Propylene epoxidation with in situ generated H2O2 in supercritical conditions

Article abstract of DOI:10.1016/j.cattod.2013.10.004

Several bi-functional materials based on palladium nanoparticles supported onto nanocrystalline titanium silicalite zeolite (Pd@TS-1) were prepared and used as active and reusable catalysts to direct PO production from hydrogen, oxygen and propylene throu

Full text of DOI:10.1016/j.cattod.2013.10.004

Catalysis Today 227 (2014) 87–95
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Propylene epoxidation with in situ generated H O2
2
in supercritical conditions
∗
Alejandro Prieto, Miguel Palomino, Urbano Díaz , Avelino Corma
Instituto de Tecnología Química (UPV-CSIC) Universidad Politécnica de Valencia, Consejo Superior de Investigaciones Científicas, Avenida de los Naranjos
s/n., 46022 Valencia, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 26 June 2013
Received in revised form
Several bi-functional materials based on palladium nanoparticles supported onto nanocrystalline tita-
nium silicalite zeolite (Pd@TS-1) were prepared and used as active and reusable catalysts to direct PO
production from hydrogen, oxygen and propylene through one-pot two-step combined process. This
type of catalysts was able to carry out the consecutive process where palladium nanoparticles catalyzed
the formation of H2O2 that was used as intermediate reagent for propylene epoxidation catalyzed by
TS-1 nanocrystals, used as supporting active matrix. Influence of supercritical CO2 conditions in batch
reactions was evaluated together with the associated effect due to the presence of platinum promoters
2
3 September 2013
Accepted 1 October 2013
Available online 26 October 2013
Keywords:
Propylene oxide
Bi-functional catalysts
Supercritical CO2 conditions
Palladium nanoparticles
Titanium silicalite
(Pd(Pt)@TS-1). Other reaction parameters were also considered, such as palladium loading, use of acid-
ity inhibitors and modification of H2/O2/C3 ratio. Reusability studies showed the high stability of the
catalysts in consecutive catalytic cycles, their regeneration being possible.
© 2013 Elsevier B.V. All rights reserved.
1
. Introduction
Propylene oxide (PO) is an important monomer for production
difficult this reaction [7,8]. In this sense, it is reported that eth-
ylene epoxidations can be carried out employing silver catalysts
in the only presence of O2 [9]. However, when these catalysts have
been employed in the case of propylene, results have certainly been
poor [10]. So, it would be decisive to avoid the drawbacks associ-
ated to chlorohydrins and Halcon processes, being important the
development of safe and low contaminating propylene epoxidation
routes.
Regarding safety and contamination, the use of catalysts
becomes essential in the industry to prevent these problems.
Homogeneous catalysts have been widely used in the industry, but
their disadvantages, including their difficult and expensive recov-
ery, poor thermal stability, limited applicability and the cost of
catalyst losses makes desirable the use of another kind of cata-
lyst. Although heterogeneous catalysts are sometimes less selective
than homogeneous have also a wide applicability, and besides,
important advantages, including easy and cheap catalyst separa-
tion and low costs associated to catalyst losses, factors which are
enhanced by the catalyst reusability [11]. For these reasons, the
preparation of heterogeneous catalysts which would catalyze the
propylene epoxidation with high performances and selectivities
would be highly desirable [12].
of polyurethane foams, resins and propylene glycol [1]. The PO pro-
duction is about 7.5 million tons per year, which comes mainly from
the chlorohydrins and organic hydroperoxide (Halcon Method)
processes. The chlorohydrins method was originally used for the
production of ethylene oxide, but in the 1960s was replaced by a
direct oxidation method and started to be used for the PO synthesis.
In this method, propylene and chlorine react leading to a mixture
containing different types of chlorohydrins, which react with an
alkali to produce PO with high selectivity. However, this method-
ology has an important drawback, because chlorine and Ca(OH)2
are necessary as sources, and CaCl2 is produced, together with a
large amount of waste water. On the other hand, in the Halcon
process, ethylbenzene hydroperoxide or tert-butyl hydroperoxide
react with propylene to yield PO. But, in this process, methylbenzyl
alcohol or tert-butanol are obtained, which are dehydrated to form
styrene monomer and isobutylene, i.e., two compounds severely
affected by market trends [2–5].
Taking this into account, the most convenient solution would
be the direct oxidation with O2 [6], although the high reactivity of
the allylic C H bonds in the propylene precursor makes extremely
Besides the chlorohydrins and Halcon processes, the production
of PO from propylene has also been carried out through two con-
secutive steps (Scheme 1): (i) H O2 formation from hydrogen and
2
∗ Corresponding author. Tel.: +34 963877811; fax: +34 963877809.
E-mail address: udiaz@itq.upv.es (U. Díaz).
oxygen and (ii) generation of PO from propylene and previously
formed H2O2. Even, different industrial prototypes have recently
0
920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.10.004

8
8
A. Prieto et al. / Catalysis Today 227 (2014) 87–95
Pd
C3H6 conversions with medium selectivities, claiming addition-
(
(
i)
H + O
H O
2
2
2
2
ally the use of iridium and gold species as palladium promoters,
although the better results were obtained when platinum nanopar-
ticles were incorporated as promoters. Further investigations
showed that the previous impregnation of the catalyst with the
corresponding salts and the use of alcohols or ketones as solvent
for the catalyst impregnation with metallic species improved the
PO selectivity. Moreover, the autoreduction of Pd@TS-1 catalysts
under pure nitrogen reduced the olefins hydrogenation with the
reaction conditions, reducing the by-products formation. Similar
Pd/Pt@TS-1 bi-functional catalysts were also employed by Jenzer
et al. [32] in a tubular flow reactor, achieving also medium PO
yields.
TS-1
ii)
H O +
+ H O
2
2
2
O
Scheme 1. Individual reaction steps to produce propylene oxide (PO).
been reported based on the combination of two connected reac-
tion systems where the intermediate H O2 is circulated between
the reactor vessels, such as the HPPO two-step process based on
the direct synthesis of hydrogen peroxide [13–15]. Into this com-
bined reaction route, several types or polymeric resins [16] and/or
different transition metals [17,18], such as Pd, Pt and Au, supported
2
Nevertheless, the low yields obtained for the epoxidation of
onto inorganic oxides have been able to directly produce H O [19].
propylene with in situ generated H O2 to direct PO formation,
2
2
2
Additionally, titanium-silicalite zeolite (TS-1), firstly synthesized
by Taramaso et al. [20], has been an excellent solid catalyst for the
propylene epoxidation. Furthermore, in the last years, optimized
synthesis methods of nanocrystalline TS-1 zeolites, from different
synthesis routes, have allowed the development of highly active
and recyclable titanium-catalysts for PO production, using ex situ
generated H O [21].
The next step to improve the PO production would be to
combine the two individual reaction steps in only one process,
developing an efficient one-pot two-step system based on the
together with the associated difficulty to work with explosive
mixtures of hydrogen and oxygen as reaction reagents, make real-
istically unfeasible the industrial scaled of this catalytic process,
being indispensable to find possible alternatives to this one-pot
two-step combined reaction in presence of solid bi-functional cat-
alysts. Nowadays, the use of supercritical CO2 (scCO ) as a solvent
2
in different reactions is receiving considerable attention due to its
numerous advantages, including inactivity, low cost, no transfer
limitations and high solubility [33]. In the case of direct PO produc-
2
2
tion, the presence of CO could provide a reaction medium capable
2
in situ generation of H O2 and the use of bi-functional solid cat-
to solubilize large quantities of gases, being possible to work in a
non-flammable and safe environment in which to merge hydro-
gen and oxygen miscible with propylene and PO [34]. Although the
2
alysts. Nowadays, the catalytic processes which combine two or
more consecutive reactions are being very attractive from sus-
tainability point of view due to the reduction in the number of
sub-processes associated to isolation and separation tasks of inter-
mediate products [22]. Into this approach, the use of active and
reusable bi-functional catalysts which contain metallic nanopar-
ticles (NPs) embedded into inorganic oxide type-matrixes are
currently being profusely considered for different catalytic pro-
cesses [23–25].
◦
supercritical CO2 state is reached at 73.8 bar and 31 C, the proper-
ties of scCO in the region near the critical point (Tr = T/Tc = 1.0–1.1
2
and Pr = P/Pc = 1–2) are specially interesting. In this region, the sin-
gle phase system has a density which enhances the dissolution
power, the diffusivity of solutes is higher than in liquids and the
viscosity is lower, enhancing mass transfer. These three proper-
ties make the reactants easier to reach the catalyst’s active site
[35]
With the discovery by Hayashi et al. [19] of bi-functional mate-
rials formed by Au nanoparticles (NPs) supported onto titania
matrix and their catalytic performances, a new possibility for the
direct epoxidation of propylene trough in situ generated H O was
In this way, Chen et al. [36] described the use of scCO2 for the
epoxidation of propylene with external H O2 in presence of large
2
grains of TS-1 as catalyst. On the other hand, the use of scCO2 has
also been reported by Chen et al. [37a] both for the synthesis of
H O and direct PO production from hydrogen and oxygen with
2
2
opened. Specifically, this type of gold-catalysts were highly selec-
tive in the direct vapor-phase oxidation of propylene to PO in
the presence of oxygen and hydrogen, being established that Au
nanoparticles with sizes between 2 nm and 5 nm were the most
active for the PO production when titania was employed as the
2
2
promising results. In their research, it was shown that the addition
of small volumes of methanol and water as co-solvents and the use
of acidity inhibitors under scCO conditions improved the PO selec-
2
support [26,27]. However, Au@TiO catalysts suffered deactivation
tivity and the yield of the reaction, using bi-functional Pd/Pt@TS-1
catalysts. Taking account the attractive results obtained, further
studies would be necessary, using supercritical CO2 as solvent to
analyze its influence on the PO formation and the effect on reactiv-
ity of the heterogeneous catalyst that takes part in the reaction.
In the current study, several bi-functional materials based on
Pd nanoparticles supported onto TS-1 nanocrystals have been
prepared (Pd@TS-1) through different synthesis and deposition
methods, followed by the evaluation of their catalytic activity for
the PO production from hydrogen and oxygen. The Pd@TS-1 cat-
alysts were able to carry out the one-pot two-step consecutive
process where metallic nanoparticles catalyze the formation of
H O that is used as intermediate for propylene epoxidation cat-
alyzed by TS-1 nanocrystals. The effect of supercritical CO2 as
solvent, during the reaction process, together with the use of addi-
tional co-solvents was considered, seeking to enhance and optimize
the performances and selectivities to PO direct synthesis. The
physico-chemical treatment of the bi-functional catalysts through
the incorporation of Pt promoters and surface acidity inhibitors
were also considered. The Pd(Pt)@TS-1 catalysts showed high sta-
bility, their regeneration and reuse being possible after consecutive
catalytic cycles, partly recovering the initial activity.
2
phenomenon, its stability being too low in the profusely studied
gaseous phase catalytic conditions [8]. With the purpose of increas-
ing the reactivity and stability of bi-functional gold-materials, Yap
et al. [28] concluded that Au@TS-1 catalysts containing Au nano-
clusters with sizes between 1 nm and 2 nm were the most active
for the epoxidation of propylene when TS-1 is used as a supporting
matrix, being its stability also enhanced. Optimization of gold-
catalysts was recently achieved by Huang et al. [29] using aqueous
alkaline solutions to increase surface defects in TS-1 supports. In
this case, the majority of deposited gold nanoclusters exhibited
sizes below 2 nm, being homogenously distributed onto the inor-
ganic matrix with reduced metallic agglomeration. The resulting
bi-functional Au@TS-1 catalysts improved the performances for PO
production using a fixed-bed reactor. Furthermore, due to the broad
range of explosion limits of hydrogen and oxygen mixtures, which
could be explosive with hydrogen concentrations between 3.8 and
2
2
9
5.5 mol% [30], the use of an inert gas becomes essential in order
to dilute the mixture at the same time that equal molar amounts of
hydrogen and oxygen are used.
In this sense, Laufer et al. [31] reported the use of solid bi-
functional Pd@TS-1 catalysts in batch conditions, achieving higher

A. Prieto et al. / Catalysis Today 227 (2014) 87–95
89
2
2
2
. Experimental
[41] was obtained from the amount of N2 adsorbed at a relative
pressure of ∼0.99. External surface area and micropore volume
were estimated with the t-plot method in the t range from 3.5
to 5. The pore diameter and the pore size distribution were
obtained following the Barret–Joyner–Halenda (BJH) method [42]
on the adsorption branch of the isotherms. Scanning and trans-
mission electron microscopies (SEM and TEM) characterization
were carried out in a JEOL6300 and Philips CM10 (100 kV) micro-
scopes, respectively. Before TEM observation, the samples were
prepared by suspending the solid in water, ultrasonicating for
15 min and placing one drop on a carbon-coated copper grid (300
mesh).
.1. Catalyst preparation
.1.1. Synthesis of nanocrystalline TS-1
The preparation of TS-1 nanocrystals, used as inorganic sup-
port, was carried out by combination of micellar and hydrothermal
synthesis methodologies [38].
Specifically, 2.0 g of Tween 20 (s.d. FINE CHEM.) were dissolved
in 32 g of distilled water. This surfactant solution was added to
1
9.2 g of tetrapropylammonium hydroxide (TPA-OH, 32% aque-
ous solution) under mild stirring, resulting in the formation of
a clear transparent solution. To the above micellar solution, 36 g
of tetraethyl orthosilicate (TEOS, Aldrich) were added in a drop-
wise fashion under vigorous stirring. The stirring was continued
for 1 h more. To this clear solution, 1.808 g of tetra n-butyl titanate
2.4. Catalytic tests
(
TNBT, Aldrich) in 9.12 g of isopropyl alcohol (IPA, s.d. FINE CHEM.)
The experiments for the synthesis of PO from propylene,
were added dropwise under vigorous stirring that was contin-
ued for another 1 h. The resulting mixture still remained clear.
The mixture was then crystallized at 160 C for 18 h under auto-
employing external H O2 as oxidant, were performed in a 3 mL
2
round bottom glass reactor in which the stirring was driven
by a Teflon-coated magnetic stirrer. Known amounts of catalyst
(2.5 mg), methanol (1.2 g), and propylene (9 mmol) were added,
followed by the addition of the oxidant, 35%wt H O in water
◦
geneous pressure. The product was recovered by centrifugation,
◦
washed with distilled water and dried (100 C, 12 h). The starting
2
2
◦
mixture has a molar composition as follows: 0.03 TiO :SiO :0.12
(0.4 mmol). Then, the mixture was heated at 60 C, and the reac-
tion was carried out for 5 h. Different samples were taken during
the reaction, and were analyzed with a gas chromatographer Fisons
GC8000 provided with a flame ionization detector (FID) and a 60 m
length 0.25 mm ID, 0.25 ␮m df. Rtx-Wax column.
2
2
TPAOH:0.009 Tween 20:0.88 IPA:14.45 H O. The calcination was
2
◦
carried out in a muffle furnace at 550 C for 3 h with a heating rate
◦
of 1.5 C/min.
2.2. Metallic nanoparticles deposition
The experiments for the direct synthesis of PO with in situ
generated H O2 were carried out in a 15 mL stainless steel auto-
2
(
Pd + Pt)@TS-1 catalysts were prepared following the method-
clave coated with a Teflon beaker and with a relief valve, for
safety. The stirring was driven by a Teflon-coated magnetic stirrer.
Known amounts of catalyst (15 mg), acidity inhibitor (ammonium
acetate, 0.01 g) and co-solvent (0.2 g of MeOH or a 50/50 mixture
ology described by Hancu et al. [39] For instance, to prepare
bifunctional catalyst containing 0.2% Pd and 0.02% Pt, 1.0 g of pre-
viously prepared and calcined TS-1 was placed in a glass flask and
5
mL of deionized water were added, followed by heating the solu-
of MeOH/H O) were added to the reactor, followed by the addition
2
◦
tion up to 80 C under continuous stirring. To this mixture, 36 ␮L
of propylene (2 mmol) and CO . Known amounts of oxygen and
2
of [Pd(NH ) ](NO ) (Strem Chemicals) 15%wt aqueous solution
finally hydrogen were then added to the reactor, which was then
heated up until the desired temperature, the critical point (73.8 bar
and 31 C) being surpassed. The reaction experiments were carried
3
4
3 2
and 0.34 mg of [Pt(NH ) ]Cl (Aldrich) were added. These amounts
3
4
2
◦
were changed depending on the desired catalyst. The mixture was
stirred for 24 h, the solid obtained was filtered and washed with
out for 5 h, unless otherwise stated. After that time, the reactor
was cooled down using a mixture of acetone and ice, and the
pressure was slowly released by venting, recovering the gaseous
mixture in an inert gas sampling bag. 3-pentanone was used for
recovering any product that could be in the walls of the reactor,
avoiding the employment of acetone because could be a possible
by-product.
The amount of formed products, i.e., propylene oxide, acetone,
propionaldehyde, acrolein, isopropanol, 1-methoxy-2-propanol
(MP1), 2-methoxy-1-propanol (MP2), propylene glycol (PG) and
propylene carbonate were then analyzed using a Fisons Gas chro-
matograph GC8000 provided with a FID detector and a 60 m length
0.25 mm ID, 0.25 ␮m df. Rtx-Wax column. The amounts of propane
and un-reacted propylene, oxygen and hydrogen were analyzed
using a Bruker 450-GC which contains two different independent
channels. The first one is provided with a thermal conductivity
detector (TCD) and 3 different columns: Hayesep N (0.5 m length),
Hayesep Q (1.5 m length) and a molsieve 13× (1.2 m length),
employing argon as carrier. The second one is provided with two
different flame ionization detectors (FID) and 3 different columns:
a capillary column CP-Wax (1 m length and 0.32 mm ID), a CP-
Porabond Q (25 m length and 0.32 mm ID) and other CP-Wax (5 m
length and 0.32 mm ID).
◦
deionized water and it was finally dried at 60 C overnight. Then, it
◦
was calcined at 150 C for 4 h and reduced by H at room tempera-
2
ture for 4 h. In the case of Pd@TS-1 catalysts, the procedure was the
same but Pt complex was not added.
The general acronym of the synthesized bi-functional catalysts
was Pdxx@TS-1, being xx indicative the palladium amount, per-
centage in weight, supported onto TS-1 support, i.e., 02, 04, 05 and
1
0 would correspond to 0.2%wt, 0.4%wt, 0.5%wt and 1.0%wt of Pd,
respectively. When platinum promoter was also supported onto
inorganic matrix, Pdxx(Pt)@TS-1 was the acronym used during the
discussion. In these samples, the content of platinum corresponded
approximately to 10%wt of palladium content.
2.3. Catalysts characterization
XRD analysis was carried out with a PANalytical X’Pert PRO
diffractometer in the Bragg–Brentano geometry using Cu K␣ radi-
ation. C, N and H contents were determined with a Carlo Erba 1106
elemental analyzer, while Pd and Pt contents were obtained by
means of atomic absorption spectroscopy (Spectra AA 10 Plus, Var-
ian). A Cary 5 spectrometer equipped with a diffuse reflectance
accessory was used for UV measurements. Volumetric analyses
were performed by nitrogen adsorption isotherms at 77 K with
a Micromeritics ASAP2010. Before the measurements, the sam-
Due the extensive list of possible by-products that have been
reported in literature for this kind of reaction, all the products have
been calibrated using mixtures with known composition. All the
products in the liquid phase have been identified by GC–MS and by
GC with commercial samples.
◦
ples were outgassed for 12 h at 100 C. The BET specific surface
area [40] was calculated from the nitrogen adsorption data in a
relative pressure range from 0.04 to 0.2. The total pore volume

9
0
A. Prieto et al. / Catalysis Today 227 (2014) 87–95
Table 1
Chemical analyses and textural properties of Pd(Pt)@TS-1 samples.
Samplesa
Ti (%wt)
Pd (%wt)
Pt (%wt)
SBET(m2
g
−1
)
SMICROb (m2
g
−1
)
VTOT (cm3
g
−1
)
VMICRO (cm3
−1
g )
TS-1
Pd02
Pd02(Pt)
Pd04(Pt)
Pd05(Pt)
Pd10(Pt)
1.99
2.00
1.85
1.81
2.00
1.98
–
–
–
0.03
0.10
0.09
0.13
416
391
391
397
396
387
271
236
242
251
244
240
0.298
0.315
0.319
0.379
0.376
0.319
0.132
0.116
0.120
0.123
0.120
0.118
0.23
0.18
0.40
0.50
1.03
a
Number after Pd indicates the theoretical Pd loading (0.2%wt, 0.4%wt, 0.5%wt and 1.0%wt). Theoretical Pt loading was approximately 10%wt of Pd content. All samples
were supported onto TS-1 zeolite.
b
From t-plots curves.
3
. Results and discussion
Bi-functional Pd@TS-1 materials were obtained through the
deposition of palladium species (from 0.2% to 1.0%wt) by impreg-
nation methodologies onto nanocrystalline TS-1 zeolite, used as
supporting matrix. The incorporation of additional metallic pro-
moters was similarly carried out merging palladium and platinum
salts. Following this procedure, several Pd@TS-1 and Pd(Pt)@TS-1
materials were prepared with different loading of metallic species
in function of metallic salts amount used during the synthesis pro-
cess. The chemical content of the different samples prepared is
shown in Table 1, being observed that the titanium content of TS-
1 support (∼2%wt) was not modified by the presence of metallic
species.
3.1. Synthesis and characterization of catalysts
The bi-functional Pd@TS-1 catalysts were based on palladium
nanoparticles supported onto nanocrystalline titanium silicalite
TS-1 which exhibited a double function, as inorganic matrix and
epoxidation catalyst of olefins. The modification with metallic pro-
moters, such as platinum species was also considered to improve
the catalytic activity of the solids (Pd(Pt)@TS-1).
The preparation of TS-1 nanocrystals was carried out by com-
bination of micellar and hydrothermal routes in presence of
surfactants which facilitate the formation of small crystallites
during the synthesis process [38]. XRD diffractograms clearly
evidenced that the as-synthesized and calcinated phases, remov-
ing organic structural directing agents, corresponded to crystalline
TS-1 zeolite, showing the characteristic diffraction bands of MFI-
type materials (Fig. 1). The alkaline media present in the synthesis
gel favored the formation of charged small nuclei which were
probably surrounded by surfactant molecules, avoiding crystal
growth. In fact, TS-1 nanocrystals with ∼600 nm of size and con-
taining very elevated titanium content (Si/Ti = 35, ∼2.0%wt) were
obtained (Table 1). In these samples, the totality of the titanium
was tetrahedrally coordinated into the zeolitic framework, this
fact being important to remark because it was reported that when
high amounts of titanium are incorporated, significant octahedral
extra-framework titanium was obtained, which is not active for
epoxidation processes [43,44]. UV–vis analysis and SEM inspection
confirmed the coordination state of the titanium on one hand, and
the size distribution of the TS-1 crystals on the other, as shown in
Fig. 2. The textural properties obtained from nitrogen adsorption
showed that the nanocrystalline TS-1 was microporous exhibiting a
type I plate isotherm characteristic of conventional 3D microporous
zeolites (Fig. 3 and Table 1).
(
c)
b)
(
(a)
5
10
15
20
25
30
35
40
2
Fig. 1. XRD patterns of (a) purely siliceous MFI-type zeolite, (b) as-synthesized TS-1
and (c) calcinated TS-1.
Fig. 2. UV–vis spectrum (a) and SEM micrographs (b) of calcinated TS-1 obtained through combined hydrothermal and micellar routes.

A. Prieto et al. / Catalysis Today 227 (2014) 87–95
91
Fig. 3. N2 isotherms of TS-1 and Pd(Pt)@TS-1 samples: (a) TS-1, (b) Pd02@TS-1, (c) Pd02(Pt)@TS-1, (d) Pd04(Pt)@TS-1, (e) Pd05(Pt)@TS-1 and (f) Pd10(Pt)@TS-1.
Furthermore, the crystallinity of the zeolitic phase remained
unmodified after the palladium and/or platinum deposition, such
as it is confirmed by XRD patterns of the different Pd(Pt)@TS-1
samples which maintain the diffraction bands characteristics of
MFI-type zeolites without formation of silica amorphous phases
between metallic nanoparticles although the metallic content was
increased (Fig. 6).
3.2. Catalytic activity
(
Fig. 4). It is important to point out that the deposition method-
Preliminary batch catalytic tests, using nanocrystalline TS-1
ology followed to incorporate metallic species did not imply the
generation of extra-framework titanium species because all spec-
tra of the bifunctional samples showed only one band focused at
∼
zeolite as catalyst from external H O2 and propylene, were per-
2
formed to produce PO. The results obtained clearly confirmed the
high reactivity of the titanium silicalite TS-1 zeolite support to cat-
alyze the epoxidation of propylene catalysts and the high efficiency
of the process to obtain elevated PO yields (∼80–85%) with high
selectivity (∼95–98%). This fact evidenced the validity of nanocrys-
talline TS-1 materials to be used as active supports to generate
effective bi-functional catalysts after the incorporation of palla-
dium and/or platinum nanoparticles. Moreover, the hydrothermal
stability of the nanocrystalline TS-1 zeolite support was addition-
ally studied through several catalytic runs carried out with the
same sample. Specifically, in Fig. 7, it was observed that the PO
200 nm, indicating the majority presence of titanium inserted
into the TS-1 network (Fig. 5). Only a slight reduction in the
total surface area and microporous volume was detected in the
samples when palladium and/or platinum species were finally
supported onto inorganic zeolitic matrix (Table 1), probably
due to the partial blockage of more accessible pores due to
the presence of metallic nanoparticles, exhibiting in all cases
type I isotherms characteristic of microporous zeolitic materials
(Fig. 3). This fact confirmed that the textural properties were not
substantially modified due to the presence of supported metal-
lic nanoparticles. From SEM micrographs, it was observed that
the TS-1 morphology of the samples was preserved although
palladium or platinum nanoparticles were deposited onto the
zeolitic support, the TS-1 nanocrystals being in all cases observed
around 600 nm of size. The TEM micrographs clearly evidenced
the formation of palladium nanoparticles homogeneously dis-
persed along the inorganic matrix, exhibiting diameters focused
at ∼10–20 nm, without observing agglomeration phenomenon
Fig. 4. XRD of different samples: (a) TS-1, (b) Pd02@TS-1, (c) Pd02(Pt)@TS-1, (d)
Fig. 5. UV of Pd(Pt)@TS-1 samples: (a) TS-1, (b) Pd02@TS-1, (c) Pd02(Pt)@TS-1, (d)
Pd04(Pt)@TS-1, (e) Pd05(Pt) and (f) Pd10(Pt)@TS-1 samples.
Pd04(Pt)@TS-1, (e) Pd05(Pt)@TS-1 and (f) Pd10(Pt)@TS-1.

9
2
A. Prieto et al. / Catalysis Today 227 (2014) 87–95
Fig. 6. TEM micrographs of (a) Pd02(Pt)@TS-1, (b) Pd02@TS-1, (c) Pd04(Pt)@TS-1, (d) Pd04(Pt)@TS-1 and (e) Pd10(Pt)@TS-1 samples.
yield obtained after the first reuse of catalyst is similar compared
with the results led by the activated fresh TS-1 catalyst. During
the different reuses, it was observed that PO selectivity remained
stable, although the catalytic activity at short times decreased.
However, even if its activity slightly decreased, longer reaction
times led to complete conversion of H O in the presence of propy-
to point out that it is characteristic of this type of consecutive pro-
cess that when high yields to PO are obtained, also an increase in
the amount of by-products formed during the reaction is observed,
and high selectivities are only obtained at low propylene conver-
sions, especially when gold is employed as metal for the H O2
2
synthesis from H2 and O , in gaseous phase [28,45]. When Pd cat-
2
2
2
lene to produce PO. In general, these results corroborated that
the TS-1 zeolite, used as active supporting matrix in the follow-
ing two-step reactions, was catalytically effective, stable and active
after several catalytic runs, without observing a substantial activity
loss.
alysts have been tested under gaseous phase, no PO production
has been observed [19] or very low yields have been obtained [46].
Specifically, in Scheme 2, the most typical by-products formed are
shown, indicating the reaction pathway followed in each case. In
this scheme are marked, the by-products detected in the catalytic
tests carried out in this work, this information being necessary to
correctly estimate the associated yield and selectivity to PO gener-
ation.
Considering these previous results, the next step was the study
of the catalytic processes for the direct epoxidation of propylene
with in situ generated H O , during one-pot two-step reaction
2
2
process (Scheme 1). With this purpose, batch catalytic tests were
carried out for the direct PO production from directly hydrogen,
oxygen and propylene, using before characterized bi-functional
Pd(Pt)@TS-1 nanocrystalline materials as catalysts. It is important
Due to the high number of possible products that could be
obtained in the reaction, it is decisive to find the most favorable
reaction conditions, using bi-functional solid catalysts, for both
the in situ generation of H O2 and the consecutive epoxidation of
2
propylene, in order to obtain high PO yield and selectivity, avoid-
ing the by-products formation. Supercritical CO2 (scCO ) has been
2
claimed to be a good solvent for the direct production of PO, enhanc-
ing also the H O2 efficiency for the epoxidation reaction [36]. For
2
this reason, scCO conditions were used in presence of bi-functional
2
Pd(Pt)@TS-1 catalysts. The reactions were carried out up to 80 bar
◦
and 60 C, in supercritical conditions, according the best results
described at literature [37b].
The catalytic results showed that the bi-functional Pd(Pt)@TS-1
catalysts were able to produce H O2 from H2 and O2 that reacted
2
with the propylene to generate PO, using CO2 as solvent and in
absence of another co-solvents (entry 1, Table 2). Specifically, when
the reaction process was carried out with Pd(Pt)@TS-1 catalyst,
the conversion of propylene was high (∼52%), although the PO
selectivity was low (∼6%), propane being the main reaction prod-
uct, although methoxypropanols and propylene glycol were also
detected. However, the formation of high amounts of undesirable
by-products was not avoided which implied reduced PO productiv-
ities (entry 1). With the objective to improve these results, different
reaction parameters were considered such as the incorporation of
additional co-solvents together with the presence of metallic pro-
moters.
Fig. 7. Catalytic activity and reuses of a TS-1 zeolite support prepared by micellar
route. The amount of the H2O2 was externally added according to the amount of
the weight of catalyst used in each reuse. On average, RH2 2
O
/Ti = 120. Propylene
is added in excess, using H2O2 as limiting reactant. Reaction conditions: 14 bar of
◦
pressure, 55 C, 5 h reaction time.

A. Prieto et al. / Catalysis Today 227 (2014) 87–95
93
O
MAIN GOAL
2 2
H O
Acrolein
O
O
H
2
O
2
1.5 O
H2
2
Propylene
MeOH
OH
Propylene
Propylene Oxide
Acroleic Acid
Propane
Propylene Carbonate
HO
HO
O
O
HO
Propylene Glicol
Isopropanol
MeOH
O
OH
OOH
CO2
OOH
OH
1
-hydroperoxypropan-2-ol 2-hydroperoxypropan-1-ol
2 2
H O
O
OH
H O
O
2
HO
O
4O2
Propylene Oxide
2
-methoxy-1-propanol
1-methoxy-2-propanol
O
O
C
O
H
H
Isomerizations
Carbon dioxide
Water
O
O
OH
Prop-2-en-1-ol
O
Acetone
Propanal
Methoxyethene
Scheme 2. Possible by-products formed during the direct PO formation through the in situ generated H2O2. In red, the by-products detected in this work. Some products
could be generated from more than one reaction pathway. (For interpretation of the references to color in this scheme legend, the reader is referred to the web version of
the article.)
It is described into the literature that the addition of small
amounts of promoters in the catalyst increased the production
of PO [32]. For this, some TS-1 catalysts, containing Pd and Pt
as metallic species, were prepared and tested on different con-
ditions, using MeOH as co-solvent in the reaction medium. The
results obtained showed that better performances were obtained
with the Pd02(Pt)@TS-1 catalyst in presence of MeOH, achiev-
ing PO selectivities and yields of ∼19% and ∼10%, respectively
water as additional co-solvent substantially worsened the catalytic
results (entry 3).
Despite the fact that the use of MeOH as co-solvent effectively
enhanced the formation of PO, propane still was the main product.
Moreover, it is known that the acidity of the TS-1 surface promoted
the decomposition of the produced PO, leading to by-products [31].
Additional water was also formed during the one-pot reaction and,
consequently, H CO formation is favored in presence of CO which
2
3
2
(
entry 2). However, the formation of high amounts of undesir-
increases the medium acidity, facilitating the decomposition of PO,
especially if MeOH was used as co-solvent and favoring the produc-
tion of ring-opening products, such as methoxypropanols, dropping
the PO selectivity. Taking in account all these considerations, the
reaction was carried out in the presence of acidity inhibitors such
as ammonium acetate, which is a well-known pH-controller buffer.
able by-products (methoxypropanols and propylene glycol) was
not avoided which implied reduced PO productivities. In general,
these data clearly confirmed the benefits of the use of metallic pro-
moters and co-solvents for the PO production with in situ generated
H O in supercritical CO conditions, being detected that the use of
2
2
2
Table 2
a
Catalytic tests carried out with in situ generated H2O2 using supercritical CO2 conditions in presence of Pd(Pt)@TS-1 catalysts .
C3 /H2/O2b
XC3
c
SPOd
SMPd
SPGd
SC3d
YPOe
Entry
Catalyst @TS-1
Solvent
1
2
3
4
5
6
7
8
Pd02(Pt)
Pd02(Pt)
Pd02(Pt)
Pd02(Pt)
Pd02(Pt)
Pd02
–
1/1.5/1.3
1/1/1.5
1/1.8/1.1
1/1.5/1
1/1.3/1.2
1/1.3/1.1
1/1.3/1.2
1/1/1.2
52.5
53.0
47.0
24.3
14.5
11.4
13.7
10.1
6.1
19.1
7.0
40.1
70.7
72.3
70.0
74.7
–
1.0
1.4
7.0
1.1
2.0
1.8
1.8
3.5
92.7
68.8
67.0
56.3
21.2
22.0
24.2
17.7
3.2
10.1
3.3
9.7
10.3
8.2
MeOH
10.3
19.0
0.5
3.5
3.9
3.5
3.7
MeOH + H2O
MeOH + H2O + AIf
MeOH + H2O + AI
MeOH + H2O + AI
MeOH + H2O + AI
MeOH + H2O + AI
Pd05(Pt)
Pd10(Pt)
9.6
7.5
a
◦
Reactions were carried out for 5 h with pressures around 80 bar and 60 C, feeding 2 mmol of propylene and 15 mg catalyst.
b
c
d
e
f
Gaseous molar ratios.
Propylene conversion.
Propylene oxide, methyl propanols, propylene glycol and propane selectivities.
Propylene oxide yield.
Acidity Inhibitor (CH3COONH4).

9
4
A. Prieto et al. / Catalysis Today 227 (2014) 87–95
Fig. 8. Catalytic reuses for PO direct production through in situ generated H2O2, using Pd04(Pt)@TS-1 catalyst.
In this case, the acidity was effectively controlled during the reac-
tion process, decreasing the formation of methoxypropanols and
propane which were promoted in acidic medium. In fact, the results
in Table 2 showed that the use of this acidity inhibitor enhanced the
formation of PO (yields of ∼10%), achieving selectivities up to 40%
due to the minimization of PO ring opening phenomenon, using
Pd02(Pt)@TS-1 catalyst (entry 4).
bi-functional catalyst recovered, at least partially, its initial activity
(Fig. 8).
4. Conclusions
Promising results were obtained to direct propylene oxide
(PO) production through one-pot two-step consecutive process
On the other hand, the C3 /H /O ratios used during the catalytic
2
2
from hydrogen, oxygen and propylene, using supercritical CO2
conditions in batch reactors in presence of optimal bi-functional
catalysts, based on palladium nanoparticles supported onto
nanocrystalline titanium silicalite zeolite (Pd@TS-1). These cata-
lysts were prepared and characterized, being active and reusable
to carry out the two-step process in which metallic nanoparticles
catalyzed the generation of in situ hydrogen peroxide as interme-
diate product for propylene epoxidation catalyzed by TS-1 active
matrix. The combined process was performed in presence of only
one solid and recyclable catalyst without necessity to use sepa-
rated catalytic processes, avoiding additional steps related with the
removing or separation of sub-products and/or intermediate sub-
strates. All these facts would imply a higher efficiency of global
catalytic route inside of more sustainable and environmental com-
bined process. Further improvements were achieved through the
modification of different reaction parameters, being concluded that
use of methanol as co-solvent, presence of acidity inhibitors, incor-
poration of platinum nanoparticles as promoters (Pd(Pt)@TS-1) and
employ of high amounts of oxygen in the reaction media increased
the PO yields (∼10%) and selectivities (∼70%), always working in
supercritical CO2 conditions. Partial deactivation of the catalysts
was observed during successive catalytic cycles probably due to
adsorbed organic compounds during the catalytic process, being
possible their regeneration after thermal treatments.
tests were also analyzed, being observed that when higher amounts
of O2 than H2 were used in the reactor batch, and in presence
of MeOH, H O and acidity inhibitor, propane formation, i.e., the
2
main hydrogenation product, was markedly reduced (entries 4 and
5
). Particularly, improved results were obtained when C3 /H /O2
2
molar ratios of approximately 1/1/1 were used during the combined
catalytic process, achieving PO selectivities close to 70% with PO
yields maintained around of 10% (entry 5). Moreover, the benefits
of the metallic promoters was corroborated because the catalytic
activity of Pd02@TS-1 sample, without platinum nanoparticles,
tested in the most suitable reaction conditions, implied a slight
reduction in the total propylene conversion that worsened the final
PO yield (entry 6).
Taking into account that the presence of metal nanoparticles
was necessary for the generation of H O2 from H2 and O , the
2
2
amount of metals deposited onto the TS-1 support was also stud-
ied in order to obtain as much H O2 as possible, at the same time
2
that selectivity to PO in the epoxidation of propylene remained
high. For this, several bi-functional catalysts with different Pd
loading (0.5% and 1.0%wt of Pd) were checked, being observed
that the PO selectivity was maintained (∼70%) but associated
to progressive propylene conversion decrease (entries 5 and 7).
This phenomenon was more marked when the palladium load-
ing was higher, decreasing the PO yield up to 7%, approximately
(
entry 8). These results showed that it was more convenient the
use of catalysts with palladium content lower than 0.5%wt which
exhibiting metallic nanoparticles with diameters around 15–20 nm
Acknowledgements
(
Fig. 6).
Finally, the stability of the bi-functional Pd(Pt)@TS-1 catalysts
The research leading to these results has received fund-
ing from European Community’s Seventh Framework Program,
through the Collaborative Project INCAS, Contract Nr. NMP2-LA-
2010-245988. AP and UD thank additional funds from Spanish
Government (MAT2011-29020-C02-01, Consolider-Ingenio MUL-
TICAT CSD2009-00050 and Severo Ochoa Excellence Program
SEV-2012-0267).
was evaluated through three consecutive catalytic cycles. After
each reuse, the catalyst solid was recovered and exhaustively
washed with methanol. The results obtained showed that, overall
after the third reuse, an appreciable decrease in the PO selectiv-
ity and propylene conversion were observed, confirming a partial
deactivation of bimetallic supported catalysts (Fig. 8). This fact
could be associated to incorporation of organic compounds, from
reaction media, during the catalytic process which blocked sev-
eral active sites. When these surface organic species were removed
through thermal treatments the regenerated and reactivated
References
[
1] B. Chowdhury, K.K. Bando, J.J. Bravo-Suárez, S. Tsubota, M. Haruta, J. Mol. Catal.
A Chem. 359 (2012) 21–27.

A. Prieto et al. / Catalysis Today 227 (2014) 87–95
95
[
2] F. Cavani, G. Centi, S. Perathoner, F. Trifirò, Sustainable Industrial Chemistry.
Principles Tools and Industrial Examples, Wiley-VCH, Weinheim, 2009.
3] M.J. Climent, A. Corma, S. Iborra, ChemSusChem 2 (2009) 500–506.
4] G. Centi, S. Perathoner, Catal. Today 138 (2008) 69–76.
[27] T. Hayashi, K. Tanaka, M. Haruta, Am. Chem. Soc. Div. Pet. Chem. 41 (1996)
71–74.
[28] N. Yap, R.P. Andres, W.N. Delgass, J. Catal. 226 (2004) 156–170.
[29] J. Huang, T. Takei, T. Akita, H. Ohashi, M. Haruta, Appl. Catal. B Environ. 95 (2010)
430–438.
[30] B.E.V. Schröder, H. Jansen, H. Schulze, Chem. Eng. Technol. 27 (2004)
847–851.
[31] W. Laufer, W.F.H. Hoelderich, Appl. Catal. A Gen. 213 (2001) 163–171.
[32] G. Jenzer, T. Mallat, M. Maciejewski, F. Eigenmann, A. Baiker, Appl. Catal. A Gen.
208 (2001) 125–133.
[
[
[
5] K. Weissermel, H.J. Arper, Industrial Organic Chemistry, 2nd ed., VCH Publishers
Inc., Weinheim, 1993, pp. 141–264.
[
[
[
[
6] J.R. Monnier, Appl. Catal. A Gen. 221 (2001) 73–91.
7] H. Chu, L. Yang, Q. Zhang, Y. Wang, J. Catal. 241 (2006) 225–228.
8] B. Taylor, J. Lauterbach, W.N. Delgass, Appl. Catal. A Gen. 291 (2005) 188–198.
9] R.A. van Santen, C.P.M.d. Groot, J. Catal. 98 (1986) 530–539.
[
10] J.J. Bravo-Suárez, J. Lu, A. Takahashi, M. Haruta, S.T. Oyama, J. Catal. 232 (2005)
[33] D. Hancu, E.J. Beckman, Green Chem. 3 (2001) 80–86.
[34] T. Danciu, E.J. Beckman, D. Hancu, R.N. Cochran, R. Grey, D.M. Hajnik, J. Jewson,
Angew. Chem. Int. Ed. 42 (2003) 1140–1142.
[35] A. Baiker, Chem. Rev. 99 (1999) 453–473.
[36] Y. Chen, Y. Wu, Y. Zhang, L. Long, L. Tao, M. Yang, N. Tang, J. Mol. Catal. A Chem.
352 (2012) 102–109.
8
5–95.
[
[
[
11] A.Z. Fadhel, C.L. Liotta, C.A. Eckert, Molecules 15 (2010) 8400–8424.
12] M.G. Clerici, G. Bellussi, U. Romano, J. Catal. 129 (1991) 159–167.
13] T.A. Nijhuis, M. Makkee, J.A. Moulijn, B.M. Weckhuysen, Ind. Eng. Chem. Res. 45
(
2006) 3447–3459.
[
[
14] C. Wang, B. Wang, X. Meng, Z. Mi, Catal. Today 74 (2002) 15–21.
15] G. Blanco-Brieva, M.C. Capel-Sanchez, M.P. de Frutos, A. Padilla-Polo, J.M.
Campos-Martin, J.L.G. Fierro, Ind. Eng. Chem. Res. 47 (2008) 8011–8015.
16] G. Blanco-Brieva, E. Cano-Serrano, J.M. Campos-Martin, J.L.G. Fierro, Chem.
Commun. 10 (2004) 1184–1185.
[37] (a) Q. Chen, E.J. Beckman, Green Chem. 9 (2007) 802–808;
(b) Q. Chen, E.J. Beckman, Green Chem. 10 (2008) 934–938.
[38] R.B. Khomane, B.D. Kulkarni, A. Paraskar, S.R. Sainkar, Mater. Chem. Phys. 76
(2002) 99–103.
[39] D. Hancu, E.J. Beckman, T. Danciu, US Patent 6 710 192 (2004).
[40] S.J. Gregg, K.S.W. Sing, Adsorption Surface Area Porosity, Academic Press, Lon-
don, 1982.
[41] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T.
Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619.
[42] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am. Chem. Soc. 73 (1951) 373–380.
[43] B. Taylor, J. Lauterbach, W.N. Delgass, Catal. Today 123 (2007) 50–58.
[44] A. Corma, U. Díaz, M.E. Dómine, V. Fornés, Angew. Chem. Int. Ed. 39 (2000)
1499–1501.
[
[
[
[
[
[
[
[
17] Y.F. Han, J.H. Lunsford, J. Catal. 230 (2005) 313–316.
18] R. Meiers, W.F. Hoelderich, Catal. Lett. 59 (1999) 161–163.
19] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566–575.
20] M. Taramaso, G. Perego, B. Notari, US Patent 4 410 501 (1983).
21] M.G. Clerici, P. Ingallina, J. Catal. 140 (1993) 71–83.
22] U. Díaz, D. Brunel, A. Corma, Chem. Soc. Rev. 42 (2013) 4083–4097.
23] B.E. Solsona, J.K. Edwards, P. Landon, A.F. Carley, A. Herzing, C.J. Kiely, G.J.
Hutchings, Chem. Mater. 18 (2006) 2689–2695.
[
24] S. Carrettin, P. Concepción, A. Corma, J.M. López-Nieto, V.F. Puntes, Angew.
[45] J. Lu, X. Zhang, J.J. Bravo-Suárez, T. Fujitani, S.T. Oyama, Catal. Today 147 (2009)
186–195.
Chem. Int. Ed. 43 (2004) 2538–2540.
[
[
25] G. Budroni, A. Corma, Angew. Chem. Int. Ed. 45 (2006) 3328–3331.
26] M. Haruta, M. Date, Appl. Catal. A Gen. 222 (2001) 427–437.
[46] S. Hikazudani, T. Mochida, N. Matsuo, K. Nagaoka, T. Ishihara, H. Kobayashi, Y.
Takita, J. Mol. Catal. A Chem. 358 (2012) 89–98.