A study on the carboxylation of glycerol to glycerol carbonate with carbon dioxide: The role of the catalyst, solvent and reaction conditions

Article abstract of DOI:10.1016/j.molcata.2006.05.021

Glycerol was reacted with CO₂ (5 MPa) at 450 K in presence of Sn-catalysts (n-Bu₂Sn(OMe)₂ 1, n-Bu₂SnO 2 or Sn(OMe)₂ 3), using either glycerol or tetraethylene glycol dimethyl ether (tedmg) as reaction medium. 1 was much more active than 2. 1 was demonstrated to convert into n-Bu₂Sn(glycerol-2H) 4 upon reaction with glycerol and elimination of MeOH. Monomeric 4 is proposed to be the active species in catalysis. It converted into a polymeric material with time with consequent reduction of its catalytic activity. Also, after the first catalytic cycle 4 was converted into an oligomeric material that did not contain glycerol. This also caused the reduction of the catalytic activity. 3 was able to uptake CO₂ but was not able to promote the carboxylation of glycerol. 1 and 2 also promoted the trans-esterification of dimethylcarbonate (DMC) with glycerol to afford glycerol carbonate, but at a lower rate than the direct carboxylation of glycerol. This fact seems to rule out that the carboxylation of glycerol may proceed through the preliminary formation of DMC and its subsequent trans-esterification.

Full text of DOI:10.1016/j.molcata.2006.05.021

Journal of Molecular Catalysis A: Chemical 257 (2006) 149–153
A study on the carboxylation of glycerol to glycerol carbonate with carbon
dioxide: The role of the catalyst, solvent and reaction conditions
Michele Aresta, Angela Dibenedetto∗, Francesco Nocito, Carlo Pastore
Department of Chemistry, University of Bari, via Orabona 4, Bari 70126, Italy
Received 13 March 2006; accepted 11 May 2006
Available online 23 June 2006
Abstract
Glycerol was reacted with CO2 (5 MPa) at 450 K in presence of Sn-catalysts (n-Bu2Sn(OMe)2 1, n-Bu2SnO 2 or Sn(OMe)2 3), using either
glycerol or tetraethylene glycol dimethyl ether (tedmg) as reaction medium. 1 was much more active than 2. 1 was demonstrated to convert into
n-Bu2Sn(glycerol-2H) 4 upon reaction with glycerol and elimination of MeOH. Monomeric 4 is proposed to be the active species in catalysis. It
converted into a polymeric material with time with consequent reduction of its catalytic activity. Also, after the first catalytic cycle 4 was converted
into an oligomeric material that did not contain glycerol. This also caused the reduction of the catalytic activity. 3 was able to uptake CO2 but was
not able to promote the carboxylation of glycerol.
1 and 2 also promoted the trans-esterification of dimethylcarbonate (DMC) with glycerol to afford glycerol carbonate, but at a lower rate than
the direct carboxylation of glycerol. This fact seems to rule out that the carboxylation of glycerol may proceed through the preliminary formation
of DMC and its subsequent trans-esterification.
© 2006 Elsevier B.V. All rights reserved.
Keywords: Carboxylation of glycerol; Glycerol carbonate; Tin catalysts
1. Introduction
tion of ethene oxide or propene oxide [10] formed, on turn, by
epoxidation of the relevant olefin (Eqs. (2a) and (2b) and (3a)
and (3b)):
The utilisation of glycerol for the synthesis of added value
chemicals is a theme of great industrial interest because glycerol
is formed in large amounts during the production of biodiesel
from natural triglycerids [1], and represents a waste that must
be used or eliminated. Several conversion processes [2–7] are
described in the literature that transform glycerol into useful
materials (see Scheme 1). Among them the synthesis of glycerol
carbonate plays a prominent role as the monomer could be used
for the synthesis of new functionalised polymers that might have
interesting new applications [8].
So far, the only method for the synthesis of glycerol carbonate
described in the literature was the trans-esterification of ethylene
carbonate (EC) or propene carbonate (PC). Either homogeneous
or heterogeneous catalysts [9] were used, with a mass ratio of
1:1 with respect to glycerol. Such technology is not completely
appropriate also because EC and PC are prepared by carboxyla-
(2a)
(2b)
(3a)
(3b)
TheformationoftheepoxideEqs. (2a)–(3a)isthelimitingstepin
the overall process from the point of view of yield and selectivity,
while both the carboxylation of the alkene Eq. (2b)–(3b) and the
trans-esterification of the alkene carbonate (Eq. (4)) are run with
∗
Corresponding author. Tel.: +39 0805442084; fax: +39 0805442429.
E-mail address: A.dibenedetto@chimica.uniba.it (A. Dibenedetto).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2006.05.021

150
M. Aresta et al. / Journal of Molecular Catalysis A: Chemical 257 (2006) 149–153
Scheme 1. Processes of conversion of glycerol into useful chemicals.
good yield and selectivity:
tricheatingjacket(maximumtemperature675 K). Theautoclave
was equipped with an inner glass jacket to avoid metal contam-
ination of the reagents. Molecular sieves were placed into the
glass reactor out of the contact with the liquid phase and used to
remove water from the gas phase. Infrared spectra were obtained
with a Shimadzu IR Prestige 21 spectrometer, the sample being
placed between KBr windows either neat or dispersed in Nujol.
Reaction products were analyzed by GC using a capillary col-
umn ZB-Wax (30 m × 0.25 mm) on a Agilent 6850 series gas-
chromatograph equipped with a flame ionization detector (FID),
using diphenyl ether as internal standard. GC–MS analyses were
performed using a Shimadzu GC–MS-QP5050. Elemental anal-
ysisontinwasconductedcolorimetrically[15]usingaShimadzu
UV-1601 spectrophotometer.
(4)
Therefore, findinganewdirectroutetoglycerolcarbonatewould
be very welcome, as it would avoid multi-step processes with
energy saving and waste reduction.
The direct carboxylation of glycerol (Eq. (5)) is a very inter-
esting process that would convert
2.2. Experimental section
n-Bu2Sn(OCH3)2 1 and n-Bu2SnO 2 were prepared as
described in the literature [16].
(5)
two wastes into a valuable product.
2.2.1. Synthesis of Sn(OCH3)2
This process has been reported in the literature as “not-
occurring” in presence of catalysts such as zeolithes [9]. Fol-
lowing our previous experience in the direct carboxylation of
alcohols [11–13] we have investigated the direct carboxylation
of glycerol using transition metal alkoxides. In this paper we
report the results of our studies on the carboxylation of glycerol
in presence of tin-catalysts. The formation of glycerol carbonate
takes place, also if with some limitations at the moment.
A solution of Na(OCH3) prepared reacting 0.45 g of sodium
(19 mmol) in 10 mL of methanol was placed in a three-necked
round-bottom flask equipped with a dropping funnel, a con-
denser and a magnetic stirrer. A solution of SnCl2 (2 g, 10 mmol)
in methanol (10 mL) was added dropwise, the mixture was
stirred for 1 h at 373 K, then cooled to 300 K and filtered. The
solution was concentrated in vacuo to half of its volume and the
white solid formed (NaCl) was discarded. Addition of toluene
andcoolingfavouredtheformationofasolidthatwasfilteredand
analyzed (1.4 g; 74% yield with respect to Sn). The compound
contained traces of SnCl2 difficult to eliminate. Anal. Calcd. for
C10H24O2Sn (%): C, 40.82; H, 8.16; Sn, 40.14. Found (%): C,
39.71; H, 7.74; Sn, 41.02. IR (KBr disks, Nujol): bands at 2953
(s), 2910 (s), 2852 (s), 1456 (m), 1375 (m), 1039 (s, ν(C–O)),
2. Experimental
2.1. General
All reactions were carried out under dry dinitrogen using
the vacuum-line technique. Chemicals (RP grade) were pur-
chased from Aldrich. n-Bu2SnO was used as received. Solvents
were further purified [14], distilled and stored under dinitrogen
over anhydrous calcium sulphate. CO2 (99.999%) was pur-
chased from Rivoira. High-pressure reactions were conducted
in a 100 mL stainless steel autoclave thermostated with an elec-
568 (w, νa(Sn–C2), 484 (w), 451 (m, νs(Sn–O2)) cm−1
.
2.2.2. Carbonation reaction of Sn(OCH3)2
A 30 mg of solid Sn(OCH3)2 were exposed to CO2 during
10 min in a round bottom flask, avoiding the contact with air,
at room temperature. The resulting product was washed with

M. Aresta et al. / Journal of Molecular Catalysis A: Chemical 257 (2006) 149–153
151
THF and dried in vacuo. IR (KBr disks, Nujol): bands at 2953
(s), 2910 (s), 2852 (s), 1633 (s), 1456 (m), 1373 (m), 1192 (w),
1097 (m), 1030 (m), 922 (w), 823 (m), 597 (m), 568 (w), 451 (w)
cm−1. Sn(OCH3)(OCOOCH3) was stable at room temperature
in vacuo, in fact the intensity of the CO band at 1633 cm−1 did
not decrease after pumping in vacuo for 5 h at 373 K. The CO2
uptake by solid Sn(OCH3)2 was also followed using a gas burette
and shown to stop when the CO2/Sn molar ratio equal to one
was reached.
2.2.3. Synthesis of di(n-butyl)tin glycerate
Toasolutionofdi(n-butyl)tindimethoxide(0.5 g, 1.69 mmol)
in toluene (5 mL), 0.15 g of glycerol (1.69 mmol) were added.
The solution was stirred for 24 h at room temperature. The
exchange between glycerol and methanol was monitored by gas-
chromatography measuring the methanol released (75% after
1 h, 95% in 6 h, 98% in 12 h). The resulting mixture was fil-
tered and the white solid formed was washed with THF (5 mL,
two times) and analyzed. Anal. Calcd. for C11H24O3Sn (%):
C, 40.90; H, 7.49; Sn, 36.75. Found (%): C, 41.3; H, 7.40; Sn,
36.2. IR (KBr disks, Nujol): 3272 (s), 2953 (s), 2910 (s), 2852
Fig. 1. Trend of conversion of glycerol into glycerol carbonate with n-
Bu2Sn(OMe)2 as catalyst. [PCO2 = 5 MPa, 450 K].
(c) Use of di(n-butyl)tin oxide as catalyst: the autoclave was
charged with 0.65 g of di(n-butyl)tin oxide (2.61 mmol), 4 g
of glycerol (43.5 mmol) and 5 MPa of CO2. After stirring for
the given time at 453 K, the resulting product was worked
up as in (b). The acetone phase was analyzed by GC. This
reaction was carried out under various conditions of CO2
pressure, temperature and time of reaction. The results are
summarized in Table 2 (Section 3).
(s), 1118 (m), 643 (m) cm−1
.
(d) Use of tin dimethoxide as catalyst: a 0.80 g of tin dimethox-
ide(4.4 mmol)and4 gofglycerol(43.5 mmol)werecharged
in the autoclave with 5 MPa of CO2. Using the same exper-
imental conditions reported for (a)–(c), only traces of glyc-
erol carbonate were revealed.
(e) Reaction in tetraethylen glycol dimethoxide: a 0.60 g of
di(n-butyl)tindimethoxide (2.04 mmol) and 4 g of glycerol
(43.5 mmol) were dissolved in 5 mL of tetraethylen glycol
dimethoxide. The resulting solution was placed in the auto-
clave that was charged with 5 MPa of CO2. The reaction
was carried out for 6 h at 453 K. The results are reported in
Table 1.
(f) Reaction of di(n-butyl)tin(glycerol-2H) with methanol: to a
suspension of di(n-butyl)tin(glycerol-2H) (2 g, 6.21 mmol)
in toluene (5 mL), 251 ␮L of methanol (6.21 mmol) were
added, avoiding the contact with air. The solution was stirred
for 5 h at 323 K and analyzed by GC. The conversion of
the tin-compound into di(n-butyl)tindimethoxide was eval-
uated by quantitative determination of released glycerol. It
2.2.4. Carboxylation of glycerol
(a) Direct reaction of glycerol with CO2 without catalyst: a 2.5 g
of glycerol (27 mmol) were placed in the autoclave that was
charged with 5 MPa of CO2. After stirring for 15 h at 453 K,
the resulting solution was cooled to room temperature and
extracted with 5 mL of acetone (three times). The filtered
acetone was analyzed by GC: no glycerol carbonate was
formed.
(b) Use of di(n-butyl)tindimethoxide as catalyst: a 0.78 g of
di(n-butyl)tindimethoxide (2.61 mmol) and 4 g of glycerol
(43.5 mmol) were charged in the autoclave with 5 MPa of
CO2. This reaction was carried out using various experi-
mental conditions as reported in Table 1 (see Section 3), that
summarizes the results. The resulting solution was cooled
to room temperature and extracted with 5 mL of acetone
(three times). The acetone solutions were combined and the
carbonate recovered by acetone evaporation. The trend of
formation of the carbonate is shown in Fig. 1.
Table 1
Carboxylation of glycerol using di(n-butyl)tindimethoxide as catalyst
Experiment
% Catalyst
Time (h)
PCO2 (MPa)
Temperature (K)
Solvent
Molecular sieves
Glycerol conversion
% (isolated)a
1
2
3
4
5
6
7
8
2
2
2
2
2
6
6
6
15
15
15
15
15
6
5
5
2.5
5
5
5
5
5
453
453
453
373
453
453
453
453
None
None
None
None
tedmgb
None
None
None
No
0.42 (0.35)
2.29 (1.94)
1.74 (1.49)
0.49 (0.42)
2.22 (1.88)
3.30 (2.80)
5.87 (4.93)
6.86 (5.72)
Yes
Yes
Yes
Yes
Yes
Yes
Yes
10
15
In all experiments 4 g of glycerol (43.5 mmol) and 0.78 g of catalyst (2.61 mmol) were used.
a
The reaction yield was in all cases 15–20% higher than the isolated yield.
tedmg = tetraethylene glycol dimethyl ether.
b

152
M. Aresta et al. / Journal of Molecular Catalysis A: Chemical 257 (2006) 149–153
Table 2
Carboxylation of glycerol using di(n-butyl)tin oxide as catalyst
Experiment
% Catalyst
Time (h)
PCO2 (MPa)
Temperature (K)
Solvent
Molecular sieves
Glycerol conversion
% (isolated)a
1
2
3
4
5
6
2
2
2
2
2
6
6
6
6
6
6
6
5
5
2.5
5
5
453
453
453
373
453
453
None
None
None
None
tedmgb
None
No
0.36 (0.30)
1.26 (1.12)
0.94 (0.82)
0.50 (0.41)
1.08 (0.97)
2.30 (1.92)
Yes
Yes
Yes
Yes
Yes
5
In all experiments 4 g of glycerol (43.5 mmol) and 0.65 g of catalyst (2.61 mmol) were used.
a
The reaction yield was in all cases 15–20% higher than the isolated yield.
tedmg = tetraethylene glycol dimethyl ether.
b
resulted to be: 4% after 30 min, 9% after 1.5 h, 10% after
2.5 h, 10% after 5 h.
a mixture of complexes was formed, difficult to characterise.
In methanol, for example, the retro-conversion of 4 into 1 was
twelve times slower than the reaction of n-Bu2Sn(OCH3)2 with
glycerol, but not zero. Studies are still in progress in order to
characterise the glycerate-complex in the solid state and in solu-
tion. Oligomer 4 did not incorporate CO2 at room temperature
also under 5 MPa, differently from 1 that uptakes CO2 in toluene
under 0.1 MPa of CO2 [16]. The reaction of 4 with CO2 took
place only at high temperature under 5 MPa of CO2 but only
2.2.5. trans-Esterification of DMC with glycerol
A 1:1 (molar ratio) mixture of DMC and glycerol (22 mmol
for each reagent) was reacted in an autoclave in presence of 1
(22 mmol) under the same conditions of temperature and time
as for the carboxylation of glycerol. The reaction products were
identified by mass spectroscopy and the methanol exchanged
monitored by gas-chromatography. The conversion was shown
to be 65% after 15 h.
a small amount of carboxylated complex (νCO at 1681 cm−1
)
was formed, as demonstrated by an FTIR study under pressure.
These data suggest that once monomeric 4 is formed, it can
either incorporate CO2 or oligomerize with deactivation. The
fact that glycerol carbonate was formed at a rate comparable
with the exchange glycerol–methanol, suggested that the car-
boxylation of the monomer is faster than the oligomerization.
That the monomer is the active species is also demonstrated
by the fact that if the oligomer isolated from the exchange of
1 with glycerol was used as catalyst, the catalytic activity was
strongly reduced. Once formed, the carboxylated species was
able to eliminate glycerol carbonate affording a Sn-compound
that resulted to be the same obtained when 1 acted as catalyst in
the carboxylation of methanol. It was clearly shown that in the
latter reaction, after the first cycle 1 converted into a decamer
[17] or trimer [18], depending on the nature of the alkyl group
on the alkoxo moiety, with formation of Sn–O–Sn bonds: this
conversion slows down the activity of the tin-compounds.
These results demonstrate that if an oligomer is formed, both
the reactivity of the complex towards CO2 and the activity of
the catalyst are slowed down. Therefore, one can assume that
the mechanism shown in Scheme 2 is operating.
In tedmg, the formation of glycerol carbonate was slower
than in glycerol (0.3 mol/mol of catalyst per hour) most prob-
ably because of solubility limitation of both glycerol and the
complexes.
In order to gather further evidence of the carboxylation of
glycerol and exclude that the formation of glycerol carbonate
could occur via an alternative route, we have investigated the
trans-esterification of DMC with glycerol. In fact, one could
imagine that 1 might generate DMC and the latter could be then
converted into glycerol carbonate by trans-esterification. The
reaction of DMC with glycerol to afford glycerol carbonate is
catalysedby1, buttheconversionrateislowerthantheformation
of glycerol carbonate from glycerol and CO2 in presence of 1.
3. Results and discussion
Glycerol was converted into glycerol carbonate by reaction
with CO2 in presence of n-Bu2Sn(OCH3)2 at 450 K. The reac-
tion was run either in glycerol or in tedmg as solvent (under
5 MPa of CO2). In the former conditions, glycerol carbonate
(crude product) was formed in a 1.14:1 molar ratio (see Table 1)
with respect to Sn within 15 h (isolated yield 95% with respect
to Sn) and the reaction proceeded, but with a much lower rate,
for several hours afterwards. The tin compound isolated at the
end of the reaction from the autoclave showed different prop-
erties and analyses than the starting catalyst: it resulted to be
analogous to the tin-compound recovered from the reaction of
carboxylation of methanol in presence of 1.
Therefore, we decided to investigate the behaviour of n-
Bu2Sn(OCH3)2 towards glycerol in absence and presence of
CO2 in order to understand its chemistry.
n-Bu2Sn(OCH3)2 reacted with glycerol at room tempera-
ture in toluene with elimination of methanol and formation
of n-Bu2Sn(glycerol-2H), 4. The reaction proceeded with an
acceptable rate (see Section 2) and the released methanol was
easily monitored by gas-chromatography. A white solid com-
pound slowly separated from the reacting mixture that gave
correct analyses for 4 and showed in its IR spectrum a broad
band at 3272 cm−1 due to ν(OH). This suggests that one OH
group of glycerol still bears hydrogen and is not bonded to Sn.
A cristallization of the compound was attempted using several
conditions that did not afford yet crystals or crystalline-powders
suitable for X-ray structure determination. The isolated com-
pound was most probably oligomeric as inferred on the basis of
its poor solubility in all organic solvents, except alcohols. How-
ever, in the latter solvents an exchange took slowly place and

M. Aresta et al. / Journal of Molecular Catalysis A: Chemical 257 (2006) 149–153
153
Scheme 2. Reaction mechanism of formation of glycerol carbonate from glycerol and CO2 under n-Bu2Sn(OMe)2 catalysis.
In fact, while an equimolar solution of 1 and glycerol under CO2
after 15 h afforded the total conversion of glycerol into glycerol
carbonate, the trans-esterification of DMC with an equimolar
amount of glycerol in presence of 1 gave a 65% conversion after
15 h.
References
[1] M. Aresta, A. Dibenedetto, M. Carone, T. Colonna, C. Fragale, Environ.
Chem. Lett. 3 (2005) 136–139.
[2] A. Robles Medina, L. Esteban Cerda`n, A. Gime`nez, B. Camacho Pa`ez, J.
Bioechnol. 70 (1999) 379–391.
We have also used n-Bu2SnO, 2 and Sn(OMe)2, 3 as catalysts
in the direct carboxylation of glycerol.
The latter easily reacted with CO2 to afford the emicarbon-
ate [(MeO)Sn[OC(O)OMe]]n 5 that was not able to afford any
further reaction. 2 was able to catalyse the formation of glycerol
carbonate also if with a low reaction rate (see Table 2).
[3] M. Roice, K.P. Subhashchandran, A.V. Gean, J. Franklin, V.N. Rajasekha-
ran Pillai, Polymer 44 (2003) 911.
[4] R. Wilson, B.J. van Schie, D. Howes, Food Chem. Toxicol. 36 (1998)
711–718.
[5] R. Garcia, M. Besson, P. Gallezot, Appl. Catal. A: Gen. 127 (1995)
165–176.
[6] P. Wittlich, A. Themann, K.D. Vorlop, Biotechnol. Lett. 23 (2001) 463–
466.
[7] J.M. Clacens, Y. Pouilloux, J. Barrault, Appl. Catal. A: Gen. 227 (2000)
181.
4. Conclusions
[8] V. Plasman, T. Caulier, N. Boulos, Plast., Addit. Compd. 7 (2005) 30–
33.
[9] C. Vieville, J.W. Yoo, S. Pelet, Z. Mouloungui, Catal. Lett. 56 (1998)
245–247.
[10] M. Aresta, A. Dibenedetto, L. Gianfrate, C. Pastore, J. Mol. Catal. A: Chem.
204–205 (2003) 245–252.
[11] M. Aresta, A. Dibenedetto, C. Pastore, Stud. Surf. Sci. Catal. 153 (2004)
221.
[12] M. Aresta, A. Dibenedetto, C. Pastore, ICCDU VIII, Oslo, Norway, June
20–23, 2005.
[13] M. Aresta, A. Dibenedetto, C. Pastore, Catal. Today 115 (1–4) (2006)
88–94.
[14] D.D. Perrin, W.L.F. Armarego, D.R. Perrin, Purification of Laboratory
Chemicals, Pergamon, Oxford, England, 1986.
[15] G. Charlot, Colorimetric Determination of Elements, Amsterdam, New
York, Elsevier Pub. Co, 1964.
[16] D. Ballivet-Tkatchenko, O. Douteau, S. Stutzmann, Organometallics 19
(2000) 4563–4567.
n-Bu2Sn(OMe)2 and n-Bu2SnO were able to promote the
direct carboxylation of glycerol into glycerol carbonate. The
most active among the catalysts tested was n-Bu2Sn(OCH3)2.
The carbonate was formed with an appreciable rate until a 1.14:1
molar ratio of carbonate with the catalyst was reached. After-
wards, the original catalyst was converted into an oligomer
which showed a moderate catalytic activity. The monomeric
species formed from the reaction of 1 with glycerol was able
to react with CO2 until it was a monomer. The formation of
an oligomer reduced the reactivity towards CO2 and, thus, the
catalytic activity.
Acknowledgements
The Italian authors wish to thank the University of Bari and
the EU Project TOPCOMBI for financial support. One of us
(Francesco Nocito) acknowledges a grant from TOPCOMBI.
[17] D. Ballivet-Tkatchenko, ICCDU VIII, Oslo, Norway, June 20–23, 2005.
[18] D. Ballivet-Tkatchenko, R. Burgat, S. Chambrey, L. Plasseraud, P. Richard,
J. Organomet. Chem. 691 (8) (2006) 1498–1504.