Pd/TiO2 coatings with template-controlled mesopore structure as highly active hydrogenation catalyst

Article abstract of DOI:10.1016/j.apcata.2014.12.044

Micro-structured reactors offer excellent mass and heat transport capabilities and can therefore sustain very high reaction rates and space-time-yields also for highly exothermic catalytic reactions. However, such high rates cannot be reached when the reactors are coated or filled with conventional catalysts powders. We present a strategy for the direct synthesis of highly active wall-coated supported catalysts via co-deposition of a pore template (here micelles formed from PEO-b-PPO-b-PEO) and a precursors for the metal oxide (TiCl₄) along with a compatible precursor for the active metal (PdCl₂). The obtained catalytic coatings possess a template-controlled open pore structure and excellent mechanical stability. Moreover, the active metal is highly dispersed and well-distributed across the coating also at high Pd loadings. The corresponding high activity along with rapid mass transfer enabled by the open pore system results in the best space-time-yields in the gas-phase hydrogenation of butadiene reported so far in literature for a supported catalyst.

Full text of DOI:10.1016/j.apcata.2014.12.044

Accepted Manuscript
Title: Pd/TiO2 coatings with template-controlled mesopore
structure as highly active hydrogenation catalyst
¨
¨
Author: Erik Ortel Jorg Polte Denis Bernsmeier Bjorn
Eckhardt Benjamin Paul Arno Bergmann Peter Strasser
Franziska Emmerling Ralph Kraehnert
PII:
S0926-860X(14)00806-0
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2014.12.044
APCATA 15183
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
18-7-2014
15-12-2014
23-12-2014
Please cite this article as: E. Ortel, J. Polte, D. Bernsmeier, B. Eckhardt, B. Paul, A.
Bergmann, P. Strasser, F. Emmerling, R. Kraehnert, Pd/TiO2 coatings with template-
controlledmesoporestructureashighlyactivehydrogenationcatalyst, AppliedCatalysis
A, General (2014), http://dx.doi.org/10.1016/j.apcata.2014.12.044
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

*Graphical Abstract (for review)
co-deposition /
co-assembly
ingredients
solutions
mesoporous
Pd/TiO2
deposited
mesophase
TiCl4
PdCl2
EtOH
catalyst
calcination;
reduction
template
polymer
H2O
EtOH
Page 1 of 30

*Highlights (for review)
 new strategy for the direct synthesis of highly active wall-coated catalysts
 template-controlled mesoporous Pd/TiO2 catalysts
 space-time-yields exceed values reported in literature for butadiene hydrogenation
 approach is generic in nature and can be adapted to other noble metals and supports
Page 2 of 30

Pd/TiO2 coatings with template-controlled mesopore structure
as highly active hydrogenation catalyst
Erik Ortel1, Jörg Polte2, Denis Bernsmeier3, Björn Eckhardt3, Benjamin Paul3, Arno
Bergmann3, Peter Strasser3, Franziska Emmerling1, Ralph Kraehnert3*
1 BAM Federal Institute for Materials Research and Testing, Unter den Eichen 87, D-12200
Berlin, Germany
2 Humboldt University Berlin, Brook-Taylor-Str. 2, D-12489 Berlin, Germany
3 Technical University of Berlin, Department of Chemistry, Straße des 17. Juni 124, D-10623
Berlin, Germany
*ralph.kraehnert@tu-berlin.de
Abstract
Micro-structured reactors offer excellent mass and heat transport capabilities and can
therefore sustain very high reaction rates and space-time-yields also for highly exothermic
catalytic reactions. However, such high rates cannot be reached when the reactors are coated
or filled with conventional catalysts powders. We present a strategy for the direct synthesis of
highly active wall-coated supported catalysts via co-deposition of a pore template (here
micelles formed from PEO-b-PPO-b-PEO) and a precursors for the metal oxide (TiCl4) along
with a compatible precursor for the active metal (PdCl2). The obtained catalytic coatings
possess a template-controlled open pore structure and excellent mechanical stability.
Moreover, the active metal is highly dispersed and well-distributed across the coating also at
high Pd loadings. The corresponding high activity along with rapid mass transfer enabled by
1
Page 3 of 30

the open pore system results in the best space- time-yields in the gas-phase hydrogenation of
butadiene reported so far in literature for a supported catalyst.
KEYWORDS: titanium oxide films, palladium nanoparticle, wall-coated supported catalysts,
template-controlled mesoporous materials, hydrogenation of 1,3-butadiene
2
Page 4 of 30

1. Introduction
Process intensification in heterogeneous catalysis requires optimal combinations of catalyst
and reactor design to achieve high space-time yields and to prevent mass and heat transport
limitations. Supported catalysts, i.e. active metal particles dispersed in a porous metal-oxide
support, are the most common catalyst type in industrial hydrogenation and oxidation
reactions. Optimal space-time-yields can be reached when these catalysts are used in
conjunction with so-called micro-structured reactors.[1-4]
Microstructured reactors are typically characterized by flow channels with at least one
dimension between 10 and 1000 µm, laminar flow conditions and surface areas as high as
10000 to 50000 m2/m3 .[3, 5] The small channel dimensions facilitate extremely fast mass
transfer and heat transfer (up to 10 kW m-2 K-1 for liquids)[2]. Hence, very fast reactions with
large associated heat release, e.g. many hydrogenation and oxidation reactions, can be
performed under isothermal conditions. Moreover, the small channel dimensions efficiently
quench gas phase radical chemistry, which allows reactor operation in the explosive regime of
feed gas mixtures [6-9]. However, in order to use the full potential of microstructured
reactors catalytic coatings with significantly higher activity than comparable powder catalyst
are needed.[2] An overview on methods used to deposit catalyst on structured surfaces can be
found in [10].
The most common method of immobilizing solid catalysts in microstructured reactors is the
filling or wash-coating of the microchannel walls with a slurry prepared from conventional
powder catalysts. Typically employed powders are micron sized (Pt/Al2O3, Rh/Al2O3[11],
Al2O3[12]) or ball-milled (Al2O3[13], 1.5wt% Pd/Na-Al-Si-Ox[14] commercial catalysts or
catalyst supports. Very small particle sizes of the deposited powders have been obtained with
3
Page 5 of 30

self-made small sols of oxides (Al2O3[15, 16], SiO2[15], TiO2 [15], CeO2[17]). Active
catalysts were obtained by adding an active metal to the sol either prior to deposition[16] or
after deposition[17], followed by a final calcination/reduction step.
When the sols are deposited in the presence of pore templates such as polymer micelles, wall-
coated catalyst layers with ordered porosity and highly accessible internal surface can be
obtained, where pore sizes are controlled by the size of the pore template. Here, the active
metal can be introduced after the synthesis of the support oxide using classical impregnation
(PtRu /SiO2 [18]). However, improved control over the size and local distribution of the
active metal particles of the catalytic coating is obtained when sol and pore template are
codeposited together with preformed colloidal metal particles (e.g. Pd-NP/TiO2[19]).
Unfortunately, such codeposition syntheses succeeded so far only for low metal loadings
without disruption of the templated pore structure [19].
In consequence, the activity of wall-coated catalysts in micro-structured reactors did so far not
exceed the performance of the corresponding powder materials significantly. Contrarily,
catalytic coatings often suffer from lower activity due to a detrimental influence of the
procedure used for coating the catalyst powder onto the reactor wall. Approaches that allow
the direct synthesis of wall-coated catalyst layers with an increased metal loading, high metal
dispersion and good accessibility of the active sites are therefore desirable.
We present a facile synthesis strategy for the synthesis of catalytic coatings with templated
mesoporosity and high metal dispersions at Pd loadings up to 2.5 wt% on stainless-steel
substrates. The synthesis exploits the advantages of the evaporation-induced self-assembly
(EISA), for which the preparation of common support oxides with template-controlled pore
size is well established (SiO2[20], CeO2, ZrO2[21], Al2O3[22], TiO2[23, 24], MgO[25],
4
Page 6 of 30

ZnO[26] and other oxides[27]). The new approach described in Scheme 1 allows the direct
preparation of wall-coated supported porous catalysts with small active metal particles
distributed throughout the oxides pore system. The synthesis relies on co-deposition of the
pore template (here micelles formed from PEO-b-PPO-b-PEO) and a precursors for the metal
oxide (TiCl4) along with a precursor for the active metal (PdCl2) that is chemically
compatible, soluble under the same conditions and does not precipitate. Co-deposition of the
precursors and the micelles on a substrate leads to formation of micelles and a meso-
structured film (Scheme 1). Further drying, calcination and subsequent reduction induces the
crystallisation of the pore walls (TiO2), template removal (open mesopores) and the formation
of small palladium nanoparticles (PdNP) inside the pore system. Pore system and active metal
prove to be highly accessible and active in the gas-phase hydrogenation of butadiene. The
pore size can be readily controlled by the size of the template polymer, whereas the employed
Pd amount controls the metal loading of the catalyst.
The following section reports the protocols of catalyst synthesis, analytics and testing.
Thereafter, the physicochemical properties of a mesoporous 2.5 wt% Pd / TiO2 catalytic
coating are discussed, followed by an investigation of the influence of metal loading on the
catalysts morphology, pore structure and surface area. Finally, the performance of catalysts
with different Pd loading is assessed in the hydrogenation of butadiene and compared with
literature data.
2. Experimental
2.1 Chemicals and Materials
PdCl2 (99.9 %) was obtained from Alfa Aesar and TiCl4 (99.9 %) from Acros. Ethanol
(99.9 %) was purchased from Roth. Pluronic F127 was purchased from Sigma-Aldrich. These
chemicals were used without further purification. Water was purified (18.2 MΩ • cm MilliQ,
5
Page 7 of 30

Millipore). Si wafers and steel plates (grade 1.4301) were employed as substrates for catalyst
deposition. Before dip-coating, Si wafers were cleaned with ethanol. The surface of steel
plates (dimension of 30x27 mm) was pretreated and cleaned by a procedure developed earlier
via grinding with 180-grit sandpaper, washing and calcination.[19, 28] Both Si wafers and
steel plates were calcined in air for 2 h at 600 °C before the coating step.
2.2 Synthesis of Pd/TiO2 catalytic coatings
The catalytic coatings were synthesized by codeposition of TiCl4, Pluronic F127® (PEO100-
PPO70-PEO100) and PdCl2 by dip-coating on different substrates followed by a thermal
treatment step (see Scheme 1). TiCl4 was chosen as precursor due to good solubility of PdCl2
in acidic solutions. Moreover, TiCl4 -based anatase films with templated porosity show
excellent adhesion on stainless steel substrates[29] and possess a similar mechanical stability
as nonporous titania materials.[30] The amount of PdCl2 was varied to reach 0.5wt%, 1.5wt%
and 2.5wt% Pd in TiO2 after the synthesis. Two separate solutions were prepared initially. For
the first solution 1.16 ml TiCl4 was slowly transferred into 10.0 ml ethanol. Then the required
amount of PdCl2 (7.1 mg, 21.3 mg and 36.0 mg for 0.5wt%, 1.5wt% and 2.5wt% Pd in TiO2,
respectively) was dissolved in the highly acidic TiCl4-based solution (pH < 1). After stirring
for 24 h, the solution shows a clear orange color. For the second solution 1.50 g F127 were
dissolved in 10.0 ml ethanol and 1.67 ml water. Both solutions were joined to obtain the final
dip-coating solution.
Dip-coating of all catalysts was performed with a withdrawal rate of 60 mm/min in a
controlled atmosphere of 20% relative humidity and 25 °C. Each substrate was dipped only
once. The films were subsequently dried for 4 h at 80 °C in flowing air in a tube furnace.
Then the temperature was raised in flowing air with 1 K/min to 450 °C, held constant for
20 min and followed by naturally cooling down to room temperature. After calcination the
catalysts were reduced for 6 h at 350 °C in H2/Ar atmosphere (4 vol% H2).
6
Page 8 of 30

2.3 Characterization
SEM images were collected with a JEOL 7401F scanning electron microscope. Image J
Version 1.45s (http://rsbweb.nih.gov/ij) was employed to determine pore diameter, film
thickness and particle diameter and to derive FFT plots from the SEM images. Size and
crystallinity of particles as well as film morphology and crystallinity of film fragments were
studied by TEM (FEI Tecnai G2 20 S-TWIN instrument operated at 200 kV). 2D-SAXS
pattern with a beam incident angle of β = 5° or 90° with respect to the film surface were
recorded at the μSpot beamline at BESSY II synchrotron (Berlin, Germany) with a calibrated
radiation energy of 14995 eV. The SAXS data were processed employing the software FIT2D
Version V12.077.
XRD was measured on a Bruker D8 Advance (Cu Kα radiation) with gracing incident beam
(1°). Obtained data were analyzed with the Rietveld method using TOPAS V4.2 (Bruker
AXS). Experimental settings were considered by the Fundamental Parameter approach
incorporated in the program. The different phases were refined using scale factor, lattice
parameter and crystallite size. Influence of four background parameters and zero point error
were taken into account. Crystallite sizes were calculated from peak broadening based on the
volume-weighted column height.
XPS measurements were carried out with an AXIS Ultra DLD electron spectrometer
manufactured by Kratos Analytical, UK. XPS spectra were recorded using monochromated Al
Kα excitation. Spectra were taken by setting the instrument to the hybrid lens mode and the
slot mode providing approximately a 300 x 700 µm2 analysis area. The binding energy scale
was corrected for charging using an electron binding energy of 285.0 eV for the C1s level of
aliphatic hydrocarbon.
The Pd concentration of the 2.5wt% Pd/TiO2 sample was determined by analyzing the
dissolved catalyst in a 715-ES-inductively coupled plasma (ICP-OES) system (Varian). The
7
Page 9 of 30

Pd content of 0.5wt%, 1.5wt% and 2.5wt% Pd/TiO2 samples was also analyzed with a
Cameca "Camebax-microbeam" electron microprobe at ZELMI (TU-Berlin) by quantification
of WDX spectra. To determine the coating mass, the mass depth of each film was calculated
by STRATAGem film analysis software (v 4.3) based on WDX spectra. N2 sorption was
measured at 77 K using the Autosorb-1-C automated gas adsorption station from
Quantachrome. Prior to adsorption measurement the samples were degassed for 20 h at
150 °C in vacuum. The surface area was calculated via Brunauer-Emmett-Teller (BET)
method. For the sorption studies the respective catalytic coatings were coated onto Al foil
(total coated area ca. 180 cm2). From the adsorption branch of the N2-isotherm the pore size
distribution was calculated via non-local density functional theory (NLDFT) by the model
called ‘‘N2 at 77 K on silica (cylindr. pore NLDFT adsorption branch)’’ integrated in the
Quantachrome software AS1Win.
2.4 Catalytic testing
The catalytic performance of Pd/TiO2 catalyst films in the gas-phase hydrogenation of 1,3-
butadiene was studied at temperatures between 35 and 80 °C. Films coated on both sides of
five steel plates (plate size 27 mm × 30 mm) were stacked parallel into a reactor housing with
1.5 mm distance between the plates. The total coating masses amount to 3.0 mg, 3.4 mg and
3.3 mg for 0.5wt%, 1.5wt% and 2.5wt% Pd in TiO2, respectively. A test setup and procedure
similar to the one described in [19, 31] was used. A reaction mixture consisting of 10%
butadiene (2.5 purity), 20% hydrogen (5.0 purity), and 70% nitrogen (5.0 purity) was passed
through the reactor at a flow rate of 60 mL/min (STP) at 1.05 bar. The catalyst was then
heated to 80 °C under reactive gas flow and equilibrated to reaction conditions. Thereafter,
the temperature was decreased stepwise in 8 K or 5 K increments to 35 °C with a dwell time
of 60 min for each temperature set point. Analysis of the gas products was performed
continuously every 7 min by online gas chromatography (Agilent GC 7890 equipped with
8
Page 10 of 30

FID, TCD and columns HP Plot Al2O3, Molsieve 5A, HP Plot Q and DB FFAP.) The space-
time yield (STY) was calculated as produced moles of butenes per second and kilogram of the
catalyst total mass [mol s−1 kg−1]. Arrhenius plots were calculated from butadiene conversions
(lnXbutadiene) in the interval of low butadiene conversion.
3. Results and Discussion
The following sections describe the physico-chemical properties of micelle-templated
mesoporous Pd/TiO2 films (section 3.1). Thereafter, the influence of Pd loading of
mesoporous Pd/TiO2 films on the catalysts morphology, pore structure and surface area is
discussed (section 3.2). Finally, the performance of catalysts with different Pd loading is
assessed in the hydrogenation of butadiene and compared with literature data (section 3.3).
3.1 Physico-chemical properties of mesoporous Pd/TiO2 coatings
The structure and phase composition of micelle-templated Pd/TiO2 coatings was analyzed by
TEM and SAED after calcination at 450 °C in air and reduction at 350 °C in H2/Ar flow.
Figure 1 presents TEM images of a 2.5wt% Pd/TiO2 recorded in (A) bright-field mode, (B) Z-
contrast imaging in high-angle annular dark-field mode (HAADF) and (C) the size
distribution of PdNP obtained from HAADF images. The crystalline phases are assed based
on (D) SAED recorded at the same sample position, and high resolution TEM images of (E)
titania crystals and (F) Pd nanoparticles.
The bright field (BF) TEM images (Figure 1A) show that the material is completely
penetrated by mesopores of about 9 nm size. Within the mesopore structure additional bright
circular features can be distinguished (Figure 1A). The same features appear dark in the
corresponding Z-contrast HAADF image (Figure 1B) and can be assigned to Pd-
9
Page 11 of 30

nanoparticles, i.e. a material with higher electron density. The size distribution of the particles
is centred around 6.6 ± 1.5 nm diameter (Figure 1C).
SAED analysis of the Pd/TiO2 (Figure 1D) shows numerous diffraction rings. The ring
positions (indexing in Figure 1D) are consistent with the anatase phase (PDF 21-1272). High-
resolution TEM images confirm the presence of anatase crystals throughout the film (Figure
1E). The lattice fringes observed in the images can be assigned to (101)-planes of anatase
(d = 0.352 nm) (PDF 21-1272). Moreover, HR-TEM provides also evidence for the presence
of crystalline Pd particles in the reduced sample (Figure 1F) with lattice fringes corresponding
to metallic Pd (111)-planes with d-spacing of 0.225 nm (PDF 46-1043). The detected Pd-NP
possessed nearly spherical shape.
The morphology of the catalysts pore system, the local distribution of Pd particles within the
film and the extent of pore ordering were studied by SEM imaging and SAXS analysis,
respectively. Figure 2 presents for reduced 2.5 wt% Pd/TiO2 (A) cross sectional SEM images
recorded in secondary electron (SE) mode as well as in (B) backscattered electron (BSE)
mode along with SAXS recorded in transmission with beam incident angles of (C) β = 90°
and (D) β = 5°.
SE-SEM images (Figure 2A) confirm the abundant presence of mesopore openings of about
9 nm size also on the outer surface of the reduced Pd/TiO2 film. The films cross-section
proofs that the mesopores are present throughout the film volume. Moreover, bright spots
indicate the occasional presence of Pd particles of about 13 to 15 nm diameter on the outer
film surface. The film thickness amounts to ca. 250 nm. However, the synthesis concept
allows also the preparation of catalytic coatings with higher film thickness. Krins et al. [32]
demonstrated for F127 templated mesoporous TiO2 films (without Pd incorporation) that an
increase of the film thickness up to several µm can be realize by control of the withdrawal
speed during dip-coating and by repeating the coating step.
10
Page 12 of 30

BSE-SEM micrographs (Figure 2B) allow a better assessment of the elemental Z contrast,
which reveals that Pd-NPs are homogeneously distributed in the films cross section. The Pd-
NP diameter observed in the films cross section amounts to ca. 6 to 8 nm. This particle size
agrees well with the size obtained from TEM data (Figure 1C), which probes primarily the
volume of the catalytic coating.
The data suggest that size of formed Pd nanoparticles differs depending on their location.
When the Pd particles reside within the system of ordered and abundant mesopores (pore
diameter ~ 9 nm) then their size amounts to 6 to 8 nm. If the particles reside on the outer
surface of the catalyst film then the particle size amounts to about 13 to 15 nm diameter.
However, the vast majority of the formed PdNP resides within the pore system.
We attribute the size difference between PdNP located within the mesopores and on the outer
catalyst surface to a confinement effect exerted by the pore system. Such an effect has been
reported before by Schueth and co-workers for AuNP in porous ZrO2[33] and PtNP in porous
graphitic carbon[34]. They observed that the pore system stabilized the respective metal
nanoparticles against sintering under severe thermal and electro-chemical conditions. One
possible explanation for the formation of larger particles on the outer catalyst surface is that
the nanoparticles which are not constrained by the presence of the mesopores have a higher
mobility and can therefore move and sinter into larger particles. Hence, the micelle-templated
TiO2 pore system plays a crucial role in controlling the size of the formed Pd particles.
The Pd/TiO2 features a surface area of 114 m2·g-1 (from N2 sorption), which corresponds well
with the surface area typically observed for micelle-templated TiO2.[35] Yu et al. reported
113 m²/g for TiO2 films based on titanium tetraisopropoxide / Pluronic P123 calcined at
400 °C.[36] Hence, the surface area of Pd/TiO2 is in a similar range as observed for micelle-
11
Page 13 of 30

templated TiO2 without Pd. Moreover, the high surface area of the Pd/TiO2 films implies that
mesopores are accessible to N2 gas and interconnected to each other.[28]
The mesopores are locally ordered: 2D-SAXS recorded in transmission mode with an
incoming X-ray beam perpendicular to the substrates surface (β = 90 °, Figure 2C) features an
isotropic ring corresponding to a d-spacing in x-direction of dx = 12.0 nm (qx = 0.52 nm-1).
The pattern recorded at an angle of (β = 5 °, Figure 2D) shows an ellipsoidal shape. This
shape is commonly attributed to an ordered mesostructure after isotropic shrinkage in the
direction normal to the substrate.[37, 38] A comparison of dx with the d-spacing in z-direction
(qz ≈ 1.76 nm-1; dz ≈ 3.6 nm) indicates an anisotropic shrinkage of the pore network of
approximately 70 %. The type and extent of shrinkage are consistent with literature reports for
mesoporous pure TiO2 films based on TiCl4 and Pluronic block-copolymers as template[38].
Moreover, the circular and elliptical scattering features are similar to the scattering patterns
that were assigned to a distorted cubic arrangement of micelle-templated mesopores[37].
The detected crystalline phases were further confirmed by XRD analysis. Figure 3A presents
XRD patterns for a film after calcination in air at 450 °C (i "Pd/TiO2-calc") and after
subsequent reduction in H2/Ar stream at 350 °C (ii "Pd/TiO2-calc-red"). The calcined Pd/TiO2
(i) shows broad reflections at 2 Theta positions of 25.3°, 37.8° and 48.1° that can be assigned
to the (101), (004) and (200) reflections of a crystalline TiO2 anatase (PDF 21-1272).
Moreover, reflections with lower intensity can be distinguished at 2 Theta positions of 29.4°
and 33.6° (Figure 3A, i). These signals correspond well with (100) and (002) reflections of a
tetragonal palladium oxide phase (PDF 75-584).
Also the XRD pattern of the sample (ii) reduced in H2/Ar (Figure 3A, ii) is dominated by
reflections that correspond to anatase (25.3°, 37.8° and 48.1°). However, an additional small
12
Page 14 of 30

signal appears at 2 Theta of 30.8°. This reflection can be assigned to crystalline TiO2 in
brookite phase (hkl 121) (PDF 029-1360). Moreover, the PdO signals at 2 Theta of 29.4° and
33.6° disappear during reduction, whereas new reflections at 40.1° and 46.7° indicate the
formation of metallic palladium (hkl (111) and (200), PDF 46-1043). The average crystallite
size (derived from Rietveld refinement) amounted to 6.4 nm for anatase and 6.5 nm for Pd.
These values are consistent with crystallite and particle sizes extracted from TEM images
(Figure 1).
The surface composition of the reduced Pd/TiO2 film (ii) was analyzed by XPS. Figure 3B
shows XPS spectra recorded in the Ti2p region (B) and the Pd3d region (C) of binding
energies along with the deconvoluted signals. The binding energy (BE) obtained for the
deconvoluted peaks Ti2p3/2 and Ti2p1/2 amounted to 458.9 eV and 464.6 eV, respectively. The
values are in good agreement with binding energies reported in literature for TiO2 (e.g.
458.9 eV[39] and 458.7 eV[40] for Ti2p3/2).
The spectra recorded in the Pd3d range indicate that the Pd/TiO2 films contain two different
surface species of Pd with binding energies of 335.3 eV and 336.2 eV (Pd3d5/2). The absolute
peak positions as well as the observed BE difference of ca. 1 eV allow the assignment to
metallic Pd0 for the main species and to Pd2+ for the smaller signal [41]. Hence, a fraction of
about 18 at% oxidized Pd is present on the surface of the metallic Pd particles after heat
treatment in 4 % H2/Ar and sample storage in air at ambient temperature. The presence of
small amounts of PdO in Pd/TiO2 also after hydrogen treatment has been explained by Wang
et al. with the fact that palladium can be oxidized by the oxygen provided from the titania
support[42]. Based on XPS analysis they identified both PdOx and metallic palladium species
after reduction of Pd/TiO2 samples at 400 °C in 10 % H2/He atmosphere.
13
Page 15 of 30

Analysis of the bulk composition of Pd/TiO2 catalyst with ICP-OES indicates a Pd content in
TiO2 of approximately 2.4 wt%. This value matches closely with the 2.5 wt% Pd expected
from the composition of the dip-coating solution.
The combined data confirm the successful synthesis of Pd/TiO2 coatings with template-
controlled pore size, highs surface area and small Pd nanoparticles distributed throughout the
volume of the porous network. The pore system appears to be essential in retaining a small
particle size. The synthesis strategy gives direct access to control the pore size by the template
as well as the Pd loading via the amount of co-deposited catalytic species.
3.2 Influence of the variation of the Pd loading in mesoporous Pd/TiO2 coatings
To analyze the interplay between Pd content, particle size and pore system mesoporous
Pd/TiO2 titania films with different Pd content of 0.5, 1.5 and 2.5wt% Pd were synthesized.
Compositional analysis of these films by WDX indicates that the obtained content ratio of Pd
to TiO2 of approximately 0.6 wt%, 1.5 wt% and 2.4wt% matches closely with the Pd content
expected from the composition of the dip-coating solution.
Figure 4 presents (A, B, C) SE-mode SEM images in top view as well as (D, E, F) Z-contrast
SEM images in cross-section view for Pd/TiO2 films with different Pd content after
calcination and reduction. All films feature the expected mesopores (dark spots in Figure
4A/B/C) and Pd particles (bright spots). The films feature pore diameters of 9.4 nm, 9.2 nm
and 9.9 nm, respectively. Hence, the obtained pore size is independent of Pd loading within
the range of experimental accuracy. The ordering of the templated mesopores can be accessed
from the corresponding FFT images (insets in Figure 4 A, B and C). Distinct rings of about
12 nm diameter can be observed independent of the Pd loading. Hence, the film surface
14
Page 16 of 30

consists of locally ordered mesopores imprinted by the F127 polymer. Moreover, in the range
of 0.5 to 2.5 wt% Pd-loading the local ordered pore system remains intact.
SEM analysis reveals also the size and local distribution of Pd particles. The number of
particles on the films outer surface appears to increases with Pd-loading from 0.5 wt% (Figure
4A) to 2.5 wt% (Figure 4C). However, the size of these Pd particles does not change
significantly with diameters of 13.8 ± 3.7 nm, 13.0 ± 3.0 nm and 14.1 ± 4.5 nm for 0.5, 1.5
and 2.5 wt% Pd, respectively. The film cross-section images (Figure 4 D, E, F) recorded in Z-
contrast SEM-mode confirm a homogeneous Pd-particle distribution also in the volume of the
titania film. The Pd particle diameters within the film amount to 6.1 ± 1.2 nm, 6.1 ± 2.6 nm
and 7.8 ± 2.3 nm for 0.5, 1.5 and 2.5 wt%, respectively, suggesting a slight increase in the
size of the particles with increasing Pd-loading. A comparison between the size of Pd particles
on the films outer surface (ca. 14 nm) and inside the films volume (ca. 6 - 8 nm) confirms that
pore confinement inside the templated pore system stabilizes smaller Pd particles for all
studied Pd loading.
N2 physisorption analysis was used to study the effect of Pd loading on the catalysts surface
area. The accessible surface area obtained by BET analysis yields values of 116 m2·g-1 for a
TiO2 coating without Pd and 107 m2·g-1 , 120 m2·g-1 and 114 m2·g-1 for 0.5 wt%, 1.5 wt% and
2.5 wt% Pd/TiO2, respectively. Hence, the addition of Pd induces no significant changes in
the catalysts surface area. The pore diameters calculated from the adsorption branch of the N2-
isotherm via NLDFT amounts to 7 nm for the catalyst without Pd and 6 nm for all three
Pd/TiO2 catalysts. Hence, the addition of Pd causes only a slight decrease in the pore
diameter. Table 1 summarizes the physicochemical characteristics of the catalyst derived from
WDX, SEM, XRD and physisorption analysis.
15
Page 17 of 30

Table 1: Summary of the physicochemical characteristics of the Pd/TiO2 catalyst.
Pd loading in
TiO2
Pd particle diameter crystallite sizes
SEM XRD
outer film
pore
pore
diameter N2-BET diameter
inside
nominell WDX the film
SEM
Pd anatase top-view
surface
area
surface
NLDFT
wt%
0
0.5
1.5
2.5
wt%
-
0.6
1.5
2.4
nm
-
nm
-
nm
-
-
nm
-
-
nm
10.2
9.4
9.2
9.9
m2 g-2
116
107
120
nm
7
6
6
6
6.1 ± 1.2 13.8 ± 3.7
6.1 ± 2.6 13.0 ± 3.0
7.8 ± 2.3 14.1 ± 4.5
-
-
6.5
6.4
114
3.3 Catalytic performance in Butadiene hydrogenation
The catalytic performance of the mesoporous Pd/TiO2 coatings with 0.5, 1.5 and 2.5 wt% Pd
loading was studied in the gas phase hydrogenation of butadiene passing a feed gas containing
10% butadiene and 20% H2 in N2 at constant flow rate of 60 mL/min over stacks of 5 catalyst
plates and analyzing at different reaction temperatures the composition of the product gas.
Figure 5 plots for each catalyst (A) butadiene conversion vs. temperature, (B) selectivity to
butane, 1-butene and the sum of all formed butenes (S1-butene + Strans-2-butene + Scis-2-butene) vs.
butadiene conversion, and (C) the obtained space-time yield to butene (STYbutene) vs. the Pd
loading along with literature values for reference.
All catalysts are active (Figure 5A) and selective (Figure 5B). Butadiene conversion increases
with increasing reaction temperature and increasing Pd loading (Figure 5A). The butadiene
conversion at 48 °C amounted to 1.8%, 8.1%, and 24.7% for 0.5wt%, 1.5wt% and 2.5wt%
Pd/TiO2 catalysts, respectively. The stability of all three catalysts was measured by running
the reaction for 12 h at constant flow conditions at 80°C. The change in butadiene conversion
amounted to less than 2.0 % (relative) for all three catalysts. The catalysts were therefore
stable during extended lab-scale tests. Moreover, also the selectivity to the individual products
(compared at the same butadiene conversion) agrees between the differently loaded catalysts
16
Page 18 of 30

within the limits of experimental accuracy. All catalysts form at low conversion 1-butene
(ca. 62%), trans-2-butene (ca. 31%), cis-2-butene (ca. 7 %) and only traces of n-butane (<1%)
(Figure 5B). With butadiene conversion exceeding ca. 50% all catalysts start to form
significant amounts of butane, whereas the selectivity to butene decreases. Both, the
selectivity values at low conversion as well as the trend towards increased total hydrogenation
at higher conversion agree well with literature data on Pd-based catalysts.[31] Pattamakomsan
et al.[43] studied materials prepared via conventional impregnation of porous Al2O3 with Pd
(II) acetylacetonate. Their 0.5 wt.% Pd catalyst with Pd particle sizes of about 2 to 6 nm
produced about 50 to 60% 1-butene at low conversion. We previously observed ca. 55%
1-butene selectivity for a Pd-NP/TiO2 catalyst synthesized from colloidal PdNPs.[19]
Arrhenius plots were constructed for all catalysts in the interval of low butadiene conversion,
i.e. 35 to 56 °C. The data could be fitted well by straight lines and yielded activation energies
of about 60, 62 and 68 kJ/mol for 0.5wt%, 1.5wt% and 2.5wt%, respectively. These values
are in line with activation energies reported in literature for Pd/Al2O3 catalyst (48–66
kJ/mol[44], 75 kJ/mol[45]) as well as Pd/TiO2 (62 kJ/mol[19]). The comparison of selectivity
and activation energy indicates that the developed mesoporous Pd/TiO2 catalysts show the
typical behavior of Pd-based supported catalysts.
Figure 5C evaluates at 48°C the influence of Pd loading on the space time yield (STY) for
butenes and references these values to literature data typically obtained at 50°C. The STY was
calculated as produced moles of butenes per second and kilogram of the catalyst total mass
[mol s−1 kg−1]. The STY increases with Pd loading for the mesoporous Pd/TiO2 catalysts from
0.013 to 0.060 and 0.183 mol s-1 kg-1 for 0.5wt%, 1.5wt% and 2.5wt% Pd, respectively (black
circles in Figure 5C). The increase with Pd loading is slightly faster than linear. This effect
could be related both to particle size effects[46] or to parts of the Pd being inaccessible to the
17
Page 19 of 30

reaction at lower Pd loading. Particle size effects are less likely due to the fact that differently
loaded Pd catalysts did not differ significantly in Pd particle size. However, the catalysts were
synthesized from solutions of premixed Ti and Pd precursors. The precursors of Pd and Ti are
thus likely to be homogeneously distributed in the deposited mesophase (see Scheme 1). The
formation of crystals of palladium and titania species occurs only during the subsequent
calcination step. If a fraction of the Pd atoms remains in the pore walls during this
crystallization step, they would not be able to contribute to the catalytic reaction. The fraction
of Pd atoms that are trapped would be higher for lower total amounts of Pd, i.e. for lower Pd-
loading in the supported catalyst, offering one possible explanation for the observed activity
trend.
Figure5C also compares the performance of mesoporous Pd/TiO2 with literature data.
Pattamakomsan et al.[40] obtained by for their Pd-Al2O3 catalyst loaded with 0.5 wt.% Pd an
STY of 0.016 mol s-1 kg-1. We reported [19] for a 0.5wt% PdNP/TiO2 catalyst based on a Ti-
complex precursor and colloidal PdNP a STY of 0.019 mol s-1 kg-1. Lee et al.[47] reported for
a TiO2-modified Pd-SiO2 catalysts with 1.0wt% Pd an STY of ca. 0.061 mol s-1 kg-1. Hence,
for the low Pd loadings reported in literature a similar activity can be reached with the
mesoporous Pd/TiO2 catalysts. However, with the developed synthesis homogeneous
mesoporous Pd/TiO2 coatings can be realized with a Pd loading as high as 2.5 wt% while
retaining a small Pd size, a high surface area and good accessibility to the active sites. The
2.5wt% Pd/TiO2 catalyst therefore reaches about three times higher space time yields to
butene than reported in literature so far. Moreover, the very fast and highly exothermal
butadiene hydrogenation was realized under isothermal conditions and in the absence of mass
transfer limitation. The developed material therefore allows a significantly better utilization of
the available reactor space without compromising the intrinsic activity of palladium.
18
Page 20 of 30

4. Conclusion
The developed synthesis protocol facilitates the synthesis of highly active catalytic coatings
and gives direct access to control the catalysts loading of active species as well as the
opportunity to control the pore size by the template polymer. The synthesis exploits the
advantages of the EISA approach and extends it to the preparation of supported Pd/TiO2
catalyst by codeposition of an oxide precursor and a pore template together with ions of a
noble metal. The derived catalytic coatings show a high accessible surface area and metal
dispersion up to at least 2.5wt% Pd loading. Confinement effects exerted by the templated
mesopores during calcination and reduction appear to be essential to retain small particle sizes
in the bulk of the catalyst. No detrimental effects of Pd on the properties of the porous TiO2
support (film integrity, pore structure, surface area) were observed. The coatings proof to be
highly active and selective in the gas phase reaction of butadiene hydrogenation. As a result
of the obtained high Pd loading, small particle size and good diffusional access provided by
the open pore system, the highest space- time-yield to butene reported so far in literature for a
supported catalyst was achieved. Despite high reaction rates and strong exothermicity, the
reaction can be performed under isothermal conditions and without signs of diffusion
limitation.
The presented data provide a proof-of-concept for the synthesis of catalytic coatings for
micro-structured reactors via simple co-deposition of the template with precursors for catalyst
support and active phase. EISA-based syntheses have been reported already for numerous
typical support oxides. Developing these synthesis routes further to incorporate also different
active metal species will provide access to a wide range of supported catalysts with high metal
loading and accessible pore structure.
19
Page 21 of 30

Acknowledgment
The authors acknowledge ZELMI at Technical University Berlin for support in material
analytics and Dieter Treu at BAM for XPS analysis. RK thanks in particular Einstein
Foundation Berlin for generous support provided by an Einstein-Junior-Fellowship (EJF-
2011-95). DB, EO and RK acknowledge also funding from BMBF (FKZ 03EK3009).
References
[1]
[2]
V. Hessel, B. Cortese, M. de Croon, Chem. Eng. Sci. 66 (2011) 1426-1448.
V. Hessel, S. Hardt, H. Löwe, Chemical Micro Process Engineering, Wiley-VCH Verlag GmbH & Co.
KGaA, 2005, pp. 66-124.
[3]
[4]
[5]
[6]
L. Kiwi-Minsker, A. Renken, Catal. Today 110 (2005) 2-14.
G. Kolb, H. Pennemann, R. Zapf, Catal. Today 110 (2005) 121-131.
B. Bromley, V. Hessel, A. Renken, L. Kiwi-Minsker, Chem. Eng. Technol. 31 (2008) 1162-1169.
C. Liebner, J. Fischer, S. Heinrich, T. Lange, H. Hieronymus, E. Klemm, Process Saf. Environ. Prot. 90
(2012) 77-82.
[7]
A. Kursawe, D. Hönicke, Catal. Commun. 2 (2001) 347-351.
[8]
M.T. Janicke, H. Kestenbaum, U. Hagendorf, F. Schuth, M. Fichtner, K. Schubert, J. Catal. 191 (2000)
282-293.
[9]
T. Inoue, M.A. Schmidt, K.F. Jensen, Ind. Eng. Chem. Res. 46 (2007) 1153-1160.
V.r. Meille, Appl. Catal., A General 315 (2006) 1-17.
[10]
[11]
[12]
[13]
R. Zapf, G. Kolb, H. Pennemann, V. Hessel, Chem. Eng. Technol. 29 (2006) 1509-1512.
E. Simsek, A.K. Avci, Z.I. Onsan, Catal. Today 178 (2011) 157-163.
S.A. Schmidt, N. Kumar, B. Zhang, K. Eranen, D.Y. Murzin, T. Salmi, Ind. Eng. Chem. Res. 51 (2012)
4545-4555.
[14]
[15]
[16]
[17]
[18]
H. Qiu, L. Bednarova, W.Y. Lee, Appl. Catal., A 314 (2006) 200-207.
K. Haas-Santo, M. Fichtner, K. Schubert, Appl. Catal., A 220 (2001) 79-92.
H. Chen, L. Bednarova, R.S. Besser, W. Lee, Appl. Catal., A 286 (2005) 186-195.
A. Ates, P. Pfeifer, O. Görke, Chem. Ing. Tech. 85 (2013) 664-672.
O. Muraza, E.V. Rebrov, T. Khirnyak, B.F.G. Johnson, P.J. Kooyman, U. Lafont, M. De Croon, J.C.
Schouten, Chem. Eng. J. 135 (2008) 99-103.
[19] E. Ortel, S. Sokolov, C. Zielke, I. Lauermann, S. Selve, K. Weh, B. Paul, J. Polte, R. Kraehnert, Chem.
Mater. 24 (2012) 3828-3838.
[20]
[21]
C.J. Brinker, Y.F. Lu, A. Sellinger, H.Y. Fan, Adv. Mater. 11 (1999) 579-+.
E.L. Crepaldi, G.J. de A. A. Soler-Illia, A. Bouchara, D. Grosso, D. Durand, C. Sanchez, Angew.
Chem. Int. Ed. 42 (2003) 347-351.
[22]
M. Kuemmel, D. Grosso, U. Boissiere, B. Smarsly, T. Brezesinski, P.A. Albouy, H. Amenitsch, C.
Sanchez, Angew. Chem. Int. Ed. 44 (2005) 4589-4592.
[23]
E. Ortel, A. Fischer, L. Chuenchom, J. Polte, F. Emmerling, B. Smarsly, R. Kraehnert, Small 8 (2012)
298-309.
[24]
S. Sokolov, E. Ortel, R. Kraehnert, Mater. Res. Bull. 44 (2009) 2222-2227.
[25]
B. Eckhardt, E. Ortel, J. Polte, D. Bernsmeier, O. Görke, P. Strasser, R. Kraehnert, Adv. Mater. 24
(2012) 3115-3119.
[26]
B. Eckhardt, E. Ortel, D. Bernsmeier, J. Polte, P. Strasser, U. Vainio, F. Emmerling, R. Kraehnert,
Chem. Mater. 25 (2013) 2749-2758.
[27]
[28]
[29]
[30]
C. Sanchez, C. Boissiere, D. Grosso, C. Laberty, L. Nicole, Chem. Mater. 20 (2008) 682-737.
S. Sokolov, E. Ortel, J. Radnik, R. Kraehnert, Thin Solid Films 518 (2009) 27-35.
E. Ortel, S. Sokolov, R. Kraehnert, Microporous and Mesoporous Mater. 127 (2009) 17-24.
T. Rahman, R. Liu, E. Ortel, R. Kraehnert, A. Antoniou, Appl. Phys. Lett. 104 (2014).
20
Page 22 of 30

[31]
T. Cukic, R. Kraehnert, M. Holena, D. Herein, D. Linke, U. Dingerdissen, Appl. Catal., A 323 (2007)
25-37.
[32]
N. Krins, M. Faustini, B. Louis, D. Grosso, Chem. Mater. 22 (2010) 6218-6220.
[33]
[34]
P.M. Arnal, M. Comotti, F. Schüth, Angew. Chem. Int. Ed.45 (2006) 8224-8227.
C. Galeano, J.C. Meier, V. Peinecke, H. Bongard, I. Katsounaros, A.A. Topalov, A. Lu, K.J.J.
Mayrhofer, F. Schüth, J. Am. Chem. Soc. 134 (2012) 20457-20465.
[35]
[36]
[37]
[38]
J.H. Pan, X.S. Zhao, W.I. Lee, Chem. Eng. J. 170 (2011) 363-380.
J.C. Yu, X.C. Wang, X.Z. Fu, Chem. Mater. 16 (2004) 1523-1530.
W. Ruland, B.M. Smarsly, J. Appl. Crystallogr. 40 (2007) 409-417.
E.L. Crepaldi, G. Soler-Illia, D. Grosso, F. Cagnol, F. Ribot, C. Sanchez, J. Am. Chem. Soc. 125 (2003)
9770-9786.
[39]
D. Courcot, L. Gengembre, M. Guelton, Y. Barbaux, B. Grzybowska, J. Chem. Soc., Faraday Trans. 90
(1994) 895-898.
[40]
2951.
[41]
[42]
[43]
R. Sanjinés, H. Tang, H. Berger, F. Gozzo, G. Margaritondo, F. Lévy, J. Appl. Phys. 75 (1994) 2945-
T.H. Fleisch, G.J. Mains, J. Phys. Chem. 90 (1986) 5317-5320.
X. Wang, G. Wu, N. Guan, L. Li, Appl. Catal., B 115 (2012) 7-15.
K. Pattamakomsan, E. Ehret, F. Morfin, P. Gelin, Y. Jugnet, S. Prakash, J.C. Bertolini, J. Panpranot, F.
Aires, Catal. Today 164 (2011) 28-33.
[44]
[45]
[46]
[47]
D. Seth, A. Sarkar, F.T.T. Ng, G.L. Rempel, Chem. Eng. Sci. 62 (2007) 4544-4557.
J. Silvestre-Albero, G. Rupprechter, H.J. Freund, Chem. Commun. (2006) 80-82.
J. Silvestre-Albero, G.n. Rupprechter, H.-J. Freund, J. Catal. 240 (2006) 58-65.
D.C. Lee, J.H. Kim, W.J. Kim, J.H. Kang, S.H. Moon, Appl. Catal., A 244 (2003) 83-91.
List with the caption to figures:
Scheme 1. Illustration of the synthesis approach for Pd/TiO2 catalytic coatings with template
controlled mesopore structure. Synthesis solutions containing the Ti precursor, Pd precursor
and the template polymer Pluronic F127 are mixed, deposited via dip-coating and then dried
to obtain a micelle-structured mesophase. The dried mesophase is calcined to remove the
template polymer and crystallize the pore walls, and finally reduced to obtain metallic Pd
particles inside the mesoporous titania.
Figure 1. TEM micrographs of mesoporous Pd/TiO2 film with a Pd-loading of 2.5wt%
calcined at 450°C in air and reduced at 350°C in 4vol% H2/Ar. A) bright field TEM image
(dark spots correspond to PdNP). B) Z-contrast image in high-angle annular dark-field mode
(HAADF) (white spots correspond to PdNP). C) histogram of the Pd-NP size distribution. D)
selected area electron diffraction pattern. Indexing corresponds to TiO2 anatase (PDF 21
1272) E, F) high resolution TEM images. Lattice fringe spacing can be assigned to TiO2
anatase (E) and Palladium (F) (PDF 46-1043).
Figure 2. Cross sectional SEM micrographs in secondary electron mode (A) and in z-contrast
backscattered electron mode (B) with sample tilt of 45° and 2D-SAXS analysis recorded in
transmission mode with an incoming X-ray beam of β=90° (C) and β=5° (D) of Pd/TiO2 film
with a Pd-loading of 2.5wt% calcined at 450°C in air and reduced at 350°C in 4vol% H2/Ar.
Figure 3: XRD patterns (A) and XPS spectra (B and C) of Pd/TiO2 film with a Pd-loading of
2.5wt%. XRD patterns (A) show the film after calcination in air for 20 min at 450 °C
(bottom) and after reduction in H2/Ar stream for 6 h at 350 °C (top). Reflection position
assigned according to literature to TiO2 anatase (PDF 21 1272), Brookite (PDF 029-1360),
tetragonal PdO (PDF 75-584) and metallic Pd (PDF 46-1043). XPS spectra show the film
after reduction for the Ti2p region (B) and the Pd3d region (C). Symbols indicate
experimental data, lines the corresponding peak deconvolution.
21
Page 23 of 30

Figure 4. SEM analysis in top view in secondary-electron mode (A,B,C) and in cross-section
view in z-contrast mode (D, E, F) of the Pd/TiO2 catalytic layers with Pd-loading of 0.5wt%
(A, D); 1.5wt% (B, E) and 2.5wt% (C, F) after calcination at 450°C in air and reduction at
350°C in 4vol% H2/Ar.
Figure 5. Activity and selectivity of the developed mesoporous Pd/TiO2 catalysts in the gas
phase hydrogenation of 1,3-butadiene. A) conversion over temperature; B) selectivity over
butadiene conversion for 0.5wt%, 1.5wt% and 2.5wt% Pd/TiO2 catalysts (feed gas containing
10% butadiene and 20% H2 in N2, flow rate of 60 mL/min, studied temperature range: 35 to
80°C); C) Space time yields (STY) for butenes at 48 °C against Pd-loading for the
mesoporous Pd/TiO2 catalysts and for representative Pd catalysts at ca. 50 °C reported in
literature.
22
Page 24 of 30

Figure Scheme 1 - Synthesis approach for mesoporous Pd/TiO2 coatings
co-deposition /
co-assembly
ingredients
solutions
mesoporous
Pd/TiO2
deposited
mesophase
TiCl4
PdCl2
EtOH
catalyst
calcination;
reduction
template
polymer
H2O
EtOH
Scheme 1. Illustration of the synthesis approach for Pd/TiO2 catalytic coatings with template controlled mesopore structure.
Synthesis solutions containing the Ti precursor, Pd precursor and the template polymer Pluronic F127 are mixed, deposited via
dip-coating and then dried to obtain a micelle-structured mesophase. The dried mesophase is calcined to remove the template
polymer and crystallize the pore walls, and finally reduced to obtain metallic Pd particles inside the mesoporous titania.
Page 25 of 30

Figure 1 - TEM analysis of mesoporous 2.5wt% Pd/TiO2
25
dPd-NP = 6.6 ± 1.5 nm
A
B
C
20
15
10
5
Pd-NP
0
0 2 4 6 8 10 12 14 16 18 20
100 nm
100 nm
Pd-NP diameter / nm
TiO2
Anatase
E
D
F
204
105/211
Pd (111)
200
004
101
TiO2
Anatase
(101)
2 nm-1
2 nm
2 nm
Figure 1. TEM micrographs of mesoporous Pd/TiO2 film with a Pd-loading of 2.5wt% calcined at 450°C in air and reduced at 350°C
in 4vol% H2/Ar. A) bright field TEM image (dark spots correspond to PdNP). B) Z-contrast image in high-angle annular dark-field
mode (HAADF) (white spots correspond to PdNP). C) histogram of the Pd-NP size distribution. D) selected area electron diffraction
pattern. Indexing corresponds to TiO2 anatase (PDF 21 1272) E, F) high resolution TEM images. Lattice fringe spacing can be
Page 26 of 30
assigned to TiO2 anatase (E) and Palladium (F) (PDF 46-1043).

Figure 2 - Film morphology and pore ordering
A
B
x
x
y
y
z
z
Si-Wafer
Si-Wafer
100 nm
100 nm
C
D
90°
5°
z
x
0.5 nm-1
0.5 nm-1
y
y
Figure 2. Cross sectional SEM micrographs in secondary electron mode (A) and in z-contrast backscattered electron mode (B)
with sample tilt of 45° and 2D-SAXS analysis recorded in transmission mode with an incoming X-ray beam of β=90° (C) and β=5°
Page 27 of 30
(D) of Pd/TiO2 film with a Pd-loading of 2.5wt% calcined at 450°C in air and reduced at 350°C in 4vol% H2/Ar.

Figure 3 - Bulk and surface composition of 2.5wt% Pd/TiO2
A
B
C
Pd3d
Anatase
(101)
Ti2p
335.3 eV
458.9 eV
340.6 eV
Anatase
(004)
Anatase
(200)
Ti4+
Brookite
(211)
Pd0
(111)
Pd0
(200)
Pd0
464.6 eV
PdO
336.2 eV
PdO
(002)
341.4 eV
(100)
ii) Pd/TiO2-calc-red
i) Pd/TiO2-calc
Pd2+
468 466 464 462 460 458 456 454
348 345 342 339 336 333 330
20 25 30 35 40 45 50
binding energy / eV
binding energy / eV
2 theta / °
Figure 3: XRD patterns (A) and XPS spectra (B and C) of Pd/TiO2 film with a Pd-loading of 2.5wt%. XRD patterns (A) show the film
after calcination in air for 20 min at 450 °C (bottom) and after reduction in H2/Ar stream for 6 h at 350 °C (top). Reflection position
assigned according to literature to TiO2 anatase (PDF 21 1272), Brookite (PDF 029-1360), tetragonal PdO (PDF 75-584) and
metallic Pd (PDF 46-1043). XPS spectra show the film after reduction for the Ti2p region (B) and the Pd3d region (C). Symbols
indicate experimental data, lines the corresponding peak deconvolution.
Page 28 of 30

Figure 4 - Influence of varied Pd loading (SEM)
0.5 wt% Pd / TiO2
1.5 wt% Pd / TiO2
2.5 wt% Pd / TiO2
A
B
C
100 nm
100 nm
100 nm
dPd-NP = 13.8 ± 3.7 nm
dPd-NP = 13.0 ± 3.0 nm
dPd-NP = 14.1 ± 4.5 nm
film surface:
D
E
F
100 nm
100 nm
100 nm
dPd-NP = 6.1 ± 1.2 nm
dPd-NP = 6.1 ± 2.6 nm
dPd-NP = 7.8 ± 2.3 nm
cross-section:
Figure 4. SEM analysis in top view in secondary-electron mode (A,B,C) and in cross-section view in z-contrast mode (D, E,
F) of the Pd/TiO2 catalytic layers with Pd-loading of 0.5wt% (A, D); 1.5wt% (B, E) and 2.5wt% (C, F) after calcination at
Page 29 of 30
450°C in air and reduction at 350°C in 4vol% H2/Ar.