%, respectively. This
excellent activity may result from the intrinsic structural and electronic features of the active sites, i.e.,
2 2
O pro-
2
2 2
O
2
+
0
the extremely small and monodispersed Pd nanoparticles with a high Pd /Pd ratio, which were realized
by combining the selective adsorption of metal precursor cations on a negatively charged activated car-
bon surface and the subsequent homogeneous surface deposition of palladium hydroxide by the hydrox-
ide ions that are slowly generated upon urea decomposition. The catalytic activity was significantly
affected by the oxygen groups of the activated carbon support. The carboxyl groups do not efficiently sup-
Keywords:
Hydrogen peroxide
Pd/C
Selective adsorption deposition
Surface modification
Catalyst poisoning
press the unfavorable H-OOH dissociation but rather accelerate the H
bonds with H . Moreover, a sharp decrease in the reaction rates of H
thesis of H was observed with the increase in the number of carboxyl groups on the activated carbon
O
2 2
hydrolysis by forming hydrogen
O
2 2
2 2
O hydrogenation and direct syn-
2 2
O
surface. This loss of activity, as confirmed by acid treatment and olefin hydrogenation experiments,
implies that the carboxyl groups in close proximity to the active sites have a detrimental effect by hin-
dering or poisoning the active sites.
Ó 2018 Elsevier Inc. All rights reserved.
1
. Introduction
routes; for instance, the anthraquinone process involves various
energy-intensive steps, the use of toxic compounds, and trouble-
Hydrogen peroxide (H
2
O
2
) is one of the greenest and most
some separation processes that generate waste materials [1,7].
Furthermore, mass production of hydrogen peroxide and ensuing
intensive concentration of up to 70 wt% are essential to ensure eco-
important oxidants, with applications ranging from bleaching
agents, detergents, disinfectants, and wastewater treatment agents
to semiconductor purification solutions and liquid rocket propel-
lants [1]. In the petrochemical industry, hydrogen peroxide has
been used in a number of oxidation reactions including the epoxi-
dation of propylene to propylene oxide [2,3]. Recently, it was also
reported that hydrogen peroxide could be effectively used in the
low temperature oxidation of methane to methanol [4–6]. As the
2 2
nomic efficiency [8]. As several reactions require 2–10 wt% H O as
oxidant [9], and the transportation of highly concentrated hydro-
gen peroxide is potentially dangerous, there are strong needs for
the development of an efficient process enabling the onsite pro-
2 2
duction of H O on a controllable scale [1,7,8,10].
In this context, the direct synthesis of hydrogen peroxide from
hydrogen and oxygen has drawn considerable attention because
hydrogen peroxide can in principle be produced through a simple
reaction route without any byproducts, except for water. In line
with the promising aspects of the reaction from an environmental
as well as economic viewpoint, the last decade witnessed a number
of studies focusing on realizing unique and sophisticated catalyst
designs [1,7,10]. For example, a number of novel approaches have
been envisaged: (i) fine tuning of active metal properties by bime-
tal formation [11–18], phase control [19–23], or ligand capping
[24]; (ii) suppression of side reactions by adopting core-shell
2 2
decomposition or hydrogenation of H O only produces water,
hydrogen peroxide has attracted much attention from an environ-
mental point of view in addition to its importance as an efficient
oxidizing agent.
Despite its environmentally benign nature, most hydrogen per-
oxide production has relied on environmentally unfavorable
⇑
Corresponding author.
E-mail address: ymchung@kunsan.ac.kr (Y.-M. Chung).
These authors contributed equally to this work.
1
https://doi.org/10.1016/j.jcat.2018.06.024
021-9517/Ó 2018 Elsevier Inc. All rights reserved.
0

1
26
S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
structures [25,26], carbon diffusion coating layers [27], as well as
solid acid supports [28–35]; and judicious combinations of these
approaches. However, despite considerable efforts over the last
decade, the direct synthesis of hydrogen peroxide remains a chal-
lenging reaction because the hydrogen peroxide yield under intrin-
sically safe and non-corrosive conditions is still unsatisfactory and
H
2
PdCl
water was mixed with a solution of the anionic metal precursor
(PdCl dissolved in 0.2 M HCl), and the mixed solution was stirred
4
[40–43]. Activated carbon (1 g) suspended in deionized
2
at 298 K for 2 h (Pd intake 5 wt%). After heating to 353 K, 0.5 M
NaOH solution was added to the suspension until the pH of the
mixed solution reached 12. After additional stirring for 2 h, a
reduction reaction was performed by bubbling hydrogen
(20 mL/min) at the same temperature for 2 h under vigorous stir-
ring. During these steps, the temperature of the suspension was
maintained at 353 K using a double-jacked reactor connected to
a circulator. At the end of the reduction, the catalyst was filtered
at room temperature, thoroughly washed with deionized water,
and dried overnight at 393 K.
the use of caustic additives and H
and safety issues [36,37].
2 2
/O mixtures causes corrosion
Herein, we propose an efficient catalyst design based on a selec-
tive adsorption deposition method to achieve high reaction perfor-
mance in the direct synthesis of hydrogen peroxide. The formation
of extremely small and monodispersed Pd nanoparticles on an acti-
vated carbon support was realized by the combination of the selec-
tive adsorption of metal precursor cations on the negatively charged
activated carbon surface and the homogeneous surface deposition of
palladium hydroxide by hydroxide ions that are slowly generated
upon urea decomposition. As will be discussed, the as-prepared
Pd/C catalyst showed superior reaction performance compared to
the Pd/C catalysts prepared by conventional method. The excep-
tional catalytic activity is comparable to that of previously reported
Pd catalysts supported on various carbon materials.
Moreover, by performing systematic characterizations and test
reactions, we elucidated the crucial role of the oxygen groups (par-
ticularly the carboxyl groups) of the activated carbon surface in the
catalytic performance. Most related studies focused on the changes
in the catalytic activity with respect to the different physicochem-
ical properties of various carbon materials [27,38,39]. We believe
that the systemically designed selective adsorption deposition
method can be extended to other catalyst designs that require
highly dispersed active metal sites.
The preparation of a Pd/C catalyst using the cationic palladium
3 4 3 2
precursor Pd(NH ) (NO ) was carried out by a modified selective
adsorption deposition method [44,45]. Activated carbon was sus-
pended in deionized water, and the suspension was adjusted to
3
pH 4 by the addition of HNO . After adding the tetraamminepalla-
dium(II) nitrate solution at 298 K, the mixed solution was stirred
for 2 h to allow the selective adsorption (Pd intake 5 wt%). After
the temperature was raised to 353 K, urea was added under vigor-
ous stirring. The pH of the resulting mixture gradually increased
and finally reached a constant value after 12–24 h; therefore, the
deposition procedure was continued for 24 h. The pH of the sus-
pension was monitored during this step, and the amount of urea
was adjusted to maintain the final pH at ca. 6.3. After the deposi-
tion, the slurry was filtered, thoroughly washed with deionized
water, and dried overnight at 393 K. The sample was reduced in
the presence of mixed N
473 K for 3 h (heating rate = 5 K/min).
2 2 2
/H gas (50 mol% H , 100 mL/min) at
2
.4. Characterization of the surface-modified activated carbon and
2
. Experimental
catalyst
2
.1. Chemicals
The textural properties of the surface-modified activated car-
bons, such as the specific surface area, pore volume, and pore
diameter, were determined by N adsorption using a BELSORP-
Max (BEL, Japan) at 77 K. Prior to adsorption, the samples were
degassed at 423 K overnight under vacuum. The distribution of
the oxygen-containing functional groups on the activated carbon
surface was determined by titration following Boehm’s method
Catalyst carrier grade activated carbon (Norit, surface area
2
3
2
1
007 m /g, pore volume 0.87 cm /g, pore diameter 3.47 nm) was
received from Carbot. Palladium chloride (99.9%) and tetraam-
minepalladium(II) nitrate solution (5.0 wt% as Pd) were supplied
by Strem-Chemicals. Cerium(IV) sulfate standard solution (0.25 N
in 2–3 N sulfuric acid), ferroin indicator (0.1 wt% solution in
water), sodium bromide (>99%), 1-decene (>97%), and heptane
[
46]. The temperature-dependent evolution of the decomposition
gases from an activated carbon domain was monitored by
temperature-programmed desorption-mass spectrometry (TPD-
Mass, BEL-CAT-Mass, Japan). Before the experiment, the samples
were dried using helium flowing at the rate of 100 mL/min at
(
(
(
(
(
99%) were purchased from Sigma-Aldrich. Hydrogen peroxide
34.5%), acetic acid (99.5%), nitric acid (60 wt%), hydrochloric acid
35–37%), sulfuric acid (96%), urea (99%), sodium hydroxide
98.0%), sodium carbonate anhydrous (99.5%), sodium bicarbonate
99%), and methanol (99.5%) were supplied by Samchun Chemicals
3
1
73 K for 2 h. The temperature of the samples was increased to
273 K at a ramping rate of 5 K/min, and the decomposition gases
in Korea. All the chemicals were used as-received without further
purification.
were detected using a mass spectrometer. Diffuse reflectance infra-
red Fourier transform spectra (DRIFTS) of the surface-modified
activated carbon were recorded using
a FT-IR spectrometer
2.2. Surface modification of the activated carbon support
(
Nicolet iS50 with DTGS detector, Thermo Fischer Scientific, USA)
equipped with a diffuse IR chamber (Pike Technology, USA).
UV–visible spectra were collected using a Hitachi U-2900 UV–Vis
spectrophotometer. Zeta potential measurements were carried
out in a wide pH range 2–12 using a zeta potential analyzer
For the introduction of oxygen-containing functional groups on
the surface, activated carbon was refluxed for 6 h in acid or acid/
oxidant solutions of different concentrations: 1 wt% HNO , 10 wt
HNO , and 10 wt% HNO /10 wt% H . After the surface modifi-
3
%
3
3
2 2
O
(
ELS-8000, Otsuka Electronics, Japan). The Pd content loaded on
cation, the resulting activated carbon was thoroughly washed with
water and dried in an oven at 383 K for 12 h. The samples were
stored in a desiccator under nitrogen before use.
the activated carbon was determined by inductively coupled
plasma atomic emission spectroscopy (ICP-AES) using a Perkin-
Elmer OPTIMA 7300 DV. For ICP analysis, palladium was extracted
3
using HNO /HCl mixture solution under microwave digestion. The
2
.3. Preparation of the Pd/C catalysts using anionic and cationic metal
Pd content on an activated carbon support was measured three
times and the average value was used. The oxygen/carbon atomic
ratio of the surface-modified activated carbon and the binding
energies of Pd in the Pd/C catalysts were determined by X-ray pho-
toelectron spectroscopy (MultiLab 2000, Thermo VG Scientific, UK)
precursors
The Pd/C catalyst was prepared via a well-known deposition-
precipitation method with the anionic palladium precursor

S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
127
with a pass energy of 30 eV, a step increment of 0.1 eV, and an Al
2.6. Olefin hydrogenation
anode. The XPS peak positions were referenced to the carbon 1s
peak at 284.6 eV. Powder X-ray diffraction (PXRD) analysis was
The hydrogenation of 1-decene was carried out in a liquid phase
mode under atmospheric pressure. First, 20 mg of catalyst, 15 mL
of 1-decene, and 15 mL of heptane were introduced to a 100-mL
three-necked round bottom flask equipped with a condenser and
a thermocouple. After heating to 323 K in a static mode, the
carried out with a Rigaku diffractometer using a CuK
source (40 kV and 60 mA). Transmission electron microscopy
TEM) analysis was performed using a transmission electron
microscope (Philips, Tecnai G2 F30, USA).
a radiation
(
reaction was started by stirring (600 rpm) with H
2
bubbling
(
5 mL/min) and maintained for 1 h. At the end of the reaction
2
.5. Direct synthesis of hydrogen peroxide, natural decomposition, and
run, the reaction product was extracted with a syringe, filtered
with a syringe filter, and analyzed with a GC (Agilent 7890A,
USA) equipped with an HP-5 capillary column (30 m ꢀ 0.32 mm
hydrogenation
The direct synthesis of hydrogen peroxide from hydrogen and
oxygen was carried out in a 100-mL high-pressure autoclave reac-
tor (Büchi Picoclave, Switzerland) connected to a circulator (Julabo
F25-HE, Germany). In a typical run, the catalyst (20 mg) was dried
at 393 K for 4 h before use, and then introduced to the reactor with
ID) and a flame ignition detector (FID) using N as a carrier gas.
2
3. Results and discussion
2
2
3
5 g of a methanol/water mixture (80 wt% methanol) containing
0 ppm NaBr. The reactor was tightly closed, pressurized up to
0 bar with nitrogen in static mode, and evacuated to remove
3.1. Surface modification of activated carbon and its characterization
The characteristics of a supporting material, in particular the
surface nature of the support, play a crucial role in determining
the efficiency of a metal-supported catalyst. Therefore, a range of
surface-modified activated carbon supports were prepared by
treating activated carbon with an acid or acid/oxidant solution of
different concentrations: 1 wt% HNO , 10 wt% HNO , and 10 wt%
residual gases. The purging step was repeated twice. During the
purging, the reaction temperature was adjusted to 275 K. After
the purging process, the nitrogen gas in the reactor was replaced
with reactant gases by co-feeding 5 mol% H
2
2 2
/N (30 mL/min) and
0 mol% O /N (10 mL/min) to the reactor using two high pressure
2
2
3
3
mass flow controllers (Brooks 5850E, Netherland), and the reactor
was pressurized again to 30 bar with the mixed gases. After 15 min
at a pressure of 30 bar, a reaction run was initiated by stirring
3 2 2
HNO /10 wt% H O . Table S1 shows the changes in the textural
properties of the activated carbon induced by the different surface
treatments. The surface-modified activated carbon mostly retained
its highly porous nature irrespective of the treatment conditions. A
more concentrated acid treatment may increase the amount of
oxygen-containing groups and enhance the acidity of the activated
carbon support [47]. However, further acid treatment was not per-
formed to avoid significant deformation of the pore structure of the
activated carbon, which may impact the effect of the surface nature
of activated carbon on the reaction performance.
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.jcat.2018.06.024.
Fig. 1 shows the changes in the concentration of oxygen groups
and the oxygen/carbon atomic ratio of the activated carbon surface
with the acid treatment process that were determined by titration
and XPS analysis, respectively. The titration result indicates that
the distribution and concentration of the oxygen groups on the
activated carbon surface strongly depend on the treatment condi-
tions. A small amount of lactone and hydroxyl groups were present
on the untreated activated carbon surface; however, the oxy-
genates on the surface of the activated carbon treated with 1 wt%
HNO3 consisted of hydroxyl and carboxyl groups, and the total
oxygen group concentration of the sample was ca. 11 times higher
than that of the untreated one. When activated carbon was treated
(
1200 rpm), and maintained for 0.5–2 h, typically 1 h. During the
reaction, the mixed gases were continuously supplied to the reac-
tor at total flow rate of 40 mL/min (H /O /N (mol%) =
.8/5.0/91.2), and the reaction pressure was maintained using a
a
2
2
2
3
back pressure regulator (Tescom, USA). At the end of the reaction
run, the catalyst was carefully separated from the reaction product
by filtration, washed thoroughly with methanol/water, and dried
for further use. The amount of hydrogen peroxide produced was
determined by titration with a cerium(IV) sulfate standard solution
and ferroin indicator. To determine the change in hydrogen con-
centration during the reaction, the off-gas collected in a gas bag
was analyzed with a gas chromatograph (GC, Agilent 6890, USA)
equipped with a Carboxen-1010 PLOT capillary column (30 m ꢀ
0
.53 mm ID) and a thermal conductivity detector (TCD) using Ar
as a carrier gas. Hydrogen conversion and selectivity were calcu-
lated according to the following equations using the hydrogen per-
oxide and hydrogen concentrations determined by the titration
and GC analysis, respectively. The hydrogen peroxide productivity
is defined as the amount of hydrogen peroxide produced per hour
2 2
divided by the Pd content (mmol H O /g Pd.h).
Moles of hydrogen reacted
Moles of hydrogen supplied
3
with 10 wt% HNO , the total oxygen group content was slightly
Hydrogen Conversion ð%Þ ¼
Hydrogen Selectivity ð%Þ ¼
ꢀ 100
reduced, but the carboxyl group content significantly increased
at the expense of the reduction in the hydroxyl groups. This result
suggests that the initially formed hydroxyl groups might be trans-
formed to carboxyl groups by further reaction with the acid. In
Moles of hydrogen peroxide produced
Moles of hydrogen reacted
activated carbon treated with 10 wt% HNO
amount of hydroxyl groups slightly increased, which indicates that
preferentially promotes hydroxyl group formation on the
3 2 2
and 10 wt% H O , the
ꢀ
100
2 2
H O
The H
reaction conditions, except for the use of 1 wt% of hydrogen
peroxide under nitrogen. In the H hydrogenation reaction,
mol% H /N (30 mL/min) and N (10 mL/min) gases were sup-
plied to the reactor containing 1 wt% of hydrogen peroxide in
methanol/water mixed solution. The H feed-in rate was the same
2
O
2
decomposition test was carried out under the same
activated carbon surface.
The XPS analysis also indicates that the O/C atomic ratio
gradually increases with the concentration of the acid, which is
consistent with the titration results, i.e., a larger number of
oxygen-containing groups can be incorporated on the activated
carbon surface by an acid treatment of higher concentration.
2 2
O
5
2
2
2
2
as that for the direct synthesis of hydrogen peroxide. The decom-
posed and hydrogenated hydrogen peroxide concentrations were
also determined by titrating the reaction mixture.
3
Although the activated carbon treated with 1 wt% HNO showed
the highest total oxygen group content (determined by chemical
titration), the O/C atomic ratio was considerably lower than that

1
28
S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
3 4
NH )
]2+, in combination with surface-modified activated carbon.
(
The catalysts are listed in Table 1.
The preparation of the Pd/C#Ax series using the anionic
palladium precursor was carried out via
a
conventional
deposition-precipitation (DP) method [43,52], in which the
2
ꢁ
[
PdCl
4
]
ions were first adsorbed on the activated carbon surface,
then deposited as palladium hydroxide species upon the addition
of NaOH, and finally reduced by hydrogen. On the other hand,
the Pd/C#Cx series using the cationic palladium precursor [Pd
2
+
(
NH
3
)
4
]
was prepared by the selective adsorption deposition
(SAD) method. In this SAD method, urea was used instead of NaOH
for the slow generation of hydroxide ions upon urea decomposi-
tion, which is advantageous for the gradual increase in the hydrox-
ide ion concentration during the palladium hydroxide deposition
step [45,53].
The surface charge of a support is of prime importance for the
efficient and selective loading of active metal precursors on a sup-
porting material [43,54]. Therefore, the zeta potential of the acti-
vated carbon was measured in a wide pH range before catalyst
preparation, and the isoelectric point (IEP) of the activated carbon
was found to be approximately pH 2.8, as shown in Fig. S3. Based
on the IEP of the activated carbon support, the initial pH of the
metal ion adsorption step was adjusted to 1.6 and 4 for the anionic
Fig. 1. Changes in the concentration of surface functional groups and the O/C
atomic ratio of activated carbon upon acid and/or oxidant treatment.
3 3
of activated carbon treated with 10 wt% HNO or 10 wt% HNO /10
2 2
wt% H O .
2
ꢁ
2+
(
4 3 4
[PdCl ] ) and cationic ([Pd(NH ) ] ) precursors, respectively.
To examine the changes in the oxygen group content over the
entire activated carbon domain under various acid treatment con-
Representative transmission electron microscopy (TEM) images
of the Pd/C catalysts and the changes in the Pd nanoparticle size
distribution with respect to the preparation method are presented
in Figs. 2–4. The formation of Pd nanoparticles was significantly
affected by the metal deposition method. For example, relatively
large and polydispersed nanoparticles (3–6.5 nm) were observed
in the Pd/C#Ax series prepared by the DP method, but very small
Pd nanoparticles (1–1.2 nm) were obtained in the Pd/C#Cx series
regardless of the oxygen group content of the activated carbon
support. This result clearly demonstrates the superiority of the
SAD method for the formation of extremely small and monodis-
persed Pd nanoparticles on an activated carbon support.
2
ditions, the temperature-dependent evolution of CO and CO from
the acid-treated activated carbon was monitored by TPD-Mass. As
shown in Fig. S1(a), the two peaks observed around 850–950 K and
1
000–1150 K in the CO-TPD-Mass analysis may result from the
decomposition of phenol/ether and carbonyl functionalities,
respectively. On the other hand, the CO evolution at relatively
2
lower temperatures (450–750 K) shown in Fig. S1(b) may be
ascribed to the thermal decomposition of carboxylic groups, and
the weak shoulder around 800–950 K implies the existence of car-
boxylic anhydrides and lactones [48,49]. The substantially reduced
evolution of CO and CO
of the 1 wt% HNO -treated activated carbon compared to those of
the activated carbon treated with 10 wt% HNO or 10 wt%
HNO /10 wt% H clearly shows that acid treatment of higher
2
gases during the thermal decomposition
The mechanism of metal deposition on activated carbon, in
terms of the adsorption behavior of the palladium precursor ions
and their subsequent transformation to palladium hydroxide spe-
cies, should be investigated to understand the critical role of the
metal deposition method on the Pd nanoparticle size and distribu-
tion, as described in Fig. 5. With the DP method, the initial pH of
the solution was adjusted to ca. 1.6 to allow the formation of a pos-
3
3
3
2 2
O
concentration greatly increases the oxygen content of an activated
carbon domain.
The DRIFTS-IR analysis presented in Fig. S2 also confirms the
different distributions of the oxygen groups on the activated car-
bon surface with respect to the acid treatment conditions. The car-
2
ꢁ
itively charged surface; thus, the [PdCl
4
]
ions can be adsorbed on
ꢁ1
the activated carbon surface by chemisorption as well as
physisorption, and/or ion exchange. However, during the deposi-
tion step, the pH of the solution was raised to ca. 12 by the
addition of NaOH. Indeed, at high pH, the interactions between
bonyl (C@O) stretching band at ca. 1780 cm
is ascribed to
carboxylic acid or its anhydride form. The broad band in the range
ꢁ1
1
100–1450 cm may result from the C@O stretching in carboxylic
acid and the lactone groups and/or from the OAH bending in the
2
ꢁ
ꢁ
1
the [PdCl
4
]
ions and the activated carbon support become
phenol groups. The peak around 1600 cm
is attributed to
weaker because the active carbon surface may develop a negative
the overlap between the stretching bands of polyaromatic C@C in
the highly conjugated carbonyl and/or quinone groups [50,51].
The apparent peaks in these three regions, observed for the acti-
2ꢁ
charge, and the non-selective adsorption of the [PdCl
4
]
ions may
increase. In addition, the high concentration of hydroxide ions
3 3
vated carbon treated with 10 wt% HNO and 10 wt% HNO /10 wt
%
2 2
H O , clearly indicate that large amounts of carboxyl groups
were newly introduced on the external surface of the activated car-
bon. The result closely matches the information obtained from the
titration, XPS, and TPD-mass experiments.
Table 1
Pd/C catalysts with various metal precursors and surface treatments of the activated
carbon support.
Metal precursor
Anionic (H PdCl
Surface treatment of support
Untreated
Catalyst
2
4
)
Pd/C#A1
Pd/C#A2
Pd/C#A3
Pd/C#A4
3
.2. Effect of metal deposition method on physicochemical properties
1
1
1
wt% HNO
3
0 wt% HNO
0 wt% HNO
of catalysts
3
3
/10 wt% H
2
O
O
2
To investigate the effect of the metal deposition method on the
Cationic (Pd(NH
3
)
4
3
(NO )
2
)
Untreated
Pd/C#C1
Pd/C#C2
Pd/C#C3
Pd/C#C4
reaction performance of a Pd/C catalyst in the direct synthesis of
hydrogen peroxide, various Pd/C catalysts were prepared using
two oppositely charged palladium precursors, [PdCl
1 wt% HNO
3
1
0 wt% HNO
3
3
2
ꢁ
10 wt% HNO
/10 wt% H
2
2
4
]
and [Pd

S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
129
Fig. 2. TEM images of Pd/C#Ax catalysts; the insets indicate the histograms of Pd nanoparticle size.
causes the fast and heterogeneous exchange of chloride and
hydroxide ligands, which results in the formation of various palla-
dium hydroxide species including polynuclear hydroxo-palladium
complexes [52]. Moreover, the nucleation and subsequent growth
of palladium hydroxide in solution may be more severe because
a high degree of local supersaturation of the growing nuclei is
inevitable [45].
nucleation of palladium hydroxide in the solution but also greatly
increases the possibility of surface deposition of the selectively
2
+
adsorbed [Pd(NH
3
)
4
]
ions [45,53]. Therefore, the mechanism of
the SAD method can be explained as follows: (i) the selective
2
+
adsorption of [Pd(NH
)
3 4
]
on the negatively charged activated car-
by
bon surface, (ii) the homogeneous surface deposition of Pd(OH)
x
slow exchange of the ammonia ligands with the hydroxide ions
As presented in Fig. 6, the formation of palladium hydroxide
was strongly affected by the pH of the solution. The dark reddish
precipitate observed for all the samples (Fig. 6(a)–(f)) indicates
the formation of palladium hydroxide. The precipitates turned
black after reduction, which indicates the formation of Pd0 (not
shown). The different degrees of precipitation within such a nar-
row pH range clearly suggest that the rate of formation of palla-
dium hydroxide is very sensitive to the pH. The UV–vis spectra
continuously generated from urea decomposition, and (iii) the
0
reduction of the uniformly deposited Pd(OH)
x
to Pd . Through
these processes, ultra-small and monodispersed Pd nanoparticles
can be supported on an activated carbon surface.
Fig. 7 clearly shows that the amount of Pd impregnated on an
activated carbon support strongly depends on the nature of the
support as well as the metal deposition method. In the Pd/C#Ax
series, similar amounts of Pd were impregnated regardless of the
(
Fig. 6(g)–(i)) show the gradual decrease and eventually the com-
oxygen group content of the activated carbon support, which
2ꢁ
2ꢁ
plete disappearance of the [PdCl
4
]
absorption band with increas-
implies that the non-selective adsorption of [PdCl
4
]
ions on the
ing pH, which also confirms that palladium hydroxide formation is
greatly accelerated at a higher pH. Therefore, the formation of large
and polydispersed palladium nanoparticles by the conventional DP
method might be attributed to the non-selective adsorption of
negatively charged activated carbon surface was prevalent. On
the contrary, the impregnated Pd content of the Pd/C#Cx series sig-
nificantly increased as the number of oxygen groups on the acti-
vated carbon increased, and the metal loading efficiency of Pd
reached 100% on the activated carbon treated with 10 wt%
2ꢁ
[
4
PdCl ]
ions on the negatively charged activated carbon surface
as well as to the rapid formation of various palladium hydroxide
species at pH as high as 12.
HNO
3 2 2
/10 wt% H O . The result suggests that the Pd content on
the activated carbon was determined during the initial selective
adsorption of [Pd(NH
vated carbon support.
2
+
With the SAD method, the use of a cationic palladium precursor
3
)
4
]
ions on the oxygen groups of the acti-
2
+
(
3 4
[Pd(NH ) ] ) is clearly advantageous in terms of selective and
homogenous adsorption because the negatively charged activated
carbon surface is maintained throughout the catalyst preparation
steps. Moreover, the slow generation of hydroxides by the
hydrolysis of urea not only effectively suppresses the unfavorable
Moreover, the carboxyl groups generated by the acid treatment
were especially advantageous to increase the selective adsorption.
The difference in the Pd content of the Pd/C#Cx series with the use
of urea may imply that the deposition of non-selectively adsorbed

1
30
S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
Fig. 3. TEM images of Pd/C#Cx catalysts; the insets indicate the histograms of Pd nanoparticle size.
the conventional DP method in terms of the formation of extre-
mely small and monodispersed palladium nanoparticles on the
activated carbon support.
The PXRD patterns of the Pd/C catalysts are presented in Fig. S4.
The broader bandwidth of the Pd [1 1 1] phase observed for the
Pd/C#Ax series indicates the formation of larger Pd nanoparticles,
which is consistent with the TEM analysis result. The [1 1 1]/[2 0 0]
ratio of the Pd crystallites was measured to examine the effect of
the preparation method on the crystallinity of the Pd/C catalysts.
However, there was no distinguishable difference with respect to
either the deposition method or the oxygen content of the acti-
vated carbon support. The Pd [1 1 1]/[2 2 0] phase ratio was not
measured because the characteristic diffraction peak ca. 2h = 68°
overlapped with that of the activated carbon.
XPS analysis was also carried out to investigate the electronic
states of Pd. The Pd(3d) region and deconvoluted chemical states
of the selected Pd/C catalysts are shown in Fig. S5. The four decon-
voluted peaks of Pd(3d5/2) and Pd(3d3/2) indicate that complete
reduction was hardly achieved, regardless of the preparation
method. However, the degree of reduction was significantly
affected by the preparation method. For Pd/C#A1, the peak areas
Fig. 4. Pd nanoparticle size distribution based on the metal precursor and surface
treatment of the activated carbon support.
0
of Pd for both the Pd(3d5/2) and Pd(3d3/2) regions were larger than
2
+
palladium ions is also partially involved. However, the increasing
Pd content with increasing number of carboxyl groups on the acti-
vated carbon surface clearly demonstrates that the surface deposi-
those of Pd . On the other hand, for Pd/C#C1 and Pd/C#C4, the lar-
2
+
ger Pd peaks of the Pd(3d5/2) and Pd(3d3/2) spin-orbit doublet
around 337.4 and 342.8 eV (compared to the reduced Pd peaks of
335.4 eV for 3d5/2 and 340.8 eV for 3d3/2) imply that a large portion
of palladium exists in an oxidized form in the Pd/C catalysts pre-
2+
tion of the selectively adsorbed [Pd(NH
3
)
4
]
ions is predominant.
Therefore, the SAD method, which combines the selective adsorp-
tion of a cationic palladium precursor on a negatively charged acti-
vated carbon surface with the homogeneous formation of
palladium hydroxide on the adsorbed sites, may be superior to
2
+
0
pared by the SAD method. The similar Pd /Pd ratios between
Pd/C#C1 and Pd/C#C4 also suggests that the oxygen content of
the activated carbon surface has little impact on the reduction of

S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
131
Fig. 5. Postulated mechanism for the formation of Pd nanoparticles on an activated carbon support; (a) conventional deposition-precipitation method and (b) homogeneous
adsorption deposition method.
Fig. 6. Formation of palladium hydroxide under various pH conditions. Optical image at (a) pH 5.2, 0 min; (b) pH 5.2, 60 min; (c) pH 6.5, 0 min; (d) pH 6.5, 60 min; (e) pH 7.5,
0
min; and (f) pH 7.5, 60 min; UV–vis spectra at (g) pH 5.2, (h) pH 6.5, and (i) pH 7.5 (i). The experiments were carried out at 293 K under atmospheric pressure.

1
32
S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
Fig. 7. Changes in Pd metal loading efficiency depending on the metal precursor
and surface treatment of the activated carbon support (metal loading efficiency (%)
=
actual Pd/intake Pd ꢀ 100).
Pd. Although Pd(OH)
x
may be present, the PdO peaks (ꢂ336.4 for
3
d5/2 and ꢂ341.8 eV for 3d3/2) were not deconvoluted because they
totally overlapped with the other peaks [55].
The difference in the electronic states of Pd with the prepara-
tion method may be attributed to the different metal deposition
mechanisms. For the DP method, it is likely that relatively larger
palladium hydroxide species may be non-selectively and rapidly
deposited on the outer surface where reduction may be easier. In
the SAD method, however, the reduction of the selectively depos-
ited palladium hydroxide on the oxygen groups of the activated
carbon surface seems more difficult.
Fig. 8. Reaction performance of Pd/C catalysts in terms of (a) H
2
O
2
productivity and
conversion rate and selectivity. Reaction conditions: catalyst (20 mg), N /H
2
(mol%) = 3.8/5.0/91.2 (30 bar), total gas flow rate = 40 mL/min, 80 wt% metha-
(
b) H
2
2
2
/
O
3.3. Direct synthesis of hydrogen peroxide
nol/water solvent (25 g, 20 ppm NaBr), 275 K, reaction time = 1 h.
The direct synthesis of hydrogen peroxide was carried out in the
presence of the various Pd/C catalysts prepared by the DP and SAD
methods, and the activity results are shown in Fig. 8. Both the
metal deposition method and the nature of the activated carbon
support played a critical role in the reaction performance of the
Pd/C catalyst. The SAD method was exceptionally superior to the
Higher Pd loadings on an activated carbon support bearing a
large number of oxygenate groups may be responsible for the
reduced catalytic activity [39,56]. However, the Pd content of
Pd/C#C3 and C4 were higher than those of Pd/C#A3 and A4, but
the catalytic activity of the former was superior to that of the lat-
2
conventional DP method to obtain higher H conversion rate and
2
ter; particularly, the H selectivity of Pd/C#C4 was 9 times higher
selectivity, and thus the productivity of Pd/C#C1 was 12 times
higher than that of Pd/C#A1. According to a density functional the-
ory (DFT) study, Pd clusters smaller than 2.5 nm may selectively
than that of Pd/C#A4. This result clearly demonstrates the promis-
ing aspect of the SAD method in realizing an efficient catalyst for
the direct synthesis of hydrogen peroxide.
Fig. S6 shows the time-dependent changes in the catalytic activ-
ity of Pd/C#C1 for the direct synthesis of hydrogen peroxide in the
presence or absence of a halide additive (NaBr). In the presence of
the halide ion, the initial catalytic activity was very promising:
catalyze H
may preferentially promote O
formation of H O [56]. The effect of the size of the active metal par-
2
O
2
formation, but larger Pd nanoparticles (>2.5 nm)
2
dissociation, which results in the
2
ticles on the catalytic performance was also reported for supported
Pd and AuPd catalysts [12,39]. Therefore, the excellent catalytic
activity of Pd/C#C1 may be explained in terms of the formation
of very small monodispersed Pd nanoparticles (ꢂ1 nm).
2 2 2
after 30 min, the H O productivity and H selectivity reached
8
606 mmol H /g Pd.h and 95.1%, respectively. As presented in
2 2
O
Table 2, the exceptional reaction performance of the Pd/C#C1 cat-
alyst is comparable to that of previously reported Pd catalysts sup-
ported on various carbon materials, such as activated carbon
The nature of the activated carbon support also significantly
affected the catalytic performance regardless of the catalyst prepa-
ration method. As shown in Fig. 8(b), both the H
and the H selectivity were greatly reduced as the oxygen group
content of the activated carbon surface increased, which resulted
in a significant reduction in H productivity (Fig. 8(a)). As there
2
conversion rate
[
11,24,57], carbon nanotubes [58,59], carbon black [39], ordered
mesoporous carbon [60], and carbon nanofiber [61].
As expected, although the H conversion increased, the H
productivity and H selectivity significantly decreased as the reac-
tion time was increased, regardless of the use of a halide ion. The
significant reduction in both H productivity and H selectivity
is obviously attributed to the subsequent H decomposition
and hydrogenation as well as the non-selective oxidation of H to
O, as described in Fig. S7. The gradual increase in H conversion
2
2
2 2
O
2 2
O
were no apparent differences in the Pd nanoparticle size and distri-
bution upon acid treatment, it is difficult to rationalize the differ-
ent catalytic activities in terms of Pd nanoparticle size. As will be
discussed later, the carboxyl groups incorporated by the acid treat-
ment may have a significant impact on the reaction performance of
a catalyst in addition to the Pd nanoparticle size.
2
2
O
2
2
2 2
O
2
H
2
2

S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
133
Table 2
Comparison of the catalytic activity of Pd/C#C1 with recently reported Pd catalysts supported on various carbon materials.
Catalyst
Pd (wt%)
Productivity (mmol H
2
O
2
/g Pd.h)
H
2
Sel. (%)
T (°C)/P (bar)/Time (h)
Additives
Ref.
Pd/C#C1a
1.2
5
0.9
2.5
5
1
0.3
1
8606
1040
1422
129
5600
210
140
6397
8400
95.1
42
45
74
–
2/30/0.5
2/40/0.5
r.t./10/1.5
10/1/4
5/20/1.5
2/40/0.5
r.t./30/0.33
2/30/0.25
0/40/0.5
20 ppm NaBr
–
This work
[11]
[58]
[39]
[60]
[61]
[59]
[57]
[24]
Pd/ACa
b
Pd/N-CNT
Pd/CBc
250 mL H
2
SO
4
4
0.03 M HCl
0.04 M HCl
–
d
Pd/CMK-3
Pd/N-CNFe
–
b
Pd/N-CNT
68
54
80
125 mL H
2
SO
Pd/ACa
0.03 M H
–
2 4
SO
f
a
Pd/HHDMA /AC
0.6
a
b
c
d
e
f
Activated carbon.
Carbon nanotubes.
Carbon black.
Ordered mesoporous carbon.
Carbon nanofibers.
Hexadecyl(2-hydroxyethyl)dimethylammonium dihydrogenphosphate.
during the reaction implies that H
genation as well as in the direct synthesis of H
The remarkably improved reaction performance with the addi-
tion of a halide ion suggests that the unfavorable non-selective oxi-
2 2
dation of H to H O was effectively suppressed by the halide ion.
For example, the halide ion may poison the non-selective surface
of Pd and thus inhibit the dissociative chemisorption of dioxygen
on the palladium surface at the expense of the decrease in H
version [10]. However, the similar reduction trends in H
ity irrespective of the presence of Br ions implies that the gradual
2
is also used in the H
2
O
2
hydro-
22%, respectively. The low activity of the Anionic/SAD Pd/C catalyst
may be understood in terms of the formation of large Pd nanopar-
2 2
O .
2
+
0
ticles with low Pd /Pd ratio as similar as those of Pd/C#A1. The
result clearly confirms that selective adsorption of the cationic pal-
ladium ions on the negatively charged activated carbon surface is
of prime importance for the formation of very small and monodis-
2
+
0
persed Pd nanoparticles with a high Pd /Pd ratio, and thus for
attaining high catalytic activity for the direct synthesis of hydrogen
peroxide.
2
con-
2
selectiv-
ꢁ
2 2 2
decrease in H selectivity over time is mainly ascribed to H O
decomposition and hydrogenation rather than to the direct oxida-
tion of hydrogen.
The use of halide ion was also advantageous to improve the
reaction performance of Pd/C#A1 catalyst. As shown in Fig. S8,
2 2 2
the H O productivity and H selectivity were greatly increased
as the amount of halide ions increased in the reaction medium.
The improved catalytic activity may be resulted from the poisoning
of the non-selective surface of Pd by the halide ions as supported
2
by the sharp reduction in the H conversion. The addition of halide
ions showed more pronounced effect on the catalytic activity of
3
.4. Hydrogen peroxide decomposition and hydrogenation
One of the challenging issues in the direct synthesis of hydrogen
peroxide is that all the reaction pathways are thermodynamically
favorable, and the active metal, such as Pd, may catalyze H for-
2 2
O
mation as well as its subsequent decomposition and hydrogena-
tion. Therefore, the reaction performance of a catalyst for the
direct synthesis of hydrogen peroxide may be determined by the
combination of its activities for these reactions. In this context,
the activity of the Pd/C catalysts in the decomposition and hydro-
genation of H O was examined. As shown in Fig. 9, both the H O
2 2 2 2
decomposition and hydrogenation rates were greatly affected by
the metal deposition method and the nature of the activated
Pd/C#A1 compared to that of Pd/C#C1. In the presence of 20
ppm NaBr, the H O productivity and H selectivity of Pd/C#A1
2 2 2
were 8.7 and 17.4 times higher than those obtained without NaBr.
However, the volcano-shaped dependence of the catalytic activity
on the halide ion content implies that there may be an optimum
level for the halide ion addition and that it is difficult to overcome
the low activity of Pd/C#A1 catalyst solely by the addition of NaBr.
To investigate the important role of the selective adsorption of a
palladium ion on the activated carbon support in determining the
catalytic activity, another Pd/C catalyst was prepared via the SAD
2 4
method using the anionic palladium precursor, H PdCl (Anionic/
SAD Pd/C). As shown in Fig. S9, not only the palladium precursor
types but also the catalyst preparation methods exerted crucial
effect on the characteristics of Pd nanoparticles. The relatively
smaller Pd nanoparticles (3.3 ± 0.6 nm) of the Anionic/SAD Pd/C
than those of the Pd/C#A1 (4.8 ± 1.5 nm) implies that the gradual
generation of hydroxide ions by urea decomposition is advanta-
geous in terms of the control of nanoparticle size. However, the
similar electronic nature of the Anionic/SAD Pd/C with that of the
2
+
0
Pd/C#A1 (low Pd /Pd ratio) suggests that the palladium hydrox-
ide species formed by the ligand exchange of the non-selectively
2
ꢁ
and weakly adsorbed [PdCl
less of the preparation methods.
The activity test of the Anionic/SAD Pd/C catalyst is presented in
Fig. S10. The H productivity and H selectivity of the Anionic/
SAD Pd/C catalyst were found to be 589 mmol H /g Pd.h and
4
]
ions may be easily reduced regard-
Fig. 9. H
prepared with different metal precursors and/or surface treatment methods of the
activated carbon support. Reaction conditions: catalyst (20 mg), 1 wt% H in 80
wt% methanol/water solvent (25 g, 20 ppm NaBr), 30 bar, 275 K, reaction time = 1 h.
Gas flow rate for H decomposition = 40 mL/min (N ). Gas flow rate for H
hydrogenation = 30 mL/min (5 mol% H /N ) and 10 mL/min (N ).
2 2
O decomposition and hydrogenation in the presence of Pd/C catalysts
2 2
O
2
O
2
2
2
O
2
2
2 2
O
2 2
O
2
2
2

1
34
S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
carbon support, but showed opposite trends based on the nature of
the activated carbon support regardless of the metal deposition
method.
H
2
selectivity by enhancing the hydrophilic nature of the support
and/or by promoting H hydrolysis. In this regard, the carboxyl
groups may retard H formation and hydrogenation by poison-
2 2
O
O
2 2
The Pd/C#Ax catalyst showed
H
2
O
2
decomposition rates
ing the active sites rather than by restricting the access of the reac-
tants to the active sites.
6
.6–12.1 times higher than those of Pd/C#Cx, which may be due
0
2+
to the higher Pd /Pd ratio of the Pd/C#Ax catalysts, as observed
in the XPS analysis. In addition, various AuPd catalysts with higher
0
2+
3.5. Effect of the surface nature of the activated carbon support on the
reaction performance
Pd /Pd ratios showed lower H
to accelerate the unfavorable H
Pd nanoparticle size also affects the H
and larger Pd nanoparticles (>2.5 nm) have a lower H
because of faster H decomposition [56]. Therefore, the faster
decomposition rate with the Pd/C#Ax series may be ascribed
to the presence of relatively larger Pd nanoparticles with higher
2
selectivity owing to their ability
decomposition [27,62,63]. The
decomposition rate,
selectivity
2 2
O
2 2
O
To clarify the influence of the carboxyl groups of the activated
carbon support on the catalytic activity, the direct synthesis of
hydrogen peroxide over Pd/C#C1 was carried out in the presence
of sulfuric or acetic acid. To maintain a constant concentration of
protons, the pH was kept at 3.5 regardless of the acid additive used.
As shown in Fig. 10, when sulfuric acid was introduced, the
2
2 2
O
2 2
H O
0
2+
Pd /Pd ratios.
Another interesting aspect of the experiment is the increase in
increased number of protons effectively suppressed H
position, and both the H productivity and H
increased while H conversion remained constant. On the other
hand, with the addition of acetic acid, H conversion and selectivity
decreased, which resulted in a significant reduction in H pro-
ductivity. The H conversion gradually increased with reaction
time because of the consumption of hydrogen in the subsequent
hydrogenation; however, the observed reduction in H con-
version upon addition of acetic acid indicates that the catalyst par-
tially lost its activity. Moreover, the simultaneous decrease in H
conversion and selectivity implies that acetic acid may also pro-
mote H hydrolysis. Although the decreased H conversion is
attributed to the poisoning of the Pd active sites by acetic acid,
the reduced H selectivity is mainly ascribed to the enhanced
decomposition rather than to poisoning.
To further confirm the poisoning effect of the carboxyl groups of
2
O
2
decom-
the H
vated carbon surface, especially carboxyl groups, regardless of
the catalyst preparation method. The natural breakdown of H
increases as the hydrophilicity of the surface increases, and thus,
a hydrophobic surface is beneficial to prevent H decomposition
10,64–66]. Therefore, the enhanced hydrophilicity of an activated
carbon surface from the incorporation of various oxygenates may
be responsible for the faster H decomposition rates. Although
2 2
O decomposition rate with the oxygen content of the acti-
2
O
2
2
selectivity
2 2
O
2
2
2 2
O
2 2
O
[
2
2
H O
2
2
2 2
O
there are no apparent difference in the total amount of oxygen
groups on the activated carbon surface treated with different acid
concentrations, as shown in the titration experiment (see Fig. 1),
2
2
O
2
2
2 2
the faster H O decomposition rates over Pd/C#C3 and #C4 com-
pared to Pd/C#C2 suggest that the carboxyl groups have a larger
influence on the formation of a hydrophilic surface than the hydro-
xyl groups.
2
2 2
H O
The weak electrolytic nature of carboxyl groups may be respon-
the activated carbon surface on the catalytic activity, hydrogena-
tion of 1-decene was carried out using the Pd/C#Ax and Pd/C#Cx
catalysts. As shown in Fig. 11, the reaction performance strongly
depends on both the metal deposition method and the surface nat-
ure of the support, and the reduction in the catalytic activity of the
Pd/C#Cx series was more significant than that of Pd/C#Ax.
The changes in the reaction performance of these catalysts for
olefin hydrogenation follow the same trend as that for the direct
synthesis and hydrogenation of hydrogen peroxide. The result
clearly demonstrates that the carboxyl groups near Pd may notice-
ably exert a detrimental effect on the catalytic activity by hinder-
ing or poisoning the active sites. The negative influence of the
carboxyl groups on the active sites in combination with the
sible for the increased H
activated carbon surface. In the presence of a strong electrolyte
such as H SO , H decomposition can be effectively suppressed
because the H-OOH dissociation may be retarded by the excess
protons [1,10,31,32,67]. With carboxylic acid, the hydrogen bonds
2 2
O decomposition rate on a hydrophilic
2
4
2 2
O
between the carboxyl groups and H
hydrolysis to H O, which may offset the benign effect of the pro-
tons on the suppression of H-OOH dissociation.
Contrary to H decomposition, the H hydrogenation rate
decreased as the number of oxygenates on the activated carbon
increased. In particular, the reduction in the H hydrogenation
rate was much larger for the Pd/C#Cx series than the Pd/C#Ax ser-
ies. The steep decrease in the H hydrogenation rate is very sim-
ilar to the rapid decrease in the reaction rate of the direct synthesis
of H in the presence of the same catalyst, which suggests that
2 2 2 2
O may accelerate H O
2
2
O
2
2 2
O
2 2
O
2 2
O
2 2
unfavorable acceleration of H O hydrolysis by the formation of
2 2
O
the carboxyl groups near the Pd nanoparticles may have a detri-
mental effect on both reactions. It is assumed that the carboxyl
groups near the active Pd metal sites may act as poison for the
active sites and/or restrict the reactants’ access to the active sites.
In the Pd/C#Cx series, the undesirable influence of the carboxyl
groups on the catalytic activity is more prominent because the
ultra-small Pd nanoparticles are located in close proximity to the
carboxyl groups on the activated carbon surface, which resulted
in a sharp decrease in reaction rates compared to the Pd/C#Ax ser-
ies in which larger and polydispersed Pd nanoparticles mainly exist
on the outer surface [43,52]. A similar negative effect of carboxyl
groups on reaction performance has been reported [39]. Vannice
and coworkers also reported that carbon might contaminate bulk
and surface Pd and therefore suppress chemisorption on the Pd
2 2
surface [38]. Considering the increase in the H O decomposition
rate as the number of carboxyl groups increases, the carboxyl
groups has little impact on the intrinsic catalytic activity of Pd
2 2
for H O decomposition. However, increasing the number of
Fig. 10. Changes in the catalytic performance of Pd/C#C1 in the presence of
carboxyl groups on an activated carbon surface may decrease
different acids.

S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
135
of ultra-small and monodispersed Pd nanoparticles with a high
2
+
0
Pd /Pd ratio resulted in one of the highest reaction performances
among previously reported Pd catalysts supported on various car-
bon materials. The high initial H
of 8606 mmol H /g Pd.h and 95.1%, respectively, were achieved
in the presence of the Pd/C#C1 catalyst after 30 min under intrin-
sically safe and non-corrosive conditions. The excellent H pro-
ductivity and H selectivity of the Pd/C#C1 catalyst prepared by
2 2 2
O productivity and H selectivity
2 2
O
2 2
O
2
the selective adsorption deposition method were 12 and 7.1 times
higher, respectively, than those of the Pd/C#A1 catalyst prepared
by a conventional deposition-precipitation method.
In addition, the oxygen groups, particularly the carboxyl groups,
of the activated carbon surface, which were incorporated by acid
treatment, have a huge impact on the reaction performance of
the catalyst, regardless of the preparation method. Increasing the
number of carboxyl groups on the activated carbon surface may
lead to the acceleration of H
the acid addition experiment, H
tively suppressed in the presence of a strong electrolyte such as
SO . However, carboxylic acid, which is a weak electrolyte, does
not effectively suppress H-OOH dissociation but rather promotes
hydrolysis by forming hydrogen bonds with H . Further-
more, the sharp decrease in the direct synthesis and hydrogenation
rates of H , as observed in olefin hydrogenation experiments,
2
O
2
decomposition. As confirmed by
Fig. 11. Effect of surface treatment of Pd/C catalysts on hydrogenation of 1-decene.
2 2
O
decomposition can be effec-
H
2
4
H
2
O
2
2 2
O
2 2
O
suggests that the carboxyl groups near the active sites exert a neg-
ative influence on the catalytic activity by hindering or poisoning
the active sites. These detrimental effects of the carboxyl groups
may be responsible for the significant reduction in the reaction
performance of catalysts as the carboxyl group content of the acti-
vated carbon support increases.
Acknowledgement
Fig. 12. Illustration of the unfavorable effect of carboxyl groups in accelerating
hydrolysis as well as hindering or poisoning the active metal sites.
This research was supported by Basic Science Research Program
through the National Research Foundation of Korea (NRF) funded
by the Ministry of Education (2016R1D1A3B02006928). This work
was also financially supported by the Kunsan National University’s
Long-term Overseas Research Program for Faculty Member in the
year 2017. Y.M.C. also thanks Seokjin Kwon and Seungyeol Oh for
their experimental assistance.
2 2
H O
2 2
hydrogen bonds with H O resulted in a significant reduction in
the catalytic performance as the amount of carboxyl groups on
the activated carbon support increased, as described in Fig. 12.
3.6. Catalyst recycling
References
The catalyst recyclability was also tested by repeated reaction
[
1] J.M. Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Hydrogen peroxide
synthesis: an outlook beyond the anthraquinone process, Angew. Chemie -
Int. Ed. 45 (2006) 6962–6984, https://doi.org/10.1002/anie.200503779.
2] G.F. Thiele, E. Roland, Propylene epoxidation with hydrogen peroxide and
titanium silicalite catalyst: activity, deactivation and regeneration of the
catalyst, J. Mol. Catal. A Chem. 117 (1997) 351–356, https://doi.org/10.1016/
S1381-1169(96)00266-X.
3] D.T. Bregante, N.E. Thornburg, J.M. Notestein, D.W. Flaherty, Consequences of
confinement for alkene epoxidation with hydrogen peroxide on highly
dispersed group 4 and 5 metal oxide Catalysts, ACS Catal. 8 (2018) 2995–
cycles using Pd/C#C1 under the same reaction conditions. The cat-
alyst was carefully filtered at the end of each reaction cycle. The
recovered catalyst was washed thoroughly with water and metha-
nol, dried overnight at 383 K, and then reused. As shown in
Fig. S11, the activity of the recycled catalyst diminished slightly,
but after the third recycle run, the catalytic activity was main-
tained. It seems that that the reduced activity of the recycled cat-
alysts may be caused by the partial leaching of Pd during the
first reaction or recycle step. Sintering of Pd nanoparticles or accu-
mulation of halide ions on the palladium surface as a reason for the
[
[
[
3010, https://doi.org/10.1021/acscatal.7b03986.
4] C. Hammond, M.M. Forde, M.H. Ab Rahim, A. Thetford, Q. He, R.L. Jenkins, N.
Dimitratos, J.A. Lopez-Sanchez, N.F. Dummer, D.M. Murphy, A.F. Carley, S.H.
Taylor, D.J. Willock, E.E. Stangland, J. Kang, H. Hagen, C.J. Kiely, G.J. Hutchings,
Direct catalytic conversion of methane to methanol in an aqueous medium by
using copper-promoted Fe-ZSM-5, Angew. Chemie - Int. Ed. 51 (2012) 5129–
deactivation can be excluded because the H
2 2
O productivity as
well as H selectivity was maintained during the recycle runs.
2
5133, https://doi.org/10.1002/anie.201108706.
[
[
[
5] M.H. Ab Rahim, M.M. Forde, R.L. Jenkins, C. Hammond, Q. He, N. Dimitratos, J.A.
Lopez-Sanchez, A.F. Carley, S.H. Taylor, D.J. Willock, D.M. Murphy, C.J. Kiely, G.J.
Hutchings, Oxidation of methane to methanol with hydrogen peroxide using
supported gold-palladium alloy nanoparticles, Angew. Chemie - Int. Ed. 52
(2013) 1280–1284, https://doi.org/10.1002/anie.201207717.
6] C. Williams, J.H. Carter, N.F. Dummer, Y.K. Chow, D.J. Morgan, S. Yacob, P.
Serna, D.J. Willock, R.J. Meyer, S.H. Taylor, G.J. Hutchings, Selective oxidation of
methane to methanol using supported AuPd catalysts prepared by stabilizer-
4
. Conclusion
An efficient Pd/C catalyst design for the direct synthesis of
hydrogen peroxide was realized by the selective adsorption
deposition method, in which a cationic palladium precursor such
as [Pd(NH
activated carbon surface and subsequently converted to palladium
hydroxide in a homogeneous manner with hydroxide ions that
were slowly generated upon urea decomposition. The formation
2
+
3
)
4
]
was selectively adsorbed on a negatively charged
free sol-immobilization, ACS Catal.
0.1021/acscatal.7b04417.
8 (2018) 2567–2576, https://doi.org/
1
7] J.K. Edwards, G.J. Hutchings, Palladium and gold-palladium catalysts for the
direct synthesis of hydrogen peroxide, Angew. Chemie - Int. Ed. 47 (2008)
9192–9198, https://doi.org/10.1002/anie.200802818.

1
36
S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
[
[
8] PERP Report - Propylene Oxide, Nexant, New York, 2003.
[33] Y.-M. Chung, Y.-R. Lee, W.-S. Ahn, Direct synthesis of hydrogen peroxide from
hydrogen and oxygen over Pd-supported metal-organic framework catalysts,
Bull. Korean Chem. Soc. 36 (2015) 1378–1383, https://doi.org/10.1002/
bkcs.10267.
9] J.K. Edwards, S.J. Freakley, R.J. Lewis, J.C. Pritchard, G.J. Hutchings, Advances in
the direct synthesis of hydrogen peroxide from hydrogen and oxygen, Catal.
Today 248 (2015) 3–9, https://doi.org/10.1016/j.cattod.2014.03.011.
[
[
10] C. Samanta, Direct synthesis of hydrogen peroxide from hydrogen and oxygen:
an overview of recent developments in the process, Appl. Catal. A Gen. 350
[34] N. Lee, Y.-M. Chung, Direct synthesis of hydrogen peroxide from hydrogen and
3 8
oxygen over Pd-supported HNb O metal oxide nanosheet catalyst, Res. Chem.
Intermed. 42 (2016) 95–108, https://doi.org/10.1007/s11164-015-2311-z.
[35] A. Gervasini, P. Carniti, F. Desmedt, P. Miquel, Liquid phase direct synthesis of
(
2008) 133–149, https://doi.org/10.1016/j.apcata.2008.07.043.
11] J.K. Edwards, B. Solsona, E. Ntainjua, A.F. Carley, A.A. Herzing, C.J. Kiely, G.J.
Hutchings, Switching off hydrogen peroxide hydrogenation in the direct
synthesis process, Science 323 (2009) 1037–1041, https://doi.org/
2 2
H O : activity and selectivity of Pd-dispersed phase on acidic niobia-silica
supports, ACS Catal.
acscatal.7b01083.
[36] D.W. Flaherty, Direct synthesis of H
7
(2017) 4741–4752, https://doi.org/10.1021/
from H and O on Pd catalysts: current
1
0.1126/science.1168980.
[
12] J.K. Edwards, S.J. Freakley, A.F. Carley, C.J. Kiely, G.J. Hutchings, Strategies for
designing supported gold-palladium bimetallic catalysts for the direct
synthesis of hydrogen peroxide, Acc. Chem. Res. 47 (2013) 845–854, https://
doi.org/10.1021/ar400177c.
2
O
2
2
2
understanding, outstanding questions, and research needs, ACS Catal. 8 (2018)
1520–1527, https://doi.org/10.1021/acscatal.7b04107.
[37] M. Selinsek, B.J. Deschner, D.E. Doronkin, T.L. Sheppard, J.-D. Grunwaldt, R.
Dittmeyer, Revealing the structure and mechanism of palladium during direct
synthesis of hydrogen peroxide in continuous flow using operando
[
13] 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, Palladium-tin
catalysts for the direct synthesis of H
2
O
2
with high selectivity, Science 351
spectroscopy, ACS Catal.
acscatal.7b03514.
8 (2018) 2546–2557, https://doi.org/10.1021/
(
2016) 965–968, https://doi.org/10.1126/science.aad5705.
[
[
14] S. Wang, K. Gao, W. Li, J. Zhang, Effect of Zn addition on the direct synthesis of
hydrogen peroxide over supported palladium catalysts, Appl. Catal. A Gen. 531
[38] N. Krishnankutty, J. Li, M.A. Vannice, The effect of Pd precursor and
pretreatment on the adsorption and absorption behavior of supported Pd
catalysts, Appl. Catal. A Gen. 173 (1998) 137–144, https://doi.org/10.1016/
S0926-860X(98)00173-2.
[39] B. Hu, W. Deng, R. Li, Q. Zhang, Y. Wang, F. Delplanque-Janssens, D. Paul, F.
Desmedt, P. Miquel, Carbon-supported palladium catalysts for the direct
synthesis of hydrogen peroxide from hydrogen and oxygen, J. Catal. 319 (2014)
15–26, https://doi.org/10.1016/j.jcat.2014.08.001.
[40] P. Albers, SIMS/XPS Investigations on activated carbon catalyst supports, J.
Catal. 150 (1994) 368–375, https://doi.org/10.1006/jcat.1994.1355.
[41] E. Auer, A. Freund, J. Pietsch, T. Tacke, Carbons as supports for industrial
precious metal catalysts, Appl. Catal. A Gen. 173 (1998) 259–271, https://doi.
org/10.1016/S0926-860X(98)00184-7.
[42] P. Albers, R. Burmeister, K. Seibold, G. Prescher, S.F. Parker, D.K. Ross,
Investigations of palladium catalysts on different carbon supports, J. Catal.
181 (1999) 145–154, https://doi.org/10.1006/jcat.1998.2279.
(
2017) 89–95, https://doi.org/10.1016/j.apcata.2016.10.023.
15] N.M. Wilson, Y.-T. Pan, Y.-T. Shao, J.-M. Zuo, H. Yang, D.W. Flaherty, Direct
Synthesis of H on AgPt Octahedra: The importance of Ag–Pt coordination
for high H selectivity, ACS Catal. 8 (2018) 2880–2889, https://doi.org/
0.1021/acscatal.7b04186.
16] N.M. Wilson, P. Priyadarshini, S. Kunz, D.W. Flaherty, Direct synthesis of H
on Pd and Au Pd clusters: understanding the effects of alloying Pd with Au, J.
Catal. 357 (2018) 163–175, https://doi.org/10.1016/j.jcat.2017.10.028.
17] S. Maity, M. Eswaramoorthy, Ni–Pd bimetallic catalysts for the direct synthesis
of H – unusual enhancement of Pd activity in the presence of Ni, J. Mater.
2
O
2
2 2
O
1
[
[
[
2 2
O
x
1
2 2
O
Chem. A. 4 (2016) 3233–3237, https://doi.org/10.1039/C6TA00486E.
18] P. Tian, X. Xu, C. Ao, D. Ding, W. Li, R. Si, W. Tu, J. Xu, Y.F. Han, Direct and
selective synthesis of hydrogen peroxide over palladium–tellurium catalysts
at ambient pressure, ChemSusChem. 10 (2017) 3342–3346, https://doi.org/
1
0.1002/cssc.201701238.
[43] M.L. Toebes, J.A. Van Dillen, K.P. De Jong, Synthesis of supported palladium
catalysts, J. Mol. Catal. A Chem. 173 (2001) 75–98, https://doi.org/10.1016/
S1381-1169(01)00146-7.
[
[
[
19] B. Zhou, L.-K. Lee, US Patent 6,168,775, 2001.
20] B. Zhou, M. Rueter, S. Parasher, US Patent 7,001,807, 2006.
21] S. Kim, D.W. Lee, K.Y. Lee, Direct synthesis of hydrogen peroxide from
hydrogen and oxygen over single-crystal cubic palladium on silica catalysts, J.
[44] L.A.M. Hermans, J.W. Geus, Interaction of nickel ions with silica supports
during deposition-precipitation, Stud. Surf. Sci. Catal.
https://doi.org/10.1016/S0167-2991(09)60208-1.
3 (1979) 113–130,
Mol. Catal.
A Chem. 383–384 (2014) 64–69, https://doi.org/10.1016/
j.molcata.2013.11.021.
[45] M.L. Toebes, M.K. Van Der Lee, L.M. Tang, M.H.H. In’T Veld, J.H. Bitter, A.J. Van
Dillen, K.P. De Jong, Preparation of carbon nanofiber supported platinum and
ruthenium catalysts: comparison of ion adsorption and homogeneous
deposition precipitation, J. Phys. Chem. B. 108 (2004) 11611–11619, https://
doi.org/10.1021/jp0313472.
[
[
[
[
[
[
[
[
[
[
[
22] S. Kim, D.W. Lee, K.Y. Lee, Shape-dependent catalytic activity of palladium
nanoparticles for the direct synthesis of hydrogen peroxide from hydrogen and
oxygen, J. Mol. Catal. A Chem. 391 (2014) 48–54, https://doi.org/10.1016/
j.molcata.2014.03.026.
23] H.E. Jeong, S. Kim, M.G. Seo, D.W. Lee, K.Y. Lee, Catalytic activity of Pd
[46] S.L. Goertzen, K.D. Thériault, A.M. Oickle, A.C. Tarasuk, H.A. Andreas,
octahedrons/SiO
2
for the direct synthesis of hydrogen peroxide from hydrogen
2
Standardization of the Boehm titration. Part I. CO expulsion and endpoint
determination, Carbon 48 (2010) 1252–1261, https://doi.org/10.1016/
j.carbon.2009.11.050.
and oxygen, J. Mol. Catal. A Chem. 420 (2016) 88–95, https://doi.org/10.1016/
j.molcata.2016.03.043.
24] G.M. Lari, B. Puértolas, M. Shahrokhi, N. López, J. Pérez-Ramírez, Hybrid
palladium nanoparticles for direct hydrogen peroxide synthesis: The key role
of the ligand, Angew. Chem. - Int. Ed. 56 (2017) 1775–1779, https://doi.org/
3
[47] J.S. Noh, J.A. Schwarz, Effect of HNO treatment on the surface acidity of
activated carbons, Carbon 28 (1990) 675–682, https://doi.org/10.1016/0008-
6223(90)90069-B.
[48] E. Papirer, J. Dentzer, S. Li, J.B. Donnet, Surface groups on nitric acid oxidized
carbon black samples determined by chemical and thermodesorption
analyses, Carbon 29 (1990) 69–72, https://doi.org/10.1016/0008-6223(91)
90096-2.
[49] Y. Otake, R.G. Jenkins, Characterization of oxygen-containing surface
complexes created on a microporous carbon by air and nitric acid treatment,
Carbon 31 (1993) 109–121, https://doi.org/10.1016/0008-6223(93)90163-5.
[50] P.E. Fanning, M.A. Vannice, A DRIFTS study of the formation of surface groups
on carbon by oxidation, Carbon 31 (1993) 721–730, https://doi.org/10.1016/
0008-6223(93)90009-Y.
[51] J. Figueiredo, M.F. Pereira, M.M. Freitas, J.J. Órfão, Modification of the surface
chemistry of activated carbons, Carbon 37 (1999) 1379–1389, https://doi.org/
10.1016/S0008-6223(98)00333-9.
1
0.1002/anie.201610552.
25] M.G. Seo, S. Kim, H.E. Jeong, D.W. Lee, K.Y. Lee, A yolk-shell structured
Pd@void@ZrO catalyst for direct synthesis of hydrogen peroxide from
hydrogen and oxygen, J. Mol. Catal. A Chem. 413 (2016) 1–6, https://doi.org/
0.1016/j.molcata.2015.12.004.
2
1
26] M. Seo, D.-W. Lee, S.S. Han, K.-Y. Lee, Direct synthesis of hydrogen peroxide
from hydrogen and oxygen over mesoporous silica-shell-coated, palladium-
2
nanocrystal-grafted SiO nanobeads, ACS Catal. 7 (2017) 3039–3048, https://
doi.org/10.1021/acscatal.7b00388.
27] S. Yook, H.C. Kwon, Y.G. Kim, W. Choi, M. Choi, Significant roles of carbon pore
and surface structure in AuPd/C catalyst for achieving high chemoselectivity in
direct hydrogen peroxide synthesis, ACS Sustain. Chem. Eng. 5 (2017) 1208–
1
216, https://doi.org/10.1021/acssuschemeng.6b02595.
28] G. Li, J. Edwards, A.F. Carley, G.J. Hutchings, Direct synthesis of hydrogen
peroxide from and and in situ oxidation using zeolite-supported
(2007) 247–250, https://doi.org/10.1016/
[52] V.A. Semikolenov, Modern approaches to the preparation of ‘‘palladium on
charcoal” catalysts, Russ. Chem. Rev. 61 (1992) 168, https://doi.org/10.1070/
RC1992v061n02ABEH000938.
H
2
O
2
catalysts, Catal. Commun.
j.catcom.2006.06.021.
8
[53] M. Dixit, G.N. Subbanna, P.V. Kamath, Homogeneous precipitation from
29] M. Sun, J. Zhang, Q. Zhang, Y. Wang, H. Wan, Polyoxometalate-supported Pd
nanoparticles as efficient catalysts for the direct synthesis of hydrogen
peroxide in the absence of acid or halide promoters, Chem. Commun. (2009)
solution by urea hydrolysis: a novel chemical route to the
a-hydroxides of
nickel and cobalt, J. Mater. Chem.
10.1039/JM9960601429.
6 (1996) 1429–1432, https://doi.org/
5
174, https://doi.org/10.1039/b910317a.
[54] M.A. Fraga, E. Jordão, M.J. Mendes, M.M.A. Freitas, J.L. Faria, J.L. Figueiredo,
Properties of carbon-supported platinum catalysts: Role of carbon surface
sites, J. Catal. 209 (2002) 355–364, https://doi.org/10.1006/jcat.2002.3637.
[55] G. Blanco-Brieva, E. Cano-Serrano, J.M. Campos-Martin, J.L.G.J.L.G. Fierro,
Direct synthesis of hydrogen peroxide solution with palladium-loaded
sulfonic acid polystyrene resins, Chem. Commun. (2004) 1184–1185, https://
doi.org/10.1039/b402530j.
30] S. Park, S.H. Lee, S.H. Song, D.R. Park, S.H. Baeck, T.J. Kim, Y.M. Chung, S.H. Oh, I.
K. Song, Direct synthesis of hydrogen peroxide from hydrogen and oxygen over
palladium-exchanged insoluble heteropolyacid catalysts, Catal. Commun. 10
(
2009) 391–394, https://doi.org/10.1016/j.catcom.2008.10.002.
31] Y.-M. Chung, Y.-T. Kwon, T.J. Kim, S.-H. Oh, C.-S. Lee, Direct synthesis of H2O2
catalyzed by Pd nanoparticles encapsulated in the multi-layered
polyelectrolyte nanoreactors on
a
charged sphere, Chem. Commun. 47
[56] P. Tian, L. Ouyang, X. Xu, C. Ao, X. Xu, R. Si, X. Shen, M. Lin, J. Xu, Y.F. Han, The
(
2011) 5705–5707, https://doi.org/10.1039/c1cc11180a.
origin of palladium particle size effects in the direct synthesis of H
2
O
2
: is
32] J. Kim, Y.M. Chung, S.M. Kang, C.H. Choi, B.Y. Kim, Y.T. Kwon, T.J. Kim, S.H. Oh,
C.S. Lee, Palladium nanocatalysts immobilized on functionalized resin for the
direct synthesis of hydrogen peroxide from hydrogen and oxygen, ACS Catal. 2
smaller better?,
jcat.2016.12.004.
J Catal. 349 (2017) 30–40, https://doi.org/10.1016/j.
[57] J. Gu, S. Wang, Z. He, Y. Han, J. Zhang, Direct synthesis of hydrogen peroxide
from hydrogen and oxygen over activated-carbon-supported Pd–Ag alloy
(
2012) 1042–1048, https://doi.org/10.1021/cs300090h.

S. Lee et al. / Journal of Catalysis 365 (2018) 125–137
137
catalysts, Catal. Sci. Technol.
C5CY00813A.
58] S. Abate, R. Arrigo, M.E. Schuster, S. Perathoner, G. Centi, A. Villa, D. Su, R.
6
(2016) 809–817, https://doi.org/10.1039/
carbon-supported Au-Pd catalysts for the direct synthesis of hydrogen
peroxide, J. Catal. 292 (2012) 227–238, https://doi.org/10.1016/j.
jcat.2012.05.018.
[
[
[
[
Schlögl, Pd nanoparticles supported on N-doped nanocarbon for the direct
synthesis of H O from H and O , Catal. Today 157 (2010) 280–285, https://
2 2 2 2
doi.org/10.1016/j.cattod.2010.01.027.
[63] D. Gudarzi, W. Ratchananusorn, I. Turunen, M. Heinonen, T. Salmi, Promotional
effects of Au in Pd-Au bimetallic catalysts supported on activated carbon cloth
(ACC) for direct synthesis of H
2 2 2 2
O from H and O , Catal. Today 248 (2015) 58–
59] R. Arrigo, M.E. Schuster, S. Abate, G. Giorgianni, G. Centi, S. Perathoner, S.
Wrabetz, V. Pfeifer, M. Antonietti, R. Schlögl, Pd supported on carbon nitride
boosts the direct hydrogen peroxide synthesis, ACS Catal. 6 (2016) 6959–6966,
https://doi.org/10.1021/acscatal.6b01889.
68, https://doi.org/10.1016/j.cattod.2014.04.011.
[64] L. Fu, K.T. Chuang, R. Fiedorow, Selective oxidation of hydrogen to hydrogen
peroxide, Stud. Surf. Sci. Catal. 72 (1992) 33–41, https://doi.org/10.1016/
S0167-2991(08)61656-0.
60] T. García, S. Agouram, A. Dejoz, J.F. Sánchez-Royo, L. Torrente-Murciano, B.
[65] Q. Liu, J.C. Bauer, R.E. Schaak, J.H. Lunsford, Supported palladium
Solsona, Enhanced
nanoparticles on ordered mesoporous carbons, Catal. Today 248 (2015) 48–
7, https://doi.org/10.1016/j.cattod.2014.03.039.
2
H O
2
production over Au-rich bimetallic Au-Pd
nanoparticles: an efficient catalyst for the direct formation of H
and O , Angew. Chem. - Int. Ed. 47 (2008) 6221–6224, https://doi.org/10.1002/
anie.200801517.
[66] B. Hu, Q. Zhang, Y. Wang, Pd/graphite as a superior catalyst for the direct
synthesis of hydrogen peroxide from H and O , Chem. Lett. 38 (2009) 256–
257, https://doi.org/10.1246/cl.2009.256.
2 2
[67] N.M. Wilson, D.W. Flaherty, Mechanism for the direct synthesis of H O on Pd
2 2 2
O from H
2
5
61] 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, Depressing the
2
2
hydrogenation and decomposition reaction in H
AuPd on oxygen functionalized carbon nanofibers, Catal. Sci. Technol. 6 (2016)
94–697, https://doi.org/10.1039/C5CY01880C.
2 2
O synthesis by supporting
6
clusters: heterolytic reaction pathways at the liquid-solid interface, J. Am.
Chem. Soc. 138 (2016) 574–586, https://doi.org/10.1021/jacs.5b10669.
[
62] J.K. Edwards, J. Pritchard, M. Piccinini, G. Shaw, Q. He, A.F. Carley, C.J. Kiely, G.J.
Hutchings, The effect of heat treatment on the performance and structure of