Synthesis and catalytic testing of Lewis acidic nano zeolite Beta for epoxide ring opening with alcohols

Article abstract of DOI:10.1016/j.apcata.2019.03.009

Lewis acidic zeolites are robust catalytic materials that are capable of a range of important chemistry for fine chemical production, including the epoxide ring opening with alcohols. The crystalline structure imposes diffusion limitations for large molecules that are overcome through reducing the particle size to produce nano zeolite beta substituted with Lewis acidic tin (nSnBeta). nSnBeta is characterized using standard methods to demonstrate that tin is efficiently incorporated into the zeolite framework with open and closed sites similar to the micron-sized conventional Sn-Beta (cSnBeta). While nSnBeta exhibits comparable catalytic activity to cSnBeta for reactions involving small substrates such as epichlorohydrin and methanol, an improvement in catalytic performance is observed for nSnBeta relative to cSnBeta when using large substrates such as 1,2-epoxyoctane and ethanol. nSnBeta and cSnBeta convert a cyclic epoxide too large to enter the pores to similar extents, indicating that the improved performance is not associated with an increased number of catalytic sites on the external surface. Catalyst reuse experiments demonstrate that organic accumulation on the material reduces catalytic activity that can be partially restored through calcination. Overall, the results demonstrate that nano zeolite beta can be synthesized with small particle sizes (i.e., 150 nm) to help overcome diffusion limitations for bulky substrates for the alcohol ring opening of epoxides.

Full text of DOI:10.1016/j.apcata.2019.03.009

Applied Catalysis A, General 577 (2019) 28–34
Contents lists available at ScienceDirect
Applied Catalysis A, General
journal homepage: www.elsevier.com/locate/apcata
Synthesis and catalytic testing of Lewis acidic nano zeolite Beta for epoxide
ring opening with alcohols
⁎
Aamena Parulkar, Alexander P. Spanos, Nitish Deshpande, Nicholas A. Brunelli
Ohio State University, William G. Lowrie Department of Chemical and Biomolecular Engineering, 151 W. Woodruff Ave., Columbus, OH 43210, United States
A R T I C L E I N F O
A B S T R A C T
Keywords:
Lewis acidic zeolites are robust catalytic materials that are capable of a range of important chemistry for fine
chemical production, including the epoxide ring opening with alcohols. The crystalline structure imposes dif-
fusion limitations for large molecules that are overcome through reducing the particle size to produce nano
zeolite beta substituted with Lewis acidic tin (nSnBeta). nSnBeta is characterized using standard methods to
demonstrate that tin is efficiently incorporated into the zeolite framework with open and closed sites similar to
the micron-sized conventional Sn-Beta (cSnBeta). While nSnBeta exhibits comparable catalytic activity to
cSnBeta for reactions involving small substrates such as epichlorohydrin and methanol, an improvement in
catalytic performance is observed for nSnBeta relative to cSnBeta when using large substrates such as 1,2-
epoxyoctane and ethanol. nSnBeta and cSnBeta convert a cyclic epoxide too large to enter the pores to similar
extents, indicating that the improved performance is not associated with an increased number of catalytic sites
on the external surface. Catalyst reuse experiments demonstrate that organic accumulation on the material
reduces catalytic activity that can be partially restored through calcination. Overall, the results demonstrate that
nano zeolite beta can be synthesized with small particle sizes (i.e., 150 nm) to help overcome diffusion limita-
tions for bulky substrates for the alcohol ring opening of epoxides.
Alcohol ring opening of epoxides
Nano zeolites
Lewis acidic zeolites
Pore-limited diffusion
1
. Introduction
Advances in catalysis using Lewis acidic materials are transforming
limitations within zeolites at the expense of the structural stability of
these materials. In certain cases, the presence of mesopores can also
affect the shape selectivity and therefore precise design of catalysts is
required [24]. An interesting alternative is to reduce the particle size.
With the synthesis of MFI nano-sheets, it has been demonstrated that
nano zeolites can significantly improve the catalytic performance
[25,26].
the ability to produce chemicals in a sustainable manner. Indeed, Lewis
acidic zeolite catalysts containing heteroatoms such as tin (Sn), tita-
nium, zirconium, and hafnium in different zeolites frameworks [1–4]
have been shown to catalyze a variety of industrially important che-
mical reactions such as biomass upgrading [5–7], Meerwein-Ponndorf-
Verley (MPV) [8,9], Baeyer-Villiger (BV) oxidation [10,11], Diels-Alder
chemistry [12], aldol condensation [13,14], epoxidation [4,15], and
ring-opening of epoxides [16,17]. This impressive array of chemistry
can have limited reaction scope because of the small pores associated
with zeolites that can introduce significant internal mass transfer lim-
itations. Therefore, it is an important challenge to design catalytic
materials to overcome these limitations to further expand the scope of
available chemical reactions.
As the catalytic behavior for zeolite catalysts is strongly dependent
on the type of framework, the pore size and network can influence the
selectivity of a chemical reaction. Even with the nanoscale dimensions,
MFI nanomaterials have limited catalytic activity for certain reactions,
given the small pore size (10 membered ring, 10MR). While MFI has a
10MR pore architecture, zeolite beta has a 12MR in the framework,
providing a catalyst structure that can be employed for larger sub-
strates. Therefore, developing methodologies to reduce particle size for
zeolite beta is of importance, especially with Lewis acidic heteroatom
incorporation.
Current research is investigating methods to overcome internal mass
transfer limitations through controlling material synthesis. The primary
goal has been to reduce the internal diffusion path length through ei-
ther creating hierarchical zeolites or nano zeolites [18–23]. Hier-
archical zeolites provide a simple and efficient way to mitigate diffusion
The synthesis procedure can be modified to tune the particle size for
zeolite Beta. While zeolite crystallization is a complex phenomenon and
not yet completely understood, studies have sought to understand the
effect of different synthesis parameters on the crystallization [27–29].
⁎
Corresponding author.
E-mail address: brunelli.2@osu.edu (N.A. Brunelli).
https://doi.org/10.1016/j.apcata.2019.03.009
Received 30 December 2018; Received in revised form 11 March 2019; Accepted 17 March 2019
Available online 18 March 2019
0926-860X/ © 2019 Elsevier B.V. All rights reserved.

A. Parulkar, et al.
Applied Catalysis A, General 577 (2019) 28–34
To reduce particle size for zeolite Beta, several methods have been in-
vestigated, including the addition of seeds, confined space synthesis,
novel structure directing agents (SDAs), and dry-gel interconversion
2. Experimental methods
2.1. Synthesis procedure for nSiBeta and nSnBeta
[
21,30–35]. While these studies demonstrate the benefits of nanosized
zeolite, they are focused on either purely siliceous or aluminosilicate
zeolite beta. Translating this work for synthesizing Lewis acidic zeolites
can be challenging since incorporating heteroatoms like Sn, Zr, and Hf
can significantly influence zeolite crystallization. Indeed, Sn inclusion is
known to slow the crystallization of zeolite beta, especially at high Sn
concentration [36]. The slow crystallization results in larger particle
size as well as increased intergrowth. These micron-sized particles of
Sn-beta can pose diffusion limitations when utilizing bulky substrates
The pure silica form of zeolite beta with nanoscale dimensions
(nSiBeta) is synthesized using the procedure previously reported in
literature [21]. Briefly, the SDA is synthesized as detailed in the sup-
plementary information (SI) according to previous methods [21]. Tet-
raethyl orthosilicate (TEOS; 5.02 g) is added to a round-bottom flask. In
a separate flask, 18.06 g of 9.5 wt% TMP(OH) hydroxide solution is
2
weighed and then added to TEOS slowly with rapid stirring, main-
taining the solution overnight at room temperature to ensure complete
hydrolysis of TEOS. The mixture is evaporated under reduced pressure
on a rotary evaporator with heating to 40 °C to remove excess water and
ethanol. Three cycles of evaporation are performed with addition of
5 mL DI water in between each cycle. After the final cycle, the weight of
synthesis gel is adjusted by adding DI water to achieve a gel composi-
[
37].
Recent studies have demonstrated synthesis of nano-sized Sn-Beta
through both hydrothermal and post-synthetic routes [22,38,39]. Post-
synthetic routes to nano-Sn-Beta utilized either partially crystalline Si-
beta or dealuminated beta as the starting material. While post-synthetic
routes can be efficient, these methods tend to produce materials that are
less active than hydrothermally crystallized zeolites [40]. It has also
been demonstrated that nano Sn-Beta can be produced through hy-
drothermal methods using N-cyclohexyl-N,N-dimethyl cyclohex-
anaminium hydroxide as the SDA [39]. This SDA could be used under
hydroxide-mediated conditions and resulted in a particle size of Sn-Beta
of ∼500 nm. Overall, these methods reduced the particle size below a
micron, but the particles are still larger than 200 nm. Recent work with
nano-Sn-MFI demonstrated the diffusion limitations for the epoxide
ring–opening reaction with 1,2-epoxyhexane and methanol were not
substantially reduced until the particle size is reduced below 200 nm
2
+
tion of 1 SiO
2
:0.15 TMP :25 H O. The synthesis gel is transferred to a
2
Teflon liner and 5 wt% of calcined Si-Beta seeds are added to facilitate
crystallization. The Teflon liner is finally sealed inside an acid digestion
vessel and placed in an oven preheated to 100 °C equipped with rota-
tion at 35 RPM. The material is allowed to crystallize for 5 days after
which the reactor is removed from the oven and quenched using tap
water. The solids are separated by centrifugation at 9000 RPM for
10 min and the eluent is removed. The solids are then washed 5 times
using 40 mL of DI water for each cycle. The washed solids are dried in
an oven at 80 °C overnight followed by calcination at 550 °C (ramp rate
of 2 °C/min) for 10 h in air using a Lindberg Blue M Moldatherm Box
Furnace.
[
17]. Therefore, it is important to investigate additional synthetic
methods that can achieve catalytic particles with a size smaller than
00 nm.
An SDA that has been demonstrated for successful synthesis of
For synthesizing nano zeolite Beta with tin (nSnBeta), the synthesis
2
procedure is modified to include SnCl
4
·5H O as the tin source. While
2
TEOS hydrolyzes in the presence of TMP(OH)
SnCl ·5H O is prepared by dissolving the required amount of tin (de-
pending on desired Si/Sn ratio) source in 1–2 mL of DI water. One hour
2
, a separate solution of
zeolite Beta with ∼140 nm particle size is 4,4′-trimethylenebis(N-me-
thyl, N-benzyl piperidinium) cations (TMP) [21]. However, this SDA
has only been utilized for purely siliceous or aluminum containing
zeolite Beta. The observed reduction in particle size while utilizing TMP
as the SDA can be partly attributed to the basic conditions of the
synthesis gel [41]. Therefore, this makes TMP a possible SDA for ex-
amining synthesis of Lewis acidic nano Beta zeolite.
4
2
after the initial mixing of TMP(OH)
2
and TEOS, the SnCl solution is
4
added to the mixture dropwise under vigorous stirring. The remaining
steps of the procedure are similar to those for pure nSiBeta material.
The final gel composition for nSnBeta materials is 1 SiO :x Sn:0.15
2
2
+
TMP :25 H O, where x = 0.005 or 0.01 for a theoretical Si:Sn of
2
Lewis acidic zeolites, both post-synthetically and hydrothermally
synthesized, including Sn-H-SSZ-13, Sn-Beta, Zr-Beta, and Hf-Beta, have
been shown to catalyze ring-opening of epoxides using amines, water,
and alcohols as the nucleophiles [16,42–44]. Specifically, ring-opening
of epoxides with alcohols results in formation of β-alkoxy alcohols,
which find application as intermediates in pharmaceutical and solvent
industry [45]. The activity for ring-opening of epoxides with alcohols
can be enhanced by using Lewis acidic zeolites [43]. However, the
small pores associated with the zeolite framework may introduce dif-
fusion limitations that necessitate utilization of nano zeolites to enable
conversion of a broad scope of substrates.
200:1 and 100:1. Similar synthesis with fluoride mediated conditions
did not produce a crystalline product even after 90 days of crystal-
lization time.
2.2. Synthesis procedure for conventional Sn-Beta
The synthesis is performed using the procedure reported previously
[43]. Briefly, TEOS (24.48 g) is slowly added to 26.48 g TEAOH (35 wt
% aqueous solution). The mixture is stirred for 90 min until a single-
phase solution is formed. In a separate vial, the tin precursor solution is
prepared by dissolving 0.206 g SnCl
4
·5H O in 1.6 mL DI water. The Sn
2
In this work, the synthesis of nano zeolite Beta substituted with tin
solution is added dropwise to the TEOS/TEAOH mixture. The mixture is
allowed to hydrolyze overnight (20–24 h). The hydrolyzed mixture is
concentrated using a rotovap three times to remove ethanol and some
water, adding 20 g of DI water after each rotovapping cycle to remove
the ethanol completely. DI water (1.28 g) is added to the final synthesis
gel. The synthesis gel is then transferred to a Teflon-lined 200 mL acid-
digestion vessel (Parr Inst. Comp.). Hydrofluoric acid (HF, 51 wt%
aqueous solution; 2.0 mL) and 350 mg of calcined Si-Beta seeds are
(
nSnBeta) is examined using hydrothermal conditions with TMP cations
as the SDA to produce particles less than 200 nm. The incorporation of
Sn at Si to Sn ratio of 100 and 200 is investigated and compared to
traditional hydrothermal synthesis of Sn-Beta. The materials are char-
acterized using a battery of standard techniques to investigate particle
size, tin incorporation efficiency, and the nature of the Sn species. The
nSnBeta catalysts are tested and compared to conventional SnBeta
(
cSnBeta) and nanoSnMFI (nSnMFI) catalysts for epoxide ring opening
added to the synthesis gel and the mixture is stirred using a Teflon rod.
−
reaction of 1,2-epoxyoctane with ethanol. The stability and reusability
of the catalyst is tested by performing recycling experiments. Overall,
the work demonstrates the synthesis of nSnBeta and the catalytic cap-
abilities of these materials to overcome diffusion limitations.
The final gel composition is 1 SiO
2
:0.005 Sn:0.54 F :0.54 TEA:7.5 H
2
O.
The acid digestion vessel is sealed and placed in a preheated oven at
140 °C with rotation at 35 RPM. The material is allowed to crystallize
for 30 days. After the necessary crystallization time, the reactor is
29

A. Parulkar, et al.
Applied Catalysis A, General 577 (2019) 28–34
removed from the oven and quenched under tap water. The formed
solids are filtered and washed with 1 L DI water. The filtered solids are
dried in an oven at 80 °C overnight and then calcined in air at 550 °C for
are characterized for elemental analysis by Galbraith Laboratories using
inductively coupled plasma-optical emission spectroscopy (ICP-OES) to
determine the weight percent of tin in the materials.
1
0 h to remove the SDA.
2.5. Catalytic testing
2.3. Synthesis procedure for nanoSnMFI (nSnMFI-100)
The catalytic testing is performed in a two neck 10 mL round bottom
Nano SnMFI (nSnMFI) is synthesized using a procedure described
(RB) flask equipped with a condenser, a magnetic stir bar, and a
septum. The RB is filled with 2 mL of a solution containing 0.4 M ep-
oxide and diethylene glycol dibutyl ether (DGDE) as an internal stan-
dard in neat alcohol. A sample (40 μL) is taken and diluted with acetone
(∼2 mL) to serve as the initial concentration data point. The required
amount of catalyst is then added to the RB to achieve an epoxide:Sn of
250:1. After adding the catalyst, the RB is immersed in a silicone oil
bath, pre-heated to the desired temperature of 60 °C. This process in-
troduces a brief non-isothermal step into the process that has a negli-
gible effect on the calculated conversion since the steady state tem-
perature is reached in less than two minutes. At specific times, a sample
(40 μL) is withdrawn from the reaction mixture using a reusable
stainless-steel needle and is filtered using a small plug of silica and
diluted with acetone. The samples are analyzed using gas chromato-
graphy (Agilent, 7820A) equipped with flame ionization detector (GC-
FID) and the conversion is computed using the internal standard
method. Product identification is done using GC–MS (equipped with
Agilent HP-5 ms Ultra Inert column) in positive ion mode (electron
ionization).
previously [17]. Briefly, 20.81 g of TEOS is mixed with 18.30 g of
TPAOH and 29.77 g of DI water. In a separate vial, 0.5 g of SnCl
4
·5H O
2
is mixed with 1.0 g of DI water. This solution is added to the TEOS/
TPAOH mixture to give a molar ratio of the precursors of 1 TEOS: 0.02
Sn: 0.36 TPAOH: 23 H O. The synthesis gel is hydrolyzed for 24 h with
2
stirring at room temperature and is then heated by submerging the flask
in an oil bath, pre-heated to the 80 °C. The materials are allowed to
crystallize for 20 days. The crystals are diluted with DI water and se-
parated using centrifugation at 8500 rpm (Beckman Coulter Alegra X-
3
0 centrifuge equipped with F0850 rotor model). After drying over-
night, the materials are calcined at 560 °C for 3 h in air.
2.4. Material characterization
The materials are analyzed using a battery of standard character-
ization techniques, including nitrogen physisorption, powder X-ray
diffraction (PXRD), scanning electron microscopy (SEM), diffuse re-
flectance infrared Fourier transform spectroscopy (DRIFTS), diffuse
reflectance ultra-violet visible spectroscopy (DRUVS), thermogravi-
metric analysis-differential scanning calorimetry (TGA-DSC), and ele-
mental analysis. The textural properties are characterized using a
Micromeritics 3Flex surface characterization analyzer. The samples are
first degassed on a Micromeritics SmartVacPrep sample preparation
2.6. Catalytic material reuse
The recycle experiment to test the reusability of catalyst is per-
formed using the same catalytic testing procedure described above with
the exception that the reaction is scaled up by a factor of four to fa-
cilitate recovering of catalyst for subsequent tests and the test is per-
formed using a 25 mL RB. For each test, the amount of the reaction
mixture is scaled to achieve a ratio of epoxide:Sn of 250:1 (0.4 mol%
Sn). After each test, the catalyst is separated from the reaction mixture
by centrifuging at 8500 RPM for 15 min. The separated catalyst is
further washed using the same alcohol used in reaction (five times the
volume of reaction mixture) and then dried at 90 °C overnight before
using it for a subsequent cycle.
−
3
device at 140 °C under vacuum (10
mmHg) for 16 h followed by in
mmHg). The nitrogen sorption isotherms of
situ degassing of samples on the 3Flex instrument for 4 h at 140 °C
−
5
under vacuum (5 × 10
degassed samples are recorded at liquid nitrogen temperatures
(
∼77 K). The surface area and micropore volume of the materials are
reported using BET and the t-plot method, respectively. The PXRD is
collected using a Bruker powder X-ray diffractometer in flat plate re-
flection, Bragg Brentano optics mode using Cu Kα1–Kα2 radiation (λ =
1
.540 and 1.544 Å) at 40 kV, 40 mA, and room temperature. The par-
ticle size and the morphology of the materials is analyzed using a FEI
Nova 400 NanoSEM. SEM samples are prepared through dispersing the
calcined material (∼2 mg) in 100 μL methanol followed by deposition
on conductive carbon tape. Prior to SEM analysis the samples are
sputter-coated with ∼8 nm thick layer of gold-palladium alloy using
EMS 150 T S sputter coater. DRIFTS is performed using a Nicolet iS50
spectrometer equipped with MCT-A liquid nitrogen cooled detector (32
3. Result and discussion
3.1. Synthesis and material characterization
Initial work demonstrates the successful synthesis of nano-Beta
zeolites at a Si:Sn ratio of 100:1 and 200:1. The materials are char-
acterized using PXRD, demonstrating the successful formation of zeolite
Beta. Compared to the typical hydrothermal synthesis, the PXRD peaks
for nSnBeta are broader than Sn-Beta, consistent with the idea that
small particles are present for nSnBeta. It should be mentioned that the
Sn incorporation lengthens the crystallization time as compared to pure
Si-Beta synthesized using the same procedure. The crystallization time
is determined to be ∼15 days for nSnBeta-200 and > 45 days for
nSnBeta-100. While crystallization time increased, SEM analysis of
nSnBeta materials reveals that the slower crystallization did not result
in a significant increase in particle size for these materials, as shown in
Fig. 1. The particle size and textural properties of all the materials
synthesized are reported in Table 1. The nSnBeta materials exhibit
higher external surface areas as compared to conventional Sn-Beta
(cSnBeta-200). This difference is as expected given the smaller particle
size of nSnBeta. The micropore volume is similar for these materials and
is comparable to that of conventionally synthesized Sn-Beta. For the
nSnBeta materials, the nitrogen physisorption isotherms have a hys-
teresis that can be attributed to interparticle pores that are common in
the synthesis of nanomaterials as shown in Figure S1. The synthesis also
−
1
scans at 2 cm
resolution). The DRIFTS set up includes a Praying
Mantis (Harrick Scientific Products, Inc.) with a high temperature re-
action chamber consisting of zinc selenide (ZnSe) windows. The ma-
terial is initially degassed in situ at 500 °C for 90 min under nitrogen
flow before cooling to 25 °C. Similar to previous work [46], the material
is pulsed with deuterated-acetonitrile using a VICI 6-port valve
equipped with 100 μL sample loop. The probe molecule is allowed to
desorb under nitrogen flow while increasing the temperature from 25 °C
to 125 °C in steps, holding for 10 min at each temperature. The IR
spectra are collected using a background of the degassed material at the
same temperature before dosing. The DRUV–vis spectra are collected on
Evolution 300 UV–vis spectrometer with a resolution of 2 nm at a rate
-
1
of 10 nm s with pure silica analogues of materials as the background.
®
TGA is performed on a STA 449 F5 Jupiter (NETZSCH instruments)
−
1
−1
under flowing air (20 mL min ) and nitrogen (20 mL min ; protec-
−
1
tive gas) at a ramp rate of 10 °C min from 30 °C to 900 °C followed by
1
a 5 min hold at 900 °C. Solution phase H NMR spectroscopy of the
synthesized SDA is performed using Bruker Advance III 400 MHz NMR
spectrometer, using deuterated water as the solvent. All of the materials
30

A. Parulkar, et al.
Applied Catalysis A, General 577 (2019) 28–34
Fig. 1. SEM images of cSnBeta-100 (left), nSnBeta-100 (center), and nSnBeta-200 (right). The cSnBeta image is collected at 5 μm scale and nSnBeta images are 1 μm
scale.
affects the morphology in addition to the particle size. For nSnBeta,
SEM images provide visual indication that the material consists of
berry-like clusters with an overall size of the cluster of ˜150 nm. The
primary particle size is estimated from the pXRD data using the
Scherrer equation (correcting for instrument broadening) to be ap-
proximately 20–40 nm. From SEM analysis, the conventional Sn-Beta is
found to have a truncated bipyramidal morphology with a particle size
of 1–5 μm, consistent with previous work.
sites are present. Through deconvoluting the DRIFTS data, the ratio of
open to closed sites is determined, as shown in Table 1. cSnBeta (Figure
S4) is found to have the expected open:closed of 0.5:1, but the nSnBeta-
100 and nSnBeta-200 have ratios of 1:1 and 1.5:1, respectively. On-
going work is examining the differences in the open to closed site ratio
for the epoxide ring opening reaction. The large silanol peak however
indicates the higher density of silanols on these materials compared to
cSnBeta, making the nSnBeta less hydrophobic as compared to the
fluoride mediated synthesis of cSnBeta. This is also corroborated by a
larger mass loss corresponding to physisorbed water on the TG curve for
nSnBeta depicted in Figure S5. For nSnMFI, the DRIFTS spectrum
consists of two peaks that can be attributed to acetonitrile adsorption
on silanols and closed Sn sites (Figure S6). Overall, the results from
DRIFTS and DRUVS analysis confirm that Sn is successfully in-
corporated in the framework of nSnBeta.
The materials are further characterized using elemental analysis,
DRUVS, and DRIFTS to investigate the incorporation of Sn into
nSnBeta. Previously, nano SnMFI zeolite synthesis resulted in low Sn
incorporation efficiency [17]. Interestingly, the synthesis conditions
used for nSnBeta does not affect the Sn incorporation efficiency. The
actual Si:Sn ratio for nSnBeta100 and nSnBeta200 is computed as 118
and 170, respectively. This is comparable to conventional Sn-Beta
synthesis (Table 1). Understanding the differences in Sn incorporation
for different zeolites represents an important challenge, but it is outside
the scope of the present work.
3.2. Catalytic testing
DRIFTS and DRUVS analysis are performed to investigate the pre-
sence of framework Sn species. These techniques have been used pre-
viously to confirm the presence of framework Sn species [17,46–48].
Initial catalytic testing examined the ring-opening reaction of epi-
chlorohydrin with methanol. These substrates and the resultant pro-
ducts are relatively small, making it possible to compare catalytic
performance of materials with minimal impact of internal or external
mass transfer. Indeed, our previous work has demonstrated that these
substrates react similarly when using the tin-substituted small pore
zeolite Sn-MFI [17]. Analysis of catalytic behavior for materials with
three different particle sizes of 80, 200, and 500 nm demonstrate that
internal mass transfer limitations are minimal for these reactants. The
catalytic testing of nSnBeta-200 and cSnBeta-200 shows that both cat-
alysts result in complete conversion of epichlorohydrin in 24 h, as
shown in Fig. 2. These catalysts result in similar regioselectivity for the
terminal ether of 96%, consistent with our previous work [17,43]. The
DRUVS analysis can be used to identify presence of tin oxide (SnO ), as
2
it has a specific signature resulting in a peak at 280 nm. This peak is
absent in the nSnBeta200, however for nSnBeta100 a low intensity
broad peak is observed at 280 nm indicating that higher Sn in-
corporation can result in formation of SnO2, as shown in Figure S2.
DRIFTS analysis with CD CN as the probe molecule is performed to
3
obtain information about the types of Sn sites present in the material
[
47,49]. Conventionally, Sn-Beta is reported to have two types of Sn
sites, open and closed that result in adsorbed CD CN having a distinct
3
shift in the IR peak for the CeN bond. For nSnBeta, the DRIFTS spec-
trum consists of a large peak corresponding to silanols along with a
small peak corresponding to framework Sn species (Figure S3). The
DRIFTS spectra can be deconvoluted assuming both open and closed
initial turnover frequency (TOF
Sn per hour is calculated for all of the materials. Both cSnBeta-200 and
nSnBeta-200 have similar TOF of 193 (moles epoxide per mole Sn per
0
) of moles epoxides converted per mole
0
Table 1
Summary of characterization of catalysts, including textural properties and composition.
Surface areaa (m /g)
2
Micropore volume (cm /g)
b
3
Theoretical Si/Sn
Actual Si/Snc
Open:Closed Sites
Sample
nSnBeta-100
806
740
480
468
594
670
0.22
0.23
0.20
0.20
0.14
0.21
100
200
100
200
50
116
170
111
237
118
172
1:1
nSnBeta-200
1.5:1
0.5:1
0.5:1
0:1
cSnBeta-100
cSnBeta-200
nSnMFI-100
nSnBeta-200 Regenerated d
200
–
a
b
c
Calculated using BET method for nitrogen physisorption data.
Calculated using t-plot method for nitrogen physisorption data.
Data obtained from ICP-OES analysis.
d
Material is tested in three catalytic cycles before calcining to remove accumulated organic species.
31

A. Parulkar, et al.
Applied Catalysis A, General 577 (2019) 28–34
Fig. 2. Comparison of conversion of epichlorohydrin with methanol using
nSnBeta-100, nSnBeta-200, cSnBeta-100, and cSnBeta-200. The amount of
catalyst added is scaled to achieve an epoxide:Sn of 250:1. The reaction con-
ditions are: 0.4 M epichlorohydrin in 2 mL methanol, 10 μL DGDE (internal
standard), 60 °C, and 600 RPM.
Fig. 4. Comparison of the conversion of 1,2-epoxyoctane over time (reacted
with ethanol) for nSnBeta-200 for catalyst reusability testing. Fresh catalyst is
used for cycle 1 that is recovered via centrifugation and used in a second cat-
alytic cycle. After cycle 3, the catalyst is regenerated through calcination before
testing again (dashed line). The amount of catalyst added is scaled to achieve an
epoxide:Sn of 250:1. The reaction conditions are: 0.4 M 1,2-epoxyoctane in
ethanol with DGDE (internal standard) using a temperature of 60 °C and a stir
rate of 600 RPM.
opening of 1,2-epoxyoctane with ethanol. For this reaction, the con-
version over time is examined using cSnBeta-100 as a catalyst, as shown
in Fig. 3. Initially, cSnBeta-100 is able to convert the epoxide with a
high rate (initial turnover frequencies are reported in Table S1), but the
apparent catalytic activity decreases after a conversion of 20%. The
plateauing is associated with the pore filling as the longer diffusion
length makes it difficult for the products to diffuse out of the catalyst.
This is confirmed by analyzing the catalyst post-reaction using TGA-
DSC (Figure S7). A mass loss of ∼15% is observed between 200–700 °C,
consistent with presence of organic content in the pores.
This catalytic behavior is similar to conversion-over-time behavior
observed previously for the ring-opening of 1,2-epoxyhexane with
methanol using SnMFI [17]. Since synthesizing nano MFI zeolites
overcame diffusion limitations for 1,2-epoxyhexane and methanol,
nano MFI (nSnMFI-100) is tested for catalytic activity for 1,2-epox-
yoctane and ethanol. As shown in Fig. 3 nSnMFI-100 initially converts
the epoxide rapidly, but the rate of conversion slows down over time,
reaching only 60% with a selectivity of 51% for terminal ether in 24 h.
This suggests that the 10MR of MFI are limiting the overall catalytic
performance.
Fig. 3. Comparison of conversion of 1,2-epoxyocatane with ethanol using
nSnBeta-100, nSnBeta-200, nSnMFI-100, cSnBeta-100, and cSnBeta-200. The
amount of catalyst added is scaled to achieve an epoxide:Sn of 250:1. The re-
action conditions are: 0.4 M 1,2-epoxyoctane in 2 mL ethanol, 50 μL DGDE
The observed performance can be improved through examining a
catalytic material with the pore size of zeolite Beta and a small particle
size. Interestingly, nSnBeta-200 shows a significant improvement in
catalytic performance reaching > 99% conversion of 1,2-epoxyoctane
in 24 h with 57% selectivity for terminal ether, as shown in Fig. 3. This
demonstrates the benefit of nano-zeolites with more open frameworks
for bulkier substrates. When nSnBeta-100 is tested for the same sub-
strates, it shows slightly lower catalytic activity as compared to
(
internal standard), 60 °C, and 600 RPM.
hour; cSnBeta-200) and 168 (nSnBeta-200). In comparison, nSnBeta-
00 has a lower TOF value of 129 than nSnBeta-200 and cSnBeta. This
difference can be attributed to the presence of SnO in nSnBeta-100,
which is observed using DRUVS and has previously been demonstrated
1
0
2
to have limited catalytic activity. The formation of SnO
2
appears to be
nSnBeta-200. This difference can be attributed to the presence of SnO
2
an issue for the high loadings of Sn. Indeed, lower catalytic activity is
also observed for the conventional cSnBeta-100 materials where the
species in nSnBeta-100, as SnO
reaction [43].
2
has limited catalytic activity for this
TOF
0
of 165 is calculated, which is lower than the TOF
0
found for
With the reduction in particle size, it becomes important to de-
termine if the improvement in catalytic activity is associated with re-
ducing the particle size or if the synthesis results in catalytic sites on the
external surface of the material. To test this, nSnBeta-100 and nSnBeta-
200 are tested for ring opening of 1,2-epoxycyclododecane with
cSnBeta-200.
After establishing the similarity of the different zeolites for small
substrates, the effect of particle size on catalytic performance is tested
by comparing the nano and conventional Sn-Beta materials for ring
32

A. Parulkar, et al.
Applied Catalysis A, General 577 (2019) 28–34
Fig. 5. Characterization of fresh nSnBeta-200 catalyst compared to the regenerated catalyst after recycle experiment. (a) SEM images at 1 μm magnification of fresh
nSnBeta-200 and (b) regenerated nSnBeta-200. (c) Nitrogen physisorption isotherms for the fresh and regenerated nSnBeta-200 as well as (d) the DRUV spectra for
the fresh and regenerated nSnBeta-200.
methanol. This substrate is selected considering that it is too bulky to
enter the pores of zeolite beta. Therefore, the observed catalytic activity
would be associated with the surface sites. Both nSnBeta and cSnBeta
catalysts result in similar conversion of 5% (cSnBeta) and 8% (nSnBeta)
conversion of 1,2-epoxycyclododecane in 24 h. The slight difference in
conversion could reflect that nSnBeta has a greater external surface
area than cSnBeta, but the difference is generally considered to be
within the margin of error for these measurements. This appears to be
associated with catalytic sites on surface since a similar reaction in the
absence of catalyst results in no conversion of the bulky epoxide. This
indicates that nSnBeta and cSnBeta have similar amounts of catalytic
sites on the external surface and that the majority of the active sites are
within the pores of zeolites and not on the surface.
the framework. The catalyst recovered after third cycle is analyzed
using TGA-DSC and results are shown in Figure S8. A mass loss of
∼10% is observed between 200–500 °C confirming accumulation of
organic content in the pores of the catalyst. This is consistent with the
organic content observed after one catalytic cycle (Figure S9). To test if
the activity can be recovered, the organic content is removed by cal-
cining the catalyst at 550 °C. In addition, the Sn content in the reused
catalyst is analyzed using ICP-OES giving a Si:Sn ratio of 172 as com-
pared to 170 for the fresh catalyst. This result confirms that leaching is
not the main cause of deactivation. The recovered calcined catalyst is
tested for ring opening of 1,2-epoxyoctane with ethanol and the amount
of catalyst is adjusted to keep epoxide: Sn ratio of 250:1. The catalytic
activity for regenerated catalyst is lower compared to the fresh catalyst.
This can be caused either by loss of crystallinity, increase in particle size
3.3. Catalyst reuse testing
because of aggregation, or increase in the amount of SnO . To under-
2
stand the cause of decreased catalytic activity, the calcined catalyst
after the third cycle of recyclability test is analyzed using PXRD, ni-
trogen physisorption, SEM, and DRUVS. The results show that the
materials remain crystalline and the particle size remains constant, as
shown in Figure S10 and Fig. 5, respectively. From nitrogen physi-
sorption, it is determined that the surface area decreases by approxi-
mately 10% relative to the fresh catalyst (Table 1). While this is a
These results are promising provided that the catalyst can be reused.
Using 1,2-epoxyoctane and ethanol as substrates, the conversion versus
time data for the first cycle is similar to the small-scale tests performed,
as shown in Fig. 4. However, it is observed that activity as well as final
conversion reduces for subsequent cycles. The reduction in catalytic
activity could be attributed to pore filling and/or leaching of Sn from
33

A. Parulkar, et al.
Applied Catalysis A, General 577 (2019) 28–34
measurable decrease, it is less than the decrease in initial TOF. The
initial decrease in TOF from 99 to 54 per hour appears to be correlated
References
with an increase in the formation of SnO
2
in the recycled catalyst, as
[1] P. Wolf, C. Hammond, S. Conrad, I. Hermans, Dalton Trans. 43 (2014) 4514.
[
[
[
2] Y. Wang, J.D. Lewis, Y. Roman-Leshkov, ACS Catal. 6 (2016) 2739.
shown in Fig. 5. While the fresh catalyst does not have a peak in the
DRUV spectrum at 280 nm that would be consistent with SnO , the
DRUV spectrum for the catalyst recovered from one reaction cycle is
determined to have a peak at 280 nm that could be consistent with SnO
formation. Since our previous work indicates that SnO has limited
catalytic activity [43], the observed decrease in catalytic performance
is likely associated with SnO formation. In comparison, cSnBeta has
3] A. Corma, M. Renz, M. Boronat, P. Concepcio, Catal. Today 121 (2007) 39.
4] T. Blasco, M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martínez, J.A. Perdigón-
Melón, S. Valencia, A. Martinez, J. Perdigon-Melon, S. Valencia, J. Phys. Chem. B
2
1
02 (1998) 75.
2
[
5] R. Bermejo-Deval, M. Orazov, R. Gounder, S.-J. Hwang, M.E. Davis, ACS Catal. 4
2
(
2014) 2288.
[6] R. Bermejo-Deval, R. Gounder, M.E. Davis, ACS Catal. 2 (2012) 2705.
[7] M. Moliner, Y. Roman-Leshkov, M.E. Davis, Proc. Natl. Acad. Sci. 107 (2010) 6164.
2
[
[
8] A. Corma, M.E. Domine, L. Nemeth, S. Valencia, J. Am. Chem. Soc. 124 (2002)
previously been demonstrated to retain high catalytic activity over
multiple catalytic cycles [43]. This difference in catalyst reusability
between nSnBeta and cSnBeta suggests that the low defect framework
produced in hydrothermal conditions in the presence of fluoride is more
stable than the framework produced in hydrothermal conditions in the
presence of hydroxide. Despite the initial decrease in activity, nSnBeta
is able to retain moderate catalytic activity upon calcination and
achieve high conversion. Overall, the catalytic activity at high con-
version of nSnBeta appears higher than the cSnBeta, demonstrating the
merit of synthesizing Lewis acidic nanozeolites.
3
194.
9] J. Dijkmans, W. Schutyser, M. Dusselier, B.F. Sels, Chem. Commun. 52 (2016) 6712.
[10] A. Corma, L.T. Nemeth, M. Renz, S. Valencia, Nature 1433 (2001) 423.
[
[
11] M. Boronat, P. Concepción, A. Corma, M. Renz, S. Valencia, J. Catal. 234 (2005)
11.
1
12] J.J. Pacheco, M.E. Davis, Proc. Natl. Acad. Sci. U.S.A. 111 (2014) 8363.
[13] J.D. Lewis, S. Van De Vyver, R. Rodríguez, S. Van de Vyver, Y. Roman-Leshkov,
Angew. Chemie Int. Ed. 54 (2015) 9835.
[
14] M. Su, W. Li, T. Zhang, H.S. Xin, S. Li, W. Fan, L. Ma, Catal. Sci. Technol. 7 (2017)
3
555.
[15] S. Hwang, W.S. Chi, S.J. Lee, S.H. Im, J.H. Kim, J. Kim, J. Memb. Sci. 480
(
2015) 11.
[
[
16] B. Tang, W. Dai, G. Wu, N. Guan, L. Li, M. Hunger, ACS Catal. 4 (2014) 2801.
17] A. Parulkar, R. Joshi, N. Deshpande, N.A. Brunelli, Appl. Catal. A Gen. 566
(2018) 25.
4
. Summary
[
[
18] K. Zhang, M.L. Ostraat, Catal. Today 264 (2016) 3.
19] J. Pérez-Ramírez, C.H. Christensen, K. Egeblad, C.H. Christensen, J.C. Groen, Chem.
Soc. Rev. 37 (2008) 2530.
Lewis acidic nSnBeta zeolites are synthesized by using a low tem-
perature synthesis with TMP(OH)
2
as the SDA. It is demonstrated that
[20] D. Verboekend, J. Pérez-Ramírez, Catal. Sci. Technol. 1 (2011) 879.
[
[
[
[
21] O. Larlus, S. Mintova, S.T. Wilson, R.R. Willis, H. Abrevaya, T. Bein, Microporous
Mesoporous Mater. 142 (2011) 17.
the final particle size of nSnBeta is not influenced by the amount of Sn
incorporated into the material since a particle size of ∼150 nm is ob-
served for both nSnBeta-100 and nSnBeta-200. The catalytic testing of
nSnBeta using epichlorohydrin ring opening with methanol shows that
nSnBeta material are active catalysts for epoxide ring opening reaction.
The real advantage of particle size reduction is observed for ring
opening of bulky 1,2-epoxyoctane with ethanol where enhanced cata-
lytic activity is observed for nSnBeta as compared to cSnBeta materials.
Catalytic testing with 1,2-epoxycyclododecane with methanol demon-
strates that the catalytic sites are located inside the pores of the zeolite.
The nano zeolite (nSnBeta) can be used in multiple catalytic cycles with
a moderate loss in catalytic activity after each cycle. The catalytic
performance of nSnBeta especially at high conversions is higher than
cSnBeta even for the reused catalyst. The catalytic activity of the used
catalyst can be significantly recovered by regeneration with calcination.
Overall, nSnBeta is a highly active catalyst for the epoxide ring opening
reaction involving bulky substrates.
22] X. Yang, Y. Liu, X. Li, J. Ren, L. Zhou, T. Lu, Y. Su, ACS Sustain. Chem. Eng. 6 (2018)
8256.
23] B. Modhera, M. Chakraborty, P.A. Parikh, R.V. Jasra, Cryst. Res. Technol. 44 (2009)
3
79.
24] M.S. Holm, E. Taarning, K. Egeblad, C.H. Christensen, Catal. Today 168 (2011) 3.
[25] M. Choi, K. Na, J.Y. Kim, Y. Sakamoto, O. Terasaki, R. Ryoo, Nature 461 (2009)
246.
[
[
[
26] J. Kim, J. Chun, R. Ryoo, Chem. Commun. 51 (2015) 13102.
27] M.A. Camblor, A. Mifsud, J. Pérez-Pariente, Zeolites 11 (1991) 792.
28] M.A. Camblor, J. Pérez-Pariente, Zeolites 11 (1991) 202.
[29] O. Larlus, V. Valtchev, Microporous Mesoporous Mater. 93 (2006) 55.
[
[
30] Y. Kalvachev, T. Todorova, L. Dimitrov, V. Valtchev, Microporous Mesoporous
Mater. 177 (2013) 11.
31] I. Schmidt, C. Madsen, C.J.H. Jacobsen, Inorg. Chem. 39 (2000) 2279.
[32] Z. Zhu, H. Xu, J. Jiang, H. Wu, P. Wu, ACS Appl. Mater. Interfaces 9 (2017) 27273.
[
[
[
[
33] R. Martínez-Franco, C. Paris, M.E. Martínez-Armero, C. Martínez, M. Moliner,
A. Corma, Chem. Sci. 7 (2016) 102.
34] X. Cheng, J. Mao, X. Lv, T. Hua, X. Cheng, Y. Long, Y. Tang, J. Mater. Chem. A 2
(2014) 1247.
35] M.M. Reddy, M.A. Kumar, P. Swamy, M. Naresh, K. Srujana, L. Satyanarayana,
A. Venugopal, N. Narender, Green Chem. 15 (2013) 3474.
36] S. Tolborg, A. Katerinopoulou, D.D. Falcone, I. Sadaba, C.M. Osmundsen,
R.J. Davis, E. Taarning, P. Fristrup, M.S. Holm, J. Mater. Chem. A 2 (2014) 20252.
37] S. Valencia, A. Corma, Stannosilicate Molecular Sieves (1999).
38] P. Li, G. Liu, H. Wu, Y. Liu, J.G. Jiang, P. Wu, J. Phys. Chem. C 115 (2011) 3663.
39] H. Xu, X. Wang, P. Ji, H. Wu, Y. Guan, P. Wu, Inorg. Chem. Front. 5 (2018) 2763.
40] C. Hammond, S. Conrad, I. Hermans, Angew. Chem. Int. Ed. 51 (2012) 11736.
41] H. Robson, J. Garces, H. Kessler, A.N. Kotasthane, H. Lechert, B. Schoeman,
S. Wilson, Verified Syntheses of Zeolitic Materials (2001).
Competing interest
[
[
[
[
[
The authors declare no competing interest.
Acknowledgements
[
[
[
42] B. Tang, W. Dai, X. Sun, G. Wu, N. Guan, M. Hunger, L. Li, Green Chem. 17 (2015)
1744.
We gratefully acknowledge the American Chemical Society
Petroleum Research Fund (ACS-PRF 55946-DNI5), the National Science
Foundation (NSF Career 1653587), and the Ohio State University
Institute for Materials Research (OSU IMR FG0211 and EMR-G00018)
for their financial support.
43] N. Deshpande, A. Parulkar, R. Joshi, B. Diep, A.R. Kulkarni, N.A. Brunelli, J. Catal.
3
70 (2019) 46.
44] W. Dai, C. Wang, B. Tang, G. Wu, N. Guan, Z. Xie, M. Hunger, L. Li, ACS Catal. 6
(2016) 2955.
[
[
45] G. Prestat, C. Baylon, M.P. Heck, C. Mioskowski, Tetrahedron Lett. 41 (2000) 3829.
46] S. Roy, K. Bakhmutsky, E. Mahmoud, R.F. Lobo, R.J. Gorte, ACS Catal. 3 (2013)
5
73.
Appendix A. Supplementary data
[47] V.L. Sushkevich, I.I. Ivanova, A.V. Yakimov, J. Phys. Chem. C 121 (2017) 11437.
[
[
48] J.W. Harris, M.J. Cordon, J.R. Di Iorio, J.C. Vega-Vila, F.H. Ribeiro, R. Gounder, J.
Catal. 335 (2016) 141.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.apcata.2019.03.009.
49] W.R. Gunther, V.K. Michaelis, R.G. Griffin, Y. Roman-Leshkov, J. Phys. Chem. C
120 (2016) 28533.
34