Enhanced catalytic activity for fructose conversion on nanostructured niobium oxide after hydrothermal treatment: Effect of morphology and porous structure

Article abstract of DOI:10.1016/j.molcata.2016.01.021

This paper reports the synthesis of a new class of catalysts based on niobium using Filter Cake (FC) as precursor, which is more affordable economically than other niobium sources commonly used. FC is obtained before the necessary purification to obtain niobic acid (Nb₂O₅?nH₂O), making the catalyst synthesis process less costly. Its modification by hydrothermal treatment in the presence of different agents (oxalic acid or hydrogen peroxide) allowed the obtention of nanostructured materials in the form of rods or spheres with larger BET area, crystallinity and acidity than the precursor. Their catalytic potential were evaluated in fructose dehydration reaction in aqueous media, aiming the obtention of 5-hydroxymethylfurfural (5-HMF). The production of furan compounds from sugars has become a process of great interest in recent years because it is related to the search for more sustainable sources of energy, since carbohydrates are the predominant part of biomass. In optimum conditions, it was possible to obtain 22% yield of 5-HMF in aqueous medium and 47% yield of 5-HMF using DMSO as solvent.

Full text of DOI:10.1016/j.molcata.2016.01.021

Accepted Manuscript
Title: Enhanced catalytic activity for fructose conversion on
nanostructured niobium oxide after hydrothermal treatment:
Effect of morphology and porous structure
Author: Nayara T. do Prado Talita E. Souza Alan Rodrigues
T. Machado Patterson P. Souza Robson S. Monteiro Luiz C.A.
Oliveira
PII:
S1381-1169(16)30021-8
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2016.01.021
MOLCAA 9753
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Revised date:
Accepted date:
23-7-2015
18-1-2016
19-1-2016
Please cite this article as: Nayara T.do Prado, Talita E.Souza, Alan Rodrigues
T.Machado, Patterson P.Souza, Robson S.Monteiro, Luiz C.A.Oliveira, Enhanced
catalytic activity for fructose conversion on nanostructured niobium oxide after
hydrothermal treatment: Effect of morphology and porous structure, Journal of
Molecular Catalysis A: Chemical http://dx.doi.org/10.1016/j.molcata.2016.01.021
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.

Enhanced catalytic activity for fructose conversion on
nanostructured niobium oxide after hydrothermal treatment: Effect
of morphology and porous structure
Nayara T. do Pradoa, Talita E. Souzaa, Alan Rodrigues T. Machadoa, Patterson P. Souzab,
Robson S. Monteiroc, Luiz C. A. Oliveiraa*
aDepartamento de Química, ICEx, Universidade Federal de Minas Gerais, Belo Horizonte – MG,
Brazil CEP 31270-901. Tel.: +55 31 34096384; Fax: +55 31 34095700. E-mail address:
luizoliveira@qui.ufmg.br
bCentro Federal de Educação Tecnológica de Minas Gerais, Belo Horizonte - MG, Brazil CEP
30421-169.
cCatalysis Consultoria Ltda., Rio de Janeiro, RJ Brazil CEP 22793-081
Corresponding Author
*luizoliveira@qui.ufmg.br
1

Graphical Abstract
Synopsis: A new class of catalysts based on niobium was hydrothermally synthesized using
Filter Cake (FC) as a precursor. The use of different agents (oxalic acid or hydrogen peroxide)
allowed the obtention of nanostructured materials in the form of rods or spheres with larger BET
area, crystallinity and acidity than the precursor. Their potential was evaluated in fructose
dehydration reaction in aqueous media and in DMSO, obtaining 22% and 47% yield of 5-HMF,
respectively.
Highlights
 A new class of catalysts based on niobium was hydrothermally synthesized using Filter
Cake as a precursor.
 The use of different agents (oxalic acid or hydrogen peroxide) allowed the formation of
nanostructured materials in the form of rods or spheres.
 Their potential was evaluated in fructose dehydration reaction obtaining 41% yield of 5-
HMF.
2

Abstract
This paper reports the synthesis of a new class of catalysts based on niobium using Filter Cake
(FC) as precursor, which is more affordable economically than other niobium sources commonly
used. FC is obtained before the necessary purification to obtain niobic acid (Nb2O5.nH2O),
making the catalyst synthesis process less costly. Its modification by hydrothermal treatment in
the presence of different agents (oxalic acid or hydrogen peroxide) allowed the obtention of
nanostructured materials in the form of rods or spheres with larger BET area, crystallinity and
acidity than the precursor. Their catalytic potential were evaluated in fructose dehydration
reaction in aqueous media, aiming the obtention of 5-hydroxymethylfurfural (5-HMF). The
production of furan compounds from sugars has become a process of great interest in recent
years because it is related to the search for more sustainable sources of energy, since
carbohydrates are the predominant part of biomass. In optimum conditions, it was possible to
obtain 22% yield of 5-HMF in aqueous medium and 47% yield of 5-HMF using DMSO as
solvent.
Keywords: Niobium oxide; fructose; dehydration; 5-hydroxymethylfurfural
1. INTRODUCTION
The use of biomass for obtaining chemical derivatives and fuels has emerged as an
important research area in recent years due to the fact that it is a renewable source of great
abundance.[1], [2]
In particular, the carbohydrate class - predominant compounds in the biomass - shows
considerable potential to provide important intermediates for the chemical industry. However,
the economic feasibility of processes based on the use of biomass critically depends on the
selectivity of the products obtained, since the carbohydrates are highly functionalized molecules
and therefore, can undergo a variety of reactions.[3]
3

Carbohydrates dehydration leads to the formation of furan compounds such as furfural
and 5-hydroxymethylfurfural (5-HMF), which are key intermediates for obtaining several
products of petrochemical interest.[4]
5-HMF is a compound consisting of a furan ring containing an aldehyde group and a
hydroxymethyl group at positions 2 and 5. It can be obtained through three dehydration reactions
of carbohydrates C-6 and C-5 derived from biomass and might act as a green chemical
intermediate for obtention of a wide range of compounds current from petroleum.[4]
Rehydration of 5-HMF enables the obtention of levulinic acid and formic acid, which are
commodity chemicals. Furthermore, the levulinic acid is the precursor of liquid fuel γ-
valerolactone. The oxidation of the hydroxymethyl group allows the obtention of 2,5-
diformylfuran and 2,5-furandicarboxylic acid (FDCA). FDCA is particularly interesting as a
polymeric replacement for terephthalic and isophthalic acids which are used to obtain
polyamides, polyesters and polyurethanes.[5]
5-HMF may also suffer undesirable degradation and polymerization to form insoluble
polymers known as humins through reaction with itself and with other monosaccharides.[6]
Therefore, an efficient production of 5-HMF requires minimization of undesirable side
reactions, such as its rehydration. This influences the efficiency of the process and the generation
of humins, which has been identified as the main cause of the low 5-HMF yields reported in the
majority of papers present in the literature.
In view of the above, different homogeneous and heterogeneous catalysts have been
applied in this reaction. Solid acid catalysts have been widely studied due to their high acidity,
easy separation from the reaction medium, reusability, among other properties. In particular, the
use of systems based on niobium oxide (Nb2O5) have played an important role in various
catalytic reactions due to its high acidity and catalytic stability.
Wang et al. evaluated the catalytic activity of mesoporous niobium phosphate on fructose
dehydration using water as solvent and obtained a maximum yield of 45% for 5-HMF and 58%
of fructose conversion in optimal conditions (130°C, 30 minutes).[7] Nb2O5 obtained by the
calcination of niobium acid in a temperature of 400°C allowed total fructose conversion with
yield of 86% for 5-HMF using DMSO as solvent at 120°C for 2 hours.[8]
Such examples demonstrate the great potential possessed by niobium-based systems.
However, its preparation in the form of nanocrystals with high surface area and acidity still
4

represents a big challenge. Some synthesis methodologies that aim the obtention of niobium
nanoparticles have been used, such as the sol-gel method followed by calcination or precipitation
method in aqueous ammonia.[9–12] However, the crystallization promoted under hydrothermal
conditions has been identified as a good strategy to obtain nanocrystals of metal oxide using
milder temperatures and reaction conditions.
Therefore, we report in this work the synthesis of a new class of catalysts obtained from
the hydrothermal treatment of the precursor Filter Cake (FC) with oxalic acid or hydrogen
peroxide and their use in the dehydration reaction of fructose in aqueous medium.
2. EXPERIMENTAL
2.1 Synthesis of Catalysts
2.1.1. Thermal treatment in presence of oxalic acid
The synthesis of Nb2O5 nanoparticles was performed using the hydrothermal method using
stainless steel autoclave BERGHOF® BR-100 containing Teflon cup and magnetic stirrer set at
300 rpm. 1g of the precursor material (Filter Cake, kindly supplied by CBMM) was dispersed in
40 mL of oxalic acid aqueous solution (0.2 mol/L). The suspension was placed in hydrothermal
cell and the temperature adjusted to 220°C and autogenous pressure. The studied times of the
hydrothermal treatment were 4 and 12h, and the materials identified as AO4h and AO12h,
respectively. A white solid was obtained as a final product, which was washed with distilled
water until pH=7, centrifuged and dried at 70°C for 12 hours.
2.1.2. Thermal treatment in presence of H2O2
For this synthesis, 1g of precursor material (Filter Cake) was dispersed in 40 mL of water. To
this dispersion, hydrogen peroxide (Synth, 50% v/v) was added in a molar ratio of 10:1 H2O2:Nb,
resulting in a yellow color, which is indicative of the niobium-peroxo complex formation. The
hydrothermal treatment was carried out at 150°C at autogenous pressure and times of 4 and 12h
were evaluated, being the materials identified as PH4h and PH12h, respectively. In the end, a
white solid was obtained and washed with distilled water until pH=7, centrifuged and dried at
70°C for 12 hours.
5

2.2. Catalytic Reactions
The fructose dehydration reactions were conducted in batch mode using a 15mL glass reactor
with a stainless steel screw top under autogenous pressure and constant magnetic agitation (500
rpm). In a typical run, 10mL of fructose aqueous solution (20, 40 or 60 g/L) was placed in
contact with the appropriate amount of catalyst (0.1 or 0.2g). Time zero was noted when the
temperature reached the adequate value (35, 70 or 130oC). After the reaction time, the reactor
was quickly submerged in a water bath cooled with ice. After cooling, the sample was
centrifuged and the supernatant liquid was analyzed by gas chromatography-mass spectrometry
and high-performance liquid chromatography, as described below. The reaction conditions used
were chosen based on other studies[13], where they obtained significant results of conversion
and selectivity. For reuse tests, the catalyst was rinsed with water, separated by centrifugation
and dried at 100oC for 3h. Then, the catalyst was again tested under the same experimental
conditions used in the original reaction system.
2.3. Products Characterization
The remaining fructose concentration was determined using calibration curve through high-
performance liquid chromatography with refractive index detector Waters 410 and Supelco
Nucleosil 100 NH2 5μm (25 cm × 4.6 mm) column. Acetonitrile: water (60:40) was used as
mobile phase at a flow rate of 1.0 mL min-1.
The products obtained were analyzed by gas chromatography-mass spectrometry using the
equipment Agilent, HP-5MS column (5% polymethylphenylsiloxane) 30 m x 250 µm x 0.25 µm.
The temperature profile used for the heating furnace was initially set to 40°C with permanence of
1 minute, and then increased to 90°C at a heating rate of 3°C/min, and maintained at that
temperature for 10 minutes.
The concentration of 5-HMF was determined by calibration curve. The product yield and the
fructose conversion were based on the initial sugar concentration and calculated as:
6

For reactions using DMSO as solvent, NMR spectroscopy was also used to characterize the
products. 300µL of the reaction product and 200µL DMSO-d6 containing gallic acid as internal
standard (1.50 mg.mL-1) were transferred to 5 mm NMR tube for analysis. 1H NMR experiments
were carried out on a Bruker Avance DRX400 (9.4Tesla) equipped with an inverse multinuclear
5 mm probehead. An exponential weighting function corresponding to 0.3 Hz line broadening
was applied to the free induction decay before applying Fourier transformation. The spectra were
processed and analyzed using TopSpin 1.3 software. 13C NMR experiment as well as the
following two-dimensional NMR spectra were also obtained (HSQC (heteronuclear single-
quantum coherence – 1H x 13C) and HMBC (heteronuclear multiplebond correlation – 1H x 13C)).
2.4. Characterization of the materials
The X-ray diffraction patterns were obtained on a Shimadzu equipment, model XRD-7000 X-
ray diffractometer, equipped with copper tube, 30 mA current and 30 kV voltage. The scan rate
used was 4° min-1.
The analyses of transmission electron microscopy (TEM) were performed at the Microscopy
Center-UFMG using the Transmission Electron Microscope Tecnai G2-20-SuperTwin FEI-200
kV.
Thermogravimetric analyses were performed using the Shimadzu-TGA50H equipment. About
3.0 mg of the sample was continuously heated in a temperature range of 25 to 700°C with a
heating rate of 10°C min-1 in air atmosphere and flow of 50 mL min-1.
The Raman scattering measurements were performed at room temperature in a Bruker-Senterra
apparatus using a 532 nm laser with 10 mW power, a laser aperture of 50 µm and 50x
magnification objective, an integration time of 10s and 15 co-additions.
The textural properties of the materials were evaluated by the technique of
adsorption/desorption of nitrogen at 77K using the Quantachrome Autosorb iQ2 equipment. The
specific surface area was calculated using the BET method in a low relative pressure region. The
pore size distribution was calculated from the isotherm using the BJH model.
The acidity properties of the catalysts were characterized by Temperature Programmed
Desorption method (TPD) using NH3 as probe molecule. A Quantachrome ChemBET-3000
equipment containing TCD detector using a current of 150 mA and an attenuation of 32 was
used. Approximately 0.2 g of catalyst was treated at 100°C for 60 min in a continuous flow of
7

helium (80 mLmin-1) and NH3 adsorption was then held at 50°C. The TPD-NH3 profiles were
obtained at a heating rate of 10°C min-1.
3. RESULTS AND DISCUSSION
3.1. Characterization of the catalysts
Figure 1 shows the diffraction patterns of X-ray for samples subjected to hydrothermal
treatment as well as the precursor Filter Cake. The patterns resemble the monoclinic system
Nb2O5 reported by Tamura et al.[14] as well as the orthorhombic system of hydroxide niobium
oxide (Nb3O7(OH)) according to Izumi et al.[15] The characteristic peak at 2Ɵ = 22.54o
corresponds to the system hkl (0 4 0).
The increase in crystallinity observed after the treatment can be attributed to the
rearrangement of the niobium’s polyhedral units caused by the occurrence of dissolution and
recrystallization during the process of hydrothermal synthesis.[16] Furthermore, according to the
literature,[17] the presence of oxalic acid prevents the formation of skewed chains due to their
chelation to metal polyhedra through oxygen atoms, that is, the two carboxyl groups of the oxalic
acid would assist in preventing the disorderly polycondensation when binding to the metal center
(Figure 2), helping to control nucleation and morphology.
The materials were analyzed by thermogravimetry upon heating in room atmosphere. The
results are shown in Figure 3.
Analyses by thermogravimetry allowed to observe a common mass loss event for all
catalysts in a temperature range up to 100°C, approximately. This result can be attributed to the
loss of adsorbed water on the oxide surface.
The materials treated with hydrogen peroxide also showed a second mass loss event at a
temperature of approximately 170oC, as can be observed in the DTG curves. This event should
be associated with the decomposition of the peroxo groups, as reported by Passoni et al.[18] This
result demonstrates, therefore, that peroxo groups were incorporated into the material by the
utilization of hydrogen peroxide. This is very interesting since several authors show that the
peroxo group is highly active in several types of catalytic reactions.[19–21]
Moreover, no mass loss events were observed due to the decomposition of organic
matter, demonstrating that all oxalic acid has been hydrolyzed during the process.
8

Thermogravimetric analysis also shows that the catalysts are thermally stable at the temperatures
chosen for the reaction tests.
Raman spectroscopy was performed to confirm the niobium oxide structure of the
catalysts. Its importance lies in the fact that this technique is very sensitive to the structure and
binding order of metal oxides, especially in the region corresponding to the metal-oxygen
stretching modes.[22] In this way, Figure 4 shows the Raman spectrum of the catalysts.
Most of the compounds based on niobium oxide have NbO6 octahedral structures, which
have distortions, so observed bands in the region 500-800 cm-1 correspond to stretching NbO6
structures slightly distorted. The presence of bands is also observed in the region 200-300 cm-1,
which are assigned to angular deformation modes of Nb-O-Nb. The presence of bands in a
higher frequency region is related to a higher niobium-oxygen bond order. Thus, the bands
observed in the region of 900-1200 cm-1 are assigned to the symmetric stretching mode of Nb=O
surface sites. The fact that the band observed at approximately 992cm-1 is well defined is related
to the presence of a more ordered structure, which is in accordance with the certain degree of
crystallinity presented by materials as evidenced by X-ray diffractometry analyses.[23–25]
Figure 5 shows the transmission electron microscopy images (TEM) of the niobium
oxides synthesized under hydrothermal conditions and also the precursor Filter Cake. These
images clearly show that the use of oxalic acid strongly influenced the structural features and the
particle size. Nanorods with approximately 5nm of diameter were formed upon oxalic acid
treatment.
The interplanar spacing of 0.394 nm for the catalyst AO4h corresponds to the (0 4 0)
plane related to Nb2O5 phase, which relates to the maximum peak intensity in 2Ɵ = 22.54°
observed in the XRD analysis.[14]
A similar morphology has been described for tungsten and titanium oxides also
synthesized under hydrothermal conditions in the presence of oxalic acid.[26],[27] However, the
fact that the growth mechanism allowed the obtention of nanorods is not yet well
understood.[17]
Figure 5 also shows that the PH4h and PH12h catalysts have spherical nanoparticles
agglomerates containing crystalline regions, which is the same morphology observed by Leite et
al.[28] using a similar synthesis procedure. The use of H2O2 in combination with the
hydrothermal treatment has the advantage of obtaining crystalline nanomaterials under mild
9

reaction conditions besides being a clean method due to absence of any potentially toxic by-
product or contaminants that could interfere in the catalytic tests. The results obtained by
transmission electron microscopy show that the precursor Filter Cake had their morphology
deeply modified by the addition of H2O2 or oxalic acid. The modification performed under mild
hydrothermal conditions can significantly affect their catalytic properties, since some reactions
are sensitive to morphology and type of defects generated on the surface of catalysts.
It is also important to mention that the majority of the studies in the literature use more
drastic reaction conditions and different precursors to obtain nanomaterials based on Nb2O5. For
example, Lu et al.[29] obtained Nb2O5 nanorods by hydrothermal treatment of metallic Nb
powder for three days in an autoclave. This way, obtainment of Nb2O5 nanoparticles of 5nm
order using Filter Cake in the presence of oxalic acid and mild reaction conditions is not reported
in the literature.
The textural properties of the catalysts are summarized in Table 1. N2
adsorption/desorption isotherms (a) and the pore size distribution profile (b) obtained by the
Barrett-Joyner-Halenda method (BJH) are shown in Figure 6. As observed in the morphology
changes, hydrothermal treatment has also promoted a strong modification in the porous structure.
As a result, the new catalysts showed higher BET area than the precursor Filter Cake (Table 1).
All the isotherms were classified as type IV according to the IUPAC classification, which is
assigned to the multilayer adsorption of mesoporous solids.[30] It is noteworthy that the PH12h
catalyst exhibits a hysteresis that can be classified as type H2, very common for porous inorganic
oxides, being attributed to the pore connectivity effects.[31] Treatment with H2O2 showed a
more severe modification, generating pores with radio between 18 to 48 Å. These pore sizes can
facilitate the diffusion of larger molecules to the active sites and consequently favor the catalytic
process.
Besides that, it is known that the acidity of the materials is of fundamental importance for
the sugars conversion study into 5-HMF. Due to this, the materials were also characterized by
temperature programmed desorption of ammonia (TPD-NH3) (Figure 7).
The TPD-NH3 experiment was carried out to evaluate the change in total acidity and acid
strength after the hydrothermal treatment. The temperature at which ammonia desorbs
determines the strength of the acid sites. Based on this desorption temperature, the acid sites can
be classified as weak (100- 200°C), moderate (200-400°C) and strong (> 400°C).[33] From the
10

TPD-NH3 profiles (Figure 7) and Table 1 it is possible to observe that the hydrothermal
treatment provided an increase in total acidity, so that the AO4h catalyst has the largest
integrated area under the curve, indicating greater desorption of ammonia and higher acidity
compared to the other catalysts and to the precursor Filter Cake.
Though the total acidity of Filter Cake is much lower than the other samples
hydrothermally treated, the acid sites showed higher acid strength. The lower temperature shift
upon treatment might indicate a surface modification of the acid sites which weakly bond with
ammonia. The total acidity increase upon treatment might be related to the increase of BET area
and also to the morphology change, which might play an important role in how this sites can be
available.
Another aspect that can be observed is the decrease of the number of acid sites with the
increase of hydrothermal treatment time. Such behavior may be related to the reduction of BET
area, described above.
On the other hand, the catalysts treated with H2O2 present practically only strong sites
and also a smaller number of acid sites. Furthermore, the reduction in area shown by peaks at
about 300°C may be related to the generation of peroxo groups, which replaces acid groups in
materials.[34] The moderate acidity indicated by the peak in the region of 200 to 400°C can be
attributed to Brønsted acid sites.[35] This way, the reduction of that peak area is related to the
reaction of the Brønsted acid sites with hydrogen peroxide to generate the oxidant groups.
Comparatively , the acidity of the catalysts reported here was higher than, or similar to,
those obtained by other authors, who also using a hydrothermal synthesis, obtained niobium
oxide from the precursor niobic acid (Nb2O5.nH2O).[16] In that case, the treatment was
conducted at a temperature of 135oC for 3 days and the amount of Nb used corresponding to 0.25
mmol. The material was then calcined at 400°C for 4h in air atmosphere and presented a number
of acid sites of 25 µmol g-1. The author also reports the hydrothermal synthesis of niobium oxide
using niobic acid in combination with oxalic acid in the same conditions as before. The obtained
material showed a number of acid sites of 78 µmol g-1, which is close to the value obtained for
AO4h catalyst. Nevertheless, it is noteworthy that the differential of the materials reported in this
work is the use of a more economically viable precursor.
11

3.2. Catalytic Tests
The synthesized catalysts were evaluated in fructose dehydration reaction aiming the
formation of 5-HMF. The performance of several heterogeneous catalysts have been examined in
this reaction, however, most studies use organic solvents or ionic liquid as the reaction
medium.[36,37]
Thus, it was decided to use water as a solvent due to the significant advantages it offers
compared to the solvents mentioned. Water is a green solvent, has high solubility for sugars and
its disposal in the environment is generally not impactful. Although dehydration reactions in
aqueous media may deactivate most solid acid catalysts, niobium-based materials have been
presented water-tolerant acidic properties, which provide stability and has afforded good
catalytic results.[38–40]
Figure 8 shows initial studies conducted in the presence of catalysts hydrothermally
treated. For comparative purposes, Filter Cake, Filter Cake treated with neat water and the
commercial niobic acid HY-340 (CBMM) calcined at 500oC for 3 hours were also used as a
control. As more significant result in the reaction conditions employed, 71% conversion and a
yield of 22% for 5-HMF was obtained. This result is similar to that obtained by Wang et al.[41],
who evaluated the catalytic performance of a series of porous niobium phosphate catalysts
synthesized under hydrothermal conditions using CTAB as a template. The use of catalysts
named NbPO-pH2 and NbPO-pH7 in fructose dehydration reactions carried out in aqueous
medium at 130oC for 60 minutes allowed to obtain conversions of 58 and 68% and yields of 45
and 34%, respectively.
As expected, in aqueous systems, dehydration of fructose is generally non-selective,
leading to many byproducts besides 5-HMF. [42] Among the by-products identified by GC-MS
the presence of furyl hydroxymethyl ketone, furfural, 1-(2-furanyl)-ethanone, 2,5-hexadione, 1-
hydroxy-2-propanone and formic acid can be cited. It is noteworthy that although identified,
such products have not been quantified. Therefore, the low yield observed in most of the
reactions could be attributed to both the formation of these by-products as well as the formation
of humins, as reported in the literature.[39]
We also observed a significant increase in conversion comparing the catalysts used as
control (FC, FC hydrothermal, HY340) with those treated hydrothermally in the presence of
12

oxalic acid and hydrogen peroxide. This demonstrates that the modifications imposed by the
hydrothermal treatment described above exerted influence on catalysis.
It is very surprising that the AO4h catalyst showed the lowest conversion among those
treated hydrothermally, despite it having high acidity and BET area. This result demonstrates
that other characteristics of the catalyst play greater or equal influence on the studied reaction,
such as morphology, effective acidity, mesoporosity and defects generated on the surface of
catalyst. This way, it is not possible to associate the activity directly to just one of these
variables.
In face of this initial results and in addition to characterization data that suggest the
greater surface area of the AO4h catalyst and the presence of certain mesoporous feature for
PH12h, it was decided to evaluate the effect that some factors have on the catalytic activity using
these two materials in order to achieve optimization of reaction conditions.
3.2.1. Influence of fructose concentration
Comparing the data presented in Figure 9, it can be seen that the higher conversion values
are obtained when using a lower sugar concentration (20g/L). Moreover, one can see that for
higher fructose concentrations the HMF yield decreases slightly. This result is in agreement with
reports in the literature.[6,43] It was also very surprising that at 60g/L the value of conversion
was similar to the yield, leading to a selectivity close to 100%. At higher concentrations, the
probability of rehydration to the formation of byproducts is reduced. Moreover, as the amount of
available acid sites should have been insufficient to promote a higher conversion at 60g/L, the
same reason may be associated with a reduction of secondary reactions for formation of humin
and other byproducts.
3.2.2. Influence of the amount of catalyst
Increased catalyst mass exerted greater influence on the activity of FC and AO4h
catalysts (Figure 10). On the other hand, there was a reduction in the yield of 5-HMF in the
presence of higher catalyst loading.
The reason for obtaining a higher yield using a lower mass of catalyst may be associated
with lower availability of acid sites that promote 5-HMF rehydration. In face of this result, the
13

catalyst mass of 0.1g, fructose concentration of 20 g/L and the PH12h catalyst were chosen to
continue the study of the other reaction variables.
It is noteworthy that although the AO4h catalyst presents higher acidity, it has not
presented the best catalytic performance despite this feature being extremely important for
reactions as studied here. The best results were obtained for the PH12h catalyst. This fact may be
related to its mesoporous feature, which would facilitate the access of the substrate to the active
site.
3.2.3. Temperature Influence
The results of fructose conversion and 5-HMF yield in function of the reaction
temperature are shown in Table 3.
These results suggest that the reaction temperature is a crucial parameter for 5-HMF
formation. The conversion increases from 6 to 71% by increasing the temperature from 70oC to
130oC, and there was no formation of 5-HMF at temperatures of 35 and 70oC.
Kinetic studies show that the conversion of fructose into 5-HMF has an activation barrier
of approximately 140 kJ/mol, in contrast to activation energy associated with undesirable side
reactions (~ 60 kJ/mol).[44] Therefore, the 5-HMF yield tends to increase with a higher
temperature, but its value is influenced by the nature of the reaction system, whereas in aqueous
systems fructose dehydration is usually non-selective, leading to the formation of many by-
products. In view of results presented, 130oC was chosen as the optimum temperature reaction.
3.2.4. Influence of Reaction Time
As seen on Figure 11, the conversion increases with reaction time and reaches a plateau
at the time of 2 hours.
Data in the literature has demonstrated that long reaction times contribute to the
formation of by-products due to decomposition of 5-HMF in humins, levulinic acid, formic acid
and other condensation products.[45] This explains, therefore, the low yield values obtained for
reaction times longer than 3 hours.
Carlini et al.[13] using solid acid catalysts based on niobium (niobic acid and niobium
phosphate) in an aqueous medium and at 100°C evaluated dehydration of fructose, sucrose and
inulin aiming at the formation of 5-HMF. The results showed that shorter reaction times (30
14

minutes) give higher selectivities for 5-HMF (>90%) and low substrate conversion (about 30%).
However, the increase in reaction time made possible the increase of sugar conversion (~ 60%)
and reduced the 5-HMF selectivity.
To investigate further the 5-HMF formation, the reaction mixture was analyzed by NMR
1
spectroscopy. Figure 12 shows the H NMR spectrum for the dehydration fructose with PH12h
in water after 2h at 130oC. The peaks between 3 and 4.5 are attributed to fructose, while the
peaks at 8.35 and 9.45 are assigned to formic acid and 5-HMF, respectively.[46] The signal for
the –COOH group proton of formic acid (H-1’) was too small and some papers attribute that
behavior to the fact that proton signal of formic acid did not resolve well in a reaction mixture or
disappeared due to its consumption in undefined side reactions.[47]
It was also carried out 13C NMR analysis (Figure 13) for the reaction and for 5-HMF
standard. Such spectra confirmed the production of 5-HMF with the corresponding peaks
identified as C-1 to C-6. As in the 1H NMR spectrum, the presence of formic acid as side product
was also evidenced, which can be explained by the rehydration of 5-HMF. Moreover, the self-
polymerization of 5-HMF molecules or cross-polymerization of 5-HMF with fructose is also
possible,[48] particularly due to the fact that the 5-HMF yield is not increased upon prolonged
reaction as shown previously.
Under optimum reaction conditions, the stability of the catalyst was evaluated in
subsequent reactions, given its importance. Figure 14 shows the results of four reuses. After each
catalytic cycle, the PH12h catalyst was recovered by centrifugation, washed with water and dried
at 100oC for 3h.
We can see, therefore, the successive deactivation of the catalyst. The literature has
pointed out the formation of sugar-derived insoluble humin polymers and its deposition on the
catalyst as the main cause of the deactivation.[49]
As previous research indicated that solvent plays a very important role in the reaction,
dimethyl sulfoxide (DMSO) was used as solvent in the previously optimized conditions in order
to evaluate its effect on the catalytic performance of the materials reported here.
As Figure 15a demonstrates, catalytic performance of the catalysts was noticeably
affected by the solvent. Higher yields for 5-HMF are obtained in DMSO than in water. Fructose
was almost completely converted over all the catalysts since it was not possible to observe its
1
signal in H NMR spectra (Figure 15b). Typical HMBC e HSQC spectra were also obtained for
15

reaction using PH12h as catalyst and they are shown in supplementary information. The yield of
5-HMF reaches the maximum (47%) on the PH12h catalyst. Clearly, the catalysts in DMSO are
also an effective system in promoting dehydration of fructose to 5-HMF. No rehydration
byproducts such as levulinic acid and formic acid were detected.
Although there are no concrete and well-understood reasons, some studies have
demonstrated what would be the contribution of these aprotic solvents for the dehydration
reaction.
Previous studies have suggest that the use of DMSO as solvent increases the fructose
conversion rate and decreases the rate of unwanted side reactions, thus resulting in higher
selectivity. This aspect is attributed to the capacity of DMSO to associate with water molecules
that are generated during dehydration, which prevents the reaction to proceed beyond 5-HMF to
form FA and LA.[50]
Moreover, the higher conversion rate is related to the tautomeric equilibrium
composition. In the presence of DMSO equilibrium towards a higher proportion of furanose
forms. As such, a faster reaction could reduce the amount of condensation by-products generated
through the reaction of 5-HMF with fructose or another intermediary.[51]
4. CONCLUSIONS
In summary, nanostructured niobium oxide was successfully hydrothermally synthesized
with the use of oxalic acid and hydrogen peroxide under mild conditions. The hydrothermal
synthesis along with structure-directing agents mentioned played an important role in the
structure of the materials. It was possible to obtain nanoparticles with different morphologies.
This synthesis method still promoted the modification of porous properties with increase in BET
area and made possible the increase of acidity, which have significant influence on the type of
reaction studied.
The obtained results demonstrate that the use of the new class of catalysts described in
this study showed significant catalytic potential in fructose dehydration reaction in both aqueous
medium and DMSO as solvent aiming to obtain 5-HMF. The use of PH12h catalyst afforded
yields of up to 22% and 47% under optimal reaction conditions in aqueous medium and DMSO,
respectively.
16

ACKNOWLEDGEMENTS
The authors gratefully acknowledge CNPq e FAPEMIG for financial support, as well as
the Microscopy Center-UFMG and the infrastructure of the Chemistry Department of UFMG
and CEFET.
REFERENCES
[1] T. Wang, M.W. Nolte, B.H. Shanks, Catalytic dehydration of C6 carbohydrates for the
production of hydroxymethylfurfural (HMF) as a versatile platform chemical, Green
Chem. 16 (2014) 548–572.
[2] J.P. de Mesquita, P.S. Patricio, C.L. Donnici, D.F.S. Petri, L.C.A. Oliveira, F.V. Pereira,
Hybrid layer-by-layer assembly based on animal and vegetable structural materials:
multilayered films of collagen and cellulose nanowhiskers, Soft Matter 7 (2011) 4405-
4413.
[3] S. Caratzoulas, D.G. Vlachos, Converting fructose to 5-hydroxymethylfurfural: a quantum
mechanics/molecular mechanics study of the mechanism and energetics., Carbohydr. Res.
346 (2011) 664–72.
[4] R.J. Van Putten, J.C. Van Der Waal, E. De Jong, C.B. Rasrendra, H.J. Heeres, J.G. De
Vries, Hydroxymethylfurfural, a versatile platform chemical made from renewable
resources, Chem. Rev. 113 (2013) 1499–1597.
[5] S. Siankevich, G. Savoglidis, Z. Fei, G. Laurenczy, D.T.L. Alexander, N. Yan, et al., A
novel platinum nanocatalyst for the oxidation of 5-Hydroxymethylfurfural into 2,5-
Furandicarboxylic acid under mild conditions, J. Catal. 315 (2014) 67–74.
[6] A. a. Rosatella, S.P. Simeonov, R.F.M. Frade, C. a. M. Afonso, 5-Hydroxymethylfurfural
(HMF) as a building block platform: Biological properties, synthesis and synthetic
applications, Green Chem. 13 (2011) 754-793.
17

[7] Y. Zhang, J. Wang, J. Ren, X. Liu, X. Li, Y. Xia, et al., Mesoporous niobium phosphate:
an excellent solid acid for the dehydration of fructose to 5-hydroxymethylfurfural in
water, Catal. Sci. Technol. (2012) 2485–2491.
[8] F. Wang, H.Z. Wu, C.L. Liu, R.Z. Yang, W.S. Dong, Catalytic dehydration of fructose to
5-hydroxymethylfurfural over Nb 2O5 catalyst in organic solvent, Carbohydr. Res. 368
(2013) 78–83.
[9] Y. Zhou, Z. Qiu, M. Lü, A. Zhang, Q. Ma, Preparation and characterization of porous
Nb2O5 nanoparticles, Mater. Res. Bull. 43 (2008) 1363–1368.
[10] R. Brayner, F. Bozon-Verduraz, Niobium pentoxide prepared by soft chemical routes:
morphology, structure, defects and quantum size effect, Phys. Chem. Chem. Phys. 5
(2003) 1457–1466.
[11] M. Ristić, S. Popović, S. Musić, Sol-gel synthesis and characterization of Nb2O5
powders, Mater. Lett. 58 (2004) 2658–2663.
[12] E.R. Camargo, M.G. Dancini, The oxidant peroxo method ( OPM ) as a new alternative
for the synthesis of lead-based and bismuth-based oxides, J. Mater. Res. 29 (2013) 131–
138.
[13] C. Carlini, M. Giuttari, A. Maria Raspolli Galletti, G. Sbrana, T. Armaroli, G. Busca,
Selective saccharides dehydration to 5-hydroxymethyl-2-furaldehyde by heterogeneous
niobium catalysts, Appl. Catal. A Gen. 183 (1999) 295–302.
[14] S. Tamura, High-Pressure Phase Research on Nb205, J. Mater. Sci. 7 (1972) 298–302.
[15] Izumi, Hydrothermal Synthesis and Characterization of Nb30,(OH), Zeitschrift Fuer
Anorg. Und Allg. Chemie. 441 (1978) 196–204.
[16] T. Murayama, J. Chen, J. Hirata, K. Matsumoto, W. Ueda, Hydrothermal synthesis of
octahedra-based layered niobium oxide and its catalytic activity as a solid acid, Catal. Sci.
Technol. 4 (2014) 4250–4257.
18

[17] Q.D. Truong, T.H. Le, J.-Y. Liu, C.-C. Chung, Y.-C. Ling, Synthesis of TiO2
nanoparticles using novel titanium oxalate complex towards visible light-driven
photocatalytic reduction of CO2 to CH3OH, Appl. Catal. A Gen. 437-438 (2012) 28–35.
[18] L.C. Passoni, M.R.H. Siddiqui, A. Steiner, I. V. Kozhevnikov, Niobium peroxo
compounds as catalysts for liquid-phase oxidation with hydrogen peroxide, J. Mol. Catal.
A Chem. 153 (2000) 103–108.
[19] D. Bayot, M. Degand, B. Tinant, M. Devillers, Spectroscopic and structural
characterizations of water-soluble peroxo complexes of niobium(V) with N-containing
heterocyclic ligands, Inorganica Chim. Acta. 359 (2006) 1390–1394.
[20] D. Bayot, M. Devillers, Peroxo complexes of niobium(V) and tantalum(V), Coord. Chem.
Rev. 250 (2006) 2610–2626.
[21] N.T. Prado, F.G.E. Nogueria, A.E. Nogueira, C.A. Nunes, R. Diniz, L.C.A. Oliveira,
Modified niobia as a new catalyst for selective production of dimethoxymethane from
methanol, Energy and Fuels. 24 (2010) 4793–4796.
[22] I.E. Jehng, Jih-Mirn Wachs, Structural Chemistry and Raman Spectra of Niobium Oxides,
Chem. Mater. 3 (1991) 100–107.
[23] R.M. Pittman, A.T. Bell, Raman Studies of the Structure of Nb2O5/TiO2, J. Phys. Chem.
97 (1993) 12178–12185.
[24] I.E. Wachs, J.-M. Jehng, G. Deo, H. Hu, N. Arora, Redox properties of niobium oxide
catalysts, Catal. Today. 28 (1996) 199–205.
[25] O.F. Lopes, E.C. Paris, C. Ribeiro, Synthesis of Nb2O5 nanoparticles through the oxidant
peroxide method applied to organic pollutant photodegradation: A mechanistic study,
Appl. Catal. B Environ. 144 (2014) 800–808.
19

[26] V.B. Patil, P.V. Adhyapak, S.S. Suryavanshi, I.S. Mulla, Oxalic acid induced
hydrothermal synthesis of single crystalline tungsten oxide nanorods, J. Alloy. Compd.
590 (2014) 283–288.
[27] M. Wang, Q. Zeng, B. Zhao, C. Chen, G. Liu, D. He, Controlled synthesis of
anatase/tungstite heterogeneous nanomaterials induced by oxalic acid, Catal. Commun. 48
(2014) 60–64.
[28] E.R. Leite, C. Vila, J. Bettini, E. Longo, Synthesis of niobia nanocrystals with controlled
morphology., J. Phys. Chem. B. 110 (2006) 18088–90.
[29] M. Wei, H. Luo, K. Wei, Synthesis of Nb2O5 nanorods by a soft chemical process, J.
Nanomater. (2009) 1–5.
[30] F. Rouquerol, J. Rouquerol, K.S.W. Sing, P. Llewellyn, G. Maurin, J. Rouquerol, et al.,
Adsorption by Powders and Porous Solids. Principles, Methodology and Applications,
Second, Elsevier, 2014.
[31] M. Kruk, M. Jaroniec, Gas Adsorption Characterization of Ordered Organic - Inorganic
Nanocomposite Materials, Chem. Mater. 13 (2001) 3169–3183.
[32] S.R. Amador, Estudio del sistema Pt/zeolita + alúmina en las reacciones de reformación
de gasolinas, Universidad Autónoma Metropolitana, 2000.
[33] G.D. Yadav, A.D. Murkute, Preparation of the novel mesoporous solid acid catalyst
UDCaT-4 via synergism of persulfated alumina and zirconia into hexagonal mesoporous
silica for alkylation reactions, Adv. Synth. Catal. 346 (2004) 389–394.
[34] L.C.A. Oliveira, M.F. Portilho, A.C. Silva, H.A. Taroco, P.P. Souza, Modified niobia as a
bifunctional catalyst for simultaneous dehydration and oxidation of glycerol, Appl. Catal.
B Environ. 117-118 (2012) 29–35.
[35] N. Topsoe, Infrared and temperature-programmed desorption study of the acidic
properties of ZSM-5-type zeolites, J. Catal. 70 (1981) 41–52.
20

[36] J. Wang, W. Xu, J. Ren, X. Liu, G. Lu, Y. Wang, Efficient catalytic conversion of fructose
into hydroxymethylfurfural by a novel carbon-based solid acid, Green Chem. 13 (2011)
2678–2681.
[37] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Metal chlorides in ionic liquid solvents
convert sugars to 5-hydroxymethylfurfural., Science. 316 (2007) 1597–1600.
[38] S. Hayashi, M. Hara, K. Nakajima, Y. Baba, R. Noma, M. Kitano, et al., Nb2O5 . nH2O
as a Heterogeneous Catalyst with Water-Tolerant Lewis Acid Sites, J. Am. Chem. Soc.
133 (2011) 4224–4227.
[39] T. Armaroli, G. Busca, C. Carlini, M. Giuttari, A.M. Raspolli Galletti, G. Sbrana, Acid
sites characterization of niobium phosphate catalysts and their activity in fructose
dehydration to 5-hydroxymethyl-2-furaldehyde, J. Mol. Catal. A Chem. 151 (2000) 233–
243.
[40] T. Okuhara, Water-Tolerant Solid Acid Catalysts, Chem. Rev. 102 (2002) 3641−3666.
[41] Y. Zhang, J. Wang, X. Li, X. Liu, Y. Xia, B. Hu, et al., Direct conversion of biomass-
derived carbohydrates to 5-hydroxymethylfurural over water-tolerant niobium-based
catalysts, Fuel. 139 (2015) 301–307.
[42] M. Bicker, D. Kaiser, L. Ott, H. Vogel, Dehydration of D-fructose to
hydroxymethylfurfural in sub- and supercritical fluids, J. Supercrit. Fluids. 36 (2005) 118–
126.
[43] Y. Román-Leshkov, J.N. Chheda, J. a Dumesic, Phase modifiers promote efficient
production of hydroxymethylfurfural from fructose., Science. 312 (2006) 1933–1937.
[44] C. Moreau, A. Finiels, L. Vanoye, Dehydration of fructose and sucrose into 5-
hydroxymethylfurfural in the presence of 1-H-3-methyl imidazolium chloride acting both
as solvent and catalyst, J. Mol. Catal. A Chem. 253 (2006) 165–169.
21

[45] F.S. Asghari, H. Yoshida, Kinetics of the decomposition of fructose catalyzed by
hydrochloric acid in subcritical water: Formation of 5-hydroxymethylfurfural, levulinic,
and formic acids, Ind. Eng. Chem. Res. 46 (2007) 7703–7710.
[46] J. Zhang, E. Weitz, An in Situ NMR Study of the Mechanism for the Catalytic Conversion
of Fructose to 5-Hydroxymethylfurfural and then to Levulinic Acid Using 13C Labeled D-
Frutose, ACS Catal. 2 (2012) 1211–1218.
[47] S. De, S. Dutta, B. Saha, Microwave assisted conversion of carbohydrates and
biopolymers to 5-hydroxymethylfurfural with aluminium chloride catalyst in water, Green
Chem. 13 (2011) 2859–2868.
[48] X. Qi, M. Watanabe, T.M. Aida, R.L. Smith, Jr., Efficient process for conversion of
fructose to 5-hydroxymethylfurfural with ionic liquids, Green Chem. 11 (2009) 1327–
1331.
[49] H. Yan, Y. Yang, D. Tong, X. Xiang, C. Hu, Catalytic conversion of glucose to 5-
hydroxymethylfurfural over SO42−/ZrO2 and SO42−/ZrO2–Al2O3 solid acid catalysts,
Catal. Commun. 10 (2009) 1558–1563.
[50] R.M. Musau, R.M. Munavu, The preparation of 5-hydroxymethyl-2-furaldehyde (HMF)
from d-fructose in the presence of DMSO, Biomass. 13 (1987) 67–74.
[51] P. DAIS, Intramolecular hydrogen-bonding and solvation contributions to the relative
stability of the furanose form of d-fructose in dimethyl sulfoxide, Carbohydr. Res. 169
(1987) 159–169.
22

Figure 1. Diffraction patterns of X-ray of the catalysts synthesized (* = Nb2O5 (JCPDS - ICDD
file 27-1312); # = Nb3O7(OH) (JCPDS - ICDD file 31-928) ♦= Nb2O5(JCPDS - ICDD file 7-
61)).
Figure 2: Scheme of a possible action mechanism of the oxalic acid. Adapted from TRUONG et
al.[17]
23

Figure 3. Thermogravimetric analyses in air atmosphere for catalysts based on niobium.
Figure 4. The Raman spectrum of the catalysts Filter Cake, AO4h, AO12h, PH4h and PH12h.
24

Figure 5. Images of transmission electron microscopy for the samples AO4h, AO12h, PH4h,
PH12h and Filter Cake (FC).
25

(a)
(b)
Figure 6. (a) N2 adsorption/desorption isotherms and (b) the respective pore size distribution of
catalysts.
Figure 7. TPD-NH3 profiles for Filter Cake and for catalysts synthesized hydrothermally. The
number of acid sites was determined from the standard beta zeolite with a specified number of
sites of 472 µmol g-1.[32]
26

Figure 8. Fructose dehydration reactions in the presence of different catalysts. (Reaction
conditions: 0.1g of catalyst, 130oC, 2h, 20g/L of fructose).
Figure 9. Influence of the fructose concentration on conversion and 5-HMF yield using a
temperature of 130°C, reaction time of 2 hours and 0.1 g of catalyst.
27

Figure 10. Influence of the catalyst mass on the fructose conversion and 5-HMF yield using a
temperature of 130°C, reaction time of 2 hours and 20g/L of fructose.
Figure 11. The influence of reaction time on fructose conversion and 5-HMF yield using a
130°C, sugar concentration of 20g/L and 0.1g of catalyst PH12h.
28

1
Figure 12. H NMR spectrum of product obtained from dehydration reaction of fructose with
PH12h in water after 2h at 130oC. (D2O, 400 MHz). TSP-d4 was used as an internal standard.
Figure 13. 13C NMR spectra for (a) fructose dehydration reaction and (b) HMF standard as
comparison. (D2O, 400 MHz).
29