Direct synthesis of H2O2 over acid-treated Pd/C catalyst derived from a Pd-Co core-shell structure

Article abstract of DOI:10.1016/j.cattod.2019.09.038

Although a wide range of Pd/C catalysts has been developed for the direct synthesis of H₂O₂, low H₂O₂ yields remain a challenging issue. In this study, we propose a novel catalyst design involving the formation of Pd@Co(BO₃)₂ core-shell nanoparticles (NPs) on an activated carbon support followed by the cobalt borate shell etching with H₃PO₄. The non-noble metal shell efficiently suppressed unfavorable aggregation of Pd NPs, resulting in the formation of small and monodisperse Pd NPs after stripping the cobalt. The use of H₃PO₄ was also beneficial to obtain a high H₂O₂ yield by decreasing the metallic nature of Pd and enabling the acid-treated activated carbon to act as a solid acid support. The features of the acid-treated Pd/C catalyst derived from the Pd@Co(BO₃)₂ core-shell were superior enough for achieving extremely high activity in the direct synthesis of H₂O₂. The H₂O₂ productivity of 5721 mmol-H₂O₂/g-Pd.h with 88.1% H₂O₂ selectivity reported herein is one of the best values among Pd/C catalysts developed to date. It was also demonstrated that the cobalt borate shell should be completely removed because cobalt accelerated H₂O₂ decomposition and hydrogenation.

Full text of DOI:10.1016/j.cattod.2019.09.038

Catalysis Today xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Direct synthesis of H O over acid-treated Pd/C catalyst derived from a Pd-
2
2
Co core-shell structure
⁎
Seungsun Lee, Young-Min Chung
Department of Nano & Chemical Engineering, Kunsan National University, 558 Daehak-ro, Kunsan, 54150, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Keywords:
2 2 2 2
Although a wide range of Pd/C catalysts has been developed for the direct synthesis of H O , low H O yields
Direct synthesis of H
Core-shell structure
Non-noble metal sacrificial method
Pd/C
2
O
2
remain a challenging issue. In this study, we propose a novel catalyst design involving the formation of Pd@Co
BO core-shell nanoparticles (NPs) on an activated carbon support followed by the cobalt borate shell etching
with H PO . The non-noble metal shell efficiently suppressed unfavorable aggregation of Pd NPs, resulting in the
formation of small and monodisperse Pd NPs after stripping the cobalt. The use of H PO was also beneficial to
obtain a high H yield by decreasing the metallic nature of Pd and enabling the acid-treated activated carbon
to act as a solid acid support. The features of the acid-treated Pd/C catalyst derived from the Pd@Co(BO core-
shell were superior enough for achieving extremely high activity in the direct synthesis of H . The H
productivity of 5721 mmol-H /g-Pd.h with 88.1% H selectivity reported herein is one of the best values
among Pd/C catalysts developed to date. It was also demonstrated that the cobalt borate shell should be com-
pletely removed because cobalt accelerated H decomposition and hydrogenation.
(
3 2
)
3
4
3
4
Solid acid support
2 2
O
3 2
)
2
O
2
2 2
O
2
O
2
2 2
O
2 2
O
1
. Introduction
thermodynamic and safety challenges. As described in Scheme 1, all
side-reactions such as direct H oxidation, H decomposition, and
hydrogenation to H O, are energetically favorable. Furthermore,
all reactions can be promoted by the same active metal catalysts such as
Pd and Pt. Therefore, it is extremely difficult to obtain high H O yields
2 2
for this type of reaction.
2
2 2
O
Hydrogen peroxide (H
2
O
2
) is a popular and important chemical
H
2
O
2
2
with wide ranging applications as a commodity as well as in medical
clinics, cosmetics, wastewater treatment, mining, textiles, paper
bleaching, and chemical syntheses [1]. Recently, as industrial require-
ments for H
that the market for H
CAGR) of 6.0% until 2024 [2]. Since it was first synthesized through
the reaction of barium peroxide with hydrochloric acid by Thenard in
818, a variety of chemical and electrochemical routes for the pro-
2
O
2
as a clean oxidant continually increase, it is forecasted
Another significant problem is that the H
must be minimized to prevent the potential explosion of the H
mixture, even in the presence of a diluent gas. The strict restriction of
the H concentration (< 5 vol% H ) causes the insufficient concentra-
tion of H during the abovementioned reaction, resulting in low H
yields. It is well-known that promoters such as sulfuric acid and halide
ions can improve the H yield, and a number of studies have been
2
concentration in a reactor
2
O
2
will expand at a compound annual growth rate
2
/O
2
(
2
2
1
2
2 2
O
duction of H
duction of H
2
O
2
have been developed [1,3]. Current commercial pro-
mainly relies on the anthraquinone autoxidation pro-
2
O
2
2 2
O
cess (AO process) consisting of multiple energy-intensive steps using
excessive toxic working solutions and generating unwanted wastes.
Therefore, there is great demand for an alternative process that over-
comes the undesirable nature of the AO process from an energy and
environmental perspective.
reported regarding the elucidation of their roles in DSHP [8–14].
However, the use of these caustic promoters may result in other pro-
blem because of the necessity for complex separation and/or post-
treatment steps.
To solve these difficult challenges of DSHP, extensive research re-
garding the development of an elegant catalyst design has been un-
dertaken [3,6,7,15]. A key factor for realizing an efficient catalyst de-
sign is fine-tuning of the physicochemical properties of the active
2 2 2 2
Direct synthesis of H O from H and O (DSHP) has emerged as a
promising process by virtue of its high atom utilization efficiency and
green nature owing to its straightforward reaction and generation of
water as the only by-product [3–7]. Despite the potential advantages of
2 2
metals to preferentially promote H O formation, and palladium is
2 2
DSHP, it suffers from insufficient H O yields due to its intrinsic
considered to be the most suitable active metal for this purpose [16].
⁎
Corresponding author.
E-mail address: ymchung@kunsan.ac.kr (Y.-M. Chung).
https://doi.org/10.1016/j.cattod.2019.09.038
Received 18 June 2019; Received in revised form 29 August 2019; Accepted 23 September 2019
0920-5861/ © 2019 Elsevier B.V. All rights reserved.
Please cite this article as: Seungsun Lee and Young-Min Chung, Catalysis Today, https://doi.org/10.1016/j.cattod.2019.09.038

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
separated via centrifugation, washed thoroughly with deionized water,
and dried at 393 K for 12 h. To obtain a crystalline Co (BO ) phase, an
3 3 2
additional heat treatment was performed at 823 K for 5 h under ni-
trogen.
Preparation of the Pd@Co
manner similar to that of Co
precursor, K PdCl , with different Co/Pd ratios (denoted as Co
where x = 2–10). For Co Pd, Co(CH COOH) ∙4H O (0.576 mmol) was
added to 50 mL of deionized water containing dissolved K PdCl
0.096 mmol). Subsequently, a 0.03 M aqueous NaBH solution was
3
(BO
3
)
)
3 2
2
core-shell NPs was performed in a
in the presence of a palladium
Pd,
3
(BO
2
4
x
6
3
2
2
2
4
(
4
added under vigorous stirring. After extended stirring for 1 h, a black
precipitate was recovered, washed twice, and dried.
The preparation method for Co
Pd/C catalysts in the presence of an activated carbon support.
PdCl (0.384 mmol) and Co(CH COOH) ∙4H O (2.3 mmol) were se-
quentially added to 100 mL of deionized water in which 2 g of activated
x
Pd was also used to prepare the
Co
x
K
2
4
3
2
2
carbon was suspended. After reduction with NaBH
tained.
4 6
, Co Pd/C was ob-
Scheme 1. Direct synthesis of H
2
O
O.
2
as well as its subsequent and parallel side-
reactions for the formation of H
2
For cobalt borate etching, Co
PO solution and vigorously stirred for 2 h to completely leach cobalt
out from the Pd/C. The cobalt was removed, and the phosphoric acid-
treated Pd/C catalyst was denoted as (Co )Pd/C (where x = 2–10). The
)Pd/C catalyst was further calcined ((Co )Pd/C-Cal) or reduced
)Pd/C-Red) under air (523 K, 3 h) or 50 vol% H /N gas (473 K,
h), respectively.
For comparison, simple Pd/C catalyst was prepared by conventional
adsorption-reduction method [31]. Activated carbon suspended in
deionized water was mixed with aqueous H PdCl solution (Pd uptake
wt%). After stirring at room temperature for 2 h, reduction was car-
x
Pd/C was re-dispersed in a 30 wt%
H
3
4
Among the various palladium catalysts, palladium/carbon (Pd/C) cat-
alysts have been widely used because of the advantages of carbonac-
eous materials such as their low cost, availability, inertness, and sta-
bility towards caustic additives. However, because simple Pd/C
catalysts generally exhibited low H O yields, strategies such as bime-
2 2
tallization [17–21], surface modification by acid treatment [22], ligand
introduction [23], and the use of well-defined carbon supports [24–27]
x
(
(
Co
(Co
6
6
6
2
2
3
2 2
have been developed to improve H O formation efficiently. Recently,
2
4
we reported that a highly active Pd/C catalyst for the DSHP could be
prepared by selective adsorption deposition [28] or sequential ligand
exchange [29], wherein the palladium precursor could selectively ad-
sorb on a charged carbon surface. Consequently, small and mono-
disperse Pd nanoparticles (NPs) were successfully loaded on an acti-
vated carbon support.
In this work, a new Pd/C catalyst design for DSHP has been realized
using a non-noble metal sacrificial method. The acid-treated Pd/C
catalyst derived from a pre-formed Pd-Co core-shell structure on an
activated carbon support showed promising activity for DSHP under
intrinsically safe reaction conditions. Based on extensive characteriza-
tion of prepared Pd/C catalysts combined with activity tests, key factors
for catalysis were elucidated.
2
4
ried out by the addition of NaBH under vigorous stirring. The resulting
catalyst was filtered, washed and dried for further use. Actual Pd con-
tent (1.45 wt%) was determined by ICP analysis.
2.3. Characterization
Powder X-ray diffraction (PXRD) analysis of the prepared samples
was performed using a Malvern Panalytical Empyrean high resolution
X-ray diffractometer. X-ray photoelectron spectroscopy (XPS) analysis
was used to investigate the binding energies of Co and Pd in the samples
on a Thermo VG Scientific MultiLab 2000 apparatus with an Al anode
and 30 eV of pass energy (step increment = 0.1 eV). The size of the Pd
NPs and their distributions were determined by analyzing transmission
electron microscopy (TEM) images obtained from Philips Tecnai G2 F30
transmission electron microscope. High-angle annular dark-field scan-
ning transmission electron microscopy (HAADF-STEM) images obtained
using a Jeol JEM 2200FS field emission transmission microscope.
Chemical compositions of the prepared catalysts were examined by
energy-dispersive X-ray spectroscopy (EDS; HORIBA 7953-H) and field
emission-scanning electron microscopy (FE-SEM; Hitachi S-4800). The
Pd and Co contents on the activated carbon support were determined
using a Perkin-Elmer OPTIMA 7300 DV inductively coupled plasma
atomic emission spectroscopy (ICP-AES) instrument.
2. Experimental
2.1. Chemicals
Activated carbon was obtained from Carbot Corporation (Norit,
2
catalyst carrier grade, surface area = 1007 m /g). Potassium tetra-
chloropalladate(Ⅱ) (K
Co(CH COOH) ∙4H O, 98%) were supplied by Strem-Chemicals and
Alfa Aesar, respectively. Ferroin indicator (0.1 wt% in H O), sodium
borohydride (NaBH , > 96%), and sodium bromide (NaBr, > 99%)
were purchased from Sigma-Aldrich. Cerium(IV) sulfate (Ce(SO
.1 N standardized solution) was obtained from Merck Millipore.
Hydrogen peroxide (H , 34.5%), methanol (CH OH, 99.9%), and
85%) were purchased from Samchun
2 4
PdCl , 99%) and cobalt(Ⅱ) acetate tetrahydrate
(
3
2
2
2
4
4 2
) ,
0
2.4. Direct synthesis of H
2 2
O
2
O
2
3
phosphoric acid (H
Chemicals, Korea.
3 4
PO ,
The DSHP reaction was performed as previously described [28,29].
Briefly, DSHP was performed at 275 K in a jacketed high pressure re-
actor connected to a circulator. Typically, 20 mg of the catalyst and
25 mL of methanol/water mixture (v/v = 4/1) containing 20 ppm NaBr
were added to the reactor. After sweeping the residual gases, the re-
2.2. Catalyst preparation
Pd@Co (BO core-shell NPs on activated carbon and an acid-
3
3
)
2
2 2 2
actor was filled with mixed reactant gases (H /O /N (vol%) = 3.8/
treated Pd/C catalyst were prepared based on the modified non-noble
metal sacrificial approach reported previously [30]. First, cobalt borate
5.0/91.2), and pressurized to 30 bar using mass flow controllers. After
the reactor pressure was stabilized at 30 bar, the reaction was com-
menced with vigorous stirring (1200 rpm) using a magnetically-driven
stirrer. The reaction was maintained for 1 h or extended to 2–4 h.
During the reaction, the total flow rate of the mixed gases was
(
3 3 2 4
Co (BO ) ) was prepared by addition of 0.03 M aqueous NaBH solu-
tion to cobalt(Ⅱ) acetate tetrahydrate (0.576 mmol) dissolved in 50 mL
of deionized water. After stirring for 2 h, a black precipitate was
2

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
2
maintained at 40 mL/min. To investigate the change in H concentra-
tion during the reaction, the off-gas was analyzed using an Agilent 6890
gas chromatograph (GC) equipped with a Carboxen-1010 PLOT capil-
lary column and thermal conductivity detector (TCD). The H
tent was measured by titration using an acidic cerium(IV) sulfate so-
lution and ferroin indicator. H conversion and H selectivity were
calculated based on the determined H and H concentrations by the
volumetric titration and GC analysis, respectively.
2 2
O con-
2
2 2
O
2
O
2
2
2 2
2.5. H O decomposition and hydrogenation
H
2
O
2
decomposition and hydrogenation tests were performed at
73 K under atmospheric pressure using the prepared catalyst (20 mg)
in a methanol/water mixed solution in the presence of
2
2 2
and 1 wt% H O
nitrogen or hydrogen gas (10 mL/min), respectively. During the ex-
periment, the liquid sample was periodically extracted and titrated to
monitor changes in the H
2
O
2
concentration.
Fig. 2. PXRD changes during catalyst preparation.
3. Results and discussion
and that Pd was exposed to the external surface.
3 2
Fig. 3. also indicates that the formation of the Pd@Co(BO ) core-
3
.1. Catalyst preparation and characterization
shell structure depended on the Co/Pd ratio. With increasing Co/Pd
ratio, the Pd (111) peak decreased in intensity (Fig. 3(a)). However, the
peak intensity of the Pd (111) phase was maintained even when the Co/
Pd ratio was > 6, indicating that perfect encapsulation was hardly
achievable and a certain portion of Pd would be exposed even at a high
Co/Pd ratio. A similar trend in the PXRD change was also observed in
the presence of activated carbon indicating that the existence of support
has little effect on the formation of the core-shell structure (Fig. 3(b)).
The formation of the Pd@Co(BO ) core-shell structure was further
3 2
confirmed by the elemental composition analysis of the catalyst surface
by XPS. As shown in Fig. 4, the metal composition of the catalyst sur-
face was greatly changed after cobalt etching. For Co Pd/C, while there
6
Generally, the Pd/C catalyst can be easily prepared by addition of a
reducing agent after adsorption of a palladium precursor on the carbon
support. However, the formation of large Pd NPs with broad particle
size distributions owing to uncontrollable particle size growth during
reduction is a major reason for the low activity of conventional Pd/C
catalysts for DSHP [32]. Moreover, Pd NPs in the internal micropores of
the activated carbon are hardly accessible and can cause severe mass
transfer resistance. To suppress unwanted aggregation and diffusion
problems of this preparation method, a non-noble metal sacrificial
method [30] was applied. As described in Fig. 1, catalyst preparation
consisted of two steps: i) formation of the Pd@Co
structure on the activated carbon and ii) subsequent cobalt borate shell
removal and acid-adsorption by addition of H PO . During the first
reduction step, the Pd NPs were surrounded by Co (BO which was
formed in-situ by the reaction of the non-noble metal precursor, Co
CH COOH) ∙4H O with NaBH . Because Co (BO can be dissolved in
PO solution, the non-noble metal shell was easily etched from the
catalyst upon addition of H PO . The color change of solution to pale
pink indicated cobalt leaching. During the second stripping process,
PO was strongly adsorbed on the activated carbon support, even
though after thorough washing. The resulting acid-treated activated
carbon can act as a solid acid support.
Fig. 2 shows the PXRD peak changes during each catalyst prepara-
tion step. While the as-prepared Co (BO ) exhibited an amorphous
3 3 2
structure, the peak positions of the annealed sample were well-matched
with the literature data (JCPDS: 25-0102), confirming the formation of
3
(BO
3
)
2
core-shell
were no appreciable peaks corresponding to palladium species, two
distinctive binding energy peaks for cobalt species were observed
(Fig. 4(a)). These cobalt peaks at 784.5 and 802 eV were ascribed to the
2p3/2 and 2p1/2 orbitals of Co2+ in cobalt borate, respectively [33].
This result indicates a high concentration of cobalt species on the ex-
ternal surface of the catalyst, with most of the palladium surface cov-
3
4
3
3 2
)
(
H
3
2
2
4
3
3 2
)
ered with Co
Pd@Co(BO
After stripping with H
3
(BO
core-shell structure.
PO , the intensities of the peaks corre-
3 2
) , strongly demonstrating the formation of the
3
4
3 2
)
3
4
3
4
H
3
4
sponding to the cobalt species were greatly reduced, and the char-
acteristic peaks of palladium species appeared (Fig. 4(b)). The high
intensity of the Pd XPS spectra confirmed that cobalt was successfully
removed from the catalyst, and palladium was exposed to the external
surface. In addition, the Pd 3d XPS spectra indicated two different states
2
+
of orbital doublets assigned to Pd
(∼342.8 eV (3d3/2) and 337.5 eV
0
(3d5/2)) and Pd (∼340.7 eV (3d3/2) and 335.3 eV (3d5 2/ 2)) [34]. These
+
0
Co
suggested that a significant portion of Pd was covered with Co
In contrast, the intense Pd (111) peak for (Co )Pd/C clearly indicated
that the cobalt borate shell was successfully removed from the catalyst
3
(BO
3
)
2
. For Co
6
Pd and Co
6
Pd/C, the weak Pd (111) peak at 40°
results indicate that palladium exists as a mixture of Pd /Pd in (Co
Pd/C, suggesting that the Pd NPs may be partially oxidized by the
PO treatment.
EDS analysis was also carried out to examine the elemental
6
)
3
(BO
3 2
) .
H
3
4
6
3 4
Fig. 1. Preparation of the acid-anchored Pd/C catalyst via formation of a Pd@Co core-shell structure and subsequent cobalt removal with H PO .
3

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
x x
Fig. 3. PXRD changes as a function of the Co/Pd ratio: (a) Co Pd series and (b) (Co )Pd/C series (x = 2–10).
disappeared and a new phosphorous peak was observed in the EDS
spectrum of (Co )Pd/C (Fig. 5(d)). EDS mapping of (Co )Pd/C offers
6
6
additional information regarding the distribution of each element after
etching. While no detectable peak was observed in its EDS spectrum, Co
mapping indicates that residual cobalt remained after cobalt removal,
as suggested by the XPS analysis (Fig. 5(e)). In contrast, EDS mappings
of Pd and P suggest that palladium was slightly aggregated, but phos-
phoric acid was widely anchored on the activated carbon support
during etching (Fig. 5(f) and (g)). The EDS analysis with other char-
acterization methods clearly demonstrated the formation of the Pd@Co
(
BO
non-noble metal shell removal and simultaneous strong adsorption of
PO during the second etching step.
To examine the changes in the NP size in the catalysts during the
3 2
) core-shell structure during the first reduction step and successful
H
3
4
various preparation steps, TEM and HAADF-STEM analyses were per-
formed. Representative images and particle size distributions of the
catalysts provided in Figs. 6 and 7 clearly show that the NP size varied
depending on the preparation step. From Fig. 7(a) and (e), the forma-
tion of large and polydisperse NPs (18.3 ± 2.3 nm) with two distinct
lattice distances (Fig. 6(a)) suggests that the palladium and cobalt co-
Fig. 4. Changes in the XPS spectra before and after cobalt etching: (a) Co (2p)
and (b) Pd (3d).
existed in Co
was remarkably reduced (5.6 ± 0.9 nm), indicating that the formation
of the Pd@Co(BO core-shell structure effectively suppressed Pd NP
growth during the reduction, which was advantageous for the
6
Pd/C. After cobalt removal, the NP size and distribution
composition of the catalyst surface. As shown in Fig. 5(a) and (b), co-
balt was well dispersed on the surface of Co Pd/C. The detection of Pd
6
also indicated that not all Pd was surrounded by the cobalt borate shell,
as indicated by the XRD analysis (Fig. 5(c)). In contrast, the cobalt peak
3 2
)
Fig. 5. EDS analysis of Co
6
Pd/C (a–c) and (Co
6
)Pd/C (d–g) catalysts: EDS spectra (a), Co mapping (b), and Pd mapping (c) of Co
6
)Pd/C.
6
Pd/C; EDS spectra (d), Co mapping
(e), Pd mapping (f), and P mapping (g) of (Co
4

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
6 6 6 6
Fig. 6. TEM and HAADF-STEM images of the prepared catalysts: (a) Co Pd/C, (b) (Co )Pd/C, (c) (Co )Pd/C-Cal, and (d) (Co )Pd/C-Red.
formation of small Pd NPs with a narrow size distribution (Figs. 6(b)
and 7(b) and (e)). The formation of larger Pd NPs during the prepara-
tion of conventional Pd/C via adsorption-reduction method also sug-
gests the effectiveness of core-shell strategy in regulating the growth of
Pd NP during reduction (Fig. S1).
Fig. 7(c) and (e) indicates that the partial aggregation of Pd NPs oc-
curred during heat treatment. However, smaller NP sizes with narrow
distribution (4.9 ± 0.9 nm) observed in (Co )Pd/C-Red (Fig. 7(d) and
6
(e)) suggests that post-reduction with hydrogen negligibly affected the
sintering of the Pd NPs.
Moreover, the cobalt borate shell was successfully removed without
any apparent aggregation of Pd NPs during the following stripping step.
The slightly increased Pd NP size with a broader size distribution
To investigate changes in the electronic nature of Pd upon calci-
nation and reduction post heat-treatment, XPS analysis was performed
again. As shown in Fig. 8(b), the strong Pd 3d5/2 peak at 335 eV and Pd
(
8.3 ± 1.6 nm) of the calcined catalyst, (Co
6
)Pd/C-Cal, shown in
6
3d3/2 peak at 342.8 eV in the XPS spectra of (Co )Pd/C-Cal correspond
Fig. 7. Histogram and particle size distributions of the metal nanoparticles incorporated on the activated carbon support: histogram for (a) Co
c) (Co )Pd/C-Cal, and (d) (Co )Pd/C-Red; (e) related particle size distribution.
6 6
Pd/C, (b) (Co )Pd/C,
(
6
6
5

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
2 2
Pd NP size exists in which a certain palladium phase selective for H O
formation is preferentially exposed. In this context, NP sizes within the
sub-nano dominating region (1.5–5 nm including error range) is su-
perior for obtaining high H
the (Co )Pd/C catalyst is slightly larger than that specified as the sub-
nano dominating region, it effectively promoted the selective formation
of H . In addition, the electronic features of Pd may affect H se-
lectivity. As the amount of reduced palladium species increases, H
decomposition and hydrogenation rates will accelerate [27,28,36,37].
2 2
O yields. While the average Pd NP size in
6
2
O
2
2 2
O
2 2
O
Therefore, the presence of the mixed Pd2+/Pd phase with high Pd
0
2+
concentration is advantageous for suppressing unfavorable side reac-
tions.
The post-treatment of (Co
lytic activity. For the (Co )Pd/C-Cal sample, all reaction performances
decreased due to the enlarged Pd NP size, despite the higher oxidation
state of palladium after calcination. After reduction, the H conversion
was greatly increased at the expense of a steep decrease in the H
selectivity. The catalytic behavior of (Co )Pd/C-Red can be rationalized
in terms of the enhanced metallic nature of Pd by H reduction. In other
words, Pd -rich active sites accelerated the unwanted H hydro-
genation with additional H consumption. Considering that the reaction
rate of H O hydrogenation is much faster than that of H O formation
6
)Pd/C significantly influenced the cata-
6
2
2 2
O
6
6
Fig. 8. XPS spectra of the Pd(3d) region of the prepared catalyst: (a) (Co )Pd/C,
2
(b) (Co
6
)Pd/C-Cal, and (c) (Co
6
)Pd/C-Red.
0
2 2
O
2
to Pd2+, suggesting that Pd was significantly oxidized by the calcina-
2
2
2 2
tion. After reduction, the enhanced intensity of the characteristic peaks
2 2
and is responsible for the 70–85% decrease in H O selectivity [38], the
0
of Pd at 340.7 eV for Pd 3d3/2 and at 335.3 eV for Pd 3d5/2 implied that
accelerated H O2 hydrogenation rate accounts for the observed in-
2
2
+
0
the Pd /Pd ratio was considerably decreased, enhancing the metallic
nature of Pd by H reduction (Fig. 8(c)).
creased H2 conversion and decreased H O2 selectivity.
2
2
The significant influence of H O2 hydrogenation on the catalytic
2
performance was confirmed by investigation of the time-dependent
changes in H
the (Co )Pd/C catalyst. A considerable increase in H
significantly decreased H
run for 4 h, indicating that H
2
O
2
productivity, H
2
O
2
selectivity, and H
2
conversion over
conversion with
3
2 2
.2. Direct synthesis, decomposition, and hydrogenation reaction of H O
6
2
2
O
2
selectivity was observed after an extended
hydrogenation became dominant with
The reaction performances of a range of catalysts for DSHP are
summarized in Fig. 9. Both H production rate and selectivity were
quite low in the presence of the Co Pd/C catalyst, which can be at-
tributed to the covering of most of the active palladium surface with the
cobalt borate shell. After cobalt removal, the catalytic activity was
greatly improved, with a H
Pd.h with 81% H selectivity. The reaction performance of the (Co
Pd/C catalyst was impressive compared to that of the conventional Pd/
C catalyst (586 mmol-H /g-Pd.h with 5.8% H selectivity) [32].
The excellent activity of the (Co )Pd/C catalyst can be rationalized
by NP size and electronic nature of Pd. The former can be explained by
the relation between Pd NP size and the active site structure proposed
by Tian et al [35]. According to their classification, an optimal range for
2 2
O
2 2
O
increasing reaction time (see Fig. 10).
6
A significant challenge for DSHP is that all reaction routes described
in Scheme 1 are thermodynamically feasible and are promoted by the
same active metal such as Pd. Indeed, as shown in Fig. 11, the nature of
the catalyst significantly influenced the decomposition and hydro-
2 2 2 2
O production rate of 5721 mmol-H O /g-
2
O
2
6
)
genation of H
rates were observed for the Co
cobalt accelerated the side-reactions and was responsible for the poor
yield. However, for the (Co )Pd/C catalyst, both the rates of H
decomposition and hydrogenation were considerably reduced, high-
2
O
2
. The fastest H
2
O
2
decomposition and hydrogenation
6
Pd/C catalyst, clearly indicating that
2
O
2
2 2
O
6
H
2
O
2
6
2 2
O
2 2
lighting the detrimental effect of cobalt on H O yield. The undesirable
Fig. 9. Direct synthesis of H
2
O
2
over the Pd/C catalysts prepared herein.
6

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
6
Fig. 10. Changes in the activity of the (Co )Pd/C catalyst during an extended run.
2 2 2 2
Fig. 11. (a) H O decomposition and (b) H O hydrogenation over the Pd/C catalysts prepared herein.
x x
Fig. 12. (a) Catalytic activity of the Co Pd/C series for the DSHP and (b) pH change of the reaction solution in the presence of the Co Pd/C catalysts (x = 2–10).
side-reactions were suppressed by the calcination of (Co
oxidation state of Pd increased.
6
)Pd/C, as the
catalyze the side reactions as other cobalt species did [39–41]. More-
over, as basicity of the reaction solution increased, the produced H
would be more rapidly decomposed or hydrogenated to form H O.
For the (Co )Pd/C series as shown in Fig. 13, greatly improved re-
action performances were obtained regardless of the Co/Pd ratio, in-
dicating that cobalt removal was indispensable to obtain a high H
2 2
O
2
x
3.3. Factors affecting reaction performance of the catalysts
2 2
O
Fig. 12 shows the effect of the Co/Pd ratio on the catalytic activity
and pH of the reaction solution in the presence of the Co Pd/C catalyst.
All indicators of catalyst reaction performance decreased as the Co/Pd
ratio increased. In particular, the H selectivity became very low. The
poor activity of the Co Pd/C catalysts can be attributed to the negative
yield. Moreover, the low pH of the reaction solution (ca. 3.5; Fig. 13(b))
suggested that phosphoric acid was strongly anchored on the activated
carbon support, which was beneficial for suppressing the charge se-
x
2
O
2
−
+
paration of H
conversion with Co/Pd ratio observed in the activity tests of (Co
series indicated that the residual amount of cobalt as well as the Pd
2
O
2
to HO
2
and H [5,15]. The gradual decrease of H
2
x
x
)Pd/C
effects of cobalt on the DSHP reaction by promoting side-reactions and
increasing the pH (Fig. 12(b)). It is possible that the cobalt borate might
7

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
x x
Fig. 13. (a) Catalytic activity of the (Co )Pd/C series for the DSHP and (b) pH change of the reaction solution in the presence of (Co )Pd/C catalysts (x = 2–10).
x x
Fig. 14. Schematic illustration of the different catalysis modes on (a) Co Pd/C and (b) (Co )Pd/C catalysts (x = 2–10).
content of catalyst impacted the reaction performance (Fig. S2).
Based on the discussion above, it is clear that cobalt plays an im-
portant role in determining the reaction performance of the catalyst by
virtue of its dual functions. The in-situ formed cobalt borate shell
important.
Acknowledgements
around palladium during reduction with NaBH
4
efficiently suppressed
This research was supported by Basic Science Research Program
through the National Research Foundation of Korea (NRF) funded by
the Ministry of Education (2016R1D1A3B02006928).
the uncontrollable particle size growth of palladium. The formation of
2
+
0
small Pd NPs with a mixed Pd /Pd phase after cobalt removal from
the Pd@Co(BO core-shell structure was essential for obtaining a high
yield. However, cobalt should be removed to a significant extent
during etching because it considerably decreased the H selectivity of
a catalyst by facilitating the H decomposition and hydrogenation, as
3 2
)
H
2
O
2
Appendix A. Supplementary data
2 2
O
2
O
2
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.cattod.2019.09.038.
described in Fig. 14. It should also be noted that the stable anchoring of
phosphoric acid during cobalt stripping enabled the acid-treated acti-
vated carbon to act as a solid acid support during the DSHP reaction.
References
4. Conclusion
[1] S.J. Gustaaf Goor, Jurgen Glenneberg, hydrogen peroxide, Ullmann’s Encycl. Ind.
Chem. 18 (2007) 393–427.
[
2 2
2] H O (Hydrogen Peroxide) Measurement Market - Global Industry Analysis, Size,
Herein, an unprecedented Pd/C catalyst design was proposed to
Share, Growth, Trends and Forecast (2016) 2024. Transparency Market Research,
2016.
overcome the unsatisfactory reaction performances of conventional Pd/
C catalysts for promoting the direct synthesis of H . The catalyst was
realized by: i) formation of the Pd@Co (BO core-shell structure on
an activated carbon and ii) stripping of the cobalt borate shell with
PO . The non-noble metal shell surrounding the Pd NPs effectively
hindered unfavorable sintering of Pd during reduction, allowing small
and monodisperse Pd NPs to form. Moreover, H PO significantly in-
[
[
3] V.S.S. Ranganathan, Catalysts 8 (2018) 379–400.
4] J.M. Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Angew. Chemie - Int. Ed. 45
2 2
O
3
3 2
)
(2006) 6962–6984.
[5] C. Samanta, Appl. Catal. A Gen. 350 (2008) 133–149.
[
[
6] R.J. Lewis, G.J. Hutchings, ChemCatChem. 11 (2019) 298–308.
7] F. Menegazzo, M. Signoretto, E. Ghedini, G. Strukul, F. Menegazzo, M. Signoretto,
E. Ghedini, G. Strukul, Catalysts 9 (2019) 251–282.
H
3
4
3
4
[8] Y.F. Han, J.H. Lunsford, J. Catal. 230 (2005) 313–316.
[9] Q. Liu, J.H. Lunsford, J. Catal. 239 (2006) 237–243.
10] V.R. Choudhary, C. Samanta, J. Catal. 238 (2006) 28–38.
11] C. Samanta, V.R. Choudhary, Appl. Catal. A Gen. 330 (2007) 23–32.
12] V.R. Choudhary, P. Jana, Appl. Catal. A Gen. 352 (2009) 35–42.
fluenced the electronic nature of Pd and the property of support by
decreasing the metallic nature of Pd and imparting acidity to the acti-
vated carbon support, respectively. The resulting acid-treated Pd/C
[
[
[
2
+
0
catalyst bearing small and mixed Pd /Pd NPs exhibited promising
reaction performance for directly catalyzing the formation of H . The
)Pd/C catalyst (5721 mmol-
/g-Pd.h and 88.1%) surpassed those of conventional Pd/C catalyst
/g-Pd.h and 5.8%). The elegant catalyst design based
[13] T. Deguchi, H. Yamano, M. Iwamoto, Catal. Today 248 (2015) 80–90.
[
14] G. Gallina, J. García-Serna, T.O. Salmi, P. Canu, P. Biasi, Ind. Eng. Chem. Res. 56
2017) 13367–13378.
15] D.W. Flaherty, ACS Catal. 8 (2018) 1520–1527.
[16] R.B. Rankin, J. Greeley, ACS Catal. 2 (2012) 2664–2672.
17] J.K. Edwards, B. Solsona, E. Ntainjua, N.A.F. Carley, A.A. Herzing, C.J. Kiely,
G.J. Hutchings, Science 323 (2009) 1037–1041.
2 2
O
(
H
H
(
2
O
O
2
2
productivity and selectivity of (Co
6
[
2
[
[
2 2
586 mmol-H O
on a non-noble metal sacrificial method may offer a versatile solution to
18] S.J. Freakley, Q. He, J.H. Harrhy, L. Lu, D.A. Crole, D.J. Morgan, E.N. Ntainjua,
other reactions in which metal dispersion and support acidity are
8

S. Lee and Y.-M. Chung
Catalysis Today xxx (xxxx) xxx–xxx
J.K. Edwards, A.F. Carley, A.Y. Borisevich, C.J. Kiely, G.J. Hutchings, Science 351
2016) 965–968.
[28] S. Lee, H. Jeong, Y. Chung, J. Catal. 365 (2018) 125–137.
(
[29] S. Lee, Y.M. Chung, Mater. Lett. 234 (2019) 58–61.
[
[
[
19] S. Wang, K. Gao, W. Li, J. Zhang, Appl. Catal. A Gen. 531 (2017) 89–95.
20] F. Li, Q. Shao, M. Hu, Y. Chen, X. Huang, ACS Catal. 8 (2018) 3418–3423.
21] N.M. Wilson, P. Priyadarshini, S. Kunz, D.W. Flaherty, J. Catal. 357 (2018)
[30] Y. Chen, Q.L. Zhu, N. Tsumori, Q. Xu, J. Am. Chem. Soc. 137 (2015) 106–109.
[31] H. Jeon, Y.-M. Chung, Appl. Cat. B Environ. 210 (2017) 212–222.
[32] N. Lee, Y.-M. Chung, Res. Chem. Intermed. 42 (2016) 95–108.
[33] C. Xie, Y. Wang, D. Yan, L. Tao, S. Wang, Nanoscale. 9 (2017) 16059–16065.
[34] 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, ACS Catal. 2 (2012) 1042–1048.
1
63–175.
[
[
[
[
22] B. Hu, W. Deng, R. Li, Q. Zhang, Y. Wang, F. Delplanque-Janssens, D. Paul,
F. Desmedt, P. Miquel, J. Catal. 319 (2014) 15–26.
23] G.M. Lari, B. Puértolas, M. Shahrokhi, N. López, J. Pérez-Ramírez, Angew. Chemie -
Int. Ed. 56 (2017) 1775–1779.
24] T. García, S. Agouram, A. Dejoz, J.F. Sánchez-Royo, L. Torrente-Murciano,
[35] P. Tian, L. Ouyang, X. Xu, C. Ao, X. Xu, R. Si, X. Shen, M. Lin, J. Xu, Y.F. Han, J.
Catal. 349 (2017) 30–40.
[36] J.K. Edwards, J. Pritchard, M. Piccinini, G. Shaw, Q. He, A.F. Carley, C.J. Kiely,
G.J. Hutchings, J. Catal. 292 (2012) 227–238.
[37] D. Gudarzi, W. Ratchananusorn, I. Turunen, M. Heinonen, T. Salmi, Catal. Today
248 (2015) 58–68.
B. Solsona, Catal. Today 248 (2015) 48–57.
25] 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. 6
(
2016) 694–697.
[38] I. Huerta, J. García-Serna, M.J. Cocero, J. Supercrit. Fluids 74 (2013) 80–88.
[39] S. Hanaoka, J.-M. Lin, M. Yamada, Anal. Chim. Acta 426 (2001) 57–64.
[40] N.-A.M. Deraz, Mater. Lett. 57 (2002) 914–920.
[
26] R. Arrigo, M.E. Schuster, S. Abate, G. Giorgianni, G. Centi, S. Perathoner,
S. Wrabetz, V. Pfeifer, M. Antonietti, R. Schlögl, ACS Catal. 6 (2016) 6959–6966.
27] S. Yook, H.C. Kwon, Y.G. Kim, W. Choi, M. Choi, ACS Sustain. Chem. Eng. 5 (2017)
[
[41] M. Makhlouf, B.M. Abu-Zied, T.H. Mansoure, Appl. Surf. Sci. 274 (2013) 45–52.
1
208–1216.
9