New efficient and long-life catalyst for gas-phase glycerol dehydration to acrolein

Article abstract of DOI:10.1016/j.jcat.2011.05.014

Zirconium and niobium mixed oxides have been shown to be selective catalysts for dehydration of glycerol to acrolein, at 300 °C in the presence of water. The catalysts exhibit a selectivity to acrolein of approximately 72%, at nearly total glycerol conversion. More selective catalysts have been designed for this reaction, but whereas all of these are almost completely deactivated after only 24 h, ZrNbO catalysts still exhibit 82% conversion efficiency after 177 h on stream, its acrolein selectivity remaining unimpaired. Catalysts have been characterized by various techniques, showing that active and selective sites are weak or moderately acid Br?nsted sites, whose presence is related to that of polymeric niobium oxide species in interaction with the zirconia support. The key stability of the catalysts has been attributed to the neutralization of Lewis acid of uncovered zirconia, which are unselective coke initiator sites.

Full text of DOI:10.1016/j.jcat.2011.05.014

Journal of Catalysis 281 (2011) 362–370
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
New efficient and long-life catalyst for gas-phase glycerol dehydration to acrolein
a
b
c
P. Lauriol-Garbey a, J.M.M. Millet a, , S. Loridant , V. Bellière-Baca , P. Rey
⇑
a Institut de Recherches sur la Catalyse et l’Environnement de Lyon, IRCELYON, UMR5256 CNRS-Université Claude Bernard Lyon 1, 2 avenue A. Einstein,
F-69626 Villeurbanne cedex, France
b Rhodia, 52 rue de la Haie Coq, 93308 Aubervilliers cedex, France
c Adisseo, Antony Parc 2, 10 Place du Général de Gaulle, 92160 Antony, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 23 June 2011
Zirconium and niobium mixed oxides have been shown to be selective catalysts for dehydration of glyc-
erol to acrolein, at 300 °C in the presence of water. The catalysts exhibit a selectivity to acrolein of
approximately 72%, at nearly total glycerol conversion. More selective catalysts have been designed for
this reaction, but whereas all of these are almost completely deactivated after only 24 h, ZrNbO catalysts
still exhibit 82% conversion efficiency after 177 h on stream, its acrolein selectivity remaining unim-
paired. Catalysts have been characterized by various techniques, showing that active and selective sites
are weak or moderately acid Brønsted sites, whose presence is related to that of polymeric niobium oxide
species in interaction with the zirconia support. The key stability of the catalysts has been attributed to
the neutralization of Lewis acid of uncovered zirconia, which are unselective coke initiator sites.
Ó 2011 Elsevier Inc. All rights reserved.
Keywords:
Niobium oxide supported on zirconia
Heterogeneous acid catalysis
Interaction with support
Stability on stream
Dehydration
Glycerol
Acrolein
1. Introduction
lead to contradictory results concerning the nature and strength
of the catalytic acid sites, which are efficient for the reaction. One
With the irrevocable decrease in the availability of raw fossil
materials, a changeover of the chemical industry to renewable feed-
stock is needed. Because it will be produced in large quantities, as a
co-product with biodiesel, glycerol is a potential initial building
block for the production of many chemicals. Numerous studies have
thus been undertaken to develop means to exploit glycerol in the
most valuable manner. Among the more intensively studied exam-
ples of such developments, is the dehydration of glycerol to yield
acrolein, which is an important intermediate chemical for the
agro-industry [1–3].
Although a liquid-phase medium, including sub and supercriti-
cal water, has been proposed to run the reaction, the gas-phase re-
mains the most economically viable method, even though a diluted
aqueous solution of glycerol has to be used as starting material, and
vaporized above 250 °C [2,3]. Several acidic catalysts have been
proposed in recent years, for the gas-phase selective conversion
of glycerol into acrolein [4–20]. Three types of catalyst allow the
reaction to run successfully: (i) zeolites [4–9], (ii) heteropolyacids
supported or unsupported [10–14], and (iii) supported oxides
[15–18]. All of these exhibit high catalytic properties, at tempera-
tures in the range between 260 °C and 350 °C, with an acrolein
selectivity ranging between 70% and 80% at total glycerol conver-
sion. Although most of the studies appear to agree on the fact that
Brønsted acid sites are more efficient than Lewis acid sites, they
study reported that the catalyst’s efficiency in acrolein synthesis
was enhanced by medium acidity (À8.2 < H0 < 3.0) [21], whereas
another study concluded that weak acidity was more appropriate
[20], and yet another reported that the best selectivity to acrolein
was obtained with strong acidic catalysts [14]. Given these discrep-
ancies, it is likely that other parameters, such as the number and
distribution of acid sites at the surface of the catalysts, may be
the key to the catalyst’s properties. Tolerance to water has also been
proposed as a key parameter in the case of heteropolyacids [14].
Although most catalytic systems lead to a high selectivity to
acrolein at total glycerol conversion, very few maintain their cata-
lytic properties for more than 5–10 h. Although the catalyst deac-
tivation is due to extensive coke deposition on its surface, attempts
to limit the rate of coking by adding O2 or H2 in the gas phase, and
modifying (or not) the catalysts, have not been convincingly suc-
cessful [14,22]. Catalyst regeneration by means of coke combustion
has also been periodically attempted ex situ, but until present this
has again met with only limited success.
In the present study, we report on the development of new cat-
alysts, which are both efficient for the dehydration of glycerol into
acrolein, and more stable than those already proposed [4–20].
These catalysts, which correspond to NbZrO mixed oxides, are de-
scribed in this paper. The most efficient NbZrO mixed oxide cata-
lysts appear to be complex, with several phases. A comparative
study of various pure or supported phases was thus needed in or-
der to identify and characterize the active phase of such catalysts.
The effects of different catalytic reaction parameters or catalyst
characteristics have been studied, to optimize the process and
DOI of original articles: 10.1016/j.jcat.2011.05.013, 10.1016/j.jcat.2011.03.005
⇑
Corresponding author. Fax: +33 4 72 44 53 99.
E-mail address: Jean-marc.millet@ircelyon.univ-lyon1.fr (J.M.M. Millet).
0021-9517/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2011.05.014

P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
363
demonstrate possible regeneration of the catalyst. Finally, the
characterization of the acid–base properties of these catalysts has
been undertaken, in an effort to determine which of the catalysts’
sites were active, selective, or unselective.
respond to a commercial tungstated zirconia from Daiichi Kigenso
Kagaku Kogyo (reference Z-1104) and a commercial ZSM5 from
Zeochem (reference Zeocat PZ2/50H). The references are respec-
tively denoted ZrWO-DKK and HZSM-5.
2. Experimental
2.2. Catalysts characterization
2.1. Preparation of catalysts
Metal contents of the mixed oxides were determined by atomic
absorption (ICP) and their specific surface areas were determined
using the BET method with nitrogen adsorption. Powder X-ray
diffraction (XRD) patterns were obtained using a BRUKER D5005
2.1.1. Preparation of ZrNbO catalysts
The preparation method is based on the work of Kantcheva et al.
[23], consisting in impregnating hydrated zirconium oxide with a
peroxo-niobate solution. However, in order to avoid the use of per-
oxo-niobate leading to a strongly exothermic reaction difficult to
control, the impregnation was conducted in acidified water with
ammonium oxalato-niobate. Twenty-four mmol of ammonium
oxalato-niobate Nb(NH4)(C2O4)2NbOÁxH2O (Aldrich, 99.99%) were
first dissolved into 65 mL of de-ionized water acidified at
pH < 0,5 with concentrated nitric acid. The solution was then
heated about 40 min at 45 °C in order to dissolve the niobium com-
pound and cooled at room temperature before the addition of
hydrated zirconia. The reaction medium was maintained for 24 h
at 25 °C under stirring, and the solid was separated by filtration.
The solid obtained were calcined under air at 600 °C for 2 h
(heating rate 5 °C minÀ1). The heating protocol involved two steps
at 350 and 450 °C to eliminate the remaining nitrates. The samples
obtained were denoted as ZrNbO-x, with x the Zr/Nb ratio.
diffractometer with Cu Ka radiation (k = 0.154184 nm). They were
recorded with 0.02 (2h) steps over 10–80 angular range with 10 s
counting time per step. The unit cell of the zirconia phase was
refined using Bruker Topas P program.
Raman spectra were recorded using a LabRam HR spectrometer
(Jobin Yvon-Horiba). The exciting line at 514.53 nm of an Ar–Kr
2018 RM laser (Spectra Physics) was focused on several points of
the samples with a long working X50 objective. The laser power
on samples was limited to 1 mW. Several points were analyzed
to control the samples homogeneity at the micron scale.
Surface analyses by X-ray photoelectron spectroscopy (XPS)
were performed in a KRATOS Axis Ultra DLD equipped with a mag-
netic immersion lens, a hemispherical analyzer, and a delay line
detector. Experiments were carried out using a monochromated
Al K
a
X-ray source with a pressure below 5 Â 10À8 Pa in the anal-
ysis chamber. XPS spectra of Zr3d5/2, Nb3d5/2, O1s, and C1s levels
were measured at 90° (normal angle with respect to the plane of
the surface). A charge neutralizer was used to control charges
effects on samples.
2.1.2. Preparation of bulk and supported pure phases
Tetragonal zirconia (t-ZrO2) has been prepared by calcination of
a zirconium hydroxide also synthesized by co-precipitation at pH
8.8 from a 0.4 M solution of zirconyl nitrate (ZrO(NO3)2ÁxH2O Al-
drich, 99%) upon addition of an ammoniac solution (28%). The pre-
cipitate formed was maintained in the solution under stirring for
1 h and separated by centrifugation. It was washed several times
with de-ionized water to eliminate nitrates and dried over night
at 110 °C before grinding and calcination at 500 °C for 2 h.
Monoclinic zirconia (m-ZrO2) has been obtained from Saint-Gobain
Norpro (reference: XZ16075).
A sample of the ZrxÀ2Nb2O2x+1 solid solution with x = 9 referred
as Zr7Nb2O19, has been prepared using a sol–gel process [24].
Briefly, 10 g of Pluronic P-123 surfactant (BASF) was added under
stirring to 120 mL of anhydrous ethanol (Fluka). About 87.5 mmol
of ZrCl4 (Aldrich, 99.5%) and 25 mmol of NbCl5 (Aldrich, 99%) were
added to the solution (Zr/Nb molar ratio equal to 3.5). The mixture
was stirred for 30 min and maintained at 60 °C for one week. The
brown solid obtained was calcined at 350 °C for 20 h to decompose
the surfactant and then at 740 °C for 10 h to crystallize the phase
and eliminate the possible traces of chlorine.
Samples were characterized by solid-state 13C CPMAS NMR
(100.62 MHz, pulse time 3.5 ls, contact time 3 ms, 10 kHz spinning
rate) using a Bruker DSX-400 spectrometer and a 4 mm CPMAS
probe to obtain some information on the coke formed at the sur-
face of the catalysts after catalytic testing.
TPD of ammonia was performed on an apparatus Autochem
2910 from micromeritics, with 100 mg of catalysts. The samples
were pre-treated 30 min at 300 °C under helium (30 mL minÀ1
)
and the successively treated in a flux of 5% of ammonia in helium
(30 mL minÀ1) for 15 min at room temperature and in helium for
1 h at 100 °C. The temperature of the calorimeter furnace was then
programmed with a heating rate of 10 °C minÀ1 at 30 mL minÀ1 he-
lium flow rate. The concentration of ammonia desorbing from the
sample was continuously monitored by gas chromatography. The
output was then plotted against temperature to get the TPD profile.
2.3. Catalysts testing
The prepared solids were tested for dehydration of glycerol at
300 °C under atmospheric pressure using a glass plug-flow mi-
cro-reactor. The catalyst volume (4.5 mL) was chosen to obtain
complete or near complete glycerol conversion with reference cat-
alysts. Prior to the reaction, the samples were pre-heated at 300 °C
for 15 min in flowing N2 (75 mL minÀ1). A 20 wt% aqueous glycerol
solution was vaporized with an inert gas flow to the reactor using a
Bronkhorst controlled evaporator mixer system. The aqueous glyc-
erol solution was fed at a rate of 3.8 g hÀ1, and the inert gas flow
(He/Ne: 90/1) was 75 mL minÀ1 giving a molar relative composi-
tion glycerol/water/inert gas of 2.3/46.3/51.4 (GHSV = 1930 hÀ1).
CO and CO2 were analyzed online, using a gas chromatograph
equipped with a TCD detector and using a Carboxen 1000 column,
Ne was used as internal standard. Organic substrates were con-
densed at low temperature during the reaction and analyzed off
line using gas chromatographs equipped with FID detectors and a
Nukol column or a ZBwax plus column.
Two zirconia supported Nb2O5 catalysts have been prepared by
impregnation. The support corresponds to a mixture of tetragonal
and monoclinic phases with a specific surface area of 95 m2 gÀ1. It
was obtained calcining zirconium hydroxide at 600 °C for 2 h. Ten
gram of the support was added to a 50 mL aqueous solution
containing 2.5 and 5.9 mmol of ammonium oxalato-niobate,
respectively, (corresponding to a Nb surface density of 3.3 and
6.6 atom nmÀ2 or 30% and 60% of a theoretical Nb2O5 monolayer).
The solution was aged for 1h30 at room temperature under
stirring, evaporated under vacuum, and calcined at 600 °C for 2 h
(heating rate 5 °C minÀ1). The solids will be designated as NbOx-
0.3/ZrO2 and NbOx-0.6/ZrO2.
2.1.3. Reference catalysts
Reference catalysts have been used to compare the catalytic
properties of the prepared catalysts. These reference catalysts cor-

364
P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
3. Results
3.1. Study of ZrNbO catalysts
Table 1 provides an overview of the detailed catalytic properties
obtained with the ZrNbO catalyst at 300 °C, under standard reaction
conditions. These data were obtained after 48 h on stream. The cat-
alysts needed between 10 and 20 h to reach their optimal selectiv-
ity, which then remained constant. During this induction period,
very strong and thus unselective acid sites should be neutralized
by coking. The best selectivity at or near to total conversion, equal
to 72%, was obtained with the ZrNbO-12 catalysts. This is slightly
lower than the best figures found in the literature, which lie above
80%. However, the catalysts appeared to be very stable on stream, as
shown in Fig. 1. For comparison, the evolution of the conversion on
reference phases described in the literature (HZSM-5 and ZrWO-
DKK), tested under the same conditions, is shown in the same fig-
ure. It should be noted that the catalytic performances obtained
with the latter phases were similar to those reported in the litera-
ture, which tends to confirm the correctness of our data acquisition
system. It can be seen in Fig. 1 that the increase in Nb content has a
positive effect on the stability of the catalysts. However, when the
time on stream was increased, all of the catalysts lost their activity,
whereas their selectivity to acrolein remained constant, with no
significant changes observed in their by-product distribution.
The ZrNbO catalysts were characterized using X-ray diffraction
and Raman spectroscopy. The X-ray diffraction patterns of the
ZrNbO samples reveal two phases, corresponding to m and t-
ZrO2 (Fig. 2). The relative m-ZrO2 content increased when the Nb
content decreased. This result is in agreement with previous stud-
ies showing that niobium helps in the stabilization of the tetrago-
nal form [24,25]. The cell parameters of the m and t-ZrO2 phases
have been calculated for the different compounds (Table 2). They
are comparable for the three catalysts, and lower than those of
the initially prepared phases without Nb. It can be noticed that
the tetragonality of the t-ZrO2 phase (the c/a ratio) increased
slightly with the Nb content. Similar behavior was already ob-
served with increasing substitution of t-ZrO2 by under-sized tetra-
valent cations [26].
Fig. 1. Evolution of glycerol conversion and selectivity to acrolein as a function of
time on stream at 300 °C for the ZrNbO-12 catalyst and for HZSM-5 and ZrWO-DKK
references. Flow rate of the 20% in weight glycerol solution: 3.8 g hÀ1, flow rate of
inert gas: 75 mL minÀ1, volume of catalyst: 4.5 mL.
of short Zr–O bonds, respectively [27], appeared with a shift of
9 cmÀ1 toward a higher frequency when compared to pure zirco-
nia, thus indicating an increase in the zirconium–oxygen bond
strength. This result can be explained by the formation of a solid
solution of Nb into ZrO2. In addition to the bands attributed to t-
ZrO2 (150, 278, 316, 457, and 649 cmÀ1) and to m-ZrO2 (179, 190,
223, 332, 378, 474, 558, and 558 cmÀ1), the room temperature
spectra revealed a large band centered at 774 cmÀ1, which has
previously been attributed to stretching frequencies of niobium–
oxygen (Nb–O) bonds in a nearly perfect NbO4 environment
[28]. A similar band was observed in the ZrO2–Y2O3–Nb2O5
Raman spectroscopy was also used to characterize the ZrNbO
catalysts. The spectra recorded at room temperature and at
600 °C are shown in Figs. 3 and 4, respectively. The room temper-
ature spectra were similar and confirmed the presence of both
tetragonal and monoclinic zirconia. The Raman bands at 278
and 649 cmÀ1 involve O–Zr–O bending and Zr–O stretching force
constants of long Zr–O bonds and O–Zr–O bending force constants
Table 1
Catalytic performances of the ZrNbO catalysts. Reaction temperature 300 °C, glycerol
aqueous solution (20 wt%) flow rate 3.8 g hÀ1, inert gas flow rate 75 mL minÀ1
.
Catalyst
ZrNbO-12
ZrNbO-19
ZrNbO-31
Specific surface area (m2 gÀ1
Mass of catalyst (g)
Reaction time (h)
Glycerol conversion (%)
Selectivity (%) to
Acrolein
)
55
7.5
48
99
53
7.5
48
74
7.5
48
100
100
72
59
51
Acetaldehyde
Propanal
Acetone
2,3-Butanedione
Allylic alcohol
Hydroxyacetone
Phenol
3.8
2.1
1.1
0.7
1.1
11.3
0.9
1.5
–
5.8
3.3
1.0
0.8
1.9
13.0
0.9
2.2
–
7.2
3.9
2.8
0.8
3.5
24.0
0.5
2.1
–
CO
CO2
Fig. 2. X-ray diffraction patterns of the catalysts (a) ZrNbO-31, (b) ZrNbO-19 and
ZrNbO-12 (c) catalysts.
Carbon balance (%)
95
89
97

P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
365
Table 2
Cell parameters of the m- and t-ZrO2 of the ZrNbO and NbOx/ZrO2 catalysts calculated from X-ray diffraction data.
Catalyst
t-ZrO2 cell parameters
m-ZrO2 cell parameters
a (nm)
c (nm)
V (nm3)
a (nm)
b (nm)
c (nm)
b (°)
V (nm3)
NbOx-0.3/ZrO2
NbOx-0.6/ZrO2
ZrNbO-31
ZrNbO-19
ZrNbO-12
0.3631(3)
0.3634(5)
0.3594(1)
0.3598(1)
0.3594(1)
0.5092(5)
0.5094(5)
0.5171(3)
0.5177(3)
0.5180(2)
0.0671
0.0672
0.0667
0.0669
0.0669
0.5155(9)
0.5158(9)
0.5145(3)
0.5135(2)
0.5140(3)
0.5247(7)
0.5250(7)
0.5233(4)
0.5236(2)
0.5236(4)
0.5266(4)
0.5265(8)
0.5257(3)
0.5257(2)
0.5246(4)
81.301(2)
81.347(2)
81.168(3)
81.129(9)
81.129(3)
0.1407
0.1408
0.1398
0.1384
0.1383
Fig. 5. Schematic representation of Zr environment in t-ZrO2 with 4 oxygen at
0.2453 nm and 4 oxygen at 0.2072 nm forming a regular tetrahedron (all O–Zr–O
equal to 103.215).
Fig. 3. Raman spectra of (a) zirconia (b) ZrNbO-31, (c) ZrNbO-19, and (d) ZrNbO-12
at room temperature in ambient air.
of the NbO4 tetrahedra in the case of Nb substitution to Zr (Fig. 5).
The stretching frequencies of the niobium–oxygen (Nb–O) bonds
of the supported niobium species are normally observed in the
range between 850 and 930 cmÀ1 in ambient air [30]. However,
the presence of hydration water leads to van der Waals interac-
tions, which weaken and enlarge the corresponding bands, such
that it is preferable to dehydrate the compound, in order to high-
light these bands. The Raman spectra were thus recorded under
dry air conditions at 600 °C (Fig. 4). All of these are characterized
by a band at 977 cmÀ1, which is intermediate between the fre-
quencies observed for niobium oxide supported on zirconia with
a low content (1%, 958 cmÀ1), and that with a high content (5%,
988 cmÀ1), which almost corresponds to a monolayer [31]. The
catalysts should thus contain niobium oxide supported on zirco-
nia as polymers. The nuclearity of these polymers should vary
strongly, as evidenced by the high width of the bands. Finally, it
should be emphasized that during Raman spectroscopy measure-
ments, several areas of each catalyst were analyzed under the
microscope. Point-to-point comparison of the spectra has shown
good homogeneity of the catalysts and confirmed the absence of
other phases such as bulk niobium oxide. In the case of the
ZrNbO-12 sample only, very few analyses were able to reveal
the presence of the Zr7Nb2O19 phase, with its characteristic band
at 418 cmÀ1. As the presence of this phase was not detected by X-
ray diffraction, it is known to have been present in only very small
quantities.
Fig. 4. Raman spectra of the catalyst at 600 °C under dry air (a) ZrNbO-31, (b)
ZrNbO-19, and (c) ZrNbO-12.
3.2. Study of pure bulk and supported phases
system and was attributed to NbO4 tetrahedra in ZrO2 [29]. In
tetragonal zirconia, Zr cations have eightfold coordination to oxy-
gen with 4 O at 0.207 nm and 4 O at 0.245 nm. The 4 first oxygen
anions form regular tetrahedra, which would fit with the presence
In order to understand the origin of the catalytic properties of
the catalysts, the phases detected by X-ray diffraction or Raman
spectroscopy were synthesized in their pure form a and tested as

366
P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
the presence of both monoclinic and tetragonal zirconia. They are
similar, apart from the fact that very small peaks corresponding
to Nb2O5 (ICSD: 027-1003) can be seen in the NbOx-0.6/ZrO2 pat-
tern. The homogeneous spreading of niobium at the surface of al-
ready crystallized zirconia is difficult, and niobium oxide is
formed before a quantity of niobium equivalent to a monolayer
can be deposited.
The pure phases and supported catalysts were tested as cata-
lysts (Table 3). The pure zirconia phases appeared to be active,
but not selective to acrolein, and hydroxyacetone was the major
product formed on these phases. The Zr7Nb2O19 phase was both ac-
tive and selective, although less than the ZrNbO catalysts, and
markedly less stable. The catalytic properties of the NbOx/ZrO2 cat-
alysts with different Nb contents (À0.3 and À0.6) appeared to be
similar. The solids are more active, selective, and stable than the
pure ZrO2 phases. However, their performance did not attain that
of the ZrNbO catalysts. The similarities in catalytic performance
of the two NbOx/ZrO2 catalysts can be explained by the fact that,
because of the formation of bulk Nb2O5 in the case of the catalysts
with the richer Nb content, the two catalysts most likely have sim-
ilar niobium dispersion characteristics at the surface of zirconia.
Since the bulk niobium oxide phase obtained after heat treatment
at 600 °C is poorly active in the reaction [21], it is understandable
that the catalytic data are the same for the two catalysts. It should
be noted that the literature reports homogeneous dispersion of
niobium at the surface of zirconia, with a thickness greater than
0.6 monolayer [32]. However, most of the reported data were ob-
tained on a nearly pure monoclinic phase, which is not the case
of the solids containing large quantities of the tetragonal phase de-
scribed here.
3.3. Surface characterization of the catalysts
The results obtained on the pure phases and the NbOx/ZrO2 cat-
alysts lead to the conclusion that the catalytic properties of the
ZrNbO catalysts are related to the surface niobium polymer species
formed by the new preparation method. It was thus important to
achieve improved characterization of the surface of the catalysts.
This was done using XPS spectroscopy and ammonia TPD.
The results obtained with XPS spectroscopy are presented in
Table 4. They show that the surface of the ZrNbO was richer in
Nb than the bulk material, which is in agreement with Raman
spectroscopy showing the presence of surface polymeric niobium
oxide species. Furthermore, analysis of the C1s signal revealed the
presence of a carbonate species (288.8 0.001 eV) at the surface
of the catalysts, the quantity of which decreased when the Nb con-
tent increased. The carbonate species results from the adsorption
of atmospheric CO2 by the strongly basic oxygen ions (Lewis type)
of the uncovered zirconium oxide. It should be pointed out that the
Zr7Nb2O19 mixed oxide has a relatively high surface content of
strong basic sites.
Fig. 6. X-ray diffraction pattern of the pure phases A: (a) Zr7Nb2O17, (b) m-ZrO2, (c)
t-ZrO2 and of the supported NbOx/ZrO2 catalysts B: (a), NbOx-0.3/ZrO2, (b) NbOx-0.6/
ZrO2.
catalysts, as well as catalysts corresponding to niobium oxide sup-
ported on zirconia. The X-ray diffraction pattern of Zr7Nb2O19, the
two ZrO2 phases (m and t), and two Nb supported catalysts with
different niobium contents are presented in Fig. 6. The pattern of
the mixed zirconium and niobium compound corresponds well
to the pure phase (ICSD: 012-2312). The patterns of the Nb
supported catalysts (NbOx-0.3/ZrO2 and NbOx-0.6/ZrO2) reveal
Fig. 7 and Table 5 show the ammonia TPD results on different
samples. All of the thermograms have two broad maxima: the first
of these is found at approximately 170 °C, and the second occurs in
Table 3
Catalytic performances of the single phases and NbOx supported zirconia-based catalysts. Reaction temperature 300 °C, glycerol aqueous solution
(20 wt%) flow rate 3.8 g hÀ1, inert gas flow rate 75 mL minÀ1
.
Catalyst
SA (m2 gÀ1
)
m (g)
TOS (h)
Conv. (%)
SAC (%)
SHA (%)
CB (%)
Zr7Nb2O19
NbOx-0.3/ZrO2
NbOx-0.6/ZrO2
m-ZrO2
41
88
84
53
98
7.5
8.9
8.9
7.5
3.8
8
8
7
8
10
88
99
100
97
60
35
35
7
7.6
5.7
2.6
36.0
26.0
88
74
66
65
82
t-ZrO2
73
21
SA = specific surface area, m = mass of catalyst, TOS = time on Stream, Conv = conversion of glycerol, SAC = selectivity to acrolein, SHA = selectivity
to hydroxyacetone, CB = carbon balance.

P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
367
Table 4
Table 6
XPS analyses of different studied catalysts. The surface Nb/Zr ratio is compared with
the bulk ratio calculated from results of chemical analyses. B.E.: binding energy.
Catalytic performances of the ZrNbO catalysts at different reaction temperature,
glycerol aqueous solution (20 wt%) flow rate 3.8 g hÀ1
75 mL minÀ1, catalyst volume 4.5 mL (7.5 g).
, inert gas flow rate
Catalyst
Nb/Zr
B.E. CCO3 (eV)
CO3/Zr
Reaction temperature (°C)
Glycerol conversion (%)
Reaction time (h)
Selectivity (%) to
Acrolein
Acetaldehyde
Propanal
Acetone
2,3-Butanedione
Allylic alcohol
Hydroxyacetone
Phenol
CO
CO2
280
64
24
290
93
28
300
100
32
Surface
Bulk
(m + t) ZrO2
NbZrO-31
NbZrO-12
NbOx-0.3/ZrO2
Zr7Nb2O19
0.00
0.15
0.73
0.10
0.41
0.00
0.03
0.09
0.03
0.28
288.6
288.7
288.9
288.7
288.9
0.13
0.10
0.06
0.06
0.07
71.0
1.5
1.0
0.2
0.0
1.1
16.6
0.1
1.5
–
71.0
2.4
1.4
0.3
0.3
1.3
16.3
0.3
2.2
–
70.0
4.5
2.6
0.8
0.8
1.3
11.8
1.1
2.1
–
Carbon balance (%)
96
97
100
Fig. 7. NH3-TPD thermograms of ammonia on different solids: (a) (m + t)ZrO2, (b)
NbOx-0.3/ZrO2, (c) ZrNbO-12, (d) Zr7Nb2O19
.
the range between 350 and 400 °C. No shift in the peaks’ maxima
was observed, from one sample to the other, although it is difficult
to define the position of the second peak as a consequence of its
broad profile. Compared with the Nb containing solids, ZrO2 exhib-
ited a more intense second peak, which can be attributed to stron-
ger acid sites. The density of strong acid sites is lower on the NbOx/
ZrO2 solid and decreased even further on the ZrNbO-12 catalyst
presenting a higher Nb surface content. The Zr7Nb2O19 has a smal-
ler number of weak and strong acid sites.
Fig. 8. Evolution of glycerol conversion and selectivity to acrolein as a function of
time on stream at 300 °C for the ZrNbO-12 catalyst with inert gas flow rates of
75 mL minÀ1 (black and white triangles) and 200 mL minÀ1 (black and white
squares). Flow rate of the 20% in weight glycerol solution: 3.8 g hÀ1, flow rate of
inert gas: 75 mL minÀ1, and volume of catalyst: 4.5 mL.
selectivity to acrolein was obtained, and the reaction temperature
was decreased successively to 290 °C, and then to 280 °C, for test-
ing. When the temperature decreased the activity of the catalyst
decreased, but the selectivity to acrolein did not change. In parallel,
the selectivity to hydroxyacetone increased, and that to acetalde-
hyde, 2,3-butanediol and phenol decreased. This can be explained
by the fact that hydroxyacetone is decomposed to a lesser extent
and that the latter compounds are the products of this decomposi-
tion. Because several of these products are suspected to be coke
formation precursors, it is probable that a lower reaction temper-
ature led to an improved temporal stability of the catalysts.
The effect of flow rate of the diluting gas was studied by testing
the same catalyst under the same conditions, with the exception of
3.4. Study of several reaction parameters and regenerability
The influences of the reaction temperature and the flow rate of
the diluting gas on the catalytic properties of the catalysts were
studied.
The catalytic properties of the ZrNbO-12 catalyst obtained at
different temperatures are presented in Table 6. This catalyst was
first tested at 300 °C for 24 h. After that time, the optimal
Table 5
Results of NH3 TPD on different catalysts.
the nitrogen flow, which was increased from 75 to 200 mL minÀ1
.
Catalyst
Volume of NH3
adsorbed
First peak
temperature
(°C)
Total acid site
concentration
The glycerol contact time was thus decreased from 1.9 to 1.0 s.
As seen in Fig. 8, the increase in nitrogen flow rate had no influence
on the selectivity to acrolein, but slowed the deactivation rate.
With a 200 mL.minÀ1 nitrogen flow rate, the glycerol conversion
still remained at approximately 90% after 200 h, with selectivity
to acrolein of 71%. The increase in nitrogen flow rate delayed the
(mL gÀ1
)
(mmol mÀ2
)
ZrNbO-12
4.8
2.9
9.0
161
170
168
211
3.5
3.0
4.8
4.5
ZR7Nb2O19
NbOx-0.3/ZrO2
(m + t) ZrO2
11.4

368
P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
rate, were observed. The catalysts are thus fully regenerated by
the heat treatment in air.
The carbon deposited at the surface of the ZrNbO catalyst was
characterized by 13C NMR testing. The spectrum obtained with a
ZrNbO catalyst after catalytic testing was characterized by two
large bands around 127 and 32 ppm corresponding to aromatic
and aliphatic carbons, respectively (Fig. 10) [33]. The two narrow
peaks at 64 and 74 ppm were attributed to adsorbed glycerol, or
products derived from glycerol such as polyglycerols or acetaliza-
tion products of glycerol, adsorbed onto the surface of the catalyst.
All of these detected species can be expected to contribute to the
deactivation of the catalysts. It is believed that coke formation oc-
curs on strong catalyst acid sites, and that polyglycerols or acetal-
ization products of glycerol are formed and remain adsorbed on
basic sites [34,35]. Both types of site are certainly involved in the
deactivation of the catalysts.
4. Discussion
The results obtained show that the ZrNbO catalysts prepared
according to the protocol described in the experimental section
are efficient for the dehydration of glycerol to acrolein. With a total
conversion of only 72%, they do not have the best selectivity to
acrolein, when compared to other catalysts such as heteropoly-
acid-based catalysts, whose selectivity exceeds 80% [36]. It should
be emphasized that it is not straightforward to compare ZrNbO
catalysts with other catalysts, tested under the same conditions,
since the former have not reached their optimal selectivity at the
time when the latter have already undergone very strong deactiva-
tion. This feature is related to the incomparable stability of the
ZrNbO catalysts, which maintain a glycerol conversion greater than
90% for more than 200 h (Fig. 8), whereas other catalysts are al-
most completely deactivated after the same time on stream. It
should be mentioned that the catalytic sites in the catalysts might
not be all working at the beginning of the testing since it is difficult
to adjust the mass of catalyst to have exactly 100% glycerol conver-
sion. It is thus possible that the stability may be over-estimated by
some hours, but this should be very limited compare with the total
time on stream.
Fig. 9. Evolution of the glycerol conversion and acrolein selectivity as a function of
time at 300 °C for 150 h before and after an in situ regeneration at 450 °C in air for
1 h. Flow rate of the 20% in weight glycerol solution: 3.8 g hÀ1, flow rate of inert gas:
75 mL minÀ1, and volume of catalyst: 4.5 mL.
formation of coke by reducing the residence time of coke precur-
sors in the reactor.
Although the ZrNbO catalysts appeared to be relatively stable,
the introduction of a regeneration step into the process using these
catalysts needs to be envisaged. It was thus important to deter-
mine whether regeneration based on a simple heat treatment in
an oxidative atmosphere, designed to burn the coke formed at
the surface of the catalysts, could be efficient. After 150 h on
stream, the reaction flow was switched to an air-flow
(75 mL minÀ1), the temperature was increased to 450 °C and main-
tained at this temperature for 1 h. The catalyst was then cooled
down to the catalytic reaction temperature, and the initial catalytic
reaction flow re-introduced. The variations in glycerol conversion
and selectivity to acrolein are plotted as a function of time, before
and after this treatment (Fig. 9). It can be seen that the catalytic
properties were recovered following the heat treatment. The same
initial period for selectivity optimization and the same deactiva-
tion as a function of time on stream, with the same deactivation
Characterization of the ZrNbO catalysts has shown that the ac-
tive species correspond to polymeric NbOx species at the surface of
the zirconia support. Such supported species have already been
shown to generate Brønsted acid sites, which are active and selec-
tive for the dehydration reaction [32,37].
Ammonia TPD experiments clearly show that these acid sites
are, in general, relatively weak. If the catalytic properties of the
ZrNbO catalysts are related to the polymeric species distributed
over the surface of zirconia, the NbOx/ZrO2 catalyst should exhibit
similar catalytic properties. However, the latter type of catalyst is
neither as selective nor as stable. The characterization clearly
showed that for NbOx/ZrO2 catalysts, contrary to the case of ZrNbO
catalysts, no solid solution of Nb was formed into ZrO2. The forma-
tion of this solid solution thus appeared to be a key element in
obtaining highly efficient catalysts.
The formation of the solid solution of Nb into ZrO2 is related to
the preparation protocol used for the synthesis of the catalysts,
which facilitated the formation of such a solid solution. The prep-
aration was conducted in water at pH < 0.5. At this pH, the niobium
species present should correspond to the anionic species
[NbO(C2O4)(OH)2(H2O)]À or [NbO(OH)4(H2O)]À [38], which could
efficiently exchange with [OH]À groups of the zirconium hydroxide
Zr(OH)4 precursor. Such an exchange is not possible, or remains
very limited, for catalysts prepared by the impregnation of ZrO2
supports, with a prior heat treatment at 600 °C, as in the case of
the NbOx/ZrO2 catalysts.
Fig. 10. 13C CP-MAS n.m.r. spectrum of coke on spent catalyst.

P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
369
The dissolution of Nb is accompanied by a decrease in the lat-
tice parameters of the zirconium oxide phase. This may be related
to the fact that the cationic size of Nb5+ (0.070 nm) is smaller than
the cationic size of the host Zr4+ (0.084 nm) ion. These results are in
agreement with those given in the literature [39], which also re-
port that at thermodynamic equilibrium the solid solution of Nb
in ZrO2 reached a concentration of at least 5% (Zr/Nb = 9.5).
Although the starting solid composition of the catalysts could only
allow the formation of such a solid solution, the latter is not com-
pletely formed. Because the Nb and Zr homogeneity in the formed
precursor is not optimal, and because the heating temperature, i.e.,
600 °C, is not sufficiently high, a completely homogeneous solid
solution is not formed and part of the niobium remains at the sur-
face of the catalyst in the form of a polymeric species.
The significant formation of the solid solution certainly allows
the number of strong acid sites (Lewis type acid sites) present at
the surface of the ZrO2 support phase to be decreased, as well as
the number of basic sites as evidenced by XPS spectroscopy. These
later sites have been shown to be unselective for glycerol dehydra-
tion to acrolein [16,40]. They may also be responsible like the strong
acid sites, but to a lower extend, for the low stability of the cata-
lysts. With that respect several studies have shown that the coke
loadings were higher on most acidic catalysts [20,21,41]. The deac-
tivation of catalysts for glycerol dehydration, through the deposi-
tion of carbon on their surface during the catalytic reaction, may
also be influenced by several other parameters such as the reaction
temperature, the reactant feed, and the porosity of the support. For
the ZrNbO catalysts, a pore size of approximately 7 nm and a reac-
tion temperature of 300 °C appeared to be efficient parameters for
long-term stability. A more complete study would be needed, to
determine whether these are the optimum values for long-term
stability. A high diluting gas flow rate, which reduces the residence
time of coke precursors in the reactor, has also been shown to de-
crease the rate of deactivation. For such a case, a compromise be-
tween the slower deactivation and the lower productivity needs
to be found.
The bulk Nb2O5 phase and Nb2O5 supported on silica have both
been shown to be active and selective catalysts for the dehydration
of glycerol [18,41]. It is thus not surprising that niobium oxide sup-
ported on zirconia is an efficient catalyst for this reaction. Zirconia
certainly has an effect on the strength of the supported species, or
on their stabilization, which leads to better catalytic properties. It
is noteworthy that niobium-based oxides have mainly Brønsted
acid sites, but also Lewis acid sites at their surface [42]. Although
most papers agree that Brønsted acid sites are the most active
and selective for the dehydration of glycerol, the precise influence
of Lewis acid sites remains unknown. Studies have been under-
taken in order to better characterize the acid–base properties of
the new catalyst described here and to ascertain the effect of a zir-
conia support for the niobium species, as well as the role of Lewis
acid sites.
It is important to note that a previous study of a Nb2O5-based
catalyst for an acid catalyzed reaction reported that it was partic-
ularly resistant to coking [42]. It is thus credible that the high sta-
bility of the catalyst is related not only to the neutralization of the
Lewis acid and basic sites of the zirconia support, by incorporation
of niobium into zirconia, but also to an intrinsic property of the
niobium oxide species.
various discrepancies can be found in the literature. Some authors
have reported that the presence of tungsten oxide species on
monoclinic zirconium oxide does not activate the catalyst for isom-
erization reactions, because they exhibit only a weak acidity
[43,44], whereas other authors did not note any difference [45].
Presumably, in the study reported here, since the strength of the
active acid sites is weak, the nature of the zirconia phase may
not be a key factor, unless the non-selective zirconia sites are
neutralized.
The compound Nb2Zr6O17, which is a solid solution composition
with a modulated structure of the type Nb2ZrxÀ2O2x+1, with
7.1 6 x 6 10.3 [46], is characterized by a relatively high selectivity
to acrolein. This may be related to the presence of mainly weak
acid sites at its surface, with even fewer strong acid sites than
the most selective ZrNbO-12 catalyst. However, conversion on this
catalyst is not stabilized as a function of time, when on stream.
This may be explained by the presence of a higher content of basic
sites, as detected by XPS characterization, which could be respon-
sible for deactivation. Experiments conducted on basic catalysts
have shown that these types of catalyst are poorly active and
unstable [41].
5. Conclusion
This study has demonstrated that ZrNbO mixed oxides are
selective catalysts for the dehydration of glycerol to acrolein at
300 °C. These new catalysts appear to be particularly efficient,
since they deactivated only very slowly by coking. The efficiency
of the catalysts is related to that of the catalytic acid sites contrib-
uted by the niobium species, supported on zirconia, but also to the
neutralization of the non-selective sites of the support. The method
used to prepare the catalysts appears to be determinant in gener-
ating both features, since it allows the formation of a solid solution
of Nb in ZrO2, evidenced both by X-ray diffraction and Raman spec-
troscopy, and the spreading of polymeric niobium oxide species at
the surface of this solid solution. The results show that the same
catalyst could not be reproduced simply by grafting niobium oxide
onto the surface of zirconia. The deactivation has been shown to be
related to the formation of coke on strong Lewis acid sites and
eventually to the adsorption of polyglycerol on basic sites at the
surface of the catalysts. Both types of non-selective sites present
at the surface of the uncovered zirconia should be annihilated by
the dissolution of Nb into zirconia, which explains the high stabil-
ity of the prepared NbZrO catalysts. Although deactivation is
strongly slowed on these catalysts, it still occurs. However, a sim-
ple treatment with flowing air at a higher temperature than the
reaction temperature was found to be sufficient to completely
regenerate the deactivated catalysts back to their original level of
activity. This is a positive point for a possibly industrialized pro-
cess. Finally, it is noteworthy that the catalysts’ composition was
not fully optimized, such that there is room for further improve-
ment. The reaction conditions could certainly also be optimized.
Acknowledgments
We gratefully acknowledge ADISSEO for financial support. We
thank C. Lorentz and P. Delichere for 13C NMR and XPS analyses.
The formation of the solid solution of Nb in ZrO2 has implica-
tions in terms of the nature of the zirconia phase formed, since
Nb is an effective agent for the stabilization of t-ZrO2, when pre-
pared above 600 °C [24]. In this respect, the relative t-ZrO2 content
increased considerably, in accordance with the Nb content of the
catalysts. However, it is rather difficult to ascertain the influence
of the ZrO2 structure (monoclinic or tetragonal) on the catalytic
performance of the catalysts. In the case of WOx/ZrO2 catalysts,
References
[1] M. Pagliaro, M. Rossi, The future of glycerol: new usages for a versatile raw
material, RSC Green Chem. Ser. (2008).
[2] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem. 10 (2008)
13.
[3] B. Katryniok, S. Paul, M. Capron, F. Dumeignil, ChemSusChem 2 (2009) 719.
[4] C.J. Zhou, C.J. Huang, W.G. Zhang, H.S. Zhai, H.-L. Wu, Z.S. Chao, Stud. Surf. Sci.
Catal. 165 (2007) 527.

370
P. Lauriol-Garbey et al. / Journal of Catalysis 281 (2011) 362–370
[5] A. Corma, G.W. Huber, L. Sauvanaud, P. O’Connor, J. Catal. 257 (2008) 163.
[26] F. del Monte, W. Larsen, J.D. Mackenzie, J. Am. Ceram. Soc. 83 (2000) 628.
[27] P. Bouvier, H.C. Gupta, G. Lucazeau, J. Phys. Chem. Sol. 62 (2001) 873.
[28] F.D. Hardcastle, I.E. Wachs, Solid State Ionics 45 (1991) 201.
[29] D.Y. Lee, J.W. Yang, D.J. Kim, Ceram. Int. 27 (2001) 291.
[30] J.M. Jehng, I.E. Wachs, J. Mol. Catal. 67 (1991) 369.
[6] W. Suprun, M. Lutecki, T. Haber, H. Papp, J. Mol. Catal. A: Chem. 309 (2009) 71.
[7] C.J. Jia, Y. Liu, W. Schmidt, A.H. Lu, F. Schueth, J. Catal. 269 (2010) 71.
[8] M. Okuno, E. Matsunami, T. Takahashi, H. Kasuga, M. Okada, M. Kirishik, WO
2007132926, 2007.
[9] X.Z. Li, CN 101070276, 2007.
[31] L.J. Burcham, J. Datka, I.E. Wachs, J. Phys. Chem. B 103 (1999) 6015.
[32] T. Onfroy, G. Clet, M. Houalla, J. Phys. Chem. B 109 (2005) 4588.
[33] D.A. Netzel, F.P. Miknis, J.M. Mitzel, Tiejun Zhang, P.D. Jacobs, H.W. Haynes Jr.,
Fuel 15 (12) (1996) 1397.
[34] M.T. Martinez, M.D. Martinez, J. Osacar, J. Fuel Process. Technol. 18 (1988) 51.
[35] J.M. Clacens, Y. Pouilloux, J. Barrault, Appl. Catal. A 227 (2002) 181.
[36] B. Katryniok, S. Paul, M. Capron, C. Lancelot, V. Bellière-Baca, P. Rey, F.
Dumeignil, Green Chem. (2010).
[10] E. Tsukuda, S. Sato, R. Takahashi, T. Sodesawa, Catal. Commun. 8 (2007) 1349.
[11] H. Atia, U. Armbruster, A. Martin, J. Catal. 258 (2008) 71.
[12] L. Ning, Y. Ding, W. Chen, L. Gong, R. Lin, Y. Lü, Q. Xin, Chin. J. Catal. 29 (2008)
212.
[13] J.L. Dubois, Y. Magatani, K. Okumura WO 2009127889, 2009 (Arkema).
[14] A. Alhanash, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. A: Chem. 378
(2010) 11.
[15] J.L. Dubois, C. Duquenne, W. Hoelderich, J. Kervennal, WO 2006087084, 2006.
[16] A. Ulgen, W. Hoelderich, Catal. Lett. 131 (2009) 122.
[17] W. Hoelderich, A. Uelgen, DE 102008027350, 2009.
[18] N.R. Shiju, D.R. Brown, K. Wilson, G. Rothenberg, Top. Catal. 53 (2010) 1217.
[19] H. Redlingshoefer, C. Weckbecker, K. Huthmacher, A. Doerflein, WO
200809254, 2008.
[37] T. Onfroy, G. Clet, S.B. Bukallah, T. Visser, M. Houalla, Appl. Catal. A 298 (2006)
80.
[38] E. Asselin, T.M. Ahmed, A. Alfantazi, Corrosion Sci. 49 (2007) 694.
[39] J.C. Ray, A.B. Panda, C.R. Saha, P. Pramanik, J. Am. Ceram. Soc. 86 (2003) 514.
[40] A.K. Kinage, P.P. Upare, P. Kasinathan, Y.K. Wang, J.S. Chang, Catal. Commun. 11
(2010) 620.
[20] Q. Liu, Z. Zhang, Y. Du, J. Li, X. Yang, Catal. Lett. 127 (2009) 419.
[21] S.H. Chai, H.-P. Wang, Y. Liang, B.-Q. Xu, Green Chem. 9 (2007) 1130.
[22] F. Wang, J.L. Dubois, W. Ueda, J. Catal. 268 (2009) 260.
[23] M. Kantcheva, H. Budunoglu, O. Samarskaya, Catal. Commun. 9 (2008) 874.
[24] D. Lu, B. Lee, J. Kondo, K. Domen, Micropor. Mesopor. Mater. 203 (2004) 75.
[25] L. Mestres, M.L. Martinez-Sarrion, O. Castano, J. Fernandez-Urban, Z. Anorg.
Allg. Chem. 627 (2001) 294.
[41] S.H. Chai, H.-P. Wang, Y. Liang, B.-Q. Xu, J. Catal. 250 (2007) 342.
[42] Q. Sun, Y. Fu, H. Yang, A. Auroux, J. She, J. Mol. Catal. A 275 (2007) 183.
[43] Y. Huang, B. Zhao, Y. Xie, Appl. Catal. A 172 (1998) 327.
[44] F. Di Gregorio, N. Keller, V. Keller, J. Catal. 256 (2008) 159.
[45] J.C. Yori, J.M. Parera, Catal. Lett. 65 (2000) 205.
[46] K. Fütterer, S. Schmid, J.G. Thompson, R.L. Withers, Acta Crystallogr. B51
(1995) 688.