Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia

Article abstract of DOI:10.1016/j.cattod.2014.12.033

In this work Pd (0.3 or 0.6?wt.%) was supported on both Zr_xM_1–xO₂ (M?=?La, Y, Ce) and on mechanical mixtures of CeO₂ and ZrO₂. The synthesized catalysts were characterized by XRD, TPR, AAS and CO chemisorption and tested for the direct synthesis of hydrogen peroxide in a high pressure semibatch apparatus. The reactants conversion was limited in order to avoid mass-transfer limitations. No selectivity enhancers of any kind were used and the all the materials were halide free. Small metal particles were obtained (1–2.6?nm). Supports with smaller pore diameters leaded to larger Pd particles, which in turn were found to preferentially support the formation of the peroxide. Moreover, supports with higher reducibility favored the production of H₂O₂, probably due to an easier reduction of the active metal, essential to achieve high selectivity. Notwithstanding the absence of enhancers, the specific activity and selectivity recorded were very high.

Full text of DOI:10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Direct synthesis of H2O2 over Pd supported on rare earths promoted
zirconia
A. Bernardinia, N. Gemob, P. Biasib,c,∗, P. Canua, J.P. Mikkolab,c, T. Salmib, R. Lanzad
a Universitá degli Studi di Padova, Dipartimento di Ingegneria Industriale, Via Marzolo n. 9, 35131 Padova, Italy
b Åbo Akademi University, Laboratory of Industrial Chemistry and Reaction Engineering, Process Chemistry Centre, Åbo Akademi University, Biskopsgatan 8,
20500 Turku/Åbo, Finland
c Umeå University, Department of Chemistry, Chemical-Biochemical Centre (KBC), Technical Chemistry, Umeå University, SE-90187 Umeå, Sweden
d KTH—Royal Institute of Technology, Department of Chemical Engineering and Technology, Chemical Technology, SE-100 44 Stockholm, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
In this work Pd (0.3 or 0.6 wt.%) was supported on both ZrxM1–xO2 (M = La, Y, Ce) and on mechanical
mixtures of CeO2 and ZrO2. The synthesized catalysts were characterized by XRD, TPR, AAS and CO
chemisorption and tested for the direct synthesis of hydrogen peroxide in a high pressure semibatch
apparatus. The reactants conversion was limited in order to avoid mass-transfer limitations. No selec-
tivity enhancers of any kind were used and the all the materials were halide free. Small metal particles
were obtained (1–2.6 nm). Supports with smaller pore diameters leaded to larger Pd particles, which in
turn were found to preferentially support the formation of the peroxide. Moreover, supports with higher
reducibility favored the production of H2O2, probably due to an easier reduction of the active metal,
essential to achieve high selectivity. Notwithstanding the absence of enhancers, the specific activity and
selectivity recorded were very high.
Received 13 October 2014
Received in revised form
11 December 2014
Accepted 15 December 2014
Available online xxx
In memory of the Charismatic Professor
Benedetto Corain (July 8th
1941–September 24th 2014), a catalyst for
our research activities and ideas.
© 2015 Elsevier B.V. All rights reserved.
Keywords:
Palladium
H2O2 direct synthesis
Zirconia
Ceria
Rare earths
1. Introduction
consumption route of the peroxide produced [5–7]. The direct syn-
thesis of water also contributes to impair the selectivity [5]. Catalyst
Hydrogen peroxide is considered one of the most environ-
mentally friendly and efficient oxidizing agents available to the
industry. Its direct production from H2 and O2 (direct synthesis,
DS) has been attracting considerable interest because of the poten-
tially reduced costs and flexibility of the process [1,2]. Although
the DS is in principle the simplest route to the peroxide, the reac-
tions involve multiple consecutive and parallel paths leading to
water (Scheme 1), which is by far thermodynamically favored. This
adversely affects the selectivity and limits the yield of H2O2.
Therefore, the selectivity towards H2O2 is the first issue for a
viable industrial application. Many Authors investigated the reac-
tion conditions, trying to suppress the undesired side reactions
[1–4]. The hydrogenation of hydrogen peroxide appears the major
design is the main tool to improve the selective synthesis of H2O2.
Researchers investigated the catalyst properties to limit the unde-
sired side reactions [1,2]. Pd has proven to be the best active metal,
also on standard supports such as carbon, SiO2 or Al2O3. Mineral
acids and halides (either Cl− or Br−) are known to be effective selec-
tivity enhancers [1,2], especially in conjunction, but corrosion and
metal leaching could become relevant issues. In addition, halide
ions can be detrimental in down-stream applications.
Catalytic supports are known to play an important role in the
selectivity towards the peroxide because of the interaction with
the active metal. Hutchings et al. stated that the major parameter
to favor the peroxide formation is a low isoelectric point, so that C
was favored upon Al2O3, TiO2 and MgO [8]. Acidic supports were
found very promising also in others studies, specifically deposit-
ing Pd on a mesoporous ion exchange resin functionalized with
sulfonic groups [9,10]. More recently, Abate et al. [11,12] inves-
tigated Pd supported on SiO2, SBA15-15, ␥-Al2O3 and N-CNT and
found no apparent relationship between reactivity and selectiv-
ity and the acidity of the support, concluding that their data better
∗
Corresponding author at: Åbo Akademi Univeristy, Department of Chemical
Engineering, Laboratory of Industrial Chemistry, Biskopsgatan 8, 20500 Turku/Åbo,
Finland. Tel.: +358 22154555.
E-mail address: bpierdom@abo.fi (P. Biasi).
http://dx.doi.org/10.1016/j.cattod.2014.12.033
0920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
Table 1
2
List, composition and suppliers of the materials used.
Label
Support (wt.%)
Supplier
Product code/CAS
Pd2
Pd3
Pd4
Pd5
Pd6
Pd7
ZrO2/CeO2 82.5/17.5
Ce doped ZrO2
ZrO2/CeO2 50/50
ZrO2/CeO2 17.5/82.5
ZrO2/Y2O3 92/8
ZrO2/CeO2/La2O3
77.5/17.5/5
MEL Chemicals
MEL Chemicals
Evonik/Alfa Aesar
Evonik/Alfa Aesar
MEL Chemicals
MEL Chemicals
XZO 802
XZO
0857
XZO 1012
XZO 892
Scheme 1. Reactions involved in the direct synthesis of H2O2.
CeO2
ZrO2
Evonik
Alfa Aesar
VP AdNano® Ceria 50
1314–23–4
Pd8
correlated with the metal particle size [11]. Pd-based catalysts sup-
ported on silica were also successfully tested by Strukul et al. [13].
In their work, the authors concluded that SBA-15 was the most
suitable support due to its ability to give the ideal Pd dispersion.
Menegazzo et al. [14,15] compared Pd and Pd-Au catalysts over
SiO2, ZrO2, ZrO2/SO4−, CeO2, CeO2/SO4−, confirming that Pd/SiO2
gave the highest selectivity [14]. The authors also proposed that a
lower selectivity was achieved with the other supports because of
the smaller Pd particles formed, preferentially leading to the for-
mation of water [14,16]. It has also been shown that the reduced
form of Pd is apparently more active and selective in the H2O2 direct
synthesis [9,17], with Pd oxide been almost inactive [17].
from Alfa Aesar, was used as metal precursor. The metal loading
was 0.3 wt.% for the samples prepared using MEL Chemicals sup-
ports and 0.6 wt.% for the remaining ones. A higher metal loading
was chosen for the latter samples to assure a high enough activity,
since they are not expected to form a solid solution, which repre-
sents a more active support. All the support used, their composition
and the catalyst prepared are listed in Table 1. About 8 g of each
support were calcined in air atmosphere at 550 ◦C for 5 h, with a
heating rate of 5 ◦C/min. The calcination conditions were chosen
in order to assure the formation of a solid solution between ceria
and zirconia. Indeed, according to MEL Chemicals, their supports
form a solid solution upon calcination at 500–550 ◦C. The calcined
supports were afterwards impregnated with a solution of Pd(NO3)2.
The impregnation was done in 3 steps on samples 2, 4, 5 and 8, while
the remaining supports needed 6 (sample 7) and 7 steps (samples
3 and 6), due to their low pore volume.
Following these inspiring results, in this work we tried to
investigate the relationship between the support properties, the
deposition of the active metal phase and the activity and selec-
tivity of the resulting catalysts. To the best of our knowledge, no
studies have yet been done using ceria and zirconia mixtures and
solutions (ZrxCe1–xO2) or zirconia stabilized with rare earths. These
materials present unique features, such as increased dispersion
[18] and O2 mobility [19], which favors the active metal activity,
especially in conditions in which the reducibility of the catalytic
systems plays an important role [20]. Furthermore, the properties
of these materials can be easily changed and adapted to specific
applications. Indeed, their surface area, pore volume, particle size,
homogeneity etc, can be tuned with different preparation proce-
dures [21,22]. An increased metal–support interaction should also
allow decreasing the metal loading, which would reduce the cost
of the catalysts. We focused on Pd supported on mixed oxides
(ZrO2 and CeO2) and rare earth elements (Y2O3 and La2O3) with
different surface properties in order to access different particle
size and metal dispersions. Both homogeneous solid solutions and
mechanical mixtures of ZrO2 with different promoters were tested.
Catalysts prepared by incipient wetness impregnation were charac-
terized by X-ray diffraction, temperature programmed reduction,
CO chemisorption, atomic absorption spectroscopy and nitrogen
adsorption analyses. Hydrogen peroxide direct synthesis was per-
formed in a novel high pressure semibatch apparatus. Although
they are known to favor the selectivity toward the peroxide, nei-
ther promoters nor stabilizers (halides and acids) were introduced
in the reaction environment to focus on the support–Pd interac-
tion solely. In order to avoid any traces of halides, all catalysts were
prepared from halide free chemicals. The support properties were
related to the structure of the active metal particles, which in turns
was found to affect the activity and selectivity toward the peroxide.
The surface area, pore volume and the pore average diame-
ter of each support are reported in Table 2. The powders were
dried at 90 ◦C in a ventilated oven for 2 h after each impregna-
tion step. All the samples were calcined in air at 500 ◦C for 4 h,
increasing the temperature at a rate of 5 ◦C/min. Finally, before the
catalytic tests, all the catalysts were reduced in a furnace in pure
H
2 flow (50 N ml/min) at 300 ◦C for 2 h, increasing the temperature
at 5 ◦C/min in 50 N ml/min argon flow.
2.2. Catalyst characterization
The surface area and porosity analyses were performed with
a Micromeritics ASAP 2000 unit. The samples were de-gassed by
evacuation at 250 ◦C for a minimum of 4–5 h prior to analysis. Data
were collected at liquid nitrogen boiling temperature (77 K). The
surface area was calculated by the BET (Brunauer–Emmett–Teller)
method with data collected at relative pressures between 0.06
and 0.2. The total pore volume was measured at p/p0 of 0.99.
The pore diameter distribution was determined using the BJH
(Barrett–Joyner–Halenda) method.
X-ray diffraction patterns (XRD) were obtained using a Siemens
D5000 diffractometer with a Cu K␣ monochromatic radiation, a
scanning range 2ꢀ between 10 and 90◦ and a step size of 0.02◦ and
a time/step ratio of 1.0 s.
Temperature programmed reduction (TPR) tests were per-
formed with a Micromeritics Autochem 2910 equipped with a TCD
detector. The samples were flushed in helium at 250 ◦C for 2 h prior
to analysis and cooled down to room temperature. The analysis was
run using 5% H2 in Ar up to 700 ◦C, with a heating rate of 5 ◦C/min.
The metal dispersion and average particle size were esti-
mated by means of CO chemisorption, using a Micromeritics ASAP
2020 unit. The samples were evacuated in helium at 100 ◦C for
30 min, then at 350 ◦C for 15 min. After the evacuation, the powder
was reduced with H2 at 350 ◦C for 2 h. The analysis was performed
dosing CO at 35 ◦C, assuming a CO/Pd stoichiometric ratio of 1, after
a further evacuation step.
2. Materials and methods
2.1. Catalyst preparation
Seven Pd catalyst samples were prepared by incipient wetness
impregnation of different support powders. A 1 wt.% Pd on carbon
commercial catalysts was purchased by Sigma Aldrich and used as
received. Most of the supports were provided by MEL Chemicals,
while others were obtained by mixing, crushing together and cal-
cining ceria and zirconia powders supplied by Evonik and bought
from Alfa Aesar, respectively. Palladium (II) nitrate hydrate 99.9%,
The samples metal loading was measured by atomic absorption
spectroscopy (AAS) with a Perkin Elmer 1100 B equipped with a
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
3
Table 2
Surface features of the supports used.
Sample
BET area (m2/g)*
Pore volume (cm3/g)*
Pore Ø (nm)*
AAS metal loading (%)
Dispersion (%)
Particle size (nm)
Metal surf. area (m2/gmetal
)
Pd2
Pd3
Pd4
Pd5
Pd6
Pd7
Pd8
51.7
93.4
31.4
44.2
71.6
102.0
71.3
0.216
0.115
0.198
0.310
0.125
0.118
0.290
16.7
4.9
25.2
28.0
7.0
0.29
0.34
0.66
0.62
0.26
0.30
0.54
48.5
55.6
61.9
80.7
52.5
42.4
99.1
2.3
2.0
1.8
1.4
2.1
2.6
1.1
216
248
276
359
234
189
441
4.7
16.3
*
Bare support.
Pd Lumina Hollow Cathode Lamp. 20 mg of dried sample were dis-
solved in an acquaregia solution and then diluted with deionized
water.
concentrations were rescaled on the moles of Pd to take into
account the different content of active metal in the catalysts.
ꢀ
ꢁ
ꢀ
ꢁ
H2O2 selectivity SH
and specific conversion CH
were
O
O
2
2
2
2
calculated as follows:
2.3. Reactor set up
H O
[
]
2
2
SH
=
× 100
]
2
O
2
2
H O + H O
[
]
[
2
2
2
The catalysts were tested in a slurry methanol solution, within
a 300 ml autoclave operated in semibatch mode. A schematic of the
apparatus is given in Fig. 1.
H O + H O
[
]
[
]
2
2
2
2
CH
=
× VL × 100
O
2
2
n˙ H × t
Methanol (200 ml) was introduced in the reactor first, followed
by the gas reagents. Throughout the experiments, the gas mix-
ture was continuously bubbled into the static liquid phase by a
gas diffuser via 3 mass flow controllers. The gas composition was
76.5–20–3.5 mol% N2 O2 H2, respectively, with a total flow of
300 N ml/min. Before the introduction of the catalyst, the tem-
perature was decreased at 10 ◦C and stirring started (1000 rpm),
allowing the pressure to reach the desired value (50 bar). Pressure
inside the reactor was kept constant via a back pressure controller.
The gas composition at the outlet was measured by gas chro-
matography analysis to ensure the vapor–liquid equilibrium was
reached before the test was initiated. Last, the catalyst (60 mg) was
introduced via a dedicated chamber. The reactions were assumed
to start as the catalyst was lead into the reactor. Note that introduc-
ing the catalyst last ensured no in situ oxidation/reduction of the
catalyst due to a potential contact with the reagents before the
beginning of the experiment. Moreover, this procedure allows for a
very precise identification of the beginning of the reactions. Liquid
samples were collected at increasing time up to 5 h via a sampling
valve. The gas outlet was equipped with a condenser (operating at
−18 ◦C) to ensure that no methanol left the reactor with the out-
going gas flow. The efficiency of the condenser was verified by gas
chromatography analysis, and no methanol was detected in the gas
flow leaving the system.
2
where [H2O2] and [H2O] are the concentration of peroxide and
water, respectively, VL is the liquid volume in the reactor and n˙ H
2
is the inlet molar flow of H2, t is time. Prior to the introduction of
the catalyst, water content of the liquid phase was measured and
deducted from the subsequent analysis.
3. Results and discussion
3.1. Porosity analysis
Being rather different in composition, the calcined supports
exhibit a rather wide range of BET surface area, from 30 to over
100 m2/g. The average pore diameter range is also quite wide, from
about 5 up to 28 nm, as well as the total pore volume that affected
the number of steps necessary for the impregnation of the Pd pre-
cursor. As expected, samples with high surface area present low
pore volume and pore diameter (see Table 2).
3.2. XRD
Fig. 1 shows the XRD spectra of catalysts prepared using MEL
supports. All the supports formed a solid solution between zir-
conia and the promoter present in the powder. Indeed, both Pd2
and Pd3 show peaks of tetragonal Zr0.84Ce0.16O2 as main phase.
Pd2 also shows some low peaks due to residual amounts of ZrO2,
which are not at all present in the pattern of Pd3. Pd6 shows the for-
mation of tetragonal Zr0.9Y0.1O1.95, while the Pd7 sample exhibits
overlapping peaks originated by the formation of cubic zirconium
lanthanum oxide (Zr0.9La0.1O1.95) and tetragonal zirconium cerium
oxide (Zr0.9Ce0.1O2).
H2O2 and H2O concentrations were determined by iodo-
metric and Karl–Fischer titrations, respectively. The measured
The patterns of the remaining samples are reported in Fig. 3,
where 2 references patterns are also included. CeO2–ZrO2 mix is a
sample of ceria and zirconia that were mixed and crushed together,
but not calcined, while the pattern on top of the figure is represen-
tative of a solid solution between ceria and zirconia. The spectra
in Fig. 3 shows that these samples are more crystalline compared
to those prepared with MEL supports. Indeed the peaks in Fig. 2
are broader and less intense, probably due both a loss of crys-
tallinity and a decrease of the particle size of the supports as already
reported for this kind of materials [23–25]. As expected the sam-
ples prepared by mixing, crushing and calcining different amounts
of ceria and zirconia, did not form a solid solution, but rather main-
tained the character of distinct phases. However, both the patterns
of Pd4 and Pd5 show a marked reduction of the peaks due to pure
Fig. 1. Schematic of the experimental apparatus: 1—reactor; 2—condenser; 3—high
pressure pump; 4—sampling valve; 5—catalyst chamber; MFC—mass flow con-
troller; BPC—back pressure controller.
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
4
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
Pd7
Pd6
Pd3
Pd2
10
20
30
40
50
2θ [deg]
60
70
80
90
Fig. 2. XRD patterns of the Pd samples supported on MEL Chemicals powders. Pd2: ZrO2/CeO2 82.5/17.5, Pd3: Ce doped ZrO2, Pd6: ZrO2/Y2O3 92/8, Pd7: ZrO2/CeO2/La2O3
77.5/17.5/5.
Zr0.84Ce0.16O2
CeO2-ZrO2 mix
Pd8
Pd5
Pd4
10
20
30
40
50
2θ [deg]
60
70
80
90
Fig. 3. XRD patterns of the Pd samples supported on ceria-zirconia mixtures. Pd4: CeO2/ZrO2 17/83. Pd5: CeO2/ZrO2 50/50, Pd8: CeO2. For comparison, the pattern of a non
calcined CeO2/ZrO2 17/83 is also reported, as well as the pattern of a solid solution of ceria and zirconia.
ZrO2. This can be noted mainly by observing the peak at 31.54◦ and
the pair of overlapping peaks at 24.1 and 24.5◦, which becomes
much smaller in the spectra of Pd4 and Pd5. As a matter of fact,
the formation of some Ce0.75Zr0.25O2 is also detected in both these
samples. The main peaks, though, are due to the presence of cubic
ceria and monoclinic ZrO2.
outside the pores, this resulting in bigger particles. In general, the
dispersion of all the samples is very high, ranging from 50 up to
almost 100%.
3.4. Atomic absorption spectroscopy
No peaks of Pd were detected, due to both the small particles
size and the low amount of metal loaded on the samples.
The results of the measurement of the metal loading are
reported in Table 2 and show that for all the samples the actual load-
ing is rather close to the nominal, desired one; this being 0.3 wt.% for
the samples supported on MEL Chemicals powders and 0.6 wt.% for
the remaining samples. The measurements were repeated 5 times
on each sample and the relative standard deviation was always
smaller than 2%.
3.3. CO chemisorption
As highlighted in Table 2, the support physical features con-
siderably affect the metal dispersion and the average particle size.
There seem to be a trend according to which the dispersion is both
directly dependent on the pore diameter and inversely proportional
to the surface area of the support. In particular, it appears quite clear
that supports with larger pore diameter allowed the formation of
smaller particles, thus favoring the metal dispersion. Conversely, on
supports with small pores, the Pd particles are less likely to grow
3.5. TPR profiles
All the prepared sample show a clear negative peak centered
around 70 ◦C. Another smaller peak (barely detectable in some
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
5
cases) is centered at around 48 ◦C. Both the peaks slightly shift their
position between 68 and 76 ◦C and 46 and 49 ◦C.
inside the reactor. Because of the reactions, the composition of
the outlet gas (and hence inside the reactor) dropped, causing a
dropped in the liquid composition, which in return impaired the
reaction rates and hence the conversion. In other words, because
of the reactions, the liquid phase goes through a transient from
a higher to a lower H2 (and O2) content, that caused a decreased
in the H2 conversion. Moreover, the decreased composition in the
liquid phase caused a buffer (reservoir) of H2 in the methanol. This
quantity of H2 was available to the reactions but is not included
in the definition of conversion given in the previous section. This
contributed to the very high conversion values observed at low
contact time (<10 min). The effect is only noticeable at low time
on stream (<10 min), as the H2 buffer is soon consumed. Includ-
ing the effect of the H2 reservoir is in principle possible, but it
would require the precise knowledge of the vapor–liquid equilib-
rium of the complex system H2 O2 N2–methanol and numerical
modelling of the reactor, which goes beyond the scopes of the
present work. Nonetheless, the reported conversion values are sig-
nificant to quantitatively evaluate the activity of the synthesized
catalysts.
A 1 wt.% Pd on carbon commercial catalyst was tested as a
benchmark. Apparently, all our synthesized catalysts had a lower
conversion. However, the Pd content of the commercial was much
higher, so that the conversion compare well with our original cata-
lysts when normalized. Furthermore, most our catalysts reached
a much higher peroxide specific concentration and selectivity,
with values up to 4 times larger. The catalysts supported on MEL
supports (Pd2, Pd6, Pd7) showed the highest H2 conversion. In
particular, the catalysts Pd2, Pd6 and Pd7 had the same con-
version throughout the experiment, reaching values around 20%
after 5 h of reaction, whereas the other catalysts showed 9% con-
version. Note that the catalysts supported on MEL supports had
half the Pd content of the others (Table 2), so that their rel-
ative activity is actually much higher than the other catalysts.
These conversion values are apparently rather low. As a com-
parison, Menegazzo et al. [14] found values between 45% and
54% at 5 h reaction, depending on the calcination temperature
of a 1.5 wt.% SiO2/Pd catalyst; the same research group reported
values between 40% and 90% with Pd catalysts (2.5 wt.%) sup-
ported on Zirconia and Ceria[15] (5 h reaction). However, in the
present study the experimental conditions (in particular the ratio
between the H2 flow rate and the catalyst quantity) were cho-
sen so that the conversion never exceeded 30% for high time on
stream to avoid any possible mass transfer limitation. Moreover,
the conversion strongly depends on the total active metal used in
the experiments, so that values are more significant when nor-
malized. Hence, in our study specific conversions in the range
For the samples supported on MEL supports (Fig. 4), other palla-
dium reduction peaks are detected at varying temperatures. Indeed,
Pd2 shows a peak between 300 and 400 ◦C; Pd3 shows a peak start-
ing at about 170 ◦C and overlapping with another very wide one at
400 ◦C, the latter likely due to the support. Pd6 shows a reduction
peak centered at 307 ◦C, while sample 7 exhibits a peak at 93 ◦C. All
the peaks below 80 ◦C are due to the decomposition of Pd ␤-hydride
[26–28]. In fact, PdO can be reduced at temperatures as low as
−15 ◦C, leading to the formation of ␤-hydride that later decomposes
releasing H2. The reduction of PdO occurs in before the analysis
starts, when H2 is dosed at room temperature until the TCD base
line stabilizes. When the TPR run starts, the freshly formed hydride
decomposes, originating the negative peaks at around 70 ◦C.
The peaks between 80 and 350 ◦C are due to Pd(II) non reducible
at room temperature, as also reported in [29–31]. In particular,
according to the cited studies, the high temperature reduction
peaks can be assigned to two dimensional PdO that can be formed
on the surface in case of low Pd content. It is also reported that
this kind of PdO structure is formed in presence of highly dis-
persed metal, due to small crystallite size. This is in agreement with
the high dispersion values detected and the consequent low par-
ticles size of our samples. Interestingly, no Pd reduction peaks are
detected for the samples at higher Pd load (Fig. 5). Only a more
pronounced negative peak at 75 ◦C is detected. Other peaks are
detected at much higher temperature, but they are very small and
they seem to correspond to those obtained in the reduction of the
bare supports. It appears that the loaded Pd is suppressing the
reduction of the support.
3.6. Experimental results
All experiments were carried out in a semibatch reactor, where
the reagents were continuously bubbled through the static liquid
phase. Prior to the experimental campaign, the reactor setup was
tested for eventual external mass transfer limitations. Given that
the gas to liquid mass transfer is independent of the specific cata-
lyst, increasing amounts of a commercial 5% Pd on carbon catalyst
(Degussa) were tested in similar reaction conditions. Since a high
H2 conversion was the aim of these experiments, a catalyst with
a higher Pd content compared to our synthesized materials was
deliberately chosen. A linear conversion of H2 (the limiting reagent)
was observed up to 80% (Supplementary material, Fig. 1), con-
firming that the reaction was not affected by gas to liquid transport
limitations. Hence, all experiments herein reported were carried
out at H2 conversion below 80%. Experimental results are reported
in Fig. 6 in terms of peroxide specific concentration, selectivity and
conversion as a function of time.
In all experiments, hydrogen peroxide concentration leaned
towards a steady state value. At the same time, the selectiv-
ity dropped, meaning that the water concentration (not shown)
steadily increased. This is expected, since hydrogen peroxide is a
reaction intermediate and water is the final product (Scheme 1);
in a semibatch apparatus the accumulation of peroxide leads to
an increase of the hydrogenation and disproportionation rates, so
that its concentration will reach an equilibrium value, at the same
time dropping the selectivity. This evolution is particularly evident
with the catalysts showing the highest H2 conversion. Interest-
ingly, the conversion decreased with the time on stream (Fig. 6C),
which is in apparent contradiction with the increasing peroxide
(and water) concentrations. This effect is due to the solubility of
H2 in the liquid phase. As soon as the catalyst was introduced, H2
started to react, creating a sink of H2 that was refilled by the inlet
gas flow. The gas phase behaved like a continuous stream reactor,
where the composition at the outlet is equal to the composition
ꢂ
ꢅ
ꢃ
ꢄ
of 46–100
molH
×
molH
2
−1 × mmol
consumed
supplied
−1
Pd
(
)
2
were obtained at 5 h time on stream, whereas the in the men-
tioned studies [14,15] the specific conversion was between 24
ꢂ
ꢅ
ꢃ
ꢄ
consumed
supplied
−1
Pd
and 40 molH
×
molH
−1 × mmol
. The
(
)
2
2
same catalysts Pd2, Pd3, Pd6 and Pd7 also showed the highest
H2O2 specific concentration (Fig. 6A), with values of 3.5 and 8.5
ꢀ
ꢁ
−1
MH
mmolPd
for the groups Pd2, Pd3 and Pd6, Pd7, respec-
O
tively. 2Lower values are normally reported in the literature. For
2
instance, Abate et al. [12] obtained peroxide specific concentra-
ꢀ
ꢁ
−1
tions in the range 2.1–5.2 MH
mmolPd
at 4 h contact time
O
in a 6.5 bar semibatch apparatu2s;2Menegazzo et al. [14,15] obtained
ꢀ
ꢁ
−1
values between 0.2 and 4.2 MH
mmolPd
in atmospheric
O
conditions at 5 h contact time, dep2ending on the catalyst. However,
a higher peroxide concentration value is expected in our system.
Indeed, due to the higher pressure (50 bar), a higher reagent con-
centration was reached, allowing achieving a higher production
2
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
6
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
Pd2
Sp2
Pd3
Sp3
Pd6
Sp6
Pd7
Sp7
0
100
200
300
400
500
600
T (°C)
Fig. 4. TPR curves of the fresh samples and of the supports for the 0.3% Pd samples.
Pd4
Sp4
Pd5
Sp5
Pd8
Sp8
0
100
200
300
400
500
600
T (°C)
Fig. 5. TPR curves of the fresh samples and of the supports for the 0.6% Pd samples.
rate. Interestingly, the catalysts with the highest specific perox-
ide concentration showed also the highest selectivity (Fig. 6B).
Nonetheless, the selectivity was quite low compared to the val-
ues obtained in the literature in similar apparatus [12,14,15,32]. It
is worth noticing that in all literature studies promoters and stabili-
zers (such as bromine and/or acids) were introduced in the reaction
slurry to enhance selectivity and productivity. No such chemicals
were used in this study, in order to isolate the effect of the catalyst
properties on the selectivity.
In the following, the experimental results are discussed in rela-
tion to the catalyst’ properties. Interestingly, the sample with the
lowest Pd content (Table 2) showed the highest selectivity and
specific peroxide concentrations. More specifically, this relates to
the interaction between the active metal and the support, where
smaller pore diameters allowed the formation of larger Pd parti-
cles. In order to quantify this effect, in Fig. 7 the selectivity and the
specific peroxide concentration are reported as a function of the
Pd particle size. Since the measurements varied with time (Fig. 6),
comparisons were carried out at 1 h and 5 h time on stream.
Selectivity and peroxide concentration evolutions were inde-
pendent on the contact time, suggesting that the choice of the
reaction time for the comparison can be arbitrary. The larger the
Pd particle, the higher the concentration and the selectivity toward
the desired product obtained. This suggests that larger particles of
the active metal suppressed the formation of water, favoring the
production of the peroxide. It has been proposed [14,16] that O2
adsorbs associatively on less energetic Pd particles, thus favoring
H2O2 formation. On the contrary, highly unsaturated sites, such as
Pd atoms at a corner or an edge cleave the O O bond, leading to the
non-selective production of H2O. The morphological properties of
the catalysts presented in this work support this hypothesis, where
large and thus less energetic particles favored the direct synthesis
of H2O2. Fig. 7 highlights the superiority of the samples supported
on solid solutions of ZrO2 with its promoters. Indeed, the samples
that perform better are those that presented a ZrxM1–xO2 structure
(M = La, Y, Ce). The Pd2 sample seems to be an exception, since it
exhibits activity and selectivity that are generally more in line with
the performance of the samples that did not form a solid solution.
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
7
25
12
9
A
B
20
15
10
5
6
3
0
0
0
100
200
Time (min)
300
0
100
200
Time (min)
300
C
60
50
40
30
20
10
0
0
100
200
300
Time (min)
Fig. 6. Experimental results in terms of specific H2O2 concentration (A), selectivity (B) and conversion (C): ♦, Pd2; *, Pd3; x, Pd4; ꢀ, Pd5; ꢀ, Pd6; + Pd7; ꢁ, Pd8; − Pd/C
commercial.
However, its H2O2 productivity after 1 h is perfectly aligned with
the general trend. The reason for the lower performance of Pd2
might be the presence of residual amounts of ZrO2 in its support
structure, as evidenced by the XRD pattern. This fact can also be cor-
related with the redox behavior of the samples. Indeed, Pd2 shows
a much smaller Pd reduction peak than any other sample that forms
a solid ZrxM1–xO2 solution (see Fig. 4). Conversely, Pd2 shows a TPR
profile that is closer to that of the samples supported on mechani-
cal mixtures of ceria and zirconia, which in turn show much lower
activity and selectivity. It is well know that metallic catalysts sup-
ported on homogenous solid solution of ZrO2 with promoters such
Ce, Y and La, become more reducible than when supported on pure
ZrO2 [21,22]. The redox properties of the metal/support system
are very important in determining the final performance of the
catalyst. In particular, more reducible systems can achieve higher
selectivity and production of H2O2 [9,17]. Furthermore, the pro-
moters also stabilize the physical properties, preserving a higher
surface area upon exposure to high temperature. Observing the
supports features reported in Table 2, it is clear that Ce did not
have an as good stabilizing effect for the Pd2 sample as for the Pd3
one. Indeed, the Pd2 surface area is only 52 vs 93 m2/g for Pd3.
Accordingly, the average pore diameter and the total pore volume
are also different. Looking closer, one can note that the better per-
forming samples (Pd3, Pd6 and Pd7) present a very low average
pore diameter (5–7 nm). On the contrary, the remaining samples,
whose performance are lower, present an average pore diameter
in the range of 16–30 nm. The same observation stands for the
total pore volume, with the better performing catalysts showing
a value that is about half or less than the remaining samples. This is
reflected on the size of the Pd particles, with the exception of Pd2.
We conclude that both the support physical features and the
global redox properties strongly affect the behavior of the cata-
lyst, with bigger metal particles and higher reducibility favoring
the production of H2O2.
25
15
1h reacꢀon
5h reacꢀon
20
10
5
15
10
5
0
0
4.0
12
10
8
3.0
2.0
1.0
6
4
2
0.0
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Parꢀcle size(nm)
Parꢀcle size(nm)
Fig. 7. Selectivity (top) and specific peroxide concentration (bottom) after 1 h (left) and 5 h (right) time on stream as a function of the Pd particle size of the catalyst.
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033

G Model
CATTOD-9416; No. of Pages8
ARTICLE IN PRESS
8
A. Bernardini et al. / Catalysis Today xxx (2015) xxx–xxx
[5] N. Gemo, P. Biasi, P. Canu, T.O. Salmi, Mass transfer and kinetics of H2O2 direct
4. Conclusions
synthesis in a batch slurry reactor, Chem. Eng. J. 207–208 (2012) 539–551.
[6] I. Huerta, J. García-Serna, M.J. Cocero, Hydrogenation and decomposition
kinetic study of H2O2 over Pd/C catalyst in an aqueous medium at high CO2
pressure, J. Supercrit. Fluids 74 (2013) 80–88.
Novel catalysts for the direct synthesis of hydrogen peroxide
were prepared supporting Pd on mixed oxides (ZrO2 and CeO2) and
rare earth elements (Y2O3 and La2O3). These materials were chosen
because of their capability of increasing both the metal dispersion
and the O2 mobility (which favors the active metal activity). Fur-
thermore, their properties (surface area, pore volume, particle size,
homogeneity, etc) are easily tunable changing the preparation pro-
cedures. XRD analysis showed the formation of ZrxM1–xO2 (M = La,
Y, Ce) solid solutions for the supports supplied by MEL Chemi-
cals, but the persistence of separate ceria and zirconia phases for
the supports prepared by mixing and calcining different amounts
of ceria and zirconia. All the samples achieved very high disper-
sion (50–99%) resulting from a very small average particle size
(1–2.6 nm). Through TPR it was shown that the supports features
affect the reducibility of the metal–support system and this result
in rather different catalytic performance. More easily reducible
samples exhibited higher performance. Hydrogen peroxide direct
synthesis was performed in a high pressure semibatch reactor in the
absence of promoters and stabilizers (acid and hides). H2 conver-
sion and peroxide concentration obtained were higher than usually
reported in the literature using similar apparatus. Specific conver-
sion and selectivity up to 4 times higher compared to a commercial
Pd/C catalyst were achieved, and up to double the values reported
in literature in works where enhancers were used. The synthe-
sis reaction was run in a broad range of product concentration,
with the peroxide leaning towards a steady state value due to the
subsequent reactions. Larger Pd particles were obtained upon sup-
ports with smaller pore diameters. In turns, the larger Pd particles
were more selective toward the peroxide, suggesting that highly
unsaturated sites, such as Pd atoms at a corner or an edge largely
present on small particles, cleave the O O bond, leading to water.
Furthermore, supports with higher reducibility favored the pro-
duction of H2O2, probably due to an easier reduction of the active
metal, essential to achieve high selectivity. These observations lead
to conclude that highly reducible supports giving larger metal par-
ticles are most desirable in the preparation of catalysts for the H2O2
direct synthesis.
[7] Y. Voloshin, A. Lawal, Kinetics of hydrogen peroxide reduction by hydrogen in
a microreactor, Appl. Catal., A: Gen. 353 (2009) 9–16.
[8] E.N. Ntainjua, J.K. Edwards, A.F. Carley, J.A. Lopez-Sanchez, J.A. Moulijn, A.A.
Herzing, C.J. Kiely, G.J. Hutchings, The role of the support in achieving high
selectivity in the direct formation of hydrogen peroxide, Green Chem. 10 (2008)
1162–1169.
[9] C. Burato, S. Campestrini, Y. Han, P. Canton, P. Centomo, P. Canu, B. Corain,
Chemoselective and re-usable heterogeneous catalysts for the direct synthesis
of hydrogen peroxide in the liquid phase under non-explosive conditions and
in the absence of chemoselectivity enhancers, Appl. Catal., A: Gen. 358 (2009)
224–231.
[10] G. Blanco-Brieva, E. Cano-Serrano, J. Campos-Martin, J.L.G. Fierro, Direct syn-
thesis of hydrogen peroxide solution with palladium-loaded sulfonic acid
polystyrene resins, Chem. Commun. 0 (2004) 1184–1185.
[11] S. Abate, S. Perathoner, G. Centi, Performances of Pd nanoparticles on different
supports in the direct synthesis of H2O2 in CO2-expanded methanol, Top. Catal.
54 (2011) 718–728.
[12] S. Abate, P. Lanzafame, S. Perathoner, G. Centi, SBA-15 as a support for palla-
dium in the direct synthesis of H2O2 from H2 and O2, Catal. Today 169 (2011)
167–174.
[13] E. Ghedini, F. Menegazzo, M. Signoretto, M. Manzoli, F. Pinna, G. Strukul, Meso-
porous silica as supports for Pd-catalyzed H2O2 direct synthesis: effect of the
textural properties of the support on the activity and selectivity, J. Catal. 273
(2010) 266–273.
[14] F. Menegazzo, M. Signoretto, G. Frison, F. Pinna, G. Strukul, M. Manzoli, F. Boc-
cuzzi, When high metal dispersion has a detrimental effect: hydrogen peroxide
direct synthesis under very mild and nonexplosive conditions catalyzed by Pd
supported on silica, J. Catal. 290 (2012) 143–150.
[15] F. Menegazzo, P. Burti, M. Signoretto, M. Manzoli, S. Vankova, F. Boccuzzi, F.
Pinna, G. Strukul, Effect of the addition of Au in zirconia and ceria supported
Pd catalysts for the direct synthesis of hydrogen peroxide, J. Catal. 257 (2008)
369–381.
[16] S. Abate, G. Centi, S. Melada, S. Perathoner, F. Pinna, G. Strukul, Preparation,
performances and reaction mechanism for the synthesis of H2O2 from H2 and
O2 based on palladium membranes, Catal. Today 104 (2005) 323–328.
[17] Q. Liu, K. Gath, J. Bauer, R. Schaak, J. Lunsford, The active phase in the direct
synthesis of H2O2 from H2 and O2 over Pd/SiO2 catalyst in a H2SO4/ethanol
system, Catal. Lett. 132 (2009) 342–348.
[18] S. Eriksson, S. Rojas, M. Boutonnet, J.L.G. Fierro, Effect of Ce-doping on Rh/ZrO2
catalysts for partial oxidation of methane, Appl. Catal., A: Gen. 326 (2007) 8–16.
[19] S. Boullosa-Eiras, T. Zhao, D. Chen, A. Holmen, Effect of the prepara-
tion methods and alumina nanoparticles on the catalytic performance of
Rh/ZrxCe1–xO2 Al2O3 in methane partial oxidation, Catal. Today 171 (2011)
104–115.
[20] M. Pudukudy, Z. Yaakob, Catalytic aspects of ceria zirconia solid solution:
Part-II: An overview on recent developments in the heterogeneous catalytic
applications of metal loaded ceria–zirconia solid solution, Pharma Chem. 6
(2014) 224.
[21] M. Pudukudy, Z. Yaakob, Catalytic aspects of ceria–zirconia solid solution: Part-
I: An update in the synthesis, properties and chemical reactions of ceria zirconia
solid solution, Pharma Chem. 6 (2014) 188.
[22] Y. Han, J. Zhu, Surface science studies on the zirconia-based model catalysts,
Top. Catal. 56 (2013) 1525–1541.
[23] R. Pérez-Hernández, L.C. Longoria, J. Palacios, M.M. Aguila, V. Rodríguez, Oxida-
tive steam reforming of methanol for hydrogen production over Cu/CeO2–ZrO2
catalysts, Energy Mater. 3 (2008) 152–157.
[24] R. Pérez-Hernández, A. Gómez-Cortés, J. Arenas-Alatorre, S. Rojas, R. Mariscal,
J.L.G. Fierro, G. Díaz, SCR of NO by CH4 on Pt/ZrO2–TiO2 sol–gel catalysts, Catal.
Today 107–108 (2005) 149–156.
Acknowledgments
This work is part of the activities Process Chemistry Centre (PCC)
financed by the Åbo Akademi University (ÅA). Dr. Pierdomenico
Biasi gratefully acknowledges the Kempe Foundations (Kempe Stif-
telserna) for financial support. In Sweden, also the Bio4Energy
program is acknowledged.
Appendix A. Supplementary data
[25] R. Pérez-Hernández, F. Aguilar, A. Gómez-Cortés, G. Díaz, NO reduction with
CH4 or CO on Pt/ZrO2–CeO2 catalysts, Catal. Today 107–108 (2005) 175–180.
[26] Z. Paál, P.G. Menon, Hydrogen Effects in Catalysis: Fundamental and Practical
Applications, Marcel Dekker, New York, U.S.A., 1988.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.cattod.
2014.12.033.
[27] Z. Karpin´ ski, Catalysis by supported, unsupported, and electron-deficient pal-
ladium, Adv. Catal. 37 (1990) 45–100.
[28] P.C. Aben, Palladium areas in supported catalysts: determination of palladium
surface areas in supported catalysts by means of hydrogen chemisorption, J.
Catal. 10 (1968) 224–229.
References
[29] S. Melada, R. Rioda, F. Menegazzo, F. Pinna, G. Strukul, Direct synthesis of hydro-
gen peroxide on zirconia-supported catalysts under mild conditions, J. Catal.
239 (2006) 422–430.
[30] N.S. Babu, N. Lingaiah, R. Gopinath, P.S. Sankar Reddy, P.S. Sai Prasad,
Characterization and reactivity of alumina-supported Pd catalysts for the
room-temperature hydrodechlorination of chlorobenzene, J. Phys. Chem. C 111
(2007) 6447–6453.
[31] A. Barrera, M. Viniegra, S. Fuentes, G. Díaz, The role of lanthana loading on
the catalytic properties of Pd/Al2O3–La2O3 in the NO reduction with H2, Appl.
Catal., B: Environ. 56 (2005) 279–288.
[1] C. Samanta, Direct synthesis of hydrogen peroxide from hydrogen and oxygen:
an overview of recent developments in the process, Appl. Catal., A: Gen. 350
(2008) 133–149.
[2] J. Garcia-Serna, T. Moreno, P. Biasi, M.J. Cocero, J. Mikkola, T.O. Salmi, Engineer-
ing in direct synthesis of hydrogen peroxide: targets, reactors and guidelines
for operational conditions, Green Chem. 16 (2014) 2320–2343.
[3] G. Centi, S. Perathoner, S. Abate, Direct synthesis of hydrogen peroxide: recent
advances, in: Modern Heterogeneous Oxidation Catalysis, Wiley-VCH Verlag
GmbH & Co. KGaA, Weinheim, Germany, 2009, pp. 253–287.
[4] J.M. Campos-Martin, G. Blanco-Brieva, J.L.G. Fierro, Hydrogen peroxide synthe-
sis: an outlook beyond the anthraquinone process, Angew. Chem. Int. Ed. 45
(2006) 6962–6984.
[32] T. Moreno, J. García-Serna, P. Plucinski, M.J. Sánchez-Montero, M.J. Cocero,
Direct synthesis of H2O2 in methanol at low pressures over Pd/C catalyst:
semi-continuous process, Appl. Catal., A: Gen. 386 (2010) 28–33.
Please cite this article in press as: A. Bernardini, et al., Direct synthesis of H2O2 over Pd supported on rare earths promoted zirconia,
Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2014.12.033