Efficient solid acid catalysts based on sulfated tin oxides for liquid phase esterification of levulinic acid with ethanol

Article abstract of DOI:10.1016/j.apcata.2018.04.041

Tin oxide nanomaterials prepared by hydrothermal synthesis at 100 °C or 140 °C with or without template and further calcination step were modified with sulfate groups by post synthesis treatment. The catalysts were characterized by X-ray powder diffraction (XRD), N₂ physisorption, UV Vis spectroscopy, TG analysis, XPS and solid state NMR spectroscopy. The acidity of the materials was characterized by temperature programmed desorption (TPD) of ammonia. The catalytic performance of nanosized SnO₂ catalysts and their sulfated analogues was studied in levulinic acid (LA) esterification with ethanol. Sulfated materials show significantly higher activity compared to non-sulfated ones. It was found that the synthesis parameters (temperature, template) for preparation of the parent SnO₂ nanoparticles influence significantly their textural properties and have a pronounced effect on the structural characteristics of the obtained sulfated tin oxide based materials and their catalytic performance in levulinic acid esterification. Skipping the calcination step during the preparation of parent SnO₂ samples synthesized without template resulted in the formation of new, highly crystalline phase based on hydrated tin(IV) sulfate [Sn(SO₄)₂.xH₂O], tin(IV) bisulfate [Sn(HSO₄)₄.xH₂O] and/or tin(IV) pyrosulfate [Sn(S₂O₇).xH₂O] species in the sulfated nanomaterials with superior catalytic performance. The formation of this new and catalytically very active phase not reported so far in the literature for sulfated tin oxide-based materials is discussed. The catalytically active sites for esterification of levulinic acid with ethanol is suggested to result from the formation of strong Br?nsted and Lewis acid sites with high density in the newly registered phase. The results indicate that the chemical structure and catalytic performance of the obtained sulfated tin oxide based materials strongly depend on the treatment of the SnO₂ nanoparticles before the sulfation procedure.

Full text of DOI:10.1016/j.apcata.2018.04.041

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
AppliedCatalysisA,General560(2018)119–131
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Efficient solid acid catalysts based on sulfated tin oxides for liquid phase
esterification of levulinic acid with ethanol
Margarita Popovaa,⁎, Pavletta Shestakovaa,⁎, Hristina Lazarovaa, Momtchil Dimitrova,
Daniela Kovachevab, Agnes Szegedic, Gregor Malid, Venkata Dasireddyd, Blaž Likozard,
Nicole Wildee, Roger Gläsere
a
Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, 1113, Sofia, Bulgaria
Institute of General and Inorganic Chemistry, Bulgarian Academy of Sciences, 1113, Sofia, Bulgaria
Research Centre for Natural Sciences, Institute of Materials and Environmental Chemistry, Hungarian Academy of Sciences, 1117, Budapest, Hungary
National Institute of Chemistry, Hajdrihova 19, 1001, Ljubljana, Slovenia
b
c
d
e
Universität Leipzig, Institute of Chemical Technology, 04103, Leipzig, Germany
A R T I C L E I N F O
A B S T R A C T
Keywords:
Tin oxide nanomaterials prepared by hydrothermal synthesis at 100 °C or 140 °C with or without template and
further calcination step were modified with sulfate groups by post synthesis treatment. The catalysts were
characterized by X-ray powder diffraction (XRD), N2 physisorption, UV Vis spectroscopy, TG analysis, XPS and
solid state NMR spectroscopy. The acidity of the materials was characterized by temperature programmed
desorption (TPD) of ammonia. The catalytic performance of nanosized SnO2 catalysts and their sulfated ana-
logues was studied in levulinic acid (LA) esterification with ethanol. Sulfated materials show significantly higher
activity compared to non-sulfated ones. It was found that the synthesis parameters (temperature, template) for
preparation of the parent SnO2 nanoparticles influence significantly their textural properties and have a pro-
nounced effect on the structural characteristics of the obtained sulfated tin oxide based materials and their
catalytic performance in levulinic acid esterification. Skipping the calcination step during the preparation of
parent SnO2 samples synthesized without template resulted in the formation of new, highly crystalline phase
based on hydrated tin(IV) sulfate [Sn(SO4)2.xH2O], tin(IV) bisulfate [Sn(HSO4)4.xH2O] and/or tin(IV) pyr-
osulfate [Sn(S2O7).xH2O] species in the sulfated nanomaterials with superior catalytic performance. The for-
mation of this new and catalytically very active phase not reported so far in the literature for sulfated tin oxide-
based materials is discussed. The catalytically active sites for esterification of levulinic acid with ethanol is
suggested to result from the formation of strong Brønsted and Lewis acid sites with high density in the newly
registered phase. The results indicate that the chemical structure and catalytic performance of the obtained
sulfated tin oxide based materials strongly depend on the treatment of the SnO2 nanoparticles before the sul-
fation procedure.
Nanosized tin oxide
Sulfated catalysts
Ethyl levulinate
Esterification
Biomass
Solid state NMR
1. Introduction
valuble compounds that can be used as fuel additives, solvents and
plasticizers. In particular, ethyl levulinate can directly be used up to
In the recent years, lignocellulosic biomass has been intensively
studied for production of chemicals, fuels and energy. Levulinic acid
and its esters are promising platform chemicals for the production of a
broad range of sources for the biofuel, polymer and fine chemicals in-
dustry [1–5]. Levulinic acid (LA) is generally produced by the acid-
catalyzed hydrolysis of cellulose, and can be converted into levulinate
esters, γ-valerolactone, 1,4-pentanediol and 5-nonanone (via pentanoic
acid) as well as diphenolic acid as an intermediate for the synthesis of
epoxy resins and poly-carbonates [6–20]. Levulinate esters are also
5 wt. % as a diesel-miscible biofuel in regular diesel car engines without
modification because of its physicochemical properties similar to those
of fatty acid methyl esters (FAME), i.e., biodiesel [6]. Therefore ethyl
levulinate has the potential to decrease the consumption of petroleum-
derived fossil fuels. Generally, levulinate esters are produced by ester-
ification reactions using mineral acids such as HCl, H2SO4 and H3PO4,
resulting in the high yield of the corresponding products within a short
reaction time. However, in such homogeneously catalyzed reactions,
the acid catalysts are unrecyclable and often lead to technological
⁎
Corresponding authors.
E-mail addresses: mpopova@orgchm.bas.bg (M. Popova), psd@orgchm.bas.bg (P. Shestakova).
https://doi.org/10.1016/j.apcata.2018.04.041
Received 7 January 2018; Received in revised form 18 April 2018; Accepted 28 April 2018
Availableonline30April2018
0926-860X/©2018ElsevierB.V.Allrightsreserved.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
M. Popova et al.
AppliedCatalysisA,General560(2018)119–131
problems, e.g., use of a large volume of base for neutralization and
corrosion of equipment. The replacement of homogeneously catalyzed
reactions with heterogeneously catalyzed analogues in which the cat-
alysts are easily separable and reusable is thus highly desirable
[20–30]. In comparison to zeolites, over sulfated metal oxide, e.g.,
SO42−/Nb2O5, SO42−/TiO2, SO42−/SnO2, as catalysts higher conver-
sions (up to 44% conversion on SO42−/SnO2 at 70 °C) in the levulinic
acid esterification is registered due to the presence of stronger acid
sites. The acid site strength of SO42−/SnO2 is reported to be higher than
that of SO42−/ZrO2 [2]. The increase in the surface acidity strength of
the modified oxides leads to the raise in activity of the catalyst. The
acidity generated by modification of SnO2 with sulfate anions is even
stronger than pure sulfuric acid. Thus it can be consider as super acid
catalysts for many commercially important reactions, such as hydro-
cracking of paraffins, dehydration of alcohols, esterification, alkylation
of olefins and protection of aldehydes, ketones and alcohols and olefins
[31]. However, a detailed explanation on the nature of the active acid
sites (Brønsted and Lewis) and the mechanism of their formation is still
missing.
Different procedures are suggested in the literature for the pre-
paration of sulfated tin oxide with strong acid sites. H. Miyazaki et al.
[32] reported the preparation of a highly active sulfated tin oxide from
tin oxide gel, precipitated by the hydrolysis of SnCl4. A. A. Dabbawala
et al. [33] revealed the role of different sulfur content (1–8 wt%) during
thermal decomposition of stannous sulfate during the preparation of
sulfated tin oxides]. The sulfation procedure was carried out by im-
pregnation of tin(IV) hydroxide with 0.5 M sulfuric acid resulting in the
preservation of tin(IV) oxide tetragonal phase, decrease of the crystal-
lite size and increase in the specific surface area. The effect of sulfate
content and calcination temperature on the structure, acidity and cat-
alytic activity of tin oxide was studied by A.S. Khder et al. [34] as well.
They found that the sulfation enhances the concentration of Brønsted
2. Experimental part
2.1. Synthesis of nanosized SnO2 materials
Nanosized SnO2 samples were synthesized by precipitation of SnCl4
solution with 20% NH4OH in the presence or absence of template fol-
lowed by hydrothermal treatment. In a typical preparation, N-hex-
adecyl-N,N,N-trimethylammoniumbromide (12.0 g) was dissolved in
100 ml distilled water. To this solution was added slowly and under
vigorous stirring a second solution of SnCl4 (7.60 g) in 50 ml distilled
water. Then the temperature was raised to 50 °C and stirred for 30 min
before adding dropwise 40 ml NH3 (12.5%). The resulting mixture was
stirred overnight at 50 °C. Then it was transferred into an autoclave and
treated at 100 °C or 140 °C for 24 h. The hydrothermal treatment was
followed by filtration of the solution and washing with distilled water,
then drying at room temperature was applied. Some of the samples
were calcined up to 500 °C with a ramp of 1° per minute and a swelling
time of 15 h at the final temperature. In the case of template-free
samples the SnCl4 (7.60 g) was added to 100 ml distilled water instead
of 50 ml.
2−
2.2. Functionalization of nanosized SnO2 by SO4 groups
Nanosized SnO2 samples were mixed with 10% wt. H2SO4 solution
(40 ml/1 g SnO2). The suspension was dried at ambient temperature
and calcined at 300 °C for 3 h. The samples after sulfation are denoted
as SO42−/SnO2(x)(y) or SO42−/SnO2(x)T(y), where x is temperature of
hydrothermal synthesis; y is the calcination temperature and T in-
dicates the use of template during the synthesis.
2.3. Characterization
acid sitesand increases the strength of Lewis acidity when 20 wt.%
Powder X-ray diffraction (XRD) patterns were collected within the
range of 10–80° 2θ with a constant step of 0.02° 2θ and counting time of
1 s/step on Bruker D8 Advance diffractometer equipped with Cu Kα
radiation and LynxEye detector. Mean crystallite sizes were determined
by the Topas-4.2 software package using the Laue formula for the in-
tegral breadth (IB) of the diffraction peaks and fundamental parameters
peak shape description for the instrumental broadening and dif-
fractometer geometry.
Nitrogen physisorption measurements were carried out at −196 °C
using Tristar 3000 Micromeritics volumetric adsorption analyzer.
Before the analysis, the samples were outgassed under high vacuum for
2 h at 250 °C. The pore-size distributions were calculated from the
desorption isotherms by the BJH method.
2−
SO4
groups are applied and calcination is carried out at 550 °C.
In the already published papers the SnO2 phase detected in sulph-
ated materials showed similar crystalline properties to the parent SnO2
materials thus suggesting that the sulfation procedures affected only
surface of SnO2 nanoparticles while their bulk structure remained un-
changed. The increased catalytic activity of the sulfated tin oxides is
considered to be a result of the formation of SO42−/SnO2 chelate
2−
structures of different coordination modes between SO4
and SnO2
which is associated with an increased concentration of Brønsted and
Lewis acid sites. However the problem related to the low stability in
liquid phase reactions and the considerable leaching of the active phase
of surface SO42−/SnO2 species during the catalytic experiments which
is more pronounced in comparison to SO42−/ZrO2 remains a major
challenge [2,35,36]. Despite the strong acidity of the sulfated metal
oxides their application is limited due to the leaching of sulfate groups
during the reaction which is the most important drawback of this type
of catalysts [33,34]. Moreover, to the best of our knowledge the exact
chemical and structural nature of the active sites in sulfated tin oxide
nanoparticles are still not fully clarified in the literature.
In the present study, a series of sulfated tin oxide based catalysts and
their catalytic activity in the esterification of levulinic acid with ethanol
were studied. We used different approaches to prepare active sulfated
species in tin oxide based nanomaterials and to stabilize them on the
support during the catalytic reaction. The parent SnO2 materials were
synthesized by hydrothermal treatment with or without further calci-
nation step and in the presence or absence of a template. The influence
of the textural and morphological properties of the parent SnO2 pre-
pared under different synthesis conditions on the properties of the
sulfated SnO2 based catalysts was investigated.
Ammonia temperature-programmed desorption (NH3-TPD) was
carried out using a Micromeritics 2920 Autochem II Chemisorption
2−
Analyser. The catalyst was pre-treated at 500 °C (at 300 °C for SO4
/
SnO2(100) and SO42−/SnO2(140)) under the stream of helium for
60 min. Then the temperature was decreased to 80 °C. A mixture of
9.8% NH3 in He was passed over the catalyst at a flow rate of 25 mL/
min for 60 min. The excess NH3 was removed by purging with helium
for 25 min. The temperature was then raised gradually to 900 °C by
ramping at 10 °C/min under the flow of helium and desorption data
were recorded. The TCD signals were calibrated using various gas
concentrations of NH3 ranging from 0 to 10 wt. % NH3 in He. The
desorbed amount of NH3 was determined continuously in a thermal
conductivity cell and by absorption in a trap containing 0.05 M H2S04
followed by titration with 0.05 M NaOH solution.
The surface chemical composition of selected samples was analyzed
by X-ray photoelectron spectroscopy (XPS). The measurements were
carried out on AXIS Supra electron spectrometer (Kratos Analitycal
Ltd.) using monochromatic Al Ka radiation with photon energy of
1486.6 eV. The energy calibration was performed by normalizing the
C1s line of adsorbed adventitious hydrocarbons to 284.6 eV. The
binding energies (BE) were determined with an accuracy of
0.1 eV.
120

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
M. Popova et al.
AppliedCatalysisA,General560(2018)119–131
Fig. 1. XRD patterns of the obtained initial (A) and sulfated (B) SnO2 samples.
The chemical composition of the samples was determined monitoring
the areas and binding energies of O1s, Sn3d and S2p photoelectron
peaks. Using the commercial data-processing software of Kratos
Analytical Ltd., the concentrations of the different chemical elements
(in at.%) were calculated by normalizing the areas of the photoelectron
peaks to their relative sensitivity factors.
The bulk chemical analysis was performed by classical chemical
methods. Hydrogen and tin content were defined by the speed method
for determination of compounds containing carbon, hydrogen and one
metal component, while sulfur content was determined by oxygen
pyrophoric incineration method. The oxygen content was calculated as
the difference to 100%.
SEM images and elemental analyses were made by a JEOL-JSM-
6390 scanning electron microscope equipped with package (standard-
less method) of Oxford Instruments. Spectra collecting conditions: ac-
celerating voltage – EDS with (Li, Si) detector. The characteristic X-Ray
spectra were analyzed with the INCA software 20 KV, integral mode of
spectra collection, current-50 spot size and collecting time 300 s.
119Sn solid state NMR spectra of precursor SnO2 samples were re-
corded on a Bruker Avance II + 600 NMR spectrometer operating at
600.01 MHz 1H frequency (223.75 MHz for 119Sn), using 4 mm solid
state CP/MAS 1H/X dual probehead. The samples were loaded in 4 mm
zirconia rotors and spun at magic angle spinning (MAS) rates of 14 and
11 kHz to identify the isotropic chemical shift values. NMR spectra were
measured with single-pulse sequence, 16 K time domain data points,
spectrum width of 2000 ppm, 6000 scans and a recycle delay of 5 s. The
spectra were referenced with respect to neat SnCl4 (isotropic chemical
shift −145 ppm). Exponential window function was applied (line
broadening factor 100) prior to Fourier transformation.
pulses was equal to one sample-rotation period (50 μs). During acqui-
sition high-power XiX proton decoupling was applied. Recycle delay
between consecutive scans was 80 s, and numbers of scans ranged be-
tween 800 and 4200 for different samples. The spectra were referenced
with respect to neat SnCl4 (isotropic chemical shift −145 ppm).
FT-IR experiments were performed with Nicolet Compact 640
spectrometer by the self-supported wafer technique with pyridine (Py)
(7 mbar) as probe molecule. Self-supported pellets (10 × 20 mm) were
pressed from the samples, placed into the IR cell, heated up to 350 °C in
high vacuum (10–6 mbar) with a rate of 10 °C/min and dehydrated for
1 h. Following 30 min contact with Py at 100 °C the sample was evac-
uated subsequently at 100, 200 and 300 °C, for 30 min. After each
evacuation step a spectrum was recorded at IR beam temperature with
a resolution of 2 cm−1. The spectra were normalized to 15 mg/cm2
weight of the wafers for comparison.
2.4. Catalytic experiments
Prior to the catalytic experiments the samples were pre-treated ex-
situ in oven for 1 h at 140 °C at static conditions. In a typical experi-
ment, the reactor was charged with 1 g LA (Sigma-Aldrich, 98%), 7 ml
etannol (Sigma-Aldrich, ≥99.8%) and 0.050 g powder catalyst (2.5 wt.
% catalyst/LA) while the LA/ethanol weight ratio was maintained 1:5.
The reactor was placed in an oil bath, heated under stirring with
200 rpm at the reaction temperature of 70 °C for 7 h. The thermocouple
was positioned in the reaction mixture for accurate measurement of the
reaction temperature. Samples were taken every hour from the reaction
mixture starting from 1 h reaction time and analyzed using HP-GC with
119Sn solid state NMR spectra of sulfated samples were recorded on
a Varian VNMRS spectrometer operating at 599.53 MHz 1H frequency
(223.40 MHz for 119Sn), using 3.2 mm solid state CP/MAS 1H/X pro-
behead. The samples were loaded in 3.2 mm zirconia rotors and spun at
MAS rates of 20 kHz. The spectra were measured with single-pulse and
Hahn-echo sequences, with 90- and 180-degree pulses of 1.9 μs and
3.8 μs, respectively. In the echo experiments the delay between the two
a WCOT FUSED SILICA 25 m × 0.25 mm COATING CP-SIL 43CB
column. The catalytic activity of all the studied samples in 7 h reaction
time was represented as levulinic acid conversion (%) and TOF (cal-
culated using the amount of the acid sites obtained by the TPD of NH3
versus reaction time). The applied mass balance was based on C-con-
taining products and reactants.
121

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
M. Popova et al.
AppliedCatalysisA,General560(2018)119–131
3. Results and discussion
Table 1
Element distribution from SEM-EDX analysis.
3.1. Powder X-ray diffraction (XRD)
Samples
Element distribution, at. %
O
S
Sn
XRD data of all parent nanosized tin oxide samples show reflections
typical of tetragonal SnO2 phase, i.e. cassiterite, identified by PDF#41-
1445 (Fig. 1A). In the case of hydrothermally treated samples (100 °C
and 140 °C) without further calcination step, the reflections are much
broader due to the formed very small tin oxide crystallites (2–3 nm).
XRD data also support that their crystalline morphology is not well
defined, and the presence of an amorphous phase cannot be excluded.
In contrast, the hydrothermally treated SnO2 samples calcined at 500 °C
(SnO2(100)T(500), SnO2(140)T(500), SnO2(100)(500) and SnO2(140)
(500)) are characterized by well-defined crystallites of about 10 nm. It
can be also observed that the application of template does not influence
significantly the size of the crystallites (Fig. 1A).
After treatment with sulfuric acid all samples were calcined at
300 °C (Fig. 1B). No observable differences in the XRD patterns of the
SO42−/SnO2(100)T(500) and SO42−/SnO2(140)T(500) samples in
comparison to their non-sulfated analogues was detected. This indicates
that the sulfuric acid does not substantially change the already well
crystallized tin dioxide phase formed in the presence of the template.
However, powder diffraction patterns of SO42−/SnO2(100)(500) and
SO42−/SnO2(140)(500) samples synthesized without template show
additional reflections indicating the appearance of a new crystalline
phase, the latter being better defined for SO4/SnO2(140)(500) (Fig. 1B).
However, very substantial changes with the initial tin dioxide phase
SO42−/SnO2(100)
84.88
67.13
9.49
3.12
5.62
29.75
SO42−/SnO2(100)(500) – agglomerates of
small particles
SO42−/SnO2(100)(500) – new phase with large
crystallites
80.77
12.28
6.95
In order to get more information about the chemical composition of
this new phase, SEM-EDX and XPS analyses (Figs. 2, S1) were per-
formed in combination with TG experiments of the selected sample.
3.2. Scanning electron microscopy/energy X-Ray spectroscopy (SEM/EDX)
and X-Ray photoelectron spectroscopy (XPS) analysis
The SEM images of SO42−/SnO2(100) showed that the new phase
mentioned above consists of randomly distributed well shaped big rod-
like particles (Fig. 2A). The SEM-EDX analysis gives the following
chemical composition of this material: O – 84.9 at.%, S – 9.5 at.%, Sn
5.6 at.% (Table 1). Thus the S:Sn molar ratio in the new crystalline
phase of SO42−/SnO2(100) sample is 1.7:1 which implies that ap-
proximately one Sn atom shares two SO4 moieties.
SEM images of SO42−/SnO2(100)(500) sample show the presence of
two phases – a main phase consisting of large irregular aggregates
composed of smaller (∼8 nm) particles, and a second phase re-
presenting regions with the new crystalline phase (Fig. 2B). The SEM-
EDS analyses show that the aggregates and the new phase with the large
crystallites have different element composition (Table 1). The sulfur
content of the aggregates in the main phase is very low with S:Sn molar
ratio of 1:10 (3.1 at. % S and 29.8 at. % Sn) while in the regions of the
new phase the S:Sn molar ratio corresponds to 1.8:1. The letter value is
practically identical with that of determined for the SO4/SnO2(100)
sample where only the new phase is present. This result leads us to the
conclusion that the new phase could be present in all acid treated
samples. However, its formation and crystallization was strongly fa-
voured in the sulfated materials obtained from non-calcined hydrated
SnO2 samples, where some residual water can be preserved.
occurred during the sulfuric acid treatment of the hydrothermally ob-
2−
tained and non-calcined samples SO42−/SnO2(100) and SO4
/
SnO2(140) (Fig. 1B). The peaks corresponding to the finely dispersed tin
dioxide disappeared completely at the expense of the appearance of a
well crystallized new phase with layered-type organization. These re-
sults demonstrate that during the treatment with sulfuric acid, the very
small tin dioxide crystallites (around 2–3 nm) undergo complete
transformation into this new, very well defined crystalline phase
(Fig. 1B). Detailed examination of the powder diffraction data showed
that this pattern could not be assigned to any known phase in the da-
tabase ICDD-PDF-2(2014). The preliminary analysis showed that the
pattern seems to be comprised of only one phase without presence of
any other known tin-containing phases. Indexing procedure of powder
diffraction pattern of the new phase resulted in monoclinic lattice
containing most probably pyrosulfate type species. This crystal phase is
characterized by unit cell parameters of a = 16.914 Å, b = 10.127 Å,
c = 12.167 Å and β = 91.08°, and a unit cell volume that exceeds 2000
Å3. To perform crystal structure determination from powder diffraction
data, a complete knowledge about the exact composition and formula
unit of the phase is crucial for building the initial structure model.
The chemical analysis of SO42−/SnO2(100) shows the following
bulk element distribution: H – 2.7 at. %; S – 13.1 at. %, Sn – 6.4 at.%, O
– 77.8 at.%. The performed bulk chemical analysis of SO42−/SnO2(100)
shows similar chemical composition of this material to that obtained by
SEM-EDS analysis.
The surface chemical composition of SO42−/SnO2(100) sample was
studied by XPS analysis (Figure S1). The registered binding energy
Fig. 2. SEM images of SO42−/SnO2(100) (A) and SO42−/SnO2(100)(500) (B).
122