Design of well balanced hydrophilic-lipophilic catalytic surfaces for the direct and selective monoesterification of various polyols

Article abstract of DOI:10.1039/b418509a

The transesterification process is a well known reaction of organic chemistry. However, the monoesterification of unprotected polyols such as glycerol or sucrose is much more complex and the design of selective catalysts is becoming a huge challenge in order to avoid many protection and deprotection steps, harmful for the cost and the environmental impact of the resulting process. In this study, we showed that the control of the hydrophilic-lipophilic balance of heterogeneous catalysts is a crucial key in order to tune both the catalyst activity and the monoester selectivity. Indeed, whereas homogeneous guanidine led to low selectivity toward monoesters, its anchorage on a hydrophilic solid support such as silica allowed us to prepare two basic hybrid organic-inorganic materials able to selectively afford monoesters in high yield and in an environmentally-friendly process, at low temperature and starting from an equimolecular mixture of unprotected polyols and various fatty methyl esters. The Royal Society of Chemistry and the Centre National de la Recherche Scientifique 2005.

Full text of DOI:10.1039/b418509a

View Article Online / Journal Homepage / Table of Contents for this issue
P A P E R
Design of well balanced hydrophilic–lipophilic catalytic surfaces
for the direct and selective monoesterification of various polyolsw
ˆ
´ ¨
Ghizlane Kharchafi, Franc¸ ois Jerome,* Isabelle Adam, Yannick Pouilloux and Joel Barrault
´
Laboratoire de Catalyse en Chimie Organique, UMR 6503 CNRS-Universite de Poitiers, ESIP,
40 avenue du Recteur Pineau, 86022 Poitiers, France. E-mail: francois.jerome@univ-poitiers.fr;
Fax: þ33 5 49 41 40 52; Tel: þ33 5 49 41 33 49
Received (in Montpellier, France) 9th December 2004, Accepted 5th April 2005
First published as an Advance Article on the web 20th May 2005
The transesterification process is a well known reaction of organic chemistry. However, the
monoesterification of unprotected polyols such as glycerol or sucrose is much more complex and the design
of selective catalysts is becoming a huge challenge in order to avoid many protection and deprotection steps,
harmful for the cost and the environmental impact of the resulting process. In this study, we showed that the
control of the hydrophilic–lipophilic balance of heterogeneous catalysts is a crucial key in order to tune both
the catalyst activity and the monoester selectivity. Indeed, whereas homogeneous guanidine led to low
selectivity toward monoesters, its anchorage on a hydrophilic solid support such as silica allowed us to
prepare two basic hybrid organic–inorganic materials able to selectively afford monoesters in high yield and
in an environmentally-friendly process, at low temperature and starting from an equimolecular mixture of
unprotected polyols and various fatty methyl esters.
Introduction
selectivities higher than 90% were only obtained for conver-
sions lower than 20% mainly because, beyond 110 1C, the fatty
acid was able to autocatalyze the reaction.4c Indeed, at 110 1C
but without catalyst, glycerol esters were produced, after 16
hours, in more than 30% yield and with a monoglyceride
selectivity higher than 47% making difficult the development
of a highly selective process. Up to now, only methods invol-
ving enzymes were able to quantitatively produce glycerol
esters with a monoglyceride selectivity higher than 95% since
enzymes were able to catalyze the reaction at room tempera-
ture avoiding the autocatalysis of the reaction by the fatty
acids.8 However, this chemical route requires a large excess of
glycerol, five fold excess, and a high dilution of the fatty acids
(0.1 M) to limit the production of di- and triglycerides.
Due to the exhaustion of the earth petroleum oil reserves and
the excessive production of greenhouse gases, intensive efforts
over the last decade were directed towards the chemical
potentialities of agroresources.1 However, molecules extracted
from biomass are often polyfunctionalized and the design of
selective catalysts is becoming a tremendous challenge in order
to avoid many protection and deprotection steps harmful for
the cost and the environmental impact of the process.2
Glycerol is industrially available in large scale from the
hydrolysis or methanolysis of vegetable oils and its use as
raw starting material is of great interest for the industrial
development of triglycerides (biofuels, biolubricants,
cosmetics. . .).3 The glycerol transformation into fatty mono-
esters is one of the viable solutions since monoglycerides find
significant applications as safe and biodegradable emulsifiers in
industry.3
In order to avoid the autocatalysis of the reaction, our group
extensively studied the monoglyceride production through a
transesterification process between glycerol and various fatty
5b
methyl esters.5 Metal oxides such as MgO, La2O3 and Mg-
In the literature, monoglycerides are synthesized either by
reaction of glycerol with fatty acids4 or fatty methyl esters5 and
by glycerolysis of vegetable oils.6 Hydroxide or carbonate
anions and p-toluenesulfonic acid are the most commonly used
homogeneous catalysts.7 However, these catalysts lead to the
formation of monoesters in a limited purity due to the im-
portant side production of di- and triglycerides and soap.7
Moreover, at the end of the reaction, acidification of the
reaction mixture affords salts, making the purification work-
up difficult. Heterogeneous catalysts offer many advantages
like easy separation from the reaction products and recycling.
Acidic anion exchange resins,4a mesoporous materials functio-
nalized with sulfonic groups4b,f–h and zeolithes4d,e are among
the most efficient catalysts for the esterification of glycerol with
fatty acids. These catalysts afforded, at 110 1C and under
solvent free conditions, glycerol esters in more than 90% yield
but with a poor monoglyceride selectivity of 40% which
rapidly dropped to 20% at total conversion.4 Monoglyceride
AlMCM-415c were mainly investigated. As expected these solid
catalysts led, at 80% conversion, to higher monoglyceride
selectivities (55%) than those previously obtained. However,
due to the greater lipophilicity of the fatty methyl esters
compared to that of fatty acids, these catalytic processes
require high temperature (220 1C) or the use of a solvent in
order to allow a better interaction between glycerol and fatty
methyl esters.
In an earlier paper, we reported that polystyrene bearing
guanidine groups as strongly basic catalytic sites (PS-TBD with
TBD ¼ triazabicyclo[4.4.0.]dec-5-ene) was an efficient solid
catalyst for the transesterification of various fatty methyl esters
with glycerol and that without solvent and at a lower tempera-
ture (110 1C) than those usually required.5a Starting from a
stoichiometric mixture of glycerol 1 and methyl dodecanoate
3a, this solid catalyst afforded glycerol esters in more than 96%
yield and with 48% selectivity toward monoglyceride 1a
(Scheme 1). However, when heating of the reaction mixture
was maintained after total consumption of 3a or when more
hydrophobic fatty methyl esters such as 3b–d were used, the
w Electronic supplementary information (ESI) available: experimental
details. See http://www.rsc.org/suppdata/nj/b4/b418509a/
T h i s j o u r n a l i s & T h e R o y a l S o c i e t y o f C h e m i s t r y a n d t h e
C e n t r e N a t i o n a l d e l a R e c h e r c h e S c i e n t i f i q u e 2 0 0 5
928
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4

View Article Online
Scheme 1 Catalytic esterification of glycerol and sucrose over sup-
ported guanidine.
selectivity toward 1a–d rapidly dropped, limiting the viability
of the process. This change in selectivity resulted from the
hydrophobic nature of PS-TBD which interacted with mono-
glycerides 1a–d more strongly than with glycerol, leading to a
poor monoglyceride selectivity.5a
Based on these preliminary results, we report in this paper
that the grafting of guanidine derivatives on a hydrophilic
framework such as silica affords a very versatile catalyst able to
produce monoglycerides 1a–d or monosucroesters 2a with
higher selectivity, and that at lower temperature, than conven-
tional catalysts, in a more environmentally-friendly process
and directly starting from an equimolar mixture of polyol and
various fatty methyl esters.
Scheme 2 Synthetic procedure for KG-TBD and HMS-TBD: i) non
porous silica, toluene, reflux, 24 h; ii) TBD, toluene, reflux, 12 h; iii)
methanol, triethylamine, 01C, 12 h; iv) tetraethoxysilane, dodecyla-
mine, water, ethanol; v) HMDS, reflux, 24 h.
groups for trimethoxysilane groups (Scheme 2).12 The replace-
ment of the trichlorosilane by the trimethoxysilane groups was
easily monitored by 1H NMR spectroscopy, where the appari-
tion of an intense peak localized at 3.60 ppm and assigned to
the –Si(OCH3)3 group was observed. This prior transformation
of the trichlorosilane groups was necessary in order to avoid an
important, harmful, evolution of HCl during the next stage. In
the second step, the HMS-Cl was prepared by copolymeriza-
tion of tetraethoxysilane (TEOS) with the trimethoxy[4-(chlor-
omethyl)phenyl]silane, in a molar ratio TEOS–organosilane of
15, in the presence of dodecylamine as structure directing agent
(Scheme 2). Our procedure was inspired by the method de-
scribed by Macquarrie et al. who gave rise to the so-called
HMS silica.13 The organic loading of HMS-Cl was determined
by elemental analysis data which indicated the presence of 2.02
mmol of chlorobenzyl groups per gram of silica.
Then, for both HMS-Cl and KG-Cl solids, the guanidine
heterogeneization procedure was similar. TBD species were
anchored to the functionalized silica by nucleophilic attack on
the benzylic position, affording the KG-TBD and the HMS-
TBD catalysts (Scheme 2). However, as described in the
literature for the PS-TBD catalyst, a guanidine quaternization
occurred during the grafting process due to the in situ evolution
of HCl and so consequently the unavoidable partial protona-
tion of the guanidine sites.9d,11e,14 This guanidine quaterniza-
tion was clearly observed by infrared spectroscopy. Indeed, the
IR spectra of KG-TBD and HMS-TBD exhibited two bands
located at 1642 and 1618 cmꢀ1 and respectively attributed to
the nCQN stretching of the guanidine and guanidinium chloride
groups (Fig. 1). In order to deprotonate all guanidinium sites,
the two hybrid silica-TBD materials were treated with a
solution of 0.1 M free TBD in acetonitrile. After this post
treatment, the IR spectra of both KG-TBD and HMS-TBD
exhibited only one single band located at 1642 cmꢀ1, indicating
that no more guanidine sites were protonated (Fig. 1).
Results and discussion
Catalyst preparation
Gelbard et al. extensively studied the guanidine moiety and its
use as a basic catalyst in various reactions.6b,9 Based on his
results, several researchers grafted guanidine to the surface of a
solid support in order to develop strongly basic heterogeneous
catalysts.10,11 However, if the catalytic activity of these materi-
als is now widely studied, their use as selective catalysts or
selective sorbents was much less investigated.
A
triazabicyclo[4.4.0.]dec-5-ene (TBD) derivative was
anchored to two different functionalized silicas prepared by
post-synthesis grafting and by a copolymerization procedure
affording KG-TBD and HMS-TBD, respectively. Usually, 3-
chloropropyltrimethoxysilane and 3-glycidoxypropyltrimetho-
xysilane are used as linkers between the guanidine and the silica
surface.11 However, in order to strictly show the influence of
the silica framework on the catalyst activity and the monoester
selectivity, we deliberately chose the same linker, i.e. the benzyl
group, as in the case of the PS-TBD catalyst. The synthetic
procedures for the two basic hybrid organic–inorganic materi-
als were inspired by the methods described in the literature
(Scheme 2).
For the preparation of KG-TBD, the first step was the
reaction of the trichloro[4-(chloromethyl)phenyl]silane with
the non porous silica (KG-60) in refluxing toluene (Scheme 2).
Elemental analysis of the collected KG-Cl revealed the presence
of 0.94 mmol of chlorobenzyl groups per gram of silica.
For the HMS-TBD synthesis, the trichloro[4-(chloro-
methyl)phenyl]silane was first reacted with methanol in the
presence of triethylamine in order to exchange the chlorosilane
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4
929

View Article Online
96% yield, the two HMS-TBD and KG-TBD catalysts were
totally inactive and no trace of glycerol esters was detected.
We assumed that this difference of reactivity between PS-
TBD, KG-TBD and HMS-TBD was the result of strong
hydrophilic interactions. Indeed, due to the greater hydrophi-
licity of the silica framework compared to that of polystyrene,
we suggested that glycerol was strongly adsorbed all around
the HMS-TBD and KG-TBD catalysts making impossible the
approach of the fatty methyl esters 3a–d to the TBD sites.
In order to validate this hypothesis, a phase transfer agent
was co-added with the HMS-TBD and KG-TBD catalysts.
Addition of 10 mol% of cetyltrimethylammonium bromide
(CTAB) to the reaction mixture immediately started the reac-
tion and glycerol esters were produced in more than 45% yield,
with a monoglyceride selectivity of 90%. However, no signifi-
cant difference in monoglyceride selectivity was observed be-
tween the PS-TBD and the two silica-TBD catalysts since the
CTAB broke all the hydrophilic interactions expected to raise
the monoglyceride selectivity.
Catalysis in DMSO. In order to gather more information
about the key role played by the silica framework, the glycerol
was diluted in a hydrophilic solvent with the aim of limiting its
adsorption to the HMS-TBD catalyst and making easier the
approach of the methyl dodecanoate to the grafted TBD
catalytic sites. Dimethyl sulfoxide (DMSO) was chosen as the
hydrophilic solvent because this latter is today one of the rare
organic solvents still authorized in the food and cosmetic
industries. DMSO was progressively added to the reaction
mixture containing the HMS-TBD solid and, when the glycerol
concentration reached 2 mol Lꢀ1, the reaction immediately
started. It is worth noting that below this concentration no
reaction occurred, which confirmed that the HMS-TBD–gly-
cerol interaction was strong enough to totally inhibit the
catalytic process. Based on this result, the following catalytic
tests were carried out, all at 110 1C, under a static nitrogen
atmosphere, in the presence of 12 mol% of grafted guanidine
and starting from unprotected glycerol 1 (10 mmol) and methyl
dodecanoate 3a (10 mmol) diluted in 5 mL of DMSO.
Fig. 1 FT-IR spectra of (A) crude HMS-TBD and (B) HMS-TBD
after deprotonation.
The guanidine content of both catalysts was determined by
thermal analysis. For both catalysts, thermal analysis revealed
three different weight losses. The first weight loss, between
20 1C and 160 1C (ca. 5%), corresponded to the removal of
water from the solids whereas the two following weight losses,
which began at 250 1C and continued up to 650 1C, corre-
sponded to the loss of the benzyl guanidine groups from the
surface. In this range, the weight loss for KG-TBD was 20.8%
and 38.6% for HMS-TBD which, respectively, corresponded
to an organic loading of 0.9 mmol gꢀ1 and 1.7 mmol gꢀ1. These
results are in excellent agreement with those obtained from
elemental analyses (see ESIw).
Whereas, under these conditions, the PS-TBD catalyst led to
a very low formation of glycerol esters (2.1 mmol hꢀ1 per
equiv. TBD), the two silica-TBD solids initially catalyzed more
than 7 times faster and rapidly led to the formation of glycerol
esters in more than 96% yield (Table 1, Fig. 2).
As in the case of solvent free conditions, this important
difference in reactivity between HMS-TBD, KG-TBD and PS-
TBD catalysts was still the result of hydrophilic interactions.
Indeed, due to its greater hydrophilicity, HMS-TBD interacts
more strongly with glycerol than does PS-TBD, thus enhancing
the production of glycerol esters. Curiously, the KG-TBD
catalyst exhibited an intermediate catalytic activity since this
latter catalyzed faster than PS-TBD (15 mmol hꢀ1 per equiv.
TBD vs. 2.1 mmol hꢀ1 per equiv. TBD) but slower than HMS-
Catalyst activities
Catalysis in solvent free conditions. Both KG-TBD and
HMS-TBD solids were first tested as catalysts under a nitrogen
atmosphere and solvent free conditions. Typically, unprotected
glycerol 1 (10 mmol) was reacted with 10 mmol of a fatty
methyl ester 3a–d in the presence of 0.12 equivalent of sup-
ported guanidine (Scheme 1). The reaction mixture was stirred
at 110 1C in order to distil out the methanol produced in situ
and shift the equilibrium towards the formation of glycerol
esters.
Surprisingly, whereas PS-TBD was an efficient catalyst
leading to the formation of glycerol esters 1a–d in more than
Table 1 Impact of the surface hydrophilicity on the monoester selectivity and the catalyst activity
Molar yieldc[d] (%)
Activity (mmol hꢀ1
per equiv. TBD)a
Selectivitye[f]
(%)
Entry
Catalyst
Time (h)
Conversionb (%)
Mono-
Di-
Tri-
Acid
1
2
3
4
a
HMS-TBD
KG-TBD
25
15
16
2.1
20
40
96
96
91
96
60
53
48
44
14 [28]
20 [40]
16 [32]
19 [38]
0 [0]
0 [0]
8
3
9
5
81 [68]
73 [57]
74 [59]
67 [48]
b
Si-HMS-TBD
PS-TBD
40
220
0.5 [1.5]
3 [9]
Glycerol (10 mmol), methyl dodecanoate 3a (10 mmol), DMSO (5 mL), 0.12 equiv. TBD, 110 1C, under a static nitrogen atmosphere. Con-
version was calculated with respect to the fatty methyl ester: conversion ¼ 100 ꢁ (nmono þ 2ndi þ 3ntri þ nacid)/(nmono þ 2ndi þ 3ntri þ nacid
þ
c
d
nremaining ester). Molar yield based on glycerol: % product ¼ 100 ꢁ nproduct/ninitial glycerol
.
Molar yield based on the fatty methyl ester: % di ¼
e
100 ꢁ 2ndi/ninitial ester and % tri ¼ 100 ꢁ 3ntri/ninitial ester
.
Selectivity of glycerol transformation into mono-, di- and triglycerides: 100 ꢁ nmono
/
f
(nmono þ ndi þ ntri). Selectivity of fatty methyl ester transformation into mono-, di- and triglycerides: 100 ꢁ nmono/(nmono þ 2ndi þ 3ntri).
930
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4

View Article Online
Fig. 2 Comparison of initial reaction rate between PS-TBD, KG-
TBD and HMS-TBD catalysts.
Fig. 3 Monoglyceride selectivity vs. methyl dodecanoate conversion
for PS-TBD and HMS-TBD.
TBD (15 mmol hꢀ1 per equiv. TBD vs. 25 mmol hꢀ1 equiv.
TBD) (Table 1, entries 1, 2 and 4). This difference in behaviour
between KG-TBD and HMS-TBD was due to the greater
hydrophilicity of HMS-TBD compared to that of KG-TBD.
Indeed, Macquarrie et al. reported that the grafting of organic
species on a non porous silica such as KG-100 led to an
important decrease in the surface hydrophilicity, even with
low organic loading.13 However, hybrid organic–inorganic
materials such as HMS functionalized with 1 mmol gꢀ1 of
organic species and prepared as above by a cocondensation
procedure, exhibited a greater hydrophilicity due to the pre-
sence of numerous free silanol groups.15 Thus, the resulting
polarity of these materials closely approaches that of purely
siliceous compounds.13
In order to clearly illustrate the great contribution of HMS-
TBD to the reaction rate, the free silanol groups present at the
HMS-TBD surface were protected. As described in the litera-
ture, the HMS-TBD catalyst was mixed with neat hexamethyl-
disilazane and stirred under reflux for 24 h affording the
Si-HMS-TBD solid (Scheme 2).16 As expected, the passivated
Si-HMS-TBD catalyst led to a decrease in the initial reaction
rate, which dropped from 25 to 16 mmol hꢀ1 per equiv. TBD
(Table 1, entry 3). This latter practically behaved like the KG-
TBD catalyst, showing that the hydrophilic–lipophilic balance
of the catalytic surfaces was a determinant parameter for an
efficient esterification of glycerol.
DMSO phase, as was the case in the enzyme route, but closely
depends on the chemical nature of the solid support. Indeed, an
increase in the hydrophilic nature of the catalyst surface
allowed a stronger interaction with glycerol than with mono-
glycerides, thus raising the monoglyceride selectivity and af-
fording a more selective process. Moreover, contrary to what
was reported with all the other catalysts, the monoglyceride
selectivity did not dramatically drop between 80% conversion
(1a selectivity: 83%) and total conversion (1a selectivity: 81%),
pushing forward the very important role played by the siliceous
framework in the monoglyceride selectivity.
Extension to other fatty methyl esters. Contrary to what was
observed in DMSO with the PS-TBD catalyst, the HMS-TBD
solid was also able to catalyze the same reaction starting from
more hydrophobic fatty methyl esters such as methyl tetra-
(3b), hexa- (3c) and octadecanoate esters (3d) (Table 2). How-
ever, a strong deactivation of the catalyst was observed with
3b–d since glycerol esters were produced in only 75% yield and
with a 1b–d monoglyceride selectivity of 80% (Table 2, entries
2–4). Nevertheless, addition of an extra amount of HMS-TBD
catalyst (0.2 mol%) allowed us to obtain similar results to
those obtained starting from 3a since glycerol esters were
produced in more than 96% yield and with a 1b–d monogly-
ceride selectivity of 78%.
We suggested that, at 110 1C and under a static nitrogen
atmosphere, traces of water present in the reaction mixture
could not be distilled out and were trapped by the grafted
guanidine leading to the formation of fatty acid and conse-
quently to catalyst poisoning. As a perfect illustration, under a
static nitrogen atmosphere, fatty acids were detected in the
reaction mixture in 3–8% yield, confirming the harmful role of
water regarding the catalytic mechanism pathway (Table 2,
entries 1–4). Due to the higher miscibility of methyl dodecano-
ate 3a in the DMSO phase, the reaction proceeded faster with
3a than with 3b–d and so consequently the catalyst deactiva-
tion was less important leading to a higher glycerol ester
production.
At the end of the reaction, the catalyst was recovered by
filtration and reactivated by washing with a solution of free
TBD in order to deprotonate all grafted guanidinium and to
remove the fatty acid from the catalyst (see experimental
section). After this treatment, the reactivated solid exhibited
the same catalytic activity as the initial HMS-TBD, confirming
that guanidine was poisoned but not destroyed during the
catalytic process.
In order to ensure that no leaching occurred during the
catalytic process, the catalyst was removed from the reaction
mixture after 20% of conversion. The filtrate was then heated
and no reaction was observed, indicating that the catalytic
process only took place on the grafted guanidine.
Remarkably, we noticed that the HMS-TBD catalyst did not
only allow an improvement of the reaction rate in DMSO but,
as expected, also played a very important role in the mono-
glyceride selectivity. Indeed, HMS-TBD afforded, at total
conversion, a very high 3a monoglyceride selectivity of 81%
(Table 1, Fig. 3). In our knowledge, HMS-TBD is the first
example of a catalyst able to reach, at total conversion and
starting from a stoichiometric mixture of unprotected glycerol
1 and fatty methyl ester 3a, a monoglyceride selectivity higher
than 81% (or 68% if selectivity is based on 3a).
As previously illustrated, KG-TBD and Si-HMS-TBD ex-
hibited a weaker hydrophilicity compared to that of HMS-
TBD and led to a lower monoglyceride selectivity (73%, Table
1). For comparison, under the same conditions, the homoge-
neous Me-TBD and PS-TBD solid catalysts led, at total
conversion, to a monoglyceride selectivity of only 67%. This
result clearly indicates that the improvement of the monogly-
ceride selectivity obtained with the HMS-TBD catalyst was not
only the result of the poor miscibility of 3a into the glycerol–
Reaction under nitrogen flow. In order to avoid the catalyst
poisoning, the catalytic reaction was carried out at 110 1C but
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4
931

View Article Online
Table 2 Synthesis of monoesters of polyols over HMS-TBD catalysta
Molar yieldc[d] (%)
Mono- Di-
Fatty Glycerol–ester
molar ratio
Conversionb
Time (h) (%)
Selectivitye[f]
Entry Polyol ester
T (1C) N2
Tri-
Acid (%)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
2
2
3a
3b
3c
3d
3a
3b
3c
3d
3a
3a
3a
3a
3a
3a
3a
1
1
110
110
110
110
110
110
110
110
40
Static
20
21
22
20
3
96
76
75
75
98
94
96
94
20
93
96
96
95
96
92
60
47
47
48
47
46
45
46
20
51
72
94
94
36
69
14 [28]
12 [24]
11 [22]
12 [24]
24 [48]
24 [48]
24 [48]
24 [48]
0 [0]
0
0
8
5
4
3
0
0
0
0
0
1
0
0
0
7
6
81 [68]
80 [66]
80 [67]
80 [67]
65 [48]
66 [49]
64 [47]
66 [49]
100 [100]
72 [55]
86 [75]
99 [98]
99 [99]
61 [40]
90 [80]
2
Static
Static
Static
Flowi
Flowi
Flowi
Flowi
Flowi
Vacuumj 48
Vacuumj 30
Vacuumj 24
Vacuumj 40
3
1
0.5 [1.5]
4
1
0
1 [3]
0
5
1
6
1
3
7
1
4
1 [3]
0
8
1
4
9
1
48
0
10
11
12
13
14
15
a
1
40
40
19 [38]
12 [24]
1 [2]
1 [3]
0
2g
5g
5h
1
40
40
0
0.5 [1]
16 [32] 7 [21]k
0
110
110
Static
Static
20
5
4
7 [14]
1 [3]
b
Polyol (10 mmol), fatty methyl ester (10 mmol), DMSO (5 mL), 0.12 equiv. TBD. Conversion was calculated with respect to the fatty methyl
c
esters: conversion ¼ 100 ꢁ (nmono þ 2ndi þ 3ntri þ nacid)/(nmono þ 2ndi þ 3ntri þ nacid þ nremaining ester). Molar yield based on glycerol: % product ¼
d
100 ꢁ nproduct/ninitial
.
Molar yield based on the fatty methyl ester: % di ¼ 100 ꢁ 2ndi/ninitial
and % tri ¼ 100 ꢁ 3ntri
/
glycerol
ester
e
f
ninitial ester
.
Selectivity of glycerol transformation into mono-, di- and triglycerides: 100 ꢁ nmono/(nmono þ ndi þ ntri). Selectivity of fatty methyl
g
ester transformation into mono-, di- and triglycerides: 100 ꢁ nmono/(nmono þ 2ndi þ 3ntri). The quantity of DMSO was adjusted in order to have
a glycerol concentration of 2 M. Enzymatic conditions: glycerol concentration of 0.8 M. 15 mL minꢀ1
not detected by HPLC and their quantity was calculated based on the remaining quantity of methyl dodecanoate.
h
i
j
. 15 mmHg. Sucrose triesters were
k
with a nitrogen flow of 15 mL minꢀ1 in order to remove water
from the reaction mixture. Under these conditions, the reaction
proceeded much more rapidly and glycerol esters 1a–b were
produced in more than 95% yield in less than 4 hours instead
of 20 hours under a static nitrogen atmosphere (Table 2, entries
5–8). As expected, at the end of the reaction, no fatty acid was
detected and the recovered catalyst has been recycled at least
five times, with only a slight decrease in its catalytic activity
mainly due to the high hygroscopy of DMSO and glycerol
which made impossible the total removal of water from the
reaction mixture. However, this increase of the reaction rate
led to a significant drop in the monoglyceride selectivity from
81% to 65%, mainly because hydrophilic interactions of HMS-
TBD with monoglycerides were less optimal. This decrease in
the monoglyceride selectivity observed when the reaction rate
increases was also reported in the case of the enzyme route. To
optimise the interactions between HMS-TBD and monoglycer-
ides, the reaction temperature was decreased to 40 1C. At this
temperature, the reaction rate was too slow since methanol and
water were removed with difficulty from the reaction mixture,
even with a nitrogen flow of 15 mL minꢀ1, and glycerol esters
were produced in only 20% yield after 48 hours of reaction
(Table 2, entry 9). In order to solve this problem, the catalytic
reaction was carried out under 15 mmHg. Whereas most
heterogeneous catalysts were totally inactive at low tempera-
ture, strongly basic HMS-TBD catalyst was able to produce
glycerol esters in more than 93% yield and with a monoglycer-
ide selectivity higher than 72% at 40 1C (Table 2, entry 10). For
comparison, mesoporous silica functionalized with sulfonic
groups or zeolithes require a higher temperature (110 1C)
and more than two times excess of glycerol to reach 70%
selectivity (or 55% if selectivity is based on the fatty methyl
esters) at total conversion.4b
The reaction mixture was then stirred at 40 1C under 15
mmHg. Using these conditions, glycerol esters were produced
in 95% yield and with a very high monoglyceride selectivity of
99%, as was obtained by Otero et al., pushing forward the
great contribution of HMS-TBD compared to the other cata-
lysts (Table 2, entry 13). However, the enzymatic route remains
a very efficient approach since enzyme catalysis is almost 50
times faster than that of HMS-TBD.8b In order to increase the
reaction rate while keeping high monoglyceride selectivity,
similar reactions were performed at 2 M. As expected, when
the glycerol concentration was raised from 0.8 M to 2 M, the
reaction rate increased and glycerol esters were quantitatively
produced in less than 24 hours instead of 48 hours (Table 2,
entries 11 and 12). Moreover, no change of yield and mono-
glyceride selectivity was observed when the glycerol concentra-
tion increased from 0.8 M to 2 M, making of HMS-TBD a very
competitive catalyst. It is worth noting that, below a five fold
excess of glycerol, monoglycerides were always produced along
with 5 mol% of diglycerides.
In all catalytic reactions described above, the regioselectiv-
ities of the reactions were similar to those reported in the
literature since 1-monoglycerides and 2-monoglycerides were
obtained in a ratio 9 : 1 whereas 1,2-diglycerides and 1,3-
diglycerides were produced in a ratio 1 : 3.
Sucrose esterification. In the course of our work, it occurred
to us that HMS-TBD could be a versatile catalyst for the
selective esterification of other polyols and for this reason we
extended our procedure to the synthesis of monosucroesters 2a.
Typically, sucrose 2 (10 mmol) was reacted with methyl
dodecanoate 3a (10 mmol) in the presence of 12 mol% of
grafted TBD and the resulting mixture was stirred at 110 1C,
under static nitrogen, in 5 mL of DMSO. Whereas under these
conditions the PS-TBD was almost totally inactive, the HMS-
TBD solid, as in the case of glycerol, afforded sucroesters in
more than 90% yield and with a monosucroester selectivity
higher than 60% (Table 2, entry 14). This lower monoester
selectivity obtained in the case of sucrose compared to that
reached starting from glycerol was due to the higher number of
hydroxyl groups of the sucrose moiety. Using four times excess
Comparison with the enzyme chemical procedure. In order to
increase the monoglyceride yield and to directly compare our
catalyst to the enzymatic chemical procedure, the HMS-TBD
catalyst was tested under the same experimental conditions as
those required for the enzyme i.e. with a glycerol–fatty methyl
ester molar ratio of 5 and a glycerol concentration of 0.8 M.8b
932
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4

View Article Online
1
of sucrose, sucroesters were produced with 92% yield and a
monosucroester selectivity of 90% (Table 2, entry 15). In the
case of sucrose, it was impossible to determine the regioselec-
tivity of the reaction mainly due to important intramolecular
transesterification reactions. This phenomenon was clearly
evidenced in the literature by Queneau et al.17 This last
example clearly demonstrated that our catalytic procedure
can be easily generalised to other polyols, affording a very
versatile catalytic pathway.
dispersion in KBr. The H NMR spectrum of the trimethoxy
[4-(chloromethyl)phenyl]silane derivative was recorded on a
Bruker Avance 300 DPX 300. Chemical shifts are expressed in
ppm relative to Me4Si. Organic content of all hybrid organic–
inorganic materials was determined by microanalysis on an
NA 2100 instrument and by thermal analysis on a SBT 2960
TA Instrument.
Chromatographic analyses
The sucrose esterification reaction progress was monitored on
a Shimadzu HPLC (SIL 10A) equipped with a column Tou-
zard & Matignon Nucleosil C8 (250 mm ꢁ 4.6 mm) and using a
methanol–water (78 : 22) mixture as eluent with a flow of 0.8
mL minꢀ1. The detector was a refractometer (Waters 2410).
This analytical technique allowed the quantification of sucrose
mono- and diesters, methyl dodecanoate and dodecanoic acid.
The glycerol esterification reaction progress was monitored
and quantified on a Varian 3300 GPC equipped with a BPX5
column (12 m ꢁ 0.22 mm) supplied by SGE. Prior to analysis,
products were silylated according to the Sahasrabudhe method
described in ref. 18. This analytical technique allowed the
quantification of glycerol, fatty methyl esters 3a–d, fatty acids,
1-monoglycerides, 2-monoglycerides, 1,2-diglycerides, 1,3-di-
glycerides and triglycerides. For both HPLC and GC, the
molar composition of the mixture was determined by an
external calibration method. In this study, the monoglyceride
selectivity was calculated by both manners: based on the
glycerol transformation into 1a–d and based on the fatty
methyl ester transformation into 1a–d, in order to make easier
the comparison with all results reported in the literature and
industrial patents (Tables 1 and 2).
Conclusion
In this paper, guanidine moieties were anchored to two differ-
ent siliceous frameworks and the catalytic activity of the
resulting solids was investigated in the selective monoesterifica-
tion of unprotected glycerol. Here, we reported that the solid
supports on which are grafted the guanidines do not act as
simple spectators but actively participate in the catalytic
process. Indeed, we found that the greater hydrophilicity of
HMS-TBD silica compared to other catalysts such as KG-
TBD or PS-TBD, allowed the production of glycerol esters in
high yield and with a much higher monoglyceride selectivity
than those usually obtained. Moreover, due to its strong
basicity, HMS-TBD was able to catalyze the reaction at
40 1C whereas most of the other solid catalysts require higher
temperatures. Compared to the reference enzymatic chemical
route, HMS-TBD gave, under the same experimental condi-
tions, a similar yield and monoglyceride selectivity. The en-
zyme chemical procedure still has the great advantage of
catalyzing the reaction much more rapidly than HMS-TBD.
However, compared to the enzymatic chemical pathway,
HMS-TBD is able to catalyze the reaction in much less toxic
solvents and starting from a much more concentrated glycerol
solution than those required for the enzyme and that without
affecting the reaction yield or the monoglyceride selectivity.
Moreover, according to the experimental conditions used, the
HMS-TBD solid can be easily recovered by filtration and
reused at least five times with only a slight decrease in its
catalytic activity.
Catalytic tests
Glycerol 1 or sucrose 2 (0.01 mol) and methyl dodecanoate 1a
(2.14 g, 0.01 mol) were charged in a Schlenk tube and diluted in
5 mL of DMSO. The reaction mixture was stirred with a
magnetic stirring bar (600 rpm) and heated to the desired
temperature. Then 12 mol% of grafted TBD (1.3 g of KG-
TBD or 0.7 g of HMS-TBD) were added and the reaction
mixture was stirred either under a static nitrogen atmosphere,
nitrogen flow (15 mL minꢀ1) or under vacuum (15 mmHg) in
order to remove the methanol and/or water from the reaction
mixture. At the end of the reaction, the catalyst was removed
by filtration. The filtrate was then diluted in water and addition
of ethyl acetate (2 ꢁ 15 mL) allowed to extract the glycerol
esters and sucroesters. The organic phases were combined and
then washed three times with water (30 mL) and dried over
MgSO4. After removal of the organic solvent, the residue was
dried under vacuum (10ꢀ1 mmHg) for 15 h at 50 1C. The
residue isolated starting from glycerol was a white powder,
whereas starting from sucrose an oily residue was obtained.
Typically, as described in the literature, mono-, di- and trigly-
cerides can be separated by silica chromatography using ethyl
acetate–heptane (20 : 100 to 100 : 0) as eluent whereas all
different sucroesters can be separated using an initial mixture
CH2Cl2–acetone–methanol–water 56 : 20 : 20 : 4 as eluent.
HMS-TBD catalyst appears to be a very versatile catalyst for
the selective monoesterification of natural polyols since this
catalyst was also effective starting from other polyols such as
sucrose and from various fatty methyl esters. Moreover, it is
worth noting that during the catalytic process no secondary
products such as polyglycerol, acrolein issued from the dehy-
dration of glycerol, or other polyol degradation products were
detected affording a very green catalytic process.
Experimental
Chemicals
Industrial glycerol (97%) and fatty methyl esters 3a–d (99%)
were kindly provided by Stearinerie Dubois and sucrose was
supplied by Beghin-Say. Dimethyl sulfoxide, tetraethoxysilane
and hexamethyldisilazane (HMDS) were purchased from Ac-
ros. Triazabicyclo[4.4.0.]dec-5-ene (TBD), dodecylamine, tri-
chloro[4-(chloromethyl)phenyl]silane and PS-TBD catalyst
were purchased from Sigma Aldrich. PS-TBD solid catalyst
was a polystyrene network crosslinked with 2% of divinylben-
zene and functionalized with 2.5 mmol gꢀ1 of benzylguanidine.
Catalyst recycling
Silica (KG: Kieselgel 60, 0.040–0.063 mm, 480–540 m2 gꢀ1
)
At the end of the reaction, the catalyst was removed by
filtration at a temperature higher than 60 1C. Under nitrogen
flow or under vacuum, the recovered catalyst was used as
collected without any further purification.
was purchased from Merck. All solvents and reagents were
used as received without any further purification. Chemical
procedures and physical properties of solid catalysts prepared
in this work are reported in the ESIw.
However, under a static nitrogen atmosphere, the HMS-
TBD was reactivated by washing this latter with a TBD
solution (0.1 M) in acetonitrile in order to remove fatty acid
from the TBD catalytic sites. Then HMS-TBD was transferred
to a cellulose extraction thimble and washed using a Soxhlet
Physical methods
IR spectra were recorded on a FT-IR Perkin Elmer (spectrum
GX) spectrophotometer and samples were prepared as a 1%
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4
933

View Article Online
apparatus with acetonitrile for 15 h before reusing. Elemental
analysis of the recovered HMS-TBD before reactivation: %C:
36.31, %H: 4.99, %N: 2.04, and after reactivation %C: 33.84,
%H: 4.57 %N: 7.09
Barrault, Appl. Catal., A, 2001, 218, 1; (c) J. Barrault, S. Banquart
and Y. Pouilloux, C. R. Chim., 2004, 7, 593.
6
(a) Y. Warabi, D. Kusdiana and S. Saka, Bioresour. Technol.,
2004, 91, 283; (b) L. M. G. Aguiar and R. M. Vargas, J. Am. Oil
Chem. Soc., 1998, 9, 199; (c) A. Corma, S. Iborra, S. Miquel and J.
Primo, J. Catal., 1998, 173, 315; (d) U. Schuchardt, R. Sercheli
and R. M. Vargas, J. Braz. Chem. Soc., 1998, 9, 199.
(a) N. O. V. Sonntag, J. Am. Oil Chem. Soc., 1992, 59, 795A; (b) S.
Kea and C. E. Walker, UK Patent 2163425, 1986; (c) V. K.
Babayan, A. K. Atikian and N. J. Boonton, US Patent 2948717,
, 1960; (d) L. Hartman, Chem. Rev., 1958, 58, 845.
(a) S. Wongsakul, P. Prasertsan, U. T. Bornscheuer and A. H.
Kittikun, Eur. J. Lipid Sci. Technol., 2003, 2, 68; (b) C. Otero, J. A.
Arcos, M. A. Berrendero and C. Torres, J. Mol. Catal. B: Enzym.,
2001, 11, 883.
Acknowledgements
7
8
9
This work was supported by the CNRS and the AGRICE
action (contract no 01 01 016). Authors are grateful to
ADEME and Region Poitou-Charentes for a PhD grant to
´
IA and also to the Morocco government for the PhD grant
to GK.
References
(a) U. Schuchardt, R. M. Vargas and G. Gelbard, J. Mol. Catal.
A: Chem., 1995, 99, 65; (b) L. M. G. Aguiar and R. M. Vargas, J.
Am. Oil Chem. Soc., 1998, 6, 755; (c) R. Ballini, D. Fiorini, R.
Maggi, P. Righi, G. Sartori and R. Sartorio, Green Chem., 2003, 5,
396.
1
(a) R. Narayan, in Emerging Technologies for Materials and
Chemicals from Biomass, ed. R. M. Rowell, T. P. Schultz, and
R. Narayan, American Chemical Society, Washington, DC, 1992;
(b) M. Eissen, J. O. Metzger, E. Schmidt and U. Schneidewind,
Angew. Chem. Int. Ed., 2002, 41, 414; (c) J. Barrault, Y. Pouilloux,
J. M. Clacens, C. Vanhove and S. Bancquart, Catal. Today, 2002,
75, 177.
10 (a) G. Gelbard and F. Vielfaure-Joly, C. R. Acad. Sci., Ser. IIc:
Chim., 2000, 563; (b) G. Gelbard and F. Vielfaure-Joly, Tetrahe-
dron Lett., 1998, 39, 2743; (c) G. Gelbard and F. Vielfaure-Joly,
React. Funct. Polym., 2001, 48, 65; (d) U. Schuchardt, R. M.
Vargas and G. Gelbard, J. Mol. Catal. A: Chem., 1996, 109, 37.
11 (a) R. Sercheli, R. M. Vargas, R. A. Sheldon and U. Schuchardt,
J. Mol. Catal. A: Chem., 1999, 148, 173; (b) S. Carloni, D. E. De
Vos, P. A. Jacobs, R. Maggi, G. Sartori and R. Sartorio, J. Catal.,
2002, 205, 199; (c) Y. V. Subba Rao, D. E. De Vos and P. A.
Jacobs, Angew. Chem., Int. Ed. Engl., 1997, 36, 2661; (d) A.
Derrien, G. Renard and D. Brunel, Stud. Surf. Sci. Catal., 1998,
117, 445; (e) A. Barbarini, R. Maggi, A. Mazzacani, G. Mori, G.
Sartori and R. Sartorio, Tetrahedron Lett., 2003, 44, 2931; (f) R.
Ballini, G. Bosica, D. Fiorini, R. Maggi, P. Righi, G. Sartori and
R. Sartorio, Tetrahedron Lett., 2002, 43, 8445; (g) A. C. Blanc, D.
J. Macquarrie, S. Valle, G. Renard, C. R. Quinn and D. Brunel,
Green Chem., 2000, 2, 283.
2
3
(a) Handbook of Green Chemistry & Technology, ed. J. H. Clark
and D. J. Macquarrie, Blackwell, Oxford, 2002; (b) J. H. Clark,
Green Chem., 1999, 1; (c) P. A. Anastas and J. C. Warner, Green
Chemistry: Theory and Practice, Oxford University Press, Oxford,
1998.
(a) M. Mittelbach and H. Junek, FR Patent 2603296, , 1988; (b) K.
Henkel, EP Patent 200982, , 1986; (c) Surfactants in Consumer
Products: Theory, Technology and Applications, ed. J. Falbe,
Springer-Verlag, New York, 1987, p. 547; (d) N. Garti, A. Aserin
and M. Fanum, Colloids Surf., A, 2000, 164, 27; (e) N. Garti, A.
Aserin and M. Fanum, Colloids Surf., A, 2000, 170, 1.
(a) Y. Pouilloux, S. Abro, C. Vanhove and J. Barrault, J. Mol.
Catal. A: Chem., 1999, 149, 243; (b) W. D. Bossaert, D. E. De Vos,
W. M. Van Rhijn, J. Bullen, P. J. Grobet and P. A. Jacobs, J.
Catal., 1999, 182, 156; (c) Y. Pouilloux, S. Metayer and J.
Barrault, C. R. Acad. Sci., Ser. IIc: Chim., 2000, 3, 589; (d) M.
4
12 R. J. P. Corriu, E. Lancelle-Beltran, A. Mehdi, C. Reye
Brandes and R. Guilard, Chem. Mater., 2003, 15, 3152.
´
, S.
13 D. J. Macquarrie, D. B. Jackson, J. E. G. Mdoe and J. H. Clark,
New J. Chem., 1999, 23, 539.
14 A. Cauvel, G. Renard and D. Brunel, J. Org. Chem., 1997, 62, 749.
15 (a) L. Mercier and T. J. Pinnavaia, Chem. Mater., 2000, 12, 188;
(b) A. Sayari and S. Hamoudi, Chem. Mater., 2001, 13, 3151.
da Silva-Machado, D. Cardoso, J. Pe
Stud. Surf. Sci. Catal., 2000, 130, 3417; (e) M. da Silva-Machado,
J. Perez-Pariente, E. Sastre, D. Cardoso and A. M. de Guerenu,
´
rez-Pariente and E. Sastre,
´
Appl. Catal., A, 2000, 203, 321; (f) I. K. Mbaraka, D. R. Radu, V.
S. Y. Lin and B. H. Shanks, J. Catal., 2003, 219, 329; (g) I. Diaz,
16 R. J. P. Corriu, E. Lancelle-Beltran, A. Mehdi, C. Reye
Brandes and R. Guilard, J. Mater. Chem., 2002, 12, 1355.
´
, S.
F. Mohino, J. Pe
´
242, 161; (h) J. Pe
rez-Pariente and E. Sastre, Appl. Catal., A, 2003,
rez-Pariente, I. Diaz, F. Mohino and E. Sastre,
´
Appl. Catal., A, 2003, 254, 173.
17 Y. Queneau, J. Fitremann and S. Trombotto, C. R. Chim., 2004, 7,
177.
18 M. R. Sahasrabudhe, J. Am. Oil Chem. Soc., 1957, 44, 376.
5
(a) F. Jerome, G. Kharchafi, I. Adam and J. Barrault, Green
Chem., 2004, 72; (b) S. Banquart, C. Vanhove, Y. Pouilloux and J.
934
N e w J . C h e m . , 2 0 0 5 , 2 9 , 9 2 8 – 9 3 4