Acetalization of acetone with glycerol catalyzed by niobium-aluminum mixed oxides synthesized by a sol–gel process

Article abstract of DOI:10.1016/j.molcata.2016.02.002

Niobium-aluminum-based catalysts were synthesized by a sol–gel process and successfully applied to the reaction of acetalization of acetone with glycerol yielding 2,2-dimethyl-1,3-dioxolane-4-methanol (solketal) and 2,2-dimethyl-1,3-dioxan-5-ol. The synthesis procedure was developed using high-throughput techniques and the materials prepared with molar ratio1 Metal (1Nb:xAl): 100H₂O: 1.5NH₄OH (in which x?=?1, 0.6, 0.3, 0.1 and 0.05) were selected to be further investigated. The obtained series of mixed oxides displayed high catalytic activity reaching glycerol conversion up to 84%, with 98% of selectivity towards solketal. The physicochemical properties of the investigated catalysts were characterized by N₂-adorption/desorption, thermal analysis, X-rays diffraction, FTIR and Raman spectroscopy, and by adsorption of pyridine monitored by FT-IR spectroscopy and TGA. The catalysts are truly heterogeneous and can be reused in consecutive runs without loss of activity.

Full text of DOI:10.1016/j.molcata.2016.02.002

Accepted Manuscript
Title: Acetalization of acetone with glycerol catalyzed by
niobium-aluminum mixed oxides synthesized by a sol-gel
process
Author: Raphael Rodrigues Dalmo Mandelli Norberto S.
Gon c¸ alves Paolo P. Pescarmona Wagner A. Carvalho
PII:
S1381-1169(16)30036-X
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2016.02.002
MOLCAA 9767
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Revised date:
Accepted date:
29-9-2015
30-1-2016
1-2-2016
Please cite this article as: Raphael Rodrigues, Dalmo Mandelli, Norberto S.Gon c¸ alves,
Paolo P.Pescarmona, Wagner A.Carvalho, Acetalization of acetone with glycerol
catalyzed by niobium-aluminum mixed oxides synthesized by a sol-gel process, Journal
of Molecular Catalysis A: Chemical http://dx.doi.org/10.1016/j.molcata.2016.02.002
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Acetalization of acetone with glycerol catalyzed by niobium-
aluminum mixed oxides synthesized by a sol-gel process
a,c*
a
b
Raphael Rodrigues , Dalmo Mandelli , Norberto S.Gonçalves , Paolo P.
Pescarmona , Wagner A. Carvalhoa
c,d
aCentre of Natural and Human Sciences, Federal University of ABC, 166 Santa Adélia
Street, Bangu, Santo André, SP, 09210-170, Brazil.
bFederal University of São Paulo, 275 Prof. Artur Riedel Street, Jd. Eldorado, Diadema,
SP, 09972-270.
cCentre for Surface Chemistry and Catalysis, KU Leuven, KasteelparkArenberg
2
3,3001 Heverlee, Belgium.
dChemical Engineering Department, University of Groningen, Nijenborgh 4, 9747 AG,
Groningen, TheNetherlands.
*Corresponding author. Tel.: +55 (11) 4996-8386; Fax: +55 (11) 4996-8386. E-mail address:
rapharod@globo.com

Graphical Abstract
Highlights




Niobium-aluminum-based mixed oxide highly active in the acetalization of
acetone with glycerol and selective towards solketal.
Set of catalysts synthesized by a sol-gel process using high-throughput
techniques to optimize the catalytic features of the solids.
Nb-O-Al bonds are crucial to the persistence of niobic acid-like structure of the
NbO species in these compositions even after thermal treatment at 823 K.
x
Hydrophilicity of the catalysts surface plays an important role in the catalytic
performance of the materials, beside the acid sites.
ABSTRACT
Niobium-aluminum-based catalysts were synthesized by a sol-gel process and
successfully applied to the reaction of acetalization of acetone with glycerol yielding
2
,2-dimethyl-1,3-dioxolane-4-methanol (solketal) and 2,2-dimethyl-1,3-dioxan-5-ol.

The synthesis procedure was developed using high-throughput techniques and the
materials prepared with molar ratio1 Metal (1Nb:xAl) : 100H O : 1.5NH OH (in which
2
4
x = 1, 0.6, 0.3, 0.1 and 0.05) were selected to be further investigated. The obtained
series of mixed oxides displayed high catalytic activity reaching glycerol conversion up
to 84%, with 98% of selectivity towards solketal. The physicochemical properties of the
investigated catalysts were characterized by N -adorption/desorption, thermal analysis,
2
X-rays diffraction, FTIR and Raman spectroscopy, and by adsorption of pyridine
monitored by FT-IR spectroscopy and TGA. The catalysts are truly heterogeneous and
can be reused in consecutive runs without loss of activity.
Keywords: Niobium; aluminum; acetalization; solketal; glycerol.
1
. Introduction
The non-renewable nature of fossil fuel reserves combined with the increasing
demand for energy has prompted humanity to look for alternative and more sustainable
types of fuels. In recent years, biofuels have been developed as a promising substitute
for the petroleum-based diesel [1]. However, the established biodiesel production
process generates ca. 10 % of glycerol as by-product, which would represent an
economic drawback to the process viability unless a valuable application is found for
the crude glycerol [2]. In this context, the conversion of glycerol into added-value
products has been widely studied and several viable synthetic routes have been
considered, such as hydrogenolysis [3], oxidation [4], carbonatation [5], etherification
[
6], esterification [7] and acetalization with either aldehydes or ketones [8-11]. The
acetalization of acetone with glycerol can produce two branched oxygen-containing
compounds. The main product is generally solketal (2,2-dimethyl-1,3-dioxolane-4-
methanol), which has potential to be directly applied as solvent, low temperature heat-
transfer fluid, surfactant and fuel additive [12-13]. The reaction can also yield the six-
membered-ring product (6MR), 2,2-dimethyl-1,3-dioxan-5-ol (Scheme 1) [14].
Scheme 1:Acetalization of acetone with glycerol.

Conventionally, the condensation of glycerol with either aldehydes or ketones is
carried out in a homogeneous Brønsted acid catalytic system [15]. However, recently
several approaches have been reported detailing the production of acetals and/or ketals
via acetalization reactions by heterogeneous catalytic processes. These studies were
stimulated by the economic and environmental advantages inherent to heterogeneous
catalysis. Solid acids such as Amberlyst-type resins [16], zeolites [9, 15, 17],
mesoporous metal-substituted silicates [10, 18], mixed oxides [19], niobic acid [20],
sulfated zirconia [21] and functionalized activated carbons [22, 23] have been
successfully applied as catalysts for the acetalization of acetone with glycerol.
Recently, niobium pentoxide (Nb
2
5
O ) has gained much attention due to its
remarkable properties, both as support and as catalyst. Niobium-oxide-based materials
own a set of attractive catalytic features including redox properties and acidity, and they
have been pointed out as active catalysts in several acid-catalyzed reactions [24]. The
catalytic properties of Nb
related to its structure. Hydrated niobium oxide compounds predominantly contain
octahedrally-coordinated NbO units with different extents of distortion due to corner-
or edge-shared NbO polyhedra and NbO tetrahedra [25, 26]. The incidence of the
highly polarized Nb=O double bond in NbO , contributes to the generation of Lewis
acidity, whereas the slightly distorted NbO , NbO and NbO are characterized by Nb-
2
O
5
.nH O, i.e. niobic acid, including acidity, are directly
2
6
6
4
4
6
7
8
O-Nb single bonds and, due to the highly polarized Nb-O bonds in these distorted
polyhedral, some of the surface OH groups function as Brønsted acid sites [27, 28]. It is
well documented that full transition of amorphous hydrated Nb
2
5
O to a crystalline form,
which typically takes place upon thermal treatment at high temperatures (above 973 K),
is followed by a decrease of its acidic properties [27]. In this direction, Jenhg and
Wachs [29] first detailed the role of niobia supported over several high surface area
metal oxides (MgO, TiO
support nature over the activity of the solids. Lewis acid sites were detected in all
supported niobia, whereas Brønsted acid sites were observed only in Nb /Al and
Nb /SiO . More specifically, the bridge Nb-OH-Al species formed through the
interaction of the NbO species with the hydroxyl groups present on the surface of
2
, ZrO
2
, SiO
2
and Al
2
O
3
) and pointed out the effect of the
2
O
5
2
O
3
2
O
5
2
x
alumina have been identified as Brønsted acid centers. In these materials, the loading of
niobia is limited to the monolayer coverage [29, 30]. Based on these specific properties,
Nb
2
O
5
/Al
2
3
O has been widely applied in several acid-catalyzed reactions such as

Friedel-Crafts alkylation [31], dealkylation of cumene [32] and methanol oxidation
29].
[
The possibility of extending the synergy between Nb and Al throughout the
framework of the mixed oxide and its contribution to the catalytic activity are still
poorly explored. In this work, a series of niobium and aluminum mixed oxides were
synthesized by a sol-gel process in order to investigate the aforementioned synergy and
relate it to the catalytic performance in the acetalization of acetone with glycerol.
Several synthetic parameters were evaluated, such as hydrolysis rate, pH and Nb/Al
molar ratio, aiming at understanding their influence on the catalytic activity of the
obtained materials in the studied reaction.
2
Experimental
2
.1 High-throughput platform
The initial screening of the synthesis of the catalysts was performed in a high-
throughput platform, equipped with a heating and stirring unit containing multiple
parallel wells that can host 10 mL glass vials [33]. The synthesis mixtures were
prepared by dispensing the desired amounts of the solutions of aluminum sec-butoxide
in sec-butanol and niobium chloride in ethanol, which were previously prepared under
2
N atmosphere in a concentration of 0.46M and 0.30M, respectively.
2
.2 Synthesis of the catalysts
The standard sol-gel synthesis procedure consisted in adding a desired amount of
a solution of each precursor into a 10 ml vial, followed by stirring of the mixture for 3
h. Afterwards, both water and a base (25% NH OH solution from Nuclear) or an acid
65% HNO solution from Vetec) were added dropwise into the synthesis vial and the
4
(
3
system was kept under stirring for further 60 min. The vial was closed and the sample
aged in an oven for 24 h at 353 K. Each composition was centrifuged in order to remove
the liquid phase. The obtained solids were dried in an oven at 353 K for 48 h (to achieve
3
-1
complete removal of the solvents) and calcined at 823 K under 100 cm min flow of
-1
3
-1
air for 5 h (heating rate of 3 K min under 100 cm min flow of N
amounts that were employed are: water/metal = 100, 50, 25 and 10, NH
.5, 0.7 and 0.3, HNO /metal = 2, 1, 0.5 and 0 and Al/Nb= 10, 5, 1, 0.6, 0.3, 0.1, 0.05.
2
). The relative
4
OH/metal =
1
3

2
.3 Characterization
The textural properties of the solids were determined by
N
2
adsorption/desorption measurements on an Autosorb 1-MP device (Quantachrome
Instruments). The samples were pre-treated at 473 K under vacuum for 3 h and the
analysis was carried out at 77 K. The surface area, pore (void) volume and pore (void)
size distribution were calculated by means of the Brunauer–Emmett–Teller (BET)
equation [34] and of the density functional theory (DFT) [35]. Powder X-ray diffraction
patterns (XRD) were measured on a Bruker D8 Advance diffractometer using a Cu Kα
radiation (1.542 Å) with a high resolution Lynxeye detector. Each sample was scanned
in a 2θ range of 20-60° with step size of 0.02°/s. Fourier-transform infrared
spectroscopy (FTIR) analysis using a Varian 3100 FT-IR spectrometer were performed
mixing the dried sample with potassium bromide (KBr) in a 1:50 weight ratio. This
mixture was ground into a fine powder, dried at 373 K for 24 h, and used to produce
thin, flat pellets by means of a manual equipment. The spectra were acquired at room
−
1
temperature by accumulating 128 scans at 4cm resolution in the range of 400−3500
−
1
cm . Raman spectra were recorded in a Raman Renishaw microscope (Model InVia)
-1
with exciting radiation at 830 nm (laser diode) in a range of 200-1200 cm with
suppression of spikes caused by cosmic rays. Before measuring each spectrum, the
sample was compacted in a glass lamina, in order to have a smooth surface, and focused
in an optical microscope. In order to estimate hydrophilicity degree of the catalysts,
thermogravimetric analysis (TGA) was performed using a Q500 TGA device (TA
3
-1
Instruments) under N atmosphere (gas flow of 50 cm min ) at a heating rate of 10 K
2
-1
min in the range of 300–1000 K. Prior to TGA analysis, the samples were equilibrated
in the open ambient environment for 48 h at 298 K (humidity of 30-40%).
The number of acid sites was determined based on the methodology proposed by
Ghesti et al.[36]. The samples (~25 mg) were placed in a glass reactor and heated up to
4
73 K under vacuum for 90 min, in order to remove physisorbed molecules, cooled
3
-1
down to 373 K and subjected to a flow of pyridine in N (100 cm min ) for 20 min. The
2
saturated samples were heated to 423 K and the temperature was held for 60 min under
reduced pressure, in order to remove the excess of probe molecule. The same procedure
was followed, skipping the pyridine adsorption step, in order to obtain a blank. Both
saturated and non-saturated samples were subjected to a thermogravimetric analysis in a
Q500 TGA device from TA Instruments. The measurements were carried out under N
2

3
-1
-1
(
50 cm min ) at 10 K min from 298 K to 1073 K and the number of acid sites was
calculated (Eq. 1).
{
[푀푝ꢀ/푀푝ꢀ423]− ꢁ푀1073/푀423ꢂ}
푛푃푦
=
(1)
푀푀푝ꢀ
where:
M
py: Mass of desorbed pyridine = Mpy423 - Mpy1073
py423: Mass of saturated sample at 423 K
M
M1073: Mass of physisorbed molecules of the non-saturated sample (blank) = M423 -
M
1073
M423: Mass of non-saturated sample (blank) at 423 K
MMpy = molecular mass of pyridine, 79.101 g mol-1
Furthermore, the population of Lewis and Brønsted acid sites of the materials
was estimated by Fourier-transform infrared spectroscopy analysis of adsorbed pyridine
at 423 K. The samples were dried at 473 K and the reference spectra were recorded.
Adsorption of pyridine on the previously dried sample took place in a glass reactor at
3
-1
3
73 K under a constant flow of a mixture of pyridine and N (100 cm min ) for 20 min
2
and then submitted to a thermal treatment at 423 K for 60 min. Afterwards,
approximately 15 mg of each sample was placed in a self-supporting disk of 7 mm
diameter and the infrared spectra were obtained using a Varian 3100 FT-IR
-1
spectrometer (512 scans and 4cm resolution). The number of acid sites was determined
-1
-1
from the integral of the bands at 1445 cm (Lewis acid sites) and at 1540 cm
Brønsted acid sites), using the molar extinction coefficients previously reported by
Tamura et al. [37].
(
2
.4 Catalytic tests
The acetalization of acetone with glycerol was carried out in 10 ml capped glass
vials under vigorous stirring (600rpm) at different temperatures (room temperature, 323
K or 353 K). In a typical catalytic test, 0.921 g (0.01 mol) of highly purified glycerol
(99%), a selected amount of acetone [i.e. 0.581 g (0.01 mol), 1.162 g (0.02 mol) or

2
.324 g (0.04 mol)] and 0.132 g (0.0015 mol) of 1,4-dioxane, used as the GC internal
standard, were weighed in a 10 mL glass vial containing 25 mg of the catalyst. The
mixture was stirred for 6 h at the selected temperature. The reaction products were
analyzed by gas chromatography with flame ionization detector (GC-FID) using an
Agilent 7890A device, equipped with a DB-WAX capillary column.
3
. Results and discussion
In order to establish the best synthetic conditions adopted to prepare a set of
niobium-aluminum mixed oxides, several parameter of the sol-gel process were
evaluated using a high-throughput platform. Both catalytic activity and reproducibility
indicated that the best conditions to synthesize a niobium-aluminum mixed oxide with
promising performance in the reaction of acetalization of acetone with glycerol imply
the use of a highly basic medium with an excess of water. The catalysts containing high
amount of aluminum oxide (Al/Nb molar ratio of 5 and 10) as well as all the solids
synthesized in acid medium were inactive (glycerol conversion lower than 5%). Based
on this preliminary high-throughput screening, a synthesis protocol was selected to
synthesize the catalysts to be further investigated on the acetalization of acetone with
glycerol. This protocol is characterized by the following molar ratio: 1Metal (1Nb:xAl) :
1
00H
2
O : 1.5NH OH (in which x = 1, 0.6, 0.3, 0.1 and 0.05).
4
3
.1 Characterization of the catalysts
The XRD patterns of the Nb
2
O
5
-yAl
2
3
O catalysts calcined at 823 K are presented
o
o
in Figure 1. The diffraction peaks of pure Nb
2
O , observed at 2θ= 22.39 , 28.40 ,
5
o
3
6.72 , 46.10º, 50.68º e 55.23º,can be indexed to the orthorhombic (T) phase of niobium
pentoxide [27, 38].On the other hand, the amorphous state is prevailing in both 1Nb:1Al
and 1Nb:0.6Al, since the presence of bulk Nb is negligible for these
compositions(and also in niobic acid, i.e. the hydrated oxide Nb .nH O). Wachs et
al.[39] claimed that the dispersion of niobia on γ-alumina takes place through the
interactions of NbO with the hydroxyl groups present on γ-Al surface up to the
monolayer coverage (about 19% (w/w)). In this work, niobium oxide persisted as an
amorphous phase up to remarkable higher load of NbO in the material. This suggests
the extensive presence of Nb-O-Al bonds within the oxides prepared with 1Nb:1Al and
Nb:0.6Al. Low intensity reflection peaks of the T-Nb crystalline phase are also
observed in 1Nb:0.3Al, 1Nb:0.1 and 1Nb:0.05Al and the presence of this crystalline
2
O
5
2
O
5
2
x
2
O
3
x
1
2
O
5

phase is due to the high load of Nb in the material, but the aforementioned Nb-O-Al
bonds are expected to be still present in significant amounts, since the full transition to
the orthorhombic phase was not observed in any of the niobium-aluminum mixed
oxides. In the case of 1Nb:0.1Al and 1Nb:0.05Al, diffraction peaks referred to
crystalline AlNbO were also observed (2θ=23.4º, 24.8º, 31.7º, 43.8º and 47.1º) [31].
4
Figure 1: XRD patterns of Niobic acid, 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al,
Nb:0.05Al and Nb
1
2
5
O .
FTIR spectra of the niobium containing materials are presented in Figure 2. The
-1
broad band in the range of 550-630 cm as well as the shoulder-like band at around 900
cm-1 have been assigned to Nb-O and Nb=O stretching modes, respectively. The latter
signal is more defined in 1Nb:0.3Al, 1Nb:0.1Al, 1Nb:0.05Al and Nb
2
5
O , and is
associated to the remarkable concentration of highly distorted niobium oxide species
containing Nb=O bonds in the crystalline phase [30, 40], in agreement with the XRD
-1
data presented in Figure 1. Accordingly, the band at around 900 cm has low intensity
in the spectra of the amorphous1Nb:1Al, 1Nb:0.6Al and niobic acid samples.
-1
-1
Moreover, the vibration bands at around ca. 3420 cm and 1610 cm are
associated to the OH stratching and bending modes of physisorbed water, respectively,
and are more pronounced in the catalysts with higher hydrophilicity [26].
Analysis by Raman spectroscopy showed characteristic bands at around 660
-1
-1
-1
cm and at 990 cm (Figures 3a and 3b). The signal at 660 cm has been assigned to
the symmetric stretching mode of the slightly distorted NbO , NbO and NbO
polyhedra, which are present in both amorphous and crystalline niobia. The band at 990
6
7
8
-1
cm has been assigned to the symmetric stretching mode of Nb=O, which is more
pronounced in crystalline Nb [25, 29]. The Raman spectra of 1Nb:1Al and
Nb:0.6Al are remarkably similar to the Raman profile observed for the amorphous
niobic acid, Nb .nH O (Figure 3a), indicating the persistence of a niobic acid-like
structure of the NbO species in these compositions even after thermal treatment at 823
K, in full agreement with FTIR and XRD data. In the case of 1Nb:0.3Al, the incipient
presence of theT-Nb crystalline phase can be recognized by the slightly pronounced
bands present in the low wave number region (200-350 cm ), which are characteristic
of the bending modes of the Nb-O-Nb bonds, typically observed in T-Nb [25]
Figure 3b). However, the characteristic band at 990 cm was not detected, likely due to
2
O
5
1
2
O
5
2
x
2
O
5
-1
2
O
5
-1
(

the low concentration of Nb=O sites in this composition. For 1Nb:0.1Al and
Nb:0.05Al, the presence of the T-Nb phase becomes clearer, as evidenced by the
Raman bands at around 260, 660 and 990 cm . Nevertheless, the presence of AlNbO
1
2
O
5
-1
4
(as suggested by XRD) cannot be disregarded since these compounds possess a highly
distorted octahedral NbO
x
structure with Raman bands in the high wave number region
-1
-1
(
850-1000 cm ) and the bands in the region of 200-300 cm assigned to the associated
bending mode might be also related to this crystalline phase [25, 31].
Figure 2: FTIR spectra of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al, 1Nb:0.05Al,
Niobic acid; Nb
2
5
O .
Figure 3: Raman spectra of (a) 1Nb:1Al, 1Nb:0.6Al and Niobic acid; (b) 1Nb:0.3Al,
Nb:0.1Al, 1Nb:0.05Al and Nb
1
2
5
O .
The acid properties of the catalysts were studied by thermal desorption of
pyridine [36], which was monitored both by FT-IR spectroscopy and by TGA. The
results with the two techniques show the same trends (Table 1). These data indicate a
decreasing number of acid sites with increasing content of Nb. The bands at 1445 cm-1
-
1
and 1540 cm in the FT-IR spectra of chemisorbed pyridine over γ-Al
at 423 K were used to quantify Lewis and Brønsted acid sites, respectively (Figure 4).
As expected, under the analysis conditions, pure γ-Al presented only Lewis acid
sites (0.42 mol kg ), which is in agreement with previous reports [28]. In the case of the
Nb:xAl oxides, most of the acid sites present on the surface are of Lewis type.
However, Brønsted acid sites were also detected in1Nb:1Al, 1Nb:0.6Al and 1Nb:0.3Al
Table 1). It is expected that the generation of weak Brønsted acid sites is related to the
bridged Nb-OH-Al but the contribution of Nb-OH-Nb species should be also relevant
30, 41]. On the other hand, Lewis acidity could stem both from the polymeric
amorphous NbO in niobic acid-like form, as from Al domains. It should be noted
2
O
3
and 1Nb:xAl
2
O
3
-1
1
(
[
x
2
O
3
that the ratio between Lewis and Brønsted acid sites was not strongly affected by the
Nb:Al ratio (for x = 1, 0.6 and 0.3, see Table 1). For the materials with a lower Al
content (1Nb:0.1Al and 1Nb:0.05Al), the low total concentration of acid sites prevented
the quantification of the Lewis and Brønsted acid species.
Table 1: Textural properties and acidity of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al
and 1Nb:0.05Al

Figure 4: Infrared spectrum of chemisorbed pyridine over γ-Al
2
3
O , 1Nb:1Al, 1Nb:0.6Al
and 1Nb:0.3Al at 423 K
Both surface area and pore volume distribution were calculated from the
nitrogen physisorption isotherm obtained by N adsoption/desorption analysis. Based on
2
the isotherms presented in Figure 5a, it is noteworthy that the samples containing higher
amount of aluminum oxide (1Nb:1Al and 1Nb:0.6Al) display a Type IV isotherm, with
an H3 hysteresis loop, which is observed with aggregates of plate-like particles giving
rise to slit-shaped mesopores or interparticle voids in the mesoscale [42].Since no
structure directing agent was used in the synthesis of the materials, the observed
porosity is attributed to interparticle voids. On the other hand, the N
2
isotherms of
1
Nb:0.3Al, 1Nb:0.1Al and 1Nb:0.05Al are classified as Type II, which is expected for
non-porous materials, and a slightly pronounced H3 hysteresis loop can be observed in
the isotherms [42] (Figure 5b).The pore size distribution of 1Nb:1Al and 1Nb:0.6Al,
(Figure 6a), is mainly in the mesoscale region below 20 nm whereas that of the
materials with lower Al content is very broad and with low dV/dD values (Figure 6b).
The surface area and the pore volume decrease significantly with increasing Nb content
in the catalysts (Table 1).
Figure 5: N
2
adsorption/desorption isotherms of (a) 1Nb:1Al, 1Nb:0.6Al and (b)
1
Nb:0.3Al , 1Nb:0.1Al, 1Nb:0.05Al .
Figure 6: Pore distribution of (a) 1Nb:1Al, 1Nb:0.6Al and (b) 1Nb:0.3Al , 1Nb:0.1Al,
Nb:0.05Al .
1
The calcination step in the synthesis of the catalysts was thoroughly investigated
by thermogravimetric analysis (TG-DTA, see Figures 7 and 8). Thermal events
observed at temperatures below 393 K are ascribed to the loss of moisture and of
remaining solvents used during the sol-gel synthesis procedures. The next events have
an endothermic nature and can be assigned to the loss of hydroxyl groups present on the
surface of the mixed oxides and on bulk Nb
oxi-hydroxy NbO (OH) species [43]. The exothermic event at approximately 830 K in
Nb:0.3Al, 1Nb:0.1Al and 1Nb:0.05Al is related to the first crystalline transition of the
niobium oxide, and the weight loss observed at this temperature might be associated
2
5
O as a consequence of the condensation of
2
1

with a change in the composition. Accordingly, no thermal event at about 830 K was
observed for the amorphous materials (1Nb:1Al and 1Nb:0.6Al).
Figure 7: Thermogravimetric analysis of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al , 1Nb:0.1Al
and 1Nb:0.05Al.
Figure 8: Differential thermal analysis of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al
and 1Nb:0.05Al.
3
.2 Catalytic tests
The conversion of glycerol to solketal via acetalization of acetone takes place
over acid sites [44], but in the case of heterogeneous catalysts the textural and surface
properties also play an important role [10]. Therefore, the synthesized Nb-Al-based acid
catalysts herein presented were investigated in the acetalization of acetone with
glycerol, and the role of the structural properties and acidity in the catalytic performance
of these mixed oxides was evaluated. First, the relationship between the glycerol
conversion and the molar ratio of glycerol to acetone was studied (Figure 9a). The
experiments were carried out at 323 K for 6 h. As expected, the conversion of glycerol
remarkably increases with the relative amount of acetone in the reaction medium
[
10,45], reaching 84% of substrate conversion with acetone:glycerol = 4:1 over
Nb:0.05Al. Fan et al.[45] applied successfully a Ti-Si-based catalyst in the
acetalization of acetone with glycerol reaching up to 86% of glycerol conversion with
0% of selectivity to solketal. Nandan et al. [46] explored the viability of a series of
1
9
sulfonated carbon–silica-meso composite materials as catalyst in the conversion of
glycerol to solketal via acetalization with acetone and achieved glycerol conversion as
high as 82% with selectivity to the main product of 99%. Based on the typical
acetalization reaction mechanism of ketal formation over Brønsted acid sites [9], it is
expected that the excess acetone shifts the equilibrium towards the products and
facilitates its chemisorption over the active centers of the catalysts, which is the first
step of the catalytic cycle, justifying the rise in glycerol conversion levels. Furthermore,
the herein synthesized mixed oxides own a remarkable population of surface Lewis acid
sites, which may contribute in catalyzing the acetalization of acetone with glycerol [10].
However, it should be kept in mind that these Lewis acid sites may convert into
Brønsted ones by reacting with the water formed during the acetalization reaction [47].

The presence of Lewis acid sites at the surface of the Nb-Al mixed oxides can also play
a role in coordinating glycerol, similarly to what has been reported on niobia [47].
The formation of the five-membered-ring of solketal involves the formation of a
bond between the carbonyl oxygen atom of acetone and the β-carbon of the glycerol,
accompanied by dehydration [10]. The less common six-membered-ring product
requires the formation of a bond between the carbonyl oxygen of acetone and the
primary carbon of glycerol. In agreement with previous reports [10], solketal was the
major product over the Nb-Al mixed oxides (Figure 9b). For comparison, pure alumina
and niobia were evaluated on the studied reaction at following conditions:
glycerol:acetone molar ratio of 1:4, 2.7 wt% of catalyst and 323K, and the former
catalyst showed no activity, whereas T-Nb
is in agreement with previous reports [38].The low activity of the crystalline T-Nb
is connected to the previously reported decrease of acidic properties upon transition
from amorphous niobia to the orthorhombic T-Nb phase [38].
2
O
5
gave 58% of glycerol conversion, which
2
O
5
2
O
5
Figure 9: Effect of the variation of the glycerol to acetone ratio on the acetalization
reaction (2.7 wt% of catalyst, 6 h at 323 K): (a) glycerol conversion and (b) solketal
selectivity.
Experiments at different temperatures (room temperature, 323 K and 353 K) were
also carried out (Table 2). Surprisingly, the catalysts with lower surface area and lower
number of surface acid sites (1Nb:0.3Al, 1Nb:0.1Al and 1Nb:0.05Al) were substantially
active in the reaction conducted at room temperature achieving up to 73% of glycerol
conversion, while 1Nb:1Al and 1Nb:0.6Al displayed markedly worse performance
under these conditions. The differences between the activity of the various catalysts was
much less marked when the temperature of the process was raised to 323 K and 353 K.
The best performance of 1Nb:0.3Al, 1Nb:0.1Al and 1Nb:0.05Al was reached at 323 K,
whereas slightly lower conversion was observed at 353 K (Table 2). Such trend has
been observed before and is attributed to the exothermic nature of the acetalization
reaction, which implies lower products concentration at equilibrium upon increase of
the reaction temperature [23, 48]. On the other hand, the catalytic performance of
1
Nb:1Al and 1Nb:0.6Al is dependent on the temperature because at lower temperature
the reaction rate over these catalysts is not sufficient to reach the equilibrium
concentrations. This might seem counterintuitive, because 1Nb:1Al and 1Nb:0.6Al
display the two largest populations of acid sites among the studied mixed oxides (Table

1
). To understand this catalytic trend, the hydrophilicity of the materials should be
considered [10]. A high hydrophilicity would prevent removal of the water formed
during the acetalization from the active sites, thus favoring the reverse reaction in which
the formed solketal is hydrolyzed back to glycerol and acetone. To evaluate the
hydrophilicity of the catalysts, thermalgravimetric analysis (TGA) was performed
(Figure 10). The weight loss observed at temperatures below 523K is assigned to the
desorption of water from the surface of the catalysts, which is related to the
hydrophilicity of the materials. This event is much more marked in the
themogravimetric analysis of 1Nb:1Al and 1Nb:0.6Al, probably due to the high
aluminum content in these materials. Combining the TGA data and the observed
catalytic trends, we propose that the hydrophilicity of 1Nb:1Al and 1Nb:0.6Al plays an
important role in limiting the catalytic performance of these materials.
Figure 10: Thermogravimetric analysis of the calcined 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al
,
1Nb:0.1Al and 1Nb:0.05Al.
These results drive us to conclude that not only the presence of acid sites is
required for catalyzing the acetalization reaction, but also a sufficient hydrophobicity of
the catalyst surface, as that in 1Nb:0.3Al, 1Nb:0.1Al and 1Nb:0.05Al, is necessary to
achieve good catalytic activity.
Table 2: Acetalization of acetone with glycerol over 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al,
1
Nb:0.1Al and 1Nb:0.05 Al at room temperature, 323 and 353 K (glycerol : acetone
molar ratio of 1:4, 2.7 wt% of catalyst, 6 h)
The activity of the catalysts as a function of time was investigated by performing
a kinetic study of the acetalization of acetone with glycerol at 323 K with
acetone:glycerol molar ratio of 4:1 (Figure11). The reaction catalyzed by 1Nb:0.05Al
reaches equilibrium conditions after 120 min (glycerol conversion of 84%), whereas the
plateau is reached at longer reaction times for the other catalysts. The difference in
catalytic activity between the materials are more marked when the initial rate is
considered, as can be seen by comparing the glycerol conversion in the first 20 min of
reaction (Figure 11).
Figure 11: Kinetic study (solid lines) and leaching test (dashed lines) of the acetalization
of acetone with glycerol catalyzed by 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al and
1
Nb:0.05Al (glycerol:acetone molar ratio of 1:4, 2.7 wt% of catalyst, 323K).

For all the catalysts tested at 323K, the selectivity to solketal was extremely high
at equilibrium conversions, but a significant amount of six-membered-ring (6MR) was
observed at lower conversions, and thus particularly in the reactions catalyzed by
1
Nb:1Al and 1Nb:0.6Al (Figure 12). It is noteworthy that selectivity to 6MR decreases
during the reaction while the selectivity to solketal increases. The energies of the five-
and six-membered-ring isomers were calculated by Mota et al. [49] and it was found
-1
that solketal is 1.7 kcal mol more stable than the six-membered-ring demonstrating
that solketal is the thermodynamically favored product of the acetalization of acetone
with glycerol. Therefore, it is expected the 6MR isomerizes to solketal throughout the
reaction, explaining the observed trend for solketal selectivity.
Figure 12: Solketal and 6MR selectivity in function of time of the acetalization of
glycerol with acetone catalyzed by 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al and
1
Nb:0.05Al (glycerol:acetone molar ratio of 1:4, 2.7 wt% of catalyst, 323K).
In order to confirm the truly heterogeneous nature of the catalysts, leaching tests
were carried out by removing the catalysts from the reaction medium after about 20 min
allowing the liquid mixture to react in the absence of the heterogeneous catalyst for
further 240 min (dashed line in Figure11). In all evaluated systems, there was no
appreciable increase in the glycerol conversion upon removing the catalyst, which
indicates the lack of leaching of the active sites from the solid catalysts, as expected for
a typical heterogeneous catalytic process.
Furthermore, the stability of the catalytic activity of 1Nb:1Al, 1Nb:0.3Al and
1
Nb:0.1Al was evaluated in consecutive runs by separating the heterogeneous catalysts
from the reaction medium after 6 h reaction time, washing the material 3 times with a
polar aprotic solvent (acetonitrile) and drying the catalysts at 353 K for 24 h. In all
cases, the catalytic activity remains approximately constant in the four catalytic runs
(Figure 13), which indicates that acetonitrile solubilize remaining reactants and products
adsorbed over the active sites avoiding any thermal treatment at high temperature to
remove those species. It should be noted that a less polar aprotic solvents as acetone or a
protic polar solvent as methanol were much less efficient in removing the reaction
residues.
Figure 13: Recycling tests of the 1Nb:1Al at 353 K, 1Nb:0.3Al and 1Nb:0.1Al at 323K
in the acetalization of glycerol with acetone (glycerol:acetone molar ratio of 1 : 4, 2.7
wt% of catalyst, 6 h).

4
. Conclusions
New niobium-aluminum mixed oxides synthesized by a sol-gel process
presented high catalytic activity in the acetalization of acetone with glycerol with high
selectivity towards the desired product (solketal). All the prepared mixed oxides of Nb
and Al display acid sites that can act as active sites for catalyzing the acetalization
reaction. However, the trends in activity cannot be explained only based on the presence
of acid sites and require taking into account the hydrophilicity of the catalysts surface,
with the more hydrophobic materials giving the best catalytic performance. The
catalysts are truly heterogeneous as no leaching of active species was observed and
retained their activity in several consecutive runs.
5
. Acknowledgements
This work was supported by FAPESP (Project No. 2013/21160-6), CNPq-SWB (Project
No. 245705/2012-0), CNPq-FWO (CNPq-Project No.490298/2009-5, FWO-Project
code: VS.025.10N, V4.520.11N) and START1 (KU Leuven, Project code:
STRT1/10/035).
6
. References
[
1] M.S.P. Gomes, M.S. Muylaert de Araújo, Renew. Sust. Energ. Rev., 13 (2009)
201-2204.
2
[
[
2] A. Beatriz, Y.J.K. Araujo, D.P. de Lima, Quim. Nova, 34 (2011) 306-319.
3] Z.W. Huang, F. Cui, J.J. Xue, J.L. Zuo, J. Chen, C.G. Xia, Catal. Today, 183 (2012)
4
2-51.
[
4] G.L. Brett, Q. He, C. Hammond, P.J. Miedziak, N. Dimitratos, M. Sankar, A.A.
Herzing, M. Conte, J.A. Lopez-Sanchez, C.J. Kiely, D.W. Knight, S.H. Taylor, G.J.
Hutchings, Angew. Chem. Int. Ed. Engl., 123 (2011) 10318-10321.
[
5] N. Ezhova, I. Korosteleva, N. Kolesnichenko, A. Kuz’min, S. Khadzhiev, M.
Vasil’eva, Z. Voronina, Petroleum Chem., 52 (2012) 91-96.
6] P.A. Celdeira, M. Gonçalves, F.C.A. Figueiredo, S.M.D. Bosco, D. Mandelli, W.A.
Carvalho, Appl. Catal. A, 478 (2014) 98-106.
[
[
[
7] S. Zhu, Y. Zhu, X. Gao, T. Mo, Y. Zhu, Y. Li, Biores.Technol., 130 (2013) 45-51.
8] A. Behr, J. Eilting, K. Irawadi, J. Leschinski, F. Lindner, Green Chem., 10 (2008)
1
3-30.

[
[
9] C.X.A. da Silva, V.L.C. Goncalves, C.J.A. Mota, Green Chem., 11 (2009) 38-41.
10] L. Li, T.I. Koranyi, B.F. Sels, P.P. Pescarmona, Green Chem., 14 (2012) 1611-
1
619.
11] B. Mallesham, P. Sudarsanam, G. Raju, B.M. Reddy, Green Chem., 15 (2013) 478-
89.
[
4
[
[
[
12] M. J. Climent, A. Corma and A. Velty, Appl. Catal., A, 263 (2004) 155–161.
13] M. J. Climent, A. Velty and A. Corma, Green Chem., 4 (2002) 565–569.
14] R. Rodrigues, M. Gonçalves, D. Mandelli, P.P. Pescarmona, W.A. Carvalho,
Catal.Sci. Technol., 4 (2014) 2293-2301.
[
[
(
[
(
[
[
[
15] J. Deutsch, A. Martin, H. Lieske, J. Catal., 245 (2007) 428-435.
16] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.-A. Poirier, C. Xu, Fuel, 128
2014) 113-119.
17] P. Manjunathan, S.P. Maradur, A.B. Halgeri, G.V. Shanbhag, J. Mol.Catal. A, 396
2015) 47-54.
18]L. Li, D. Cani, P.P. Pescarmona, Inorg. Chim. Acta, 4319 (2015) 289-296.
19] S. Zhang, Z. Zhao, Y. Ao, Appl. Catal. A, 496 (2015) 32-39.
20]T.E. Souza, M.F. Portilho, P.M.T.G. Souza, P.P. Souza, L.C.A. Oliveira,
ChemCatChem, 6 (2014) 2961-2969.
21] P.S. Reddy, P. Sudarsanam, B. Mallesham, G. Raju, B.M. Reddy, J. Ind. Eng.
Chem., 17 (2011) 377-381.
[
[
[
22] M.S. Khayoon, B.H. Hameed, Appl. Catal. A, 464-465 (2013) 191-199.
23] R. Rodrigues, M. Goncalves, D. Mandelli, P.P. Pescarmona, W.A. Carvalho,
Catal.Sci. Technol., 4 (2014) 2293-2301.
[
[
[
24] K. Tanabe, S. Okazaki, Appl. Catal. A, 133 (1995) 191-218.
25] J.M. Jehng, I.E. Wachs, Chem. Mater., 3 (1991) 100-107.
26] K. Nakajima, Y. Baba, R. Noma, M. Kitano, J. N. Kondo, S. Hayashi, M. Hara, J.
Am. Chem. Soc., 133 (2011)4224-4227
[
[
27] I. Nowak, M. Ziolek, Chem. Rev., 99 (1999) 3603-3624.
28] T. Kitano, T. Shishido, K. Teramura, T. Tanaka, J. Phys. Chem. C., 116 ( 2012)
1
1615-11625.
[
[
[
29] J.-M. Jehng, I.E. Wachs, Catal.Today, 8 (1990) 37-55.
30] J. Datka, A.M. Turek, J.M. Jehng, I.E. Wachs, J. Catal., 135 (1992) 186-199.
31] T. Shishido, T. Kitano, K. Teramura, T. Tanaka, Top. Catal., 53 (2010) 672-677.

[
32] C.L.T. da Silva, V.L.L. Camorim, J.L. Zotin, M.L.R. Duarte Pereira, A.d.C. Faro
Jr, Catal. Today, 57 (2000) 209-217.
[
[
[
[
33] P.P. Pescarmona, P.A. Jacobs, Catal. Today, 137 (2008) 52-60.
34] B.C. Lippens, J.H. de Boer, J. Catal., 4 (1965) 319-323.
35] J. Olivier, W. Conklin, M. Szombathely, Stud.Surf. Sci. Catal., 87 (1994) 81-89.
36] G.F. Ghesti, J.L. de Macedo, V.C. Parente, J.A. Dias, S.C. Dias, Microporous
Mesoporous Mater., 100 (2007) 27-34.
37] M. Tamura, K.I. Shimizu, A. Satsuma, Appl. Catal. A: Gen., 433-434 (2012) 135-
45.
38] G.S. Nair, E. Adrijanto, A. Alsalme, I.V. Kozhevnikov, D.J. Cooke, D.R. Brown,
N.R. Shiju, Catal.Sci. Technol., 2 (2012) 1173-1179.
[
1
[
[
[
39] I.E. Wachs, Catal. Today, 27 (1996) 437-455.
40]F. Houshang, H. Fatemeh, R. Rahmatollah, G. Ali, J. Clust. Sci., 25 (2014) 651-
6
66.
[
[
[
41] T. Kitano, T. Shishido, K. Teramura, T. Tanaka, Catal. Today, 226 (2014) 97-102.
42] K.S.W. Sing, in: Pure Appl.Chem., 1985, pp. 603.
43] T.E. Souza, I.D. Padula, M.M.G. Teodoro, P. Chagas, J.M. Resende, P.P. Souza,
L.C.A. Oliveira, Catal.Today, 213 (2013) 65-72.
44] B. Mallesham, P. Sudarsanam, G. Raju, B.M. Reddy, Green Chem., 15 (2013) 478-
89.
45] C.-N. Fan, C.-H. Xu, C.-Q. Liu, Z.-Y. Huang, J.-Y. Liu, Z.-X. Ye, React. Kinet.
Mech. Cat., 107 (2012) 189-202.
46]D. Nandan, P. Sreenivasulu, L.N. Sivakumar Konathala, M. Kumar, N.
Viswanadham, Microporous and Mesoporous Mater., 179 (2013) 182-190.
[
4
[
[
[
[
(
[
47]G. S. Foo, D. Wei, D. S. Sholl, C. Sievers, ACS Catalysis, 4 (2014) 3180-3192.
48] M.R. Nanda, Z. Yuan, W. Qin, H.S. Ghaziaskar, M.-A. Poirier, C.C. Xu, Fuel, 117
2014) 470-477.
49] L.P. Ozorio, R. Pianzolli, M.B.S. Mota, C.J. Mota, J. Braz.Chem. Soc., 23 (2012)
9
31-937.

FIGURE CAPTIONS
Figure 1: XRD patterns of Niobic acid, 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al,
1
Nb:0.05Al and Nb
Figure 2: FTIR spectra of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al, 1Nb:0.05Al,
Niobic acid; Nb
Figure 3: Raman spectra of (a) 1Nb:1Al, 1Nb:0.6Al and Niobic acid; (b) 1Nb:0.3Al,
Nb:0.1Al, 1Nb:0.05Al and Nb
2
O
5
.
2
O .
5
1
2
5
O .
Figure 4: Infrared spectrum of chemisorbed pyridine over γ-Al
2
3
O , 1Nb:1Al, 1Nb:0.6Al
and 1Nb:0.3Al at 423 K.
Figure5: N
Nb:0.3Al, 1Nb:0.1Al, 1Nb:0.05Al .
Figure 6: Pore distribution of (a) 1Nb:1Al, 1Nb:0.6Al and (b) 1Nb:0.3Al, 1Nb:0.1Al,
Nb:0.05Al.
2
adsorption/desorption isotherms of (a) 1Nb:1Al, 1Nb:0.6Al and (b)
1
1
Figure 7: Thermogravimetric analysis of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al
and 1Nb:0.05Al.
Figure 8:Differential thermalanalysis of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al
and 1Nb:0.05Al.
Figure 9: Effect of the variation of the glycerol to acetone ratio on the acetalization
reaction (2.7 wt% of catalyst, 6 h at 323 K): (a) glycerol conversion and (b) solketal
selectivity.
Figure 10: Thermogravimetric analysis of the calcined 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al,
1
Nb:0.1Al and 1Nb:0.05Al.
Figure 11: Kinetic study (solid lines) and leaching test (dashed lines) of the acetalization
of acetone with glycerol catalyzed by 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al and
1
Nb:0.05Al (glycerol:acetone molar ratio of 1:4, 2.7 wt% of catalyst, 323K).
Figure 12: Solketal and 6MR selectivity in function of time of the acetalization of
glycerol with acetone catalyzed by 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al and
1
Nb:0.05Al (glycerol:acetone molar ratio of 1:4, 2.7 wt% of catalyst, 323K).
Figure 13: Recycling tests of the 1Nb:1Al at 353 K, 1Nb:0.3Al and 1Nb:0.1Al at 323K
in the acetalization of glycerol with acetone (glycerol:acetone molar ratio of 1 : 4, 2.7
wt% of catalyst, 6 h).
Scheme 1: Acetalization of acetone with glycerol.

Fig. 1
Fig. 2

Fig. 3
Fig. 4

Fig. 5
Fig. 6

Fig. 7
Fig. 8

Fig. 9a
Fig. 9b

Fig. 10
Fig. 11

Fig. 12
Fig. 13

Scheme. 1
Table 1: Textural properties and acidity of 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al, 1Nb:0.1Al
and 1Nb:0.05Al.
Total
pore Total acid Lewis acid Brønsted acid
Surface area
sites*
a
sites
(mol kg )
0.43
0.10
0.09
*b
sites
*b
(mol
Catalysts
volume
2
-1
(m g )
3
-1
-1
-1
-1
(
cm g )
(mol kg )
0.65
0.12
0.10
0.089
0.094
kg )
0.031
0.012
0.008
n.d
1
1
1
1
1
Nb:1Al
182
36
14
<10
<10
0.35
0.12
0.07
0.05
0.06
Nb:0.6Al
Nb:0.3Al
Nb:0.1Al
n.d
Nb:0.05Al
n.d
n.d
*
Thermal desorption of pyridine monitored by (a) TGA and (b) FT-IR spectroscopy;
n.d = not detected.

Table 2: Acetalization of acetone with glycerol over 1Nb:1Al, 1Nb:0.6Al, 1Nb:0.3Al,
Nb:0.1Al and 1Nb:0.05 Al at room temperature, 323 and 353 K (glycerol:acetone
molar ratio of 1:4, 2.7 wt% of catalyst, 6 h).
1
Glycerol
conversion (%)
Selectivity to solketal (%)
Selectivity
6MR (%)
to Temperature
(K)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Nb:1Al
36
26
69
66
73
62
52
82
83
84
78
69
75
77
77
80
75
95
93
93
94
90
97
98
98
97
97
97
97
97
20
25
5
Nb:0.6Al
Nb:0.3Al
Nb:0.1Al
Nb:0.05Al
Nb:1Al
298
7
7
6
Nb:0.6Al
Nb:0.3Al
Nb:0.1Al
Nb:0.05Al
Nb:1Al
10
3
3
23
2
2
3
Nb:0.6Al
Nb:0.3Al
Nb:0.1Al
Nb:0.05Al
3
3
3
353
3