Direct oxidative transformation of glycerol to acrylic acid over Nb-based complex metal oxide catalysts

Article abstract of DOI:10.1016/j.cattod.2015.07.016

W-V-Nb-O complex metal oxides having a structure like that of orthorhombic Mo₃VO_x were found to be an efficient catalyst for the gas-phase direct oxidative transformation of glycerol to acrylic acid. The catalysts with various compositions were facilely synthesized by hydrothermal method. The W-Nb-O catalyst without V component, mainly promoted the glycerol dehydration, giving acrolein selectively. Since Nb-O catalyst showed poor activity for the reaction, the introduction of W into the Nb-O catalysts was found effective for improving the activity for the dehydration of glycerol to acrolein. This is mainly due to the increase of Br?nsted acidity by the combination of W and Nb. When V was incorporated into the framework of W-Nb-O, the resulting catalyst was found to prominently promote the formation of acrylic acid in the glycerol transformation in the presence of O₂, while showed no effect on the dehydration of glycerol to acrolein. It was also found that V acts as an oxidation site only when glycerol is completely reacted. The optimum elemental composition for the reaction was W_2.2V_0.4Nb_2.4O₁₄ giving ca. 46% one path yield of acrylic acid and this catalytic performance was further improved by the surface modification with phosphoric acid, achieving ca. 60% one path yield of acrylic acid.

Full text of DOI:10.1016/j.cattod.2015.07.016

Catalysis Today 259 (2015) 205–212
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Direct oxidative transformation of glycerol to acrylic acid over
Nb-based complex metal oxide catalysts
Kaori Omataa, Keeko Matsumotob, Toru Murayamab, Wataru Uedac,∗
a Department of Materials Science and Engineering, Suzuka National College of Technology, Shiroko-cho, Suzuka 510-0294, Japan
b Catalysis Research Center, Hokkaido University, N-21, W-10, Sapporo 001-0021, Japan
c Department of Material and Life Chemistry, Kanagawa University, 3-27-1, Rokkakubashi, Kanagawa-ku, Yokohama 221-8686, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 3 March 2015
Received in revised form 4 July 2015
Accepted 13 July 2015
Available online 31 August 2015
W–V–Nb–O complex metal oxides having a structure like that of orthorhombic Mo3VOx were found to
be an efficient catalyst for the gas-phase direct oxidative transformation of glycerol to acrylic acid. The
catalysts with various compositions were facilely synthesized by hydrothermal method. The W–Nb–O
catalyst without V component, mainly promoted the glycerol dehydration, giving acrolein selectively.
Since Nb–O catalyst showed poor activity for the reaction, the introduction of W into the Nb–O catalysts
was found effective for improving the activity for the dehydration of glycerol to acrolein. This is mainly
due to the increase of Brönsted acidity by the combination of W and Nb. When V was incorporated into the
framework of W–Nb–O, the resulting catalyst was found to prominently promote the formation of acrylic
acid in the glycerol transformation in the presence of O2, while showed no effect on the dehydration of
glycerol to acrolein. It was also found that V acts as an oxidation site only when glycerol is completely
reacted. The optimum elemental composition for the reaction was W2.2V0.4Nb2.4O14 giving ca. 46% one
path yield of acrylic acid and this catalytic performance was further improved by the surface modification
with phosphoric acid, achieving ca. 60% one path yield of acrylic acid.
Keywords:
Glycerol
Acrylic acid
Hydrothermal synthesis
Nb
W
V
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
catalysts by carbon deposition. Some zeolites or oxide catalysts
give acrolein in yields of ∼80% with byproducts such as hydrox-
Biomass utilization aiming at CO2 reduction and decreasing
dependency of oil has been expanding. Under this situation, the
amount of biodiesel production has been increasing year after year
[1]. Glycerol is a main byproduct in biodiesel production by trans-
esterification of plant oils or animal fat with methanol, and has
been produced heavily at a relatively low price. Therefore, the
transformation of glycerol into other desirable chemicals by var-
ious reactions such as oxidation [2], hydrogenolysis [3], reforming
[4], esterification [5,6], etherification [7] and others [8,9] has been
attempted by many researchers. The dehydration of glycerol to
acrolein, which is an important intermediate for chemical and agri-
cultural industries, is one of the most valuable reactions.
yacetone, propanal and others [10,11,14–16]. On the other hand,
heteropoly acid catalysts show higher yield (∼98%) of acrolein [13].
However, regeneration of heteropoly acid catalysts by combustion
of coke is difficult because of their low thermal stability.
Acrolein is oxidized to acrylic acid, and then used in the
manufacture of plastics, coatings, adhesives and so on. Direct trans-
formation of glycerol to acrylic acid brings about simplification
of process and reduction of energy. Standard enthalpy change of
the glycerol dehydration to acrolein and the acrolein oxidation
to acrylic acid are 18.0 kJ mol−1 and −298.3 kJ mol−1, respectively
[17]. The direct transformation of glycerol may be energy-efficient
process if the energy exhausted in the latter reaction can be used
for the former reaction.
As catalysts for the direct transformation of glycerol to acrylic
acid, Mo–V–Te–Nb–O having an orthorhombic structure [18],
W–V–O [18,19] and W–V–Nb–O [20,21] having a hexagonal tung-
sten bronze structure have been reported. The catalyst which gave
the highest acrylic acid yield was a hexagonal W–V–Nb–O among
the proposed catalysts. Chieregato et al. [20] has reported that the
incorporation of Nb in W–V–O structure generates strong acid sites
Various solid acid catalysts such as zeolites [10,11], heteropoly
acids [12,13] and mixed metal oxides [14–16] have been applied
to this reaction in the gas phase. The main problems within
this process are the formation of byproducts and deactivation of
∗
Corresponding author.
E-mail address: uedaw@kanagawa-u.ac.jp (W. Ueda).
http://dx.doi.org/10.1016/j.cattod.2015.07.016
0920-5861/© 2015 Elsevier B.V. All rights reserved.

206
K. Omata et al. / Catalysis Today 259 (2015) 205–212
and improves the acrylic acid yield. According to their latest report,
the yield of acrylic acid reached to 50.5% by optimization of the
reaction conditions [21].
2.2. Catalyst characterization
Surface area of the catalysts was estimated by BET method
where nitrogen physisorption amount was measured at 77 K with
a BEL max 00094 (BEL Japan Inc.). Prior to the measurement, the
samples were evacuated at 473 K for 2 h. Powder X-ray diffraction
(XRD) pattern of the catalysts was recorded on RINT2200 (Rigaku)
with Cu K␣ radiation (tube voltage: 40 kV, tube current: 20 mA).
Scanning transmission electron microscope (STEM) images were
obtained with a HD-20000 (HITACHI). The composition of catalysts
was determined by inductively coupled plasma atomic emission
spectroscopy (ICP-AES). The acid amount of catalysts was mea-
sured by NH3-TPD with a TPD apparatus (BEL Japan Inc.). Prior to
the measurement, the samples were pretreated under He flow at
673 K for 2 h. NH3 was adsorbed on the catalysts at 473 K. Acidity
of catalysts was measured by FT-IR spectroscopy (PARAGON 1000,
Perkin Elmer) of adsorbed pyridine with a furnace cell with CaF2
windows, containing a self-supporting disk of sample. The sam-
ples were pretreated in a vacuum at 623 K. Pyridine was adsorbed
onto the samples at 373 K and after evacuation at 523 K for 1 h, the
adsorption spectrum was recorded. The spectrum of adsorbed pyri-
dine on sample in the presence of water vapor (4.6 Torr) was also
recorded.
Nevertheless, the values have not come up to that in the possible
process with a separated glycerol dehydration and acrolein oxi-
dation. (For example, Cs2.5H0.5PW12O40 catalyst gives an acrolein
yield of 98% in the glycerol dehydration [13]. On the other hand,
Mo–V based catalyst gives acrylic acid yield of 98% in the acrolein
oxidation [22].) The direct transformation of glycerol to acrylic
acid is desirable but still a challenging process because not only
improvement of selectivity for each reaction by catalyst design but
also tune of optimum conditions for each reaction is required to
achieve higher acrylic acid yield.
We have recently reported the orthorhombic-Mo3VOx and
trigonal-Mo3VOx catalysts which exhibit a high yield of acrylic
acid of more than 90% in the acrolein oxidation [23]. Because the
orthorhombic and trigonal phases show similar catalytic prop-
erties, we have advocated their local structure, which is the
layered structure comprising 6- and 7-rings of metal-oxide octa-
hedral and pentagonal columns, is important for high oxidation
catalysis. We have also reported that W–Nb–O catalysts synthe-
sized by hydrothermal method have a layered structure similar to
orthorhombic Mo–V–O [24] and gave acrolein in high yield in gas-
phase glycerol dehydration in the presence of water and oxygen
[25].
If a complex oxide has both enough acidity and acrolein-
oxidizing capability comparable to orthorhombic Mo–V–O, it is
expected to be an efficient catalyst for the direct transformation of
glycerol to acrylic acid. In the present work, we applied the complex
metal oxides composed of W, Nb and V having the orthorhombic-
like structure as a catalyst for the first time to the direct oxidative
glycerol transformation to acrylic acid, and investigated effects
of constituent materials on the selectivity for the glycerol trans-
formation. And then we further tried to improve the catalytic
performance by surface modification with phosphoric acid.
2.3. Catalytic testing
Transformation of glycerol was carried out in a vertical fixed-bed
reactor. A mixture of 0.02–0.80 g of catalyst powder and sea sand
(catalyst + sea sand = 2 ml) was set in a Pyrex glass tube with 12 mm
internal diameter and pretreated at 573 K in flowing 20% O2/N2 for
0.5 h before the reaction. A glycerol-water 1:5 (mol/mol) solution
was fed into the top of reactor with a micro feeder. Pure N2 or O2/N2
was introduced with mass flow controllers. The molar composition
of reaction gas was glycerol/N2/H2O = 5/70/25 (absence of O2) or
glycerol/O2/N2/H2O = 5/14/56/25 (presence of O2). Reaction prod-
ucts and unconverted glycerol in both liquid and gas phases were
collected hourly in an ice trap (273 K) and a Tedlar bag connected
at the end of the trap. Products in the liquid phase were analyzed
with FID-GC (GL science GC353, 30 m 0.32 mm TC-WAX capillary
column). Products in the gas phase were analyzed with FID-GC
(Shimadzu GC14B, Gaskuropak54 column) and TCD-GC (Shimadzu
GC8A, Porapak QS and MS-13X column). Glycerol conversion and
product selectivities and yields were calculated by Eqs. (1)–(3).
2. Experimental
2.1. Catalyst preparation
2.1.1. W–V–Nb–O by hydrothermal method
The complex metal oxide catalysts of W, V and Nb were prepared
by hydrothermal synthesis method. (NH4)6[H2W12O40]·nH2O
(Nippon Inorganic Color & Chemical Co., Ltd., >90.8% as WO3),
VOSO4·nH2O (Mitsuwa Chemical Co., Ltd., 61.9%) and Nb2O5·nH2O
(Soekawa Chemical Co., Ltd, 70.8% as Nb2O5) were used as raw
materials. The metal salts were added in 45 ml of ion-exchanged
water under stirring. This mixed suspension was put in a stainless
steel autoclave with a Teflon liner and heated at 448 K for 72 h.
The formed solid was filtered, washed with ion-exchanged water,
dried at 353 K and then calcined at 673 K for 4 h in air. The W/Nb
ratio in the preparative materials was set to be 1.35 (W 2.7 mmol,
Nb 2.0 mmol in 45 ml water) according to our previous articles
[25]. W–V–Nb–O, W–V–O, and V–Nb–O were prepared from solu-
tion consisting metals with W:V:Nb = 1.35:0.1–0.6:1, W:V = 1.35:1,
V:Nb = 0.4:1, respectively. Nb–O was also synthesized by the same
procedure as described above.
Moles of glycerol reacted
Moles of glycerol fed
Glycerol conversion (%) =
× 100
(1)
Product selectivity (%)
Moles of carbon in a product defined
Mole of carbon in glycerol reacted
=
× 100
(2)
Moles of carbon in a product defined
Moles of carbon in glycerol fed
Product yield (%) =
× 100(3)
2.1.2. Addition of P to W–Nb–O by impregnation method
Phosphoric acid-added catalysts were prepared by impreg-
nation of uncalcined W–Nb–O (W:Nb = 1.35:1) or W–V–Nb–O
(W:V:Nb = 1.35:0.3:1) with an aqueous solution of phosphoric acid,
followed by calcination at 673 K in air. The content of P was set to
be 2.5 wt% of the supports after optimization of P content.
Oxidation of acrolein was carried out in the same reac-
tor as the transformation of glycerol. Reaction conditions were:
catalyst weight, 0.2 g; reaction temperature, 458–658 K; total
flow rate, 77 ml min−1
acrolein/O2/N2/H2O = 1/15/58/26.
; molar composition of reaction gas,

K. Omata et al. / Catalysis Today 259 (2015) 205–212
207
Mo3VOx, as can be seen in the XRD patterns in Fig. 2. However,
these diffraction peaks observed on W–Nb–O were very broad,
suggesting that the crystal growth into a and b directions is very
low compared to the c-direction. To confirm this crystal state,
we collected STEM image of W–Nb–O sample and the result is
shown in Fig. 2(b). Concomitantly, W–Nb–O was observed to be
much small crystals with the nano-sizes of 10–30 nm in diame-
ter (a and b direction) and of 0.1–0.3 ␮m in length (c-direction),
while Cs0.5[Nb2.5W2.5O14] with rod-like morphology was well-
grown crystals with the sizes of 0.5–1 ␮m in diameter and of more
than 5 ␮m in length (Fig. 2(a)). By taking it into account that the lat-
tice parameters of Cs0.5[Nb2.5W2.5O14], of which crystal structure is
assumed here in the W–Nb–O case, are quite large (2.6 nm × 2.1 nm
in a-b plane), the observed broad peaks in XRD of the nano-sized
W–Nb–O might be reasonable. These facts may suggest the for-
mation of the orthorhombic structure in the W–Nb–O sample.
In order to add more evidence to this suggestion, we conducted
Rietveld analysis on the basis of the orthorhombic crystal struc-
ture of Mo3VOx, the lattice parameters of Cs0.5[Nb2.5W2.5O14], and
the crystal size of 10 nm × 10 nm × 300 nm as observed by STEM.
As shown in Fig. 2(b), the observed and the simulated diffractions
were agreed each other well (Rwp = 7.03%), while Rietveld refine-
ment with using structure models and structural parameters of
hexagonal and other layered materials did not converge. This result
evidently supports the formation W–Nb–O oxide with the structure
similar to that of Mo3VOx.
Fig. 1. XRD patterns (left: wide angle range, right: narrow range around 22◦) of (a)
W2.8Nb2.2O14, (b) W2.4V0.1Nb2.5O14, (c) W2.2V0.4Nb2.4O14, (d) W2.2V0.5Nb2.3O14, (e)
W2.2V0.9Nb1.9O14, (f) W2.8V2.2O14, (g) Nb5.0O14, (h) V1.0Nb4.0O14
.
3. Results and discussion
3.1. Structure of W–Nb–O and W–V–Nb–O
The vanadium-containing W–Nb–O, W–V–O, V–Nb–O, and
Nb–O also showed the similar XRD pattern of the W–Nb–O sam-
ple (Fig. 1(left)). In the case of the vanadium-containing W–Nb–O,
the diffraction peak ascribed to (0 0 1) plane shifted to higher angle
depending on the V content (Fig. 1(right)). This clearly indicates
that V was incorporated into the framework of W–Nb–O. XRD pat-
terns similar to that of W–Nb–O were also observed for Nb–O,
W–V–O and V–Nb–O, indicating that the same structure can be
constructed in any combinations of three elements, W, Nb and V.
Weak unknown peaks in W–V–O (Fig. 1(left) may be attributed to
a type of vanadium oxide which forms from hydrated vanadium
X-ray diffraction patterns of all the samples synthesized
hydrothermally are illustrated in Fig. 1 (left). The diffraction peaks
in all cases were observed at 22.7◦ and 46.2◦ attributed to (0 0 1)
and (0 0 2) planes of a layered structure, respectively. Since it can
be considered that these peaks are due to the linear arrange-
ment of corner-shared octahedra in c-direction, the appearance of
these clear diffraction peaks indicates the crystal formation of c-
direction. The other diffractions around 8◦, 13◦, 27◦, 35◦ were also
observed for all the samples. It should be noted that the angles
of these peaks were roughly coincident with the angle region
of the XRD peaks of well-crystallized Cs0.5[Nb2.5W2.5O14] in the
orthorhombic structure [25] which is basically the same as that of
+
oxide containing NH4 ion [26] in the course of the calcination.
Fig. 2. STEM images of (a) Cs0.5[Nb2.5W2.5O14] and (b) W2.8Nb2.2O14 and their XRD patterns with Rietveld analysis.

208
K. Omata et al. / Catalysis Today 259 (2015) 205–212
Table 1
Catalyst surface properties and their performance for the glycerol transformation in the absence of O2.
a
Catalyst
SABET (m2 g−1
)
Acid amount
Tb (K)
GLRc conv.d (%)
Yielde (%)
(␮mol g−1
)
ACRL
AA
ACAL
PRAL
HACT
ACA
COx
Others
Nb5.0O14
89
106
100
8
40
133
136
5
566
563
562
563
563
56.1
56.1
90.0
9.5
10.6
13.5
50.7
3.0
<0.1
<0.1
0.2
<0.1
<0.1
0.3
1.0
1.1
0.2
0.8
0.1
0.7
1.3
0.1
1.4
1.9
3.7
6.4
<0.1
7.8
0.1
0.2
0.9
<0.1
<0.1
0.4
0.3
0.2
0.2
0.2
42.6
36.5
29.2
6.0
V
1.0Nb4.0O14
W
W
W
2.8Nb2.2O14
2.8V2.2O14
2.2V0.5Nb2
118
148
91.7
53.8
27.7
a
BET surface area.
Reaction temperature.
GLR = glycerol.
b
c
d
e
Reaction conditions: catalyst weight, 0.2 g; flow rate, 80 ml min−1; time on stream, 1–2 h; composition of reaction gas, GLR/N2/H2O = 5/70/25 (mol%)
ACRL = acrolein, AA = acrylic acid, ACAL = acetaldehyde, PRAL = propanal, HACT = hydroxyacetone, ACA = acetic acid.
On the basis of the above structural discussion, M5O14(M = W, V,
that Brönsted acid sites are advantageous over Lewis acid sites in
the synthesis of acrolein from glycerol [1].
Nb) is suitable as the general chemical formula for the present
materials, although net charge should be balanced by either
additional proton or oxygen deficiency. We then conducted the
ICP-AES analysis of all the calcined samples. Based on the analytical
data, the samples are now denoted with the chemical formula
Effect of water addition on the acid property of W2.8Nb2.2O14
was investigated because the glycerol transformation is carried
out in the presence of additive water and formed water. Fig. 4(B)
illustrates the IR spectra of adsorbed pyridine on Nb5.0O14 and
W2.8Nb2.2O14 in the presence of water. The addition of water
decreased the intensity of the IR-band ascribed to the adsorption
of pyridine on Lewis acid sites and on the other hand increased
the intensity of IR-band ascribed to that on Brönsted acid sites in
both Nb5.0O14 and W2.8Nb2.2O14. The result indicates that Lewis
acid site is hydrated and changes into Brönsted acid site in the
presence of water. The ratio of Lewis acid that changed into Brön-
sted acid was different depending on the presence or absence W.
About 29% and 57% of Lewis acid sites changed into Brönsted acid
sites in Nb5.0O14 and W2.8Nb2.2O14, respectively. As a result, such
high Brönsted acidity of W2.8Nb2.2O14 should be beneficial for the
catalytic reaction, since it has been widely considered that Brön-
sted acid sites are responsible for the glycerol transformation to
acrolein.
In contrast to the prominent addition effect of W, the intro-
duction of V into the framework of W2.8Nb2.2O14 or Nb5.0O14 gave
almost no effect on both the conversion of glycerol and the acrolein
selectivity as can be seen in Table 1 and Fig. 3, although the acid
amount of Nb5.0O14 largely changed. It seems that the surface acid
property derived from the existence of V in the lattice and surface V
itself have no ability to contribute to the glycerol dehydration steps.
This speculation will be evident from the results in the glycerol
transformation in the presence of O2 as shown in the next.
as follows; W2.8Nb2.2O14
2.2V0.5Nb2.3O14 W2.2V0.9Nb1.9O14
1.0Nb4.0O14
,
W2.4V0.1Nb2.5O14
,
W2.2V0.4Nb2.4O14
Nb5.0O14
,
,
W
V
,
,
W2.8V2.2O14,
.
3.2. Glycerol transformation in the absence of O2
The catalytic performances in the glycerol transformation in
the absence of O2 were first examined to compare the catalytic
activity of the synthesized materials for the dehydration of glyc-
erol to acrolein. The catalytic results with specific surface area and
the number of acid sites per gram (acid amount determined by
NH3-TPD) are listed in Table 1.
The role of Nb is first considered. The surface area and the
acid amount of the non-Nb-containing W–V–O catalyst were lower
than these of the Nb-containing catalysts. The acrolein yield over
W
2.2V0.5Nb2.3O14 and W2.8V2.2O14 were 53.8 and 3.0%, respectively,
indicating the introduction of Nb augmented the acrolein yield. The
improvement of acrolein yield is due to the increase in the sur-
face area and the corresponding acid amount by the introduction
of Nb.
The acrolein yield over W2.8Nb2.2O14 was 50.7%, while Nb5.0O14
gave acrolein yield of 10.6%. Apparently the incorporation of W to
Nb5.0O14 improved the activity for the dehydration of glycerol. It
is evident that the activity enhancement is mainly caused by the
increase of surface acid amount as observed. However, according
to Fig. 3 showing the plot of the acrolein yield over Nb5.0O14 and
W2.8Nb2.2O14 against the glycerol conversion, the acrolein yield
over W2.8Nb2.2O14 was higher than that over Nb5.0O14 when com-
pared at an iso-conversion of glycerol. It can be concluded that this
yield improvement was not due to the difference of the glycerol
conversion but to the catalytic property changes by the combina-
tion of W and Nb in the structure. This fact also indicates that the
introduction of W was improved not only activity but also acrolein
selectivity.
The effect of the introduction of W to Nb5.0O14 on acid prop-
erty was further investigated with FT-IR of adsorbed pyridine on
Nb5.0O14 and W2.8Nb2.2O14. Fig. 4(A) illustrates the IR spectra of
pyridine on Nb5.0O14 and W2.8Nb2.2O14. Both the catalysts gave the
IR-bands ascribed to the adsorption of pyridine on Lewis and Brön-
sted acid sites at about 1450 and 1540 cm−1, respectively. The ratio
of Brönsted to Lewis acidity in W2.8Nb2.2O14 was larger than that in
Nb5.0O14. Therefore, it seems that the introduction of W increases
the Brönsted acidity and hence enhances acrolein selectivity in the
glycerol transformation. It has been also indicated in many papers
Fig. 3. Acrolein yield against glycerol conversion in the glycerol transforamtion.
Symbols: W2.8Nb2.2O14 ᭹; W2.2V0.5Nb2.3O14 ꢀ; Nb5.0O14 ꢀ; V1.0Nb4.0O14 ꢁ. Compo-
sition of reaction gas, GLR/N2/H2O = 5/70/25 (mol%).

K. Omata et al. / Catalysis Today 259 (2015) 205–212
209
Fig. 4. IR spectra of adsorbed pyridine on (a) Nb5.0O14 and (b) W2.8Nb2.2O14 in the absence (A) and presence (B) of water.
3.3. Glycerol transformation in the presence of O2
catalytic performance. Fig. 5 illustrates the yield of acrolein,
acrylic acid and COx in the glycerol transformation in the pres-
ence of O2 under the various contact times over the W–V–Nb–O
catalysts. The W2.8Nb2.2O14 catalyst gave maximum acrolein
The catalytic results of the glycerol transformation in the pres-
ence of O2 are listed in Table 2. The W2.8Nb2.2O14 catalyst gave
acrolein mainly in better yield compared to the reaction in the
absence of O2. Obviously oxygen assists the dehydration and
maintains the catalytic activity. When V was incorporated into
the W2.8Nb2.2O14 catalyst (W2.2V0.5Nb2.3O14), acrylic acid yield
greatly increased and acrolein formation was suppressed. This
result clearly demonstrates that the incorporated V can promote
the acrolein oxidation process to acrylic acid. Since the addition
of V into Nb5.0O14 (V1.0Nb4.0O14) simply promoted over oxida-
tion to COx, W as a component seems essential for the formation
of acrylic acid. In fact, W2.8V2.2O14 showed appreciable activity
for the formation of acrylic acid as shown in Table 2. How-
ever, the yield values over this catalyst were inferior to that
achieved over W2.2V0.5Nb2.3O14, suggesting that co-existence of
three components in the framework of orthorhombic type struc-
ture is important.
yield at W/F = 2.5 × 10−3 g min ml−1
. At contact time above
W/F = 2.5 × 10−3 g min ml−1, the yield of acrolein decreased and
the yield of COx increased with increasing the contact time. Acrylic
acid was not observed at all over the W2.8Nb2.2O14 catalyst. In
a separate experiment of acrolein oxidation over this catalyst,
it was found that acrolein was directly oxidized to COx without
formation of acrylic acid. The results indicate that acrolein was
directly oxidized to COx also under the glycerol transformation.
On the other hand, acrylic acid was formed over the V-containing
catalysts. The catalysts containing larger amount of V tend to
give maximum yields of acrylic acid at shorter contact time.
Over the W2.4V0.1Nb2.5O14 catalyst, the formations of acrolein,
acrylic acid and COx were formed and these product yield changes
indicate a consecutive reaction profile. In addition to this, the
direct conversion of acrolein to COx should be contributed to the
profile, since the formation of COx from acrolein on W–Nb–O sites
The presence of V is also important for the catalytic activ-
ity durability. We tested the time courses of acrolein and acrylic
acid yield in the glycerol transformation over W2.8Nb2.2O14 and
W2.2V0.5Nb2.3O14 in the presence of O2. The yield of acrylic acid
over W2.2V0.5Nb2.3O14 was kept constant during 5 h of the reaction
although acrolein yield over W2.8Nb2.2O14 decreased from 74.5%
to 57.7%. The amounts of carbon deposition after the reaction over
W2.8Nb2.2O14 and W2.2V0.5Nb2.3O14 catalysts measured by TG were
about 193.6 and 30.3 mg gcat−1, respectively. It is clear that V sites
decrease the carbon deposition that may cause deactivation of the
catalysts in the dehydration of glycerol. The materials causing cat-
alytic deactivation may be oxidatively decomposed at the V sites
under the reaction conditions.
The effect of constituent materials is summarized as follows: Nb
increased the surface area and the acid amount of the catalysts and
improved the activity for the dehydration of glycerol to acrolein.
W increased the Brönsted acidity of the catalysts and improved
the activity for the dehydration of glycerol to acrolein. V was the
essential element for the oxidation of acrolein to acrylic acid. The
introduction of V into the framework of Nb5.0O14 or W2.8Nb2.2O14
little affects the acrolein formation from glycerol. The co-existence
of three components in the framework of orthorhombic type struc-
ture is most indispensable for achieving high acrylic acid in the
direct oxidative transformation of glycerol.
3.4. Effect of V content on the yield of acrylic acid
Fig. 5. Effect of contact time on yield of (A) acrolein, (B) acrylic acid and (C) COx in the
The W–V–Nb–O catalysts with different V contents were
tested in order to further elucidate the role of added V on the
glycerol transformation in the presence of O2 over W2.8Nb2.2
O14 ᭹; W2.4V0.1Nb2.5O14
ꢀ; W2.2V0.4Nb2.4O14ꢁ; W2.2V0.5Nb2.3O14 ꢀ; W2.2V0.9Nb1.9O14 ×.

210
K. Omata et al. / Catalysis Today 259 (2015) 205–212
Table 2
Catalytic performance in the glycerol transformation in the presence of O2.
Catalyst
Ta (K)
GLR conv.b(%)
O2 conv. (%)
Yield (%)
ACRL
AA
ACAL
PRAL
HACT
ACA
COx
Others
W2.8Nb2.2O14
V1.0Nb4.0O14
565
628
628
596
97.2
100
100
4.9
80.0
55.0
51.9
72.3
4.4
3.3
<0.1
1.4
18.2
33.7
1.8
0.8
0.8
0.5
0.7
<0.1
<0.1
<0.1
1.6
<0.1
<0.1
<0.1
0.4
3.5
14.0
15.2
2.9
83.7
53.2
42.7
17.5
6.0
10.5
7.0
W2.8
V2.2O14
W
2.2V0.5Nb2.3O14
100
0.9
a
Reaction temperature.
b
Reaction conditions: catalyst weight, 0.2 g; flow rate, 80 ml min−1; time on stream, 1–2 h; composition of reaction gas, GLR/O2/N2/H2O = 5/14/56/25 (mol%).
Table 3
Glycerol transformation over phosphoric acid-added W–Nb–O and W–V–Nb–O catalysts.
Catalyst
W/F/gcat (min ml−1
)
GLR conv. a (%)
O2 conv. (%)
Yield (%)
ACRL
AA
0.1
0.3
46.2
36.6
44.8
59.2
ACAL
PRAL
HACT
ACA
COx
Others
W
2.8Nb2.2O14
2.5 × 10−3
2.5 × 10−3
6.7 × 10−3
1.0 × 10−2
6.7 × 10−3
1.0 × 10−2
98.9
100
100
100
100
100
6.5
5.6
48.6
53.4
33.0
45.3
74.5
81.8
3.5
0.5
15.3
0.5
2.9
2.1
1.4
0.4
2.6
0.3
1.0
0.6
<0.1
<0.1
0.1
0.4
0.5
<0.1
<0.1
<0.1
<0.1
0.7
0.3
12.7
14.2
7.8
3.8
2.7
33.8
44.8
21.5
22.3
15.5
11.7
2.4
4.0
7.9
2.5 wt%PO4/W2.8Nb2.2O14
W
2.2V0.4Nb2.4O14
2.5wt%PO4/W2.2V0.4Nb2.4O14
<0.1
8.2
9.5
a
Reaction conditions: set temperature of the furnace, 558 K; composition of reactant gas, glycerol/O2/N2/H2O = 5/14/56/25 (mol%).
is unavoidable. The catalyst which exhibited the maximum yield
of acrylic acid was W2.2V0.4Nb2.4O14 and gave an acrylic acid yield
of 46.2% at W/F = 6.7 × 10−3 g min ml−1. The further addition of V
resulted in large COx formation.
because the reaction rate of the glycerol dehydration to acrolein
(pathway 1) is much higher than that of the glycerol oxidation
(pathway 5). Then, to yield acrylic acid by the oxidation of acrolein
(pathway 3), the high reaction temperature or the long contact
time is required. Under the high reaction temperature or long
contact time, however, a high yield of acrylic acid cannot be
expected due to the occurrence of the direct formation of COx
from acrolein (pathway 2) over much available W–Nb–O site.
On the other hand, although W2.2V0.9Nb1.9O14 gives acrylic acid
in a high selectivity in the acrolein oxidation, the oxidation of
glycerol to COx (pathway 4) inevitably proceeds over excess V
sites. W2.2V0.4Nb2.4O14 exhibited the highest yield of acrylic acid
as a whole because undesirable oxidation pathways (pathway 2, 4
and 5) are retarded by well-balanced cooperation of the W–Nb–O
network and the incorporated V.
In Fig. 6, the conversion of oxygen and the yields of acrylic
acid, acrolein and COx are plotted against the glycerol conversion.
Over W2.2V0.9Nb1.9O14, the conversion of oxygen more than 20%
was observed even under the condition that glycerol conversion
was less than 100%. The W2.2V0.5Nb2.3O14 catalyst also showed a
similar catalytic result. This oxygen consumption should be of the
oxidation of glycerol because acrylic acid was formed after glyc-
erol conversion reached 100%. In W2.8Nb2.2O14, W2.4V0.1Nb2.5O14
and W2.2V0.4Nb2.4O14, on the other hand, the conversion of oxy-
gen was much less than 5% even under the condition that glycerol
conversion was almost 100%. That is to say, the oxidation of
glycerol hardly proceeds over these catalysts. This is the rea-
son why W2.2V0.4Nb2.4O14 gave higher acrylic acid yield than
3.6. Oxidative glycerol transformation over phosphoric
acid-added W–V–Nb–O
W2.2V0.5Nb2.3O14 or W2.2V0.9Nb1.9O14 where excess V may have
a strong activity for the oxidation of glycerol as well as the desired
products to unidentified products.
Based on the above results and discussion on the role of each
catalytic element, one may think that a control of oxidation activ-
ity of V site is still necessary for improving acrylic acid yield in the
oxidative glycerol transformation. We, therefore, tried to introduce
phosphate anions on the surface of the W–V–Nb–O catalyst [27]
because we can expect an interaction between V site and phos-
phate anions to form vanadium phosphate like species and also
another interaction between W site and phosphate anions to form
heteropoly-type unit in the framework.
More interesting pints from Fig. 6 is that the formation of acrylic
acid and COx began just after glycerol was converted completely.
This result strongly suggests that glycerol having high boiling point
(563 K) can exclusively adsorb on the surface V oxidative sites and
inhibit the access of acrolein. This is why W–Nb–O site can work for
dehydration independently until glycerol is completely consumed
and then V in the framework of W–Nb–O starts working for the
oxidation of accumulated acrolein to acrylic acid.
3.5. Reaction pathway of the glycerol transformation over
W–V–Nb–O catalysts
Reaction pathways of the glycerol transformation on
W–V–Nb–O are shown in Scheme 1. On W–Nb–O framework
site, two pathways, pathway 1 where acrolein is formed by
the dehydration of glycerol and pathway 2 where the formed
acrolein is oxidized to COx, subsequently proceed. On V sites
in the W–Nb–O framework, pathway 3 where acrylic acid is
formed by the oxidation of acrolein can take place, followed by
pathway 4 where acrylic acid is oxidized to COx. Pathway 5 is
the oxidation of glycerol proceeding on V site. As shown in Fig. 5,
over W2.4V0.1Nb2.5O14, acrolein forms in a relatively high yield
Scheme 1. Reaction scheme for the transformation of glycerol to acrylic acid over
W–V–Nb–O catalysts.

K. Omata et al. / Catalysis Today 259 (2015) 205–212
211
Fig. 6. O2 conversion (A) and yield of acrolein (B), acrylic acid (C) and COx (D) as the function of the glycerol conversion. Reaction conditions: temperature, 558 K; composition
of reaction gas, glycerol/O2/N2/H2O = 5/14/56/25 (mol%). Symbols: W2.8Nb2.2O14 ᭹; W2.4V0.1Nb2.5O14 ꢀ; W2.2V0.4Nb2.4O14 ꢁ; W2.2V0.5Nb2.3O14 ꢁ; W2.2V0.9Nb1.9O14 ×.
We
simply
synthesized
and 2.5 wt%PO4/W2.2V0.4Nb2.4O14
two
materials,
well as the acrolein formation. Phosphoric acid seems to interact
with surface V site to moderate the oxidation ability and to
suppress the pathway 4 mainly.
2.5 wt%PO4/W2.8Nb2.2O14
,
by impregnation method as described in Section 2. The catalytic
performance of W2.8Nb2.2O14 and 2.5 wt%PO4/W2.8Nb2.2O14 in the
glycerol transformation was first examined and the results are
shown in Table 3. Both the glycerol conversion and the acrolein
yield were appreciably increased by the phosphoric acid addition,
and the 2.5 wt%PO4/W2.8Nb2.2O14 catalyst achieved the acrolein
yield of 81.8%. This improvement seems due to the changes of
surface acidity. It was observed that the number of acid sites
per gram largely increased by the addition of phosphoric acid to
244 ␮mol g−1 and the ratio of Brönsted to Lewis acidity increased
from 0.4 to 1.5. This is probably due to a formation of heteropoly-
type unit over the surface. Moreover, it was observed in the FT-IR
study that water substantially decreased the Lewis acid sites and on
the other hand increased the Brönsted acid sites (B/L ratio to be 23.4
over 2.5 wt%PO4/W2.8Nb2.2O14). This result indicates that Lewis
acid sites change into Brönsted acid sites in the presence of water.
4. Conclusions
While the orthorhombic-like W–Nb–O gives acrolein selec-
tively in the glycerol transformation, acrylic acid formed from
glycerol directly in the presence of oxygen by the introduction
of V into the W–Nb–O. W increased the Brönsted acidity of the
catalysts, so that the activity for the dehydration of glycerol to
acrolein was enhanced. V in the framework little affected the
acrolein formation from glycerol but exhibited high selectivity of
acrylic acid in the oxidation of acrolein. V appears to be essential
element for the oxidation of acrolein to acrylic acid. By chang-
ing the V content in W–Nb–O, a maximum yield of acrylic acid
from glycerol 46.2% was achieved over W2.2V0.4Nb2.4O14 under
W/F = 6.7 × 10−3 g min ml−1 at the reaction temperature of 580 K.
The addition of phosphoric acid to W2.8Nb2.2O14 increased the
acid amount and the Brönsted acidity of the W2.8Nb2.2O14 catalyst.
In addition to the acidity change, phosphoric acid interacts with
V sites for suppressing the sequential oxidation of acrylic acid
to COx. As a result, acrylic acid yield in the glycerol transforma-
tion increased significantly. The 2.5 wt%PO4/W2.2V0.4Nb2.4O14
catalyst gave acrylic acid yield of 59.2% directly from glycerol
under oxidative condition (W/F = 1.0 × 10−2 g min ml−1, reaction
temperature 594 K). The catalytic performance evidences provided
in the present work may prove a possibility of the industrial
realization of the direct transformation of glycerol to acrylic
acid.
The
same
effect
of
the
phosphoric
acid
addi-
tion was also observed in the W2.2V0.4Nb2.4O14 and
2.5 wt%PO4/W2.2V0.4Nb2.4O14 for the direct glycerol transfor-
mation to acrylic acid and the results are also shown in Table 3.
The addition effect is prominent and the attained maximum yield
of acrylic acid was 59.2% under a suitable contact time condition,
which is the highest reported for the direct transformation of
glycerol to acrylic acid [18–21]. Of additional importance is that
the yield of acrylic acid was kept almost constant during 5 h of the
reaction.
In order to elucidate the role of phosphoric acid in this course of
the transformation, we conducted acrolein oxidation over W2.2V0.4
Nb2.4O14
and
2.5 wt%PO4/W2.2V0.4Nb2.4O14
.
It
was
observed that the maximum acrylic acid yield was 64.4%
(581 K) over W2.2V0.4Nb2.4O14 and 71.7% (577 K) over
References
2.5 wt%PO4/W2.2V0.4Nb2.4O14
enhances the selectivity to acrylic acid in the acrolein oxidation as
, indicating that phosphoric acid
[1] B. Katryniok, S. Paul, V. Bellière-Baca, P. Reye, F. Dumeignil, Green Chem. 12
(2010) 2079.

212
K. Omata et al. / Catalysis Today 259 (2015) 205–212
[2] S. Gil, M. Marchena, L. S-Silva, A. Romero, P. Sánchez, J.L. Valverde, Chem. Eng.
[16] F. Cavani, S. Guidetti, L. Marinelli, M. Piccinini, E. Ghedini, M. Signoretto, Appl.
Catal. B 100 (2010) 197.
J. 178 (2011) 423.
[3] Y. Kusunoki, T. Miyazawa, K. Kunimori, K. Tomishige, Catal. Commun. 6
(2005) 645.
[4] A.C.-C. Chang, W.-H. Lin, K.-H. Lin, C.-H. Hsiao, H.-H. Chen, H.-F. Chang, Int.
Hydrogen Energy J. 37 (2012) 13110.
[17] J.A. Dean, Lange’s Handbook of Chemistry, 1985, pp. 1–9.
[18] J. Deleplanque, J.-L. Dubois, J.-F. Devaux, W. Ueda, Catal. Today 157 (2010) 351.
[19] M.D. Soriano, P. Concepción, J.M.L. Nieto, F. Cavani, S. Guidetti, C. Trevisanut,
Green Chem. 13 (2011) 2954.
[5] M.S. Khayoon, B.H. Hameed, Appl. Catal. A 433 (2012) 152.
[6] M. Kapoor, M.N. Gupta, Process Biochem. 47 (2012) 503.
[7] H.J. Lee, D. Seung, K.S. Jung, H. Kim, I.N. Filimonov, Appl. Catal. A 390 (2010)
235.
[20] A. Chieregato, F. Basile, P. Concepción, S. Guidetti, G. Liosi, M.D. Soriano, C.
Trevisanut, F. Cavani, J.M.L. Nieto, Catal. Today 197 (2012) 58.
[21] A. Chieregato, M.D. Soriano, F. Basile, G. Liosi, S. Zamora, P. Concepción, F.
Cavani, J.M.L. Nieto, Appl. Catal. B 150 (2014) 37.
[8] A. Martin, M. Richter, Eur. J. Lipid Sci. Technol. 113 (2011) 100.
[9] T. Tago, T. Masuda, WO 2011108509 (2011).
[22] Y. Moro-oka, W. Ueda, Adv. Catal. 40 (1994) 233.
[23] M. Sadakane, K. Kodato, T. Kuranishi, Y. Nodasaka, K. Sugawara, N. Sakaguchi,
T. Nagai, Y. Matsui, W. Ueda, Angew. Chem. Int. Ed. 47 (2008) 2493.
[24] T. Murayama, N. Kuramata, S. Takatama, K. Nakatania, S. Izumia, X. Yi, W.
Ueda, Catal. Today 185 (2012) 224.
[25] K. Omata, S. Izumi, T. Murayama, W. Ueda, Catal. Today 201 (2013) 7.
[26] Y. Oka, T. Yao, N. Yamamoto, Nippon Seramikkusu Kyokai Gakujutsu
Ronbunshi 98 (1990) 1365.
[10] Y.T. Kim, K.-D. Jung, E. Duck Park, Appl. Catal. A 393 (2011) 275.
[11] C.-J. Jia, Y. Liu, W. Schmidt, A.-H. Lu, F. Schüth, J. Catal. 269 (2010) 71.
[12] E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catal. Commun. 8 (2007) 1349.
[13] A. Alhanash, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. A 378 (2010) 11.
[14] M. Massa, A. Andersson, E. Finocchio, G. Busca, F. Lenrick, L.R. Wallenberg, J.
Catal. 297 (2013) 93.
[15] A. Ulgen, W.F. Hoelderich, Appl. Catal. A 400 (2011) 34.
[27] K. Omata, K. Matsumoto, T. Murayama, W. Ueda, Chem. Lett. 43 (2014) 435.