Ethanol aerobic and anaerobic oxidation with FeVO4 and V2O5 catalysts

Article abstract of DOI:10.1016/j.apcata.2018.11.013

This study compares the aerobic and anaerobic transformation of ethanol using FeVO₄ and V₂O₅ catalysts. Despite their different structure, the two oxides showed very similar catalytic performances and their main product was acetaldehyde. However, in the absence of oxygen, the catalysts produced an equimolar amount of ethane and acetaldehyde, and this aspect has been little studied in the literature. In-situ XPS and DRIFT spectroscopy studies showed that the active species for the disproportionation of the alcohol into ethane and aldehyde was the reduced V³⁺ ion; nevertheless, the Fe in the FeVO₄ catalysts was responsible for directing the reduction of metals toward the formation of a Fe-V-O spinel phase which was homogeneous and more stable than V₂O₅. Moreover, an in-situ DRIFT spectroscopy study showed that ethanol adsorbs in different ways on the surface of the catalysts during the reduction of samples (anaerobic reaction), forming H-bonded and dissociated ethoxy species, and giving rise to new surface OH groups that participate in the aldehyde/alkane formation. To conclude, a new mechanism of hydrogen transfer for the anaerobic ethanol disproportionation into ethane and acetaldehyde is proposed. This research completes the picture about ethanol oxidation to acetaldehyde on V-based catalysts, demonstrating that the catalytic behavior is mainly affected by the oxidation degree of the vanadium species which, in turn, depends on the reaction environment, and not on the structure itself.

Full text of DOI:10.1016/j.apcata.2018.11.013

Accepted Manuscript
Title: Ethanol aerobic and anaerobic oxidation with FeVO4
and V2O5 catalysts
Authors: Andrea Malmusi, Juliana Velasquez Ochoa,
Tommaso Tabanelli, Francesco Basile, Carlo Lucarelli,
Stefano Agnoli, Francesco Carraro, Gaetano Granozzi,
Fabrizio Cavani
PII:
S0926-860X(18)30567-2
DOI:
Reference:
https://doi.org/10.1016/j.apcata.2018.11.013
APCATA 16882
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
16 September 2018
13 November 2018
14 November 2018
Please cite this article as: Malmusi A, Velasquez Ochoa J, Tabanelli T, Basile F,
Lucarelli C, Agnoli S, Carraro F, Granozzi G, Cavani F, Ethanol aerobic and anaerobic
oxidation with FeVO4 and V2O5 catalysts, Applied Catalysis A, General (2018),
https://doi.org/10.1016/j.apcata.2018.11.013
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.

Ethanol aerobic and anaerobic oxidation with FeVO4 and V2O5 catalysts
Andrea Malmusi,1,2,§ Juliana Velasquez Ochoa,1,§ Tommaso Tabanelli,1 Francesco Basile1, Carlo
Lucarelli,3 Stefano Agnoli,4 Francesco Carraro,4 Gaetano Granozzi,4 Fabrizio Cavani1,2
1Dipartimento di Chimica Industriale “Toso Montanari”, Università di Bologna, Viale Risorgimento
4, 40136 Bologna (Italy)
2Consorzio INSTM, Research Unit of Bologna, Firenze (Italy)
3Dipartimento di Scienze ed Alta Tecnologia, Università degli Studi dell’Insubria, Via Valleggio 9,
22100 Como (Italy)
4Dipartimento di Scienze Chimiche, Università di Padova, via Marzolo 1, Padova 35131 (Italy)
§ These authors contributed equally
Graphical abstract
Highlights
 FeVO4 is active and selective in the oxidative dehydrogenation of ethanol to acetaldehyde
 In the absence of oxygen it reduces to a spinel compound containing V3+ and Fe2+/Fe3+ ions
 The spinel catalyzes the disproportionation of ethanol to ethane and acetaldehyde
1

 The spinel shows a catalytic performance similar to that one V2O3, but the former is more
stable
 A mechanism for the intermolecular disproportionation of ethanol is proposed
Abstract: This study compares the aerobic and anaerobic transformation of ethanol using FeVO4
and V2O5 catalysts. Despite their different structure, the two oxides showed very similar catalytic
performances and their main product was acetaldehyde. However, in the absence of oxygen, the
catalysts produced an equimolar amount of ethane and acetaldehyde, and this aspect has been little
studied in the literature. In-situ XPS and DRIFT spectroscopy studies showed that the active species
for the disproportionation of the alcohol into ethane and aldehyde was the reduced V3+ ion;
nevertheless, the Fe in the FeVO4 catalysts was responsible for directing the reduction of metals
toward the formation of a Fe-V-O spinel phase which was homogeneous and more stable than
V2O5. Moreover, an in-situ DRIFT spectroscopy study showed that ethanol adsorbs in different
ways on the surface of the catalysts during the reduction of samples (anaerobic reaction), forming
H-bonded and dissociated ethoxy species, and giving rise to new surface OH groups that participate
in the aldehyde/alkane formation. To conclude, a new mechanism of hydrogen transfer for the
anaerobic ethanol disproportionation into ethane and acetaldehyde is proposed. This research
completes the picture about ethanol oxidation to acetaldehyde on V-based catalysts, demonstrating
that the catalytic behavior is mainly affected by the oxidation degree of the vanadium species
which, in turn, depends on the reaction environment, and not on the structure itself.
Keywords: ethanol oxidation; ethanol disproportionation; iron vanadate; acetaldehyde; ethane
Introduction
The development of a new bio-based chemical industry is based mainly on the valorization of bio-
alcohols, mainly ethanol, due to the possibility of obtaining it directly from biomass fermentation.
The gas-phase selective oxidation of ethanol to acetaldehyde is still the subject of investigation and
has been studied with a wide range of catalysts [1–8]; vanadium and molybdenum oxides, however,
continue to be the preferred catalysts due to their high activity and selectivity [9–14]. In the
2

oxidation of alcohols, bulk vanadates show a greater activity and less likelihood of decomposition
at higher temperatures compared to bulk mixed metal molybdates. Specifically, vanadates such as
FeVO4 are interesting because of their greater activity and structural stability; in fact, the
incorporation of Fe causes the stabilization of V, decreasing its volatility [12]. As for supported
catalysts, sometimes metal vanadates such as FeVO4 and CeVO4 are preferred to V2O5 in
applications such as the selective catalytic reduction (SCR) of nitric oxide, due both to their higher
melting point (850°C and 1100°C, respectively) compared to bulk V2O5, and to their higher
resistance to deactivation after impregnation of SiO2-WO3-TiO2; with bulk vanadium oxide, V ions
become mobile already over 690°C [15].
However, most of the studies on the transformation of alcohols to aldehydes with vanadium-based
oxides have been conducted under aerobic conditions (to perform the so-called oxidative
dehydrogenation, ODH), and little attention has been paid to their behavior in an inert/reducing
atmosphere. Nevertheless, it was recently demonstrated that some vanadium (and molybdenum)
oxides catalyze the formation of ethane and acetaldehyde from ethanol in equivalent amounts
[16,17]. Also, the formation of some ethane from ethanol was observed by us in the case of Fe-
containing catalysts [18]. Therefore, it was interesting to study the behavior of a mixed Fe-V oxide
catalyst because the two metals might have a synergic effect during the anaerobic ethanol
transformation.
We conducted a systematic study of the catalytic behavior of FeVO4 under different conditions
(aerobic and anaerobic, with and without steam), and performed the catalyst characterization before
and after reaction, compared to reference V2O5. A key point of the research, however, was the
possibility to study materials surface under real working conditions (in-situ), since it was observed
that once in contact with air, the reduced phases were easily reoxidized: an event which is an
intrinsic limitation for the interpretation of most data in literature, since they are often obtained ex-
situ.
3

A surprising result of our study was that the catalytic performance of FeVO4 and V2O5, after the
equilibration and development of the stable structure under working conditions, was very similar,
despite the formation of completely different crystalline structures. This similarity was not due to a
surface enrichment in V, as in the case of methanol oxidative dehydrogenation to formaldehyde [9]
and the oxidative dehydrogenation of ethanol with V-Fe-Sb-O catalysts [19]. In fact, when deprived
of oxygen, FeVO4 was reduced by ethanol, thus forming a spinel structure where V was present as
V3+; the latter was the active species for ethanol disproportionation to acetaldehyde and ethane.
Experimental
Catalyst preparation
FeVO4 was prepared by co-precipitation in an aqueous basic environment, followed by filtration,
drying, and calcination. Briefly, a solution made by dissolving 23.5 g Fe(NO3)3 nonahydrate in 50
mL of distilled water was added to a solution made of 6.7 g NH4VO3 and 2.9 g H2C2O4 dissolved in
50 mL of water to obtain a Fe/V atomic ratio equal to 1:1. Afterwards, the pH was adjusted to 6.8
by adding 25% ammonia solution; a precipitate was obtained, which was then aged for 1h, filtrated,
washed in 2 L of distilled water and dried at 120°C overnight. Finally, the solid was calcined at
650°C for 3 hours. The surface area after calcination was 8 m2/g.
V2O5 99.9% and Fe3O4 97%, 100-50 nm particle size, were provided by Sigma-Aldrich. The same
metal oxides were also synthesized, in order to compare their redox properties with those of
corresponding commercial samples. The magnetite sample (Fe3O4) was synthesized by co-
precipitation at 50°C from an aqueous solution of FeSO4·7H2O and FeCl3 with NaOH added
dropwise until pH>10. Afterward, the solution was vigorously stirred for 3 h, and the precipitate
was washed with water and filtered under vacuum. The drying temperature was 80 °C, to avoid the
oxidation of Fe2+. Lastly, the sample was annealed at 450 °C for 8 h in N2 flow.
4

Vanadium pentoxide (V2O5) was synthesized by calcination of ammonium metavanadate at 650 °C
for 3h.
Before use all catalysts were pelletized, crushed, and sieved to obtain a material with particle size
between 0.395 and 0.400 mm.
Catalyst characterization
Fresh and spent catalysts were characterized by means of X-Ray Diffraction (XRD), Raman, and IR
spectroscopy. XRD patterns of catalyst powders were collected on a Bragg/Brentano Philips PV
1710 operating with a Cu Kα wavelength; the signal was acquired between 5 and 80° 2 with an
acquisition time of 1 s every 0.1° 2. The specific surface area of fresh catalysts was determined by
the one-point BET technique on a Fisons Sorpty 1750. SEM images were collected and analyzed by
an EDX probe to estimate element distribution and molar ratio. Data were registered with a Zeiss
EP EVO 50, equipped with an Oxford Instruments INCA ENERGY 350 EDX probe. General
conditions of the microscope were: EHT 20 KeV, high vacuum (10-6 Pa) or variable vacuum
between 60 and 100 Pa. The EDX probe works with Mn Kα radiation with a resolution of 133 eV.
H2 Temperature-Programmed Reduction (TPR) was performed in an Autochem II 2920 -
TPD/TPR/TPO Micromeritics instrument. 100–200 mg of catalysts were pre-treated at 200°c for 30
min under He flow. After cooling down to 50°C, a reducing stream of 5% H2 in Ar was sent at 30
ml min-1. The temperature increased at 10 °C min-1 until 800°C and hold there for further 20
minutes.
Pseudo in-situ XPS
Pseudo in-situ X-ray photoemission spectroscopy (XPS) data were acquired using non-
monochromatized Mg Kα (1253.6 eV) X-ray source. The measurements were carried out in an
ultra-high-vacuum (UHV) system (base pressure of 5x10-10 mbar) equipped with a custom-made
high-pressure cell (base pressure: 10-7 mbar) for sample heating and controlled gas dosing. The
5

spectra were acquired at RT on drop casted films of powder samples from methanol dispersions.
For each sample O 1s and V 2p and/or Fe 2p photoemission lines and V L23M23M45 Auger
spectrum were recorded sequentially after introduction in the UHV systems (fresh), after annealing
in UHV for 15 min at 300°C (to check if a temperature induced reduction was present), after
exposure in static condition to ethanol (5 mbar) at 300°C for 5 and 15 minutes. After each step, the
sample was transferred back in the UHV analysis chamber and the photoemission spectra were
acquired. Moreover, the V 2p3/2 photoemission lines were separated into individual components
(after Shirley background removal) using symmetrical Voigt functions and non-linear least squares
routines for the χ2 minimization. Finally, photoemission spectra of spent catalysts (ex-situ spent)
measured as received were collected with the same procedure.
DRIFT –MS experiments
The IR apparatus used was a Bruker Vertex 70 with a Pike DiffusIR cell attachment. Spectra were
recorded using a MCT detector after 128 scans and 2 cm−1 resolution. The mass spectrometer was
an EcoSys-P from European Spectrometry Systems. Samples were pre-treated at 450°C in a He
flow (10 ml min-1) for 45 min, in order to remove any molecules adsorbed on the material. Samples
were then cooled down to 300°C and ethanol was fed at 0.6 ml min-1, vaporized, and sent to the
DRIFTS cell using either He or air as the carrier gas. Spectra were recorded continuously, every
minute. The following selected mass spectroscopy signals (m/z) were monitored continuously with
time (and temperature): 2, 16, 25, 28, 29, 30, 31, 40, 41, 43, 44, 45, 56, 58, 59, 60, and 61. By
combining the information obtained from several different m/z signals, it was possible to obtain
unambiguous information on the various products formed.
Catalytic tests
The catalytic activity was evaluated in a lab-scale fixed-bed glass reactor, at atmospheric pressure.
Ethanol or ethanol-water azeotropic mixtures were continuously fed to the reactor by a perfusion
6

pump which injects the fluid into a vaporization chamber connected to the reactor. The carrier gas
(nitrogen) and, if needed, oxygen were fed to the reactor by a Brooks 5850E mass-flow meter.
Typical reaction conditions were as follows: time factor (measured at room temperature) 0.5 g s/ml,
total gas flow 60 ml/min; feed 5% ethanol in N2. After approximately 2 h reaction time, data were
taken at each temperature level.
Unconverted ethanol and reaction products were determined by on-line chromatography on a
Agilent 7890A GC equipped with an HP-Molesieve and an HP-PLOT U connected with a TC
detector.
A blank test was conducted by feeding the reaction mixture to the empty reactor; the thermal
dehydration of ethanol to ethylene was observed at 350°C and 400°C with a product yield of 1%
and 5% at the two temperatures, respectively.
Ethanol conversion is calculated by comparing the number of moles of reactant at the reactor outlet
per unit time with the number of moles of reactant at the reactor inlet. Yield of products is
calculated by dividing the number of moles of each product by the number of moles of reactant at
the reactor inlet; the value also includes the ratio of stoichiometric coefficients. Selectivity to each
products is calculated by dividing the corresponding yield by the reactant conversion. C balance is
calculated by dividing the sum of selectivities to all products; C loss (or “missing C”) is the
difference: 1 – C balance.
Results and Discussion
Reactivity under anaerobic conditions
Figure 1 shows the catalytic behavior of FeVO4 at increasing temperatures under anaerobic
conditions. It can be seen that the main products were ethane and acetaldehyde (both with a
selectivity in the 20-43% range) and that the temperature rise led to an increased ethanol conversion
and selectivity of some minor products, such as ethyl acetate. When the same experiment was
performed with bulk V2O5 and Fe3O4 (Figures 2 and 3), the behavior of the former resembled that
7

of the FeVO4 catalyst, with some small differences such as the lesser formation of C4 compounds
and the poorer carbon balance observed with the V2O5. Conversely, the main products with Fe3O4
were acetone (at low temperature), ethyl acetate and acetaldehyde (at intermediate temperature),
and ethylene, acetone, and CO2 at high temperature. CO formed with a selectivity of 2% at 400°C;
other C4 compounds were absent; C balance in this case was between 50 and 65%.
In order to understand the similar performance shown, the two catalysts were evaluated at the same
temperature (300°C), during longer time-on-stream. It was observed that FeVO4 showed a more
pronounced variation in performance during the first hour, whereas V2O5 was more stable (Figure
4). Nevertheless, at steady state, the product distribution was similar for both materials, but the
former catalyst showed a slightly higher selectivity to ethane (49 vs 43% for V2O5). The fact that
the product selectivity was practically independent from the reaction temperature and ethanol
conversion clearly indicates that ethane and acetaldehyde were not formed by consecutive reactions,
but in a single step from a common intermediate; this is in line with our previous observations in
the case of ferrites [18] and with the studies of Ueda and co-workers in the case of vanadium oxides
[16,17] for a reaction with the following stoichiometry:
2CH3CH2OH  CH3CHO + CH3CH3 + H2O.
This result also rules out the formation of alkane by a consecutive reaction such as the
hydrogenation of ethylene. In literature, it has been proposed that a surface enrichment of V in
FeVO4 during the oxidation of methanol to formaldehyde (in aerobic conditions) is responsible for
catalytic activity [9]. In the absence of oxygen, however, our catalyst changed its structure during
the reaction with ethanol, and the catalyst used no longer showed the triclinic structure of FeVO4
but a spinel phase, as evidenced by the comparison of XRD patterns for fresh (top) and spent
catalysts (middle) in Figure 5. Moreover, there was no segregation of other crystalline phases for
single metals (FeOx or VOx).
8

Figure 1. Effects of temperature on ethanol conversion and product distribution with FeVO4
catalyst. Reaction conditions: feed 5% ethanol in N2; W/F 0.5 g s/ml. Symbols: ethanol conversion
(), selectivity to acetaldehyde (), ethane (), butyraldehyde (), crotyl alcohol (),
crotonaldehyde (), acetone (), ethylacetate (), and CO2 (). CO formed in traces only.
Results were taken at each temperature after the steady state performance had been reached.
Figure 2. Effects of temperature on ethanol conversion and product distribution with V2O5 catalyst.
Reaction conditions: feed 5% ethanol in N2; W/F 0.5 g s/ml. Symbols: ethanol conversion (),
selectivity to acetaldehyde (), ethane (), butyraldehyde (), crotonaldehyde (), acetone (),
ethylacetate (), and CO2 (). Only by-products with a selectivity higher than 1% are shown.
Results were taken at each temperature after the steady state performance had been reached.
9

Figure 3. Effects of temperature on ethanol conversion and product selectivity with Fe3O4 catalyst.
Reaction conditions: feed 5% ethanol, rest N2; W/F 0.5 g s/ml. Symbols: ethanol conversion (),
selectivity to acetaldehyde (), ethylene (), acetone (), ethylacetate (), and CO2 ().Results
were taken at each temperature after the steady state performance had been reached.
Figure 4. Effects of time-on-stream on ethanol conversion and product distribution with FeVO4
(left) and V2O5 (right) catalysts. Reaction conditions: T 300°C, feed 5% ethanol in N2; W/F 0.5 g
s/ml. Symbols: ethanol conversion (), selectivity to acetaldehyde (), ethane ()
In fact, the pattern of the spent FeVO4 was characterized by broad and intense peaks at
approximately 30.5, 35.5, 43.5, 58, and 63 2 degrees, similar to those for magnetite Fe3O4 and -
Fe2O3 (maghemite) [20,21]. The structural formula of Fe3-xVxO4 spinels (0x2) has been described
3+
as (Fe2+Fe1- 3+)A(Fe1- 2+Fe1- 3+Vx )BO4, with =x/2, A, and B representing tetrahedral and



octahedral sites, respectively [22–26]. Thus, if V is in 3+ oxidation state and Fe/V ratio is still 1:1,
the stoichiometry of this spinel would be 퐹푒 퐹푒 푉3+. A further characterization with Raman
2+
3+
1.0
0.5 1.5
(Figure S1 in Supplementary Information) and FTIR spectroscopy (Figure S2) showed that the
reduced phase formed in-situ is a pure single phase. In fact, FTIR spectrum after reaction showed
the characteristic bands of a Fe-V spinel structure: in particular, the positions of the bands at around
598 and 470 cm-1 correspond to those reported in literature for spinel compounds of the type Fe3-
10

xVxO4 with x>1 [27]. On the other hand, the XRD pattern for the V2O5 after reaction (Figure 5,
bottom) showed the formation of a peculiar phase that does not fit exactly with any of the known
polymorphs of the reduced vanadium oxides (VO2, V2O3 or VO) or suboxides (V6O13, V4O9).
Conversely, this phase might correspond to a newly described polymorph called ω-VO2. Details as
to the nature of this new V4+ oxide crystalline compound are being studied more in detail by our
group.
Figure 5. XRD patterns of FeVO4 catalyst fresh (top), and used after reaction at 300°C (center), and
comparison with V2O5 after reaction at 300°C (bottom).
In order to confirm the oxidation state of the metals in used catalysts, pseudo in-situ XPS
measurements were performed (Figure 6, S12). Such characterization is fundamental to avoid air-
induced re-oxidation of the catalysts. We have first analyzed the as prepared samples: from the
examination of the V 2p photoemission line (having a V 2p3/2 binding energy (BE) of 517.1 eV) we
confirm an oxidation state of V+5 in both materials [28]. Moreover, in the case of FeVO4, the BE of
Fe 2p3/2 (711.5 eV) and the occurrence of the characteristic satellites peaks confirm the presence of
Fe3+ [29].
11

After the exposure of FeVO4 to ethanol at 300°C for 5 minutes we observed that most of the Fe3+
species are reduced to Fe2+ and that the V5+ ions are reduced to V3+/V4+ (515.7 eV). Similarly, in the
case of the V2O5, the V species are reduced to V3+/V4+. Longer exposure to ethanol at 300°C did not
show further reduction of Fe and V (as an example the Fe 2p photoemission line of FeVO4 after 15
minutes of ethanol exposure is reported in Figure 6). It should be noted that on the basis of
photoemission data, it is rather difficult to distinguish between V3+ and V4+ species, since V2O3 and
VO2 show similar multiplets (see the deconvolution into single chemically shifted components of
the V 2p3/2 photoemission lines in Figure S12) [30]. A more detailed picture can be obtained by the
analysis of the Auger V LMM spectra (Figure 7), whose line shape is diagnostic of the V oxidation
state: it is evident that after the exposure to ethanol, the FeVO4 sample exhibits a lower residual
amount of V5+ species with respect to V2O5 [31]. Actually, in reduced FeVO4 the V3+ component is
largely predominant, whereas in reduced V2O5 the V5+ component (at a binding energy of 787.9 eV)
is more intense than that of reduced V species (i.e. V3+/V4+ at a binding energy of 780.7 eV,).
Indeed, the Auger spectrum of the reduced V2O5 phase corresponds well to that reported in the
literature for VO2 [31]. Therefore, it can be supposed that the presence of Fe drives the reduction of
the oxides toward the formation of a spinel compound, where V is stabilized as V3+, whereas V2O5
forms an oxide, which on average is less reduced (mainly V4+, but with some V3+ species remaining
on the surface). The XPS data also confirm a Fe/V surface atomic ratio close to unity, meaning the
absence of V segregation on the surface for the reduced spinel. A very important observation is that
the characterization performed in-situ provides a clear evidence of the reduction of the material
even after a short period of time (5 minutes). Conversely, when the same measurements are
performed ex-situ on the spent catalysts, a quick re-oxidation of the materials is observed (Figure
6).
12

Figure 6. Fe 2p (Left) and V 2p and O 1s (Right) photoemission lines of fresh, exposed to 5 mbar
of ethanol at 300°C and ex-situ spent catalysts.
Figure 7. V L23M23M45 Auger lines for the FeVO4 and V2O5 samples before and after in-situ
reduction with ethanol at 300°C.
Studies by Ueda et al. [16,17] concluded that V2O3 is selective for ethanol disproportionation
compared to V5+ and V4+ oxides. However, V3+ – which is the result of the in-situ reaction with this
13

alcohol (as in our case) – is more active than V2O3, and this might indicate that oxygen vacancies
occurred during the reduction process with the alcohol-enhanced ethanol conversion.
In the case of FeVO4, even if a completely different structure formed, the presence of V3+ made the
catalyst active for the reaction, again confirming the role of reduced vanadium as the active species.
Figure S3 shows the SEM-EDX maps of the used FeVO4 catalyst after reaction at 300°C. In this
image, the atomic distribution of V and Fe appears homogeneous, as a further indication that there
was no phase segregation of single oxides for the reduced catalyst.
Experiments were also conducted in the presence of steam; in this case, the catalyst was selective to
acetaldehyde at low temperature (>80% selectivity with a C balance of around 85%; Figure S4).
This indicates that water was able to keep the catalyst oxidized at this temperature (whereas, in the
absence of steam, it was reduced and formed ethane and acetaldehyde by ethanol
disproportionation; see Figure 1). When the temperature was raised, however, the selectivity to
acetaldehyde declined, while it increased to ethane, until the two compounds formed with similar
selectivity at 300°C, as expected from the stoichiometry for ethanol disproportionation. This means
that over 250°C the catalyst underwent reduction, and that steam could not reoxidize it; indeed,
steam reoxidation of reduced Fe oxide is thermodynamically disfavored at high temperatures [32].
A further increase in temperature led to a decline of acetaldehyde selectivity, with the formation of
acetone and CO2 as major by-products. At 400°C, the C balance was close to 80%.
Characterization by means of XRD of used catalysts after reaction with ethanol and steam is shown
in S5. The pattern is similar to the one recorded with ethanol only; FeVO4 was reduced to the spinel
phase after operation at 300°C.
Reactivity under aerobic conditions
The effect of temperature on the catalytic performance of FeVO4 under aerobic conditions was also
studied. The results obtained (Figure 8) reproduce the data reported in literature for methanol
oxidation to formaldehyde [9,12,20,21]. At 200°C acetaldehyde was produced with a selectivity of
14

97%, but with an ethanol conversion of only 45%. At higher temperatures, acetaldehyde selectivity
declined mainly to the advantage of COx. Also, the selectivity to acetic acid increased, reaching a
maximum of 7% at 350°C. There was an increasing trend of ethylene formation occurring at 300
and 350°C, but blank experiments (see Experimental) demonstrated that part of this ethylene was
due to thermal dehydration. Selectivity to C4 compounds (crotonaldehyde and butyraldehyde) was
always lower than 2%. For all experiments, the C balance was over 90%.
Figure 8. Effects of temperature on ethanol conversion and product selectivity with FeVO4 catalyst.
Reaction conditions: feed 5% ethanol, 5% O2, 90% N2; W/F 0.5 g s/ml. Symbols: ethanol
conversion (), selectivity to acetaldehyde (), ethylene (), crotonaldehyde (), butyraldehyde
(), acetone (), diethylether (), acetic acid (), CO () and CO2 (). Results were taken after
around 2 h of tos.
It was interesting to note that, after reaction at 300°C, the catalyst was still oxidized, as can be seen
in the XRD patterns shown in Figure 9. Indeed, even though O2 was fed in excess compared to the
stoichiometric requirement for ethanol oxidative dehydrogenation to acetaldehyde, due to the
formation of several by-products its conversion was almost total at 300°C. The prevailing high
oxidized state for V (and Fe) ions under the used conditions was also confirmed by the absence of
ethane as a reaction product. When the temperature was increased up to 400°C, however, the used
catalyst appeared to be reduced and the spinel phase had formed (Figure 9).
15

Figure 9. XRD patterns of FeVO4 catalyst; from top to bottom: calcined, used after reaction at
300°C, and 400°C, in ethanol + O2.
In the presence of both steam and oxygen the product distribution was very similar to that observed
with O2 only (Figure S6). After the reaction at 400°C, however, the catalyst was still oxidized
(which was not the case with O2 alone). In fact, the used sample was still orange after the reaction
and its characterization by means of Raman spectroscopy (Figure S7) confirmed that in this case the
catalyst was still oxidized.
In- Situ DRIFTS-MS Ethanol adsorption
Figure 10 shows the DRIFT spectra of the FeVO4 catalyst before and during the adsorption of
ethanol at 300°C at a different time-on-stream. Since the structure of the catalyst is changing with
time (due to the reduction to the spinel phase), it is not possible to subtract the initial spectrum of
the catalyst. It can be observed from the figure that after just 5-10 minutes, some bands appear due
to adsorbed ethanol molecularly adsorbed (approx. 1382, 2898, and 2974 cm-1) [11,18,33] and
dissociated (ethoxy form, bands at 2873, and 2925 cm-1) [8,32,34], as well as an adsorbed
acetaldehyde band at 1764 cm-1; all these bands are not present in the catalyst spectrum (see a close-
16

up of the CH region in Figure S8). After 60 minutes of adsorption, the disappeared low-frequency
bands (862, 957, and 1015 cm-1, attributed to V=O stretching of short vanadyl bonds in the distorted
VO4 units) [35] confirm the structure modification (reduction) of the catalyst, in agreement with the
reactivity tests that showed a stabilization of product distribution after more or less one hour, i.e. the
time needed to reduce the sample. The formation of the adsorbed species is accompanied by the
decreased intensity of a band assigned to a combination band of the terminal V-O(H) bond at 1347
cm-1 (that appears negative at increasing time) [36]. This means that this group is also involved in
the adsorption of ethanol and its disproportionation to acetaldehyde and ethane. This process is
accompanied by the formation of new hydrogen bonds, which is reflected in a broad band observed
at 3265 cm-1.
Figure 10. In-situ DRIFTS spectra of FeVO4 catalyst during ethanol adsorption at 300°C
(anaerobic conditions) compared with a V2O3 reference spectrum under the same conditions.
Mass spectrometer signals, recorded during this experiment, well reproduced catalytic tests (Figure
S9). In fact, both ethane and ethanol signals increased, thus suggesting a decreased ethanol
17

conversion and a reduction of the catalyst. Acetaldehyde signal decreased slightly, probably
because during catalyst reduction there was a progressively lesser contribution of the oxidative
dehydrogenation and, conversely, acetaldehyde was formed by ethanol disproportionation. As for
the mass signal trend of water, this decreased over time, and the decrease appeared to be parallel to
the increase of the ethane mass signal. This means that in the first hour water was produced from
ethanol oxidative dehydrogenation to acetaldehyde and further to CO2 (the latter has been detected
in traces in both DRIFT/MS and reactivity experiments). Then, after the stabilization time,
conversely, water formed from ethanol disproportionation.
When the same experiment was performed in the presence of air (Figure S10), results confirmed
that the structure of the catalyst was stable with time (no reduction occurred in these conditions).
Moreover, spectra did not reveal any OH band formation, thus indicating the absence of significant
new OH groups which formed, conversely, when the sample was reduced with ethanol. This may
indicate that this moiety plays a role in the alkane formation. Ueda and co-workers [16] proposed a
mechanism for ethane formation over stoichiometric V catalysts, and the adsorption of two ethanol
molecules on vicinal sites was the first step. Since we also observed the formation of ethoxy groups,
however, we suggest that ethanol might adsorb on the catalyst surface in two different ways:
dissociated or hydrogen-bonded.
Experiments were also performed with bulk V2O5 in the absence of O2 (Figure 11). The catalyst was
reduced after 30 minutes only, as evidenced by the fading of characteristic V=O overtone bands for
crystalline V2O5 (two bands at 1970 and 2016 cm-1) as well as the formation of bands between 1500
and 2000 cm-1 that correspond to oxidized ethanol species such as acetates (bands at 1440 and 1540
cm-1) and other compounds with C=O groups (bands between 1600-1800 cm-1) that afterwards
desorb as it was observed during the reactivity test (Figure 2, secondary products). This reduction
time is shorter than that required for FeVO4 under the same conditions. In fact, reactivity
experiments revealed that V2O5 formed ethane much faster than FeVO4. However, even though this
catalyst reduced faster, its final reduction degree was lower, as it was observed by XRD and XPS
18

tests. Another confirmation of this fact was obtained by comparing the spectra of spent catalysts in
the region of vanadyl overtones with those taken by feeding ethanol onto stoichiometric-reduced
vanadium oxides (green lines on Figures 10 and 11); that shows that the spectrum of the V2O5
sample after reduction presents some features similar to those of a V4+ oxide, whereas the spectrum
of the reduced FeVO4 sample was closer to the one of V3+ oxide (See a close-up in figure S11).
With this catalyst, the band of new H-bonded OH groups (approx. 3200 cm-1) developed during
reduction, as was the case for FeVO4.
Figure 11. In-situ DRIFTS spectra of V2O5 catalyst during ethanol adsorption at 300°C (anaerobic
conditions) compared with a VO2 reference spectrum under the same conditions.
Thermal-Programmed-reduction experiments
19

Hydrogen-TPR experiments were performed for FeVO4, V2O5 and Fe3O4 samples; TPR profiles are
shown in Figure 12. The results indicate that FeVO4 is more easily reduced than V2O5, a result
which is in contrast with the similar reactivity of the two catalysts under anaerobic conditions.
Moreover, when compared to magnetite, it is observed that FeVO4 does not present the reduction
peak attributed to the reduction of segregated (oxidized) hematite to magnetite and this confirms
that the hematite amount in FeVO4 sample is negligible.
Figure 12. Comparison of the H2-TPR profile of FeVO4 compared to V2O5 and Fe3O4.
We also compared the H2-TPR profiles of two synthesized V2O5 and Fe3O4 samples, in order to
check whether their reducibility properties were different from those of the corresponding
commercial samples used for reactivity experiments. TPR profiles (Figure S13) demonstrated that
the reducibility of both metal oxides was similar in the two cases.
Mechanistic model
The information collected enabled us to outline a general picture capable of explaining the
differences between bulk FeVO4 and V2O5, based on the reaction conditions used, in either the
presence or absence of oxygen as represented in Figure 13. With oxygen, V5+ species is active in
the oxidative dehydrogenation of ethanol to acetaldehyde in both cases; it is reported in literature
20

that the segregation of vanadium oxide on the surface of FeVO4 occurs during methanol
oxidehydrogenation; therefore, the V oxide-rich amorphous layer is the true active phase [9]. On the
other hand, when there is no oxygen in the gas phase (under anaerobic conditions), the alcohol is
capable of reducing V2O5 to bulk VO2 (V4+) with surface V3+ species, whereas FeVO4 is
transformed into a spinel compound where V is present only as V3+, while Fe is partially reduced to
Fe2+. The presence of V3+ in the two different structures makes these materials similarly active in
the equimolar production of acetaldehyde and ethane.
Acetic acid, COx
Acetic acid, COx
Ethanol
Ethanol
Ethanol
Ethanol
Fe1.5V1.5O4
V2O3 -  VO2
Ethane + acetaldehyde
FeVO4
Ethane + acetaldehyde, C loss
V2O5
FeVO4
(bulk)
V2O5 (surface)
Acetaldehyde + H2O
Acetaldehyde + H2O
Ethanol + O2
Ethanol + O2
Figure 13. Similarities and differences of the reaction pathways and final products for the FeVO4
and V2O5 reaction with ethanol depending on the environment.
Concerning the reaction mechanism for the less studied anaerobic reaction, it has been observed
that there are two main steps: in the first one the alcohol reduces the catalyst producing adsorbed
oxidized species (acetates, aldehydes and ketones) that afterwards desorb as COX and secondary
products. After the stabilization time (when the Vanadium is already reduced) in-situ and operando
DRIFT spectroscopy suggests that the adsorption of two ethanol molecules, one as an ethoxy and
one hydrogen bonded on the V-OH moieties generated during V reduction; facilitates the H-transfer
from one molecule onto the other one, and the disproportionation into ethane and acetaldehyde as
proposed in Figure 14. These results indicate that hydrogen elimination from the adsorbed alkoxides
on vanadia-based bulk catalysts is best described as a hydride transfer reaction, as has been stated as
well in the case of supported V2O5 [37]. Recently, this type of reaction has proved to be useful as a
21

catalytic tool for the reduction of bio-based building blocks, e.g. furfural and
hydroxymethylfurfural, with a FeVO4 catalyst [38].
Figure 14. Mechanism proposed for the transformation of ethanol into equimolar amounts of ethane
and acetaldehyde over reduced V oxides.
Conclusions
This research completes the information available on ethanol transformation into acetaldehyde on
V-based catalysts under different conditions. In the absence of oxygen, an equimolar amount of
ethane is produced together with the aldehyde; this reaction depends mainly on the oxidation degree
of the vanadium compound and not on the structure itself. In-situ XPS and DRIFT spectroscopy
showed that the active species for disproportionation of the alcohol was the reduced V3+ ion;
however, Fe in the FeVO4 catalyst was responsible for directing the reduction of metals toward the
formation of a Fe-V-O spinel phase which was homogeneous and more stable than V2O5.
In conclusion, a hydrogen transfer mechanism for the transformation of ethanol to equimolar
amounts of ethane and acetaldehyde over reduced V oxides is suggested.
References
[1]
G. Zhao, F. Yang, Z. Chen, Q. Liu, Y. Ji, Y. Zhang, Z. Niu, J. Mao, X. Bao, P. Hu, Y. Li,
22

Nat. Commun. 8 (2017) 1–8.
[2]
[3]
P. H. Rana, P. A. Parikh, New J. Chem. 41 (2017) 2636–2641.
J. Velasquez Ochoa, F. Cavani, in:, F. Cardona, C. Parmeggiani (Eds.), Transit. Met. Catal.
Aerob. Alcohol Oxid., Royal Society of Chemistry, Cambridge, 2014, pp. 203–230.
K. Y. Koltunov, V. I. Sobolev, Catal. Ind. 4 (2012) 247–252.
[4]
[5]
[6]
[7]
A. Chieregato, J. M. López Nieto, F. Cavani, Coord. Chem. Rev. 301–302 (2015) 3–23.
R. D. Weinstein, A. R. Ferens, R. J. Orange, P. Lemaire, Carbon N. Y. 49 (2011) 701–707.
T. V. Andrushkevich, V. V. Kaichev, Y. A. Chesalov, A. A. Saraev, V. I. Buktiyarov, Catal.
Today 279 (2017) 95–106.
[8]
[9]
V. V. Kaichev, Y. A. Chesalov, A. A. Saraev, A.Y. Klyushin, A. Knop-Gericke, T. V.
Andrushkevich, V. I. Bukhtiyarov, J. Catal. 338 (2016) 82–93.
K. Routray, W. Zhou, C. J. Kiely, I. E. Wachs, ACS Catal. 1 (2011) 54–66.
[10] I. E. Wachs, L. Briand, US Pat. 7,193,117 (2007).
[11] I. E. Wachs, Appl. Catal. A Gen. 391 (2011) 36–42.
[12] R. Häggblad, M. Massa, A. Andersson, J. Catal. 266 (2009) 218–227.
[13] R. Häggblad, J. B. Wagner, S. Hansen, A. Andersson, J. Catal. 258 (2008) 345–355.
[14] M. Massa, R. Häggblad, S. Hansen, A. Andersson, Appl. Catal. A 408 (2011) 63–72.
[15] A. Marberger, D. Ferri, M. Elsener, A. Sagar, C. Artner, K. Schermanz, O. Kröcher, Appl.
Catal. B 218 (2017) 731–742.
[16] Y. Nakamura, T. Murayama, W. Ueda, ChemCatChem 6 (2014) 741–744.
[17] Y. Nakamura, T. Murayama, W. Ueda, J. Mol. Catal. A 394 (2014) 137–144.
[18] J. Velasquez Ochoa, C. Trevisanut, J.-M. M. Millet, G. Busca, F. Cavani, J. Phys. Chem. C
117 (2013) 23908–23918.
[19] B. Mehlomakulu, T. T. N. Nguyen, P. Delichre, E. Van Steen, J. M. M. Millet, J. Catal. 289
(2012) 1–10.
[20] M. Massa, R. Häggblad, A. Andersson, Top. Catal. 54 (2011) 685–697.
[21] R. Häggblad, S. Hansen, L. R. Wallenberg, A. Andersson, J. Catal. 276 (2010) 24–37.
[22] V. Nivoix, B. Gillot, Mater. Chem. Phys. 63 (2000) 24–29.
[23] V. Nivoix, B. Gillot, Chem. Mater. 12 (2000) 2971–2976.
[24] V. Nivoix, F. Bernard, E. Gaffet, P. Perriat, B. Gillot, Powder Technol. 105 (1999) 155–161.
[25] V. Nivoix, B. Gillot, Solid State Ionics 111 (1998) 17–25.
[26] B Gillot, V Nivoix, Mat. Res. Bull. 34 (1999) 1735–1747.
[27] M. Nohair, D. Aymes, P. Perriat, B. Gillot, Vib. Spectrosc. 9 (1995) 181–190.
[28] G. Silversmit, D. Depla, H. Poelman, G.B. Marin, R. De Gryse, J. Electron. Spectrosc. Relat.
23

Phenom. 135 (2004) 167-175.
[29] M.C. Biesinger, B.P.Payne, A.P.Grosvenor, L.W.M. Lau, A.R. Gerson, R.St.C. Smart, Appl.
Surf. Sci. 257 (2011) 2717-2730.
[30] Demeter, M.; Neumann, M.; Reichelt, W. Surf. Sci. 454 (2000) 41−44.
[31] G.A. Sawatzky, D. Post, Phys. Rev. B 20 (1979) 1546–1555.
[32] T. Kodama, N. Gokon, Chem. Rev. 107 (2007) 4048–4077.
[33] G. Busca, L. Lietti, G. Ramis, F. Berti, Appl. Catal. B Environ. 18 (1998) 1–36.
[34] M. M. Branda, A. H. Rodríguez, P. G. Belelli, N. J. Castellani, Surf. Sci. 603 (2009) 1093–
1098.
[35] H. Tian, I. E. Wachs, L. E. Briand, J. Phys. Chem. B 109 (2005) 23491–23499.
[36] A. Garcia, N. Yan, A. Vincent, A. Singh, J. M. Hill, K. T. Chuang, J.-L. Luo, J. Mater.
Chem. A 3 (2015) 23973–23980.
[37] T. Feng, J. M. Vohs, J. Phys. Chem. B 108 (2004) 5647–5652.
[38] L. Grazia, D. Bonincontro, A. Lolli, T. Tabanelli, C. Lucarelli, S. Albonetti, F. Cavani, Green
Chem. 19 (2017) 4412–4422.
24