Selective oxidation of propylene to acrolein by hydrothermally synthesized bismuth molybdates

Article abstract of DOI:10.1016/j.apcata.2014.05.038

Hydrothermal synthesis has been used as a soft chemical method to prepare bismuth molybdate catalysts for the selective oxidation of propylene to acrolein. All obtained samples displayed a plate-like morphology, but their individual aspect ratios varied with the hydrothermal synthesis conditions. Application of a high Bi/Mo ratio during hydrothermal synthesis afforded γ-Bi₂MoO₆ as the main phase, whereas lower initial bismuth contents promoted the formation of α-Bi₂Mo ₃O₁₂. Synthesis with a Bi/Mo ratio of 1:1 led to a phase mixture of α- and γ-bismuth molybdate showing high catalytic activity. The use of nitric acid during hydrothermal synthesis enhanced both propylene conversion and acrolein yield, possibly due to a change in morphology. Formation of β-Bi₂Mo₂O₉ was not observed under the applied conditions. In general, the catalytic performance of all samples decreased notably after calcination at 550 °C due to sintering.

Full text of DOI:10.1016/j.apcata.2014.05.038

Accepted Manuscript
Title: Selective oxidation of propylene to acrolein by
hydrothermally synthesized bismuth molybdates
Author: Kirsten Schuh Wolfgang Kleist Martin Høj Vanessa
Trouillet Pablo Beato Anker Degn Jensen Greta R. Patzke
Jan-Dierk Grunwaldt
PII:
S0926-860X(14)00372-X
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2014.05.038
APCATA 14858
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
1-3-2014
22-5-2014
27-5-2014
Please cite this article as: K. Schuh, W. Kleist, M. Hoj, V. Trouillet, P. Beato, A.D.
Jensen, G.R. Patzke, J.-D. Grunwaldt, Selective oxidation of propylene to acrolein by
hydrothermally synthesized bismuth molybdates, Applied Catalysis A, General (2014),
http://dx.doi.org/10.1016/j.apcata.2014.05.038
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.

Selective oxidation of propylene to acrolein by hydrothermally
synthesized bismuth molybdates
Kirsten Schuh, Wolfgang Kleist, Martin Høj, Vanessa Trouillet, Pablo Beato, Anker Degn Jensen, Greta
R. Patzke and Jan-Dierk Grunwaldt*
Kirsten Schuh, Dr. Wolfgang Kleist, Prof. Dr. Jan-Dierk Grunwaldt
Karlsruhe Institute of Technology (KIT), Institute for Chemical Technology and Polymer Chemistry
(ITCP) and Institute of Catalysis Research and Technology (IKFT), Engesserstr. 20, 76131 Karlsruhe
(Germany)
Dr. Martin Høj, Prof. Dr. Anker Degn Jensen
Technical University of Denmark (DTU), Department of Chemical & Biochemical Engineering, Søltofts
Plads, 2800 Kgs. Lyngby (Denmark)
Prof. Dr. Greta R. Patzke
University of Zurich (UZH), Institute of Inorganic Chemistry, Winterthurerstrasse 190, CH-8057 Zurich
(Switzerland)
Vanessa Trouillet
Karlsruhe Institute of Technology (KIT), Institute for Applied Materials (IAM) and Karlsruhe Nano
Micro Facility (KNMF), Hermann-von-Helmholtz-Platz 1, 76344 Eggenstein-Leopoldshafen (Germany)
Dr. Pablo Beato
Haldor Topsøe A/S, Nymøllevej 55, 2800 Kgs. Lyngby (Denmark)
KEYWORDS: Bismuth molybdates; hydrothermal synthesis; unsupported catalysts; selective
propylene oxidation; acrolein
ABSTRACT: Hydrothermal synthesis has been used as a soft chemical method to prepare bismuth
molybdate catalysts for the selective oxidation of propylene to acrolein. All obtained samples
displayed a plate-like morphology, but their individual aspect ratios varied with the hydrothermal
synthesis conditions. Application of a high Bi/Mo ratio during hydrothermal synthesis afforded γ-
Bi2MoO6 as the main phase, whereas lower initial bismuth contents promoted the formation of α-
Bi2Mo3O12. Synthesis with a Bi/Mo ratio of 1:1 led to a phase mixture of α- and γ-bismuth molybdate
showing high catalytic activity. The use of nitric acid during hydrothermal synthesis enhanced both
propylene conversion and acrolein yield, possibly due to a change in morphology. Formation of β-
Bi2Mo2O9 was not observed under the applied conditions. In general, the catalytic performance of all
samples decreased notably after calcination at 550 °C due to sintering.
1
Page 1 of 36

1. Introduction
The typical commercial catalyst for the selective oxidation of propylene to acrolein and for propylene
ammoxidation to acrylonitrile is based on bismuth molybdates as the active species in combination
with additional elements such as Co, Ni, Fe, Cr, Al or alkali metals [1, 2]. Transition metal molybdates,
especially CoMoO4 or Fe2(MoO4)3 and their solid solutions, increase the ability to exchange electrons
and oxygen anions due to the effective activation of molecular oxygen by Co2+ and Fe3+. They change
the morphology of the catalyst under the reaction conditions along with recently reported synergistic
effects of the different phases [3-5]. Other elements, including alkali metals, W, V, Nb etc. are usually
added to enhance the catalyst life time and its mechanical strength. They also lead to minor
improvements in activity and selectivity [6, 7].
Various bismuth molybdates including α-Bi2Mo3O12, β-Bi2Mo2O9 and γ-Bi2MoO6 have been prepared
and tested as oxidation catalysts [8]. Whereas γ-Bi2MoO6 exhibits an Aurivillius-type structure with
alternating [Bi2O2]2+ slabs and layers of corner sharing MoO6 octahedra, both the α- and the β-phase
can be considered as defective fluorite structures consisting of tetrahedral MoO4 motifs. At
temperatures above 570 °C the metastable γ’’-phase is formed which is further transferred to the
high-temperature γ’-phase at 640 °C. γ´-Bi2MoO6, which is also known as the γ(H)-phase, also has a
fluorite like structure based on MoO4 tetrahedra [9]. The metastable phase β-Bi2Mo2O9 can usually
be accessed in the temperature range from 540 to 665 °C [10].
Despite extensive preceding studies, there is an ongoing debate in literature about the relative
catalytic activity of these phases [11, 12]. The group of Keulks [4, 11] found the γ-Bi2MoO6 phase to
be most active, whereas Burrington and Grasselli [13] stated that the β-phase showed the highest
rates for propylene oxidation. Carson et al. [14] prepared the different bismuth molybdate phases by
co-precipitation at pH = 7 and found that the catalytic activity for propylene oxidation decreased in
the following order: α > γ > β. They further concluded that a mechanical mixture of α-Bi2Mo3O12 and
γ-Bi2MoO6 was more active and selective in propylene oxidation to acrolein than the according pure
phases. This synergistic effect was confirmed by Bing et al. [12], who suggested that the γ-phase
3
Page 2 of 36

generated the active oxygen species, while the α-phase contributed the selective sites for acrolein
formation. According to Le et al. [15] the synergistic effect was only observed in mixtures containing
the γ-phase in close contact to the other phases.
The most common method to prepare bismuth molybdate catalysts is co-precipitation [16, 17]
followed by calcination or solid-state reaction [18] at temperatures up to 1000 °C, but also sol-gel
synthesis [19] or spray drying [20] were applied. These synthesis routes require harsh conditions and
high temperatures to obtain crystalline products, which may influence the catalytic activity of the
resulting phase. van Well et al. [21] found that the activity of γ-Bi2MoO6 prepared by co-precipitation
and spray drying strongly depended on the calcination conditions. Alternatively γ-Bi2MoO6 has been
synthesized under hydrothermal conditions applying cetyltrimethylammonium bromide (CTAB) as a
surfactant [22] or using sodium, chlorine or fluorine containing reactants [23, 24], which may remain
in the product and influence its catalytic performance. Hydrothermal processes are typical soft
chemistry ('chimie douce') approaches which provide convenient access to advanced materials of
high crystallinity, high purity, narrow size distribution and a low degree of aggregation [8]. Beale and
Sankar [25] prepared α- and γ-bismuth molybdate in a one step hydrothermal synthesis from Bi2O3
dissolved in nitric acid and ammonium molybdate dissolved in ammonium hydroxide. Synthesis with
β-phase stoichiometry (Bi/Mo = 2:2) and variation of the pH value did not yield β-Bi2Mo2O9 under
hydrothermal conditions so that heat treatment was required [8].
In the present study, mild hydrothermal techniques were applied to produce various bismuth
molybdate phases with different morphologies and corresponding phase mixtures. The influence of
calcination on the phase composition and the catalytic activity for propylene oxidation to acrolein
has been studied in detail. However, to the best of our knowledge hydrothermally synthesized
bismuth molybdates have not yet been applied for the selective oxidation of propylene to acrolein. In
the following, we compare the properties of catalysts emerging from this flexible synthetic route to
samples synthesized by conventional co-precipitation.
4
Page 3 of 36

2. Experimental section
2.1.
Synthesis.
The bismuth molybdate materials were synthesized by hydrothermal synthesis, while reference
samples were obtained from conventional co-precipitation. All chemicals were analytical grade and
used without further purification.
In a typical hydrothermal synthesis 10 mmol Bi(NO3)3 ∙ 5H2O and the stoichiometric amount of
(NH4)6Mo7O24 ∙ 4H2O (Bi/Mo ratios in the range of 0.5 – 3) were dissolved in 100 ml deionized water.
These samples will be referred to here as HT_BixMoy with x/y being the Bi/Mo-molar ratio.
Additionally for a ratio of Bi/Mo = 1:1 the pH was adjusted from 0.9 to 4 by addition of ammonium
nitrate solution (HT_Bi1Mo1_pH4). 5 ml of nitric acid was added to the solution containing a Bi/Mo
ratio of 1:1 and 2:1 referred to as HT_BixMoy_HNO3. Precursors with a Bi/Mo = 2:1 ratio were
moreover dissolved in 80 ml of 25 vol.% acetic acid to lower the pH value from 0.8 to 0.2
(HT_Bi2Mo1_HOAc). The resulting solutions were heated in sealed 250 ml autoclaves (Berghof) with
Teflon inlays at 180 °C for 24 h in an oven. After cooling to room temperature, the solid product was
separated by filtration, washed with water, ethanol and finally with acetone. The resulting powder
was dried at room temperature and ambient pressure. The samples prepared with a Bi/Mo ratio in
the range 0.5 – 2 in water were also calcined at 550 °C for 4 h.
The corresponding co-precipitated samples were synthesized according to Carrazán et al.[6] using
(NH4)6Mo7O24 ∙ 4H2O dissolved in NH4OH and Bi(NO3)3 ∙ 5H2O dissolved in HNO3 at pH = 7. The
resulting solid material was calcined at 450 °C to yield the α- and the γ-phases and at 680 °C to obtain
the β-phase. Experiments are referred to as CP_BixMoy_CT where CT is the calcination temperature
in °C.
2.2.Characterization.
Powder X-ray diffraction (PXRD) measurements were recorded using a Bruker D8 Advance powder
diffractometer in the range 2θ = 8 - 80° (step size 0.016 °) with Cu Kα radiation (Ni-filter, 45 mA, 35
5
Page 4 of 36

kV) on rotating sample holders. Raman spectra were recorded with a Horiba Jobin Yvon
spectrometer (LabRam) attached to an Olympus microscope (BX 40) using a 632.8 nm laser in the
range 100 - 1100 cm-1 on an object slide without pretreatment of the sample.
The specific surface area (SSA) was measured by nitrogen adsorption at its boiling point (Belsorp II
mini, BEL Japan Inc.) using multipoint BET theory in the p/p0 = 0.05 – 0.3 range.
For scanning electron microscopy (SEM), performed on a Quanta 200 ESEM (FEG) microscope at the
Centre of Electron Nanoscopy (CEN) at DTU, samples were deposited on a carbon foil on aluminium
stubs and coated with carbon to improve the conductivity.
Surface analysis by X-ray photoelectron spectroscopy (XPS) was performed with a K-Alpha
spectrometer (ThermoFisher Scientific) using a microfocused Al Kα X-ray source (400 µm spot size).
Data acquisition and processing using Thermo Avantage software is described elsewhere. [26] Charge
compensation during analysis was achieved using electrons of 8 eV energy and low energy argon ions
to prevent any localized charge build-up. Spectra were fitted with one or more Voigt profiles (binding
energy uncertainty: ±0.2 eV). [27, 28] Scofield sensitivity factors were applied for quantification. [29]
The energy scale was calibrated to the binding energy of C 1s (C-C, C-H) at 285.0 eV controlled by
means of the well known photoelectron peaks of metallic Cu, Ag and Au, respectively.
The bulk composition of the catalysts was determined by optical emission spectrometry with
inductively coupled plasma (ICP-OES, Agilent 720/725-ES). The plasma was created by a 40 MHz high-
frequency generator and argon was applied as the plasma gas. For the ICP-OES each sample was
dissolved in 6 ml concentrated HNO3, 2 ml concentrated HCl and 0.5 ml H2O2 in a microwave (at 600
Watt for 45 minutes).
Quantitative nitrogen analysis was carried out using hot gas extraction method (LECO TC 600). The
samples were heated in a graphite crucible under flowing helium and thermally decomposed. The
amount of N2 gas was determined by a heat conductivity detector. Each measurement was repeated
twice and a standard deviation was calculated.
6
Page 5 of 36

2.3.Catalytic test reaction.
Catalytic activities were measured in a continuous flow fixed bed reactor, a U-shaped quartz reactor
with 4 mm inner diameter. The catalyst powders were crushed and sieved to 150 – 300 µm particles
and 500 mg of sample was loaded in the reactor and stabilized with quartz wool. The quartz reactor
was connected to a commercial test unit (ChimneyLab Europe) with calibrated mass flow controllers
(Brooks) and placed in an oven (Watlow) [30]. A thermocouple was placed inside the reactor just
touching the catalyst bed to measure the reaction temperature and a pressure transducer placed
upstream of the reactor measured the actual reaction pressure. The catalysts were pre-oxidized in
dry air at 550 °C for the calcined hydrothermally synthesized samples and at 300 °C for all other
samples. Activity tests were performed using a gas composition of C3H6/O2/N2 = 5/25/70 and flows of
50, 80, 120 Nml/min. The gas analysis was performed with a dual channel GC–MS (Thermo Fischer)
with a TCD detector for quantifying N2, O2, CO and CO2 and a FID detector parallel to the MS for
identifying and quantifying saturated and unsaturated light hydrocarbons and oxygenated
byproducts. The measured concentrations were corrected for expansion of the gas due to
combustion using the nitrogen signal as internal standard, before calculating the conversion of
propylene and the selectivity for acrolein.
3. Results and Discussion
3.1.
Samples synthesized in pure aqueous hydrothermal media.
3.1.1. Characterization of the as-prepared bismuth molybdates.
Five different samples were prepared under hydrothermal conditions in pure water with Bi/Mo ratios
from 0.5 – 3, and their phase composition was analyzed by powder X-ray diffraction (PXRD) and
Raman spectroscopy. Only two different bimetallic phases were present in the products summarized
in Table 1: α-Bi2Mo3O12 and γ-Bi2MoO6. Application of a high Bi/Mo ratio led to the formation of the
7
Page 6 of 36

bismuth-rich phase, γ-Bi2MoO6. Decreasing initial bismuth contents led to increasing amounts of α-
Bi2Mo3O12 in the as prepared samples. The diffraction pattern of the sample prepared with a high
excess of bismuth (Bi/Mo = 3:1) displays characteristic reflections at 2θ = 28.3, 23.7 and 46.9 ° which
can be assigned to γ-Bi2MoO6 (PDF 21-102 [31]) [32] and cubic BiO2-x (PDF 47-1057) [33]. The
crystallinity of HT_Bi3Mo1 was lower than the crystallinity of the other four samples, as indicated by
the broader reflections in the diffraction pattern (Figure 1). This goes hand in hand with a
significantly higher surface area than observed for the other four samples (30 m²/g compared to 6 - 8
m²/g). The diffraction pattern of the sample synthesized with Bi/Mo = 2:1 (Figure 1a) indicates γ-
Bi2MoO6 as the main phase and additionally the presence of α-Bi2Mo3O12 (low intensity reflection at
27.9°) and Bi6O6(OH)3(NO3)3 ∙ 1.5H2O (10.3° and 31.3°, PDF 53-1038) [34]. The corresponding Raman
spectrum of this sample depicted in Figure 1b confirmed the presence of γ-Bi2MoO6 as the main
phase in this sample. The observed Raman bands at 848, 808, 792, and 714 cm-1 were assigned to
Mo-O stretching frequencies of the distorted MoO6 octahedra of the Aurivillius layered structure [32,
35]. Additionally, bands at 352, 323, 294 and 282 cm-1 were found, which correspond well to those
reported for γ-Bi2MoO6 in the literature [36]. The presence of α-Bi2Mo3O12 was evident from the
band at 901 cm-1, and the vibration at 1031 cm-1 could be assigned to a nitrate containing phase. This
nitrate containing phase also led to a broader band at 782 – 808 cm-1 [37]. Quantitative nitrogen
analysis yielded 0.36 wt.% (± 0.02 wt.%) N in this sample, which agrees well with the presence of a
nitrate phase.
The ICP-OES measurement of HT_Bi2Mo1 showed that the product contained bismuth and
molybdenum in the ratio 2:1, which corresponds well to the applied ratio.
Whereas distinction of the three different phases may be difficult with PXRD methods due to their
adjacent main reflections (α: 27.9 °, β: 27.8 °, γ: 28.3 °) and the formation of the β-phase is easily
overlooked, Raman spectra were recorded for a more precise and complementary identification of
these three different bismuth molybdate phases.
8
Page 7 of 36

The sample prepared with Bi/Mo = 1:1 contained a mixture of γ-Bi2MoO6 and α-Bi2Mo3O12 according
to PXRD measurements, whereas the Raman spectrum only displayed the α-phase. While X-ray
diffraction measurements provide information of quantitative phase composition of larger particles,
Raman spectra often only show the strongest bands which may overlap with band of other phases.
The Raman spectrum of α-Bi2Mo3O12 exhibited six bands between 1000 – 800 cm-1 which can be
attributed to the Mo-O stretching modes of different tetrahedral species, i.e. 955 cm-1 (a3), 925 cm-1
(a1), 906 cm-1 (a2), 856 cm-1 (a1), 840 cm-1 (a2) and 816 cm-1 (a3). All of these six bands were present in
the spectra of HT_Bi1Mo1 as well as in the spectra of the samples synthesized with a Bi/Mo ratio
below 1:1 (Figure 1b).
The X-ray diffraction pattern and the Raman spectrum of the sample with an initial Bi/Mo ratio 2:3
showed the characteristic features of α-Bi2Mo3O12 and indicated in addition the presence of a
bismuth-free molybdenum oxide side phase (reflection at 25.7 °, Raman shift: 993 cm-1). Calculation
of the Bi/Mo ratio from the values obtained from ICP-OES measurements yielded to 0.7 (± 10%)
(Table 1), corresponding to the Bi/Mo ratio in the α-phase. This indicates that, additionally to α-
Bi2Mo3O12 and an ammonium-containing molybdenum oxide phase, another bismuth oxide phase
should be present which could not be detected by PXRD or Raman spectroscopy. Generally, such
small amounts of bismuth oxides are difficult to identify from diffraction patterns due to the overlap
of some of their reflections with those of bismuth molybdates. van Well et al. [21] also claimed that
excess bismuth in form of β-Bi2O3 or β-Bi2Mo2O9 cannot be seen by X-ray diffraction when the
samples were calcined at temperatures lower than 500 °C.
The Raman spectrum of the sample synthesized with a high excess of molybdenum (HT_Bi1Mo2)
illustrates the presence of the main γ-phase together with a band at 973 cm-1 which corresponds well
to NH3(MoO3)3 in agreement with PXRD data (reflections at 19.4 and 25.7°).
SEM images of all samples showed a plate-like morphology (Figure 2) with variations in aspect ratio
and size. Synthesis with a large excess of bismuth led to the smallest particle size (Figure 2a and b),
resulting in a relatively high specific surface area (30 m²/g). The particles of sample HT_Bi2Mo1 were
9
Page 8 of 36

rectangular with a side length of 1 - 2.5 µm and a thickness of ca. 200 nm (Figure 2c). Synthesis with
Bi/Mo = 1:1 led to smaller plates (500 nm) with rounded edges (Figure 2e and f). SEM images of
HT_Bi2Mo3 (Figure 2g and h) revealed a mixture of plates with a side length around 1 µm and larger
needles / rods. HT_Bi1Mo2 (Figure 2d) morphologically displays two types of particles as well: the
plate-like α-Bi2Mo3O12 (400 nm – 1 µm in size) and the wedge-like / rod-like ammonium molybdenum
oxide phase visible at the bottom of the image.
3.1.2. Catalytic performance of the as prepared bismuth molybdates.
The catalytic activity measurements for propylene oxidation of the as prepared samples are depicted
in Figure 3. At 320 °C propylene conversion for all five samples remained below 10%. The sample
prepared with a high excess of bismuth (HT_Bi3Mo1) exhibited propylene conversion up to 7.3% but
the selectivity for acrolein was very low and mainly COx was formed. HT_Bi2Mo3 and HT_Bi1Mo2
yielded similar propylene conversion (2.6 – 7.2%) but HT_Bi1Mo2, which additionally to α-Bi2Mo3O12
also contained NH3(MoO3)3, exhibited lower acrolein selectivity (36 - 56% compared to 60 – 63%).
Additionally to acrolein also CO2, CO and acetaldehyde were formed. The synthesized sample with
Bi/Mo = 2:1 (HT_Bi2Mo1) showed the highest acrolein yield at all three flows (120, 80 and 50
Nml/min). The surface areas of the hydrothermally synthesized samples with Bi/Mo = 0.5 - 2 were
very similar (6 – 8 m²/g; see Table 1) and therefore calculating propylene conversion on the basis of
the surface area inside the reactor did not change the relative activities of the measured samples. At
360 °C all samples showed very similar activities resulting in a propylene conversion between 10%
and 22% depending on the flow and accordingly on the contact time. Besides acrolein (60 - 90%) only
COx (mainly CO2) and traces of acetaldehyde were formed. The sample prepared with a large excess
of bismuth (HT_Bi3Mo1) converted less propylene than the other four samples at 360 °C and was
almost inactive for propylene oxidation. This sample was also not selective for acrolein and mainly
produced hexadiene and COx. This agrees well with previous literature reports on the formation of
hexadiene in the presence of Bi2O3 [38, 39]. HT_Bi2Mo1 also contained a bismuth oxide phase but its
concentration may be too low to influence the catalytic activity of the sample. The presence of γ-
10
Page 9 of 36

Bi2MoO6 as main product goes hand in hand with a slightly higher selectivity at lower conversion
compared to the samples synthesized with lower Bi/Mo ratio. At higher temperature (400 °C) the
propylene conversion increased up to 32% and the difference between the measurement points for
each sample at similar conditions also raised (Figure 3b). The sample prepared with Bi/Mo = 1:1 was
slightly more active than HT_Bi2Mo1 and their catalytic activity deteriorated with decreasing Bi/Mo
ratio. The selectivity for acrolein was never below 60% and remained relatively constant with
increasing propylene conversion indicating that acrolein further reacted to COx only very slowly on
these catalysts. At temperatures above 400 °C the bismuth molybdates deactivated probably due to
a loss of surface area loss and morphological changes. An increase of the reaction temperature from
440 °C to 520 °C did not result in an increase in propylene conversion and acrolein yield.
Note that as for most transition metal oxide catalysts the initial phase probably changed during the
catalytic activity measurements from 320 – 520 °C, so that the phases summarized in the second
column of Table 1 (“as-prepared”) are not those actually present in the reactor during propylene
conversion. We thus investigated the phase composition after the catalytic process as well.
Figure 4 shows the powder diffraction pattern of the hydrothermally synthesized samples after
catalytic tests at temperatures up to 520 °C. Comparisons of the diffraction pattern and also of the
Raman spectra of fresh and used catalysts indicate that the main phase of the catalysts did not
change during the application in selective oxidation of propylene. Post-catalytic HT_Bi3Mo1
consisted of γ-Bi2MoO6 and Bi8Mo3O21, which is orthorhombic and presents a basic fluorite-type
structure [40, 41]. The minor phases in samples HT_Bi2Mo1 and HT_Bi2Mo3 indicated by only one or
two low intensity reflections disappeared, whereas for the sample synthesized with a high excess of
molybdenum (HT_Bi1Mo2) the hexagonal NH3(MoO3)3-phase was transformed into α-MoO3. SEM
images in Figure 5 revealed that the plate-like morphology of the samples could not be preserved
during the application for selective oxidation of propylene at 320 – 520 °C. Only the sample
synthesized with the Bi/Mo ratio 2:1 partly retained its plate-like morphology (Figure 5d). All samples
exhibited similar morphologies after the catalytic tests, whereas their particle sizes decreased with
11
Page 10 of 36

increasing bismuth content. Generally, thermal treatment led to particle agglomeration and a
significant loss of surface area (< 1 m²/g). Post-catalytic samples only contained α-Bi2Mo3O12 and / or
γ-Bi2MoO6, except for sample HT_Bi1Mo2, where an excess of Mo led to the additional formation of
α-MoO3.
3.2.
Influence of the calcination procedure on catalyst properties and activity
3.2.1. Characterization of calcined samples.
Hydrothermally synthesized samples with Bi/Mo ratios 0.5 – 2.0 (pure water) were calcined at 550 °C
for 4 h to increase the phase purity of the products and to evaluate the transformation of the phases
at elevated temperatures. Figure 6a depicts the PXRD patterns of the four calcined samples, and the
corresponding phase composition as well as the Bi/Mo ratios determined by ICP-OES and XPS are
summarized in Table 2. For the sample synthesized with a Bi/Mo ratio 2:1 additionally to γ-Bi2MoO6
the β-phase was formed during calcination at 550 °C. The presence of β-Bi2Mo2O9 is indicated by
characteristic reflections at 27.9° and 31.8° in the diffraction pattern in Figure 6a and by the band at
885 cm-1 in the Raman spectrum (Figure 6b). As the β-phase is considered stable between 540 and
665 °C, formation of β-Bi2Mo2O9 was also expected for the calcined sample emerging from Bi/Mo =
1:1. Surprisingly, calcination of HT_Bi1Mo1 only improved the crystallinity of the α- and γ-bismuth
molybdate mixture, but did not lead to the formation of β-Bi2Mo2O9. This suggests that higher
calcination temperatures or longer treatment times should have been applied and according
investigations are under way. Li et al. [8] could not synthesize the β-phase directly by hydrothermal
synthesis but after calcination at 560 °C. Beale and Sankar [25] showed by combined EDXRD/XAS that
phase-pure β-Bi2Mo2O9 could be produced from a hydrothermally synthesized precursor material by
heat treatment at temperatures above 500 °C.
After calcination at 550 °C for 4 h the sample synthesized from a Bi/Mo ratio 2:3 only contained α-
Bi2Mo3O12. HT_Bi1Mo2_calc containing a larger excess of molybdenum additionally showed
characteristic features of α-MoO3 in the diffraction pattern (Figure 6a; reflections at 23.3°, 25.7°,
27.3°, 33.7°) and in the Raman spectra (Figure 6b; 993 cm-1). Bi/Mo ratios found in the bulk of the
12
Page 11 of 36

products synthesized with Bi/Mo = 2:1 and 2:3 were consistent with the applied ratio and did not
change during calcination (Table 2), thereby indicating that no molybdenum or bismuth losses
occurred during thermal treatment. Comparison of the metal ratio in the bulk determined by ICP-OES
and on the surface (XPS) showed that both values were practically identical after calcination. The
surface composition of the catalyst was determined from the peak areas of Mo 3d5/2 = 232.7 eV and
Bi 4f7/2 = 159.5 eV and a representative spectrum is shown in the supporting information. XPS
measurement of the as prepared HT_Bi2Mo1 revealed enrichment with bismuth on the surface,
which might be due to the presence of Bi2O3.96. However, the activity for propylene conversion was
sufficiently high and no formation of hexadiene due to the presence of Bi2O3 could be observed [38,
39]. Hence, it is more likely that the γ-bismuth molybdates have an excess of superficial bismuth,
which is in line with literature [20, 42]. Bing et al. [12] also observed a decrease of the surface Bi/Mo
ratio for γ-Bi2MoO6 prepared by co-precipitation during calcination at 420 °C. van Well et al. [21]
calcined spray dried and co-precipitated samples with a theoretical Bi/Mo ratio of 2.0 at 550 °C and
determined similar surface ratios to the values found for our hydrothermally synthesized samples:
the high Bi/Mo ratio of 3.0 determined for the as prepared HT_Bi2Mo1 and the value of 1.9 for the
calcined sample containing additional amounts of the β-phase (see Table 2).
3.2.2. Catalytic performance before and after calcination.
Next, the four calcined samples HT_Bi2Mo1_calc, HT_Bi1Mo1_calc, HT_Bi2Mo3_calc and
HT_Bi1Mo2_calc were applied in the catalytic oxidation of propylene where they exhibited a lower
propylene conversion and, accordingly, a lower acrolein yield than the non-calcined samples under
the same conditions (flow range 50 – 260 Nml/min, 460 – 520 °C). Figure 7 compares the acrolein
yield and the propylene conversion of the as prepared and calcined catalysts at 480 °C and 120
Nml/min. The acrolein yield for the calcined samples decreased with increasing molybdenum content
to 1.0% at corresponding propylene conversion of 1.7% for HT_Bi1Mo2_calc.
For HT_Bi2Mo3 propylene conversion decreased from 15.8% to 2.3% due to calcination and the
acrolein selectivity increased from 82% to 92%. Comparing the phase composition of this sample
13
Page 12 of 36

before and after calcination suggests that the phase change was responsible for the deactivation. The
ammonium molybdenum oxide phase in HT_Bi2Mo3 was removed in the course of calcination and
phase pure α-Bi2Mo3O12 was obtained (Figure 7). This ammonium molybdenum phase should have a
lower catalytic activity in propylene oxidation and a lower selectivity to acrolein than α-Bi2Mo3O12
[13]. Therefore the mixture of ammonium molybdenum oxide and α-Bi2Mo3O12 should be less active
and less selective than pure α-Bi2Mo3O12. However, calcination of the other three samples also led to
a deactivation in propylene oxidation. For HT_Bi2Mo1 the β-phase was formed and for HT_Bi1Mo2
hexagonal NH3(MoO3)3 was changed to orthorhombic MoO3. HT_Bi1Mo1 did not undergo phase
changes during calcination but the deactivation through thermal treatment was more significant than
for the other three samples. Although propylene conversion decreased for the calcined samples
compared to the as prepared materials, the selectivity decreased only for HT_Bi1Mo1 and
HT_Bi1Mo2. For HT_Bi2Mo1 and HT_Bi2Mo3 acrolein selectivity increased from 69 to 80% and from
82 to 92% but the resulting acrolein yield was always lower for the calcined than the as-prepared
sample.
van Well et al. [21] reported that calcination time and temperature strongly influenced the activity of
the bismuth molybdenum catalysts. They proposed surface enrichment with bismuth after
calcination at higher temperature or longer time as an explanation. This hypothesis could not be
confirmed here, because the bismuth concentration on the surface decreased during calcination (see
Table 2). A direct correlation between the Bi/Mo ratio on the surface given in Table 1 and the
corresponding propylene conversion could thus not be derived. An alternative explanation could be
the decrease of the surface area, which was ≤ 1 m²/g for all calcined samples, leading to low
propylene conversion.
Comparison of SEM images in Figure 8 and Figure 2 revealed that the particles sinter and
agglomerate during calcination. Only the sample prepared from an initial Bi/Mo ratio 2:1 retained a
plate-like morphology, but the thickness increased during calcination and the aspect ratio changed
from rectangular to quadratic (Figure 8a). In summary, only HT_Bi2Mo1 partially preserved its
14
Page 13 of 36

morphology during calcination and this sample showed the best catalytic performance compared to
the other three calcined materials.
3.3.
Comparison of different hydrothermally prepared catalysts with co-
precipitated bismuth molybdates.
3.3.1. Characterization of hydrothermally prepared and conventionally synthesized
catalysts.
The diffraction patterns of the bismuth molybdate samples prepared from different Bi/Mo ratios by
hydrothermal synthesis or co-precipitation are shown in Figure 9. Samples prepared with a Bi/Mo
ratio 2:1 all contained mainly γ-Bi2MoO6 except for the co-precipitated sample calcined at 680 °C,
which consisted of γ”-Bi2MoO6, whereas α-Bi2Mo3O12 was the main phase for Bi/Mo = 2:3. The
samples prepared with Bi/Mo = 1:1 consisted of a mixture of the α-phase with either β- or γ-bismuth
molybdate.
As already discussed in section 3.1, hydrothermal synthesis at 180 °C with Bi/Mo = 2:1 prepared in
pure water led to the formation of γ-Bi2MoO6 with some other Mo- or Bi-containing phases as
impurity. Addition of acetic acid did also result in the formation of γ-Bi2MoO6, albeit with α-
Bi2Mo3O12 as a side product. This sample featured agglomerates of flat plates with round-shaped
edges illustrated in Figure 10a and b. Phase pure γ-Bi2MoO6 was obtained from synthesis in the
presence of nitric acid. The plates of HT_Bi2Mo1_HNO3 (Figure 10c) were rectangular with well
defined steps and edges on the surface, which may be beneficial for the selectivity of the catalytic
reaction. Co-precipitation followed by calcination at 450 °C yielded pure γ-Bi2MoO6 as well, whereas
γ”-Bi2MoO6 (PDF No. 33-208 [31]) [9] was formed during calcination at 680 °C as indicated by the
reflections at 27.4, 31.1, 31.9 and 45.1°. Hydrothermal synthesis with Bi/Mo = 1:1 did not lead to the
formation of the β-phase but to a mixture of α- and γ-phase for all three samples. Co-precipitation of
bismuth molybdates with Bi/Mo = 1:1 and calcination at 680 °C resulted in a mixture of β-Bi2Mo2O9
and α-Bi2Mo3O12.
15
Page 14 of 36

N2 physisorption measurements showed that the co-precipitated samples and the hydrothermally
synthesized samples exhibited similar surface areas (6 - 9 m²/g; cf. Table 3).
The Bi/Mo ratios determined by ICP-OES and XPS are summarized in Table 3. For both preparation
methods the applied ratio corresponded to the bulk concentration actually present in the prepared
catalyst. A loss of molybdenum, which was found by van Well et al. [21] during precipitation, could
not be confirmed here, probably because higher pH values were applied in the present study.
Comparison of the ICP-OES results with the corresponding ratios calculated from XPS shows that the
samples containing γ-Bi2MoO6 exhibit an excess of bismuth on the surface. HT_Bi2Mo1 displayed the
highest Bi excess, whereas heat treatment (e.g. calcination of the co- precipitated samples) led to a
reduced Bi/Mo ratio on the surface (Table 3).
3.3.2. Catalytic performance comparison of hydrothermally synthesized and co-
precipitated samples.
The catalytic performance of all materials during the selective oxidation of propylene to acrolein at
320 °C and 360 °C is summarized in Figure 11. At 320 °C the co-precipitated sample synthesized with
Bi/Mo = 2:1 which consisted of γ-Bi2MoO6 showed the highest catalytic activity with propylene
conversion of 5 - 11% at acrolein selectivities of 59 – 69%. HT_Bi2Mo1 which additionally to γ-
Bi2MoO6 contained α-Bi2Mo3O12 also exhibited relatively high propylene conversion (3 – 9%) at
acrolein selectivities of 58 – 100%. In Figure 11b the catalytic activities are reported for the surface
area inside the reactor instead for a fixed weight of catalyst as in Figure 11a. Comparison of these
two figures shows that the catalytic activities of the different samples still follow the same trend
except CP_Bi2Mo3_450. The co-precipitated sample synthesized with Bi/Mo = 2:3 exhibited higher
propylene conversion per surface area than the other bismuth molybdates due to its low surface
area (1 m²/g). The two samples calcined at 680 °C (CP_Bi1Mo1_680 and CP_Bi2Mo1_680) with
surface areas < 1 m²/g showed very low catalytic activity in the selective oxidation of propylene and
also low propylene conversion per surface area at 320 °C (Figure 11a and b). An increase in reaction
temperature to 360 °C resulted in higher propylene conversion for all samples (up to 27%) and
16
Page 15 of 36

acrolein selectivities of mainly 60 – 70%. Comparison of the samples synthesized with the ratio Bi/Mo
= 2:1 evidenced that the samples HT_Bi2Mo1_HNO3 and CP_Bi2Mo1_450, both synthesized using
nitric acid and consisting of pure γ-Bi2MoO6 showed the highest activity, whereas the co-precipitated
sample was less selective than the hydrothermally synthesized sample. This difference in selectivity
could be associated with the morphology, which was well defined for HT_Bi2Mo1_HNO3 but random
for co-precipitated samples [43, 44]. The samples synthesized under hydrothermal conditions with
acetic acid, which partially led to the formation of α-Bi2Mo3O12, showed relatively low activity in
propylene oxidation and low selectivity for acrolein. In general, Figure 11c suggests that the γ-phase
was more active than the α-phase as the two most active samples consisted only of γ-Bi2MoO6. The
two samples synthesized with nitric acid but different Bi/Mo ratios (HT_Bi2Mo1_HNO3 and
HT_Bi1Mo1_HNO3) both showed relatively high catalytic activity but HT_Bi1Mo1_HNO3, containing
a
mixture of α-Bi2Mo3O12 and γ-Bi2MoO6, was slightly less active and selective than
HT_Bi2Mo1_HNO3. In Figure 11d acrolein selectivity is given as a function of propylene conversion
per surface area. CP_Bi2Mo1_450 and HT_Bi2Mo1_HNO3 still exhibited high propylene conversion
per surface area but again the co-precipitated sample synthesized with Bi/Mo = 2:3 obtained the
highest values for propylene conversion per surface area (5 – 10.4 %/m²) due to the low surface area
of this sample (1 m²/g). Also at 360 °C both samples calcined at 680 °C (CP_Bi2Mo1_680 and
CP_Bi1Mo1_680) were inactive for propylene oxidation, due to the high temperature treatment and
the very low surface area (< 1 m²/g). As calcination at 550 °C strongly decreased the surface area and
the activity of the prepared samples without affording β-Bi2Mo2O9, prolonged calcination at higher
temperatures would probably not exert a productive influence either. Whereas direct hydrothermal
synthesis of β-Bi2Mo2O9 without calcination was not possible, the obtained mixtures (HT_Bi1Mo1,
HT_Bi1Mo1_HNO3, HT_Bi1Mo1_pH4) still exhibited relatively high activity at 360 °C.
According to Table 3, where the maximum acrolein yield for each sample and the corresponding
propylene conversion are summarized, HT_Bi1Mo1_pH4, HT_Bi1Mo1_HNO3 and HT_Bi2Mo1_HNO3,
i.e. the samples synthesized under hydrothermal conditions with nitric acid or ammonia, were the
17
Page 16 of 36

most active catalysts. One reason for this superior performance could be the incorporation of
nitrogen in the catalyst, but the analysis of the nitrogen content of the samples HT_Bi2Mo1,
HT_Bi2Mo1_HOAc and HT_Bi2Mo1_HNO3 resulted in 0.4 wt.%, 0.1 wt.% and 0.1 wt.% N and does
not correlate with the catalytic performance. The highest nitrogen content was found for the sample
synthesized in the absence of acid, probably due to the presence of a nitrate phase, which
disappeared during calcination. HT_Bi2Mo1_HOAc, which is composed of γ-Bi2MoO6 and α-
Bi2Mo3O12, also contained nitrogen. This indicated that either the Mo or the Bi precursor or both
acted as main nitrogen source instead of HNO3 or NH3. The increased activity of the samples
synthesized with nitric acid could not be explained in terms of nitrogen incorporation but by the
difference in phase composition and morphology due to the different acids applied (no acid, acetic
acid or nitric acid). Addition of a strong inorganic acid (nitric acid) resulted in different particle
morphology compared to a weak organic acid (acetic acid; cf. Figure 10) and therefore the catalytic
performance of the resulting bismuth molybdate was different.
A direct correlation of the Bi/Mo ratio on the surface determined by XPS (see Table 3) with propylene
conversion and acrolein yield was not possible. Hence, the phase composition and the surface area of
the samples determine their catalytic performance. Comparison of the samples HT_Bi2Mo1_HNO3
and HT_Bi1Mo1_HNO3 demonstrated that the catalytic performance of these two samples was
identical at 360 °C although they showed different phase composition. A synergistic effect between
the α- and the γ-phase could not be confirmed comparing HT_Bi1Mo1_HNO3 to HT_Bi2Mo1_HNO3,
which only contained γ-Bi2MoO6.
Comparison of the catalytic performance of the co-precipitated samples calcined at 450 °C showed
that γ-Bi2MoO6 (CP_Bi2Mo1_450) converted more propylene and yielded more acrolein than α-
Bi2Mo3O12 (CP_Bi2Mo3_450). Generally, comparison of catalysts prepared under the same conditions
shows that the activity decreased in the following order: γ-Bi2MoO6 > α-Bi2Mo3O12. However, a
different order of activity has been suggested in preceding studies, namely: β-Bi2Mo2O9 ≥ α-
Bi2Mo3O12 > γ-Bi2MoO6 [2, 45-47]. Krenzke and Keulks [48] as well as Batist et al. [49] observed yet
18
Page 17 of 36

another trend (γ-Bi2MoO6 ≥ β-Bi2Mo2O9 > α-Bi2Mo3O12), confirming the α-phase as the least active
phase.
CP_Bi2Mo1_680 changed its phase composition from metastable γ”-Bi2MoO6 to a mixture of
orthorhombic γ´-Bi2MoO6 and γ-Bi2MoO6. This transformation usually occurs between 550 – 640 °C
[9, 10]. Except CP_Bi2Mo1_680, HT_Bi2Mo1 and HT_Bi2Mo3 (cf. section 3.1) all other samples were
stable during the activity measurements up to 520 °C (see powder X-ray diffraction pattern in Figure
S3 Supporting Information).
For the highly active HT_Bi2Mo1_HNO3 propylene conversion and acrolein selectivity was measured
at two different oxygen concentrations (5% O2 with C3H6/O2/N2 = 5/5/90 and 25% O2 with C3H6/O2/N2
= 5/25/70; see Fig. S4 in ESI) which evidenced that at 320 °C and 360 °C the catalytic performance
(activity and selectivity) was strongly influenced by the oxygen concentration present in the reaction
gas. Higher oxygen concentration (25%) resulted in higher propylene conversion and lower acrolein
selectivity. At 400 °C and 440 °C propylene conversion was independent of the oxygen concentration.
This indicates that at temperatures below 400 °C reoxidation of the catalyst is the rate determining
step. Comparison of the XP spectra of some representative samples before and after the catalytic
test measurements did not result in a shift of the binding energies for bismuth (159.5 eV) and
molybdenum (232.6 eV) indicating that bismuth is present as Bi3+ [50] and molybdenum as Mo6+ [51]
before and after the catalytic test reaction. This shows that the used catalyst was completely
reoxidized (see Fig. S5 in ESI), which was also demonstrated in air [27]. The Bi/Mo ratio on the
surface of the used samples were lower than the corresponding ratio of the fresh bismuth
molybdates (cf. Table 3 and Figure S5 Supporting Information) which can be explained by the
relatively high temperatures in the reactor (up to 520 °C; cf. 3.3.1). XPS of the used catalysts also
proofed that coke formation on these bismuth molybdates can be neglected as the carbon amount
on the surface of the catalyst before and after the catalytic tests did only differ in the range of the
measurement inaccuracy.
19
Page 18 of 36

4. Conclusion
The catalytic performance of bismuth molybdates strongly depends on the crystalline phases present
in the catalysts and their morphologies which emerge from the selected synthetic method. During
hydrothermal synthesis application of a high Bi/Mo ratio led mainly to the formation of γ-Bi2MoO6,
while applying a Bi/Mo ratio < 1 resulted in α-Bi2Mo3O12 as the main phase. A mixture of α- and γ-
phase was obtained with precursor ratios of Bi/Mo = 1:1. As hydrothermal synthesis did not provide
access to the one-step formation of β-Bi2Mo2O9 under the parameters applied in the present study,
the materials synthesized in pure water were furthermore calcined at 550 °C. Even these harsh
conditions resulted only in partial formation of the β-phase, and the process can probably be driven
to completion only at the expense of considerable loss of surface area and catalytic performance.
The catalytic activity of the freshly hydrothermally prepared catalysts was significantly higher than
after their calcination at 550°C. Furthermore, the obtained acrolein yield decreased regardless of the
phase composition. Generally, for the investigation of the relative activities of the different bismuth
molybdate phases catalysts should be prepared under the same conditions including pre-
conditioning of the precursor materials. Therefore we used as prepared materials or applied the
same heat treatment to compare the catalytic performance of the different bismuth molybdate
phases. In addition, highly active catalysts could be prepared under hydrothermal conditions applying
initial Bi/Mo ratios of 1:1 or 2:1 with nitric acid or ammonia as additives. Generally, these catalysts
were more selective at comparable propylene conversion than co-precipitated γ-Bi2MoO6.
Consequently, extension of this method to industrially more relevant Bi-Mo-O systems containing,
e.g. Co, Fe or V as additional elements is a promising concept.
ACKNOWLEDGMENTS
We thank the Danish Research Council for financial support in the framework of the DSF proposal
“Nanoparticle synthesis for catalysis”. We gratefully acknowledge Hermann Köhler at the Institute of
Catalysis Research and Technology (IKFT) at KIT for performing ICP-OES measurements and Dr.
Thomas Bergfeldt at the Institute for Applied Materials in the Chemical Analysis group (IAM-AWP) at
20
Page 19 of 36

KIT for the quantitative nitrogen analysis. The Center for Electron Nanoscopy (CEN) at DTU is
gratefully acknowledged for the possibility to use the scanning electron microscopes (SEM) and the
help and support during the measurements. G. R. P. acknowledges financial support by the Swiss
National Science Foundation (SNSF Professorship PP00P2-13348/1) and by the University of Zurich.
Electronic supporting information:
The supporting information contains a representative XPS (before and after use), Raman spectra of
samples prepared by hydrothermal synthesis and co-precipitation. Moreover, powder X-ray
diffraction patterns of the samples before and after application in selective oxidation (cf. section 3.3)
and the catalytic performance in the selective oxidation of propylene at different oxygen
concentrations are reported.
References
[1]
J.L. Callahan, R.K. Grasselli, E.C. Milberge, H.A. Strecker, Erdöl und Kohle Erdgas, Petrochem.
24 (1971) 349.
[2]
R.K. Grasselli, J.D. Burrington, Adv. Catal. 30 (1981) 133-163.
[3]
[4]
Y. Morooka, W. Ueda, Adv. Catal. 40 (1994) 233-273.
L.D. Krenzke, G.W. Keulks, J. Catal. 64 (1980) 295-302.
[5]
[6]
[7]
W. Ueda, Y. Morooka, T. Ikawa, I. Matsuura, Chem. Lett. (1982) 1365-1368.
S.R.G. Carrazán, C. Martin, V. Rives, R. Vidal, Appl. Catal., A. 135 (1996) 95-123.
D.H. He, W. Ueda, Y. Morooka, Catal. Lett. 12 (1992) 35-44.
[8]
[9]
H. Li, K. Li, H. Wang, Mater. Chem. Phys. 116 (2009) 134-142.
D.J. Buttrey, T. Vogt, U. Wildgruber, W.R. Robinson, J. Solid State Chem. 111 (1994) 118-127.
D.J. Buttrey, D.A. Jefferson, J.M. Thomas, Philos. Mag. A. 53 (1986) 897-906.
J.R. Monnier, G.W. Keulks, J. Catal. 68 (1981) 51-66.
Z. Bing, S. Pei, S. Shishan, G. Xiexian, J. Chem. Soc., Faraday Trans. 86 (1990) 3145-3150.
J.D. Burrington, R.K. Grasselli, J. Catal. 59 (1979) 79-99.
D. Carson, M. Forissièr, J.C. Védrine, J. Chem. Soc., Faraday Trans. 1. 80 (1984) 1017-1028.
M.T. Le, W.J.M. van Well, P. Stoltze, I. van Driessche, S. Hoste, Appl. Catal., A. 282 (2005) 189-
[10]
[11]
[12]
[13]
[14]
[15]
194.
[16]
[17]
[18]
[19]
G.W. Keulks, J.L. Hall, C. Daniel, K. Suzuki, J. Catal. 34 (1974) 79-97.
F. Trifiro, R.D. Scarle, H. Hoser, J. Catal. 25 (1972) 12.
R.P. Rastogi, A.K. Singh, C.S. Shukla, J. Solid State Chem. 42 (1982) 136-148.
L.M. Thang, L.H. Bac, I. van Driessche, S. Hoste, W.J.M. van Well, Catal. Today. 131 (2008)
566-571.
[20]
[21]
1-8.
[22]
[23]
[24]
[25]
[26]
M.T. Le, J. Van Craenenbroeck, I. Van Driessche, S. Hoste, Appl. Catal., A. 249 (2003) 355-364.
W.J.M. van Well, M.T. Le, N.C. Schiodt, S. Hoste, P. Stoltze, J. Mol. Catal. A: Chem. 256 (2006)
J.Q. Yu, A. Kudo, Chem. Lett. 34 (2005) 1528-1529.
H. Kodama, F. Izumi, J. Cryst. Growth. 50 (1980) 515-520.
K. Zheng, Y. Zhou, L. Gu, X. Mo, G.R. Patzke, G. Chen, Sens. Actuators, B. 148 (2010) 240-246.
A.M. Beale, G. Sankar, Chem. Mater. 15 (2003) 146-153.
K.L. Parry, A.G. Shard, R.D. Short, R.G. White, J.D. Whittle, A. Wright, Surf. Interface Anal. 38
(2006) 1497-1504.
21
Page 20 of 36

[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
270.
[35]
[36]
[37]
[38]
[39]
[40]
J.-D. Grunwaldt, M.D. Wildberger, T. Mallat, A. Baiker, J. Catal. 177 (1998) 53-59.
A. Ayame, K. Uchida, M. Iwataya, M. Miyamoto, Appl. Catal., A. 227 (2002) 7-17.
J.H. Scofield, J. Electron. Spectrosc. Relat. Phenom. 8 (1976) 129-137.
M. Høj, A.D. Jensen, J.-D. Grunwaldt, Appl. Catal., A. 451 (2013) 207-215.
http://icsd.fiz-karlsruhe.de/ (October 2013).
F. Pertlik, J. Zemann, Fortschr. Mineral. 60 (1982) 162-163.
B. Begemann, M. Jansen, J. Less. Common Met. 156 (1989) 123-135.
A.N. Christensen, M.-A. Chevallier, J. Skibsted, B.B. Iversen, J. Chem. Soc. Dalton (2000) 265-
F. Theobald, A. Laarif, A.W. Hewat, Ferroelectr. 56 (1984) 219-237.
F.D. Hardcastle, I.E. Wachs, J. Phys. Chem. 95 (1991) 10763-10772.
R.P. Oertel, R.A. Plane, Inorg. Chem. 7 (1968) 1192-1196.
B. Grzybowska, J. Haber, J. Janas, J. Catal. 49 (1977) 150-163.
H.E. Swift, J.E. Bozik, J.A. Ondrey, J. Catal. 21 (1971) 212.
E. Vila, A.R. Landa-Canovas, J. Galy, J.E. Iglesias, A. Castro, J. Solid State Chem. 180 (2007)
661-669.
[41]
P.J. Bereciartua, F.J. Zuniga, J.M. Perez-Mato, V. Petricek, E. Vila, A. Castro, J. Rodriguez-
Carvajal, S. Doyle, Acta Crystallogr., Sect. B: Struct. Sci. 68 (2012) 323-340.
[42]
[43]
[44]
J.M. Herrmann, M. Eljamal, M. Forissier, React. Kinet. Catal. Lett. 37 (1988) 255-260.
M.T. Le, W.J.M. van Well, I. van Driessche, S. Hoste, Can. J. Chem. Eng. 83 (2005) 336-343.
A.P.V. Soares, L.D. Dimitrov, M. de Oliveira, L. Hilaire, M.F. Portela, R.K. Grasselli, Appl. Catal.,
A. 253 (2003) 191-200.
[45]
[46]
[47]
[48]
[49]
[50]
J.F. Brazdil, D.D. Suresh, R.K. Grasselli, J. Catal. 66 (1980) 347-367.
E. Godard, E.M. Gaigneaux, P. Ruiz, B. Delmon, Catal. Today. 61 (2000) 279-285.
I. Matsuura, R. Schut, K. Hirakawa, J. Catal. 63 (1980) 152-166.
L.D. Krenzke, G.W. Keulks, J. Catal. 61 (1980) 316-325.
P.A. Batist, J.F.H. Bouwens, G.C.A. Schuit, J. Catal. 25 (1972) 1.
L. Chen, E. Aarcon-Lado, M. Hettick, I.D. Sharp, Y. Lin, A. Javey, J.W. Ager, J. Phys. Chem. C.
117 (2013) 21635-21642.
[51] J.G. Choi, L.T. Thompson, Appl. Surf. Sci. 93 (1996) 143-149.
22
Page 21 of 36

Tables
Table 1: Characterization of hydrothermally prepared samples with different Bi/Mo ratios before and
after the catalytic tests by PXRD, Raman spectroscopy and BET (main phases in bold letters).
Sample
Phases according to PXRD
Phases according to Raman
spectroscopy
Specific
surface area
(BET) [m²/g]
as-prepared
after use
as-prepared
after use
as-prepared
γ-Bi2MoO6,
cubic BiO2-x
γ-Bi2MoO6,
Bi8Mo3O21
γ-Bi2MoO6,
γ´-Bi2MoO6
HT_Bi3Mo1
γ-Bi2MoO6
30
γ-Bi2MoO6, α-
Bi2Mo3O12,
Bi6O6(OH)3(NO3)3∙
1.5H2O
γ-Bi2MoO6,
HT_Bi2Mo1
HT_Bi1Mo1
HT_Bi2Mo3
γ-Bi2MoO6
presence of
γ- Bi2MoO6
n.d.
6
8
7
-
NO3
γ-Bi2MoO6 and α- γ-Bi2MoO6 and
α-Bi2Mo3O12
Bi2Mo3O12
α-Bi2Mo3O12
α-Bi2Mo3O12,
cubic
H0.68(NH4)2Mo14.16
O4.34∙6.92H2O
α-Bi2Mo3O12,
hexagonal
α-Bi2Mo3O12,
α-Bi2Mo3O12
α-Bi2Mo3O12
MoO3
α-Bi2Mo3O12,
α-MoO3
α-Bi2Mo3O12,
NH3(MoO3)3
α-Bi2Mo3O12,
HT_Bi1Mo2
6
MoO3
NH3(MoO3)3
23
Page 22 of 36

Table 2: Characterization of hydrothermally prepared samples before and after calcination at 550 °C
for 4 h. Phases were identified by PXRD, the ratio of bismuth to molybdenum in the bulk and on the
surface were calculated from ICP-OES (atom %) and XPS analysis.
Sample
Phases according to PXRD
Bi/Mo ratio
bulk
Bi/Mo ratio
surface
as-prepared
Calcined at
550°C for 4 h
Calcined and as-prep. calc as-prep. calc
used
γ-Bi2MoO6, α-
Bi2Mo3O12,
Bi6O6(OH)3(NO3)3∙
1.5H2O
γ-Bi2MoO6,
some β-
Bi2Mo2O9
HT_Bi2Mo1
HT_Bi1Mo1
HT_Bi2Mo3
γ-Bi2MoO6
2.0
n.d.
0.7
2.0
n.d.
0.7
3.0
1.3
0.8
1.9
n.d.
0.7
γ-Bi2MoO6 and
α-Bi2Mo3O12
γ-Bi2MoO6 and γ-Bi2MoO6 and
α-Bi2Mo3O12
α-Bi2Mo3O12
α-Bi2Mo3O12,
cubic
H0.68(NH4)2Mo14.16
O4.34∙6.92H2O
α-Bi2Mo3O12
α-Bi2Mo3O12
α-Bi2Mo3O12,
hexagonal
NH3(MoO3)3
α-Bi2Mo3O12,
α-Bi2Mo3O12,
HT_Bi1Mo2
0.5
n.d.
n.d.
n.d.
MoO3
MoO3
24
Page 23 of 36

Table 3: Characterization and catalytic data of the differently prepared samples. All samples are as-
prepared and non-calcined except the co-precipitated samples (calcination temperature is included
in the sample name).
Sample
Maximum Corresponding
T
Specific
Bi/Mo
Bi/Mo ratio
surface b
yield
(C3H4O)
Conversion
(C3H6)
surface ratio bulk a
area
[at.%/at.%]
(BET)
[at.%/at.%]
fresh
%
%
°C
[m²/g]
fresh
used
CP_Bi2Mo1_450
CP_Bi2Mo1_680
25.3
5.8
33.9
7.3
491
482
6
2.0
2.3
2.1
n.d.
1.9
< 1
n.d.
n.d.
HT_Bi2Mo1_HNO3
HT_Bi2Mo1_HOAc
HT_Bi2Mo1
33.6
20.1
23.1
3.6
41.2
26.8
33.3
4.3
450
405 c
445
521
486
489
493
483
443
6
5
1.8
n.d.
2.0
2.8
n.d.
3.0
n.d.
2.6
6
CP_Bi1Mo1_680
HT_Bi1Mo1_HNO3
HT_Bi1Mo1_pH4 d
HT_Bi1Mo1
< 1
9
n.d.
n.d.
n.d.
n.d.
0.7
n.d.
1.2
n.d.
1.1
38.4
36.6
31.0
15.3
21.9
43.8
45.5
39.9
18.8
28.9
7
n.d.
1.3
n.d.
n.d.
n.d.
0.7
8
CP_Bi2Mo3_450
HT_Bi2Mo3
1
n.d.
0.8
7
0.7
a: The calculated error accounts to around 10%
b: Average of two XPS measurements at different spots
c: Maximum test temperature 400 °C for HT_Bi2Mo1_HOAc
25
Page 24 of 36

d: pH value adjusted with NH3
Figures
Figure 1: Characterization of the samples prepared with different Bi/Mo ratio under hydrothermal
conditions without calcination by X-ray diffraction (a) and Raman spectroscopy (b). The different
phases in pattern (a) are indicated as follows: α-Bi2Mo3O12 (triangles), γ-Bi2MoO6 (circles), NH3(MoO3)3
(squares), BiO2-x (diamond).
26
Page 25 of 36

Figure 2: SEM images of hydrothermally synthesized samples with different Bi/Mo ratio without
further calcination a-b) HT_Bi3Mo1, c) HT_Bi2Mo1, d) HT_Bi1Mo2, e-f) HT_Bi1Mo1, g-h) HT_Bi2Mo3.
27
Page 26 of 36

Figure 3: Catalytic performance of hydrothermally prepared samples with different Bi/Mo ratios at
320 °C (a) 360 °C (b) and 400 °C (c) without calcination. 500 mg of the as prepared samples (150 –
300 µm) were pre-treated in the reactor in synthetic air at 300°C and the catalytic performance was
measured using a gas composition of C3H6/O2/N2 = 5/25/70 and flows of 50, 80, 120 Nml/min.
28
Page 27 of 36

Figure 4: PXRD pattern of hydrothermally synthesized samples after propylene oxidation at
temperatures up to 520 °C, illustrating that no major changes occurred in the main phase after
catalysis; α-Bi2Mo3O12 (triangles), γ-Bi2MoO6 (circles), α-MoO3 (squares) and Bi8Mo3O21 (diamond).
29
Page 28 of 36

Figure 5: SEM images of a) HT_Bi3Mo1, b) HT_Bi2Mo3 and c-d) HT_Bi2Mo1 after catalytic propylene
oxidation at temperatures up to 520 °C.
30
Page 29 of 36