Fluoride-free synthesis of a Sn-BEA catalyst by dry gel conversion

Article abstract of DOI:10.1039/c4gc02457e

Sn containing molecular sieves with BEA topology (Sn-BEA) are active Lewis catalysts for a large variety of redox reactions. However, the synthesis of Sn-BEA often requires the use of toxic chemicals (e.g. hydrofluoric acid). In this study, we demonstrated that an active Sn-BEA catalyst can be directly synthesized in a non-fluoride medium via a dry gel conversion method with the aid of seed crystals. It was shown that seeding of zeolite BEA crystals can facilitate crystal growth and lead to BEA topology with framework Sn. Furthermore, it was found that ion-exchange with ammonium ions is indispensable to retain the crystal structure during calcination. The catalytic activity of the Sn-BEA catalyst was evaluated by the isomerization of glucose and reaction of pyruvaldehyde in the aqueous phase, and compared with the conventional hydrophobic Sn-BEA synthesized in a fluoride medium. The catalytic activity of the Sn-BEA synthesized in the absence of fluoride was found to be lower than that of the conventional Sn-BEA for catalyzing the isomerization of glucose in the aqueous phase, which could be attributed to its hydrophilic surface. For the reaction of pyruvaldehyde in the aqueous phase, the Sn-BEA synthesized in this study showed similar activity to the conventional hydrophobic Sn-BEA, indicating that the catalytic activity of Sn-BEA catalysts largely depends on the type of reactions. The current work provides the first example of direct preparation of a Sn-BEA catalyst in a caustic medium.

Full text of DOI:10.1039/c4gc02457e

Green Chemistry
View Article Online
View Journal
PAPER
Fluoride-free synthesis of a Sn-BEA catalyst by dry
gel conversion†
Cite this: DOI: 10.1039/c4gc02457e
Chun-Chih Chang, Hong Je Cho, Zhuopeng Wang, Xuanting Wang and Wei Fan*
Sn containing molecular sieves with BEA topology (Sn-BEA) are active Lewis catalysts for a large variety
of redox reactions. However, the synthesis of Sn-BEA often requires the use of toxic chemicals (e.g.
hydrofluoric acid). In this study, we demonstrated that an active Sn-BEA catalyst can be directly syn-
thesized in a non-fluoride medium via a dry gel conversion method with the aid of seed crystals. It was
shown that seeding of zeolite BEA crystals can facilitate crystal growth and lead to BEA topology with
framework Sn. Furthermore, it was found that ion-exchange with ammonium ions is indispensable to
retain the crystal structure during calcination. The catalytic activity of the Sn-BEA catalyst was evaluated
by the isomerization of glucose and reaction of pyruvaldehyde in the aqueous phase, and compared with
the conventional hydrophobic Sn-BEA synthesized in a fluoride medium. The catalytic activity of the Sn-
BEA synthesized in the absence of fluoride was found to be lower than that of the conventional Sn-BEA
for catalyzing the isomerization of glucose in the aqueous phase, which could be attributed to its hydro-
philic surface. For the reaction of pyruvaldehyde in the aqueous phase, the Sn-BEA synthesized in this
study showed similar activity to the conventional hydrophobic Sn-BEA, indicating that the catalytic activity
of Sn-BEA catalysts largely depends on the type of reactions. The current work provides the first example
of direct preparation of a Sn-BEA catalyst in a caustic medium.
Received 17th December 2014,
Accepted 2nd March 2015
DOI: 10.1039/c4gc02457e
www.rsc.org/greenchem
Sn-BEA is water-tolerant and capable of catalyzing isomeriza-
tion of glucose to fructose in the aqueous phase,7 a key inter-
Introduction
Zeolites have been extensively used in catalysis, ion-exchange mediate step for the conversion of biomass into chemicals and
and separation due to their highly crystalline microporous fuels.20 Similarly, Sn-BEA has also been used to catalyze the
structures and a large variety of available frameworks and isomerization of pentose and triose sugars to form xylulose,
compositions.1–4 In addition to Brønsted acid catalysts, mole- methyl lactate and lactic acid.6,8,21 The outstanding Lewis
cular sieves containing tetrahedrally coordinated Sn, Ti and acidity of the Sn-BEA catalyst in the aqueous phase is likely
Zr have been used for Lewis acid-catalyzed redox reactions.5–11 due to the hydrophobic environment created by the defect-free
In particular, Sn-BEA molecular sieve, a crystalline stannosili- siliceous surface.15,22,23
cate with BEA topology, exhibit an extraordinary Lewis acidity
Although Sn containing molecular sieves show intriguing
for selectively activating functional groups of organic mole- catalytic properties, typical synthesis of Sn-BEA involves the
cules involved in fine chemical syntheses.12–18 The catalyst has utilization of toxic hydrofluoric acid and a lengthy synthesis
shown promising activity for the catalytic oxidation of satu- procedure (up to 40 days),7,10,24 leading to several environ-
rated and unsaturated ketones through the Baeyer–Villiger mental and practical concerns. Significant efforts have been
reaction and the reduction of carbonyl compounds with made to improve the synthesis process of Sn-BEA, such as
secondary alcohols through the Meerwein–Ponndorf–Verley using less toxic chemicals25 and shortening the crystallization
reaction.9,10,19 Moreover, Moliner et al. recently reported that time by a seeded growth method.26 However, fluoride medium
is still used in all syntheses. Although more benign chemicals
such as ammonium fluoride (NH4F) and sodium fluoride
(NaF) can be potentially used in the synthesis, a non-fluoride
Department of Chemical Engineering, University of Massachusetts Amherst,
686 North Pleasant St., 159 Goessmann Lab, Amherst, MA 01003, USA.
method is still strongly desired due to potential health risks
related to chronic excessive exposure to fluorides.27 Recently,
post-synthesis approaches have been used to synthesize Sn-
BEA in the absence of fluorides. For instance, Li et al. reported
that Sn can be grafted onto dealuminated zeolite BEA by
chemical vapor deposition using tin(IV) chloride vapor;
E-mail: wfan@ecs.umass.edu; Fax: +1-413-545-1647; Tel: +1-413-545-1750
†Electronic supplementary information (ESI) available: The procedures for syn-
thesis of Sn-BEA and SnO2/Si-BEA in a fluoride medium, the XRD patterns for
+
the sample prepared in the absence of alkali ions and with NH4 salts, the SEM
image for Sn-BEA-F, the XRD pattern for SnO2/Si-BEA, product distribution for
the reactions and the 1H-NMR spectrum for fructose in the isomerization of deu-
terium-labeled glucose. See DOI: 10.1039/c4gc02457e
This journal is © The Royal Society of Chemistry 2015
Green Chem.

View Article Online
Paper
Green Chemistry
Hammond et al. proposed a solid state ion exchange method (Ti4+: 0.68 Å, Sn4+: 0.71 Å) and the incomparable hydrolysis
via mixing tin(II) acetate with dealuminated zeolite BEA; rate of the Sn source to the silica source. These differences
Dijkmans et al. utilized a solution-based method to insert could lead to an amorphous phase, incredible slow crystalliza-
Sn into dealuminated zeolite BEA.28–30 In all post-synthesis tion and non-framework Sn as shown later in this study.
methods, Sn was incorporated into the defects generated by
Herein, a direct and fluoride-free synthesis route is develo-
the dealumination process. Condensation of various ped for Sn-BEA. We demonstrated that an active Sn-BEA cata-
Sn sources with silanol defects resulted in the formation of lyst can be directly synthesized by the DGC method without
tetrahedrally coordinated Sn sites. The Sn concentration in the using any fluoride containing chemicals. The use of crystalline
samples can be easily tailored by controlling the amount of seeds and inorganic cations for charge compensation is essen-
the Sn source applied in the treatment. The materials prepared tial for the crystallization of Sn-BEA in the absence of fluoride.
from the post-synthesis approaches show interesting catalytic The inorganic cations in the as-synthesized sample must be
performances in Baeyer–Villiger oxidation, isomerization of ion-exchanged with the ammonium ions in order to maintain
triose sugar (such as dihydroxyacetone to lactates in alcohols) the crystalline structure of the formed catalysts during calcina-
and glucose. However, due to the low mobility of Sn sources tion. The synthesized Sn-BEA exhibits a hydrophilic nature
used in the syntheses and the lack of ability to control the type compared to those synthesized in the presence of fluoride.
and amount of defects generated by the dealumination
process, it is a challenge to completely diminish all defects in
the dealuminated zeolites by the incorporation of framework
Sn. This is likely one of the reasons why a very different cata-
lytic activity was observed for the Sn-BEA catalysts prepared by
Experimental
Synthesis of Sn-BEA by the DGC method
the post-synthesis methods.
A dry gel conversion method was used to synthesize Sn-BEA in
To the best of our knowledge, direct synthesis of the Sn- the absence of fluoride. In a typical synthesis, 0.0448 g of
BEA catalyst in a fluoride-free system has not been achieved so sodium hydroxide (NaOH, Fisher) was dissolved in 3.09 g of
far. This is mainly due to the difficulties in the crystallization 35 wt% tetraethylammonium hydroxide (TEAOH, Alfa Aesar or
of siliceous zeolite BEA in the absence of fluoride and incor- ZeoGen SDA 440, SACHEM) solution and 14 g of deionized
poration of tetrahedrally coordinated Sn in zeolite frameworks. water in a plastic vessel followed by adding 1 g of fumed silica
It has been known that during zeolite syntheses fluoride can (Cab-O-Sil M5, Cabot). The obtained solution was magnetically
facilitate the mineralization of silica sources and compensate stirred for 1 h at room temperature. A tin source solution con-
for the positive charges associated with the organic structure taining 0.0548 g of tin tert-butoxide (Sn(OC4H9)4, Aldrich) and
directing agents (OSDAs) employed in the syntheses.31 As a 0.254 mL of 30% hydrogen peroxide aqueous solution (H2O2,
result, the zeolites synthesized in the presence of fluoride Fisher) was mixed for 1 h prior to adding into the solution pre-
exhibit fewer defects and higher crystallinity than those pre- pared previously. The resulting solution was stirred for 2 h at
pared under conventional hydrothermal synthesis conditions. room temperature. A dealuminated zeolite BEA seed solution
In order to synthesize high silica or siliceous zeolites without (4 wt% with respect to fumed silica added, the particle size of
using fluoride, the positive charge of OSDA must be compen- the seed is around 200 nm) was then added, and the solvent in
sated by the defects located in the framework (e.g. uSi– the translucent mixture was evaporated at 80 °C with stirring
O−⋯HO–Siu), which is particularly challenging for relatively in an oil bath for 24 h, leading to a dry gel. The detailed syn-
small OSDAs, such as tetraethylammonium (TEA) used in thesis procedure for the dealuminated zeolite BEA seed solu-
zeolite BEA synthesis. The small OSDAs usually result in tion is reported in our previous work.26 The final composition
an increased number of OSDAs occluded in zeolite channels of the dry gel is SiO2:0.008SnO2:0.22TEA2O:0.034Na2O. Other
(e.g. 6 TEA per unit cell for BEA, while only 4 TPA per unit cell cations were also employed in the synthesis, including potass-
for MFI),32,33 which requires more defects for charge compen- ium hydroxide (KOH, Fisher), lithium hydroxide (LiOH, Alfa
sation and could compete with the formation of stable crystal- Aesar), ammonium nitrate (NH4NO3, Alfa Aesar), ammonium
line phases. This challenge has been partially solved by using hydrogen carbonate (NH4HCO3, Alfa Aesar) and ammonium
the dry gel conversion (DGC) method that was initially develo- hydroxide (NH4OH, Fisher). The obtained dry gel (ca. 2 g) was
ped for the ZSM-5 zeolite and later extended to several other ground into powder and put into a Teflon-lined stainless steel
zeolites including the high silica zeolite BEA.34,35 In the DGC autoclave with a cup volume of ca. 50 mL. A portion of 0.5 mL
method, crystallization of a concentrated gel is carried out deionized water was added into the autoclave in a separate
under saturated steam generated from liquid water which is Teflon cup to avoid direct contact between the dry gel and
not allowed to directly come into contact with the synthesis water. The autoclave was then placed in an oven preheated to
gel. Using this method, zeolite BEA with a wide range of Si/Al 140 °C for crystallization. The as-made sample was washed by
ratios has been successfully synthesized.35 Although incorpor- filtration with 1 L deionized water and dried at 100 °C over-
ation of other heteroatoms in zeolite frameworks by the DGC night (denoted as Sn-BEA-AM). Ion-exchange was carried out
method has also been demonstrated for TS-1 and Ti-BEA,36,37 by treating 0.25 g of as-made zeolites with 25 mL of 1 M
synthesis of Sn-BEA by the DGC method has not been success- ammonium nitrate (NH4NO3, J. T. Baker) solution for 2 h at
ful, likely due to the relatively large atomic size of Sn 80 °C. The process was repeated five times. Removal of OSDA
Green Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Green Chemistry
Paper
and ammonium ions was carried out by calcining the powder coupled plasma mass spectrometer (ICP-MS, Thermo Scientific
in a muffle furnace at 550 °C for 12 h at a ramping rate of 1 °C XSERIES 2).
min−1 (denoted as Sn-BEA-IE). Synthesis of Sn-BEA (denoted
as Sn-BEA-F, Si/Sn = 125) in a fluoride medium was achieved
by using a method reported in our previous study.26 Extra-
framework SnO2 in the siliceous zeolite BEA sample (SnO2/
Si-BEA, Si/Sn = 125) was prepared according to the literature.38
Detailed procedures are available in ESI.†
Catalytic activity tests
The catalytic activity of the obtained Sn-BEA catalysts was
studied for sugar isomerization including the conversion of
glucose into fructose and conversion of pyruvaldehyde (PA)
into lactic acid (LA) in water. For glucose isomerization, 0.8 g
of a 10 wt% aqueous glucose (Acros) solution was mixed with
0.068 g of the catalyst in a closed glass vial. After the reaction
mixture was heated at 95 °C for 15 min on an aluminum
heating block, it was cooled down in ice water and filtered
with a 0.2 μm syringe filter. Conversion of PA in water was
carried out with 0.056 g of PA solution (40 wt% in water,
Aldrich) in 1 g of water with 0.02 g of the catalyst at 70 °C for
15 min. The filtrate was then analyzed on a HPLC system
(LC-20C Shimadzu) equipped with a refractive index (RI) detec-
tor. Separation of sugar isomers was carried out by running
the sample through an HPX-87H proton type column (BioRad)
at 30 °C at a flow rate of 0.6 mL min−1 with 0.005 M sulphuric
acid (H2SO4, Fisher). Turnover frequencies (TOF) of the cata-
lysts were calculated based on the product formation with a
unit of min−1 on each Sn site (the number of moles of Sn sites
is obtained by elemental analysis):
Characterization
Samples were characterized by X-ray diffraction (XRD), scan-
ning electron microscopy (SEM), transmission electron
microscopy (TEM), nitrogen sorption measurements, thermo-
gravimetric analysis (TGA), diffuse reflectance ultraviolet-
visible spectroscopy (DR UV-Vis), FT-IR spectroscopy and
inductively coupled plasma mass spectroscopy (ICP-MS). XRD
measurements were performed on a diffractometer (X’Pert Pro,
PANalytical) using Cu Kα radiation generated at 45 kV and
40 mA in the 2 theta range of 4° to 40° with a step size of
0.02°. SEM was performed on a Magellan 400 (FEI) with an
accelerating voltage of 3.0 kV, and the samples were coated
with platinum/palladium alloy prior to observation. TEM
images were recorded on a Joel JEM-2000FX (200 kV). Nitrogen
adsorption–desorption isotherms were collected on an auto-
mated gas sorption analyzer (Autosorb iQ2, Quantachrome) at
77 K. The samples were outgassed at 300 °C for 12 h under
vacuum before the measurement. The thermogravimetric ana-
lysis (TGA) was performed on a thermogravimetric analyzer
(SDT600, TA) at a ramping rate of 10 °C min−1 under helium
flow. Prior to the analysis, the samples were brought into
contact with liquid water and dried in a convection oven at
70 °C overnight. DR UV-Vis measurements were conducted on
a Cary 50 spectrometer with a diffuse reflectance attachment.
Barium sulphate was used as a background material. The
FT-IR experiment was carried out using a Praying Mantis™
DRIFT accessory (Harrick) in a high-temperature reaction cell
(Harrick) allowing control of the atmosphere, temperature and
adsorption of probe molecules over the zeolite samples. The
samples were pretreated with helium flow at 550 °C for 1 h
prior to adsorption (pyridine or deuterated acetonitrile). For
FT-IR measurements of samples with adsorbed pyridine,
helium was used to carry pyridine (Alfa Aesar) vapor for bring-
ing it into contact with the samples at 120 °C for 10 min. After
heating up to 250 °C for 1 h to desorb the weakly adsorbed pyr-
idine, the spectrum was recorded on a FT-IR spectrometer
(Equinox 55, Bruker) equipped with an MCT detector with a
resolution of 2 cm−1 at 120 °C. For deuterated acetonitrile
(CD3CN, Aldrich) adsorption, it was adsorbed on the samples
at 30 °C under flowing helium for 30 min. CD3CN was then
desorbed at 100 °C for various periods of time. The spectra
were acquired at 30 °C and normalized with the Si–O–Si over-
tones at 1800–2000 cm−1. Potassium bromide (KBr, Acros) was
used as the background in this case. Elemental analysis was
carried out in the Analytical Geochemistry Lab, Dept. of Earth
ðmoles of product formedÞ
TOFðminꢀ1Þ ¼
ðmoles of Sn in the catalystÞ ꢁ ðreaction timeÞ
Results and discussion
Synthesis parameters for Sn-BEA by the DGC method
Synthesis of Sn-BEA by the DGC method was first performed
with and without adding zeolite BEA seeds. As shown in Fig. 1,
characteristic XRD peaks corresponding to the BEA topology
are observed for the sample synthesized by adding zeolite BEA
Fig. 1 XRD patterns of the samples prepared with and without using
Sciences in the University of Minnesota using an inductively zeolite BEA seeds after dry gel conversion at 140 °C for 5 days.
This journal is © The Royal Society of Chemistry 2015
Green Chem.

View Article Online
Paper
Green Chemistry
seeds after 5 days of crystallization, revealing that highly crys- mixture. Different alkali ions (Li+ and K+) were also tested for
talline BEA zeolite has been synthesized without any impurity the crystallization of Sn-BEA by the DGC method. The XRD pat-
phase. Crystallization is apparently slow in the absence of terns from the obtained samples suggest that the zeolite BEA
zeolite seeds, and an unidentifiable reflection peak dominates phase can be obtained in all cases (Fig. 2(b)). The result clearly
the spectrum after 5 days of crystallization. Longer crystalliza- indicates that the alkali ions do not affect the crystalline struc-
tion time (20 days) does not yield a pure BEA phase. These ture of the formed zeolites even though the size of these ions
results indicate that adding zeolite BEA seeds not only facili- is very different. It is believed that the role of the alkali ions is
tates the crystallization but also directs the formation of the to compensate for the charge from a large number of frame-
desired crystal structure under the present synthesis con- work defects (uSi–O−⋯H–O–Siu) caused by the relatively
ditions. Thus, it is concluded that the seeding approach plays high charge/volume ratio of TEA+. This observation is consist-
an essential role in the crystallization of the pure Sn-BEA ent with the previous studies on the formation of high silica
phase from a dry gel. Zeolite seeds were used in all the synth- zeolite BEA in the presence of TEA+ and Na+.35,39
eses in this study later.
Fig. 3(a) reveals the XRD patterns of as-made Sn-BEA
It is also found that the use of alkali ions is crucial for the (Sn-BEA-AM), Sn-BEA after direct calcination (Sn-BEA-DC) and
crystallization of Sn-BEA via the DGC method. As revealed in Sn-BEA after ion-exchange and calcination (Sn-BEA-IE).
Fig. 2(a), when the Na2O/SiO2 ratio in the initial gel is lower Sn-BEA-AM shows typical diffraction peaks corresponding to
than 0.034, the crystallinity of the obtained samples is signifi- the zeolite BEA phase. However, much weaker peaks are
cantly lower than the samples prepared with a higher concen- observed in the XRD pattern of Sn-BEA-DC which is prepared
tration of Na+ after 5 days of crystallization.
by direct calcination without ion-exchange, indicating that the
Only very weak reflections can be observed in the samples crystalline structure of the sample was significantly damaged.
when no alkali ion was added (Fig. S1†), and the signals are The partial collapse of the crystalline structure during calcina-
likely attributed to the seed crystals present in the synthesis tion might be attributed to the inherent defects (uSiO−)
Fig. 2 XRD patterns of the samples synthesized with (a) different Na
contents and (b) different alkali ions. The crystallization was carried out
Fig. 3 (a) XRD patterns of Sn-BEA-AM, Sn-BEA-IE, and Sn-BEA-DC and
at 140 °C for 5 days, followed by ion-exchange and calcination.
(b) nitrogen sorption isotherms of Sn-BEA-DC, Sn-BEA-IE and Sn-BEA-F.
Green Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Green Chemistry
Paper
generated for compensating for the positive charges from the
TEA+ ions.
Because of no aluminum in the zeolite framework, positive
charges from the TEA+ occluded within Sn-BEA samples can
generate much more framework defects (uSiO−) than Al-BEA.
Formation of the BEA crystalline phase under these conditions
thus requires the use of alkali ions (e.g. Na+, K+ and Li+) to
compensate for the additional defects as shown previously. It
is interesting that even though the crystalline phase can be
achieved using Na+ in the synthesis solution, the uSi–O–Na
structure could not undergo condensation with adjacent
defect sites (uSi–O–H or uSi–O–Na) to form uSi–O–Siu
during calcination. As a result, the crystalline structure col-
lapsed after the removal of TEA+ (Sn-BEA-DC) by calcination.
Interestingly, it was found that the crystalline structure can be
Fig. 4 (a) SEM and (b) TEM images of Sn-BEA-IE synthesized by the
DGC method.
also attempted to replace the alkali ions (e.g. Na+, K+, Li+) by
+
NH4 ions in the synthesis mixture as the charge compensator
+
well retained by the ion-exchange of Na+ with NH4 before the
to avoid the ion-exchange process. However, no crystalline
phase was achieved under the synthesis conditions (Fig. S2†).
The ineffectiveness of the NH4 ions for stabilizing the zeolite
structure during crystallization might be due to the steric
hindrance with the TEA+ ions within the zeolite BEA
framework.
calcination. The Sn-BEA-IE sample prepared by the ion-
exchange of Sn-BEA-AM with an ammonium nitrate (NH4NO3)
solution and subsequent calcination exhibits a well-crystalline
BEA structure (Fig. 3(a)). It is believed that the Na+ and some
+
TEA+ were replaced by NH4 during the ion-exchange. Upon
+
Nitrogen sorption isotherms (Fig. 3(b)) also support the
high crystallinity of the Sn-BEA-IE sample with a micropore
volume (Vmic) of 0.177 cm3 g−1 (calculated from the t-plot),
which is comparable to the Sn-BEA-F prepared in a fluoride
further heat treatment, uSi–O–NH4 could be transformed
into uSi–O–H. Subsequent dehydration/condensation of the
silanol groups could lead to the formation of the uSi–O–Siu
bond, which helps in retaining the crystallinity and micro-
porous structure of Sn-BEA (Scheme 1). This result clearly
suggests that exchanging alkali ions from the zeolite samples
medium (Vmic = 0.185 cm3
g
−1). However, the sample after
direct calcination, Sn-BEA-DC, possesses only limited accessi-
ble micropores (Vmic = 0.033 cm3 g−1) due to the destruction of
the zeolite structure. In addition, the hysteresis loop appearing
on the Sn-BEA-IE sample from P/P0 ∼0.5–0.8 indicates the
existence of mesopores. It might be due to the surface rough-
ness and/or the voids created from the intergrowth of small
crystal domains as shown in the electron microscopy images
(Fig. 4). The result is also consistent with the higher external
+
by NH4 is indispensable for obtaining the highly crystalline
Sn-BEA catalyst by the synthesis route used in this study. The
method might also be applicable for retaining the crystallinity
of other high silica or siliceous zeolites that have a large
number of OSDAs and defects in the as-made forms. It was
surface area of the Sn-BEA-IE sample (Sext = 160 m2 g−1
;
the t-plot method) compared with that in Sn-BEA-F (Sext
=
121 m2 g−1).
Fig. 4 shows the SEM and TEM images of the Sn-BEA-IE
sample. Discrete truncated bipyrimidal crystals with a size of
around 400–700 nm can be observed in the sample (Fig. S3†).
The Sn-BEA-F catalyst synthesized in a fluoride medium
(Fig. S4†) also shows truncated bipyrimidal crystals (600 nm to
1 μm) but with highly intergrown morphology. The images
suggest that Sn-BEA-IE consists of more discrete and smaller
crystal domains than conventional Sn-BEA, which is consistent
with its higher surface area obtained from N2 sorption exper-
iments. In addition, lattice fringes can be clearly observed in
the TEM image (Fig. 4(b)), confirming that the sample is a
highly crystalline material. Based on the above observations, it
is concluded that Sn-BEA-IE features a well-crystalline structure
and a large surface area. The elemental analysis shows that Sn-
BEA-IE has a Si/Sn ratio of 102.
Characterization of Sn coordination
+
A series of characterization techniques were used to study the
structures of the Sn site in the synthesized Sn-BEA samples.
Scheme 1 Illustrative representation of the effect of NH4 ion-
exchange on the crystal structure of the final product.
This journal is © The Royal Society of Chemistry 2015
Green Chem.

View Article Online
Paper
Green Chemistry
Fig. 6 FT-IR spectrum of the adsorbed pyridine on Sn-BEA-IE. BA:
Brønsted acid. LA: Lewis acid. The spectrum was acquired after deso-
rbing pyridine at 250 °C for 1 h.
Fig. 5 DR UV-Vis spectra for Sn-BEA-IE and Sn-BEA-DC. The spectrum
of Sn-BEA-DC was shifted by 0.05 unit for clarity.
UV-Vis spectroscopy has been shown to be an effective method
to distinguish the coordination environment of Sn in zeo-
lites.28,38,40 The absorbance at ∼200–210 nm corresponds to
tetrahedrally coordinated Sn (framework Sn), the absorbance
at ∼240 nm results from extra-framework Sn species, and the
band at ∼280 nm is assigned to Sn in tin dioxide.38 Fig. 5
shows the UV-Vis spectra for Sn-BEA-IE and Sn-BEA-DC
samples. As can be seen in the plot, the intensified absorbance
at ∼200 nm suggests that most Sn in Sn-BEA-IE are located in
the framework position (SnO4 tetrahedral), while there are
several different coordination states of Sn observed in the Sn-
BEA-DC sample. It is plausible that multiple Sn species are
formed after the zeolite structure collapses during calcination.
The result also indicates that the ion-exchange/calcination
treatment can preserve the micropore structures of the zeolite
samples and the local environment of Sn. The FT-IR spectrum
of the Sn-BEA-IE sample with the adsorbed pyridine (Fig. 6)
might originate from the structural defects of the sample
(uSi–OH). An FT-IR study of CD3CN adsorption was per-
formed to assess the Lewis acidity of Sn-containing
materials.41–43 The FT-IR spectra of the adsorbed CD3CN for
different Sn-BEA samples, including Sn-BEA-IE, Sn-BEA-F and
SnO2/Si-BEA, are shown in Fig. 7. The data for Sn-BEA-IE after
different desorption times at 100 °C are shown in Fig. 7(a).
The band at 2263–2266 cm−1 can be assigned to the physically
adsorbed CD3CN, while the band at 2276 cm−1 reflects the
interaction of CD3CN and silanol groups, and the band around
2310 cm−1 is attributed to the vibration related to framework
Sn species.41,43 It has been proposed that the v(C–N) vibrations
around 2310 cm−1 can be used to distinguish the “closed”
(2308 cm−1) or “open” (hydrolyzed; 2316 cm−1) Sn sites in the
Sn-BEA framework.41 Interestingly, our data reveal a main
band at 2310 cm−1, and no other band at higher wavenumbers
is observed during the course of CD3CN desorption, even
when the desorption temperature was 150 °C. In contrast, two
bands (2308 cm−1 and 2317 cm−1) can be clearly observed
during the desorption of CD3CN for Sn-BEA-F (Fig. 7(b)),
which is consistent with previous literature.41 In addition, the
reveals that the material exhibits Lewis acidity (1452 cm−1
)
corresponding to the interaction between the isolated Sn sites
in the zeolite framework and pyridine.9 Moderate Brønsted
acidity (1546 cm−1) was also observed in the spectrum, which
Fig. 7 FT-IR spectra of the adsorbed CD3CN on (a) Sn-BEA-IE, (b) Sn-BEA-F and (c) SnO2/Si-BEA. The spectra were acquired after saturation of
CD3CN on the samples followed by different period desorption times at 100 °C, as specified in the figures.
Green Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Green Chemistry
Paper
spectra from the sample intentionally prepared with SnO2 as
an extra-framework Sn site did not show absorption in the
2305–2317 cm−1 range (SnO2/Si-BEA, Fig. 7(c); see Fig. S5 for
its XRD pattern). These results clearly indicate that the
Sn-BEA-IE sample prepared from the DGC method indeed pos-
sesses framework Sn, while the local Sn environment could be
different from the Sn-BEA-F sample prepared in the fluoride
medium. It should be noted that a recent report suggests that
the two vibration bands can also appear due to the different
packing structures of CD3CN rather than the difference in Sn
site structures.43 In addition, the locations and distribution of
Sn in zeolite frameworks might also affect their FT-IR spectra.
At this point, it can be concluded that Sn-BEA-IE does have a
framework Sn site, however the local environment of Sn and
its difference from the Sn-BEA-F are still elusive, and require
further investigation.
To further confirm the coordination environment of Sn in
the synthesized Sn-BEA-IE sample, glucose with deuterium
substituted at the C-2 position was used in the isomerization
reaction.44 The 1H-NMR spectra for the unlabelled fructose
and fructose isomerized from the labelled glucose are shown
in ESI (Fig. S6†). The absence of resonances at δ = 3.4–3.5 ppm
corresponding to the protons at the C-1 position in the
1H-NMR spectrum of fructose after the reaction indicates that
the reaction only undergoes an intramolecular hydride shift
from C-2 to C-1 in water, which follows the same mechanism
as the Sn-BEA-F.44 The findings from the 1H-NMR measure-
ment are consistent with the DR UV-Vis and CD3CN FT-IR
measurements, showing that most Sn are incorporated into
the zeolite framework with a tetrahedral coordination and are
able to catalyze the glucose isomerization reaction via an intra-
molecular hydride shift pathway. With these characterization
results, it is concluded that the Sn-BEA-IE synthesized in this
study possesses framework Sn in the siliceous BEA framework,
and the isolated metal centers exhibit Lewis acidity which
makes it capable of catalyzing the sugar isomerization in the
aqueous phase.
Fig. 8 (a) TGA curves and (b) FT-IR spectra for Sn-BEA-IE and Sn-
BEA-F. The samples were brought into contact with liquid water and
dried at 70 °C overnight prior to TG measurements. FT-IR spectra were
acquired at 30 °C after degassing at 550 °C for 1 h. The band at
3745 cm−1 represents free terminal silanols, and the bands at 3738 and
3701 cm−1 may be related to weakly hydrogen bonded terminal silanols.
clearly suggest that Sn-BEA-IE adsorbs more water (4% vs. 2%
weight loss at 400 °C), which implies a more polar surface of
Sn-BEA-IE due to the defective structure (Si–OH), and that it
shows more hydroxyl condensation at higher temperature
(∼2% vs. 0.5% weight loss above 450 °C) than Sn-BEA-F. The
FT-IR spectra in the OH stretching region for degassed
samples are shown in Fig. 8(b). Intensified absorption bands
can be observed at 3738 and 3745 cm−1, which can be assigned
to terminal silanol groups (weakly hydrogen bonded and free
Si–OH, respectively).47 A small shoulder at 3701 cm−1 corres-
ponding to weakly hydrogen bonded silanols in zeolites is also
observed for Sn-BEA-IE. By directly comparing the normalized
spectra, it is obvious that Sn-BEA-IE possesses more silanol
groups than Sn-BEA-F does, consistent with the TGA results.
These results indicate that Sn-BEA-IE has more defects and
more hydrophilic surface than the sample prepared in the
presence of fluoride.23,47
Catalytic properties
It has been suggested that the hydrophobic environment is a
key factor for adsorbing glucose from the aqueous phase into
the hydrophobic Sn-BEA catalyst and catalyzing the isomeriza-
tion reaction from glucose to fructose.23,45,46 The Sn-BEA cata-
lyst synthesized by the DGC method (Sn-BEA-IE) is expected to
have more silanol defects than the sample prepared in a fluor-
ide medium, and it exhibits a different hydrophobicity from
Sn-BEA-F. This is because the charges of the occluded TEA+
ions in the Sn-BEA-IE synthesized in a caustic medium must
be compensated by framework defects (uSi–O−), while in Sn-
BEA-F, F− can form a pair with TEA+ and generates fewer
defects. As discussed in the previous paragraph, the framework
defects (uSi–O−) can be reduced by reacting with the adjacent
silanol groups in the Sn-BEA-IE sample during ion-exchange
and calcination processes. However, Sn-BEA-F still has fewer
silanol groups (uSi–OH) than Sn-BEA-IE, as shown in TGA
and FT-IR spectra (Fig. 8). The TG experiments (Fig. 8(a))
The catalytic activity of the Sn-BEA-IE sample was examined
by glucose isomerization and conversion of PA into lactic acid
and compared with hydrophobic Sn-BEA-F. The reaction
results are summarized in Table 1. The product distributions
This journal is © The Royal Society of Chemistry 2015
Green Chem.

View Article Online
Paper
Green Chemistry
Table 1 Turnover frequencies (TOF, min−1) of Sn-BEA catalysts for
sugar isomerization reactions
pyruvaldehyde reaction in the aqueous phase. The difference
in catalytic activity could be attributed to the different hydro-
phobic nature and local Sn environments in the catalysts pre-
pared by different synthesis routes. The current work provides
an alternative and more environmentally friendly approach to
synthesize the Sn-BEA catalyst.
Sn-
BEA-IE
Sn-
BEA-F
Substrate
Reaction scheme
Pyruvaldehydea
1.20
0.25
1.52
0.99
Glucoseb
Acknowledgements
The authors acknowledge the financial support from the
Center for Catalysis and Energy Innovation, an Energy Frontier
Research Center funded by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences under award
number DE-SC0001004. We thank Prof. Gonghu Li in the
University of New Hampshire for his kind help in the DR
UV-Vis measurements.
a 70 °C, 15 min, 20 mg catalyst with 1.06 g of 2.1 wt% PA solution
b 95 °C, 15 min, 68 mg catalyst with 0.80 g of 10 wt% glucose solution.
for different reactions are available in ESI (Tables S1 and S2†).
As expected, the TOF for the isomerization of glucose into fruc-
tose in the aqueous phase on Sn-BEA-IE is only ∼25% of that
of Sn-BEA-F. The low reaction rate is likely due to the hydro-
philic surface of the Sn-BEA-IE catalyst, which is in good agree-
ment with previous literature.45 On the other hand, the activity
of Sn sites in Sn-BEA-IE for the conversion of PA into lactic
acid (LA) (1.20 min−1) is still lower than that in Sn-BEA-F
(1.52 min−1), but the difference is much less significant com-
pared to the glucose reaction. The production of LA is similar
for the two catalysts under the reaction conditions (18.3% vs.
18.9% yield). It has been known that the conversion from PA
to LA involves hydration of PA in the aqueous phase followed
by intramolecular 1, 2 hydride-shift reaction in the Lewis acid
sites of Sn-BEA.8 Hence, it is reasonable that the hydrophilic
material shows improved activity due to the enhanced adsorp-
tion of water in micropores. However, we must point out that
the local environment of Sn sites including “open/closed
sites”, their distribution and location might be different for
the two Sn-BEA catalysts as revealed in the CD3CN adsorbed
FT-IR study and this could also affect the catalytic activity of
the materials.9,48 Although it is evident that most Sn atoms are
incorporated in the framework, the discrepancies in activities
for different reactions reported here might result from the
combinatorial effect of the active site nature and surface
properties (hydrophobicity).
References
1 J. Čejka, G. Centi, J. Perez-Pariente and W. J. Roth, Catal.
Today, 2012, 179, 2–15.
2 M. Stöcker, Angew. Chem., Int. Ed., 2008, 47, 9200–9211.
3 M. A. Snyder and M. Tsapatsis, Angew. Chem., Int. Ed.,
2007, 46, 7560–7573.
4 S. G. Wettstein, D. M. Alonso, E. I. Gürbüz and
J. A. Dumesic, Curr. Opin. Chem. Eng., 2012, 1, 218–224.
5 Y. Zhu, G. Chuah and S. Jaenicke, Chem. Commun., 2003,
2734–2735.
6 V. Choudhary, A. B. Pinar, S. I. Sandler, D. G. Vlachos and
R. F. Lobo, ACS Catal., 2011, 1, 1724–1728.
7 M. Moliner, Y. Román-Leshkov and M. E. Davis, Proc. Natl.
Acad. Sci. U. S. A., 2010, 107, 6164–6168.
8 E. Taarning, S. Saravanamurugan, M. S. Holm, J. M. Xiong,
R. M. West and C. H. Christensen, ChemSusChem, 2009, 2,
625–627.
9 A. Corma, M. E. Domine and S. Valencia, J. Catal., 2003,
215, 294–304.
10 A. Corma, L. T. Nemeth, M. Renz and S. Valencia, Nature,
2001, 412, 423–425.
11 T. Blasco, M. A. Camblor, A. Corma, P. Esteve, A. Martinez,
C. Prieto and S. Valencia, Chem. Commun., 1996, 2367–
2368.
12 P. Concepción, Y. Pérez, J. C. Hernández-Garrido,
M. Fajardo, J. J. Calvino and A. Corma, Phys. Chem. Chem.
Phys., 2013, 15, 12048–12055.
Conclusions
Highly crystalline Sn-BEA with tetrahedrally coordinated Sn
was successfully synthesized in a non-fluoride medium for the
first time by a dry gel conversion method. The sample exhibits 13 C. Paris, M. Moliner and A. Corma, Green Chem., 2013, 15,
a higher surface area than the conventional Sn-BEA syn- 2101–2109.
thesized in a fluoride medium. It is found that the use of seed 14 C. M. Lew, N. Rajabbeigi and M. Tsapatsis, Ind. Eng. Chem.
crystals, addition of alkali ions and ion-exchange with Res., 2012, 51, 5364–5366.
ammonium nitrate before calcination are essential for obtain- 15 E. Nikolla, Y. Román-Leshkov, M. Moliner and M. E. Davis,
ing a highly crystalline Sn-BEA catalyst. The Sn-BEA catalyst ACS Catal., 2011, 1, 408–410.
shows a lower catalytic activity than the conventional Sn-BEA 16 T. J. Schwartz, S. M. Goodman, C. M. Osmundsen,
for glucose isomerization and a comparable activity for
E. Taarning, M. D. Mozuch, J. Gaskell, D. Cullen,
Green Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Green Chemistry
Paper
P. J. Kersten and J. A. Dumesic, ACS Catal., 2013, 2689– 33 V. Shen and A. T. Bell, Microporous Mater., 1996, 7, 187–
2693. 199.
17 W. R. Gunther, Q. Duong and Y. Román-Leshkov, J. Mol. 34 W. Xu, J. Dong, J. Li, J. Li and F. Wu, J. Chem. Soc., Chem.
Catal. A: Chem., 2013, 379, 294–302. Commun., 1990, 755–756.
18 M. Boronat, P. Concepción, A. Corma and M. Renz, Catal. 35 P. R. Hari Prasad Rao, K. Ueyama and M. Matsukata, Appl.
Today, 2007, 121, 39–44. Catal., A, 1998, 166, 97–103.
19 A. Corma, M. E. Domine, L. Nemeth and S. Valencia, J. Am. 36 X. Ke, L. Xu, C. Zeng, L. Zhang and N. Xu, Microporous
Chem. Soc., 2002, 124, 3194–3195. Mesoporous Mater., 2007, 106, 68–75.
20 J. N. Chheda, G. W. Huber and J. A. Dumesic, Angew. 37 M. Ogura, S.-i. Nakata, E. Kikuchi and M. Matsukata,
Chem., Int. Ed., 2007, 46, 7164–7183. J. Catal., 2001, 199, 41–47.
21 M. S. Holm, S. Saravanamurugan and E. Taarning, Science, 38 R. Bermejo-Deval, R. Gounder and M. E. Davis, ACS Catal.,
2010, 328, 602–605. 2012, 2, 2705–2713.
22 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 39 M. Matsukata, T. Osaki, M. Ogura and E. Kikuchi, Micro-
2411–2502.
porous Mesoporous Mater., 2002, 56, 1–10.
23 R. Gounder and M. E. Davis, AIChE J., 2013, 59, 3349–3358.
24 S. Valencia and A. Corma, U.S. Patent 5,968,473, 1999.
40 H. Y. Luo, L. Bui, W. R. Gunther, E. Min and Y. Román-
Leshkov, ACS Catal., 2012, 2, 2695–2699.
25 Z. Kang, X. Zhang, H. Liu, J. Qiu and K. L. Yeung, Chem. 41 M. Boronat, P. Concepción, A. Corma, M. Renz and
Eng. J., 2013, 218, 425–432. S. Valencia, J. Catal., 2005, 234, 111–118.
26 C.-C. Chang, Z. Wang, P. Dornath, H. J. Cho and W. Fan, 42 C. M. Osmundsen, M. S. Holm, S. Dahl and E. Taarning,
RSC Adv., 2012, 2, 10475–10477. Proc. R. Soc. London, Ser. A, 2012, 468, 2000–2016.
27 Fluoride: dose–response analysis for non-cancer effects, 43 S. Roy, K. Bakhmutsky, E. Mahmoud, R. F. Lobo and
http://water.epa.gov/action/advisories/drinking/fluoride_index.
cfm, accessed February, 2015.
28 P. Li, G. Liu, H. Wu, Y. Liu, J.-g. Jiang and P. Wu, J. Phys.
Chem. C, 2011, 115, 3663–3670.
R. J. Gorte, ACS Catal., 2013, 3, 573–580.
44 Y. Román-Leshkov, M. Moliner, J. A. Labinger and
M. E. Davis, Angew. Chem., Int. Ed., 2010, 49, 8954–
8957.
29 C. Hammond, S. Conrad and I. Hermans, Angew. Chem., 45 R. Gounder and M. E. Davis, J. Catal., 2013, 308, 176–
Int. Ed., 2012, 51, 11736–11739. 188.
30 J. Dijkmans, D. Gabriels, M. Dusselier, F. de Clippel, 46 C. Buttersack, W. Wach and K. Buchholz, J. Phys. Chem.,
P. Vanelderen, K. Houthoofd, A. Malfliet, Y. Pontikes and
B. F. Sels, Green Chem., 2013, 15, 2777–2785.
31 A. Corma and M. E. Davis, ChemPhysChem, 2004, 5, 304–
313.
1993, 97, 11861–11864.
47 T. Blasco, M. A. Camblor, A. Corma, P. Esteve, J. M. Guil,
A. Martinez, J. A. Perdigon-Melon and S. Valencia, J. Phys.
Chem. B, 1998, 102, 75–88.
32 F. Vaudry, F. Di Renzo, P. Espiau, F. Fajula and P. Schulz, 48 G. Yang, E. A. Pidko and E. J. M. Hensen, J. Phys. Chem. C,
Zeolites, 1997, 19, 253–258.
2013, 117, 3976–3986.
This journal is © The Royal Society of Chemistry 2015
Green Chem.