Template-assisted hydrothermally synthesized iron-titanium binary oxides and their application as catalysts for ethyl acetate oxidation

Article abstract of DOI:10.1016/j.apcata.2016.09.006

Iron-titanium binary oxides were synthesized by template-assisted hydrothermal method. The effect of the variation in the Fe/Ti ratio and the temperature of hydrothermal treatment on the phase composition, structural, textural and redox characteristics of thus obtained materials was studied by Nitrogen physisorption, XRD, SEM, HRTEM, XPS, Raman, UV–vis, FTIR and Moessbauer spectroscopy and TPR with hydrogen. The contribution of various features of the samples on their catalytic behaviour in total oxidation of ethyl acetate as a representative VOC was in the focus of the discussion. The binary materials represent a complex mixture of pure and substituted anatase and/or hematite phases with different dispersion and proportion, which can be controlled by both the variation in the Fe/Ti ratio and the temperature of hydrothermal treatment. The stabilization of finely dispersed hematite-like particles in the vicinity of Fe³⁺ substituted in titania was considered as an important factor for the improvement of the catalytic activity and selectivity in total oxidation of ethyl acetate and it is favoured by the lower Fe/Ti ratio and the lower temperature of the hydrothermal treatment.

Full text of DOI:10.1016/j.apcata.2016.09.006

Accepted Manuscript
Title: Template-assisted hydrothermally synthesized
iron-titanium binary oxides and their application as catalysts
for ethyl acetate oxidation
Author: Tanya Tsoncheva Radostina Ivanova Momtchil
ˇ´
Dimitrov Daniela Paneva Daniela Kovacheva Jirı Henych Petr
´ˇ
Vomacka Martin Kormunda Nikolay Velinov Ivan Mitov
ˇ
Vaclav Stengl
PII:
S0926-860X(16)30470-7
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.09.006
APCATA 16002
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
16-6-2016
7-9-2016
13-9-2016
Please cite this article as: Tanya Tsoncheva, Radostina Ivanova, Momtchil
ˇ´
´ˇ
Dimitrov, Daniela Paneva, Daniela Kovacheva, Jirı Henych, Petr Vomacka,
ˇ
Martin Kormunda, Nikolay Velinov, Ivan Mitov, Vaclav Stengl, Template-
assisted hydrothermally synthesized iron-titanium binary oxides and their
application as catalysts for ethyl acetate oxidation, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2016.09.006
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.

Template-assisted hydrothermally synthesized iron-titanium binary oxides and their
application as catalysts for ethyl acetate oxidation
Tanya Tsoncheva1*, Radostina Ivanova1, Momtchil Dimitrov1, Daniela Paneva2, Daniela
Kovacheva3, Jiří Henych4, Petr Vomáčka4, Martin Kormunda5, Nikolay Velinov2, Ivan
Mitov2, Vaclav Štengl4
1Institute of Organic Chemistry with Centre of Phytochemistry, BAS, Sofia, Bulgaria,
tsoncheva@orgchm.bas.bg
2 Institute of Catalysis, BAS, Sofia, Bulgaria
3 Institute of general and Inorganic Chemistry, BAS, Sofia, Bulgaria
4
Materials Chemistry Department, Institute of Inorganic Chemistry AS CR v.v.i., 25068 Řež,
Czech Republic
5Faculty of Sciences, University of Jan Evangelista Purkyne, Ceske Mladeze 8, 400 96 Usti
nad Labem, Czech Republic

Graphical abstract
Phase composition
Specific catalytic activity per unit Fe
1
100
98
96
94
92
0,25
0,20
0,15
0,10
0,05
0,30
0,25
0,20
0,15
0,10
0,05
-12 -10 -8
-6
-4
-2
0
2
4
6
8
10 12
Velocity, mm/s
20
30
40
50
60
70
80
90
100
G Dbl, %
T 50% Conversion, K
8Fe2Ti_413
2Fe8Ti_413
650
5Fe5Ti_413
373 K
413 K
600
550
500
20
40
60
80
100
TiO2_373 2Fe8Ti_373
Fe2O3_373
5Fe5Ti_373 8Fe2Ti_373
Fe content, mol %
Highlights
 Template assisted hydrothermal synthesis of Fe2O3-TiO2 with different composition.
 Effect of texture and structure of binary oxides on total oxidation of ethylacetate.
 Finely dispersed Fe2O3 in vicinity of Fe3+ in TiO2 facilitates catalytic activity.
 Phase composition and catalytic behavior control by Fe/Ti and synthesis temperature.

Abstract
Iron-titanium binary oxides were synthesized by template-assisted hydrothermal method. The
effect of the variation in the Fe/Ti ratio and the temperature of hydrothermal treatment on the
phase composition, structural, textural and redox characteristics of thus obtained materials
was studied by Nitrogen physisorption, XRD, SEM, HRTEM, XPS, Raman, UV-Vis, FTIR
and Moessbauer spectroscopy and TPR with hydrogen. The contribution of various features
of the samples on their catalytic behaviour in total oxidation of ethyl acetate as a
representative VOC was in the focus of the discussion. The binary materials represent a
complex mixture of pure and substituted anatase and/or hematite phases with different
dispersion and proportion, which can be controlled by both the variation in the Fe/Ti ratio and
the temperature of hydrothermal treatment. The stabilization of finely dispersed hematite-like
particles in the vicinity of Fe3+ substituted in titania was considered as an important factor for
the improvement of the catalytic activity and selectivity in total oxidation of ethyl acetate and
it is favored by the lower Fe/Ti ratio and the lower temperature of the hydrothermal treatment.
Keywords
Iron-titanium binary oxides, hydrothermal synthesis, effect of Fe/Ti ratio and temperature of
hydrothermal treatment, total oxidation of ethyl acetate.
1. Introduction
Volatile organic compounds (VOCs) are environmental pollutants emitted by human
activities, vehicles and various industrial processes which are responsible for the formation of
smog and troposphere ozone [1 and refs therein]. Nowadays, the increasing political, social
and economic attention on the environment and the quality of life has enforced the strict
monitoring of VOCs emissions and the development of efficient technologies for their

elimination. Among them, the catalytic oxidation has been considered as promising
environmentally friendly and economically beneficial approach. Transition metal oxides have
been extensively employed as catalysts because they are cost efficient, less sensitive to
poisons and with lower sintering ability as compared to noble metals. Moreover, they can be
fabricated as high surface area porous materials and their preparation in the nanoscale
provokes the appearance of some outstanding properties [2-7]. Titania and hematite are both
oxides of wide interest for electronic devices and many industrial processes due to their
chemical and biological inertness, non-toxicity, strong oxidizing/reducing power, relatively
high melting point and cost-effectiveness [8-13]. It was reported that the essential
improvement of the structural, textural, optical and electro-conductivity characteristics of the
individual oxides could be achieved by their application in a dual system. Nowadays, iron and
titania mixed oxides gained a considerable interest as absorbers of electromagnetic radiation,
micro-transformers, protective systems, photocatalysts under UV and visible light irradiation
[5,12,14 and refs therein] and demonstrated catalytic activity in various processes, such as
oxidation of cyclohexane [15] and methanol [16], α-pirene isomerization [17], etc. Number
of studies have been aimed at clarifying the effect of doping concentrations, calcination
temperature and the preparation method used on the phase composition and the related
properties of iron-doped titania materials [9 and refs therein]. The enhanced catalytic activity
of binary oxides has usually been considered to be a result from the synergistic effect between
the Lewis acidic sites (Fe3+ and Ti4+) and the redox function of iron oxide species, the
mobility of oxygen in the binary oxide lattice or the increased specific surface area as
compared to the individual oxides [8]. It has been reported that the substitution of Ti4+ by Fe3+
in titania lattice is favoured due to the similarity in their ionic radii (0.65 nm for Fe3+ and 0.61
nm for Ti4+ at coordination number 6) [7, 16]. The substitution of Ti4+ ions with Fe3+ leads not
only to distortion of TiO6 octahedra but also to the appearance of strong charge imbalance

resulting in oxygen content deficiency [11]. Typically, the situated on adjacent sites Ti3+ ions
and oxygen vacancies were assumed as the dominant defects in iron substituted titania
materials [18]. Wang et al. [19] reported that the iron ions could be located at interstice
positions in titania lattice or could occupy some of the titanium sites with the formation of
iron-titania solid solution. Besides, the iron ions in the hematite lattice can also be substituted
by Ti4+ ions [20]. Nowadays, the limited solubility of Ti4+ in hematite and Fe3+ in titania is a
question of intensive debates [21]. Higher solubility of Fe3+ in anatase phase as compared to
rutile one was reported in [22]. For the xTiO2-(1-x)Fe2O3 system obtained under
hydrothermal conditions, the formation of solid solution isostructural with hematite was
registered for 0.0<x<0.3, while isostructural of anatase solid solution was observed for
0.6<x<1.0 [21]. Zhang et al. [23] and Wang et al. [24] reported the formation of homogeneous
FexTi1-xO2 solution when TiO2 was doped with 0.3-6% iron and segregation of α-Fe2O3 over
10 at.% of iron [24]. Just the opposite, Pal et al. [25] didn’t observe α-Fe2O3 even at 50 wt.%
iron in titania. Luthra et al. [26] demonstrated the formation of pseudobrookite phase when
the amount of titania was above the solubility degree. Formation of ilmenite, pseudorutile,
pseudobrookite or mixture of hematite and TiO2 with mutual solubility depending on the
Fe/Ti ratio and the calcination temperature used has been reported in [8, 9, 16, 20, 27, and
refs. therein]. The effect of titania doping on the phase transformation of anatase to rutile has
also been a matter of controversial discussion in refs. [4, 6, 10, 13, 20, 27-29].
The incorporation of iron species into titania was carried out by several methods, e.g. co-
precipitation, microemulsion procedures, combustion, sol-gel, hydrothermal/solvothermal,
chemical synthesis, plasma pyrolysis etc. [1, 6, 11 and refs therein], but up to now the
controlled synthesis of materials with well defined structure, morphology, texture and
dispersion by a simple and reproducible method is still a challenge. In the last two decades the
synthesis procedures of mesoporous silica with high surface area and uniform mesopores

using structure-directing agents has been extended to non-silica metal oxides [30-34 and refs
therein]. The main troubles with their synthesis evolve from the high reactivity of transition
metal oxides towards hydrolysis and condensation, the collapse of structure due to redox
reactions, the possible phase transformation and crystallization processes as well as to the
high sensitivity of the synthesis procedure to external parameters which deteriorates the
reproducibility [31 and refs therein]. Despite the number of studies on the synthesis and
characterization of mesoporous titania, the effect of iron substitution on the pore structure,
phase composition and catalytic behaviour, particularly in VOCs oxidation, is still
insufficiently reported [2, 4, 13, 21, 30, 35, 36]. The aim of the current investigation is to
study the effect of Ti/Fe ratio and the variation of the temperature of hydrothermal treatment
on the phase composition, structural and textural parameters of iron-titanium binary oxides
prepared by template-assisted hydrothermal procedure. Special attention is paid on the
elucidation of the role of different materials characteristics on their catalytic behaviour in
ethyl acetate oxidation, used as a representative VOC, as a reliable way for the optimization
of catalysts synthesis parameters.
2. Experimental
2.1. Materials
The iron-titanium oxides were synthesized using template-assisted precipitation technique
followed by hydrothermal treatment. In principle, 12.0 g N-hexadecyl-N,N,N-
trimethylammoniumbromide (CTAB) were dissolved in 100 ml distilled water. To this
solution was added slowly and under vigorous stirring a second solution of TiCl4 (5.54 g) or
FeCl3.6H2O (7.89 g) in 50 ml distilled water for the monocomponent samples and a mixture
of TiCl4 and FeCl3.6H2O in amounts corresponding to Fe:Ti mol ratio of 2:8, 5:5 or 8:2 and
total 0.0292 mols of metal in case of bi-component samples. Then, the temperature was raised

to 323 K and the reaction mixture was stirred for 30 min before adding drop-wise 20 ml
NH4OH (25 %). The resulting mixture was stirred overnight at 323 K. Then it was transferred
into a polypropylene container and treated at 373 or 413 K for 24 h. The so-prepared particles
were filtrated, washed with distilled water, dried at room temperature and calcined up to 773
K with a ramp of 1 ºK/min and dwelling time of 10 h at the final temperature. The samples
were denoted as xFeyTi_T, where x/y was the Fe/Ti mol ratio and T was the temperature of
hydrothermal synthesis.
2.2. Methods of investigation
Specific surface area (ABET) and pore size distribution were measured through nitrogen
adsorption-desorption isotherms at 77 K using Beckman Coulter SA 3100 apparatus. ABET
was calculated applying the Brunauer, Emmet and Teller (BET) equation for N2 relative
pressure within the following range 0.05 < P/P0 < 0.30. Pore size distribution was determined
by NLDFT method. Powder X-ray diffraction patterns were collected within the range of 10-
80° 2θ with a constant step of 0.02° 2θ and counting time of 17.5 s per step on a Bruker D8
Advance diffractometer with Cu Kα radiation and using a LynxEye detector. Mean crystallite
size were determined with the Topas-4.2 software package using the fundamental parameters
of peak shape description including appropriate corrections for the instrumental broadening
and diffractometer geometry. For this purpose the integral line breadth approach for the
generalized treatment of the domain size broadening – βi=λ/Lvolcosθ was employed, where βi
is the integral breadth of the diffraction line i and Lvol is the volume weighted mean column
height. The UV–Vis spectra were recorded on a Jasco V-650 UV-Vis spectrophotometer
equipped with a diffuse reflectance unit. FTIR spectra in the region of skeletal vibrations were
recorded on a Bruker Vector 22 spectrometer at a resolution of 1–2 cm−1, accumulating 64–
128 scans and KBr pellets technique. Raman spectra were acquired with a DXR Raman
microscope (Thermo Fischer Scientific, Inc., Waltham, MA) using a 780 nm laser. The X-ray

photoelectron spectroscopy (XPS) analyses were measured in high-vacuum chamber
equipped with SPECS Xray XR-50 (Al cathode 1486.6 eV) and SPECS PHOIBOS 100
Hemispheric Analyzer with a 5-channels detector. The background pressure in XPS system
was under 2.5x10-9 mbar. The XPS survey-scan spectra were recorded at pass energy of 40
eV; the energy resolution was set to 0.5 eV. The individual high-resolution spectra were taken
at pass energy of 10 eV with 0.05 eV energy steps. CasaXPS software was used to fit the
high-resolution multi-components peaks. The Shirley background model was used. The atom
concentration of the compounds was calculated with RSF defined in the standard table of
CasaXPS. Before analyses the binding energy of C-C component in C 1s peak was set to
284.0 eV the expected precisions in BE are about 0.2 eV because the C1s peaks were with
low intensity they have low RSF factor of 1.
The Moessbauer spectra were obtained at room temperature and liquid nitrogen temperature
(LNT) with a Wissel (Wissenschaftliche Elektronik GmbH, Germany) electromechanical
57
spectrometer working in a constant acceleration mode. A Co/Rh (activity y25 mCi) source
and a-Fe standard were used. The experimentally obtained spectra were fitted by the least
square method.
The SEM images were collected on FEI Nova NanoSEM450 in a High-Vacuum mode using
acceleration voltage 5kV, samples were dispersed on a conductive adhesive carbon discs. The
morphology of the samples was measured by transmission electron microscopy (TEM) using
a FEI Talos F200X microscopes. As a specimen support for TEM investigations a
microscopic copper grid covered by a thin transparent carbon film was used. The powdered
sample was dispersed in ethanol and the suspension treated in an ultrasonic bath for 10 min
before deposition on the grid.

The TPR/TG (temperature-programmed reduction/thermo-gravimetric) analyses were
performed on a Setaram TG92 instrument in a flow of 50 vol% H2 in Ar (100 cm3 min-1) and
a heating rate of 5 K min-1.
2.3. Catalytic tests
The catalytic oxidation of ethyl acetate was performed in a flow type reactor (0.030 g of
catalyst) with a mixture of ethyl acetate in air (1.21 mol %) and WHSV of 100 h-1. Before the
catalytic experiments the samples were treated in argon at 373 K for 1 h. The experimental
data were obtained under temperature-programmed regime in the range of 473-773 K. The
catalyst behaviour was studied also with time on stream experiments under isothermal
conditions at selected temperature (670 K). Gas chromatographic analyses were done on a HP
5890 apparatus using carbon-based calibration. The products distribution was calculated as
CO2 (SCO2), acetaldehyde (SAA), ethanol (SEt) and acetic acid (SAcAc) selectivity by the
equation: Si= Yi/X*100, where Si and Yi were the selectivity and the yield of (i) product and
X was the conversion. For more precise comparison, the conversion was normalized to unit
surface area (SA=X/A, where X was the conversion at 600 K and A was the specific surface
area of the corresponding sample) and unit iron content (SAM= SA/M, where M was the mol
content of iron in the samples).
3. Results and discussion
3.1. Catalytic properties in ethyl acetate oxidation
The temperature dependencies of ethyl acetate conversion and CO2 selectivity for various iron
and titanium oxide materials are demonstrated in Fig.1 and the products distribution at 50%
conversion are listed in Table 1. Typically, the oxidation of ethyl acetate was observed above
550 K (Fig.1a,c) and 80-100% conversion was registered at 650-700 K. As compared to the

data reported in the literature (Table S1), the iron promoted titania exhibited above the
average level of activity in total oxidation of EA. Moreover, all binary materials represented
higher catalytic activity as compared to the individual oxides and this is well demonstrated by
the lower temperatures at which 50% conversion was achieved (Fig. S1). Besides, the
hydrothermal synthesis of the samples under milder conditions facilitated their catalytic
activity (Fig.1c, Fig.S1). The extremely high selectivity to acetaldehyde, ethanol and ethane
for pure TiO2 (Table 1) can be attributed to the simultaneous activity of acidic and relatively
weak redox sites following the mechanisms described in refs. [37, 38]. The improved
selectivity to CO2 for all iron containing materials indicates the presence of stronger redox
sites in them. The variation in the products distribution with the changes in Fe/Ti ratio (Table
1) evidences differences in the nature and/or the environment of the catalytic sites. Note the
extremely low amount of acetic acid among the products for all samples. This could be
attributed to its rapid oxidation to CO2 as compared to the others by-products (ethanol and
acetaldehyde), and probably occurs by a nucleophilic attack with oxygen from the solid lattice
following the Mars van Krevelen mechanism [37]. The absence of clear dependence of the
catalytic behavior of the samples on their composition indicates that the samples are not a
simple mixture of the individual oxides. In Figs.S2 and S3 is demonstrated the evolution of
catalytic behaviour of the samples with time on stream at selected temperature (670 K). The
catalytic properties of pure Fe2O3 remained almost unchanged while well defined tendency to
a decrease in the catalytic activity and CO2 selectivity was observed for pure TiO2. Oxidative
treatment (under air) up to 873 K of the TiO2 sample after the reaction demonstrated about
3% weight loss, which could be due to the release of hardly desorbable carbon based by-
products (coke precursors). Under isothermal conditions the catalytic activity for the binary
materials was either constant or increased with time on stream and we shall return to this
again in Section 3.4.

3.2. Structural characterization
3.2.1. Nitrogen physisorption
Nitrogen physisorption isotherms and pore size distribution for all Fe-Ti oxide materials are
presented in Fig. S4 and the corresponding textural parameters are listed in Table 2. With the
exception of Fe2O3, all isotherms were characterized with well pronounced step at about 0.7-
0.8 relative pressure due to capillary condensation of nitrogen into the pores and a hysteresis
loop of H2 type, which according to IUPAC classification is typical of mesoporous materials
with “ink-bottle” shape of the mesopores. When the hydrothermal synthesis was carried out at
373 K, the addition of small amount of iron to titania (2Fe8Ti_373) slightly increased the
BET surface area and provoked the formation of narrower and more uniform mesopores with
the preservation of pore volume. In agreement with Wang et al. [30] this could be due to
limited grain growth during heating as a result of incorporation of iron ions into titania lattice.
The observed decrease in the specific surface area and total pore volume with the increase of
iron content in the samples evidences structural changes in the samples (Table 2). A decrease
in the BET surface area and total pore volume due to the formation of larger and non-uniform
mesopores was observed for the samples obtained at higher temperature of hydrothermal
treatment (413 K). In order to elucidate the contribution of textural parameters on the catalytic
behaviour of the samples we calculated the specific catalytic activity (SA) as conversion per
unit BET surface area (Table 1). Note the high SA value for pure Fe2O3 which clearly
indicates the decisive role of iron state on the catalytic properties of the samples. The
observed changes in the SA with the variation of Fe/Ti ratio and the temperature of
hydrothermal treatment evidence a complex dependence of the catalytic behaviour of the
samples on their phase composition and textural characteristics.

3.2.2. XRD
In Fig. 5 are presented the XRD patterns of all iron and titanium oxide samples. The XRD
pattern of Fe2O3_373 contained the reflections typical of rhombohedral α-Fe2O3 at 24.2o,
o
33.2o, 35.8o, 41.9o, 49.5 , 62.5o and 63.9o 2θ (PDF 01-089-0599). The results from the
diffraction line broadening analysis showed that the hematite crystallites were relatively large
(about 108 nm). This provoked low surface area and pore volume for this material (Table 2).
The XRD pattern of pure TiO2 represented intensive reflections at 25.6o, 38.3o, 47.8o and
62.5o 2θ corresponding to the diffractions of (101), (004), (200) and (204) crystallite faces of
tetragonal body centered anatase phase (PDF 00021-1272) with average crystallite size of 17
nm (Table 3). Predominant formation of anatase phase, which under ambient conditions is
thermodynamically less stable than rutile [39], was also registered during hydrothermal
synthesis in ref. [21]. After the catalytic test, only a slight increase in the crystallite size of
titania (to about 19 nm) was detected, which could also be responsible for the observed
decrease of its catalytic activity (Fig. S2). Among the binary materials hydrothermally treated
at 373 K, a decrease in the intensity and broadening of the reflections typical of anatase was
observed for 2Fe8Ti_373, and especially for 5Fe5Ti_373. The observed slight increase in the
lattice parameters as compared to pure titania indicates that the inhibited agglomeration of
TiO2 crystals is due to substitution of Fe3+ ions in titania lattice [11]. The additional
reflections at 18.1o and 32.5o 2θ are well fitted with Fe3Ti3O10 orthorhombic iron titanium
oxide (PDF 00-047-0421). In agreement with Diamandescu et al. [21] the segregation of this
phase could be due to limited solubility of Fe in titania. This was also supported by the
appearance of reflections, typical of hematite phase, which became more intensive and
narrower with the increase in the Fe/Ti ratio for 5Fe5Ti_373. Note also the slightly lower unit
cell parameters as compared to pure hematite, especially for 2Fe8Ti_373, which could be
assigned to substitution of Ti4+ in the segregated hematite phase [20]. This trend to formation

and growth of hematite crystallites was clearly demonstrated with the further increase of the
Fe/Ti ratio (sample 8Fe2Ti_373). The observed slight increase in the dispersion of the
existing hematite phase after the catalytic test (Table 3) could be responsible for the increase
in the catalytic activity under isothermal conditions (Fig. S2). This evolution was especially
well pronounced for the samples with higher iron content. It could be due to the incorporation
of Ti4+ in the surface layer of Fe2O3 and its further progress through this layer into the bulk of
hematite particles [40]. Note that the increase in the temperature of hydrothermal treatment
provided a decrease in the ability for crystallization of Fe3T3O10 phase. It was combined with
growth in the anatase and hematite crystallites for 2Fe8Ti_413, and vice versa for the samples
with higher Fe/Ti ratio. This indicates that the substitution of Ti4+ in hematite and Fe3+ in
titania is controlled in a complex way by the temperature of hydrothermal synthesis and Fe/Ti
ratio. However, we should take into consideration that X-ray diffraction technique is not
particularly sensitive to elucidate the distribution of Ti4+ and Fe3+ in dual oxides, which is
generally assigned to the similarity in their electron configuration [40].
3.2.3. SEM and TEM
Surface morphology and particles size distribution were studied by SEM (Fig. 3) and TEM
(Fig. 4) techniques. Pure Fe2O3 crystallized in the form of cubic grains with average diameter
of 50-60 nm, while TiO2 demonstrated rough structure of smaller (about 10 nm) crystallites
(Fig.3). Similar structure was observed for 2Fe8Ti_373. According to HRTEM analyses
(Fig.4), the sample consisted predominantly of about 6-12 nm nanocrystallites with interlayer
spacing (d) of about 0.31 nm or 0.23 nm which can be assigned to (101) and (110) plane of
distorted anatase and hematite cells, respectively. With the increase of Fe/Ti ratio (samples
5Fe5Ti_373 and 8Fe2Ti_373) a progressive disappearance of particles with titania-like

morphology and crystallization of irregular in their shape hematite-like particles was observed
(Fig. 3).
3.3. Spectroscopic study
3.3.1. FTIR
In Fig. 5 are presented the FTIR spectra of iron-titanium oxide materials. The broad peak at
ca. 3600-3400 cm-1 which is observed in the spectra of all titanium containing samples can be
assigned to stretching vibration mode of tetrahedrally coordinated vacancies of Ti4+-OH type
[11, 41]. The corresponding bending vibration mode was observed at 1600 cm-1. The
absorption bands at about 3700 and 3850 cm-1 are generally assigned to Ti3+-OH octahedral
vacancies. These bands revealed formation of point defects after doping of anatase with iron
[41] and their intensity varied with Fe/Ti ratio and the temperature of hydrothermal synthesis.
The broad band about 900-600 cm-1 is assigned to bending vibrations of Ti-O-Ti unit in
titania [11]. The appearance of strong plasmon mode in the spectrum of pure TiO2 at ca. 450
cm-1 is typically assigned to the existence of great number of free carriers, introduced by
nonstoichiometric defects, like Ti3+ or Ti4+ interstitials, or O2- vacancies [40]. This mode was
especially well pronounced and also shifted to higher wavenumbers in the spectrum of
2Fe8Ti_373 which in accordance with XRD data (Table 3) could be due to the better
incorporation of Fe ions in the anatase lattice. The appearance of new absorption mode at ca.
960 cm-1 is generally assigned to Fe-O-Ti linkage [16]. The bands at ca. 457 and 562 cm-1 in
the FTIR spectrum of Fe2O3 are typical of hematite. The appearance of similar bands for all
binary materials indicates the segregation of hematite phase [29, 31]. The intensity of these
bands increases with the increase of Fe/Ti ratio which is also in accordance with the XRD
data (Table 3). The observed slight shift of the main hematite bands to lower wavenumber and
the appearance of additional peak at ca. 415 cm-1 for all binary oxides do not exclude the
incorporation of Ti4+ ions in it, as was proposed on the base of the XRD measurements [43].

3.3.2. Raman spectroscopy
In Fig. 6 are presented Raman spectra of all Ti- and Fe oxide materials. The Raman modes in
the spectrum of pure TiO2 at ca. 143 cm-1 (Eg), 198 cm-1 (Eg), 397 cm-1 (B1g), 514 cm-1 (A1g),
519 cm-1 (B1g) and 638 cm-1 (Eg) are characteristic of the anatase phase [42]. The absence of
Raman bands at 445 cm-1 and 320 cm-1 that are typical of rutile and brookite, respectively,
evidences the purity of the anatase phase. The typical bands of anatase at 638 cm-1, 514 cm-1
and 399 cm-1 still remained in the spectra of both 2Fe8Ti samples. The band at 638 cm-1
became broader and less intensive which could be due to the increase in oxygen vacancies as
well as to a decrease in the crystallite size, the latter being in consistence with the XRD
analysis [11, 43]. The observed blue shift of the maximum of the main band at 143 cm-1 after
anatase doping with iron indicates structural changes in anatase, which are probably provoked
by the formation of oxygen vacancies [44,45]. The Raman bands at 203 cm-1 and 397 cm-1 are
typically assigned to asymmetric stretching and symmetric bending vibrations of Ti-OH. All
these features became less distinguished in the spectra of 5Fe5Ti and 8Fe2Ti. The appearance
of bands at 218 cm-1 and 278 cm-1, which also existed in the spectrum of pure Fe2O3_373,
indicates the presence of hematite phase. The band at ca. 192 cm-1 which could be
distinguished in all spectra is probably due to wagging vibrations of O-Ti-O bonds of Ti2O3
[45].
3.3.3. UV-Vis
The UV-Vis diffuse reflectance spectra of the samples are shown in Fig.7. The absorption
threshold of pure titania was detected at ca. 340 nm. It is lower than the reported in the
literature (380 nm) [46] and, in accordance with SEM and TEM analyses, indicates high
dispersion of titania. Iron doping resulted in absorption edge shifting towards higher

wavelengths which could be considered as superposition of two absorption bands at 408 nm
and 550 nm. Typically these bands are related to the energy absorption of the existing 3d
electrons from Fet2g energy level to the Ti(3d) conduction band (charge transfer transition)
and to d-d transition of Fe3+(2T2g→2A2g,2Ti1g) or charge transfer transition between
interacting iron ions (Fe3++Fe3+→Fe2++Fe4+), respectively [47, 48].
3.3.4. XPS
In Figs. 8 and S5 are presented Ti2p, Fe2p and O1s core level spectra of pure TiO2 and Fe2O3
and selected binary oxides. All peak positions were referenced to the C 1s core level for C-
C/C-H components, which was taken to be at 284.5 eV (not shown). The samples
composition was calculated from the high resolution spectra and the data are listed in Table 4.
All Ti 2p core level spectra represented two intensive peaks situated at 458.4 eV and 464.2
eV. The binding energies of Ti2p3/2, which have been generally reported in the literature,
varied due to the calibration process, charging effects and chemical surrounding effects [49-
53]. Here, the Ti2p3/2 spectra were fitted according to the model reported in [50] where the
spin split was 5.7 eV and the Ti2p3/2 components were 458.5 eV for TiO2, 457.5 eV for Ti3O5
and 455.9 eV for Ti2O3. The data from the deconvolution of Ti2p3/2 spectra showed domination
of Ti species with BE similar to the one found over TiO2, but also presence of Ti species with
BE corresponding to Ti2O3 and Ti3O5 (Table 4). The amount of the latter components was
higher for the binary materials, especially for the sample with lower Fe/Ti ratio. There are
also some possible contributions from selected background model (Shirley) used for XPS
analyses. However, the lack of changes in the position of the main Ti 2p peak urged the
authors to assume limited solubility of iron in titania which contributed to small increase in
the sub stoichiometric levels. The Fe 2p peaks (Fig.8, S5) at 711.0 and 724.8 eV BE and the
satellite peak at 719.0 eV confirm predominant presence of Fe3+ in all materials [12]. But the

Fe 2p peak of Fe2O3 sample was well fitted with the models for the presence of both Fe2O3
and FeO and the calculated Fe2O3/FeO ratio was 72/28. The formation of Fe2+ was also
described and discussed in [16, 54]. The observed surface Fe/Ti ratio (Table 4) was almost
similar to the nominal one for Fe2Ti8_373 sample and exceeded it for Fe8Ti2_373 K. This
indicates almost uniform distribution of Fe in titania lattice for the former material and higher
exposure of Fe to the surface for the latter probably due to the segregation of hematite phase
(see XRD and Raman data). The observed O 1s photoelectron spectra exhibit a single peak
with BE in the interval 529.9 eV±0.1 eV (Fig. 11). Due to the asymmetry, the O 1s peaks
were deconvoluted to two components which were assigned to lattice oxygen and hydroxyl
species (Table 4). The lattice oxygen was not attributed to individual oxides as their binding
energies are in wide range in the literature, and it would be speculation. The hydroxyl species
were detected in high amount (about 22 % of the whole oxygen content) for pure Fe2O3. On
the contrary, the contribution of hydroxyl groups was about 6.2 % for pure TiO2 and their
formation could be assigned to the interaction of surface oxygen vacancies with water
molecules. For the binary materials, the amount of hydroxyl groups slightly increased with
the increase of Fe/Ti ratio (Fig.8c, S5c). In accordance with the XRD and Raman data (see
above) this could be provoked by the increase in the number of oxygen vacancies in titania as
well as by the segregation of hematite phase.
3.3.5. Moessbauer spectroscopy
In Fig. 9 are presented 57Fe Moessbauer spectra of selected samples and the corresponding
parameters of all materials are listed in Table 5. The spectra were well fitted with two sextets
(Sx1 and Sx2) and two doublets (Db1 and Db2). Only the Moessbauer spectrum of pure
Fe2O3 was fitted with one sextet (named Sx1) with IS of 0.38 mm/s, Heff=51.5 T and negative
QS which could be assigned to hematite with crystallite size above 13.5 nm and this
corresponds well with the data from XRD analysis. No significant changes with this sample

were observed after the catalytic test (Table 5). This sextet component (Sx1) was also
observed in the spectra of all binary materials with the exception of 2Fe8Ti_373. Moreover,
its relative part increased with Fe/Ti ratio increase and also with the increase of the
temperature of hydrothermal treatment. The quadrupole splitting of the doublets was found to
be 0.55-0.61 mm/s (Db1) and 0.95-1.06 mm/s (Db2). The isomer shift for both doublets was
0.32-0.37 mm/s. These data indicate presence of high spin Fe3+ ions in different environment
[11, 55, 56]. However, the interpretation of the doublet part in the literature has been rather
controversial. Sharma et al. [57] considered the higher QS value for Db2 as indication of
higher number of crystal defects. Cordishi et al. [22] and Hirano et al. [58] reported that the
doublet with QS=1.15 mm/s could be assigned to substitutional FexTi1-xO2 and that one with
QS=0.64 mm/s was typical of pseudobrookite Fe2TiO5. The latter was also assumed in [59].
Zhu et al. [9] also assigned the doublet with QS=1.13 mm/s to substitutional FexTi1-xO2 phase,
but the doublet with QS=0.71 mm/s was ascribed to α-Fe2O3 particles, which sizes were
bellow the critical one (13.5 nm) for the transition of ferromagnetic to SPM behaviour.
Similar interpretation of both doublets was reported in [21]. Guo et al. [59] fitted the
Moessbauer spectra of Fe1+xTi2−xO5 solid solution with two doublets with IS=0.4 mm/s and
QS of 0.54-0.58 mm/s and IS=0.3 mm/s and QS of 1.05 and 0.88 mm/s. Here we observed
that the doublet part of the spectra increased with the decrease in Fe/Ti ratio. In consistence
with the XRD and Raman data (Sections 3.2.2. and 3.3.3.), this could be provoked by the
substitution of Fe ions in titania lattice. In order to precise this assumption, low temperature
(LNT) Moessbauer spectra of selected materials were collected. As a whole, the doublet
component decreased in the LNT Moessbauer spectra, which could be due to the presence of
finely dispersed hematite particles with grain size in the range of 4-13.5 nm. However, the
preservation of the doublet components in the LTN spectra in a high extent does not exclude
the presence of ultrafinely dispersed hematite particles with crystallite sizes bellow 4 nm as

well as Ti-containing phases. In accordance with the XRD and Raman data (see above) they
could be substituted FexTi1−xO2 and/or iron-titanium oxide Fe3Ti3O10. In any case, we can
assume that precise quantitative elucidation of the distribution of Fe3+ in the ultrafinely
dispersed hematite particles and various Ti containing phases only on the base of the analyses
of the Db1/Db2 ratio could be rather inaccurate. The second sextet component (Sx2), which
appeared in all binary materials, is also difficult for interpretation. Its relative part was lowest
for 2Fe8Ti_373 and increased with Fe/Ti ratio and, in most cases, with the temperature of
hydrothermal synthesis. Similarly to Sx1, this component exhibited negative quadrupole
splitting but lower mean magnetic hyperfine field (Table 5). In agreement with the data
reported in ref. [58] this could be assigned to the presence of hematite particles in which Fe3+
ions are partially substituted with non-magnetic Ti4+ ions. For these ones, no Morin transition
in the low temperature Moessbauer spectra have been confirmed [55]. Indeed, in the low
temperature (LNT) spectrum of 5Fe5Ti_373 we observed only slight decrease in the Sx2
component, which evidences that the main part of this component is attributed to Ti-
substituted hematite. The observed significant increase in the Sx2 component in the LNT
Moessbauer spectrum of Fe2Ti8_413 at the expense of the doublet component, urged the
authors to assume the contribution of hematite particles with higher dispersion and defectness
in it. Thus, the Moessbauer analysis clearly indicated that the iron in all materials is in the
Fe3+ oxidative state. It is included either in pure hematite particles with different dispersion or
in binary phases on the base of anatase and hematite structures, which proportion can be
controlled both by the variation in the Fe/Ti ratio and temperature of hydrothermal synthesis.
The absence of significant changes with the Moessbauer spectra after the catalytic test (Table
5) evidences the stability of the binary materials under the influence of the catalytic medium.
3.4. TPR measurements and concluding remarks

In Fig. S6 are presented TPR-TG and TPR-DTG profiles of the studied materials. The TPR
process for pure Fe2O3 was initiated at 612 K. The observed two TPR effects with maxima
centered at 648 K and 748 K could be assigned to step-wise reduction of Fe2O3 to Fe [60].
The calculated reduction degree in the whole temperature interval was 71%. The weight loss
for pure TiO2, which was registered just above 720 K, corresponded to 7% reduction of Ti4+
to Ti3+. The main reduction effect for all binary materials appeared as a single and relatively
broad peak. Here, the reduction process was initiated at lower temperatures (553-593 K) than
for the pure Fe2O3 and the reduction degree was more than twice lower if the samples were
considered as mechanical mixtures of the individual oxides. Despite the lack of any simple
relation between the TPR features of the binary materials and their characteristics (Fe/Ti ratio,
temperature of hydrothermal synthesis), the observed results clearly indicate simultaneous
presence of easily reducible and non-reducible under the experimental conditions iron-
containing particles. The combined analyses of the TPR measurements and the data from the
other physicochemical techniques (see above) indicated reduction of hematite particles with
different dispersion. Due to the shortening of the Fe-O bond in the vicinity of Ti4+ ion, the
reduction is expected to be more difficult for the Fe3+ ions included in the binary oxide
structures such as titania substituted with Fe3+ or hematite substituted with Ti4+ [60, 61].
Remember, that on the base of the calculated specific catalytic activity (SA, Table 1) we
assumed that the observed higher catalytic activity in the binary oxides is not in a simple
dependence on the improved textural characteristics, and the state of the iron in the samples is
of primary importance. In order to ignore the effect of different amount of iron in various
materials we calculated the specific catalytic activity per unit Fe content (SAM, Table 1) and
plotted it versus the Fe/Ti ratio for various temperatures of hydrothermal synthesis (Fig.10).
For easier discussion, we also plotted the changes of SAM values versus the relative weight of

the doublet part in the Moessbauer spectra for various binary materials (Fig.10, inset). Several
important features were observed:
(i) With the exception of 2Fe8Ti_373, the SAM value for pure Fe2O3 was higher than that one
for all binary materials. XRD and Moessbauer analyses (Sections 3.2.2. and 3.3.5) showed the
presence of well crystallized and relatively large hematite particles. Thus the observed low
catalytic activity of the latter sample (Figs.1, 2) was obviously provoked by its much lower
specific surface area (Table 2) which can be improved using titania as a structure promoter.
(ii) The lower SAM values for the binary materials (Fig. 10) clearly demonstrate that part of
iron species is not active. In accordance with the TPR results (Section 3.4) and the data
reported by Chen at al. [33,34] these iron species could be hardy reducible Fe3+ ions, situated
in the vicinity of Ti4+ in the substituted phases. The presence of such species renders difficult
the realization of Mars van Krevelen mechanism.
(iii) Well defined tendency of an increase in SAM values with a decrease in Fe/Ti ratio is
observed. The analyses of Moessbauer data (Fig. 10, inset) indicate that this effect is
accompanied by an increase in the doublet part in the spectra. Thus the observed extremely
high SAM value for 2Fe8Ti_373 evidences the highest catalytic activity for the ultrafinely
dispersed hematite particles with superparamagnetic behaviour. The deviation from the
simple linear correlation between the relative weight of doublet component and SAM values
for various binary samples (Fig. 10, inset) are in accordance with our assumption that these
ultrafinely dispersed hematite particles partially contributed to the doublet part of the spectra.
On the base of XRD, Raman and XPS data (Sections 3.2.2-3.2.4) we propose their formation
in the vicinity of defects originating from the substitution of Fe3+ in titania lattice. It seems
that the amount of these active species is controlled by the limited solubility of Fe3+ in titania
which provokes segregation of bulk hematite particles (pure or substituted with Ti4+,
corresponding to Sx1 and part of Sx2 components in Moessbauer spectra, respectively, Table

5). We also do not exclude the appearance of synergistic effect between these ultrafinely
dispersed hematite-like species and titania support. The analysis of the data from the catalytic
test (Section 3.1.), particularly from the selectivity to various intermediate products (Table 1),
demonstrates hydrolysis of EA molecules on the Lewis and/or Broensted sites of titania and
further oxidation of the obtained products (Et, AcAc) via Mars-van Krevelen mechanism after
their “spillover” on hematite-like species. It seems that this synergistic effect between titania
and hematite facilitates the oxidation of the intermediate products of EA decomposition
(especially production of ethene) which prevents the catalysts against coke deposition and
ensures higher stability of the catalysts (Figure S2). The lack of significant changes in the
phase composition of the binary materials after the catalytic test (Table 5) urges the authors to
assume also stabilization of metal oxide phases against agglomeration. On the base of these
results, the observed increase in the catalytic activity with time on stream for the binary
materials (Fig. S2) could not simply be related to structural changes in the catalysts under the
oxidative reaction medium. The XPS and Raman data (see above) indicate that it could be
provoked by the formation of additional catalytic active sites due to changes in the surface
properties of the solid as a result of reoxidation of Ti3+ to Ti4+, generation of additional
hydroxyl groups by the interaction with oxygen defects, etc. The incorporation of Ti4+ in
hematite, which was facilitated for the materials with higher Fe/Ti ratio, decreases the
mobility of oxygen in the latter materials and contributes to lower catalytic activity via Mars-
van-Krevelen mechanism (Fig.10).
(iv) The increase of the temperature of hydrothermal synthesis resulted in a decrease in the
SAM values and again, this was controlled by the Fe/Ti ratio of the samples (Fig. 10). The
data from the XRD (Table 3) and Moessbauer analyses (Table 5, Fig.10) demonstrate that the
higher temperature facilitated the segregation of hematite phase. This restricted the
incorporation of Fe3+ in titania and promotes the insertion of Ti4+ ions in hematite lattice

(Section 3.2.2) which results in a decrease in SAM values. The nitrogen physisorption data
(Table 2) demonstrate a decrease in the BET surface area with the increase in the temperature
of hydrothermal synthesis which also results in a decrease of the overall catalytic activity
(Fig.1).
4. Conclusion
The phase composition of iron-titanium binary oxides prepared by template-assisted
hydrothermal procedure depends on the Fe/Ti ratio of the initial precursors and the
temperature of the hydrothermal synthesis. The lower Fe/Ti ratio and lower temperature of
hydrothermal treatment improves the textural parameters of the obtained binary materials.
This is controlled by the favourable substitution of Fe3+ in anatase phase which provokes the
crystallization of finely dispersed hematite-like species in the vicinity of the formed crystal
defects. These finely dispersed hematite-like particles possess high catalytic activity,
selectivity and stability in total oxidation of ethyl acetate probably not only due to the size
determined effect but also due to their synergistic activity with titania. The formation of these
highly active hematite-like species is controlled by the limited substitution of Fe3+ in anatase.
The increase in both Fe/Ti ratio and temperature of hydrothermal synthesis results in the
crystallization of larger, partially substituted with Ti4+ hematite particles, which decreases the
catalytic activity. The variation in the Fe/Ti ratio and the temperature of hydrothermal
synthesis was considered as reliable simple approach for the optimization of catalyst
properties.
Acknowledgements.
Financial support of project DFNI E02/2/2014, and Joint research project between ASCR and
BAS is acknowledged. The authors also acknowledge the assistance provided by the Research
Infrastructure NanoEnviCz, supported by the Ministry of Education, Youth and Sports of the
Czech Republic under Project No. LM2015073.

References
[1] X. Chen, S. A. C. Carabineiro, S.S.T. Bastosa, P. B. Tavares, J. J. M. Órfão, M. F. R.
Pereira, J. L. Figueiredo Appl. Catal. A: Gen., 472 (2014), pp. 101–112
[2] C. J. Lin, Y. H. Liou, Y. Zhang, C. L. Chen, C. L. Dong, S.Y. Chen, G. D. Stucky Appl.
Catal. B: Environ., 127 (2012), pp. 175– 181
[3] J. Cui, W. He, H. Liu, S. Liao, Y. Yue Coll. Surf. B., 74 (2009), pp 274–278
[4] M. Zhou, J. Yu, B. Cheng, H. Yu Mater. Chem. Phys., 93 (2005), pp. 159–163
[5] M. Qamar, B. Merzougui, D. Anjum, A.S. Hakeem, Z.H. Yamani, D. Bahnemann Catal.
Today., 230 (2014), pp. 158–165
[6] D. Kotadia, S. Soni Appl. Catal. A: Gen., 488 (2014), pp. 231–238
[7] K. M. Parida, G. K. Pradhan Mater. Chem. Phys., 123 (2010), pp. 427–433
[8] A. Khaleel Coll. Surf. A., 346 (2009), pp. 130–137
[9] S. Zhu, Y. Li, C. Fan, D. Zhang, W. Liu, Z. Sun, S. Wei Physica B., 364 (2005), pp. 199–
205
[10] S. Zhu,W. Liu, C. Fan, Y. Li Hyperfine Interact., 165 (2005), pp. 273–278
[11] S. D. Delekar, H. M. Yadav, S. N. Achary, S. S. Meena, S. H. Pawar Appl. Surf. Sci.,
263 (2012), pp. 536–545
[12] Z. Ambrus, N. Balazs, T. Alapi, G. Wittmann, P. Sipos, A. Dombi, K. Mogyorosi Appl.
Catal. B: Environ., 81 (2008), pp. 27–37
[13] S.K. Giri, N.N. Das Powder Technol., 239 (2013), pp. 193–198
[14] V.S. Rudnev, M.V. Adigamova, I.V. Lukiyanchuk, I.A. Tkachenko, V.P. Morozov J.
Alloys Compd., 618 (2015), pp. 623–628
[15] N. Perkas, Y. Koltypin, O. Palchik, A. Gedanken, S. Chandrasekaran Appl.Catal. A:
Gen., 209 (2001), pp. 125–130
[16] C. T. Wang, S. H. Ro Mater. Chem. Phys., 101 (2007), pp. 41–48

[17] G. Neri, G. Rizzo, S. Galvagno, G. Loiacono, A. Donato, M.G. Musolino, R.
Pietropaolo, E. Rombi Appl. Catal. A: Gen., 275 (2004), pp. 243–251
[18] H. Liu, H.T. Ma, X.Z. Li, W.Z. Li, M. Wu, X.H. Bao Chemosphere., 50 (2003), pp. 39–
46
[19] J. A. Wang, R. Limas-Ballesteros J. Phys. Chem. B., 105 (2001), pp. 9692–9698
[20] M.V. Nikolic, D. L. Sekulic, N. Nikolic, M. P. Slankamenac, O.S. Aleksic, H.
Danninger, E. Halwax, V. B. Pavlovic, P. M. Nikolic Sci. Sintering., 45 (2013), pp. 281–292
[21] L. Diamandescu, M. Feder, D. Tarabasanu-Mihaila, F. Vasiliu Appl. Catal. A: Gen., 325
(2007), pp. 270–275
[22] D. Cordishi, N. Burriesci, F. D’Alba, M. Petrera, G. Polizzotti, M. Schiavello J. Solid
State Chem., 56 (1985), pp. 182–190
[23] Y. H. Zhang, A. Reller J. Mater. Chem., 11 (2001), pp. 2537–2541
[24] C. Wang, Q. Li, R. D. Wang J. Mater. Sci., 39 (2004), pp. 1899–1902
[25] B. Pal, M. Sharon, G. Nogami Mat. Chem. Phys., 59 (1999), pp. 254–261
[26] V. Luthra, K. F. E. Pratt, R.G. Palgrave, D.E. Williams, R.P. Tandon, I.P. Parkin
Polyhedron., 29 (2010), pp. 1225–1230
[27] I. Ganesh, P. P. Kumar, A.K. Gupta, P. S. C. Sekhar, K. Radha, G. Padmanabham, G.
Sundararajan Proces. Appl. Ceramics., 6 (2012), pp. 21–36
[28] A. A. Cristobal, C.P. Ramos, P. G. Bercoff, P. M. Botta, J. M. Porto Lopeza Mat. Chem.
Phys., 133 (2012), pp. 971– 976
[29] M. R. Mohammadi, D. J. Fray Physica E., 46 (2012), pp. 43–51
[30] Y. Wang, Z. H. Jiang, F. J. Yang Mat. Sci. Eng. B., 134 (2006), pp. 76–79
[31] N. Pal, A. Bhaumik Adv. Coll. Int. Sci., 189–190 (2013), pp. 21–41
[32] D.W. Lee, B.R. Yoo J. Ind. Eng. Chem., 20 (2014), pp. 3947–3959

[33] X. Chen, S.A.C. Carabineiro, S.S.T. Bastos, P.B. Tavares, J.J.M.Orrfao, M.F.R. Pereira,
J.L. Figueiredo, J. Environ. Chem. Eng. 1 (2013), pp. 795–804
[34] X. Chen, S.A.C. Carabineiro, P.B. Tavares, J.J.M. Orrfao, M.F.R. Pereira, J.L.
Figueiredo, J. Environ. Chem. Eng. 2 (2014), pp. 344–355
[35] S.Yu, T. Zhang, Y. Xie, Q. Wang, X. Gao, R. Zhang, Y. Zhang, H. Su Int. J. Hydrogen
Energy., 40 (2015), pp. 870–877
[36] J. Cuia, W. Heb, H. Liuc, S. Liao, Y. Yue Coll. Surf. B., 74 (2009), pp. 274–278
[37] P. О. Larsson, A. Andersson Appl. Catal. B: Environ., 24 (2000), pp. 175–192
[38] A. R. Gandhe, J. S. Rebello, J. L. Figueiredo, J. B. Fernandes Appl. Catal. B: Environ.,
72 (2006), pp. 129–135
[39] Y. Liao, W. Que, Q. Jia, Y. He, J. Zhanga, P. Zhonga J. Mater. Chem., 22 (2012), pp.
7937–7944
[40] Z. Z. Djuric, O.S. Aleksic, M.V. Nikolic, N. Labusa, M. Radovanovic, M.D. Lukovic
Ceramics Int., 40 (2014), pp. 15131–15141
[41] M. Gharagozlou, R. Bayati Ceramics Int., 40 (2014), pp. 10247–10253
[42] M. Grujić-Brojčin, M. Šćepanović, Z. Dohčević-Mitrović, Z.V. Popović Sci. Sinter., 38
(2006), pp. 183–189
[43] P. Goswami, J. N. Ganguli Mater. Res. Bull., 47 (2012), pp. 2077–2084
[44] S. Kumar, N.K. Verma, M.L. Singla J. Nanomater. Biostruct., 7 (2012), pp. 607–619
[45] G. Chen, J. Chen, C. Srinivasakannan, J. Peng, Mater. Proc., 31 (2012), pp.187–191
[46] M. N. Iliev, V.G. Hadjiev, A.P. Litvinchuk Spectroscopy., 64 (2013), pp. 148–152
[47] C. T. Wang, H.L. Siao, Y.C. Chiu Surf. Coatings Technol., 232 (2013), pp. 658–665
[48] C. Adán, A. Bahamonde, I. Oller, S. Malato, A. Martínez-Arias Appl. Catal. B: Environ.,
144 (2014), pp. 269–276
[49] C. Oviedo Condens. Matter., 5 (1993), pp. 153–154

[50] M.V. Kuynetson, J.F. Yhuravlev, V.A. Gubanov J. Electron Spectrosc., 58 (1992), pp.
169–176
[51] N. Kruse, S. Chenakin Appl.Catal. A: Gen., 391 (2011), pp. 367–376
[52] K.Y. Yang, K.Z. Fung, M.C. Wanga J. Appl. Phys., 100 (2006), pp. 056102–056102
[53] A. Sandell, M.P. Andersson, M.K.J. Johansson, P. G. Karlsson, Y. Alfredsson, J.
Schnadt, H. Siegbahn, P. Uvdal Surf. Sci., 530 (2003), pp. 63–70
[54] M. Zhou, J. Yu, B. Cheng, H. Yu Mater. Chem. Phys., 93 (2005), pp. 159–163
[55] C. A. McCammon, S. A. McEnroe, P. Robinson, K. Fabian, B. P. Burton, Earth Planet.
Sci. Lett., 288 (2009), pp. 268–278
[56] S. S. Shinde, S. S. Meena, S. M. Yusuf, K. Y. Rajpure J. Phys. Chem., 115 (2011), pp.
3731–3736
[57] K. Sharma, S. S. Meena, S. Saxena, S. M. Yusuf, A. Srinivasan, G. P. Kothiyal Mater.
Chem. Phys., 133 (2012), pp. 144–150.
[58] M. Hirano, T. Joji, M. Inagaki J. Am. Ceram. Soc., 87 (2004), pp. 35–71
[59] W. Q. Guo, S. Malus, D.H. Ryan, Z. Altounian J. Phys. Condens. Matter., 11 (1999), pp.
6337–6346
[60] B. M. Corbella, J.M. Palacios Fuel., 86 (2007), pp. 113–122
[61] C. A. Sierra-Pereira, E.A. Urquieta-Gonzalez Fuel., 118 (2014), pp. 137–147

Figure captions
Figure 1. Conversion of ethyl acetate (a,c) and CO2 selectivity (b,d) on various samples
hydrothermally treated at 373 K (up) and 413 K (down)
100
80
60
40
20
0
100
80
60
40
20
0
a
b
TiO2_373
Fe2O3_373
Fe2O3_373
8Fe2Ti_373
5Fe5Ti_373
2Fe8Ti_373
TiO2_373
8Fe2Ti_373
5Fe5Ti_373
2Fe8Ti_373
550 600 650 700 750
550 600 650 700 750
Temperature, K
Temperature, K

100
80
60
40
20
0
100
80
60
40
20
0
d
c
2Fe8Ti_413
8Fe2Ti_413
5Fe5Ti_413
2Fe8Ti_413
5Fe5Ti_413
8Fe2Ti_413
550 600 650 700 750
550 600 650 700 750
Temperature, K
Temperature, K