Evidences for Different Reaction Sites for Dehydrogenation and Dehydration of Ethanol over Vanadia Supported on Titania

Article abstract of DOI:10.1002/bkcs.11709

Vanadia supported on titania were prepared with increasing vanadia loadings up to 20 wt % to study the effect of the vanadia loading on catalytic reactivity. A competitive decomposition of ethanol into ethylene (dehydration) and acetaldehyde (dehydrogenation) over the catalyst occurs over the vanadia supported on TiO₂. Dehydrogenation is generally favored over dehydration over a wide range of the vanadia loadings up to 20 wt % due to a lower energy barrier for the dehydrogenation channel. Both dehydration and dehydrogenation rates are enhanced in proportion to the vanadia loadings up to about 2 wt %. At higher vanadia loadings (>2 wt %), however, the dehydration rate decreases while the dehydrogenation rate saturates. X-ray photoelectron spectroscopic analysis reveals that reduced V and Ti species are formed at the low vanadia contents and act as strong dissociative adsorption sites for H₂O, while extended VO_x clusters as well as three-dimensional V₂O₅ islands formed at high vanadia loadings prevent the formation of such sites. Thus, the observed difference between the two reaction channels can be explained from the structural transition of the vanadia overlayers from highly dispersed small VO_x clusters into larger polyvanadates or islands. At low vanadia loadings (<2 wt %), both the edge sites and the surface sites of the small VO_x clusters grow in proportion to the loadings as the number of the clusters increases. Thus, both reaction channels are enhanced as well. At higher vanadia loadings (>2 wt %), however, the transformation into larger islands reduces the edge sites or the boundaries between the clusters and TiO, not the surface area. This implies that the active sites for the dehydrogenation are on the surfaces of the vanadia overlayers, while those for the dehydration are on the boundaries between the VO_x and TiO₂. Our results provide an additional insight into the active sites for dehydration and dehydrogenation reactions over the titania-supported vanadia catalysts.

Full text of DOI:10.1002/bkcs.11709

Article
DOI: 10.1002/bkcs.11709
BULLETIN OF THE
Y. K. Kim and S. Ryu
KOREAN CHEMICAL SOCIETY
Evidences for Different Reaction Sites for Dehydrogenation and
Dehydration of Ethanol over Vanadia Supported on Titania
†
‡,
*
Yu Kwon Kim and Seol Ryu
†Department of Chemistry and Department of Energy Systems Research, Ajou University, Suwon 16499,
Republic of Korea
‡
Department of Chemistry, Chosun University, Gwangju 61452, Republic of Korea.
*E-mail: sryu@chosun.ac.kr
Received February 26, 2019, Accepted March 11, 2019, Published online April 29, 2019
Vanadia supported on titania were prepared with increasing vanadia loadings up to 20 wt % to study the
effect of the vanadia loading on catalytic reactivity. A competitive decomposition of ethanol into ethylene
(dehydration) and acetaldehyde (dehydrogenation) over the catalyst occurs over the vanadia supported on
TiO . Dehydrogenation is generally favored over dehydration over a wide range of the vanadia loadings
2
up to 20 wt % due to a lower energy barrier for the dehydrogenation channel. Both dehydration and dehy-
drogenation rates are enhanced in proportion to the vanadia loadings up to about 2 wt %. At higher van-
adia loadings (>2 wt %), however, the dehydration rate decreases while the dehydrogenation rate
saturates. X-ray photoelectron spectroscopic analysis reveals that reduced V and Ti species are formed at
the low vanadia contents and act as strong dissociative adsorption sites for H O, while extended VO
2
x
clusters as well as three-dimensional V O islands formed at high vanadia loadings prevent the formation
2
5
of such sites. Thus, the observed difference between the two reaction channels can be explained from the
structural transition of the vanadia overlayers from highly dispersed small VOx clusters into larger
polyvanadates or islands. At low vanadia loadings (<2 wt %), both the edge sites and the surface sites of
the small VO clusters grow in proportion to the loadings as the number of the clusters increases. Thus,
x
both reaction channels are enhanced as well. At higher vanadia loadings (>2 wt %), however, the transfor-
mation into larger islands reduces the edge sites or the boundaries between the clusters and TiO, not the
surface area. This implies that the active sites for the dehydrogenation are on the surfaces of the vanadia
overlayers, while those for the dehydration are on the boundaries between the VO and TiO . Our results
x
2
provide an additional insight into the active sites for dehydration and dehydrogenation reactions over the
titania-supported vanadia catalysts.
Keywords: Vanadia/TiO , Dehydrogenation, Dehydration, Ethanol
2
2
1–23
23,24
Introduction
species,
a polymeric vanadium oxide species,
and
25,26
crystalline vanadium oxide species.
Studies of Kilos
Titania-supported vanadia (VO /TiO ) catalysts are com-
et al., indicate that VO species from monovanadate to
x
2
x
mercial materials that are used for the removal of NO by
polyvanadate structures are formed with increasing vanadia
loadings, which would grow into crystalline V O domains
at higher vanadia loadings. Here, the interaction with the
x
1–5
selective catalytic reduction with ammonia.
Also, they
2
5
13
are known to be active for various reactions including oxi-
6
–10
dative dehydrogenation (ODH) of alkane
and alcohols
support determines not only the types of vanadium oxide
11,12
13–16
21,23,27
such as methanol
and ethanol.
Especially, increas-
species
but also the catalytic activity, which can vary
28,29
ing interests on the use of ethanol as a fuel or a feedstock
to produce hydrogen have triggered many studies on the
use of various vanadia-based catalysts for the conversion of
ethanol into acetaldehyde or ethylene, with modified sup-
by orders of magnitude.
Highly selective (~80%) ODH
of ethanol to acetaldehyde is suggested to occur on
supported vanadia catalysts with a polyvanadate mono-
13,30
layer.
Such diverse structures of supported vanadia cat-
1
7,18
19
13
ports such as TiO -SiO ,
ZrO -SiO , Al O , and
2 2 2 3
alysts also act a central role in tuning the surface
2
2
1
6,20
1,31,32
32
33
with promoters such as alkali metals.
acidity,
such as Brönsted and Lewis acid, which
One of the key factors determining the catalytic activity
is the structure of vanadium oxides in the supported van-
adia catalysts, which would vary with the vanadia loading
and the type of interaction with the supporting oxide. Vana-
dium oxides can exist as isolated vanadium oxide
are also closely correlated with the nature of active sites in
34
connection with surface V O bonds and specific influ-
35
ences of oxide supports.
In this study, we investigate into how selectivity and
reactivity of ethanol toward dehydrogenation and
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Wiley Online Library
489

Article
BULLETIN OF THE
ISSN (Print) 0253-2964 | (Online) 1229-5949
KOREAN CHEMICAL SOCIETY
dehydration change for vanadia supported on TiO (p-25)
(Porapak Q, Agilent, CA, USA, 6 ft) and a thermal conduc-
tivity detector detector, operating at a He flow of 20 sccm.
Thermally vaporized ethanol in a bottle placed in a water
2
with increasing vanadia loadings up to 20 wt %. We find
an enhancement of dehydration as well as dehydrogenation
at low vanadia contents (0.3–2 wt %), which we attribute
to an increased concentration of charged defects at/near the
surface. Reactivity toward dehydrogenation is greatly
enhanced with increasing vanadia loadings up to 2 wt %
and saturates at higher vanadia loadings. We explain the
observed variations in reactivity and selectivity from the
changes in the density and distribution of the charged
defects with increasing VO densities on TiO .
ꢀ
bath maintained at 10 C, was carried out in a He flow
(40 sccm, net ethanol flux = 55 μmol/min) with O (2 sccm,
2
82 μmol/min). The catalyst powder (0.1 g) was placed in a
Pyrex tube reactor (Kwangjin Glass, Korea) and the tem-
perature of the reactor was controlled from room tempera-
ꢀ
ture up to 400 C by a ceramic heating jacket. The reactant
gas passed through the tube reactor and the composition of
the gas mixture after the reaction was analyzed by GC. A
typical GC spectrum (Figure 1) indicates well-resolved sig-
nals for O , ethylene, H O, acetaldehyde, and ethanol. All
x
2
Experimental
2
2
the GC signals for the reactant and products were found to
increase proportional to the reactant gas flux under our
experimental condition, indicating that the reactions are not
limited by the mass transfer limit. The absolute fluxes
(μmol/min) of reaction products (ethylene and acetalde-
hyde) were determined by referencing the integrated GC
peak area of a known flux of each gas. The integrated area
of the GC peak was proportional to the concentration of
each component in the gas flow.
Preparation of VO /TiO . Preparation of VO /TiO cata-
x
2
x
2
lysts (VO -p25) was performed by impregnation of the
x
2
commercial TiO (Degussa p-25, 47 m /g) powders (1 g)
2
with a concentrated nitric acid (14 M) solution containing
0
–0.322 g of NH VO (Sigma Aldrich, 99.0%) for the van-
4 3
adia loadings of 0–20 wt %. The resulting solution was
ꢀ
heated at 80 C with a constant stirring for several hours,
ꢀ
washed and dried at 100 C for 10 h. After a final calcina-
tion under an O flow at 500 C, the resulting powders
ꢀ
2
showed colors ranging from yellow to orange depending on
the vanadia loading.
Results and Discussion
Characterization of VO /TiO . Characterization of VO /
x
2
x
Figure 2 shows XRD spectra for VO -p25 with selected
vanadia loadings of 0, 2, 5, and 20 wt %. We find that the
XRD peaks assigned to p-25 (associated with anatase and
x
TiO was performed by X-ray fluorescence (XRF), X-ray
2
diffraction (XRD), UV–Vis diffuse reflectance spectroscopy
(DRS) and X-ray photoelectron spectroscopy (XPS). The
rutile TiO phases) have no change in the peak position as
2
vanadia loading (wt %) used in the present study was deter-
mined from XRF analysis (Rigaku, TX, USA, ZSX Pri-
well as in the intensity ratio with increasing vanadia load-
ings except a gradual decrease of the overall peak intensity
due to the decrease in the mole ratio of p-25 in the compo-
sition of our catalysts. At the same time, new set of XRD
peaks are observed from 5 wt %, which are assigned to
orthorhombic crystalline V O phase. Below 5 wt %, we
mus). The surface vanadium atomic density (~1.4 V
2
atoms/nm ) of our VO -p25 for the vanadia loading of 1 wt
x
%
was also determined using the measured surface area of
2
Degussa p-25 (47 m /g, our measurement). XRD patterns
2
5
of the VO -p25 were obtained using an X-ray diffractome-
x
find no distinct vanadia-related XRD feature within our
detection limit; this implies that well-dispersed small VOx
ter (Rigaku, TX, USA, Ultima III) with Cu-Kα
(
2
λ = 0.15406 nm, 40 kV, 30 mA) irradiation by scanning
clusters are uniformly dispersed over the TiO support.
2
ꢀ ꢀ ꢀ
θ from 10 to 70 at a scanning speed of 2 /min to deter-
mine the appearance of any bulk vanadia phase as well as
the crystallinity of our catalysts. UV–Vis DRS of the VOx-
p25 catalysts was obtained on a double-beam UV spectro-
photometer (Scinco, Seoul, Korea, Neosys-2000) equipped
with a BaSO -coated integrating sphere. The absorbance in
4
the photon energies of 200–800 nm was measured for the
determination of the band-gap (E ) of our powder samples
g
by applying the Kubelka-Munk model (that is, a plot of (F
1/2
(
R)hν) vs. hν). XPS measurements were performed in a
XPS system (PHI, MN, USA, 5000 VersaProbe II) with
Mg Kα (1253.6 eV) and a charge neutralization system.
The binding energies are referenced to the C 1 s peak at
284.4 eV of the surface adventitious carbon.
Thermal Reactivity Measurements. Thermal reactivity
measurements of ethanol over the VOx-p25 catalysts were
evaluated by gas chromatography (GC) (Younglin,
Anyang-si, Korea Acme 6000MC) equipped with a column
Figure 1. A typical GC spectrum showing all chemical species
identified from our reactant gas (ethanol + O in he flow) passed
2
through VOx-p25 catalysts at 450 K.
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
490

BULLETIN OF THE
Article
Dehydrogenation and Dehydration over Vanadia
KOREAN CHEMICAL SOCIETY
no H desorption is detected up to 570 K. Thus, the follow-
2
ing stoichiometric reactions of dehydrogenation and dehy-
dration are suggested as described below;
Dehydrogenation: CH CH OH +1/2O ! CH CHO + H O.
3
2
2
3
2
Dehydration: CH CH OH ! CH CH + H O.
3
2
2
2
2
Figure 4 shows the areal rates of dehydrogenation (top)
and dehydration (bottom) with increasing vanadia loadings
at selected temperatures between 380 and 460 K. Both reac-
tions are greatly enhanced with increasing temperatures. In
addition, the enhancement is in proportion to the vanadium
loading from the very initial content. A difference between
the two channels is also found; the dehydrogenation rate sat-
urates at the vanadia loadings of 2 wt % and above, while
the dehydration rate, shows a maximum at 2 wt %, then
decreases at higher vanadia loadings. The ratio between the
two channels (dehydrogenation/dehydration) is higher than
x
Figure 2. XRD patterns of VO -p25 prepared with the vanadia
5
0 at the very low vanadia loading (~0.1 wt %), indicating
loadings of 0, 2, 5, and 20 wt %. XRD peaks for bulk crystalline
phase are noted for V-loading of 5 and 20 wt %.
the selectivity is in favor of dehydrogenation. But it gradu-
ally decreases to about 20 as the vanadia loading approaches
to 2 wt %, then increases again to about 50 at 20 wt %.
The fact that the enhancement starts from the very
initial vanadia loading up to 2 wt % indicate that the ini-
tial vanadia overlayers provide the active sites for both
reactions. These coverages (0–2 wt %) corresponds to
those without distinct vanadia phase in XRD (Figure 2).
2 5
V O
Figure 3 shows the band gap of our VO -p25 catalysts
x
determined from the UV–vis DRS measurements. We find
that the band gap of p-25 (~3.0 eV) decreases in proportion
to the vanadia loading from the very early stage and slowly
approaches to about 1.7 eV above the vanadia loading of
about 10 wt %. A gradual increase in the domain size of
surface vanadia overlayers in a form of polyvanadate would
induce a gradual shift in the absorption edge in the UV–vis
36
spectra. Compared with the band gap of bulk crystalline
V O (2.2 eV), the band gap goes below 2.0 eV at above
2
5
5
wt % of vanadium. This indicates the formation of disor-
dered structure of VO clusters on TiO with possible V
x
2
incorporation into TiO in addition to large crystalline bulk
2
V O islands (which is confirmed from XRD in Figure 2).
2
5
During the cycles of reactions up to 570 K, the colors of
our catalysts remain unchanged as determined by naked
eyes and the reactivity is maintained to be constant over
cycles, suggesting that our reaction condition does not
change the surface or bulk structure of our catalysts. In
addition, we find H O desorption as a dominant byproduct;
2
2
Figure 4. Areal rates (per s/nm ) of dehydrogenation and dehydra-
Figure 3. The band gap of VOx-p25 plotted against the vanadia
loading (wt %). Dotted line is to guide the eyes.
tion are plotted vs. the vanadia loading (wt %) at different
temperatures.
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
491

BULLETIN OF THE
Article
Dehydrogenation and Dehydration over Vanadia
KOREAN CHEMICAL SOCIETY
Thus, highly dispersed VO clusters that can be formed
in the low coverage ranges provide active sites for both
reactions.
The difference in the behavior of the reaction rate
enhancement between the two reaction channels at around
as the size of the islands grow at higher vanadia loadings
(>2 wt %).
x
The measured reaction rates (in a unit of μmol/min) of
acetaldehyde and ethylene, corresponding to dehydrogena-
tion and dehydration, respectively, are plotted in the Arrhe-
nius form of ln(R) vs. 1/T, in Figure 5. The maximum
conversion of ethanol into acetaldehyde (ethylene) mea-
sured in the present experiment is in the range of 0–10%
(0–1%) for acetaldehyde (ethylene). A quite linear relation
between ln(R) and 1/T is obtained for the reaction condition
studied, suggesting that the reaction rates is not limited by
the reactant.
The enhanced reaction rates or turnover frequencies with
increasing vanadia loadings are reflected as a shift of the
linear relation in the Arrhenius plot (Figure 5) toward lower
temperatures. For dehydrogenation, the low-temperature
shift saturates above 2 wt %, while it shifts back to higher
temperatures by about 20 K above 2 wt % for dehydration.
This is consistent with the reduced reaction rates for dehy-
dration at higher vanadia loadings as indicated in Figure 4.
We can also determine the apparent energy barrier
2
wt % suggests different active sites for both channels. It
may be that the active sites for dehydrogenation are on the
surface layers of the vanadia islands, and those for dehy-
dration are at the boundary sites of the islands. Small van-
adia islands are formed in the beginning and will grow in
size at higher loadings. The surface area of the vanadia
islands can be expected to grow and saturate at higher
islands even though the size of the islands grows later. The
boundary sites (between TiO2 and vanadia) will also
increase in the beginning when small islands are formed at
low vanadia loadings. However, they will begin to decrease
(
ΔEapp) from the slopes in the Arrhenius plot, as shown in
Figure 6. For the dehydrogenation channel, they are mea-
sured to be ~ 70 ꢁ 10 kJ/mol at all vanadia loadings up to
2
0 wt %. This value is quite close to the experimentally
determined energy barrier of methanol dehydrogenation
37
over vanadia supported on TiO2. For dehydration, how-
ever, they are measured to vary depending on the vanadia
loading. The lowest value of ~ 140 kJ/mol is obtained at
0
.1–03 wt %. ΔEapp, however, increases again with increas-
ing vanadia loadings and saturates at 160–170 kJ/mol
above 2 wt %. This value is somewhat close to 175 kJ/mol
38
for ethanol dehydration over rutile TiO (110) surface, at
2
which the oxygen vacancies with excess charges act as
39
reaction sites for ethanol dehydration. The higher ΔEapp
for p-25 (~200 kJ/mol) may be due to the fact that oxygen
vacancies need to be created on fully oxidized p-25 for eth-
anol dehydration.
Figure 5. Arrhenius plot for dehydrogenation (top) and dehydra-
tion (bottom) of ethanol over VOx-p25 with the vanadia loadings
of 0–20 wt %.
Figure 6. Apparent activation energy barriers for dehydration and
dehydrogenation of ethanol over VOx-p25.
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
492

BULLETIN OF THE
Article
Dehydrogenation and Dehydration over Vanadia
KOREAN CHEMICAL SOCIETY
Considering that the catalytic reactivity is not limited by
The main O 1 s peak at 529.8 eV is due to the lattice
the gas flow, the change in the measured ΔEapp is a direct
measure of the energy barrier of a rate-limiting step. This
leads to the conclusion that the rate-limiting step for dehy-
drogenation is insensitive to the coverage of vanadia or the
size of VOx-p25, while that for dehydration is more
structure-sensitive. This result is consistent with our specu-
lation that the active sites for dehydrogenation are on the
surface of 2D vanadia islands while those for dehydration
are on the boundaries of the islands. It is likely that the
boundary sites are more sensitive to the size of the islands.
Enhanced dehydrogenation over polyvanadates has been
oxygen species in TiO . The O 1 s binding energy of bulk
2
40,41
V O also appears at a very similar position
and little
2
5
shift in O 1 s peak position is observed with increasing
vanadia loadings. The shoulder features toward higher
binding energies (with distinct peaks at 532 and 533 eV,
observed especially at 0.1–0.5 vanadia wt %) are assigned
4
3
to surface-bound OH and H O species, respectively, as
2
2
5
53,54
48
has been reported on vanadia, TiO2,
and VO /TiO .
x 2
The shoulder feature (labeled as O , at 528.5 eV) toward
s
lower binding energies of the main O 1 s peak observed
especially at 0.3–0.5 V O wt % (indicated by arrows) may
2
5
1
3
suggested in a recent study of others, suggesting that the
active sites for dehydrogenation are on the surface of
extended vanadia overlayers.
be due to surface O atoms bound to reduced Ti sites as has
5
0
been reported on VO /TiO
as well as on reduced
x
2
43,55
2− 56
titania,
in a form of O .
Chemical bonding states of our VO -p25 catalysts are
Figure 8 shows the XPS-determined surface atomic con-
x
further investigated from XPS analysis as shown in
Figure 7. Here, Ti 2p, V 2p, and O 1 s core level spectra
are shown with increasing vanadia loadings. Ti 2p at
centration ratio of Ti and V of our VO -p25 catalysts with
x
increasing vanadia loadings. A linear increase (decrease) of
V (Ti)-ratio up to 2 wt % indicates a growth of highly dis-
persed VO clusters on TiO . A deviation from the linearity
3
/2
4
+
40–42
4
58.5 eV is from Ti species in the bulk TiO .
In
2
x
2
4
+
addition to the major Ti species, a shoulder feature
appears toward lower binding energy at the vanadia load-
ings of 0.3–0.5 wt % (indicated by arrows), which can be
occurs indicates a growth of three-dimensional large V O5
2
islands on TiO2.
Based on the above XPS analysis, it is clear that highly
dispersed (and reduced) V species on TiO2 surface are
formed at 0–2 wt %. Strong interaction between vanadia
3
+
assigned to a reduced Ti species such as Ti as in Ti inter-
42
43,44
stitials or Ti O .
At higher vanadia loadings (5–20 wt
), the main Ti 2p3/2 peak decreases in intensity due to the
formation of three-dimensional crystalline V O islands.
2
3
57,58
%
and titania
and the disordered structure of the surfaces
of p-25 nanoparticles inhibit the transformation of isolated
2
5
V 2p3/2 features starts to appear at 515–516 eV from 0.3
vanadia wt %, which shifts to 517 eV as the vanadia load-
ing increases above 2 wt %. The main V 2p peak at
VO clusters into polyvanadates or even three-dimensional
x
V O islands up to the vanadia loading of 2 wt %
2
5
2
4+
3+
(or 2.8 V/nm ). The reduced V species (V and V ) are
suggested to be stabilized on the surface of titania.
3
/2
5
+
58,59
5
17 eV is assigned to V species from the bulk V O5
The
2
40,45
46,47
phase
in V O /TiO .
The V 2p3/2 features at lower
reduced V species on titania-supported vanadia could exist
2
5
2
binding energies (516–515 eV) at the very low vanadia
in various forms such as a single or an extended VO clus-
x
4+40,48,49
12,35,60–62
49
loadings are due to reduced V species such as V
ters with terminal vanadyl group (V=O),
VO2,
3+
40,50
50,63,64
and V
.
The reduced V species can originate from
V O ,
2 3
or V species incorporated into the TiO lattice
2
4+
65
highly dispersed VO clusters or V species incorporated
substituting lattice Ti in a form of Ti-O-V. Incorpora-
x
4
+ 51
into the TiO lattice (substituted for Ti ).
tion of such V species into the TiO lattice can occur dur-
2
2
ꢀ
66
ing calculation above 450 C, and would induce charged
defects into the bulk.
Figure 7. V 2p, Ti 2p, and O 1 s core level spectra of VOx-p25
catalysts with increasing vanadia loadings.
Figure 8. Surface atomic ratio of Ti and V determined from inte-
grated areas of Ti 2p and V 2p core levels in Figure 7.
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
493

BULLETIN OF THE
Article
Dehydrogenation and Dehydration over Vanadia
KOREAN CHEMICAL SOCIETY
O 1 s spectra in Figure 7 also indicate that surface O,
OH, and H O contents are higher at low vanadia loadings
interface between VO clusters and TiO surface due to the
x 2
reduced boundary sites of larger vanadia islands.
2
(0.1–2 wt %), suggesting that the highly dispersed V spe-
cies formed on TiO can act as stable binding sites for
2
Conclusion
those oxygen species. The growth of polyvanadates as well
as large crystalline V O islands (XRD in Figure 2) at high
2
5
To summarize our results, ethanol dehydrogenation is
found to be favored over dehydration on our titania-
supported vanadia catalysts. At the very low vanadia load-
ing, we find that the surface has a high density of reduced
Ti (or V) species with highly dispersed small VO clusters.
Such defective structures provide strong binding sites not
vanadia loadings reduces the binding sites for surface OH
and H O species. Thus, the binding sites for H O and OH
2
2
are likely to be distributed to the boundaries between
polyvanadates and TiO as well as those of the isolated
2
x
VO clusters. Such VO species are reported to be in a
x
x
form of a vanadium atom with a vanadyl group (V O) and
oxygen atoms (in a distorted tetrahedral geometry) that
only for H O dissociation into OH, but also for alcohols for
2
heterolytic dissociation into alkoxy and OH for oxidative
and reductive catalytic reactions into aldehyde and alkene,
respectively. At higher vanadia loadings, the growth of
extended VO structures such as polyvanadates and large
V O islands suppress the formation of such defective
structures, which reduces the dehydration channel. The
structure insensitivity of the dehydrogenation suggests the
active sites are on the surface of vanadia. We find that the
structural transition from small VOx clusters into larger
islands play a key role in controlling the reactivity and
selectivity in catalytic reactions of ethanol.
bridge the vanadium atom to the Ti atoms in the TiO sup-
2
67,68
port.
Binding sites for H O can also act as the binding
2
sites for alcohol (RO–H).
x
Ethanol may dissociate into ethoxy (CH CH O) and
3
2
2
5
hydroxyl (OH) over terminal vanadyl group or over
1
3,69
reduced Ti (or V) sites.
desorb from the surface as H O
Hydroxyls recombinatively
1
3,70
; this process consumes
2
lattice oxygen in the oxide catalyst and needs to be
replenished from O gas to keep the stoichiometric cycles
2
of catalytic reaction. Ethoxy can be released from the sur-
face by a rate-limiting decomposition into acetaldehyde
1
3,38
(
α-H elimination) or ethylene (β-H elimination).
Combining the results of XPS (Figure 7) and reactivity
Acknowledgments. This research was supported by Basic
Science Research Program through the National Research
Foundation of Korea (NRF) funded by the Ministry of Edu-
analysis (Figures 4 and 5), we conclude that the presence
of reduced V (or Ti) species on/near the surface at the low
vanadia loadings is the origin of the enhanced reactivity
and selectivity, especially in favor of dehydration. The
reduced V (or Ti) species provide excess charges to the sur-
face in various forms of point defects such as interstitials
cation,
Science
and
Technology
(NRF-
2
016R1D1A1B03931639) and the Human Resources
Development of the Korea Institute of Energy Technology
Evaluation and Planning (KETEP) grant funded by the
Korea government Ministry of Trade, industry & Energy
6
6
and oxygen vacancies. The role of such defects in cata-
lyzing catalytic dehydrogenation and dehydration has been
(
No. 20154010200820).
7
1,72
extensively studied over rutile TiO (110) surface
as
2
5
8
References
well as on a V-deposited TiO surface. The decrease in
2
the apparent energy barrier of dehydration at low vanadia
loadings also explains a direct influence of the defects in
lowering the energy barrier for the rate-limiting step (β-H
elimination) toward dehydration. No or negligible effect of
the vanadia loading on the energy barrier of dehydrogena-
tion suggests the insensitivity of the reaction pathway for
dehydrogenation to the charged defects. Here, the role of
1
2
. C.-H. Lin, H. Bai, Appl. Catal. B Environ. 2003, 42, 279.
. I. Giakoumelou, C. Fountzoula, C. Kordulis, S. Boghosian,
J. Catal. 2006, 239, 1.
3
4
. F. Tang, B. Xu, H. Shi, J. Qiu, Y. Fan, Appl. Catal. Environ.
2
010, 94, 71.
. Q. Li, H. Yang, F. Qiu, X. Zhang, J. Hazard. Mater. 2011,
92, 915.
1
the vanadyl group (V O) in the surface VO clusters is
5. P. H. Mutin, A. F. Popa, A. Vioux, G. Delahay, B. Coq, Appl.
Catal. Environ. 2006, 69, 49.
6. M. V. Martínez-Huerta, X. Gao, H. Tian, I. E. Wachs,
J. L. G. Fierro, M. A. Bañares, Catal. Today 2006, 118, 279.
x
1
2,73–75
regarded to be important.
As the TiO is saturated with VO clusters with increasing
2
x
surface VO density, dispersed (mono) VO clusters are con-
x
x
7
8
9
. A. A. Lemonidou, L. Nalbandian, I. A. Vasalos, Catal. Today
nected by V-O-V linkages to form extended (poly) vanadates
or even three-dimensional V O islands). Suppression of
2
000, 61, 333.
(
2
5
. R. Grabowski, B. Grzybowska, A. Kozłowska, J. Słoczy ꢀn ski,
K. Wcisło, Y. Barbaux, Top. Catal. 1996, 3, 277.
. A. Adamski, Z. Sojka, K. Dyrek, M. Che, G. Wendt,
S. Albrecht, Langmuir 1999, 15, 5733.
dehydration at the higher VO density is expected to be the
x
result of decreased number of active sites associated with
reduced V(or Ti) sites, which may be located on/near the
x
edges of the isolated VO clusters as well as on bare TiO2
10. M. D. Argyle, K. Chen, A. T. Bell, E. Iglesia, J. Catal. 2002,
08, 139.
surfaces. The formation of polyvanadate overlayers reduce
2
the number of edge sites with charged point defects (e.g.,
oxygen vacancies and reduced V or Ti species) at the
11. P. Forzatti, E. Tronconi, A. S. Elmi, G. Busca, Appl. Catal.
Gen. 1997, 157, 387.
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
494

BULLETIN OF THE
Article
Dehydrogenation and Dehydration over Vanadia
KOREAN CHEMICAL SOCIETY
1
1
1
1
1
1
1
1
2. H. Y. Kim, H. M. Lee, R. G. S. Pala, H. Metiu, J. Phys.
Chem. C 2009, 113, 16083.
3. B. Kilos, A. T. Bell, E. Iglesia, J. Phys. Chem. C 2009,
45. G. Silversmit, D. Depla, H. Poelman, G. B. Marin, R. De
Gryse, J. Electron Spectros. Relat. Phenomena 2004,
135, 167.
46. M. Kobayashi, R. Kuma, S. Masaki, N. Sugishima, Appl.
Catal. Environ. 2005, 60, 173.
1
13, 2830.
4. R. Tesser, V. Maradei, M. Di Serio, E. Santacesaria, Ind.
Eng. Chem. Res. 2004, 43, 1623.
5. J. H. Kwak, J. E. Herrera, J. Z. Hu, Y. Wang, C. H. F. Peden,
Appl. Catal. Gen. 2006, 300, 109.
6. R. J. Chimentão, J. E. Herrera, J. H. Kwak, F. Medina,
Y. Wang, C. H. F. Peden, Appl. Catal. Gen. 2007, 332, 263.
7. E. Santacesaria, A. Sorrentino, R. Tesser, M. Di Serio,
A. Ruggiero, J. Mol. Catal. A Chem. 2003, 204, 617.
8. N. E. Quaranta, J. Soria, V. C. Corberán, J. L. G. Fierro,
J. Catal. 1997, 171, 1.
47. J. P. Nogier, M. Delamar, Catal. Today 1994, 20, 109.
48. G. Silversmit, H. Poelman, D. Depla, N. Barrett, G. B. Marin,
R. De Gryse, Surf. Interface Anal. 2006, 38, 1257.
49. M. Sambi, G. Sangiovanni, G. Granozzi, F. Parmigiani, Phys.
Rev. B 1997, 55, 7850.
50. Q. Guo, S. Lee, D. W. Goodman, Surf. Sci. 1999, 437, 38.
51. X. Yang, C. Cao, K. Hohn, L. Erickson, R. Maghirang,
D. Hamal, K. Klabunde, J. Catal. 2007, 252, 296.
52. V. Bondarenka, S. Grebinskij, S. Kaciulis, G. Mattogno,
S. Mickevicius, H. Tvardauskas, V. Volkov, G. Zakharova,
J. Electron Spectros. Relat. Phenomena 2001, 120, 131.
53. G. Philippin, J. Delhalle, Z. Mekhalif, Appl. Surf. Sci. 2003,
212–213, 530.
9. Y.-C. Lin, C.-H. Chang, C.-C. Chen, J.-M. Jehng, S.-
G. Shyu, Cat. Com. 2008, 9, 675.
2
2
0. M. Calatayud, C. Minot, J. Phys. Chem. C 2007, 111, 6411.
1. T. Tanaka, H. Yamashita, R. Tsuchitani, T. Funabiki,
S. Yoshida, J. Chem. Soc. Faraday Trans. 1 1988, 84, 2987.
2. H. Eckert, I. E. Wachs, J. Phys. Chem. 1989, 93, 6796.
3. Z. Luan, P. A. Meloni, R. S. Czernuszewicz, L. Kevan,
J. Phys. Chem. B. 1997, 101, 9046.
54. S. Ka ꢁc iulis, G. Mattogno, A. Napoli, E. Bemporad, F. Ferrari,
A. Montenero, G. Gnappi, J. Electron Spectros. Relat. Phe-
nomena 1998, 95, 61.
55. C.-N. Huang, J.-S. Bow, Y. Zheng, S.-Y. Chen, N. Ho,
P. Shen, Nanoscale Res. Lett. 2010, 5, 972.
2
2
2
2
4. G. G. Cortez, M. A. Bañares, J. Catal. 2002, 209, 197.
5. J. P. Balikdjian, A. Davidson, S. Launay, H. Eckert, M. Che,
J. Phys. Chem. B. 2000, 104, 8931.
56. J.-C. Dupin, D. Gonbeau, P. Vinatier, A. Levasseur, Phys.
Chem. Chem. Phys. 2000, 2, 1319.
2
6. J. Spengler, F. Anderle, E. Bosch, R. K. Grasselli, B. Pillep,
P. Behrens, O. B. Lapina, A. A. Shubin, H. J. Eberle,
H. Knözinger, J. Phys. Chem. B. 2001, 105, 10772.
7. V. Brázdová, M. V. n. Ganduglia-Pirovano, J. Sauer, J. Phys.
Chem. C 2010, 114, 4983.
57. I. E. Wachs, B. M. Weckhuysen, Appl. Catal. Gen. 1997,
157, 67.
58. S. Lee, G. W. Zajac, D. W. Goodman, Top Catal 2006,
38, 127.
59. M. A. Bañares, L. J. Alemany, M. C. Jiménez,
M. A. Larrubia, F. Delgado, M. L. Granados, A. Martinez-
Arias, J. M. Blasco, J. L. G. Fierro, J. Solid State Chem.
1996, 124, 69.
2
2
2
8. I. E. Wachs, Catal. Today 2005, 100, 79.
9. G. T. Went, L.-J. Leu, R. R. Rosin, A. T. Bell, J. Catal.
1
992, 134, 492.
3
3
0. G. C. Bond, S. F. Tahir, Appl. Catal. 1991, 71, 1.
1. H. Schneider, S. Tschudin, M. Schneider, A. Wokaun,
A. Baiker, J. Catal. 1994, 147, 5.
60. B. M. Weckhuysen, D. E. Keller, Catal. Today 2003, 78, 25.
61. G. Centi, Appl. Catal. Gen. 1996, 147, 267.
62. V. Shapovalov, H. Metiu, J. Phys. Chem. C 2007, 111,
14179.
63. N. J. Price, J. B. Reitz, R. J. Madix, E. I. Solomon,
J. Electron Spectros. Relat. Phenomena 1999, 98–99, 257.
64. E. A. Kröger, F. Allegretti, M. J. Knight, M. Polcik,
D. I. Sayago, D. P. Woodruff, V. R. Dhanak, Surf. Sci. 2006,
600, 4813.
3
3
2. N. Y. Topsoe, J. A. Dumesic, H. Topsoe, J. Catal. 1995, 151, 241.
3. L. Lietti, J. L. Alemany, P. Forzatti, G. Busca, G. Ramis,
E. Giamello, F. Bregani, Catal. Today 1996, 29, 143.
4. I. E. Wachs, Catal. Today 1996, 27, 437.
5. B. Olthof, A. Khodakov, A. T. Bell, E. Iglesia, J. Phys.
Chem. B. 2000, 104, 1516.
3
3
3
3
3
3
6. A. Khodakov, B. Olthof, A. T. Bell, E. Iglesia, J. Catal.
65. F. Trifirò, Catal. Today 1998, 41, 21.
66. J.-M. Herrmann, J. Disdier, G. Deo, I. E. Wachs, J. Chem.
Soc. Faraday Trans. 1997, 93, 1655.
67. G. T. Went, L.-j. Leu, A. T. Bell, J. Catal. 1992, 134, 479.
68. X. Gao, I. E. Wachs, J. Phys. Chem. B. 2000, 104, 1261.
69. W.-C. Wu, C.-C. Chuang, J.-L. Lin, J. Phys. Chem. B. 2000,
104, 8719.
70. M. A. Henderson, Langmuir 1996, 12, 5093.
71. C. Lun Pang, R. Lindsay, G. Thornton, Chem. Soc. Rev.
2008, 37, 2328.
72. Z. Dohnálek, I. Lyubinetsky, R. Rousseau, Prog. Surf. Sci.
2010, 85, 161.
73. R. Z. Khaliullin, A. T. Bell, J. Phys. Chem. B. 2002,
106, 7832.
74. H. Y. Kim, H. M. Lee, H. Metiu, J. Phys. Chem. C 2010,
114, 13736.
1
999, 181, 205.
7. J. L. Bronkema, D. C. Leo, A. T. Bell, J. Phys. Chem. C
007, 111, 14530.
8. Y. Kim, B. Kay, J. M. White, Z. Dohnálek, Catal. Lett. 2007,
19, 1.
9. Y. K. Kim, B. D. Kay, J. M. White, Z. Dohnalek, Surf. Sci.
008, 602, 511.
2
1
2
4
4
0. Q. Wang, R. J. Madix, Surf. Sci. 2001, 474, L213.
1. J. Pouilleau, D. Devilliers, H. Groult, P. Marcus, J. Mater.
Sci. 1997, 32, 5645.
4
4
4
2. Y. K. Kim, S. Park, K.-J. Kim, B. Kim, J. Phys. Chem. C
2
011, 115, 18618.
3. P. M. Kumar, S. Badrinarayanan, M. Sastry, Thin Solid Films
000, 358, 122.
2
4. A. F. Carley, P. R. Chalker, J. C. Riviere, M. W. Roberts,
J. Chem. Soc. Faraday Trans. 1 1987, 83, 351.
75. A. Goodrow, A. T. Bell, J. Phys. Chem. C 2008, 112, 13204.
Bull. Korean Chem. Soc. 2019, Vol. 40, 489–495
© 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
495