Acetalisation of bio-glycerol with acetone to produce solketal over sulfonic mesostructured silicas

Article abstract of DOI:10.1039/b923681c

Sulfonic acid-functionalized mesostructured silicas have demonstrated excellent catalytic behaviour in the acetalisation of glycerol with acetone to yield 2,2-dimethyl-1,3-dioxolane-4-methanol, also known as solketal. This molecule constitutes an excellent compound for the formulation of gasoline, diesel and biodiesel fuels. The activity achieved with arenesulfonic acid-functionalized silica is comparable to that displayed by Amberlyst-15. Optimal production of solketal over arenesulfonic acid mesostructured silica has been established for a reaction system consisting of three consecutive 2-step batches (30 min under reflux and an evaporation step under vacuum), and using a 6/1 acetone/glycerol molar ratio. The use of lower grades of glycerol, such as technical (purity of 91.6 wt%) and crude (85.8 wt%) glycerol, has also provided high conversions of glycerol over sulfonic acid-modified heterogeneous catalysts (84% and 81%, respectively). For refined and technical glycerol the catalysts have been reused, without any regeneration treatment, up to three times, keeping the high initial activity. However, the high sodium content in crude glycerol deactivates the sulfonic acid sites by cation exchange. This deactivation is readily reversed by simple acidification of the catalyst after reaction.

Full text of DOI:10.1039/b923681c

View Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/greenchem | Green Chemistry
Acetalisation of bio-glycerol with acetone to produce solketal over sulfonic
mesostructured silicas†
Gemma Vicente,*a Juan A. Melero,b Gabriel Morales,b Marta Paniaguab and Eric Mart´ınb
Received 13th November 2009, Accepted 19th February 2010
First published as an Advance Article on the web 22nd March 2010
DOI: 10.1039/b923681c
Sulfonic acid-functionalized mesostructured silicas have demonstrated excellent catalytic
behaviour in the acetalisation of glycerol with acetone to yield 2,2-dimethyl-1,3-dioxolane-4-
methanol, also known as solketal. This molecule constitutes an excellent compound for the
formulation of gasoline, diesel and biodiesel fuels. The activity achieved with arenesulfonic
acid-functionalized silica is comparable to that displayed by Amberlyst-15. Optimal production of
solketal over arenesulfonic acid mesostructured silica has been established for a reaction system
consisting of three consecutive 2-step batches (30 min under reflux and an evaporation step under
vacuum), and using a 6/1 acetone/glycerol molar ratio. The use of lower grades of glycerol, such
as technical (purity of 91.6 wt%) and crude (85.8 wt%) glycerol, has also provided high
conversions of glycerol over sulfonic acid-modified heterogeneous catalysts (84% and 81%,
respectively). For refined and technical glycerol the catalysts have been reused, without any
regeneration treatment, up to three times, keeping the high initial activity. However, the high
sodium content in crude glycerol deactivates the sulfonic acid sites by cation exchange. This
deactivation is readily reversed by simple acidification of the catalyst after reaction.
In turn, the EU 2009/28/EC Directive establishes a 10%
1. Introduction
share of energy from renewable sources for transport in EU
Biodiesel constitutes a renewable fuel that is almost compatible
energy consumption by 2020. Conversely, another EU Di-
with commercial diesel engines and has clear benefits relative
rective (2003/96/EC) allows the Member States exemptions
to diesel fuel, including enhanced biodegradation, reduced
or reductions on excise duties so as to promote biofuels. As
toxicity and a lower emission profile.1 In a recent European
a consequence, and also taking into account the unstable
Union (EU) Directive (2009/28/EC), biodiesel is defined as a
petroleum prices, there is a growing interest in FAMEs as
methyl ester produced from vegetable or animal oil, of diesel
an alternative diesel fuel in Europe. That implies an increase
quality, to be used as biofuel. Methyl esters, usually referred
in biodiesel manufacturing in the next few years, which also
to as fatty acid methyl esters (FAMEs), are products from
means a dramatic rise in the availability of its by-product (crude
the transesterification of vegetable oils and animal fats with
glycerol). In fact, the production of this biofuel has already
methanol in the presence of an acid catalyst or a basic one.2
increased significantly in the EU in recent years, reaching a
In addition, the process yields an amount of glycerol which
production of 7.75 million tonnes in 2008.3 Therefore, the value
is equivalent to approximately 10 wt% of the total biodiesel
of glycerol has fallen as a result of its oversupply in the global
produced. The glycerol obtained at this stage is called crude
market.
glycerol and it is about 80% pure, the main contaminants
In this context, important research is currently being devel-
being soaps, salts, methanol and water. By further refinement,
oped in order to find new applications for this cheap and off-
this glycerol turns into usable and profitable grades. Thus,
grade glycerol from biodiesel plants. Some recent reviews have
technical glycerol (>90% pure) is obtained as an intermediate to
dealt with the related research into glycerol upgrading to be
high-purity glycerol (pharmaceutical-grade glycerol). It results
processed into valuable chemicals.4–9 These strategies include
essentially from a desalting process. Finally, pharmaceutical
glycerol (>99.7% pure) is produced by additional purification
through distillation.
selective oxidation to produce a range of products such as
dihydroxyacetone, glyceraldehyde, glyceric acid, glycolic acid,
hydroxypyruvic acid, mesoxalic acid, oxalic acid, and tartronic
acid; reduction to obtain 1,3-propanediol and 1,2-propanediol;
hydrogenolysis to obtain propylene glycol; dehydration to yield
acrolein or 3-hydroxypropionaldehyde; halogenation to pro-
aDepartment of Chemical and Energy Technology, ESCET, Universidad
Rey Juan Carlos, C/Tulipa´n s/n, 28933, Mo´stoles, Madrid, Spain.
duce 1,3-dichloropropanol (intermediate in the epichlorohydrin
synthesis); fermentation towards 1,3-propanediol; or polymer-
ization to obtain polyglycerols and polyglycerol esters. Apart
from these alternatives, the transformation of glycerol into fuel
oxygenates is being explored by means of etherification with
olefins or alcohols,10–17 esterification with low molecular weight
E-mail: gemma.vicente@urjc.es; Fax: +34 91 488 70 68; Tel: +34 91 488
85 31
bDepartment of Chemical and Environmental Technology, ESCET,
Universidad Rey Juan Carlos, C/Tulipa´n s/n, 28933, Mo´stoles, Madrid,
Spain
† Electronic supplementary information (ESI) available: XRD patterns;
nitrogen adsorption-desorption isotherms. See DOI: 10.1039/b923681c
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 899–907 | 899
©

View Online
acids,18–21 transesterification using low molecular weight esters18
and acetalisation with aldehydes or ketones.18,22–27 This approach
is a promising and economically viable alternative since it not
only makes a profitable use of glycerol but also increases the yield
of biofuel in the overall biodiesel production process, helping to
meet the target for energy from renewable sources for transport
in the EU Directive.
hydrophilic/hydrophobic balance as well as the strength and
concentration of acid sites,29 appear to be promising catalysts
for the acid-catalyzed transformation of bulky molecules such as
glycerol. Indeed, they have previously demonstrated an excellent
catalytic behaviour in the transformation of glycerol into fuel
components by means of etherification with isobutylene17 and
esterification with acetic acid.19
Special interest is focused on the preparation of acetals
and ketals of glycerol by acetalisation with an aldehyde or
a ketone, respectively, in the presence of an acid catalyst.
The most-used solid acids for the production of these glyc-
erol derivates have been homogeneous catalysts (mainly p-
toluenesulfonic acid),18,23,26 ion-exchange organic resins22,24,25,28
and acid zeolites.25,27 Acetals and ketals of glycerol constitute
an excellent component for the formulation of gasoline, diesel
and biodiesel fuels. These oxygenated compounds, when in-
corporated into standard diesel fuel, have led to a decrease
in particles, hydrocarbons, carbon monoxide and unregulated
aldehyde emissions.18,24 Likewise, these products can act as cold
flow improvers for use in biodiesel, also reducing its viscosity.18
This issue is of significant importance due to the growing
demand for new additives specifically for biodiesel that are
biodegradable, non-toxic and renewable. More recently, Garc´ıa
et al.26 confirmed that the addition of these compounds to
biodiesel improved the viscosity and also met the established
requirements for flash point and oxidation stability.
The main drawback of the glycerol acetalisation is the
production of water, which has to be removed in order to
hinder the reversibility of the reaction. The use of solvents
such as benzene, toluene, petroleum ether or chloroform to
increase the conversion of glycerol into acetals or ketals has
been described.25 However, this method is not very efficient in
this reaction and presents environmental problems. In addition,
Bruchmann et al.23 have used aldehyde or ketone in excess, which
is continuously removed by distillation during the reaction to
favour the irreversibility of this reaction and increase the glycerol
conversion. The levelof liquid in the reactor was kept constantby
continuously feeding dry aldehyde or ketone. Also, continuous
processes for the formation of solketal employing heterogeneous
catalysts, such as the commercial macroporous acid resins of the
Amberlyst family, have been described by others authors.24,25
More recently, da Silva et al.27 reported the use of zeolite Beta
with a Si/Al ratio of 16 as a catalyst for the acetalisation of
glycerol. The hydrophobic character of this zeolite prevents the
diffusion of the water into the pores, preserving the strength of
the acid sites and impairing the reverse reaction.
The catalytic performance of these sulfonic acid-modified
mesostructured silicas has been benchmarked with other com-
mercial acid catalysts. In addition, the process was developed
and optimised by following the factorial design and response
surface methodology. This technique is a powerful tool that
involves many advantages that have been described in previous
studies.30 Finally, the effects of different glycerol grades ranging
from crude glycerol to refined glycerol (pharmaceutical grade)
were evaluated in this acetalisation reaction.
2. Experimental section
2.1. Catalysts preparation
Propylsulfonic acid-functionalized mesostructured silica (Pr-
SBA-15) was synthesized following a previously reported
procedure.31 The molar composition of the mixture for
4
g of templating co-polymer was: 0.0369 tetraethy-
lorthosilicate (TEOS, Sigma-Aldrich), 0.0041 mercaptopropyl-
trimethoxysilane (MPTMS, Sigma-Aldrich), 0.0369 H2O2, 0.24
HCl, and ª6.67 H2O.
Arenesulfonic acid-functionalized mesostructured silica (Ar-
SBA-15) was obtained as described elsewhere.31 In this case, the
molar composition of the mixture for 4 g of co-polymer was:
0.0369 TEOS, 0.0041 chlorosulfonyl-phenyl ethyltrimethoxy-
silane (CSPTMS, ABCR), 0.24 HCl, and ª6.67 H2O. The
amount of sulfur-containing precursor in both materials
(MPTMS and CSPTMS) was established to be 10 mol% of total
silicon species.
Hydrophobized arenesulfonic acid-functionalized mesostruc-
tured silica (Hydrophobized Ar-SBA-15) was synthesized in
order to increase the surface hydrophobicity of the above-
mentioned catalyst. For this purpose, 1 g of dried arenesulfonic
acid material was subjected to a silylation procedure of capping
the free surface silanol groups (Si–OH) by reacting them with
trimethylmethoxy silane (1 g) in dry toluene under reflux (12 h).
R
Other commercial catalysts used in this work were Nafionꢀ-
SiO2 composite (SAC-13) with resin content in the range of
10–20 wt%, supplied by DuPont, and an ionic-exchange sulfonic
acid-based macroporous resin, Amberlyst 15, supplied by Rohm
and Haas. Both catalysts were ground to powder in order to
minimize mass transfer limitations and thus avoid distortions
in the catalytic results. Also, arenesulfonic and propylsulfonic-
acid-functionalized non-ordered silicas, under the commercial
In this work, the synthesis of solketal (2,2-dimethyl-1,3-
dioxolan-4-methanol) from glycerol acetalisation with acetone
(Scheme 1) was carried out, for the first time, over different
sulfonic acid-modified mesostructured silicas. These materials,
characterized by high surface areas, large uniform pores, high
thermal stability, and the possibility of controlling the surface
R
R
names SiliaBondꢀ Tosic Acid and SiliaBondꢀ Propylsulfonic
Acid, were acquired from Silicycle directly in powder form.
2.2. Catalyst characterization
In order to characterize textural properties of the synthesized
catalyst, nitrogen adsorption and desorption isotherms were
measured at 77 K using a Micromeritics TRISTAR 3000
system. The data was analyzed using the BJH model and
Scheme 1 Main reaction products in the glycerol acetalisation with
acetone.
900 | Green Chem., 2010, 12, 899–907
This journal is
The Royal Society of Chemistry 2010
©

View Online
Table 1 Physicochemical, textural and acidity-related properties for sulfonic acid-modified mesostructured silicas
Textural Properties Acid Properties
Acid capacitye/
meq g-1
a
b
d
d100 /A Pore size /A BET area/m2 g-1 Pore volumec/cm3 g-1 Wall thickness /A Sulfur
H+
Accessibilityf (%)
˚
˚
˚
Sample
Pr-SBA-15
Ar-SBA-15
Hydrophobized 114
Ar-SBA-15
104
108
81
92
83
721
712
533
1.44
1.03
0.75
39
32
48
0.95
1.05
1.01
0.94 99
1.06 100
1.04 100
a d (100) spacing, measured from small-angle X-ray diffraction. b Mean pore size (Dp) from adsorption branch applying the BJH model. c The pore
√
volume (Vp) was taken at P/Po = 0.975 single point. d Average pore wall thickness calculated by ao-pore size (ao = 2 d(100)/ 3). e Acid capacities
defined as meq of acid centers per g of catalyst (obtained either directly by titration or indirectly from sulfur content by elemental analysis). f Defined
as the ratio between H+ from acid–base titration and sulfur content from elemental analysis.
total pore volume was taken at P/Po = 0.975 single point.
Structural ordering was determined by X-ray powder diffraction
(XRD) on a PHILIPS X’PERT diffractometer using Cu-Ka
radiation. Cationic-exchange capacities corresponding to the
sulfonic acid-modified mesostructured materials (acid capacity)
were measured using 2M NaCl (aq.) as cationic-exchange agent.
The resulting suspension was potentiometrically titrated at room
temperature by dropwise addition of 0.01 M NaOH (aq). Sulfur
and organic contents were determined by means of elemental
analysis in a Vario EL III apparatus, and thermogravimetric
analysis (SDT 2960 Simultaneous DSC-TGA, from TA Instru-
ments).
Table 1 summarizes the most relevant physicochemical prop-
erties for the sulfonic acid-modified mesostructured materials.
Data from XRD and nitrogen adsorption isotherms evidence
high mesoscopic ordering and high surface areas along with
narrow pore size distributions around 8–9 nm (size enough
to avoid the steric constraints imposed by the pore size when
relatively bulky substrates such as glycerol derivatives are con-
sidered). All the materials, prepared through a co-condensation
technique, display high incorporation yields for the sulfonic
acid moieties. It must be noted that the mesostructure of the
hydrophobized Ar-SBA-15 was not modified by the silylation
procedure, although the incorporation of trimethylsilyl species
on the silica mesopores produces an increase of the apparent
wall thickness accompanied by a slight decrease in pore volume
and BET surface area.
2.3. Reaction procedure
Crude, technical and pharmaceutical grade glycerols used in the
present work were kindly provided by Acciona Biocombustibles,
from the biodiesel production plant in Caparroso (Navarra,
Spain). The rest of the reagents used in the experiments and
sample analyses, acetone (99.93% purity), solketal (98% purity)
and 1,4-butanediol (99% purity), were purchased from Sigma-
Aldrich.
Scheme 1 is a representation of the acetalisation of glycerol
with acetone. This reaction yields the five- and the six-membered
ring isomers, whose relative formation depends on the acetali-
sation position within the glycerol molecule. In this case, the
five-membered ring compound, solketal, is obtained in a ratio of
99 : 1,5 and thus will be the only product considered. In addition,
water is a by-product of the reaction. Other secondary products
(e.g. polyglycerols) were not detected.
As previously stated, it is important to note that the present
reaction possesses an unfavourable equilibrium constant. Taking
this into account, to reach high conversions of glycerol it is neces-
sary to shift the equilibrium towards the formation of solketal,
feeding acetone in excess or continuously removing the water
generated during the reaction.23 For this reason, the reaction
was carried out following a two-step batch mode of operation.
In the first step, the reaction mixture (glycerol, acetone and the
catalyst) was stirred under reflux in a 100 mL round-bottom flask
fitted with a water-cooled condenser (refluxing stage). Agitation
was fixed at 500 rpm using a magnetic stirrer, and the reaction
temperature was achieved using a thermally controlled water
bath at 70 ◦C. Then, in a second step, the water produced during
the reaction along with the excess of acetone was removed by
extraction under vacuum keeping the flask at 70 ◦C. For the
study of this reaction system, the number of two-step batches has
Additionally, some characterization data corresponding to the
commercial sulfonic acid-based catalysts used in this study with
the purpose of comparison is summarized in Table 2. In this case,
the characterization is provided by the suppliers (Rohm & Haas
for the Amberlyst resin, DuPont for SAC-13 nanocomposite,
and Sylicycle for the functionalized silicas).
Table 2 Physicochemical properties corresponding to commercial sulfonic acid-based catalystsa
Acid capacity/meq H+ g-1
BET area/m2 g-1
Pore size/A
Pore volume/cm3 g-1
Max. Op. T/◦C
˚
Catalyst
Amberlyst-15
≥4.80
1.04
0.78
0.12
53
300
—
120
301b
279b
>200
20–200b
20–200b
>100
0.44
0.38
—
>200
>200
200
R
SiliaBondꢀ Propylsulfonic Acid
R
SiliaBondꢀ Tosic Acid
R
Nafionꢀ SAC-13
a Properties provided by the suppliers. b Experimentally determined by N2 adsorption-desorption isotherm at 77 K.
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 899–907 | 901
©

View Online
been considered as an additional reaction variable. Thus, every
batch consisted of a certain reaction time under reflux followed
by vacuum extraction, and the addition of fresh acetone to start a
new cycle. Typically, weight composition of the reaction mixture
was 5 g of glycerol, from 6.3 g to 18.9 g of acetone (which also
played the role of solvent), which means from a 2/1 to 6/1
acetone/glycerol molar ratio, and a constant catalyst loading of
0.25 g (5 wt% based on glycerol).
Reaction samples were analyzed by GC (Varian 3900 chro-
matograph) using a CP-WAX 52 CB column (30 m ¥ 0.25 mm,
DF = 0.25) and a flame ionization detector (FID). Catalytic
results are shown in terms of absolute conversion of glycerol.
Solketal was the only reaction product detected by GC. The GC-
quantified amount of this compound matches with the reacted
glycerol, indicating that no other secondary products are being
obtained in significant amounts.
3. Results and discussion
3.1. Screening of catalysts
First of all, a screening of different sulfonic acid-modified
solids as catalysts was performed. This comparative study was
focused on the analysis of catalytic properties such as acid site
strength and concentration, surface hydrophobicity, and porous
structure, and their effect on the catalytic conversion of glycerol
into solketal. For all the catalysts, reaction conditions were fixed
as follows: single batch (reaction time 30 min), bath temperature
70 ◦C (under reflux), acetone/glycerol molar ratio 6/1 and
pharmaceutical grade glycerol.
Fig. 1 Screening of sulfonic acid-modified catalysts for the acetalisation
of glycerol with acetone. Reaction conditions: tr 30 min, bath tempera-
ture 70 ◦C (under reflux), acetone/glycerol molar ratio 6/1, catalyst 5
wt% based on glycerol, pharmaceutical grade glycerol.
Fig. 1 shows the conversion of glycerol achieved by each
catalyst. Three groups of catalysts are shown: arenesulfonic acid-
functionalized silicas, propylsulfonic acid-functionalized silicas,
and sulfonic acid-modified commercial catalysts. The highest
conversion (85.1%) was obtained over the sulfonic acid resin
Amberlyst-15, despite its poor structural properties and low
surface area. Acid site concentration—almost 5 meq H+ g-1
in this catalyst—appears to be the most influential factor
affecting the progression of the reaction. Also, sulfonic acid-
functionalized silicas display high conversion rates, especially
those comprised of arenesulfonic acid catalytic sites, which
provide higher activities than the corresponding propylsulfonic
silicas (e.g., 82.5% for Ar-SBA-15 vs. 79.0% Pr-SBA-15). Thus,
the higher acid strength of the phenylsulfonic acid moiety rela-
tive to that of the alkylsulfonic acid site29 can also be considered
a beneficial factor for the development of the catalytic process.
Importantly, arenesulfonic acid-modified mesostructured SBA-
15 resulted in a glycerol conversion very close to that achieved
by the resin Amberlyst-15, even with a much lower acid site
concentration.
Comparing the results obtained with the three arenesulfonic
acid-modified silicas, the best performance was achieved by the
sample Ar-SBA-15. However, the same material after surface
hydrophobization—hydrophobized Ar-SBA-15—did not result
in an improved glycerol conversion but instead in a slight
decrease. The synthesis of this modified sample was directed
to displace the reaction equilibrium by affecting the local
concentration of water molecules inside the mesopores, in an
attempt to improve the overall glycerol conversion. The lower
conversion obtained using this catalyst compared to the parent
material can be ascribed to a combination of two factors. Firstly,
the passivation process reduces the available surface area from
739 to 533 m2 g-1, and secondly the hydrophobization of the
mesopores may also affect the local concentration of other polar
molecules aside from water (dielectric constant, e0 = 80.0), such
as acetone and glycerol (dielectric constants, 20.7 and 42.5,
respectively). Hence, the use of hydrophobic catalysts in this
reaction is not a valid approach. Likewise, the arenesulfonic
acid-functionalized non-ordered silica—Tosic acid-SiO2—gave
a significantly lower glycerol conversion likely because of its
reduced surface area as a consequence of its non-structured
nature. The same trend is observed for the propylsulfonic acid-
functionalized non-ordered silica—Pr-SO3H-SiO2. This fact
justifies the use of mesostructured silicas over non-ordered
silicas. Surprisingly, comparing both commercial silicas, and in
contrast to the previously discussed results for SBA-15 materials,
the propylsulfonic one gave a higher glycerol conversion despite
the lower acid strength of its acid sites. The lower acid capacity
of the Tosic acid-SiO2 sample, 0.78 vs. 1.04 meq H+ g-1, could be
the reason behind this discrepancy, and would also confirm that
the acid site concentration is the most influential parameter.
The low conversion obtained using the composite Nafion-
SAC-13 must also be noted. The low acid capacity of this
catalyst is considered the main cause of this result. A further
effect to be taken into account is the perfluorinated nature
of its sulfonic acid sites, which makes it the catalyst with the
highest acid strength and hence the most hydrophilic material.
902 | Green Chem., 2010, 12, 899–907
This journal is
The Royal Society of Chemistry 2010
©

View Online
Table 3 Experiment matrix and experiment results for the acetalisa-
The high affinity between perfluorosulfonic acid moieties and
water molecules could produce a relatively high local water
concentration around the catalytic acid sites, thus creating
an unfavourable microenvironment from the point of view of
reaction equilibrium. This would indicate that an excessive acid
strength would be as undesirable as a low acid strength.
tion of pharmaceutical glycerol with acetone over arenesulfonic acid-
modified mesostructured silica (Ar-SBA-15) [catalyst/glycerol weight
ratio = 5%; T = 70 ◦C; tr = 30 min before each water removal]
Run Number
B
MR
IB
IMR
XG (%)
1
2
3
4
5
6
7
9
3
3
1
1
2
2
2
2
3
2
1
6/1
2/1
6/1
2/1
4/1
4/1
4/1
6/1
4/1
2/1
3/1
+1
+1
-1
-1
0
0
0
0
+1
0
+1
-1
+1
-1
0
0
0
+1
0
-1
0
89.5
61.8
82.5
54.3
80.7
78.0
75.5
86.7
85.4
59.2
73.5
3.2. Optimization of the reaction conditions
Preliminary kinetic experiments were carried out to determine
the evolution of the glycerol conversion with the reaction time
for the acetalisation of pharmaceutical glycerol with acetone
over arenesulfonic acid-functionalized mesostructured silica
(Ar-SBA-15) since this catalyst has demonstrated an excellent
catalytic behaviour in this reaction. All the catalytic tests were
performed under the same reaction conditions, increasing the
refluxing time. Fig. 2 shows that glycerol conversion reaches
a constant value (80%) after only 15–30 min, indicating that
the reaction achieves the equilibrium state. Therefore, higher
refluxing times are not necessary since glycerol conversion keeps
constant.
10
11
12
-1
B, number of 2-step batches; MR, acetone/glycerol molar ratio; I, coded
value; XG, conversion of glycerol.
Table 4 Predictive equations obtained by design of experiments
Statistical model
2
XG = 78.55 + 4.4IB + 13.9IMR + 0.17IB
-
r2 = 0.983
r2 = 0.983
(1)
(2)
2
6.33 IMR - 0.12 IBIMR
Technological model
XG = 16.9 + 3.84B + 19.76MR + 0.198B2 -
1.588MR2 - 0.0575B ¥ MR
B, number of 2-step batches; MR, acetone/glycerol molar ratio; I, coded
value; XG, conversion of glycerol.
solketal using a large excess of acetone or by removing the
water produced from reaction media. As a result, the chosen
factors were the number of 2-step batches and the molar ratio of
acetone to glycerol, MR. The lower and upper numbers of 2-step
batches were 1 and 3 and the levels of MR were 2/1 and 6/1.
Thus, the standard experimental matrix for the design is shown
in Table 3. Columns 4 and 5 represent the 0 and 1 encoded
factor levels on a dimensionless scale, whereas columns 2 and 3
represent the factor levels on a natural scale. Experiments were
run randomly to minimize errors due to possible systematic
trends in the variables. Table 3 also shows the experimental
results obtained for the glycerol conversion.
Fig. 2 Evolution of glycerol conversion with the refluxing time over
arenesulfonic acid-modified mesostructured silica (Ar-SBA-15). Reac-
tion conditions: bath temperature 70 ◦C (under reflux), acetone/glycerol
molar ratio 6/1, catalyst 5 wt% based on glycerol (0.25 g), pharmaceu-
tical grade glycerol.
From the matrix generated by the experimental data and
assuming a second-order polynomial model, equations 1 and
2 were obtained by multiple regression analysis (Table 4).
The statistical model is obtained from coded levels giving
the real influence of each variable on the process, whereas
the technological model is obtained from the real values of
the variables. Consequently, the influence of variables on the
response is discussed using the statistical model shown in eqn
(1) (Table 4).
Statistical analysis of the studied experimental range identifies
the acetone : glycerol molar ratio (IMR) as the most important
factor in the glycerol conversion response. The second factor
in importance is the number of batches (IB) followed by the
quadratic effect of the molar ratio (IMR2). The first two have
a positive effect on the glycerol conversion: an increase in the
molar ratio and the number of batches produces an increase
in the conversion of glycerol. But the enhancement of this
The production of acetal from glycerol using an arene-
SO3H-functionalized mesostructured silica (Ar-SBA-15) as the
catalyst was developed and optimized by following factorial
design and response surface methodology.32 In order to shift the
equilibrium towards higher production of solketal, improving
in this way the glycerol conversion, a two-step batch operation
methodology as described in the experimental section was used.
The experimental design applied to this study was a full 32
design (two factors, each one at three levels). The central point
experiment was repeated three times in order to determine
the variability of the results and to assess the experimental
error. The selected response was the glycerol conversion, XG,
since the main objective was to achieve, if possible, a complete
conversion of glycerol. Selection of the factors was based on
the results obtained in preliminary studies and on the fact that
the equilibrium has to be driven towards the production of
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 899–907 | 903
©

View Online
Fig. 3 Response surface and contour plots for glycerol conversion over the catalyst Ar-SBA-15 predicted by the model.
response by IMR is three times that of IB. The quadratic effect
of the molar ratio has a significant negative influence on the
glycerol conversion. This, in turn, indicates that the increase in
this operating variable does not produce a constant rise in the
glycerol conversion, because the curvature effect is significant
at high molar ratios. In addition, the rest of the coefficients,
the quadratic effects of the number of batches (IB2) and the
number of two step batches-molar ratio interaction (IB - IMR),
have no significant influence on the glycerol conversion since
their absolute values are smaller than the corresponding main
effects.
In Fig. 3, the technological model (equation 2, Table 4)
is represented as a response surface and contour plots for
predicted values of glycerol conversion over the experimental
range studied. It clearly shows an enhanced conversion of
glycerol at high acetone/glycerol molar ratios and with a high
number of batches. Thus, the optimal values are the highest
molar ratio (6/1) and the highest number of 2-step batches (3).
At these operating conditions, the glycerol conversion predicted
by the non-linear model is 90.6%.
Fig. 4 Accuracy of the predicted data relative to the experimental data
for the conversion of glycerol (catalyst: Ar-SBA-15).
molar ratio. In this sense, a large excess of acetone can drive the
equilibrium towards the production of solketal, but the amount
of acetone needed for doing so is too high to be practically
and economically effective. On the other hand, the increase in
the number of 2-step batches leads only to a slight increase of
the glycerol conversion which becomes almost constant over 3
consecutive 2-step batches.
Finally, the arithmetical average and the standard deviation
of the response were calculated from the central point replicas:
glycerol conversion 78% 2.6%. The standard deviation was
lower than 5%. Therefore, the experimental error corresponding
to the results shown in Table 4 is not excessively significant which
indicates that the models accurately represent the influence of
acetone/glycerol molar ratio and number of 2-step batches
on glycerol conversion over the experimental range studied.
Likewise, Fig. 4 shows the relationship between experimental
and predicted values. As can be observed, values calculated with
the predictive equation are very close to those obtained exper-
imentally, indicating again the high accuracy of the obtained
models.
In view of the results from the experimental design, where
a constant value of glycerol conversion was not reached, an
extension of the experimental design was made. In order to
improve the glycerol conversion response, some additional
reactions were carried out increasing the acetone/glycerol molar
ratio up to 8/1 and 10/1 and the number of 2-step batches
up to 4 and 5. Fig. 5 confirms the conclusion obtained from
the experimental design since the most influential factor on
the reaction of the glycerol acetalisation is the acetone/glycerol
3.3. Applicability to different grades of glycerol and catalyst
reusability
Another purpose of the present work was to evaluate the
possibility of using low-grade glycerol for the reaction of
acetalisation with acetone. Specifically, crude and technical
grades have been tested. Table 5 includes mass composition of
the different grades of glycerol considered in this work. The
presence of salts and/or water adversely affects the catalytic
performance. As discussed above, the presence of water in the
medium imposes a thermodynamic barrier, limiting the reaction.
Likewise, sodium cations can deactivate sulfonic acid-modified
materials by simple cationic-exchange of the catalytic protons in
the sulfonic acid moieties.
904 | Green Chem., 2010, 12, 899–907
This journal is
The Royal Society of Chemistry 2010
©

View Online
Fig. 5 Extension of the design experiment. Reaction conditions: catalyst Ar-SBA-15, bath temperature 70 ◦C (under reflux), catalyst 5 wt% based
on glycerol, tr = 30 min before each water removal. (a) Acetone : glycerol molar ratio 6/1 (b) Number of consecutive 2-step batches: 1.
Table 5 Composition of the different grades of glycerol evaluated
those performed with acid solid catalysts. Furthermore, the
result of the blank reaction for pharmaceutical glycerol notably
differs from those with technical and crude glycerols (below
10% conversion even in the optimized reaction conditions). This
is a clear indication of the negative effect of the presence of
water on the reaction progress. Nevertheless, in the presence
of Amberlyst-15 or Ar-SBA-15, high activities are obtained
even for the low-purity glycerol grades. The observed trend
is similar for both catalysts, with only a minimal decrease in
conversionwhen treatingtechnical andcrude glycerols relativeto
pharmaceutical glycerol. This behaviour can be partly attributed
to the efficient water removal under vacuum between consecutive
batches, which would minimize the detrimental effect of the
presence of water in the reaction medium as well as the presence
of a catalytic system. Concerning the presence of sodium cations,
surprisingly it does not seem to have a strong impact on the
catalytic performance. However, a certain degree of Na+/H+
cationic-exchange can be expected to occur so that part of the
actual catalytic effect would likely come from homogeneous
released protons. If that is the case, then the heterogeneous
(wt%)
Ash (%)
Glycerol grade Purity (%) Water (%) NaCl
Others MONG (%)
Pharmaceutical 99.9
0.1
6.2
8.0
< 0.001 n.d
< 0.001 0.2
5.2 0.94
n.d
2
0.06
Technical
Crude
91.6
85.8
Others: non-NaCl inorganic compounds; MONG: matter (organic) non-
glycerol; n.d., not detected.
Fig. 6 shows the glycerol conversions achieved with the
different grades of glycerol in a blank reaction—no catalyst
added—and with Amberlyst-15 and Ar-SBA-15 as the most
active catalysts in this reaction. Reaction conditions were those
above optimized in the experimental design, an acetone/glycerol
molar ratio of 6/1 and 3 two-step batches of 30 min.
As expected, reactions performed without catalyst resulted
in substantially reduced glycerol conversions as compared to
Fig. 6 Influence of the glycerol grade. Reaction conditions: tr 30 min before each water removal, 3 consecutive 2-steps batches, bath temperature
70 ◦C (under reflux), acetone/glycerol molar ratio 6/1, catalyst 5 wt% based on glycerol.
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 899–907 | 905
©

View Online
Fig. 7 Direct reutilization of catalyst Ar-SBA-15 for the different grades of glycerol. Reaction conditions: every run consists of 3 two-step batches
of 30 min, bath temperature 70 ◦C (under reflux), acetone/glycerol molar ratio 6/1, catalyst 5 wt% based on glycerol.
catalysts would be transformed into the corresponding sodium
forms, and hence they would be deactivated.
4. Conclusions
In order to confirm this hypothesis, a reutilization test using
Ar-SBA-15 as the catalyst in three consecutive reaction cycles for
each glycerol grade under the optimized reaction conditions was
performed. The reuse procedure was kept as simple as possible,
without washing or regenerating the catalyst, just separation by
Sulfonic acid-functionalized mesostructured silicas have demon-
strated an excellent catalytic behaviour in the acetalisation
of bio-glycerol and acetone to produce solketal as a fuel
component. The activities achieved in the acetalisation of
this glycerol using sulfonic acid-functionalized mesostructured
silicas, especially those comprised of arenesulfonic acid-catalytic
sites, are comparable to those displayed by the widely used
macroporous commercial acid resin (Amberlyst-15). The acid
capacity of the catalysts is the most significant factor affecting
the catalytic performance of this reaction. The acid strength
of the catalytic sites has also proved to be an influential
parameter. However, the use of hydrophobized catalysts (e.g.
hydrophobized arene- and propyl-sulfonic mesostructured ma-
terials) in the acetalisation of glycerol with acetone is not an
adequate approach, because these modified catalysts have a
lower available surface area and the hydrophobized mesopores
limit the access of highly polar glycerol and acetone molecules.
The experimental design model carried out for different levels
of two-step batch number and acetone to glycerol molar ratio
with the arenesulfonic acid-modified mesostructure catalyst
has shown that it is necessary to use the highest molar ratio
(6/1) and the highest number of two-step batches (3) in
order to maximize the conversion of pharmaceutical grade
glycerol. The glycerol conversions obtained using crude and
technical grade glycerols are similar to those achieved using
the high purity pharmaceutical grade glycerol in the presence
of arenesulfonic acid-modified mesostructured silica and the
macroporous sulfonic acid resin as catalysts. These activities are
maintained after three reaction cycles using pharmaceutical and
technical grade glycerols, in which both catalysts are reutilized
without an intermediate purification stage, indicating that these
catalysts are not deactivated. However, using crude glycerol,
both catalysts need to be regenerated in order to achieve high
glycerol conversions.
◦
centrifugation and overnight drying at 90 C before being use
again. Fig. 7 illustrates the experimental results of this study
in terms of glycerol conversion. The high activity observed
initially for technical and pharmaceutical glycerols is maintained
after 3 runs, indicating that no deactivation is produced. This
is consistent with the very low sodium content of these two
types of glycerol, which will not lead to a relevant cation-
exchange deactivation effect. However, in the case of crude
glycerol, an important catalytic deactivation is observed already
in the second run, as glycerol conversion dramatically drops to
values close to zero. The high sodium content of crude glycerol
is considered the main factor responsible for this situation,
very likely through an almost complete cation exchange of the
arenesulfonic protons by sodium ions during the first reaction
cycle.
An attempt to recover the activity of the catalyst Ar-SBA-
15 after use with crude glycerol was made using the following
procedure: separation of the solid by centrifugation, ethanol
washing under reflux for 2 h, acidification with 2M HCl (aq.)
for 30 min to produce the reverse cation exchange, washing
with deionised water until neutral pH, and drying overnight
at 90 ◦C. After this regeneration treatment, the Ar-SBA-15
catalyst was evaluated again in reaction with crude glycerol
and using the same reaction conditions shown in Fig. 7.
Resultant glycerol conversion was 77%, i.e. an almost complete
recovery of the initial activity. This confirms that the catalyst
deactivation is due to a Na+/H+ cationic-exchange, which
can be readily reversed by means of the above regeneration
procedure.
906 | Green Chem., 2010, 12, 899–907
This journal is
The Royal Society of Chemistry 2010
©

View Online
15 K. Klepacova, D. Mravec, A. Kaszonyi and M. Bajus, Appl. Catal.,
Acknowledgements
A, 2007, 328, 1–13.
16 R. S. Karinen and A. O. I. Krause, Appl. Catal., A, 2006, 306, 128–
133.
The financial support from the “Ministerio de Ciencia e
Innovacio´n” through the project CTQ2008-01396 and CENIT
Program (2006) through a research project entitled “Proyecto
de innovacio´n para el impulso del biodie´sel en Espan˜a” (PiIBE)
are gratefully acknowledged.
17 J. A. Melero, G. Vicente, G. Morales, M. Paniagua, J. M. Moreno, R.
Rolda´n, A. Ezquerro and C. Pe´rez, Appl. Catal., A, 2008, 346, 44–51.
18 J. Delgado, Spanish Patent 1 331 260, 2002.
19 J. A. Melero, R. van Grieken, G. Morales and M. Paniagua, Energy
Fuels, 2007, 21, 1782–1791.
20 M. I. Galan, J. Bonet, R. Sire, J. M. Reneume and A. E. Plec¸u,
Bioresour. Technol., 2009, 100, 3775–3778.
21 X. Liao, Y. Zhu, S. G. Wang and Y. Li, Fuel Process. Technol., 2009,
90, 988–993.
22 H. Gerard, D. Bruno, D. Isabelle, WO Patent 093 015 A1, 2005.
23 B. Bruchmann, K. Ha¨berle, H. Gruner, M. Hirn, U.S. Patent
5 917 059, 1999.
References
1 G. Vicente, M. Mart´ınez and J. Aracil, Bioresour. Technol., 2004, 92,
297–305.
2 B. Freedman, E. H. Pryde and T. L. Mounts, J. Am. Oil Chem. Soc.,
1984, 61, 1638–43.
3 European Biodiesel Board, http://www.ebb-eu.org/index.php.
4 M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi and C. Della Pina,
Angew. Chem., Int. Ed., 2007, 46, 4434–40.
5 C. Zhou, J. N. Beltramini, Y. Fana and G. Q. Lu, Chem. Soc. Rev.,
2008, 37, 527–549.
6 Y. Zheng, X. Chen and Y. Shen, Chem. Rev., 2008, 108, 5253–5277.
7 A. Behr, J. Eilting, K. Irawadi, J. Leschinski and F. Lindner, Green
Chem., 2008, 10, 13–30.
8 Y. Gu, A. Azzouzi, Y. Pouilloux, F. Je´roˆme and J. Barrault, Green
Chem., 2008, 10, 164.
24 B. Delfort, I. Durand, A. Jaecker, T. Lacome, X. Montagne, F. Paille,
U.S. Patent 0 163 949, 2003.
25 J. Deutsch, A. Martin and H. Lieske, J. Catal., 2007, 245, 428–435.
´
26 E. Garc´ıa, M. Laca, E. Pe´rez, A. Garrido and J. Peinado, Energy
Fuels, 2008, 22, 4274–4280.
27 C. X. Da Silva, V. L. Gonc¸alves and C. J. Mota, Green Chem., 2009,
11, 38–41.
28 R. Wessendorf, Petrochem., 1995, 48, 138–143.
29 J. A. Melero, R. van Grieken and G. Morales, Chem. Rev., 2006, 106,
3790–3812.
9 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107, 2411.
10 V. P. Gupta, U.S. Patent 5 476 971 to ARCO Chem. Technology, L.P,
1995.
11 D. S. Bradin, U.S. Patent 5 578 090, BRI, 1996.
12 H. Noureddini, U.S. Patent 6 015 440, 2000.
13 H. Noureddini, U.S. Patent 6 174 501, 2001.
14 A. Behr and L. Obendorf, Eng. Life Sci., 2002, 2, 185–189.
30 G. Vicente, M. Mart´ınez and J. Aracil, Bioresour. Technol., 2007, 98,
1724–33.
31 R. van Grieken, J. A. Melero and G. Morales, Appl. Catal., A, 2005,
289, 143–152.
32 G. E. P. Box, W. G. Hunter and J. S. Hunter, Statistics for
Experimenters: An Introduction to Design, Data Analysis, and Model
Building, Wiley, New York, 1978.
This journal is
The Royal Society of Chemistry 2010
Green Chem., 2010, 12, 899–907 | 907
©