Highly-efficient conversion of glycerol to solketal over heterogeneous Lewis acid catalysts

Article abstract of DOI:10.1039/c2gc16619d

The acetalization of acetone with glycerol to yield 2,2-dimethyl-1,3- dioxolane-4-methanol (solketal) was successfully catalyzed by mesoporous substituted silicates including the novel Hf-TUD-1 material. This reaction offers an attractive path for the conversion of glycerol, which is the main side-product in the synthesis of biodiesel, to a valuable compound with potential for industrial applications. The most promising among the heterogeneous catalysts employed in this work, Zr- and Hf-TUD-1 and Sn-MCM-41, display mainly Lewis acid properties as demonstrated by characterization with FT-IR analysis of pyridine adsorption, and achieve superior results compared to a reference solid acid catalyst such as Ultrastable zeolite Y. Especially the newly synthesized Hf-TUD-1, showing the highest conversion and turnover among the screened materials, is a promising catalyst for the acetalization of acetone with glycerol in a sustainable process. The excellent performance of these mesoporous catalysts is ascribed to their combination of acidity, wide pores, large specific surface area and relatively hydrophobic surface. The Royal Society of Chemistry.

Full text of DOI:10.1039/c2gc16619d

View Online / Journal Homepage
Dynamic Article Links
Green Chemistry
Cite this: DOI: 10.1039/c2gc16619d
www.rsc.org/greenchem
PAPER
Highly-efficient conversion of glycerol to solketal over heterogeneous Lewis
acid catalysts†
Li Li, Tamás I. Korányi, Bert F. Sels and Paolo P. Pescarmona*
Received 16th December 2011, Accepted 19th March 2012
DOI: 10.1039/c2gc16619d
The acetalization of acetone with glycerol to yield 2,2-dimethyl-1,3-dioxolane-4-methanol (solketal) was
successfully catalyzed by mesoporous substituted silicates including the novel Hf-TUD-1 material. This
reaction offers an attractive path for the conversion of glycerol, which is the main side-product in the
synthesis of biodiesel, to a valuable compound with potential for industrial applications. The most
promising among the heterogeneous catalysts employed in this work, Zr- and Hf-TUD-1 and Sn-
MCM-41, display mainly Lewis acid properties as demonstrated by characterization with FT-IR analysis
of pyridine adsorption, and achieve superior results compared to a reference solid acid catalyst such as
Ultrastable zeolite Y. Especially the newly synthesized Hf-TUD-1, showing the highest conversion and
turnover among the screened materials, is a promising catalyst for the acetalization of acetone with
glycerol in a sustainable process. The excellent performance of these mesoporous catalysts is ascribed to
their combination of acidity, wide pores, large specific surface area and relatively hydrophobic surface.
glycerol into valuable chemicals is a research area that has
received much attention in recent years. A lot of work has been
Introduction
The world economy heavily relies on petrochemical processes
based on the use of fossil resources for the production of fine
chemicals, bulk chemicals, polymers and fuels. Besides the
decline of fossil resources, there have been increasing concerns
about environmental pollution and global warming related to the
use of fossil fuels. Therefore, growing efforts have been directed
to the development and study of renewable feedstocks that are
able to satisfy the energy needs and the production of chemicals
while minimizing the negative impact on the environment. Bio-
diesel constitutes a renewable fuel that is compatible with com-
mercial diesel engines and has clear benefits compared with
diesel fuel, including enhanced biodegradability, reduced toxicity
and a lower emission profile.1–4 Biodiesel is defined as methyl
esters produced from vegetable or animal lipids and used as fuel
in diesel engines and heating systems. Biodiesel is produced by
the transesterification of triglycerides from vegetable oils and
animal fats, which in addition yields an amount of glycerol
equivalent to approximately 10 wt% of the total biodiesel pro-
duced.5 During the last years, the market for biodiesel increased
continuously, and this trend is expected to continue.6 The conse-
quent oversupply of glycerol caused a drop in its commercial
value. This generated a shift in the role of glycerol from end-of-
the-chain product to raw material. Therefore, the conversion of
devoted to the transformation of glycerol by various catalytic
processes involving oxidation, dehydration, hydrogenation,
etherification and esterification.7–11 One of the relevant paths for
the conversion of glycerol is represented by its condensation
with acetone, which yields the compound known as solketal
(2,2-dimethyl-1,3-dioxolane-4-methanol) (Scheme 1). This com-
pound has direct applications as a fuel additive, surfactant and
flavouring agent.12–14
Conventionally, the condensation of glycerol with acetone is
performed in the presence of large amounts of strong homo-
geneous Brønsted acid catalysts.15–19 However, all these
methods present limitations derived from the use of unrecyclable
expensive reagents, tedious work-up procedure, and the neces-
sity of neutralization of the strongly acidic media, with the pro-
duction of undesired wastes. Iridium complexes have also been
reported to be active,20 but costly, homogeneous catalysts
for this reaction. The sustainability of the process can be
increased by substituting the homogeneous catalysts with
heterogeneous ones. Recently, Amberlyst resins,21,22 zeolites,23
Centre for Surface Chemistry and Catalysis, KU Leuven, Kasteelpark
Arenberg 23, 3001 Heverlee, Belgium. E-mail: paolo.pescarmona@
biw.kuleuven.be; Fax: +3216321998
†Electronic supplementary information (ESI) available: Comparison
between heterogeneous catalysts for the conversion of glycerol to sol-
ketal. See DOI: 10.1039/c2gc16619d
Scheme 1 Solketal prepared from the acetalization of acetone with
glycerol.
This journal is © The Royal Society of Chemistry 2012
Green Chem.

View Online
K 10 montmorillonites,24 sulfonic mesostructured silicas,25 and
silica-supported heteropolyacids26 were reported to be active in
this reaction. The activity of these heterogeneous catalysts has
been generally ascribed to their Brønsted acidity.
min−1; 6 h at 80 °C; 80 to 450 °C at 1 °C min−1; 6 h at 450 °C).
After calcination, the solid was kept at 100 °C until its use in the
catalytic test.
In this work, the reaction of glycerol with acetone was carried
out for the first time over mesoporous substituted silicates (Zr-
and Hf-TUD-1, and Sn-MCM-41), in which the metal atoms
incorporated in the silicate framework can act as Lewis acid
sites.
Two possible reaction paths were envisaged with these hetero-
geneous catalysts: (i) the acetalization of acetone with glycerol to
yield solketal or (ii) the Meerwein–Ponndorf–Verley reduction/
Oppenauer oxidation (MPVO reaction) that would convert gly-
cerol and acetone into dihydroxyacetone/glyceraldehyde and 2-
propanol, respectively. Dihydroxyacetone and glyceraldehyde
could then be converted to valuable products such as lactic acid
and alkyl lactates.27–31 Our results prove that the acetalization is
the preferential path.
2. Characterization
The powder X-ray diffraction (XRD) pattern of Sn-MCM-41
was measured on a STOE Stadi P Combi diffractometer with an
image plate position sensitive detector (IP PSD) in the range
from 2θ = 0 to 60.00° with internal IP PSD resolution 0.03°, and
a scan time of 1500 s. For Zr- and Hf-TUD-1, XRD patterns
were recorded on a STOE Stadi MP diffractometer with a linear
position sensitive detector (PSD) (6° 2θ window) in the low
angle region from 2θ = 0 to 20.00°, with a step width of 0.5°,
internal PSD resolution 0.01°, and a step time of 10 s. An empty
sample holder was measured to estimate the effect of the back-
ground in the low angle region. Both instruments were equipped
with focusing Ge(111) monochromator, CuKα1 radiation with
λ = 1.54056 Å. The measurements were performed in Debye–
Scherrer mode at room temperature.
Experimental
N2 adsorption–desorption isotherms were measured at 77 K
on a Micrometrics Tristar 3000. The pore size distributions were
calculated from the adsorption isotherm using the Barrett–
Joyner–Halenda (BJH) method.33 The Brunauer–Emmett–Teller
(BET) method was used to calculate the surface area of the
samples.34
High-resolution transmission electron microscopy (HR-TEM)
was performed on a Philips FEG CM 200 operating at 200 kV.
Before examination, the samples were dispersed in ethanol and
deposited on 50 nm 300 mesh carbon-coated copper grids.
Chemical analyses of Zr- and Hf-TUD-1 and Sn-MCM-41
were performed by dissolving 50 mg of each sample in a
mixture of 0.5 mL of aqua regia and 3 mL of 40% aqueous HF.
The obtained solutions were analyzed by inductively coupled
plasma optical emission spectroscopy (ICP-OES) on a Jobin
Yvon Ultima instrument.
Thermogravimetric analyses (TGA) were carried out under
inert atmosphere (nitrogen gas flow of 90 mL min−1) using TGA
Q500 from TA Instruments, with a heating rate of 10 °C min−1
in the range 35–500 °C. Following a previously reported pro-
cedure,35 all samples were pre-treated overnight in a desiccator
containing a saturated aqueous solution of NH4Cl to reach the
maximum level of water adsorption. The hydrophilicity of the
materials was estimated from the number of water molecules
adsorbed on the surface of each sample, calculated from the
mass loss between 25 and 200 °C measured by TGA, using the
following equation:
1. Materials
Zr-, Hf- and Al-TUD-1 mesoporous materials were prepared via
synthesis methods based on those developed for materials of the
TUD-1 family.32 Triethanolamine (TEA) was used as chelating
and templating agent in a one-pot sol–gel procedure with a
molar ratio composition of 1SiO2 : 0.02MO2 : 0.5TEAOH : 1
TEA : 11H2O (with M = Zr, Hf or with 0.01Al2O3 instead of
0.02 MO2 for Al-TUD-1). Zirconium propoxide (Zr-
(OCH2CH2CH3)4, 70 wt% in 1-propanol), hafnium chloride
(HfCl4) and aluminum isopropoxide (Al(OCH(CH3)2)3), were
used as metal precursors. The metal precursor was firstly added
to 2-propanol and stirred at room temperature until complete dis-
solution. Then, the metal salt solution was added to tetraethyl
orthosilicate (TEOS) while vigorously stirring for 10 min. After
that, appropriate amounts of a mixture of triethanolamine and
water (TEA/H2O = 1 : 11) were added, followed by addition of
tetraethyl ammonium hydroxide (TEAOH) as a mineralizer
under vigorous stirring for at least 2 h until a gel was obtained.
The resulting gel was subsequently aged statically at room temp-
erature for 24 h, and then dried at 98 °C for another 24 h. The
obtained solid was gently ground and treated in a Teflon-lined
stainless steel autoclave at 178 °C for 24 h. Finally, the solid was
calcined in air at 600 °C for 10 h (heating rate of 1 °C min−1).
Sn-MCM-41 (Si/Sn = 50 in the synthesis gel) was prepared
following
a hydrothermal synthesis procedure previously
reported.27 Cetyltrimethylammonium bromide (CTAB) was used
as the template in a gel formed by tetraethyl orthosilicate
(TEOS), SnCl4·5H2O and tetramethylammonium silicate sol-
ution (molar gel composition: 1SiO2 : 0.02SnO2 : 0.31(CH3)4N-
(OH)·2SiO2 : 0.6CTAB : 38H2O). After hydrothermal treatment
in a Teflon-lined autoclave for 14 h at 140 °C, the obtained solid
was filtered, washed with deionized water, dried at 60 °C and
finally calcined in air for 6 h at 550 °C with a heating rate of
Δm
NA
n
¼
ꢀ
ð1Þ
H2O
mi SBET ꢀ M
H2O
where nH O is the number of adsorbed water molecules per nm2
2
of catalyst surface; Δm is the mass loss between 25 and 200 °C
(g); mi is the initial mass of the sample at 25 °C (g); MH O is the
2
molar mass of water (18.0153 g mol−1); SBET is the surface area
calculated from the nitrogen adsorption–desorption isotherms
using the BJH formula (except for USY, whose data was
obtained from the manufacturer) (nm2 g−1); NA is the Avogadro
constant (6.022 × 1023 mol−1).
1 °C min−1
.
Ultrastable zeolite Y [FAU] (CBV 600, Si/Al = 2.6, denoted
as USY) was obtained from PQ Zeolites. Prior to use, the
material was calcined in air at 450 °C (from 25 to 80 °C at 1 °C
Green Chem.
This journal is © The Royal Society of Chemistry 2012

View Online
The acidic properties of the synthesized samples were studied
by adsorption and temperature programmed desorption (TPD) of
pyridine by Fourier transform infrared (FT-IR) spectroscopy. The
infrared spectra were recorded on a Nicolet 6700 spectrometer
equipped with a DTGS detector (128 scans, 2 cm−1 resolution).
A self-supporting wafer of the sample was evacuated in a
vacuum infrared cell with ZnSe windows, developed in-house.
The sample was dried at 400 °C for 1 h under vacuum to remove
the adsorbed water. During cooling of the sample to 50 °C, refer-
ence spectra were recorded at 350, 250, 150 and 50 °C. The
dried sample at 50 °C was saturated with about 25 mbar of pyri-
dine vapour for 10 min; subsequently it was evacuated again for
30 min to ensure that no more physisorbed pyridine remained on
the sample. The evacuated sample containing chemisorbed pyri-
dine was subjected to TPD at 150, 250 and 350 °C for 30 min,
respectively, with a heating rate of 4 °C min−1 and the IR spectra
were recorded in situ at these temperatures. The amounts of acid
sites at 150, 250 and 350 °C were determined from the integral
intensity of the corresponding bands (Lewis acid sites:
1450 cm−1 and Brønsted acid sites: 1545 cm−1) in the IR spectra
using the molar extinction coefficients of Emeis.36 The reference
spectra were subtracted from the spectra containing bands from
adsorbed pyridine at the corresponding temperatures.
Selectivity and yield of the products were calculated based on
the following equations:
molof solketalformed
Selectivitytowardssolketalð%Þ¼
ꢀ100%
molof glycerolconverted
Yieldð%Þ¼conversionð%Þꢀselectivityð%Þ=100
For the recycling of Zr- and Hf-TUD-1 and Sn-MCM-41, the
following procedure was employed: after reaction at 80 °C for
6 h, the sample was centrifuged for 6 min at 3500 rpm to deposit
the solid catalyst. Next, the reaction solution was removed, 5 mL
of pre-heated tert-butanol was added and the sample was stirred
for 6 min. The sample was again centrifuged to deposit the cata-
lyst; then the supernatant tert-butanol solution was removed.
This washing procedure was repeated 5 times. Finally, the
samples were dried overnight at 100 °C and reused. In another
approach, after the washing steps with tert-butanol the catalysts
were treated in air at 600 °C for 10 h for Zr- and Hf-TUD-1 and
550 °C for 6 h for Sn-MCM-41 with a heating rate of 1 °C
min−1, and reused. An alternative washing procedure, including
washing the used catalyst with ethanol instead of tert-butanol
and drying them at higher temperature (150 °C) overnight, was
also carried out in order to maximize the removal of formed
water from the pores of the catalysts.
Leaching tests were performed under the general reaction con-
ditions employed for the catalytic tests (vide supra) with 25 mg
of catalyst. The solid was removed from the reaction mixture
after 30 min by centrifugation, followed by hot filtration (80 °C)
using a plastic syringe equipped with a 25 mm HPLC syringe
filter from Altech, with a pore size of 0.2 μm. The filtrate was
allowed to react for another 5 h 30 min at 80 °C. The products
were analyzed by GC both after 30 min and at the end of the
filtrate test (6 h).
3. Catalytic tests
The reaction of glycerol with acetone was carried out in tightly
closed glass vials under vigorous stirring (1000 rpm) in a mul-
tiple-well parallel reaction block. In a typical run, 25 mg of cata-
lyst was added at room temperature into a solution containing:
0.921 g of highly purified glycerol (purity 99%, 1.00 × 10−2
mol) or 1.15 g of a mixture of 80 wt% glycerol and 20 wt%
water (denoted as aqueous glycerol) (1.00 × 10−2 mol of gly-
cerol); 0.581 g of acetone (1.00 × 10−2 mol); 0.132 g of dioxane
(1.50 × 10−3 mol, as GC internal standard); and 1.48 g of tert-
butanol (2.00 × 10−2 mol, as solvent). In an extra test, a higher
acetone to glycerol ratio (2 : 1) was employed by using 2.00 ×
10−2 mol of acetone; the amounts of catalyst and other chemicals
remained as above. Then, the reaction mixture was stirred and
heated at 80 °C for 6 h. At the end of the test, the catalyst was
separated by centrifugation and the reaction solution was
analyzed by gas chromatography (GC) on a Trace GC Ultra
from Interscience (PH POR-Q column, FT-3, 10 m, 0.25 mm).
The temperature profile of the GC analysis was: 0.6 min
at 100 °C, from 100 to 270 °C at 270 °C min−1, 2.0 min at
270 °C. The assignment of the products was confirmed by gas
chromatography-mass spectrometry analysis (GC-MS) on an
Agilent 6890N Gas Chromatograph (WCOT fused silica column,
30 m, 0.25 mm) coupled to an Agilent 5973 MSN Mass
Spectrometer.
Results and discussion
Mesoporous Zr-, Hf- and Al-TUD-1, Sn-MCM-41 and micro-
porous Ultrastable Y zeolite (CBV 600, Si/Al = 2.6) were
studied as heterogeneous catalysts for the acetalization of
acetone with glycerol. The mesoporous catalysts were selected
because their combination of acid sites, wide and accessible
pores and high specific surface area was expected to be favour-
able for catalyzing the conversion of glycerol. USY, a H-zeolite
previously reported as catalyst for the acetalization of acetone
with glycerol, was employed for comparison purpose.23 The syn-
thesis of the three-dimensional mesoporous silicate Hf-TUD-1,
by using triethanolamine as template for the formation of meso-
pores and as metal complexing agent, is reported here for the
first time.
Prior to the catalytic tests, the physicochemical properties of
the synthesized materials were investigated by powder X-ray dif-
fraction, nitrogen physisorption and ICP-OES elemental analysis
(Table 1). Calcined Sn-MCM-41 displays the characteristic XRD
pattern and type IV isotherm of a material of the MCM-41
family, corresponding to a structure with long-range order, with
a narrow pore size distribution centered at 2.5 nm and with a
Conversion of glycerol was calculated based on the following
equation:
mol of glycerol converted
Conversion of glycerol ð%Þ ¼
ꢀ100%
initial mol of glycerol
large specific surface area of 724 m2 g−1 27
TUD-1 materials are
.
with mol of glycerol converted ¼ initial mol of glycerol
ꢁfinal mol of glycerol
a family of mesoporous silica-based solids with an open and
This journal is © The Royal Society of Chemistry 2012
Green Chem.

View Online
Table 1 Physicochemical properties of the studied catalysts
Si/M ratio
In the
synthesis In the
gel
Mesopore Average
SBET
volume
pore size
[nm]
Sample
producta [m2 g−1] [cm3 g−1
]
Zr-TUD-1
Hf-TUD-1
Al-TUD-1
Sn-MCM-41 50
CBV 600
50
50
50
51
53
651
715
629
724
66037
0.8
0.6
0.7
0.6
n.d.
13.3
4.0
4.3
2.5
n.d.
n.d.
54
—
2.628
a From ICP-OES analysis [M = metal]; n.d. = not determined.
easily accessible three-dimensional pore system.32 The calcined
Zr- and Hf-TUD-1 samples give a single broad peak at low
angle in the XRD pattern (Fig. 1(a)).This signal is similar to the
XRD patterns of other metal-incorporated TUD-1 materials, in
agreement with the mesoporous and disordered character of this
type of amorphous material.38–40 No evidence of crystalline
metal oxide was observed in the X-ray diffractograms: this
suggests the successful incorporation of zirconium and hafnium
into the framework of TUD-1, although the presence of amor-
phous metal oxide clusters cannot be excluded on the basis of
XRD data. The expected mesoporosity and amorphous character
was further confirmed by HR-TEM images of Zr- and Hf-
TUD-1 (Fig. 2), in which the interconnected pores are clearly
visible and not ordered, presenting a sponge-like 3-D network
characteristic of TUD-1 mesoporous materials. The N2 adsorp-
tion–desorption isotherms of Zr- and Hf-TUD-1 (Fig. 1(b) and
(c)) both present a type IV isotherm with a narrow hysteresis
loop, in good agreement with previous reports.39,41 All the pre-
pared TUD-1 materials display relatively narrow pore size distri-
butions and high BET specific surface areas (Table 1). The
difference in the average pore size between these
TUD-1 materials might be related to the different nature of the
metal precursors employed in the syntheses and to their rates of
hydrolysis and condensation. The elemental analysis (ICP-OES)
indicated that the Si/M ratios in the calcined mesoporous
materials are basically the same as those in the initial synthesis
gel, confirming the general validity of the synthesis method.38,42
The heterogeneous catalysts were tested in the reaction of gly-
cerol with acetone. Solketal, the product of the acetalization reac-
tion, was observed for all catalysts while no product of an
MPVO reaction or any other side-product was identified. In prin-
ciple, the acetalization reaction can yield both the five- and the
six-membered-ring solketal, depending on the acetalization pos-
ition within the glycerol molecule. In our case, complete selec-
tivity towards the five-membered-ring solketal was observed
with all catalysts, in line with data reported before for the same
reaction over heterogeneous Brønsted acid catalysts.21–26
Notably, the mesoporous substituted silicates Zr- and Hf-
TUD-1 and Sn-MCM-41 give higher conversion of glycerol to
solketal than the aluminosilicates USY and Al-TUD-1 (Table 2).
The novel Hf-TUD-1 gives the best results among all the studied
catalysts, and its excellent performance is demonstrated not only
by the highest level of conversion of glycerol to solketal (52%
after 6 h), but also by its turnover frequency (TOFmetal = number
Fig. 1 XRD patterns of Zr and Hf-TUD-1 (a). Nitrogen adsorption–
desorption isotherms and corresponding pore size distribution (insert) of
Zr-TUD-1 (b) and of Hf-TUD-1 (c).
of molglycerol converted per molmetal sites per hour), which is more
than one order of magnitude higher than that of USY (Table 2).
Green Chem.
This journal is © The Royal Society of Chemistry 2012

View Online
Fig. 2 HR-TEM images of Zr- and Hf-TUD-1.
Table 2 Catalytic conversion of glycerol and acetone to solketal
TOF
Based on
Based on
number of acid
sitesc
Conversion of
glycerola (%)
number of
Entry Catalyst
metal sitesb
Fig. 3 Synthesis of solketal through the acetalization of acetone with
glycerol: conversion of glycerol as a function of reaction time and leach-
ing tests. With all the tested catalysts the reaction proceeds with com-
plete selectivity towards the formation of the five-membered-ring
solketal product.
1
2
3
4
5
6
Zr-TUD-1
Hf-TUD-1
Al-TUD-1
Sn-MCM-41 42
USY
46
52
28
127
217
61
102
12
339
320
105
319
54
36
1
No catalyst
—
—
a Reaction conditions: 1.00 × 10−2 mol of glycerol, 1.00 × 10−2 mol of
acetone (acetone–glycerol = 1 : 1), 1.50 × 10−3 mol of dioxane and 2.00
× 10−2 mol of tert-butanol, 25 mg of catalyst (3 wt% relative to
glycerol), 6 h at 80 °C. b Turnover frequency is moles of glycerol
converted per mole of metal sites (Zr, Hf, Al and Sn, respectively) of the
catalyst per hour. The number of metal sites for each catalyst is
calculated based on the ICP-OES analysis of the calcined products (see
Table 1) except in the case of Al-TUD-1, for which the Si/Al ratio in the
synthesis gel was used. c Turnover frequency is moles of glycerol
converted per mole of acid sites (sum of Lewis and Brønsted acid sites)
of the catalysts per hour. The amount of acid sites of each catalyst is
determined by pyridine TPD-IR at 150 °C (see Table 3). In both
cases,b,c the TOF values were calculated using the data after 2 h of
reaction, i.e. when the reaction is far from equilibrium conditions.
A kinetic study shows that, at any reaction time, the order of
activity of the catalysts is Hf-TUD-1 > Zr-TUD-1 > Sn-
MCM-41 > USY (Fig. 3). With all tested catalysts, the conver-
sion of glycerol increases rapidly in the first 6 h of reaction, after
which the reaction rate decreases noticeably, suggesting that the
reaction is approaching equilibrium conditions.
Fig. 4 Infrared spectra of pyridine adsorbed at 150 °C over (i) Al-
TUD-1, (ii) Zr-TUD-1, (iii) Hf-TUD-1, (iv) Sn-MCM-41, and (v) USY.
understanding of the relationship between the acidity of materials
and their catalytic properties. Fig. 4 shows the FT-IR spectra of
chemisorbed pyridine at 150 °C over the TUD-1 materials, Sn-
MCM-41 and USY. The intensities of the peaks are normalized
by the sample masses. All samples exhibit bands due to strong
Lewis bound pyridine, around 1450 and 1620 cm−1, and weak
Lewis bound pyridine at 1575 cm−1 (L in Fig. 4). Peaks at 1545
and 1639 cm−1 are attributed to Brønsted acid sites (B in Fig. 4)
and the peak around 1492 cm−1 is assigned to pyridine associ-
ated with both Brønsted and Lewis acid sites (L + B in Fig. 4).43
The type and amounts of acid sites for the different catalysts are
summarized in Table 3. As anticipated, the majority of the acid
sites in the mesoporous catalysts are of Lewis type, while both
Brønsted and Lewis acid sites are present in the USY zeolite,
with the former being in larger amount. The presence of small
amounts of Brønsted acid sites in the metal-substituted
The heterogeneous nature of the mesoporous catalysts was
confirmed by a hot filtration leaching test. After 30 min of reac-
tion at 80 °C, the solids (25 mg) were removed from the reaction
mixture by centrifugation followed by filtration at 80 °C. The
filtrate was allowed to react for further 5 h 30 min at 80 °C. The
conversion of glycerol after 30 min of reaction is 5% for
Zr-TUD-1, 10% for Hf-TUD-1 and 3% for Sn-MCM-41. No
increase in the final conversion in the filtrate was observed after
further 5 h 30 min. Thus, no leaching of active species occurs
during the reaction, proving that these catalysts are truly hetero-
geneous (Fig. 3).
To correlate the catalytic behaviour of these materials with
their physicochemical properties, the catalysts were further
studied using a combination of characterization techniques.
Characterization of the catalyst surface by FT-IR spectra of
adsorbed pyridine helps to develop
a
comprehensive
This journal is © The Royal Society of Chemistry 2012
Green Chem.

View Online
mesoporous silica materials is generally associated to surface
hydroxyls in the vicinity of the Lewis acid metal sites.44,45 The
strength of Brønsted and Lewis acid sites can be estimated by
means of their amounts at different temperatures during the
temperature programmed desorption of pyridine (Table 3). For
the TUD-1 materials and Sn-MCM-41, the amounts of Brønsted
acid sites decreases with increasing temperature and becomes
almost negligible at 350 °C, while the Lewis acid sites are more
persistent as shown by the trend of increasing ratio between
Lewis and Brønsted acid sites upon increase of the temperature
(L/B ratio in Table 3). This indicates that the Lewis acid sites of
these catalysts are stronger than the Brønsted ones. This is par-
ticularly evident in the case of Hf-TUD-1, for which the L/B
ratios at 250 °C are much larger than that at 150 °C. On the
other hand, relative large amounts of Brønsted and Lewis acid
sites are detectable at 350 °C for USY, and the amount of
Brønsted acid sites is larger than that of Lewis acid sites at all
temperatures. The presence of both types of acid sites is well
known for this zeolite, obtained by steaming of an ammonium-
exchanged Na-Y zeolite that leads to the partial removal of Al
from the zeolite framework.28 The analysis of the population of
acid sites of the various catalysts suggests that Lewis acid sites
are active in the acetalization of acetone with glycerol. The total
number of acid sites in Zr- and Hf-TUD-1 and Sn-MCM-41 cor-
relates well with their catalytic activity: Hf-TUD-1 displays the
highest number of acid sites and the highest activity among these
three mesoporous catalysts. When normalizing the activity
through the number of acid sites, i.e. if the TOF is calculated
relative to the number of acid sites, we obtain very similar values
for Sn-MCM-41 and Zr- and Hf-TUD-1 (Table 2). Conversely,
when the TOF is calculated relative to the number of metal sites,
Hf-TUD-1 gives a higher value than Zr-TUD-1 and Sn-
MCM-41. This suggests that not all the metal atoms in these cat-
alysts are equally accessible, part of them being located within
the walls of the mesoporous structure and not on the surface.
Hafnium atoms, being larger than zirconium and tin, might be
preferentially located at the surface and, thus, be more likely
accessible for the reagents.
in determining the activity in the conversion of glycerol to solke-
tal. Particularly, the hydrophilicity of the catalyst should be taken
into account. Water is the side-product of this reaction
(Scheme 1). Therefore, its presence is expected to have a nega-
tive effect on the reaction progress. Indeed, when our set of cata-
lysts was tested using aqueous glycerol, which mimics the water
content of crude glycerol,6 the conversion to solketal was mark-
edly lower than with pure glycerol (Fig. 5). Previous work
showed that the acetalization reaction can be displaced towards
the product by continuously removing water from the system.21
Different solvents have been employed to separate water from
the product mixture using a Dean–Stark apparatus, so as to
increase the glycerol conversion. However, most of the
employed solvents such as benzene, toluene, chloroform and
dichloromethane may cause environmental problems and are
hazard to humans. An alternative and greener approach to
remove water from the catalytic sites is to use a sufficiently
hydrophobic catalyst. Recently, da Silva et al. attributed the high
conversion of glycerol to solketal obtained over zeolite Beta
with Si/Al = 16 to the relatively high hydrophobic character of
this material, which could drive out the formed water from the
pores, thus preserving the acid strength and minimizing the
reverse reaction.23 The hydrophilicity of the catalysts presented
in this work was studied by measuring the amount of water
Although the catalytic trends with Zr- and Hf-TUD-1 and Sn-
MCM-41 can be explained in terms of their acidity, the inferior
catalytic performance of USY and Al-TUD-1 contrasts with their
high number and strength of Brønsted and Lewis acid sites (see
data in Tables 2 and 3). The smaller size of the micropores of
zeolite USY compared to the mesopores of the other studied
materials might limit the diffusion of reagents and products and,
thus, negatively affect its catalytic activity. However, Al-
TUD-1 has similar textural features to the other mesoporous
materials, indicating that other parameters play an important role
Fig. 5 Catalytic conversion of glycerol with acetone with different
reaction conditions: 1.00 × 10−2 mol of glycerol, 1.00 × 10−2 mol of
[a]
acetone (acetone/glycerol = 1 : 1), 1.50 × 10−3 mol of dioxane and 2.00
[b]
[a]
× 10−2 mol of tert-butanol, 25 mg of catalyst, 6 h at 80 °C;
as in
but with addition of 1.30 × 10−2 mol of H2O (glycerol/H2O = 4 : 1 in
mass); [c] as in [a] but with 2.00 × 10−2 mol of acetone (acetone/glycerol
= 2 : 1).
Table 3 Type and amount of surface acid sites and Lewis/Brønsted (L/B) acidity ratios of the catalysts, as determined by pyridine TPD-IR
Lewis acidity [mmol g−1
]
Brønsted acidity [mmol g−1
]
L/B ratio
150 °C
Catalysts
Temperature
150 °C
250 °C
350 °C
150 °C
250 °C
350 °C
250 °C
350 °C
Zr-TUD-1
Hf-TUD-1
Al-TUD-1
Sn-MCM-41
USY
0.093
0.157
0.133
0.083
0.184
0.033
0.059
0.111
0.014
0.092
0.004
0.006
0.091
0.001
0.051
0.025
0.043
0.059
0.011
0.333
0.005
0.007
0.027
0.002
0.213
0.001
0.001
0.009
0.000
0.096
3.7
3.6
2.2
7.4
0.6
6.6
9.0
4.1
8.3
0.4
6.8
10.0
10.1
(∞)
0.5
Green Chem.
This journal is © The Royal Society of Chemistry 2012

View Online
Fig. 6 TGA curves and amount of water molecules adsorbed on the
surface (nH O/nm2) of the studied catalysts: (i) Al-TUD-1; (ii) USY; (iii)
Hf-TUD-1; (iv) Sn-MCM-41; and (v) Zr-TUD-1.
2
Fig. 7 Recycling of Zr- and Hf-TUD-1 and Sn-MCM-41 as catalysts
for the conversion of glycerol to solketal [filled symbols: the catalysts
were washed with tert-butanol and dried at 100 °C overnight; empty
symbols: the catalysts were washed with ethanol and dried at 150 °C
overnight].
adsorbed on the surface of the materials using thermal gravi-
metric analysis (TGA) (Fig. 6). The number of adsorbed water
molecules on the surface of each catalyst (nH O/nm2) was calcu-
2
lated according to formula (1) (see Experimental section) and
listed in Fig. 6. Although a full comparison between the hydro-
philicity of these materials is hindered by the different adsorptive
behaviour of micro- and mesoporous materials, these data quali-
tatively show that the hydrophobic character of the catalysts
decreases in the following order: Zr-TUD-1 > Sn-MCM-41 >
Hf-TUD-1 > USYSi/Al = 2.6 > Al-TUD-1. We propose that the
higher hydrophilicity of USY and, particularly, of Al-TUD-1 is
responsible for their lower catalytic activity, as the water formed
during the reaction could not be expelled, favouring the reverse
reaction (i.e., the hydrolysis of solketal).
In summary, on the basis of the characterization of our cata-
lysts we can conclude that the acetalization of acetone with gly-
cerol to yield solketal can be efficiently catalyzed by the
mesoporous Sn-MCM-41 and Zr- and Hf-TUD-1 materials. The
final catalytic activity does not depend only on the amount of
acid sites: wide pores and high surface area help optimizing the
diffusion of reagents and products and the contact with the
active sites, while a relatively hydrophobic surface favours
removal of water, minimizing the hydrolysis of the formed solke-
tal. This explains why Zr- and Hf-TUD-1 and Sn-MCM-41
display better catalytic results than USY and Al-TUD-1, despite
the high population of acid sites of the latter two. Apparently,
the novel Hf-TUD-1 material presents the most suitable balance
of all these parameters, leading to the best catalytic performance
in the conversion of glycerol to solketal.
In order to check whether the hydrolysis of the formed solke-
tal would generate only glycerol and acetone (i.e., the reverse
reaction) or also dihydroxyacetone and isopropanol, the five-
membered-ring solketal was reacted with water by using Zr-
TUD-1 (25 mg) as catalyst at 80 °C for 6 h. After the reaction,
glycerol and acetone were the only products and the conversion
of solketal was 40%.
Finally, the reusability of Zr and Hf-TUD-1 and Sn-MCM-41
was assessed in several consecutive runs. After completion of
each run, the reaction mixture was centrifuged and the deposited
catalyst was washed five times with pre-heated tert-butanol at
room temperature followed by either drying at 100 °C overnight
(1st–4th cycles) or recalcination in air (5th cycle). In the
washing case, the conversion of glycerol gradually decreased
upon recycling with all catalysts (Fig. 7, filled symbols).
However, full reactivation of the used catalysts was achieved
after recalcination. The hypothesis that the deactivation
was caused by water adsorbed in the pores of the catalysts was
tested by increasing the temperature of the drying step after
the 1st run from 100 to 150 °C, and by immediately adding
the reagents after this drying step. With all catalysts the glycerol
conversion in the 2nd run was very similar to that obtained
with the original recycling procedure (Fig. 7). This indicates
that the deactivation is caused by reaction residues deposited on
the active sites and not by adsorbed water. The calcination
efficiently removes the residues and restores the original catalytic
activity.
Comparison of our materials to other heterogeneous catalysts
reported for the conversion of glycerol to solketal is not straight-
forward due to the different conditions used for the catalytic
tests in different works (see Table S1 in the ESI†).21–26 Never-
theless, the excellent properties of our mesoporous catalysts are
demonstrated by their very high turnover and by the high gly-
cerol conversion achieved with a low catalyst loading, with a
1 : 1 ratio of acetone to glycerol and without need for removal of
water from the reaction mixture.
Aiming at an atom efficient process, a 1 : 1 ratio between
acetone and glycerol was employed in the catalytic tests pre-
sented so far in this work. Higher ratios, as those used in most
literature on this reaction (see Table S1 in the ESI†), can be
employed to achieve a higher degree of glycerol conversion to
solketal.21–26 In line with these reports, glycerol conversion
increased over our catalysts when an acetone to glycerol ratio of
2 : 1 was used (Fig. 5). We anticipate that further increase of the
conversion could be achieved by using higher acetone to gly-
cerol ratio or higher catalyst loading.
This journal is © The Royal Society of Chemistry 2012
Green Chem.

View Online
Gina Vanbutsele for assistance in the measurements of the N2
isotherms, Dr Elena Gobechiya for support in the XRD analysis,
Dr Chalida Klaysom and Prof. Johan Billen for the TEM analy-
sis, Sofie Mannaerts for the ICP-OES analysis. We also thank
the master students Sasja Helsen, Willem Gilbert and Jonas Van
Lishout for their contribution to the catalytic tests.
Scheme 2 Proposed mechanism for the reaction of glycerol and
acetone to yield solketal over Lewis acid catalysts (M is the metal acting
as a Lewis acid site).
Notes and references
1 J. Coombs and K. Hall, Renewable Energy, 1998, 15, 54–59.
2 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411–2502.
3 F. W. Lichtenthaler and S. Mondel, Pure Appl. Chem., 1997, 69, 1853–
1866.
For the acetalization of acetone with glycerol catalyzed by
solid Brønsted acids, the five-membered-ring solketal has been
proposed to form selectively through a mechanism involving a
short-lived carbenium ion intermediate.23 For the reaction over
Zr- and Hf-TUD-1 and Sn-MCM-41, we propose that the Lewis
acid metal sites play a similar role as in the Meerwein–Ponn-
drof–Verley reduction/Oppenauer oxidation reactions, by coordi-
nating and activating the carbonyl group of the acetone
(Scheme 2). Then, the carbon atom of the carbonyl can be
attacked by the primary alcoholic group of glycerol,
accompanied by the formation of a bond between the carbonyl
oxygen atom and the β-carbon of the glycerol and by dehy-
dration, leading to the formation of the five-membered-ring solk-
etal. The conversion of the Lewis acid sites of Zr- and Hf-
TUD-1 and Sn-MCM-41 into Brønsted ones is unlikely, as
demonstrated in previous work on Sn-MCM-41.46–48 The frame-
work oxygen atom next to the Lewis acid metal site could play a
role as Lewis base in activating glycerol,49 but further studies
would be necessary to evaluate if this effect is relevant in this
reaction.
4 P. Gallezot, Catal. Today, 2007, 121, 76–91.
5 F. Ma and M. A. Hanna, Bioresour. Technol., 1999, 70, 1–15.
6 C. H. C. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. Lu, Chem. Soc.
Rev., 2008, 37, 527–549.
7 S. H. Chai, H. P. Wang, Y. Liang and B. Q. Xu, J. Catal., 2007, 250,
342–349.
8 M. A. Dasari, P. P. Kiatsimkul, W. R. Sutterlin and G. J. Suppes, Appl.
Catal., A, 2005, 281, 225–231.
9 P. Gray and A. Williams, Nature, 1960, 188, 56–56.
10 K. Klepáčová, D. Mravec and M. Bajus, Appl. Catal., A, 2005, 294, 141–
147.
11 R. R. Soares, D. A. Simonetti and J. A. Dumesic, Angew. Chem., Int. Ed.,
2006, 45, 3982–3985.
12 M. J. Climent, A. Corma and A. Velty, Appl. Catal., A, 2004, 263, 155–
161.
13 M. J. Climent, A. Velty and A. Corma, Green Chem., 2002, 4, 565–569.
14 A. Sokołowski, A. Piasecki and B. Burczyk, J. Am. Oil Chem. Soc.,
1992, 69, 633–638.
15 E. García, M. Laca, E. Pérez, A. Garrido and J. n. Peinado, Energy Fuels,
2008, 22, 4274–4280.
16 H. Garcia, Chem. Today, 2008, 26, 10–12.
17 Y. J. M. W. J. Monbaliu, B. Chevalier, F. Schmidt, R. Hoogendoorn,
M. Kousemaker and C. V. Stevens, Chem. Today, 2010, 28, 8–11.
18 A. Krief, L. Provins and A. Froidbise, Tetrahedron Lett., 1998, 39, 1437–
1440.
19 P. C. Ray and S. M. Roberts, Tetrahedron Lett., 1999, 40, 1779–1782.
20 C. Crotti, E. Farnetti and N. Guidolin, Green Chem., 2010, 12, 2225–
2231.
Conclusion
21 J. Deutsch, A. Martin and H. Lieske, J. Catal., 2007, 245, 428–435.
22 J. S. Clarkson, A. J. Walker and M. A. Wood, Org. Process Res. Dev.,
2001, 5, 630–635.
23 C. X. A. da Silva, V. L. C. Goncalves and C. J. A. Mota, Green Chem.,
2009, 11, 38–41.
24 L. Roldán, R. Mallada, J. M. Fraile, J. A. Mayoral and M. Menéndez,
Asia-Pac. J. Chem. Eng., 2009, 4, 279–284.
25 G. Vicente, J. A. Melero, G. Morales, M. Paniagua and E. Martin, Green
Chem., 2010, 12, 899–907.
The novel three-dimensional mesoporous silicate Hf-TUD-1,
together with Zr-TUD-1 and Sn-MCM-41, demonstrated excel-
lent catalytic activity in the acetalization of acetone with glycerol
to produce 2,2-dimethyl-1,3-dioxolane-4-methanol (solketal).
This is the first time that mesoporous Lewis acid catalysts have
been reported to be active in such reaction. The achieved conver-
sion and turnover frequencies are superior to those obtained with
Ultrastable zeolite Y and Al-TUD-1. Wide pores, large specific
surface area, a relatively hydrophobic surface and the amount of
accessible acid sites were identified as the factors that positively
influence the catalytic performance in this reaction. The active
mesoporous materials do not suffer from leaching and can be
efficiently reused in consecutive catalytic cycles.
26 P. Ferreira, I. M. Fonseca, A. M. Ramos, J. Vital and J. E. Castanheiro,
Appl. Catal., B, 2010, 98, 94–99.
27 L. Li, C. Stroobants, K. Lin, P. A. Jacobs, B. F. Sels and P.
P. Pescarmona, Green Chem., 2011, 13, 1175–1181.
28 P. P. Pescarmona, K. P. F. Janssen, C. Delaet, C. Stroobants,
K. Houthoofd, A. Philippaerts, C. De Jonghe, J. S. Paul, P. A. Jacobs and
B. F. Sels, Green Chem., 2010, 12, 1083–1089.
29 K. P. F. Janssen, J. S. Paul, B. F. Sels and P. A. Jacobs, in Studies in
Surface Science and Catalysis, ed. Z. G. J. C. Ruren Xu and Y. Wenfu,
Elsevier, 2007, vol. 170, pp. 1222–1227.
30 P. Pescarmona, K. Janssen, C. Stroobants, B. Molle, J. Paul, P. Jacobs and
B. Sels, Top. Catal., 2010, 53, 77–85.
31 K. Lin, L. Li, B. F. Sels, P. A. Jacobs and P. P. Pescarmona, Catal. Today,
2011, 173, 89–94.
We have discovered new, highly efficient heterogeneous cata-
lyst for the synthesis of valuable solketal from a renewable feed-
stock such as glycerol, thus opening new perspectives for
sustainable utilization of this side-product of biodiesel
production.
32 J. C. Jansen, Z. Shan, L. Marchese, W. Zhou, N. v. d. Puil and
T. Maschmeyer, Chem. Commun., 2001, 713–714.
33 E. P. Barrett, L. G. Joyner and P. P. Halenda, J. Am. Chem. Soc., 1951,
73, 373–380.
Acknowledgements
34 B. C. Lippens and J. H. de Boer, J. Catal., 1965, 4, 319–323.
35 K. Lin, P. P. Pescarmona, K. Houthoofd, D. Liang, G. Van Tendeloo and
P. A. Jacobs, J. Catal., 2009, 263, 75–82.
The authors acknowledge sponsoring in the frame of the follow-
ing research programs: START1, CREA, Methusalem, IAP-PAI
(BELSPO, Federal Government), NoE IDECAT (EU). We thank
36 A. Emeis, J. Catal., 1993, 141, 347–354.
Green Chem.
This journal is © The Royal Society of Chemistry 2012

View Online
37 Zeolyst International, http://www.zeolyst.com/our-products/standard-
zeolite-powders/zeolite-y.aspx
43 B. Chakraborty and B. Viswanathan, Catal. Today, 1999, 49, 253–
260.
38 I. C. Neves, G. Botelho, A. V. Machado, P. Rebelo, S. Ramôa,
M. F. R. Pereira, A. Ramanathan and P. Pescarmona, Polym. Degrad.
Stab., 2007, 92, 1513–1519.
39 A. Ramanathan, D. Klomp, J. A. Peters and U. Hanefeld, J. Mol. Catal.
A: Chem., 2006, 260, 62–69.
40 R. Anand, M. S. Hamdy, P. Gkourgkoulas, T. Maschmeyer, J. C. Jansen
and U. Hanefeld, Catal. Today, 2006, 117, 279–283.
41 R. Anand, M. S. Hamdy, U. Hanefeld and T. Maschmeyer, Catal. Lett.,
2004, 95, 113–117.
42 S. Telalovic, A. Ramanathan, G. Mul and U. Hanefeld, J. Mater. Chem.,
2010, 20, 642–658.
44 Y. Goa, P. Wu and T. Tatsumi, J. Phys. Chem. B, 2004, 108, 8401–
8411.
45 H. Kosslick, G. Lischke, G. Walther, W. Storek, A. Martin and R. Fricke,
Microporous Mater., 1997, 9, 13–33.
46 A. Corma, M. E. Domine and S. Valencia, J. Catal., 2003, 215, 294–
304.
47 A. Corma and M. Renz, Chem. Commun., 2004, 550–551.
48 A. Corma, F. X. Llabrés i Xamena, C. Prestipino, M. Renz and
S. Valencia, J. Phys. Chem. C, 2009, 113, 11306–11315.
49 M. J. Climent, A. Corma, P. De Frutos, S. Iborra, M. Noy, A. Velty and
P. Concepción, J. Catal., 2010, 269, 140–149.
This journal is © The Royal Society of Chemistry 2012
Green Chem.