Catalytic etherification of glycerol with short chain alkyl alcohols in the presence of Lewis acids

Article abstract of DOI:10.1039/c3gc36944g

Here we report the homogeneously-catalyzed etherification of glycerol with short chain alkyl alcohols. Among the large variety of Bronsted and Lewis acids tested, we show here that metal triflates are not only the most active but are also capable of catalyzing this reaction with an unprecedented selectivity. In particular, in the presence of Bi(OTf)₃, the targeted monoalkylglyceryl ethers were obtained with up to 70% yield. Although tested Bronsted acids were also capable of catalyzing the etherification of glycerol with alkyl alcohols, they were found however less active and less selective than Bi(OTf)₃. By means of counter experiments, we highlighted that the high activity and selectivity of Bi(OTf)₃ may rely on a synergistic effect between Bi(OTf)₃ and triflic acid, a Bronsted acid that can be released by in situ glycerolysis of Bi(OTf)₃. The scope of this methodology was also extended to other polyols and, in all cases, the monoalkylpolyol ethers were conveniently obtained with fair to good yields.

Full text of DOI:10.1039/c3gc36944g

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Catalytic etherification of glycerol with short chain
alkyl alcohols in the presence of Lewis acids†
Cite this: Green Chem., 2013, 15, 901
Fei Liu,a Karine De Oliveira Vigier,a Marc Pera-Titus,b Yannick Pouilloux,a
Jean-Marc Clacens,b Floryan Decampo*b and François Jérôme*a
Here we report the homogeneously-catalyzed etherification of glycerol with short chain alkyl alcohols.
Among the large variety of Brønsted and Lewis acids tested, we show here that metal triflates are not
only the most active but are also capable of catalyzing this reaction with an unprecedented selectivity. In
particular, in the presence of Bi(OTf)3, the targeted monoalkylglyceryl ethers were obtained with up to
70% yield. Although tested Brønsted acids were also capable of catalyzing the etherification of glycerol
with alkyl alcohols, they were found however less active and less selective than Bi(OTf)3. By means of
counter experiments, we highlighted that the high activity and selectivity of Bi(OTf)3 may rely on a syner-
gistic effect between Bi(OTf)3 and triflic acid, a Brønsted acid that can be released by in situ glycerolysis
of Bi(OTf)3. The scope of this methodology was also extended to other polyols and, in all cases, the
monoalkylpolyol ethers were conveniently obtained with fair to good yields.
Received 3rd December 2012,
Accepted 28th January 2013
DOI: 10.1039/c3gc36944g
www.rsc.org/greenchem
the green house gases emissions in surfactant industries.3 Of
course, such a benefit can only be achieved by using sustain-
Introduction
With an annual production of about 14 million tons in 2011, able feedstocks that are converted through high eco-efficient
surfactants represent a huge market with a turnover estimated processes while the targeted bio-based surfactants should have
around 23 billion US dollars per year.1 Surfactants nowadays an acceptable cost performance balance.
exhibit a broad range of applications and are mainly used in
three key markets which are detergents, personal care and and a hydrophilic part, also so-called the tail and head,
other industrial applications (paints, coating, cleaning, etc.). respectively. Renewably-sourced hydrophobic moieties are
Surfactants are a class of molecules bearing a hydrophobic
Surfactants are typically classified according to the charge usually derived from tallow triglycerides that can be issued
on their hydrophilic head into four different categories: non- either from an animal or vegetal origin. Nowadays, a wide
ionic, anionic, cationic and amphoteric. It must be noted that variety of fatty acids with different chain lengths are commer-
more than 80% of the volume of surfactants are of the non- cially produced from different feedstocks such as palm oil,
ionic or anionic types.
rapeseed, soybean or coconut. These fatty acids can be then
Historically, the production of surfactants from petrochem- transformed into alcohols or amines or used directly to
icals has been favored over oleochemicals and still represents prepare surfactants.4 Owing to their large availability, fatty
the major part. In recent years, the continuous increase of the derivatives allowed renewably-sourced hydrophobic moieties to
fossil oil price has boosted the research and development on grow significantly in surfactants production to nowadays rep-
renewable resources such as oleochemicals, carbohydrates or resent almost half of the volumes.
organic acids.2 Few studies have recently pointed out that the
On the other hand, a particular attention has been also
use of such feedstocks has the potential to significantly reduce given to the search of renewably sourced hydrophiles in order
to produce 100% bio-based surfactants. In particular, non-
ionic hydrophiles capable of replacing ethoxylated products
aInstitut de Chimie des Milieux et Matériaux de Poitiers, CNRS/Université de
are highly desirable. Glycerol or carbohydrates have the poten-
Poitiers/ENSIP, 1 rue Marcel Doré, 86022 Poitiers, France.
tial to be promising renewable hydrophiles but are only scar-
E-mail: francois.jerome@univ-poitiers.fr; Fax: +33 5 49 45 33 49;
cely used today in the field of surfactants.3 If ester derivatives
Tel: +33 5 49 45 40 52
bEco-Efficient Products and Processes Laboratory, Unité Mixte de Recherche UMI
of glycerol or carbohydrates can be easily prepared, their per-
3464 CNRS/RHODIA, 3966 Jin Du Road, Shanghai 201108, China.
formances and especially their stabilities compared to ether
E-mail: floryan.decampo@ap.rhodia.com; Fax: +86 21 54425911;
derivatives are however not acceptable in many applications.
Tel: +86 21 24089391
For this reason, much effort has been made towards the
†Electronic supplementary information (ESI)
available. See DOI:
synthesis of more robust bio-based surfactants through
10.1039/c3gc36944g
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 901–909 | 901

View Article Online
Paper
Green Chemistry
etherification or reductive amination of polyols for instance. as promising molecules especially for the home and personal
In this context, catalytic synthesis of glycerol ethers has care market but they also have the potential to act as a safe
received considerable attention due to the robustness of these defoamer in various industrial applications.
chemicals and the very attractive price of glycerol (<0.6 € kg−1),
a co-product of the vegetable oil industry. In particular, alkyl
(poly)glyceryl ethers have promising applications in many
Experimental
Chemicals
industrial fields.5 For instance, they nowadays enter in the
composition of deodorants, detergents, oily foaming aerosols,
All chemicals were used as received. Glycerol (Stearinerie
cleaning cosmetics, hair dyeing agents, surface active agents,
Dubois, 99%), 1-butanol (Acros, 99%), 1-pentanol (Prolabo),
among many other applications. However, one should
1-hexanol (Merck), 1,2-propanediol (Fluka, >99%), 1,3-
mention that glycerol is scarcely used in the synthesis of such
propanediol (Acros, 98%) 1,1,1-tris(hydroxymethyl)ethane
surfactants and more reactive chemicals such as glycidol, epy-
(Acros, 97%) were used for all syntheses. All metal triflates and
chlorhydrine are generally preferred in order to get acceptable
triflimidates were kindly provided by Rhodia-Solvay. Metal
yields to the desired alkyl (poly)glyceryl ethers. The price and
chlorides were purchased from Strem Chemicals.
toxicity of these precursors however dramatically limit the
industrial emergence of these non-anionic surfactants. Hence,
Typical procedure for the catalytic etherification of glycerol
few groups have attempted the synthesis of alkyl (poly)glyceryl
with short chain alcohols
ethers directly from glycerol, a renewable, cheap and safe
organic building block.
In a typical procedure, glycerol (40 mmol) and n-butanol
(10 mmol) were mixed together in the presence of an acid cata-
lyst (6.5 mol%) prior to heating at 150 °C for 24 h in a sealed
glass reactor. Note that the solution occupies 80% of the
volume of the reactor in order to limit the presence of n-
butanol in the gas phase. At the end of the reaction, the
reactor was opened when it was still hot and a sample was
taken off for analysis.
Behr,6 Weckhuysen7 and others8 have shown that synthesis
of glycerol ethers can be achieved by catalytic telomerization of
glycerol with butadiene in the presence of palladium-phos-
phine complexes affording the corresponding unsaturated gly-
cerol monoethers with fair to good yields. Alternatively,
Lemaire and co-workers have recently reported the synthesis of
glycerol ethers by reductive alkylation of either aldehydes or
carboxylic acids in the presence of Pd/C and fair to good yields
are also reported.9 The direct acid-catalyzed etherification of
glycerol with alkyl alcohols is even more attractive since it
avoids the use of transition metals. In this context, most of the
works have been dedicated to the etherification of glycerol
with activated alcohols such as benzyl or tertiobutyl alcohols
in the presence of various acid catalysts.10 Unfortunately, alkyl
alcohols are not eligible in such processes making this route
not viable for the synthesis of surfactants. We and others have
recently investigated the acid-catalyzed etherification of gly-
cerol with alkyl alcohols.11 Although formation of glyceryl
ethers has been observed, one should mention that these pro-
cesses suffer from an important lack of selectivity requiring a
complex and energy-consuming work-up.
Analytical methods
Gas chromatography analyses were performed on a Varian
3350 equipped with an on column injector (50 °C), an FID
detector (300 °C) and an HT5 column (25 m × 0.32 mm ×
0.1 μm). Prior to analysis, the sample was silylated as follows:
typically, to 30 mg of the reaction medium was added 100 μl of
diglyme (internal standard), 500 μl of pyridine and 250 μl of
hexamethyldisilazane (HMDS). The mixture was then heated
and stirred vigorously to dissolve all products. Then, 200 μl of
chlorotrimethylsilane (TMSCl) was added and the resulting
solution was kept at room temperature for 10 minutes. Next,
the sample was centrifugated prior to performing the analysis.
NMR spectra were recorded on BRUKER ADVANCE DPX 400
spectrometers at 400.13 MHz for 1H and 100.6 MHz for 13C.
Clearly the industrial emergence of alkyl glyceryl ethers
nowadays requires deeper investigations in order to better
control the selectivity of this reaction. In these processes, few
Characterization methods
issues need to be addressed. First it is necessary to avoid the Monobutylglyceryl ethers (B1G1) are produced as a mixture of
side dehydration and self-etherification of glycerol and alkyl two regioisomers with the main etherification on the primary
alcohols, side products that are very difficult to separate from hydroxyl group. Exact determination of the regioselectivity is
the reaction medium. Secondly, polyetherification of glycerol difficult to determine in this case due to the overlapping of the
with alkyl alcohols is also unwanted since these products do two regioisomer retention times in gas chromatography. Below
not have the desired surfactant properties. Here, we wish to are provided the analytical data for the 1-monobutylglyceryl
report a selective route capable of providing alkyl glyceryl ethers.
ethers with high yields directly from glycerol and alkyl alcohols
1H NMR (400 MHz, CDCl3, ppm): 0.90 (t, J = 7.2 Hz, 3H,
with a chain length within the range of 4–6 atoms of carbon. –CH3), 1.29–1.39 (m, 2H, –CH2–), 1.51–1.58 (m, 2H, –CH2–),
Note that both chemicals are nowadays available from 2.92–3.21 (m, 2H, 2× –OH), 3.42–3.51 (m, 4H, 2× –OCH2–),
biomass, either from vegetable oils or through the fermenta- 3.57–3.68 (m, 2H, –CH2–OH), 3.84 (m, 1H, –CH–O); 13C NMR
tion of carbohydrates. Owing to their low foaming properties, (100 MHz, CDCl3, ppm): 13.9 (–CH3), 19.3 (–CH2), 31.6 (–CH2),
these short chain alkyl glyceryl ethers are nowadays recognized 64.3 (–OCH2), 70.6 (–OCH), 71.5 (–OCH2–), 72.3 (–OCH2). Note
902 | Green Chem., 2013, 15, 901–909
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
that similar NMR spectra were collected in the literature for 4.22–4.23 (m, 1H, H5), 4.38 (d, J = 4 Hz, 1H, H2), 4.63 (t, J =
monopentyl- and monohexyl-glyceryl ethers.9c Only integration 4 Hz, 1H, H3). 13C NMR (100 MHz, CDCl3, ppm): 13.8 (–CH3),
of the fatty chain differs between each spectrum.
19.1 (–CH2), 31.8 (–CH2), 69.9 (–CH2), 70.7 (–O–CH2–), 75.8
Mono ethers obtained from 1,2-propylene glycol (mixture of (–O–CH2–), 76.5 (–CH–), 80.0 (–CH–), 80.3 (–CH–), 88.3 (–CH–).
regioisomer): 1H NMR (400 MHz, CDCl3, ppm): 0.91 (t, J =
7.2 Hz, 3H, –CH3), 1.09 and 1.13 (d, J = 6.4 Hz and J = 6.0 Hz,
2H, –CH2), 1.32–1.41 (m, 2H, –CH2–), 1.52–1.59 (m, 2H,
Results and discussion
Catalysis in the presence of Lewis acids
–CH2–), 2.61 (bs, 1H, –OH), 3.17–3.51 (m, 4H, 2× –CH2),
3.53–3.58 and 3.90–3.98 (m, 1H, –CH–O). 13C NMR (100 MHz,
CDCl3, ppm): 13.7 and 13.9 (–CH3), 15.9 and 18.6 (–CH3), 19.3
and 19.4 (–CH2), 31.7 and 32.1 (–OCH2), 66.3 and 71.1
(–OCH2), 66.4 and 75.7 (–OCH), 68.6 and 76.3 (–OCH2).
Mono ethers obtained from 1,3-propanediol: 1H NMR
(400 MHz, CDCl3, ppm): 0.89 (t, J = 7.2 Hz, 3H, CH3), 1.29–1.39
(m, 2H, –CH2–), 1.50–1.57 (m, 2H, –CH2–), 1.78–1.84 (m, 2H,
–CH2–), 2.61 (bs, 1H, –OH), 3.42 (t, J = 6.8 Hz, 2H, –CH2–O–),
3.59 (t, J = 5.6 Hz, 2H, –CH2–O–), 3.75 (t, J = 5.6 Hz, 2H, –CH2–
OH). 13C NMR (100 MHz, CDCl3, ppm): 13.9 (–CH3), 19.3
(–CH2), 31.7 (–CH2), 32.0 (–CH2), 62.3 (–CH2–OH), 70.3
(–OCH2), 71.1 (–OCH2).
In a first set of experiments, the acid-catalyzed etherification
of glycerol with n-butanol was investigated. The reaction was
first performed at 150 °C in the presence of a Lewis acid cata-
lyst. The glycerol/n-butanol molar ratio was fixed to 4 in order
to limit the dimerization of n-butanol. The reaction progress
(yield, conversion, selectivity, carbon mass balance) was moni-
tored by gas chromatography using diglyme as an internal
standard. For the sake of clarity, structures and abbreviations
of the different reactants and products used in this work are
illustrated in Fig. 1.
Various Lewis acids were first tested. Results are presented
in Table 1. Although a high catalyst loading (6.5 to 24 mol%)
was employed, metal chlorides such as AlCl3, FeCl3 and BiCl3
Mono ethers obtained from tris(hydroxymethyl)ethane: 1H
NMR (400 MHz, CDCl3, ppm): 0.80 (s, 3H, –CH3), 0.90 (t, J =
7.4 Hz, 3H, –CH3), 1.29–1.39 (m, 2H, –CH2–), 1.50–1.57 (m,
2H, –CH2–), 2.86 (bs, 2× –OH), 3.39–3.42 (m, 4H, –CH2–O–
CH2–), 3.55 (d, J = 10.8 Hz, 2H, –CH2–OH), 3.68 (d, J = 10.8 Hz,
2H, –CH2–OH). 13C NMR (100 MHz, CDCl3, ppm): 13.9 (–CH3),
17.2 (–CH3), 19.3 (–CH2), 31.6 (–CH2), 40.7 (–CH2), 67.8
(2× –CH2–OH), 71.7 (–OCH2), 76.9 (–OCH2).
1
Mono ethers obtained from isosorbide: H NMR (400 MHz,
CDCl3, ppm): 0.87 (t, J = 7.2 Hz, 3H, –CH3), 1.27–1.36 (m, 2H,
–CH2–), 1.51–1.58 (m, 2H, –CH2–), 3.18 (bs, –OH), 3.37–3.64
(m, 3H, –CH2– and H4), 3.83–3.97 (m, 4H, H1 and H6),
Fig. 1 Abbreviation of chemicals used in this work.
Table 1 Catalytic etherification of glycerol with n-butanol in the presence of Lewis acidsa
Conv.
Gly (%)
Conv.
BuOH (%)
B1G1
B2G1
B2
yield (%)
G2
C %
C %
Entry
Catalyst
yieldb (%)
yieldb (%)
yieldb (%)
(BuOH)c
(Gly)c
d
1
2
3
4
5
6
7
8
FeCl3
10
6
1
0
0
4
5
12
17
23
18
30
10
6
13
30
6
0
0
20
30
31
70
80
65
91
55
52
40
9
16
4
0
0
12
19
28
45
62
48
70
32
17
28
0
0
0
0
0
0
0
0
0
2
0
7
0
0
0
2
1
1
0
0
3
5
1
10
9
8
10
3
4
2
2
2
0
0
0
1
1
5
6
6
4
8
2
1
7
69
53
—
100
—
75
80
83
79
91
87
95
64
40
75
43
100
100
100
—
100
100
100
100
93
89
85
99
71
AlCl3, 6H2Oe
BiCl3
CrCl3·6H2Of
g
ZnCl2
Nd(OTf)3
Fe(OTf)3
Sc(OTf)3
In(OTf)3
Ga(OTf)3
Al(OTf)3
Bi(OTf)3
Al(TFSI)3
Fe(TFSI)3
Bi(TFSI)h
9
10
11
12
13
14
15
20
70
a Glycerol/n-butanol molar ratio = 4, 150 °C, 24 h, 6.5 mol% of catalyst. b Mixture of regioisomers. c Carbon mass balance. d 22 mol%. e 16 mol%.
f 13 mol%. g 24 mol%. h At 190 °C.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 901–909 | 903

View Article Online
Paper
Green Chemistry
were poorly active in the catalytic etherification of glycerol with selectivity than those achieved with metal triflates were
n-butanol and the targeted B1G1 were produced with 16%, 9% obtained. Among all tested metal triflimidates (see the com-
and 4% yields, respectively, after 24 h of reaction (entries 1–3). plete list in ESI†), only Al(TFSI)3 and Fe(TFSI)3 were capable of
Additionally, at higher conversion of n-butanol (30%), the producing B1G1 with 32% and 17% yields, respectively, after
selectivity to B1G1 was rather low in the presence of 24 h of reaction at 150 °C. However, in both cases, the reaction
AlCl3·6H2O. Indeed, under these conditions, the n-butanol rate and selectivity (see carbon mass balance) were lower than
carbon mass balance was as low as 53% indicating the for- in the case of Bi(OTf)3. Note that, as compared to what was
mation of side products from n-butanol. Note that other metal observed with metal triflates, Bi(TFSI)3 was less active than
chlorides such as CrCl3 and ZnCl2, often used in acid-catalyzed Al(TFSI)3 and Fe(TFSI)3, affording the targeted B1G1 with 35%
dehydration of hexoses,12 were here totally inefficient, further yield only at 190 °C (entry 15). Taken together, these results
confirming the complexity of this reaction (entries 4, 5).
show that metal triflates are more active and more selective
To our delight, when 6.5 mol% of metal triflates were used than metal triflimidates themselves more active than metal
under similar conditions, the reaction was much more selec- chlorides. Among metal triflates, Bi(OTf)3 appears to be the
tive to the desired B1G1 (Table 1, entries 6–12). In particular, most efficient one.
in the presence of Bi(OTf)3, which is nowadays recognized as a
Next, we tried to rationalize the superior activity and selec-
“safe Lewis acid”, the B1G1 were obtained with 70% yield at a tivity of Bi(OTf)3. In water, few metal triflates are known to
conversion of 91% of n-butanol (entry 12). B2G2, B2 and G2 exist as an equilibrium mixture of metal triflates with metal
were produced in a limited amount with 7%, 10% and 8% hydroxide and triflic acid. Kobayashi classified metal triflates
yield, demonstrating that Bi(OTf)3 is selective (85% of B1G1 + according to their hydrolysis constant (Kh).14 Like water, gly-
B2G1) to alkyl glyceryl ethers. Note that the etherification of cerol is a highly polar protic solvent and one may suspect that
glycerol with n-butanol mainly occurred in the primary glycerol may lead to a glycerolysis of metal triflates inducing a
hydroxyl group of glycerol. Having this result in hand, other partial release of triflic acid along with the formation of a
metal triflates were tested. As shown in Table 1, Al, Ga, In, Sc, metal–glycerol complex (Fig. 2).
Fe and Nd triflates were also active in the etherification of gly-
Although pKh values provided by Kobayashi were collected
cerol with n-butanol, although differences of activity were in water, we observed, in a first approximation, that metal tri-
clearly observed. In all cases, the carbon mass balance on gly- flates with a pKh value lower than 4 (Bi, Ga, In, Al) were among
cerol and n-butanol remained higher than 85%, further con- the most active catalysts in this study (Table 2, entries 1–13).
firming the greater selectivity of metal triflates as compared to This result suggests that the superior activity of these metal tri-
that of metal chlorides. The order of activity was Bi > Al > Ga > flates may rely on the release of triflic acid (TfOH) in glycerol.
In > Sc > Nd. All other tested metal triflates were found ineffi- Fe and Nd, which exhibit a pKh of 9.5 and 8, respectively,
cient (see a complete list of tested metal triflates in Table 2).
however appear as two exceptions (entries 6, 7). Hence, we also
Metal triflimidates (M(TFSI)3) were also investigated as focused our attention on another parameter. Considering that
Lewis acids (Table 1, entries 13–15). In contrast to what was the release of TfOH is initially accompanied by the formation
reported earlier in water,13 in our case, lower activity and of a metal–glycerol-based complex, we next attempted to corre-
late the difference of activity of metal triflates with the covalent
radius of trivalent cations which is known to be a driving force
in the formation of metal–glycerol complexes (templating
effect). As previously observed by Djakovitch et al., glycerol was
supposed to readily form a five-membered ring complex with
metals.15 Formation rate and stability of these complexes are
of course closely depend on the size of the cation. Interest-
ingly, we observed that all active trivalent cations have a
covalent radius lower than 2 Å and preferentially lower than
1.4 Å for the most active ones supporting the possible for-
mation of a glycerol complex. We hypothesize here that metal
triflates with a trivalent cation covalent radius within the
range of 1.2–1.4 Å are more prone to glycerolysis and thus to
the release of TfOH explaining their superior activity. Fe and
Nd are very close to these criteria limits (pKh and covalent
Table 2 Covalent radius of mono-, di- and trivalent cations used and their pKh
values
Conv.
pKh
Covalent radius
of cation (A)
Entry
Cation
BuOHa (%)
valueb
1
2
3
4
5
6
7
8
Bi
Ga
Al
In
Sc
Nd
Fe
La
Y
Ce
Pr
Sm
Gd
Zn
Cu
Ca
Ag
91
80
65
70
31
30
20
0
0
0
0
0
1.09
2.60
4.97
4.00
4.30
8.00
9.50
8.50
7.70
8.30
8.10
7.90
8.00
8.96
7.53
12.85
12.00
1.17
1.22
1.21
1.42
1.74
2.01
1.52
2.07
1.90
2.04
2.03
1.98
1.96
1.22
1.32
1.76
0.45
9
10
11
12
13
14
15
16
17
0
0
0
0
0
a Collected after 24 h of reaction at 150 °C in the presence of 6.5 mol%
of metal triflate. b Collected in water.
Fig. 2 Plausible glycerolysis of Bi(OTf)3.
904 | Green Chem., 2013, 15, 901–909
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
radius) and thus promote the etherification of glycerol with (91% vs. 74% for Bi(OTf)3 and TfOH, respectively). A similar
n-butanol with only a very moderate activity. trend was observed when the catalytic reaction was conducted
It is noteworthy that mono and divalent cations such as Zn, in neat glycerol (absence of n-butanol). While in the presence
Cu, Ag and Ca are not capable of catalyzing the etherification of 6.5 mol% of Bi(OTf)3, 28% of glycerol was converted to 23%
of glycerol with n-butanol (entries 14–17). In particular, yield of G2 after 24 h of reaction at 150 °C, the reaction rate
although Zn and Cu have an optimal pKh value and a similar was divided by a factor of 2 when the same reaction was con-
covalent radius to Fe and Nd, these latter were totally ineffi- ducted in the presence of 6.5 mol% of TfOH further pushing
cient in pushing forward the formation of a key complex forward that TfOH cannot be considered as the real catalytic
between trivalent cation and glycerol, complex which is pre- species.
sumably stabilized by the presence of the three hydroxyl
With the aim of getting more insight into the possible role
groups of glycerol. This aspect is further supported at the end of TfOH in the catalytic reaction, 1,5-di-tertiobutylpyridine was
of the article where we show that replacing glycerol by a diol added at the beginning of the reaction. This sterically hin-
such as 1,2- or 1,3-propanediol led to a significant drop of the dered pyridine derivative has been reported to be capable of
Bi(OTf)3 selectivity.
selectively trapping protons from a solution containing metal
In order to discriminate between Lewis and Brønsted acid triflates.16 In particular, due to the presence of two sterically
catalysis in the etherification of glycerol with n-butanol, we hindered tertiobutyl groups, this pyridine derivative is not
next carried out few counter experiments. First of all the ether- capable of coordinating Bi(OTf)3. Interestingly, addition of
ification of glycerol with n-butanol was performed at low cata- 6.5 mol% of 1,5-di-tertiobutylpyridine at the beginning of the
lyst loading (2.17 mol%) either in the presence of Bi(OTf)3 or reaction completely inhibited the reaction showing that TfOH
TfOH (Table 3). Interestingly, under these conditions, Bi(OTf)3 does have a significant role in the reaction mechanism. Taken
was nearly two times more active than TfOH. For instance, together, these results suggest that the reaction mechanism is
after 24 h of reaction at 150 °C, B1G1 were obtained with 39% here rather complex and the formation of the real catalytic
yield with Bi(OTf)3 against 22% in the presence of TfOH. species may rely on a cooperative effect between Bi(OTf)3 and
Additionally, Bi(OTf)3 was found to be more selective than released TfOH. A similar phenomenon was previously observed
TfOH as illustrated by the n-butanol carbon mass balance by Le Roux and co-workers.17 At this stage, we are unfortu-
nately not able to clarify the reaction mechanism with certi-
tude and this aspect is now the topic of current investigations
in our groups. Results will be published in due course.
Next, we tried to optimize the catalytic reaction in the pres-
ence of Bi(OTf)3 by varying the temperature, time, catalyst
Table 3 Comparison of activity between Bi(OTF)3 and TfOH in the acid-cata-
lyzed etherification of glycerol with n-butanol and the dimerization of glycerol
loading and glycerol/n-butanol molar ratio (Table 4). An
Conv.
Conv.
B1G1
G2
C %
C %
increase of the reaction temperature from 150 °C to 170 °C
obviously had a beneficial effect on the conversion rate of
n-butanol and glycerol. At 170 °C, the conversion of n-butanol
reached 89% after only 8 hours of reaction vs. 24 h at 150 °C
(entries 2, 3). Interestingly, at such temperature, no significant
change of selectivity was observed. Only a slight decrease of
the n-butanol carbon mass balance (from 95% to 86%) was
Catalyst
Gly (%) BuOH (%) yield (%) yield (%) (BuOH) (Gly)
a
b
Bi(OTf)3
TfOHa
15
8
45
34
—
—
39
22
—
—
4
2
23
11
91
74
—
—
92
94
82
79
Bi(OTf)3
28
14
TfOHb
a Glycerol/n-butanol molar ratio = 4, 150 °C, 2.17 mol% of catalyst,
24 h. b From neat glycerol, 150 °C, 24 h, 6.5 mol% of catalyst.
Table 4 Optimization of the reaction parameters in the presence of Bi(OTf )3
Catalyst
Entry loading (mol%) (°C)
Temp. Gly/BuOH
Time Conv.
Conv.
B1G1
B2G1
B2
G2
C %
C %
molar ratio (h)
Gly (%) BuOH (%) yield (%) yield (%) yield (%) yield (%) (BuOH) (Gly)
1
2
3
4
5
6
7
8
9
10
6.5
6.5
6.5
6.5
6.5
13.0
2.2
6.5
6.5
6.5
120
150
170
190
190
170
170
150
170
170
4
4
4
4
4
4
4
4
4
2
72
24
8
2
3
No reaction
30
29
38
52
30
22
46
80
54
91
89
78
90
85
70
92
95
93
70
64
54
49
57
40
41
35
42
7
3
6
10
10
5
8
9
14
15
10
5
15
15
11
95
86
82
66
78
76
57
49
59
85
86
72
52
81
69
59
31
59
7
4
3
5
6
10
48
60
8
2
7
11
4
10
6
2
7
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 901–909 | 905

View Article Online
Paper
Green Chemistry
observed. A further increase of the reaction temperature from
170 °C to 190 °C enhanced again the conversion rate since
78% of n-butanol was now consumed within only 2 h (entry 4).
However, at 190 °C, the glycerol carbon mass balance was sig-
nificantly lowered (72%). This effect of the reaction tempera-
ture on the glycerol mass balance was even worse (52%) when
the conversion of n-butanol was extended to 90% (3 h, entry
5). Reversely, a decrease of the reaction temperature from
150 °C to 120 °C dropped the reaction rate to an unacceptable
level (entry 1). Hence, 170 °C appeared to be a good compro-
mise between reaction rate and selectivity. One should
comment that, in all cases, it was necessary to stop the reac-
tion immediately after the total conversion of n-butanol. After
a prolonged reaction, the yield of B1G1 indeed rapidly
decreased due to their acid-catalyzed dehydration or oligomeri-
zation. This effect was of course even more pronounced when
increasing the reaction temperature (entries 8, 9). We then
varied the catalyst loading. Reactions were carried out at
170 °C. Logically, when the catalyst loading was increased
from 6.5 mol% to 13 mol%, the reaction time was also conco-
mitantly decreased suggesting, in a first approximation, that
the reaction is not limited by diffusion problems (entry 6).
Reversely, a decrease of the catalyst loading by a factor 2
obviously decreased the reaction rate (entry 7). However, in
this case, the carbon mass balance was surprisingly lowered
(62% and 34% based on n-butanol and glycerol, respectively)
suggesting that, under these conditions, the catalytic for-
mation of B1G1 was not the dominant reaction anymore. This
unexpected change of selectivity when decreasing the catalyst
loading can be attributed either to (i) the thermal instability of
B1G1, i.e. in this case the formation rate of B1G1 may be lower
than their hypothetical thermal degradation, or (ii) a decrease
Fig. 3 Yields of B1G1 vs. reaction time at 150 °C, 170 °C and 190 °C.
Fig. 4
Cumulative yield of monobutylglyceryl ethers calculated on the basis
○
of glycerol after four consecutive additions of n-butanol and
carbon mass
balance based on glycerol.
of the catalyst amount mostly affects the formation rate of active and (2) an extra amount of n-butanol can be added at
B1G1 as compared to other side reactions. In order to discrimi- the end of the reaction.
nate between these two hypotheses, 1 eq. of B1G1 was heated
To this end, after completion of the reaction, an extra
in the presence of 3 eq. of glycerol at 150 °C for 24 h without amount of n-butanol was directly added to the reaction media.
the addition of any catalyst. Under these conditions, B1G1 The amount of added n-butanol was adjusted in order to main-
were stable suggesting that this decrease of selectivity cannot tain a (remaining) glycerol/n-butanol molar ratio of 4. Interest-
be attributed to the thermal degradation of B1G1. From this ingly, when using this procedure, the yield of B1G1
result, one may conclude here that a decrease of the catalyst (determined on the basis of n-butanol) remained constant
loading affects the etherification of glycerol with n-butanol to after three consecutive additions and obviously gradually
a greater extent than other side reactions such as B1G1 and increased from 18% to 35% when the yield was calculated on
glycerol oligomerization or dehydration for instance. Hence, a the basis of glycerol (Fig. 4).
catalyst loading of 6.5 mol% appears to be
compromise.
a
good
These results show that, under these conditions, the cata-
lyst is stable. However, after the fourth addition of n-butanol,
Next, we investigated the influence of the glycerol/n-butanol the selectivity to B1G1 started to drop and, in this case, the
molar ratio on the selectivity to B1G1. A decrease of the gly- conversion rate of B1G1 (dehydration, oligomerization)
cerol/n-butanol molar ratio from 4 to 2 did not dramatically becomes the dominant reaction as supported by the measure-
change the reaction rate (entry 10). The carbon mass balance ment of the carbon mass balance.
was however lowered presumably due to the formation of poly-
etherified glyceryl adducts that are unfortunately difficult to etherification of glycerol with n-butanol, we then screened the
detect and quantify using our analytical method.
catalytic activity of Brønsted acid catalysts in order to compare
Considering that TfOH has a significant activity in the
In Fig. 3 is plotted the formation of B1G1 as a function of them with Bi(OTf)3. All reactions were performed in the pres-
reaction time at 150 °C, 170 °C and 190 °C. From this figure, it ence of 6.5 mol% of Brønsted acid and heated at 150 °C for
is interesting to note that, in all cases, the yield of B1G1 did 24 h (Table 5). For all the tested Brønsted acids, the yield of
not reach a plateau suggesting that (1) the catalyst was still B1G1 was lower (46–56%) than in the case of Bi(OTf)3. In
906 | Green Chem., 2013, 15, 901–909
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Table 5 Catalytic etherification of glycerol with n-butanol in the presence of Brønsted acid catalysts or base catalystsa
Conv.
Gly (%)
Conv.
BuOH (%)
B1G1
yield (%)
B2G1
yield (%)
B2
yield (%)
G2
yield (%)
C %
(BuOH)
C %
(Gly)
Entry
Catalyst
1
2
3
4
5
6
7
TfOH
39
22
45
46
8
0
20
84
63
90
89
6
0
71
60
46
56
46
2
0
56
6
0
0
0
0
0
6
6
11
4
7
0
0
4
15
10
22
12
0
0
5
86
90
67
60
33
0
78
95
80
51
6
0
99
DBSAb
MSAc
H2SO4
Bet·HCld
Bet·H2SO4
TFSIe
d
93
a Glycerol/n-butanol molar ratio = 4, 150 °C, 6.5 mol% of catalyst, 24 h. b Dodecylbenzene sulfonic acid. c Methanesulfonic acid. d Bet =
(CH3)3N+(CH2)(COO−). e Bis(trifluoromethane)sulfonimide acid.
particular, H2SO4 was poorly selective and favours intramole-
cular dehydration reactions. Note that betaïn-hydrochloride
and betaïn hydrogenosulfate were found ineffective in the
etherification of glycerol with n-butanol, presumably due to
either their lower acidity as compared to other tested Brönsted
acids or their ability to form a deep eutectic mixture with gly-
cerol, thus lowering the acid strength of the carboxylic group
(entries 5, 6).18 Note that no esterification reaction with gly-
cerol was evidenced. Although triflimidic acid (HN(O2SCF3)2,
TFSI) was less active than Bi(OTf)3 (56% of B1G1 with TFSI vs.
70% with Bi(OTf)3 after 24 h of reaction), this Brønsted acid
catalyst was as selective as Bi(OTf)3 to alkyl glyceryl ethers
(87%, B1G1 + B2G1) and the carbon mass balance remained
higher than 93%.
Table 6 Acid-catalyzed etherification of various polyols in the presence of Bi
a
(OTf )3
Alkyl
alcohol
T
(°C)
Time
(h)
Monoether
yieldb (%)
Polyol
n-Pentanol
n-Hexanol
n-Butanol
150
150
150
24
24
24
50
32
46
n-Butanol
n-Butanol
170
150
32
48
41
38
Basic catalysts such as NaOH, K2CO3 and TBD, which are
also known to catalyze the oligomerization of glycerol, were
also tested.19 However, within a range of temperature of
150–200 °C, formation of B1G1 did not take place and only
unidentified products were observed at 200 °C (entries 8–10).
Having all these results in hand, we then examined the
scope of this route not only to determine the limit of this
a Glycerol/n-alkyl alcohol molar ratio = 4, 6.5 mol% of catalyst. b GC
yield.
the reaction was more difficult from 1,3-propanediol since, in
methodology but also to get more insight into the unprece- this case, we had to increase the reaction temperature to
170 °C. Under these conditions, 41% yield of monoether was
produced after 32 h of reaction. Under similar conditions, tris
dented selectivity observed in the presence of Bi(OTf)3
(Table 6). First of all, we tested the etherification of glycerol
with alkyl alcohols with a longer chain length such as n-penta- (hydroxymethyl)ethane was also converted to the desired
monobutylether with 38% yield. As observed with glycerol, pro-
nol and n-hexanol. Reactions were conducted at 150 °C in the
longed reaction time after 90% conversion of n-butanol led to
presence of 6.5 mol% of Bi(OTf)3. In accordance with our pre-
vious report,11b we observed a decrease of the catalyst activity a drop of the monoether selectivity. Although we have no
rational explanation yet, we would like to point out here that,
in the presence of Bi(OTf)3, oligomerization of 1,2-propane-
with an increase of the chain length presumably due to an
increase of the alkyl alcohol hydrophobicity making its coup-
ling with glycerol (highly polar molecule) more difficult. As a diol, 1,3-propanediol and tris(hydroxymethyl)ethane was more
favourable than from glycerol, explaining the lower yield of
monobutylethers obtained with these polyols and reinforcing
result, after 24 h of reaction at 150 °C, the yield of alkyl glyceryl
ethers dropped from 70% to 50% and then to 32% when going
from n-butanol to n-pentanol and n-hexanol, respectively. the key role that it may play in the hypothesized complex bi-
glycerol described above.
However, the carbon mass balance (based on glycerol)
remained higher than 91% further confirming the high selec-
tivity of Bi(OTf)3 in such reactions.
In the presence of n-butanol, Bi(OTf)3 was also capable of
catalyzing the etherification of 1,2-propanediol. At 150 °C,
Finally, we investigated the potential of this methodology
from sorbitol, a valuable polyol obtained after hydrogenation
of glucose (Fig. 5). To our delight, when neat sorbitol was
heated at 150 °C for 16 h in the presence of 3.6 mol% of
monobutylethers of 1,2-propanediol were obtained with 46% Bi(OTf)3, the dehydrative intramolecular cyclisation of sorbitol
to isosorbide (85% yield) selectively took place. Next, using the
same conditions to those described in Table 6, isosorbide can
yield. The regioselectivity of the reaction was 7/3 in favour of
the etherification in the primary hydroxyl group. Surprisingly,
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 901–909 | 907

View Article Online
Paper
Green Chemistry
2 For selected comprehensive reviews please see:
(a) A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
2411–2502; (b) P. Y. Dapsens, C. Mondelli and J. Perez-
Ramirez, ACS Catal., 2012, 2(7), 1487–1499; (c) P. Gallezot,
Chem. Soc. Rev., 2012, 41(4), 1538–1558.
3 P. Foley, A. Kermanshahi pour, E. S. Beach and
J. B. Zimmerman, Chem. Soc. Rev., 2012, 41, 1499–1518.
4 U. Biermann, U. Bornscheuer, M. A. R. Meier, J. O. Metzger
and H. J. Schäfer, Angew. Chem., Int. Ed., 2011, 50(17),
3854–3871.
5 A wide range of patents are available mainly from TAIYO
KAGAKU KK and DAICEL CHEM in recent years.
6 (a) A. Behr and M. Urschey, Adv. Synth. Catal., 2003,
Fig. 5 Catalytic dehydrative intramolecular cyclisation of sorbitol to isosorbide
and its subsequent etherification with n-butanol (13 mol% of Bi(OTf)3, isosor-
bide/n-butanol molar ratio = 4/1).
345,
1242–1246;
(b)
A.
Behr,
J.
Leschinski,
C. Awungacha, S. Simic and T. Knoth, ChemSusChem,
2009, 2, 71–76.
be etherified with n-butanol affording the monobutylether
of isosorbide with 44% yield which is a similar yield to
those obtained from 1,2-propanediol, 1,3-propanediol and
tris(hydroxymethyl)ethane. Note that the one pot conversion
of sorbitol to the desired monobutylether of isosorbide is
a feasible route but the yield of the reaction remained
rather low (18%) and formation of unidentified side
products was observed. Such a result might be due to
the difference of kinetics between the two steps, isosorbide
being much less stable than sorbitol in our experimental
conditions.
7 (a) R. Palkovits, I. Nieddu, R. J. M. Klein Gebbink and
B. Weckhuysen, ChemSusChem, 2008, 1, 193–196;
(b) R. Palkovits, I. Nieddu, C. A. Kruithof, R. J. M. Klein
Gebbink and B. Weckhuysen, Chem.–Eur. J., 2008, 14,
8995–9005; (c) R. Palkovits, A. N. Parvulescu,
P. J. C. Hausoul, C. A. Kruithof, R. J. M. Klein Gebbink
and B. Weckhuysen, Green Chem., 2009, 11, 1155–1160;
(d) P. Hausoul, J. C. Parvulescu, N. Andrei, M. Lutz,
A. L. Spek, P. C. A. Bruijnincx, R. J. M. Klein Gebbink
and B. M. Weckhuysen, ChemCatChem, 2011, 3(5),
845–852; (e) P. Hausoul, P. A. Bruijnincx, R. J. M. Klein
Gebbink and B. M. Weckhuysen, ChemSusChem, 2009,
2(9), 855–858.
8 (a) L. Conceicao, R. Bogel-Lukasik and E. Bogel-Lukasik,
Green Chem., 2012, 14(3), 673–681; (b) J. M. Lopes,
Z. Petrovski, R. Bogel-Lukasik and E. Bogel-Lukasik, Green
Chem., 2011, 13(8), 2013–2016; (c) S. Bigot, J. Lai, I. Suisse,
M. Sauthier, A. Mortreux and Y. Castanet, Appl. Catal., A,
2010, 382(2), 181–189; (d) A. Gordillo, L. Durán Pachón,
E. de Jesus and G. Rothenberg, Adv. Synth. Catal., 2009,
351(3), 325–330.
9 (a) Y. Shi, W. Dayoub, A. Favre-Réguillon, G.-R. Chen and
M. Lemaire, Tetrahedron Lett., 2009, 50, 6891–6893;
(b) Y. Shi, W. Dayoub, G.-R. Chen and M. Lemaire, Green
Chem., 2010, 12, 2189–2195; (c) M. Sutter, W. Dayoub,
E. Metay, Y. Raoul and M. Lemaire, ChemSusChem, 2012, 5,
2397–2409.
10 As representative articles please see: (a) K. Klepacova,
D. Mravec and M. Bajus, Chem. Pap., 2006, 60, 224;
(b) K. Klepacova, D. Mravec and M. Bajus, Appl. Catal., A,
2005, 294, 141–147; (c) Y. Gu, A. Azzouzi, Y. Pouilloux,
F. Jérôme and J. Barrault, Green Chem., 2008, 10, 164–167;
(d) R. Luque, V. Budarin, J. H. Clark and D. J. Macquarrie,
Appl. Catal., B, 2008, 82, 157–162.
11 (a) P. Gaudin, R. Jacquot, P. Marion, Y. Pouilloux and
F. Jérôme, Catal. Sci. Technol., 2011, 1, 616–620;
(b) P. Gaudin, R. Jacquot, P. Marion, Y. Pouilloux and
F. Jérôme, ChemSusChem, 2011, 4(6), 719–722;
(c) S. Pariente, N. Tanchoux and F. Fajula, Green Chem.,
2009, 11, 1256–1261.
Conclusions
We report here that Lewis acids such as metal triflates, and
especially Bi(OTf)3, are capable of catalyzing the etherification
of glycerol with n-butanol with an unprecedented selectivity.
Under optimized conditions, up to 70% yield of B1G1
was obtained while the carbon mass balance determined
on glycerol and n-butanol remained higher than 85%. By com-
parison, all tested Brønsted or Lewis acids are less active and
less selective, showing the attractiveness of metal triflates in
such reactions. Although elucidation of the reaction mechan-
ism clearly deserves deeper investigations, we show here, by
means of different counter experiments, that the surprising
selectivity observed with Bi(OTf)3 may rely on a synergistic
effect between Bi(OTf)3 and triflic acid that may be released by
glycerolysis of Bi(OTf)3. This aspect is the topic of current
investigations in our group and should be published in a
couple of months.
Notes and references
1 Published by Transparency Market Research (http://www.
transparencymarketresearch.com) “Specialty Surfactants
Market – Global Scenario, Raw Material and Consumption
Trends, Industry Analysis, Size, Share
2011–2017”.
&
Forecast
908 | Green Chem., 2013, 15, 901–909
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
12 As a selected review please see: M. E. Zakrzewska, E. Bogel-
Lukasik and R. Bogel-Lukasik, Chem. Rev., 2011, 111(2),
397–417.
13 S. Antoniotti, V. Dalla and E. Dunach, Angew. Chem., Int.
Ed., 2010, 49, 7860–7888.
14 (a) S. Kobayashi and K. Manebe, Acc. Chem. Res., 2002, 35,
209–217; (b) S. Kobayashi and C. Ogawa, Chem.–Eur. J.,
2006, 12, 5954–5960; (c) S. Kobayashi, M. Ueno and
T. Kitanosono, Top. Curr. Chem., 2012, 311, 1–17.
15 J. Chaminand, L. Djakovitch, P. Gallezot, P. Marion,
C. Pinel and C. Rosier, Green Chem., 2004, 6,
359–361.
9269–9277; (b) T. C. Wabnitz, J.-Q. Yu and J. B. Spencer,
Chem.–Eur. J., 2004, 10, 484–493.
17 (a) O. Mouhtady, H. Gaspard-Iloughmane, N. Roques and
C. Le Roux, Tetrahedron Lett., 2003, 44, 6379–6382;
(b) M. Peyronneau, M. Boisdon, N. Roques, S. Mazieres and
C. Le Roux, Eur. J. Org. Chem., 2004, 4636–4640;
(c) J. M. Bothwell, S. W. Krabbez and R. S. Mohan, Chem.
Soc. Rev., 2011, 40, 4649–4707.
18 (a) Q. Zhang, K. De Oliveira Vigier, S. Royer and F. Jérôme,
Chem. Soc. Rev., 2012, 41, 7108–7146; (b) K. De Oliveira
Vigier, A. Benguerba, J. Barrault and F. Jérôme, Green
Chem., 2012, 14, 285–289.
16 (a) R. F. Lambert, R. J. Hinkle, S. E. Ammann, Y. Lian, 19 J. Barrault, Y. Pouilloux, J.-M. Clacens, C. Vanhove and
J. Liu, S. E. Lewis and R. D. Pike, J. Org. Chem., 2011, 76,
S. Bancquart, Catal. Today, 2002, 75, 177–181.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 901–909 | 909