Chemicals from biomass: Synthesis of glycerol carbonate by transesterification and carbonylation with urea with hydrotalcite catalysts. The role of acid-base pairs

Article abstract of DOI:10.1016/j.jcat.2009.11.001

Synthesis of glycerol carbonate has been performed by transesterification of ethylene carbonate with glycerol catalyzed by basic oxides (MgO, and CaO), and mixed oxides (Al/Mg, Al/Li) derived from hydrotalcites. The results showed that the optimum catalyst in terms of activity and selectivity is a strong basic Al/Ca-mixed oxide (AlCaMO) which is able to catalyze the reaction at low temperature (35 °C), and low catalyst loading (0.5 wt%) giving high glycerol conversions with 98% selectivity to glycerol carbonate. When the synthesis of glycerol carbonate was carried out by carbonylation of glycerol with urea, the results showed that balanced bifunctional acid-base catalysts where the Lewis acid activates the carbonyl of the urea and the conjugated basic site activates the hydroxyl group of the glycerol were the most active and selective catalysts.

Full text of DOI:10.1016/j.jcat.2009.11.001

Journal of Catalysis 269 (2010) 140–149
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Chemicals from biomass: Synthesis of glycerol carbonate by transesterification
and carbonylation with urea with hydrotalcite catalysts. The role of acid–base pairs
*
Maria J. Climent, Avelino Corma , Pilar De Frutos, Sara Iborra, Maria Noy,
Alexandra Velty, Patricia Concepción
Instituto de Tecnologia Quimica, Universidad Politecnica de Valencia.-C.S.I.C., Camino de Vera, E-46022 Valencia, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 25 May 2009
Revised 9 September 2009
Accepted 1 November 2009
Available online 5 December 2009
Synthesis of glycerol carbonate has been performed by transesterification of ethylene carbonate with
glycerol catalyzed by basic oxides (MgO, and CaO), and mixed oxides (Al/Mg, Al/Li) derived from hydro-
talcites. The results showed that the optimum catalyst in terms of activity and selectivity is a strong basic
Al/Ca-mixed oxide (AlCaMO) which is able to catalyze the reaction at low temperature (35 °C), and low
catalyst loading (0.5 wt%) giving high glycerol conversions with 98% selectivity to glycerol carbonate.
When the synthesis of glycerol carbonate was carried out by carbonylation of glycerol with urea, the
results showed that balanced bifunctional acid–base catalysts where the Lewis acid activates the car-
bonyl of the urea and the conjugated basic site activates the hydroxyl group of the glycerol were the most
active and selective catalysts.
Keywords:
Glycerol carbonate
Glycerol transesterification
Urea carbamoilation
Hydrotalcites
Ó 2009 Elsevier Inc. All rights reserved.
Bifunctional acid–base catalysts
1. Introduction
to react with anhydrides [7] to form ester linkages or with isocya-
nates to form urethane linkages [8]. These materials are used for
Much interest has been developed in the last years for the use of
renewable feedstocks [1] and fats and oils not competing with food
can become very important players in the chemical industry for the
near future. Their competitive cost, worldwide availability, and
built-in functionality make them attractive for numerous commer-
cial applications. One application is the production of biodiesel by
transesterification processes with methanol or ethanol. The grow-
ing production of biodiesel has generated a surplus of glycerine
and its price can be lower than that of propylene glycol or sorbitol.
Glycerol is a synthesis intermediate for the preparation of a
large number of compounds via oxidation [2], etherification [3],
esterification, transesterification [4], polymerisation, etc. There-
fore, it is not surprising that a large effort has been devoted in
the last years for converting glycerol into high value-added chem-
icals [5].
the production of polyurethane protecting coatings for wood and
metal substrates. Besides, GC is the most valuable intermediate
for the production of glycidol which is a precursor for the synthesis
of polymers [9].
The main methods for the preparation of GC are based on the
reaction of glycerol with (a) a carbonate source (phosgene, a dial-
kyl carbonate [10] or an alkylene carbonate), (b) urea, and (c) car-
bon monoxide and oxygen. Traditionally, cyclic carbonates have
been prepared by reaction of glycols with phosgene, but due to
the high toxicity and corrosive nature of phosgene alternative
routes such as transesterification reaction of dialkyl or alkykene
carbonates to obtain cyclic carbonates have been explored [11–
15]. For instance, ethylene carbonate, a commercial product with
interesting physical properties (low toxicity, low evaporation rate,
biodegradability, high solvency, etc.) has been widely used as car-
bonate source for preparing GC. The transesterification between
glycerol and ethylene carbonate (EC) to produce GC is performed
with an alkaline base. For instance, using sodium bicarbonate, at
130 °C, the yield of glycerol carbonate reached 81% after 30 min
[11]. However, the main problem associated with these catalysts
is the requirement of a final neutralization step. For doing this, at
the end of the reaction, a mineral acid (phosphoric acid, sulphuric
acid, benzenesulphonic acid, etc.) is added to the system to neu-
tralize the catalysts, followed by distillation under reduced pres-
sure for the recovery of GC from the reaction mixture, which
contains reactants, salts and products.
One important glycerol derivative is the glyceryl carbonate (GC)
which is widely used as protic solvent (in resins and plastics) [6],
additive, and as chemical intermediate. Due to its low toxicity,
low evaporation rate, low flammability, and moisturizing ability,
GC is used as wetting agent for cosmetics and carrier solvent for
medical preparations.
On the other hand, the presence of a cyclic carbonate group
along with a primary hydroxylmethyl group allows the molecule
* Corresponding author.
E-mail address: acorma@itq.upv.es (A. Corma).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.11.001

M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
141
Despite the fact that heterogeneous base catalysts should allow
easy separation and recycling of the catalyst by filtration avoiding
the neutralization step and reducing waste formation, there are
few examples for the synthesis of GC by transesterification of EC
and glycerol in the presence of solid catalysts. Thus, Sugita et al.
[12] have performed the transesterification reaction in the pres-
ence of aluminium oxide as catalyst. When the reaction was car-
ried out at 135–140 °C under reduced pressure with progressive
removal of ethylene glycol, GC is obtained with 99% yield.
More recently, a process for the preparation of GC by transeste-
rification of ethylene carbonate with glycerol in the presence of an
anionic bicarbonated or hydroxylated macroporous resin (Amber-
lyst A26) and basic form zeolites (Y, X, A), has been reported in
the patent literature [13]. In this case, when the transesterification
reaction was carried out in the presence of basic KX zeolite in a
molar ratio ethylene carbonate/glycerol of 2, at 80 °C, 81% yield
of GC was obtained after 2 h reaction time. When a reactor which
allows the distillation of ethylene glycol formed was used 99%
yield of GC was obtained.
The transcarbonation between ethylene carbonate and glycerol
in supercritical CO2 using zeolites or strongly basic resins has also
been performed, though in this case the yields of GC were lower
than 25% [14]. Concerning the preparation of GC from urea and
glycerol, the same authors have patented a process for preparing
GC using metallic or organometallic salts, or supported metallic
compounds. The best results (GC yield 80%) were obtained using
a calcined ZnSO4 as catalyst with a molar ratio glycerol/urea of 1,
at 150 °C and 40 mbar, while removing the ammonia formed dur-
ing the reaction [15]. Okutsu and Kitsuki [16] have prepared GC by
reacting glycerol with urea with a metal oxide such as ZnO in the
presence of a dehydrating agent such as anhydrous MgSO4. Thus
a mixture of glycerol and anhydrous magnesium sulphate was
heated at 120 °C in the presence of nitrogen. After 2 h, urea and
zinc oxide were added. After 6 h reaction time, the conversion of
glycerol was 62–65% with selectivity close to 92%. Better results
have been recently reported by the same authors using ZnSO4 as
catalyst [17]. Performing the reaction at 130 oC, 90% yield of GC
was obtained after 10 h reaction time. However, when the reaction
was performed using MgSO4 as a catalyst and in the presence of
diethylene glycol diethyl ether as solvent only 58% yield of GC
was obtained after 24 h reaction time [18].
The Al/Mg hydrotalcite was prepared by co-precipitation at
constant pH and controlled by slow addition in a single container
of two diluted dilutions (A and B), namely, solution A containing
Mg(NO3)2 and Al(NO3)3, 1.5 M in Al + Mg with Al/Al + Mg atomic
ratio equal to 0.25, and solution B prepared by dissolving Na2CO3
and NaOH in water in such way that the ratio CO23ꢀ/(Al + Mg) is
equal to 0.66.
In order to prepare an Al/Ca hydrotalcite, the same methodol-
ogy was followed using Ca(NO3)2 as solution A in this case.
The Al/Zn/Li hydrotalcite was prepared following the same
methodology using Al(NO3)3, Zn(NO3)2 (1.5 M in Al + Zn) + LiCl
(5 mol% respect to Al + Zn) with Al/(Al + Zn + Li) atomic ratio equal
to 0.25, as solution A.
The Al/Mg/Fe hydrotalcite was prepared following the same
methodology using Al(NO3)3, Mg(NO3)2 and Fe(NO3)2 (Fe = 5 mol%
respect to Mg + Fe), solution 1.5 M in Al + Mg + Fe, with Al/
(Al + Mg + Fe) atomic ratio equal to 0.33 as solution A.
In all cases, the solutions A and B were mixed at 60 mLhꢀ1 addi-
tion rate, for 4 h, with vigorous stirring. The final pH was of 13. The
co-addition process was carried out at room temperature. The
sample was aged at 60 oC for 24 h. Finally, the precipitate was fil-
tered and washed to eliminate the alkali metals and the nitrate
ions until the pH of the washing water was 7 and then dried at
60 oC for 12 h.
The Al/Li layered double-hydroxyl carbonate ([Al2Li(OH)6]2-
CO3nH2O) with a Al/(Al + Li) molar ratio of 0.33 was prepared by
adding a hexane solution of aluminium-tri-(sec-butoxide) (17 wt/
wt%) to a lithium carbonate aqueous solution (0.55 wt/wt%) under
vigorous mechanical stirring at room temperature. The suspension
was left for 24 h at 60 °C. The Al/Li hydrotalcite was filtered and
washed until pH 7 and the solid was dried at 60 °C.
The Al/Zn hydrotalcite was synthesized by mixing an aqueous
solution of zinc nitrate hexahydrate and aluminium nitrate
nonahydrate with a basic aqueous solution containing sodium
hydroxide and sodium nitrate. The molar ratios of the reagents
Zn(NO3)26H2O/Al(NO3)29H2O/NaNO3/NaOH were 2.1:0.9:4:4.9.
The metal and the basic solutions were added together at a rate
of 1 mL/min under vigorous stirring. The resulting gel of pH 5.84
was aged at 60 °C for 18 h. The hydrotalcite formed was dried at
100 °C overnight.
The different hydrotalcite-like materials synthesized were ana-
Finally, GC has also been directly prepared by oxidative carbon-
ylation of glycerol by reacting glycerol with a gaseous mixture of
carbon monoxide and oxygen. Thus, Teles el al. [19] have described
a process that uses salts of the elements of groups Ib, IIb, and VIIb
(copper, silver, mercury, cobalt, etc.) as catalysts. More specifically,
the reaction takes place in an inert solvent such as nitrobenzene at
130 °C, with a gas mixture of CO/O2 (95:5%) at 8 bar in the presence
of copper (I) chloride as catalyst. Under these reaction conditions
and after 20 h, 96% yield of GC was obtained.
lyzed by X-ray diffraction (XRD) with a Phillips PW diffractometer
using Cu Ka radiation. Fig. 1 shows the XRD patterns of the differ-
ent hydrotalcite-like materials. As can be observed HTAl-Mg, HTAl-
Mg-Fe, and HTAl-Li showed the characteristic peaks corresponding
In this work we have studied the possibility of preparing pre-
pare GC in high yields and selectivity using solid catalysts with
adequate acid–base pairs. For doing this, we have chosen the reac-
tion of transcarbonation of glycerol with ethylene carbonate and
carbonylation of glycerol with urea.
HTAl-Zn
HTAl-Mg
2. Experimental
HTAl-Mg-Fe
2.1. Materials
HTAl-Li
AlCaMO
Glycerol (99.5%), ethylene carbonate, and urea were purchased
from Aldrich and used without further purification. The ZnO and
CaO commercial samples were purchased from Fluka. MgO was
purchased from NanoScale Materials Inc. Sn-Beta zeolite was syn-
thesized according to the literature procedures [20].
0
20
40
2θ
60
80
Fig. 1. XRD patterns of the different basic catalysts.

142
M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
to well-crystallized hydrotalcite structure. In the case of HTAl-Zn,
besides the characteristics peaks of hydrotalcite phase, some impu-
rities corresponding possibly to Zn carbonates and oxyhydroxycar-
bonates are also observed. On the other hand, the XRD pattern of
the Al/Ca material showed that it does not correspond to a
hydrotalcite structure, but to a mixture of CO3Ca and Al(OH)3
phases.
For its use as basic catalysts, hydrotalcites were calcined at
450 °C for 6 h in a nitrogen flow in order to obtain the correspond-
ing mixed oxides and denoted as HTc-Mg, HTc-Li, HTc-Zn, HTc-Zn-
Li, and HTc-Fe. The mixture of CO3Ca and Al(OH)3 was calcined at
450 °C for 6 h in a nitrogen flow and labeled as AlCaMO (Al/Ca-
mixed oxide).
In the case of Al/Mg hydrotalcite, the resulting mixed oxide was
hydrated at room temperature by adding directly 36 wt% of decar-
bonated water (MilliQ) just before the use and is denoted as HTr.
For the screening of catalytic activity all the catalysts have just
been calcined before their use.
The materials synthesized were analyzed by Inductively Cou-
pled Plasma (ICP) using a Varian 715-ES ICP Optical Emission Spec-
trometer and by X-ray diffraction (XRD) with a Phillips PW
line. In a typical experiment glycerol was placed under reaction
conditions for 10 min before urea and catalyst were added. The
amount of catalyst used was always 5% by weight of the initial
amount of glycerol. The reaction was stirred at a constant rate of
300 rpm, and heated in an oil bath at the desired temperature.
Reactions were run under a reduced pressure and in the absence
of a solvent. After the reaction was completed, methanol was
added and the catalyst removed by filtration. The catalyst was
washed with acetone to remove the adsorbed products. The filtrate
fractions were concentrated under reduced pressure and the prod-
ucts were identified by GC–MS. Product quantification was made
by using the response factor of the different compounds and tetra-
ethylene glycol (TEG) as external standard. Conversion, selectivity,
and yield were calculated with respect to glycerol.
The samples were analyzed with a Varian 3900 gas chromato-
graph equipped with
a
TR-WAX column (15 m ꢁ 0.32 mm ꢁ
0.25
lm) and flame ionization detector.
GC–MS measurements were made with an Agilent 6890N Net-
work GC System equipped with a Tracer TR-WAX column coupled
to an Agilent 5973 Network Mass Selective Detector.
diffractometer using Cu Ka radiation. Specific surface areas of the
2.4. Characterization: FTIR spectroscopy
different materials were obtained with an ASAP 2000 (Micromeri-
tics) using the BET methodology. The main characteristics of solid
catalysts are summarized in Table 1.
Transmission FTIR spectra were recorded on a BioRad FTS-40A
spectrometer. The samples were pressed into self-supporting wa-
fers and placed in a stainless steel cell for low temperature exper-
iments (CO adsorption at ꢀ175 °C) and in a quartz cell for room
temperature experiments (CHCl3 adsorption at 25 °C). Prior to
the adsorption experiments the samples were activated in situ in
a nitrogen flow at 450 °C for 3 h followed by evacuation at the
same temperature at 10ꢀ5 mbar. CO adsorption experiments were
done at ꢀ175 °C increasing the CO partial pressure (0.2–45 mbar).
After CO saturation at ꢀ175 °C, the samples were evacuated and
the spectra were collected at increasing temperature (ꢀ175 °C to
ꢀ130 °C) under dynamic vacuum conditions. CHCl3 adsorption
experiments were done at 25 °C increasing the CHCl3 partial pres-
sure (1–23 mbar). The spectra were treated and deconvoluted
using an Origin software.
2.2. Reactionprocedure: transcarbonation of ethylene glycol carbonate
with glycerol
The experiments were performed in a glass batch reactor
equipped with a condenser system. The catalyst was calcined just
before its use at 450 °C and added to the reactor at the same time
that reactants. The system of reaction was stirred and heated in a
silicone bath to the required temperature in the absence of solvent,
under nitrogen atmosphere. Samples were periodically taken out
of the reactor and separated from the catalyst by filtration and ana-
lysed by gas chromatography. The gas chromatograph (Varian
3900) was equipped with a Trocar TR-WAX column (15 m ꢁ 0.32
mm ꢁ 0.25 m) and a flame ionization detector. After the comple-
tion of the reaction, the reaction mixture was filtered and the cat-
alyst washed with acetone to wash off the products adhered on the
surface of the catalyst. The combined filtrate fractions were con-
centrated under reduced pressure and the products were identified
by GC–MS. Nitrobenzene was used as internal standard, and the
conversion, selectivity and yield were calculated with respect to
the limiting reactant.
2.5. XPS spectroscopy
XPS spectra were recorded on a SPECS spectrometer equipped
with a Phoibos 150 9MCD detector using a non-monochromatic
X-ray source (Al and Mg) operating at 200 W. The samples were
pressed into a small disc and evacuated in the prechamber of the
spectrometer at 1 ꢁ 10ꢀ9 mbar. Some of the samples have been
activated in situ in nitrogen flow at 450 °C for 3 h followed by evac-
uation at 10ꢀ8 mbar. The measured intensity ratios of components
were obtained from the area of the corresponding peaks after non-
linear Shirley-type background subtraction and corrected by the
transmission function of the spectrometer. Casa software has been
used for quantification and spectra treatment. Binding energy (BE)
of the peaks have been corrected by charging effect according to
Table 7.
2.3. Carbonylation of glycerol with urea
The reactions were performed in a 10 mL round-bottom flask
with a side arm and a 90° angle adapter to connect it to a vacuum
Table 1
Main characteristics of solid catalysts.
Catalyst
BET area
Pore volume
Al/Al + M
Molar ratio
(m2 gꢀ1
)
(cc gꢀ1
)
3. Results and discussion
HTc-Mg0.20
HTc-Mg0.25
HTc-Mg0.33
HTc-Li
AlCaMO
MgO
CaO
HTc-Zn
HTc-Fe
ZnO
199
265
275
240
50
640
12
37
0.604
0.625
0.558
0.657
0.119
0.664
0.086
0.163
1.068
0.090
0.337
0.20
0.25
0.33
0.33
0.25
–
The catalytic transesterification of a cyclic carbonate with an
alcohol involves two equilibrium steps, which typically generate
a hydroxyl alkyl carbonate (compound A, Scheme 1) as reaction
intermediate. Addition of a second molecule of alcohol (in this case
the secondary hydroxyl group of the glycerol) of the intermediate
product (A) results in the formation of the glycerol carbonate (4-
hydroxymethyl-1,3-dioxolan-2-one) and ethyleneglycol. These
two equilibrium reaction steps are presented in Scheme 1.
–
0.30
0.33
–
306
5
37
HTc-Zn–Li
0.25

M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
143
O
100
80
60
40
20
0
O
Base
CH2 CH CH2
OH OH OH
catalyst
O
O
O
O
+
HOCH2-CH2
O
CH2CHCH2OH
OH
A
O
O
O
CH2 CH2
+
O
O
HOCH2-CH2
CH2CHCH2OH
OH
OH OH
OH
A
Scheme 1.
0
50
100
150
200
Time (min)
250
300
350
400
Recently we have shown that transesterification of fatty acid
methyl esters with glycerol [4] and with polyethyleneglycol can
be performed efficiently in the presence of solid basic catalysts
such as MgO and mixed oxides [21]. Magnesium oxide with high
surface area offers strong Lewis type-base sites (O2ꢀ) associated
with Mg2 + O2ꢀ ion pairs in were the basic strength will depend
on their coordination number. When aluminum is introduced into
de brucite layer of MgO, hydrotalcites are formed. The hydrotal-
cites are layered double hydroxides (LDHs) wꢀhnich have the general
molecular formula M2x þ M3y þ ðOHÞ2ðx þ yÞAy=nmH2O, where M2+
and M3+ are a divalent and a trivalent metal ions, respectively,
and Aꢀn is an intercalated anion. Structurally, they possess bru-
cite-like (Mg(OH)2) sheets where isomorphous substitution of
Mg2+ by a trivalent cation like Al3+ occurs. The resulting excess of
positive charge in the layered network is compensated by anions,
which occupy the interlayer space along with water molecules.
By controlled thermal decomposition, LDHs are converted into
mixed oxides (HTc) with high specific surface areas and strong Le-
wis basic sites.
Taking into account our previous results and considering that
the transcarbonation is a transesterification type reaction, we have
selected a high surface area MgO and an Al/Mg-mixed oxide de-
rived from hydrotalcite with an Al/(Al + Mg) molar ratio of 0.25
(HTc-Mg0.25) as basic catalyst.
The reactions were performed in the absence of solvent, under
inert atmosphere with an EC/glycerol molar ratio of 2, at 50 °C
and using 7 wt% of catalyst. The results obtained are summarized
in Table 2. In a preliminary experiment it has been observed that
the transesterification is a slow reaction in the absence of catalyst,
giving 5% yield of GC after 5 h of reaction time.
Fig. 2. Kinetics curves of transesterification between ethylene carbonate and
glycerol in the presence of HTc-Mg (h), MgO (s) and rehydrated hydrotalcite (d).
the HTc-Mg (Fig. 2). However the HTc-Mg catalyst is more selective
than MgO which also forms glycidol as a reaction product.
It has been presented that the catalytic activity of calcined
hydrotalcites can be increased by CO2-free hydration at room tem-
perature [23]. This treatment results in the restoration of the origi-
nal layered structure, with OHꢀ as the compensating anions in the
interlayer, which can act as Brönsted basic sites. It has been re-
ported that the amount of water added on the calcined hydrotal-
cite has a strong influence on the catalytic activity because an
excess of water on the catalyst surface can poison the basic sites
decreasing catalytic activity [24]. In the reaction studied here, an
optimum in catalytic activity was found with samples with
36 wt% water content. The result obtained for the transesterifica-
tion reaction using HTr is summarized in Table 2 and Fig. 2. As
can be seen in Fig. 2 initial rates of formation of GC are similar
for HTr and HTc-Mg catalysts, indicating that in this case Brönsted
and Lewis basic sites possess similar activity. However, the final
conversion of glycerol and selectivity to GC on the HTr hydrotalcite
is lower than in the case of HTc-Mg. As can be observed in Fig. 2,
HTr exhibits a higher deactivation rate than HTc-Mg, which can
be related to a stronger adsorption of the highly polar reactants
and products involved in the reaction that may block the active
sites. On the other hand, glycidol is formed in higher amounts in
the hydrated than in the non-hydrated sample lowering, therefore,
the final selectivity.
The HTc-Mg and MgO are the most active catalysts for the
transesterification reaction. These results are consistent with the
observation that these materials have basic sites strong enough
for proton abstraction in cases where the pKa of the hydrogen is
close to 16, as it is the case of glycerol with a pKa of 14 (measured
in water at 25 °C) [22].
According with the results presented in Table 2, it appears that
mixed oxides can be suitable catalysts in terms of activity and
selectivity for the transesterification of EC and glycerol to produce
GC. Therefore it is of interest to improve the activity of the mixed
oxide to operate at lower temperature and reaction times since
cyclic carbonates are heat-sensitive, and prolonged reaction times
can promote decomposition of cyclic carbonates (EC and GC).
As can be seen, MgO with a high concentration of basic sites
performs the transesterification at higher initial reaction rate than
Table 2
Results of transesterification between EG and glycerol on different heterogeneous base catalysts.
Catalyst
Conversion glycerol (%)
Yield GC (%)
Yield glycidol (%)
Intermediate A (%)
Selectivity GC (%)
–
20
85
86
78
5
–
<1
3
15
2
5
43
96
90
87
HTc-Mg0.25
MgO
HTr
82
78
68
7
3
Reaction conditions: 16 mmol EC, 8 mmol glycerol, 7 wt% of catalyst respect to the total weight of reactants, at 50 °C, 5 h reaction time.

144
M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
3.1. Influence of chemical composition
the presence of stronger Lewis basic sites results to be very bene-
ficial and the catalysts perform the transesterification reaction
more efficiently than with HTc-Mg, the selectivity in all cases being
very high. Interestingly glycidol was not detected in the reaction
media, the selectivity to GC being above 96%.
For comparative purposes, the reaction was also performed
with CaO as basic catalyst, and as occurred with MgO, the activity
for converting glycerol was very high, but the selectivity to GC was
lower due to the formation of glycidol.
From the results presented we can conclude that a correlation
does exist between the catalytic activity and the basic properties
of the catalyst for transesterification reaction. This correlation be-
tween basic properties of solid catalysts and catalytic activity has
been observed by Figueras et al. [26] in the transcarbonation of
1-phenyletanol with diethylcarbonate in presence of CsF, KF sup-
ported on alumina and Mg/La mixed oxide. In our case the pres-
ence of stronger Lewis basic sites in the HTc-Li and AlCaMO
catalysts allows to carry out the reaction under very mild reaction
conditions and short reaction times, with excellent selectivities
and yields.
It is known that by increasing the aluminum content in an
Al2O3–MgO mixed oxide (HTc-Mg) the basic site density decreases,
while the fraction of the strong basic sites increases [25]. Therefore
and in order to improve the activity of mixed oxides for the transe-
sterification reaction three samples with Al/Al + Mg molar ratios of
0.20, 0.25, and 0.33 have been prepared. The results obtained for
the transesterification reaction with the above samples are given
in Fig. 3. As can be observed, HTc-Mg with Al/(Al + Mg) molar ratios
of 0.20 and 0.25 shown similar activity, whereas the sample with
Al/(Al + Mg) molar ratio of 0.33 shows lower activity. These results
indicate that in the samples prepared, the introduction of stronger
basic sites does not compensate the decrease in the number of ac-
tive sites produced by increasing the Al3+ content.
A further increase in basicity can be achieved by replacing Mg2+
by more electropositive ions such as Li+ and Ca2+ [4]. By doing this,
the density of negative charge on the oxygen increases, leading to
an increase of the basicity of the resulting mixed oxide. To check
the effect of this modification on catalyst activity, an HTc-Li sample
with an Al/Al + Li = 0.33 was synthesized. However, when we try to
synthesize an HTc-Ca sample with an Al/Al + Ca = 0.25 as described
in the experimental section, according to XRD analysis, a mixture
of CaCO3 + Al(OH)3 was obtained. This material was calcined (AlCa-
MO) and tested as basic catalyst in the transesterification reaction.
The transesterification of ethylene carbonate with glycerol was
tested using these samples as catalyst under the same reaction
conditions than above, and their activity resulted much higher
when compared with the Al/Mg-mixed oxide, in such way that
temperature and amount of catalyst were strongly lowered being
possible to operate at 35 °C and 0.5 wt%, respectively. The results
obtained are presented in Table 3, where it can be observed that
3.2. Catalysis deactivation
To study catalyst deactivation HTc-Mg, HTc-Li and AlCaMO
were reused several times under the same reaction conditions.
Thus, after completion of the transcarbonatation reaction the solid
was filtered, washed with methanol (5 ꢁ 2.5 mL) and acetone to
remove the products adhering to the surface of the catalysts, dried
at room temperature and reused. As can be seen in the Table 4 the
conversion slightly decreased while selectivity was maintained
after three catalytic cycles when using HTc-Mg and HTc-Li. On
the other hand, when using AlCaMO some deactivation is observed
after 1 h reaction time.
From these results obtained we can conclude that with HTc-Li
and AlCaMO catalysts, excellent yield and selectivity to GC can
be obtained under very mild reaction conditions. Moreover the cat-
alysts can be recycled, and activity is basically maintained.
100
80
60
40
20
0
Table 4
Results of the reuse of the mixed oxides in the transesterification of EC and glycerol.
Catalysts
Conversion (selectivity to GC)
t (h)
1st cycle
2nd cycle
3rd cycle
HTc-Mg0.25
HTc-Li
AlCaMO
1
1
1
4
57(98)
88(96)
89(98)
92(97)
57(98)
88(97)
74(98)
90(96)
47(97)
86(96)
72(98)
90(96)
0
50
100
150
Time (min)
200
250
300
350
Fig. 3. Influence of chemical composition of Al/Mg mixed oxides in the transeste-
rification between glycerol and ethylene carbonate using HTc-Mg with Al/Al + Mg
ratio of 0.20 (e), 0.25 (h), and 0.33 (d).
Reaction conditions: 24 mmol ethylene carbonate, 12 mmol de glycerol, 0.5 wt% of
catalyst respect to the total weight of reactants at 35 °C.
Table 3
Catalytic performance of HTc-Li, AlCaMO, CaO and HTc-Fe for the synthesis of GC.
Catalyst
HTc-Li
AlCaMO
CaO
t (min)
Glycerol
conversion (%)
Yield GC (%)
Yield glycidol (%)
Yield of intermediate
A (%)
Selectivity GC (%)
15
60
83
88
81
85
0
0
2
3
97
96
15
60
87
89
85
87
0
0
3
2
98
98
15
60
90
92
81
83
5
6
4
3
90
90
HTc-Fe
15
60
80
84
65
70
10
10
5
4
81
83
Reaction conditions: 24 mmol ethylene carbonate, 12 mmol glycerol, 0.5 wt% of catalyst with respect to the total weight of reactants at 35 °C.

M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
145
3.3. Preparation of glyceryl carbonate from urea and glycerol
25
20
15
10
5
25
20
15
10
5
HTc-Mg
Sn-Beta
Another interesting route of producing glyceryl carbonate is the
reaction between glycerol and urea. The main advantage of this
method is that urea is the readily available and cheap reactant.
In addition, ammonia formed can be easily converted to urea since
the urea synthesis is performed from ammonia and carbon dioxide.
The accepted mechanism for the reaction of urea with alcohols
involves two steps [27] and that in the specific case of glycerol it
will be as follows: first the carbamoylation of glycerol to glycerol
carbamate liberating a mol of ammonia. This step occurs at higher
reaction rate than the second step, the carbonylation of the glyc-
erol urethane to glycerol carbonate with elimination of a second
mol of ammonia (Scheme 2).
As described above, the reaction between urea and glycerol has
been performed using ZnSO4 and ZnO catalysts. High yields are re-
ported using ZnSO4 as catalyst [15,17,28], however, this salt is sol-
uble in glycerol and the reaction takes place under homogeneous
catalysis with the catalysts being partially recovered after the
reaction.
Ball et al. [27] reported that the reaction between primary and
secondary alcohols with urea to form alkyl carbonates can be im-
proved using an adequate combination of a weak Lewis acid and
a Lewis base. More recently Li et al. [29] reported that catalysts
exhibiting adequate acid and base properties were favorable to
the synthesis of cyclic carbonates from urea and diols. The authors
claim that the presence of Lewis acid sites able to activate the urea
is an important feature of the catalyst to achieve high yields of the
cyclic carbonate, and ZnO appeared to be the optimum catalyst. It
should be remarked however that with ZnO high ratios of urea
glycerol (up to 4 mol molꢀ1) are normally used.
MgO
HTc-Zn
HTc-Zn
ZnO
ZnO
Catalysts
0
0
Fig. 4. Results of the Knoevenagel condensation between benzaldehyde (10 mmol)
and malononitrile (10 mmol) 5 wt% of catalyst at room temperature (d), and
acetalization of benzaldehyde (5 mmol) with TMOF (25 mmol) 5 wt% of catalyst in
CCl4 (8 mL) at 76 °C (h).
OH
HO
NH2
H2N
O
O
C
O
H
M
O
M
Scheme 3.
Taking into account all the above-mentioned features, a mech-
anism can be proposed that involves Lewis acid site able to activate
the carbonyl group, that will be now more prone to the nucleo-
philic attack of the glycerol adsorbed on the Lewis basic site
(Scheme 3).
In order to find solid and recyclable heterogeneous catalysts
able to perform efficiently the reaction between the urea and glyc-
erol we have prepared a series of solid materials with different
acid–base properties. The reactions were performed under reduced
pressure in order to shift the thermodynamic equilibrium by
removing the ammonia formed. Reaction conditions (temperature,
amount of catalyst, pressure and molar ratio and reaction time)
were optimized using ZnO that was taken as reference. We have
selected as optimized parameters: 5 wt% of catalyst with respects
to glycerol, 145 oC reaction temperature, 30 Torr of pressure, glyc-
erol to urea molar ratio of one and 5 h reaction time. The results
obtained are summarized in Table 5.
3.4. Catalyst characterisation: acid–base properties
The relative acidity and basicity of various representative solid
catalysts of this series, was determined by two test reactions: the
Knoevenagel condensation between malononitrile and benzalde-
hyde which is typically catalyzed by bases, and the acetalization
of benzaldehyde with trimethylorthoformate (TMOF), which is a
reaction catalyzed by purely acid sites. Results of initial rates for
Knoevenagel condensation and acetalization for the catalysts with
different acid–base pairs are presented in Fig. 4. As can be observed
the order of activity for the Knoevenagel condensation which can
be associated with the order of catalysts basicity is: HTc-
Mg0.25 > MgO > HTc-Zn > ZnO. On the other hand, the order of
activity for the acetalization reaction which can be associated with
the order of catalysts acidity is: Sn-Beta > HTc-Zn ꢂ ZnO, while
HTc-Mg and MgO gave negligible activity.
O
C
O
O
OH
+ NH3
-NH3
HO
OH
+
O
HO
O
NH2
NH2-CO-NH2
OH
OH
Glycerol urethane
Glycerol Carbonate
NH2-CO-NH2
O
C
O
OH
O
O
O
NH2
O
O
N
H
(2-oxo-1,3-dioxolan-4-yl)methyl
5-(hydroxymethyl)oxazolidin-2-one
carbamate
Scheme 2.

146
M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
Table 5
Results of carbonylation of glycerol and urea in the presence of different solid
catalysts.
2145
0.02
MgO
2155
Entry
Catalyst
Glycerol
conversion (%)
GC yield
(%)
GC selectivity
(%)
0
1
2
3
4
5
6
7
8
Without
Beta-Sn
MgO
25
70
73
93
80
80
82
84
78
15
26
39
5
10
60
72
27
39
60
37
53
5
12
75
88
32
50
HTc-Mg0.25
AlCaMO
ZnOa
HTc-Znb
HTc-Zn/Li
HTc-Fe/Mg
2166
Reaction conditions: Glycerol/urea molar ratio = 1, 30 Torr, 145 °C, 5 wt% catalyst at
5 h of reaction time.
a
2% glycerol carbamate, 4% (2-oxo-1,3-dioxolan-4-yl)methyl carbamate, and 3%
of 5-(hydroxymethyl)oxazolidin-2-one and 3% unidentified products and traces of
glycidol were detected.
2300
2200
2100
2000
Wavenumber (cm-1)
b
2% glycerol carbamate, 3% (2-oxo-1,3-dioxolan-4-yl)methyl carbamate, 1% of 5-
(hydroxymethyl)oxazolidin-2-one and 3% of unidentified products, and traces of
glycidol were detected.
2152
HTc-Zn
Catalysts acidity and basicity has been also evaluated by FTIR
spectroscopy of adsorbed probe molecules: CO for surface acidity
and chloroform for surface basicity. Fig. 5 shows the IR spectra of
CO adsorbed at ꢀ170 °C at increasing equilibrium pressure on
MgO, HTc-Zn and Sn-Beta catalysts. CO interaction with coordin-
atively unsaturated surface sites causes a shift of the C–O stretch-
ing vibration to higher frequencies from that of the free molecule
(2143 cmꢀ1). This shift is associated to the Lewis acidity of the sur-
face site. Indeed a correlation between the electric field strength at
the surface site and the CO frequency shift has been proposed [30].
The IR spectra of CO adsorbed on MgO shows one band at
2155 cmꢀ1 together with a shoulder at 2166 cmꢀ1 associated to
two different types of Mg2+ surface sites [31] and an IR band at
2145 cmꢀ1 associated to physisorbed CO. On Sn-Beta sample CO
adsorption leads to IR bands at 2189 (weak, which becomes re-
solved at low CO coverages, see Fig. 5), 2177 and 2156 cmꢀ1 asso-
ciated to different co-ordinated Sn4+ surface sites [32], an IR band
at 2141 cmꢀ1 associated to OH groups according to the simulta-
neous shift of the hydroxyl groups, and an IR band at 2133 cmꢀ1
due to physisorbed CO. On HTc-Zn IR bands at 2204, 2187 and
2175 cmꢀ1 associated to CO interacting with Al3+ and Zn2+ Lewis
surface sites [33] are observed and an IR band at 2152 cmꢀ1 due
to OH groups and an IR band 2132 cmꢀ1 due to physisorbed CO.
Assignation of the IR band at 2152 cmꢀ1 to OH groups is supported
by the simultaneous shift of the hydroxyl groups. However the
contribution of some weak Lewis acid sites to the 2152 cmꢀ1 IR
band cannot be completely ruled out. Thus, from the IR spectra
of CO adsorption and the shift in the CO-stretching vibration of
the adsorbed CO molecules, a higher Lewis acidity of surface sites
on HTc-Zn and Sn-Beta samples than on MgO sample can be in-
ferred. On the other hand, while both Sn-Beta and HTc-Zn samples
show similar shift in the CO-stretching vibration of adsorbed CO
molecules (IR bands at 2189, 2173, and 2154 cmꢀ1), the higher
thermal stability of the IR bands in the Sn-Beta sample point to a
slight higher acid strength on this sample compared to the HTc-
Zn sample (Fig. 6). Moreover, Sn-Beta sample shows a higher
amount of Lewis acid surface sites than HTc-Zn sample (1.568 Le-
wis sites, a.u./mgr, in Sn Beta and 0.047 Lewis sites, a.u./mg, in
HTc-Zn). According to this results, the higher acidity observed on
Sn-Beta sample versus HTc-Zn in the acetalization of benzaldehyde
used as test reaction for order of acidity cannot only be related to
the slight difference in acid strength between both samples but
mainly to the higher amount of Lewis acid sites in the Sn-Beta
sample.
0.02
2132
2187
2204
2250
2200
2150
2100
2050
Wavenumber (cm-1)
2141
0.1
Sn-Beta
2133
2156
2177
2189
2200
2100
2000
Wavenumber (cm-1)
Fig. 5. FTIR spectra of CO adsorption at ꢀ170 °C on MgO, HTc-Zn and Sn-Beta
samples at increasing CO equilibrium pressure.
Fig. 7 shows the IR spectra of chloroform adsorbed at 25 °C at
increasing partial pressure on MgO, HTc-Zn and Sn-Beta samples.
Chloroform has been used as probe molecule for studding the ba-
sicity of the samples. The high basicity of MgO lead to a partial
decomposition of chloroform giving formiat species (IR bands at
2846 and 1600 cmꢀ1). On the other hand, the IR bands at 3020

M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
147
MgO
1217
0.02
0.001
2153
HTc-Zn
2982
2954
2173
2189
3020
2202
2846
a
b
c
d
1243
1249
2300
2250
2200
2150
2100
3200
3200
3200
3000
2800
1250
1200
Wavenumber (cm-1)
-1
Wavenumber (cm )
HTc-Zn
2157
0.001
1217
0.02
Sn-Beta
2175
2189
1226
3017
a
b
c
2982
d
2300
2250
2200
2150
2100
3000
2800
1250
1200
Wavenumber (cm-1)
Wavenumber (cm-1)
Fig. 6. FTIR spectra of CO adsorbed on MgO and HTc-Zn samples at increasing
temperatures (a) ꢀ170 °C, (b) ꢀ160 °C, (c) ꢀ153 °C, (d) ꢀ132 °C under dynamic
vacuum conditions.
0.02
3019
Sn-Beta
and 1217 cmꢀ1 associated to the
the (HCCl) deformation mode, respectively, are close to those ob-
m(C–H) stretching vibration and
m
served on liquid CHCl3 and associated to chloroform interacting
with the OH groups of the MgO surface (Cl–HO interaction)
according to the simultaneous shift of the hydroxyl groups. In
addition, the IR bands at 2982, 2954 cmꢀ1 and 1243, 1249 cmꢀ1
can be associated to chloroform interacting with Lewis basic sites
of the MgO surface. Different modes of interaction of CHCl3 in
which acid–base pairs or only base sites are involved could explain
the presence of the two IR bands [34]. On the HTc-Zn sample form-
iat species are not observed in agreement to a lower basicity. On
the other hand, IR bands at 3017 and 1217 cmꢀ1 corresponding
to CHCl3 interacting with OH groups are observed together with
another bands at 2982 and 1226 cmꢀ1 associated to chloroform
interacting with surface base sites. The lower shift of these bands
compared to those observed on MgO points to a lower basicity of
the sample. On the other hand on Sn-Beta only bands due to CHCl3
interacting with OH groups are observed (IR band at 3019 cmꢀ1).
3150
3100
3050
3000
2950
2900
Wavenumber (cm-1)
Fig. 7. FTIR spectra of CHCl3 adsorbed at 25 °C on MgO, HTc-Zn and Sn-Beta
samples at increasing equilibrium pressure.
The
m
(HCCl) deformation vibration is hidden due to the strong
3.5. Catalyst performance in the carbamoylation between urea and
glycerol
framework adsorption. Thus form the IR spectra of CHCl3 adsorp-
tion we can conclude that the basicity of MgO is higher than
HTc-Zn and both of them higher than Sn-Beta sample, which agree
with the order of activity observed in the Knoevenagel condensa-
tion reaction.
When the carbamoylation reaction between urea and glycerol
was performed in the presence of a material with strong Lewis acid
sites associated to a very weak conjugated base pair, such as that in

148
M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
the case of Sn-Beta zeolite, low glycerol conversion and low selec-
tivity to GC were observed (entry 1 in Table 5). Increasing the base
strength using MgO or MgO–Al2O3-mixed oxide derived from
hydrotalcites (entries 2 and 3 in Table 5), high conversions of glyc-
erol with very low selectivities to GC are obtained. In both cases,
polymerization of glycerol carbonate and/or glycerol into heavy
polymers is highly favored with the main responsible for the low
selectivity to glycerol carbonate observed.
the contrary we increase the basicity of the hydrotalcite by intro-
ducing Li, while maintaining the ZnO, the yield strongly decreases.
3.6. Deactivation of the Al/Zn mixed oxide
To evaluate catalyst deactivation and recycling, the Al/Zn mixed
oxide was rinsed after the first use (entry 1, Table 6), with metha-
nol, air-dried, and used again under exactly the same conditions
(entry 2). A 9% decrease in selectivity occurred. This process was
repeated with the same material and conditions in entry 3, which
gave similar selectivity than in the second cycle. The catalyst was
then extracted with methanol in a soxhlet setup, and used again
(entry 4) giving similar yield. Finally, the catalyst was calcined
after four reuses, and the catalytic results with the calcined sample
shows a slightly lower yield (entry 5) to the previous reactions. In
other words catalyst deactivation results show that there is a de-
crease in yield after the first reuse, but after this its activity re-
mains constant in further reuses.
Using a ZnO catalyst (entry 5) which has a lower basic strength
and higher acidity of the conjugated acid than MgO or Al/Mg-
mixed oxides, activity and selectivity were improved, in agreement
with the previous reports [35,29]. However, we have found that
when the strength of the Lewis acid sites (Zn) was maintained
and the basic site strength of the conjugated base was increased
by preparing an Al/Zn-mixed oxide (entry 6), activity and selectiv-
ity were the highest. Thus, the most active and selective catalyst
HTc(Al/Zn) gives 82% conversion of glycerol with 88% selectivity
to GC. In addition low amounts of the intermediate glycerol carba-
mate (2,3-dihydroxypropyl carbamate), 5-(hydroxymethyl)oxaz-
olidin-2-one, and (2-oxo-1,3-dioxolan-4-yl)methyl carbamate
(Scheme 2), besides negligible amount of unidentified by-products
were also detected working with an urea to glycerol ratio of
3.7. Surface characterisation of fresh and re-used catalysts
1 mol molꢀ1
.
Surface characterisation of three selected samples has been
done by XPS and shown in Table 7. The samples have been ana-
lysed after ‘‘in situ” activation (450 °C 3 h in nitrogen flow) and
after reaction. XPS data of ‘‘used” samples shows an increase in
the amount of carbon (C1s) due to adsorbed surface species and
some structural changes in the composition and chemical nature
of surface sites. In fact, a lower BE of Zn2p and Al2p XPS peaks
and the presence of two components related to aluminium species
in different environment are observed in the HTc-Zn sample after
reaction. Moreover an apparent increase in the Al/Zn ratio is ob-
served after reaction. However the lower ineslatic mean free path
of Zn2p electrons (0.49) versus Al2p electrons (1.124) and the pres-
ence of several layers of adsorbed surface species could contribute
to the increase in the Al/Zn ratio. On the other hand, in the Sn-Beta
sample a lower BE of Sn3d XPS peak is observed together with a
decrease in the Sn/Si ratio. The ineslatic mean free paths of Sn3d
electrons (0.87) and Si2p electrons (1.01) are quite similar, thus
the decrease in the Sn/Si ratio should be due to some structural fac-
tors. In the same way in MgO sample a lower BE of Mg2s XPS peak
is observed in the sample after reaction. The decrease in the BE of
Zn2p, Al2p, Sn3d, and Mg2s on the samples after reaction can be
related to a change in the electronic charge of the surface site,
either due to the presence of adsorbed species or due to some
structural modification of the samples after catalytic reaction,
which could explain the observed deactivation of re-used samples.
The excellent performance of this catalyst can be attributed to
the existence of well balanced acid–basic properties. Thus whereas
strong Lewis acid or basic sites promote mainly polymerization
reactions, a cooperative effect between weak metal Lewis acid sites
able to activate the carbonyl group of the urea, and Lewis basic
sites able to activate the glycerol, could be the responsible for
the high catalytic performance observed with Al/Zn oxides. If on
Table 6
Study of the reusability of HTc-Zn catalyst in the preparation of GC from urea and
glycerol.
Entry Cycles Glycerol conversion (%) GC yield (%) GC selectivity (%)
1
2
3
4
5
1
2
82
80
81
79
76
72
65
64
65
60
88
81
80
82
79
3
4a
5b
All experiments were carried out using 5 wt% of catalyst with respect to glycerol, at
30 Torr and 145 °C, for 5 h.
a
The catalyst was extracted with MeOH in a soxhlet.
The catalyst was calcined at 450 °C during 6 h in an air flow, and then with a N2
b
flow during 6 h.
Table 7
BE and surface composition determined by XPS on fresh and used samples.
Sample
BE (eV)
C1s
Surface composition
O1s
Al2p
79.6
Zn2p
Sn3d
–
Mg2
–
Si2p
O:C:(Al:Zn:Sn:Mg:Si)
59.4:3:12.1:25.5:–:–:–
HTc-Zna
Fresh
291.4
295.8
291.4
294.7
285.2
292.6
286.9
290.3
290.1
293.4
289.1
292.7
536.6
533.2
536.4
533.4
533.3
529.7
533.3
530.7
534.3
1027.5
HTc-Zna
Used
79.06
77.76
–
1026.45
–
–
57.1:28.4:8.7:5.7:–:–:–
64.0:1.36:–:–:0.37:–:34.2
60.9:12.27:–:–:0.2:–:26.6
50.4:3.8:–:–:–:45.7:–
Sn-Betab
Fresh
–
–
–
–
488.2
487.5
104
104
Sn-Betab
Used
–
–
–
MgOc
–
–
–
–
92.8
91.6
Fresh
MgOc
534.6
49.8:35.9:–:–:–:14.5:–
Used
a
b
c
Charge correction to C1s = 291.364 eV.
Charge correction to Si2p = 1044 eV.
No charge correction have been applied.

M.J. Climent et al. / Journal of Catalysis 269 (2010) 140–149
149
4. Conclusions
References
[1] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411.
[2] L. Prati, P. Spontoni, A. Gaiassi, Top. Catal. 52 (2009) 288.
[3] A.M. Ruppert, J.D. Meeldijk, B.W.M. Kuipers, B.H. Erné, B.M. Weckhuysen,
Chem. Eur. J. 14 (2008) 2016.
In order to avoid the formation of by-products, the transesteri-
fication reaction between EG and glycerol to give glycerol carbon-
ate has to be carried out at temperatures as low as possible.
Calcined hydrotalcites with pKa > 14 [36] which are able to abstract
a proton from glycerol are suitable materials to perform the transe-
sterification giving high yields of GC. On the other hand, the regen-
erated hydrotalcite with Bronsted basicity gives lower selectivity
owing to a higher production of glycidol.
Increasing the Lewis basicity of the mixed Mg–Al oxides de-
rived from hydrotalcites by substituting the Mg by Li allows to
work at much lower temperatures and catalyst loading than with
the Mg–Al mixed oxide (35 instead of 50 °C), and (0.5 instead of
7 wt%). In addition the mixed oxide obtained by calcination of a
mixture of CaCO3 + Al(OH)3 resulted the most effective catalyst
for the transesterification reaction giving selectivities of 98% to
GC.
Finally, the carbonylation of glycerol with urea is better carried
out with solid catalysts with well balanced acid–base, in where Le-
wis acid activates the carbonyl of the urea and the conjugated basic
sites activate the hydroxyl group of the glycerol. While ZnO has
been reported as a catalyst with an adequate acid–base pair, it
has been observed here that if the basicity of the conjugated base
of ZnO is increased by preparing a Zn–Al mixed oxide with a higher
conjugated base pair strength than ZnO, the reaction selectivity
and final yield to GC significantly increase working with an urea
to glycerol ratio of 1 mol molꢀ1. This allows to substitute the solu-
ble and difficult to recycle ZnO reported by the heterogeneous Zn–
Al hydrotalcite catalyst reported here.
[4] A. Corma, S.B. Abd Hamid, S. Iborra, A. Velty, J. Catal. 234 (2005) 340.
[5] C.H. Zhou, J.N. Beltramini, Y.-X. Fan, G.Q. Lu, Chem. Soc. Rev. 37 (2008) 527.
[6] A. Murase, JP 6222709, 1987.
´
[7] G. DAlelio, T. Huemmer, EP 0328150, 1967.
[8] (a) H.R. Gillis, D. Stanssens, US 5703136, 1997.;
(b) J.M. JR Whelan, R.R. Cotter, US 3072613, 1963.
[9] J.W. Yoo, Z. Mouloungui, A. Gaset, FR 9800451, 1998.
[10] R.A. Grey, US5091543, 1992.
[11] J.B. Bell, C.V. Arthur, US2915529, 1959.
[12] A. Sugita, Y. Sone, M. Kaeryama, JP06329663, 1994.
[13] Z. Mouloungui, J.-W. Yoo, C. L. Gachen, A. Gaset, EP0739888, 1996.
[14] C. Vieville, J.W. Yoo, S. Pelet, Z. Mouloungui, Catal. Lett. 56 (1998) 245.
[15] S. Claude, Z. Mouloungui, J.-W. Yoo, A. Gaset, US6025504, 2000.
[16] M. Okutsu, T. Kitsuki, E P 11566042, 2001.
[17] (a) T. Sasa, M. Okutsu, JP 2009067689, 2009.;
(b) T. Sasa, M. Okutsu, JP 2008285457, 2008.
[18] M. Okutsu, JP 2007039347, 2007.
[19] J.H. Teles, N. Rieber, W. Harder, US5359094, 1994.
[20] (a) A. Corma, L.T. Nemeth, M. Renz, S. Valencia, Nature 412 (2001) 423;
(b) T. Takewaki, L.W. Beck, M.E. Davis, J. Phys. Chem. 103 (1999) 2674.
[21] M.J. Climent, A. Corma, S.B. Abd Hamid, S. Iborra, M. Mifsud, Green Chem. 8
(2006) 524.
[22] E.M. Woolley, J. Tomkins, L.G. Hepler, J. Sol. Chem. 1 (1972) 341.
[23] (a) C. Moreau, R. Durand, A. Roux, D. Tichit, Appl. Catal. A 193 (2000) 257;
(b) D. Tichit, B. Coq, Cattech 7 (2003) 2006;
M.J. Climent, A. Corma, S. Iborra, K. Epping, A. Velty, J. Catal. 225 (2) (2004)
316.
[24] (a) J. Roelofs, A.J. van Dillen, K.P. de Jong, Catal. Today 60 (2000) 297;
(b) M.J. Climent, A. Corma, S. Iborra, A. Velty, Catal. Lett. 79 (2002) 157.
[25] F. Cavani, F.A. Trifiro, A. Vaccari, Catal. Today 11 (1991) 173.
[26] J.M. Veldurthy, F. Clacens, F. Figueres, J. Catal. 229 (2005) 237.
[27] P. Ball, H. Füllmann, W. Heitz, Angew. Chem., Int. Ed. Engl. 19 (1980) 718.
[28] J.W. Yoo, Z. Mouloungui, Stud. Surf. Sci. Catal. 146 (2003) 757.
[29] Q. Li, W. Zhang, N. Zhao, W. Wei, Y. Sun, Catal. Today 115 (2006) 111.
[30] M.I. Zaki, H. Knoezinger, Spectrochim. Acta A 43 (1987) 1455.
[31] M. Sterrer, T. Risse, H.J. Freund, Appl. Catal. A: Gen. 307 (2006) 58.
[32] D.A. Popescu, F.B. Verduraz, Catal. Today 70 (2001) 139.
[33] D. Scarano, G. Spoto, S. Bordiga, A. Zecchina, C. Lamberti, Surf. Sci. 276 (1992)
281.
Acknowledgments
The financial support from CENIT Program (2006) funded by
CDTI through the project entitled ‘‘Proyecto de innovación para el
impulso del biodiesel en España” leaded by Repsol-YPF is gratefully
acknowledged. The contribution of Mr. Pablo Ramos to the exper-
imental work is also gratefully acknowledged.
[34] E. Bosch, S. Huber, J. Weitkamp, H. Knözinger, Phy. Chem. Chem. Phys. 1 (1999)
579.
[35] B.M. Bhanage, S.-I. Fujita, Y. Ikushima, M. Arai, Green Chem. 5 (2003) 429.
[36] (a) W.T. Reichle, J. Catal. 94 (1985) 547;
(b) M.J. Climent, A. Corma, S. Iborra, J. Primo, J. Catal. 15 (1995) 60.