Glycerol acetals and ketals as bio-based solvents: Positioning in Hansen and COSMO-RS spaces, volatility and stability towards hydrolysis and autoxidation

Article abstract of DOI:10.1039/c4gc02377c

Four recently launched cyclic glycerol acetals or ketals are evaluated as bio-based solvents. Three of them are industrially available and result from the condensation of glycerol with formaldehyde, acetone and isobutyl methyl ketone. The fourth is under development and is prepared by the reaction of glycerol with benzaldehyde under heterogeneous acidic catalysis. Their solvent properties are evaluated through Hansen and COSMO-RS (COnductor-like Screening MOdel for Real Solvents) approaches, in comparison with traditional petrochemical solvents. Dioxolane- and dioxane-type isomers have close solubility parameters; however the nature of the starting aldehyde/ketone significantly impacts the solvency properties. The stability to hydrolysis depends heavily on both the aldehyde/ketone part and on the size of the ring. In acidic medium, acetals are found to be more stable than ketals and glycerol-based ketals are more stable than ethylene glycol-based ketals. In the case of benzaldehyde glycerol acetal, it is shown that the 6-membered ring isomer (dioxane-type) is approximately 8 times more stable than the 5-membered ring counterpart (dioxolane-type) at low pH. Stability towards autoxidation by O₂ is high for formaldehyde and acetone-derived acetals and drops for the other two compounds. Glycerol acetals and ketals are promising potential alternatives to some harmful solvents such as glycol ethers and aniline. This journal is

Full text of DOI:10.1039/c4gc02377c

Green Chemistry
View Article Online
View Journal
PAPER
Glycerol acetals and ketals as bio-based solvents:
positioning in Hansen and COSMO-RS spaces,
volatility and stability towards hydrolysis and
autoxidation†
Cite this: DOI: 10.1039/c4gc02377c
a
a
a
a
Laurianne Moity, Adrien Benazzouz, Valérie Molinier, Véronique Nardello-Rataj,
b,c
b,c
b,c
Mohammed Kamal Elmkaddem, Pascale de Caro, Sophie Thiébaud-Roux,
Vincent Gerbaud, Philippe Marion and Jean-Marie Aubry*
d
e
a
Four recently launched cyclic glycerol acetals or ketals are evaluated as bio-based solvents. Three of
them are industrially available and result from the condensation of glycerol with formaldehyde, acetone
and isobutyl methyl ketone. The fourth is under development and is prepared by the reaction of glycerol
with benzaldehyde under heterogeneous acidic catalysis. Their solvent properties are evaluated through
Hansen and COSMO-RS (COnductor-like Screening MOdel for Real Solvents) approaches, in comparison
with traditional petrochemical solvents. Dioxolane- and dioxane-type isomers have close solubility para-
meters; however the nature of the starting aldehyde/ketone significantly impacts the solvency properties.
The stability to hydrolysis depends heavily on both the aldehyde/ketone part and on the size of the ring.
In acidic medium, acetals are found to be more stable than ketals and glycerol-based ketals are more
stable than ethylene glycol-based ketals. In the case of benzaldehyde glycerol acetal, it is shown that the
6-membered ring isomer (dioxane-type) is approximately 8 times more stable than the 5-membered ring
Received 3rd December 2014,
Accepted 23rd December 2014
2
counterpart (dioxolane-type) at low pH. Stability towards autoxidation by O is high for formaldehyde and
DOI: 10.1039/c4gc02377c
acetone-derived acetals and drops for the other two compounds. Glycerol acetals and ketals are promis-
ing potential alternatives to some harmful solvents such as glycol ethers and aniline.
www.rsc.org/greenchem
solvents” are a class of commodity chemicals that already
1
. Introduction
1
–3
represent a mature market.
They are obtained from the
4
–8
The chemical industry is increasingly turning to so-called “bio- biomass feedstock
and experience an increased demand
based” solutions, as a remedy to dwindling fossil resources due to the regulatory constraints that restrict the use of a large
and greenhouse gas emissions, and as part of its commitment number of traditional organic solvents, such as halogenated,
to “sustainable development”. The so-called “bio-based aromatic or ethylene glycol ether derivatives.
As a by-product of biodiesel manufacture, glycerol is one of
9
–12
the top-ranking bio-building blocks
and has received
a
Université Lille 1, ENSCL, E.A. 4478 Chimie Moléculaire et Formulation,
considerable attention. An appreciable number of chemical
intermediates and commodity chemicals can now be obtained
Cité Scientifique, 59652 Villeneuve d’Ascq Cedex, France.
E-mail: Jean-Marie.Aubry@univ-lille1.fr
1
3–19
b
from glycerol through optimised catalytic routes.
Apart
Université de Toulouse, INP, LCA (Laboratoire de Chimie Agro-Industrielle),
from glycerol itself, several classes of bio-based solvents are
ENSIACET, 4 Allée Emile Monso, 31030 Toulouse, France
c
20–24
INRA, UMR 1010 CAI, 31030 Toulouse, France
currently obtained from glycerol
(Fig. 1).
d
Université de Toulouse, INP, CNRS, LGC (Laboratoire de Génie Chimique),
Hydrogenolysis of glycerol gives rise to propanediols and
ethylene glycol that are common solvents and for which an
efficient bio-derived route is competitive to the traditional petro-
ENSIACET, 4 Allée Emile Monso, 31432 Toulouse, France
e
Solvay Recherches & Innovation, Centre de Lyon, 85 rue des Frères Perret,
69192 Saint-Fons Cedex, France
†
Electronic supplementary information (ESI) available: Physicochemical pro- chemical route. Glycerol triacetate (triacetin) is obtained by
perties of acetals/ketals and various oxygenated solvents, details and discussion glycerol esterification and is used as a solvent and an additive
about the synthesis of benzaldehyde glycerol acetal 4, Hansen solubility para-
in various fields, including food and pharmaceutical formu-
lations. Glycerol carbonate is prepared by carbonation or trans-
carbonation of glycerol and is on the market for cosmetic,
personal care and medicinal applications due to its low toxicity
meters and COSMO-RS coordinates of 88 traditional solvents of Hansen’s
5
0
work, as well as the ranking of the 88 solvents with regard to their proximity to
acetals 1–4 according to Hansen’s approach and COSMO-RS. See DOI: 10.1039/
c4gc02377c
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
for the conversion of benzaldehyde with glycerol. The selectiv-
ities obtained towards 5- and 6-membered ring derivatives
(
1,3-dioxolane and 1,3-dioxane types) depend on the experi-
mental conditions. Recently, Gonzalez-Arellano et al. described
the preparation of glycerol acetals using an aluminium
substituted mesoporous material Al-SBA-15 in high yields and
tuneable selectivities towards 5- and 6-membered ring isomers
4
9
depending on the starting aldehyde/ketone.
In this work, we intend to compare four cyclic glycerol
acetals and ketals 1–4, recently launched on the industrial
scale, with petrochemical solvents having acetal or ether func-
tions 5–19. The solvent properties are evaluated through the
traditional Hansen’s approach and also via the theoretical
COSMO-RS model. Particular attention is paid to their vola-
tility, oxidability and stability towards hydrolysis in aqueous
medium since they are key parameters to be considered in the
selection of a suitable solvent. The four cyclic glycerol acetals
or ketals selected are glycerol formal 1a + 1b, 1,2-isopropyli-
dene glycerol (solketal) 2, 1,2-methylisobutylidene glycerol 3
and benzaldehyde glycerol acetal (BGA) 4a + 4b (see Table 1).
Acetals 1 and 4 are mixtures of 5- and 6-membered ring deriva-
tives (see the Experimental section). Acetal 4 was prepared
by condensation of glycerol with benzaldehyde using a macro-
porous ion exchange resin.
Fig. 1 Families of bio-based solvents derived from glycerol.
and low flammability. Finally, the cyclic acetals (respectively
ketals) are synthesised by the condensation of two alcohol
2
. Results and discussion
functions of glycerol with an aldehyde (respectively a ketone). For the sake of clarity, the synthesis of benzaldehyde glycerol
The final compound can be fully bio-based if the aldehyde or acetal 4a/4b is given in the ESI.† The aim of the present paper
ketone used is bio-sourced (for example acetaldehyde, acetone is indeed not to discuss the synthesis, all the more that a
and benzaldehyde).
recent paper describes in detail the preparation of acetals and
49
The acetals/ketals of glycerol have only recently been classi- ketals investigated in the present work. We rather focus on
fied as low toxicity bio-based solvents, even though they have the potentialities of compounds 1–4 as bio-based solvents and
been known for a long time as chemical intermediates. In par- therefore study their relevant physical and chemical pro-
ticular, 1,2-isopropylidene glycerol (solketal) is commonly perties. The solvent properties per se are evaluated through the
used in organic synthesis to introduce the protected form of traditional Hansen’s parameters and the more modern
the glycerol moiety in a molecular scaffold. Its solvent pro- COSMO-RS approach. Special interest was also paid to volatility
perties have been recently highlighted, since it is now manu- that often triggers the choice of a solvent. Finally, stabilities
25
factured on an industrial scale at a competitive cost. Several towards hydrolysis and autoxidation are examined and com-
patents describe the use of acetals/ketals of glycerol as sol- pared with other relevant solvents 5–20 (Table 1) since these
vents, coalescing agents in paints or lacquers or as coupling compounds are prone to cleave in acidic solutions and to
2
6–31
agents.
prohibited (e.g. 2-methoxyethanol) or controversial (e.g.
2
In some formulations, they advantageously replace suffer free radical chain oxidation with oxygen.
2.1. Solvent properties
-butoxyethanol) short ethylene glycol ethers. Glycerol acetals
and ketals are also produced as bio-additive fuels
3
2–35
to Glycerol acetals/ketals 1–4 are first considered using the tra-
decrease emissions of carbon monoxide, hydrocarbons and ditional Hansen’s approach. In this semi-empirical theory, the
3
2,36
unregulated aldehydes and as cold flow improvers.
solubilising properties of solvents are characterized by three
These glycerol derivatives are usually prepared by conden- parameters corresponding to dispersion δ , polar δ and
d
p
sation of glycerol with carbonyl compounds, catalysed by hydrogen bonding δ interactions, which build a three-dimen-
h
3
7–39
50,53,54
homogeneous catalysts such as gold or iridium complexes.
sional space known as the “Hansen space”.
ion exchange that are located close to each other in this space are expected
and sulfonic mesostructured to exhibit similar solubilising behaviours. This simple
silicates are also reported to be active for acetalisation or approach has a broad acceptance in the industrial community.
Solvents
4
0–42
Heterogeneous catalysts such as zeolites,
4
3,44
45,46
resins,
montmorillonites
4
7
4
3
ketalisation. Some of these catalysts (Amberlyst-36, metal It allows identifying in a rather reliable way the good solvents
3
3
48
54
oxides or silica supported catalyst ) proved to be efficient for a given solute thanks to the chemical rule of similarity,
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Table 1 Number, name and structure of the glycerol acetals/ketals 1–4 and of the various solvents studied in this work for comparison with regard
to volatility, hydrolysis and oxidability
No.
Name
Structure
No.
Name
Structure
1a
Glycerol formal (dioxolane)
10
Methylal
1
b
Glycerol formal (dioxane)
Solketal
11
12
Ethylal
2
3
4
n-Butyl acetate
1,2-Methylisobutylidene glycerol
13
14
4 1
Ethylene glycol n-butyl ether (C E )
a
Benzaldehyde glycerol acetal (dioxolane)
4 1
Propylene glycol n-butyl ether (C P )
4b
Benzaldehyde glycerol acetal (dioxane)
15
2 2
Diethylene glycol ethyl ether (C E )
5
6
1,3-Dioxolane
16
17
4 2
Diethylene glycol butyl ether (C E )
2-Methyl-1,3-dioxolane
Diethylene glycol hexyl ether (C
Tetrahydrofurfuryl alcohol
Dimethyl isosorbide
6 2
E )
7
8
2,2-Dimethyl-1,3-dioxolane
2,2,4-Trimethyl-1,3-dioxolane
18
19
9
1,4-Dioxane
20
Limonene
provided that the solubility parameters of the solute and
In parallel, acetals/ketals 1–4 have been modelled using the
solvents are known. The Hansen solubility parameters of COSMO-RS approach. We have recently used this purely predic-
the four glycerol acetals/ketals 1–4 and other relevant tive method based on quantum mechanics and statistical
petro-based acetals/ketals/ethers solvents 5–11 have been com- thermodynamics to rebuild an a priori classification of organic
5
1
puted using the recommended Y-MB method (ESI†). Fig. 2 solvents. Based on the knowledge of the chemical structure
shows the positioning of glycerol acetals/ketals 1–4 in the and without resorting to any experimental data, 10 consistent
δ /δ projection of the Hansen’s space, among classical sol- solvent families were identified. This classification was then
p
h
vents 5–11. The relative sizes of the dots account for the δd used to establish a panorama of currently available “green
7
values.
solvents”. This methodology gives access to a so-called
As expected, glycerol acetals/ketals 1–4 (green dots) have “COSMO-RS 4-dimensional space” in which each solvent is
significantly higher hydrogen bonding parameters than other identified by four coordinates F1, F2, F3 and F4 resulting from
acetalic solvents 5–8, 10,11 (red dots) due to the pending a PCA analysis. This space can be viewed as a purely predictive
hydroxyl group. The aldehyde/ketone from which they have and more consistent alternative to the 3-dimensional semi-
been prepared also has an impact on their position in the empirical “Hansen space”. The coordinates of acetals/ketals
Hansen space. Since dioxolane and dioxane isomers 1a/1b and 1–4 in the COSMO-RS space were calculated according to a pro-
7
,51
4
a/4b have almost the same Hansen parameters, each pair of cedure previously described
(see ESI†). The COSMO-RS
isomers is represented by a single dot.
approach allows positioning of solvents 1–4 in the ten classes
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
p h d
Fig. 2 Location of the four glycerol acetals/ketals 1–4 within the δ /δ projection of the Hansen space. The sizes of the dots express the δ value of
the solvent. The four solvents 1–4 are represented in green while classical solvents are in grey and close neighbours are in black. Other acetalic sol-
50
vents 5–8, 10,11 are also positioned for comparison (red dots). When not listed in Hansen’s list, solubility parameters were calculated using the
Yamamoto Molecular Break method (Y-MB).
Table 2 Closest neighbours of glycerol acetals/ketals 1–4 in the Hansen’s space (H) and in the COSMO-RS space (C). Solvents taken as references
50
are the 88 classical solvents used by Hansen to determine their solubility parameters. The ranking and coordinates of all solvents can be found in
ESI
Ranking of closest classical solvents from acetals/ketals 1–4
1a + 1b
2
3
4a + 4b
Classical solvents
H
C
H
C
H
C
H
C
Dipropylene glycol
Aniline
Diethylene glycol monomethyl ether
Butyl lactate
Tetrahydrofuran
Morpholine
1
(17)
1
—
—
—
—
—
—
1
(6)
—
—
—
—
(15)
1
—
—
—
—
—
(2)
1
—
—
—
1
(66)
—
—
(3)
—
—
—
1
—
1
—
—
—
(79)
(22)
—
—
—
—
—
established previously: solketal 2 and 1,2-methylisobutylidene the closest points, the one which has a similar diameter
glycerol 3 both belong to amphiprotic solvents, just as most (similar δ ). The nearest neighbours of 1–4 are reported in
d
alcohols, whereas glycerol formal 1 is positioned in the polar Table 2, while the ranking obtained by the other method is
protic family among short alcohols and polyols, which indicated in parentheses. For 1 and 4, only the neighbours of a
evidences its higher tendency to establish hydrogen bonds. mixture of isomers 1a/1b (34 : 66) and 4a/4b (66 : 34) are shown
Benzaldehyde glycerol acetal 4 is found at the border of the since dioxolane and dioxane isomers exhibit almost similar
amphiprotic family and the polar protic family which contains coordinates, with only slight differences in the ranking (see
furfuryl alcohol, 2-aminoethanol as well as diethylene glycol.
Both approaches can be used to identify the closest classi-
ESI†).
For several solvents, the rankings obtained by Hansen and
cal solvents by calculating the distance D separating them COSMO-RS methods are consistent. This is the case for butyl
from each glycerol acetal/ketal 1–4 in the 3D-Hansen space or lactate which is found close to 2 and 3 as well as for aniline
in the 4D-COSMO-RS space (see the Experimental part). which is close to 4 whatever the method. Both methods ident-
Solvents taken as references are the 88 conventional solvents ify several short chain glycol ethers as close neighbours of
5
0
used by Hansen to determine solubility parameters.
solketal 2, which means that this compound might be a safe
A simple method to visually locate the nearest neighbour of a alternative to this family of reprotoxic solvents. Actually, 2 is
given solvent in the Hansen space (Fig. 2) is to search among already used to substitute 2-methoxyethanol in aqueous
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Fig. 3 σ-Potentials computed with COSMO-RS for the glycerol acetals/ketals 1–4 (green curves) and their closest neighbours according to Hansen
black curves) or COSMO-RS (red curves). (a) Glycerol formal 1 (green), dipropylene glycol (black) and aniline (red). (b) Solketal 2 (green), diethylene
glycol monomethyl ether (black) and butyl lactate (red). (c) 1,2-Methylisobutylidene glycerol 3 (green), tetrahydrofuran (black) and butyl lactate (red).
d) Benzaldehyde glycerol acetal 4 (green), morpholine (black) and aniline (red).
(
(
2
8
formulations as a coupling agent. Surprisingly, several aro- branch is oriented upward. The more the solvent is basic, the
matic conventional solvents are encountered in the vicinity of more inclined downwardly the branch is. The interpretation of
1
–4 according to the COSMO-RS method. For instance, aniline the right branch is similar to the left one except that, in this
is the closest solvent of glycerol formal 1 and benzaldehyde case, the Lewis acidity is considered.
glycerol acetal 4. Presumably, the strong electronic density in
Fig. 3a shows that the σ-potential curves of glycerol formal
the acetalic cycle due to the doublets of the oxygen atoms com- 1 and its nearest neighbours obtained by Hansen and
pensates the lack of aromatic electrons of 1–3. In contrast, two COSMO-RS approaches are quite similar. They all have a poor
solvents, namely THF and morpholine are very close to 3 and 4 acidity and a medium basicity including aniline whose elec-
respectively in Hansen’s space whereas they are far apart in the tron pair localized on the nitrogen atom is much less basic
COSMO-RS space.
than that of the aliphatic amine because it is delocalized on
= 4.6 compared to pK = 10.8 for NEt ).
To shed some light on the origin of these discrepancies, we the aromatic ring (pK
a
a
3
drew in Fig. 3a–d, the σ-potential curves of glycerol acetals/ It is noteworthy that COSMO-RS suitably captures the negli-
ketals 1–4 (green curves) and of their closest solvents deter- gible acidity of the three solvents and the similarity between 1
mined either by the Hansen approach (black curves) or and aniline, what could hardly be expected by comparing the
COSMO-RS (red curves).
chemical formulas. In the case of solketal 2, Fig. 2 and 3b
A σ-potential curve finely summarizes all the information show that both approaches agree to predict that it should
necessary to apprehend the solubilising properties of the present solubilising properties similar to those of butyl lactate
5
1,54
corresponding solvent.
Here we only recall the two main and diethylene glycol monomethyl ether. Ketal 3 is predicted
information that can be inferred from a simple observation of to be close to tetrahydrofuran and butyl lactate according to
the shape of the curves. The left branch reflects the Lewis basi- the Hansen approach and COSMO-RS respectively. Looking at
city of the solvent, i.e. the presence of electron-rich sites. For the chemical formulas, one can easily infer that 3 and butyl
solvents that have no basic sites such as alkanes, the left lactate should present similar solubilising properties. On the
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
other hand, it is clear that THF and 3 must be significantly together with the one of dimethyl isosorbide 19 (DMI), a bio-
different due to the lack of alcohol function in the formula of based solvent that is located close to the VOC limit (b.p. =
5
6
THF. This hydroxyl group allows 3 to establish hydrogen bonds 236 °C and Pvap = 4.1 Pa at 25 °C). For a quantitative compari-
with basic solutes unlike THF. This difference appears clearly son of solvents volatility, the temperatures at which 50% of the
on the shape of the right branches of σ-potential curves in compound is vaporised (T50%) have been collected. They are
Fig. 3c. That of THF is directed upwards, in agreement with its compared with various well-known petrochemical or bio-based
lack of acidity, while the other two are curved slightly down- oxygenated solvents, in particular glycol ethers (Table 1),
ward. The closest neighbours of 4 are morpholine and aniline studied under the same experimental conditions. Incidentally,
according to the DHansen and DCOSMO criteria respectively. it is worth noting that the boiling points of the well-known
Looking at Fig. 3d, it appears that the σ-potential curve of mor- oxygenated solvents are strongly correlated with T50% values
pholine is very different from that of 4 especially at the level of (Fig. 5), which allows an estimation of the b.p. when experi-
the left branch which is much steeper downward. This reflects mental values are not available (see ESI†).
the higher basicity of this cyclic secondary amine (pK
a
= 8.4)
Among the four acetals/ketals under study, solketal 2 is
compared to that of aniline (pK = 4.6). It is clear that the basi- unexpectedly more volatile than glycerol formal 1, itself more
a
city of morpholine is too high to mimic the poor basicity of volatile than ketal 3. Acetal 4 derived from benzaldehyde is
acetal 4 coming from the electron pairs of oxygen atoms.
much less volatile than the others and is the only non-VOC gly-
cerol derivative. This trend generally follows the molecular
weights of the solvents. However the inversion between com-
2
.2. Volatility
Volatility is an important issue for solvents since it determines pounds 1 and 2 indicates that the structuration in the liquid
their use, recycling and impact on the environment. The direc- state is stronger for glycerol formal 1 than for solketal 2. Also,
tives in force aim at decreasing VOC (volatile organic the low volatility of 4 reflects the strong interactions in the
compound) emissions. Volatility is also a key criterion in appli- liquid state via π–π interactions due to the presence of an aro-
cations and often triggers the choice of solvents. It is thus of matic ring. It is also noteworthy that, though they have the
great interest to evaluate the volatility in comparison with tra- same number of carbon and oxygen atoms, 2 is more volatile
ditional solvents. Volatility depends on various factors, includ- than diethylene glycol monoethyl ether (C
2 2
E ). In the same
ing vapour pressure, vaporisation enthalpy, surface tension way, 3 is also more volatile than diethylene glycol hexyl ether
5
5
and molecular weight. The volatility of the glycerol acetals/ (C
ketals 1 to 4 has been evaluated by thermogravimetric analysis This indicates that a cyclic acetal function tends to increase
TGA) by measuring the weight loss when submitting each the volatility. This result is of particular interest in applications
sample to a temperature increase from room temperature to since we have shown in the previous section that glycerol
00 °C under an inert atmosphere. The thermogravimetric pro- acetals/ketals are relevant candidates for the replacement of
6 2
E ) although the latter has one additional carbon atom.
(
3
files for the four acetals/ketals 1 to 4 are presented in Fig. 4 glycol ethers.
Fig. 4 Thermogravimetric profiles of glycerol acetals/ketals 1–4 (solid lines) and dimethyl isosorbide 19 (dashed line) used as a reference com-
pound to separate VOC from non-VOC solvents.
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
Fig. 5 Correlation between T50% and boiling points of glycerol acetals and ketals 1–4 (green dots), traditional acetals/ketals/ethers 5–11 (blue dots)
and other oxygenated solvents 12–19 (black dots). The boiling point of 3 predicted from the experimental T50% value is indicated as a green empty
dot. Dimethyl isosorbide 19 indicates the limit between VOC and non-VOC compounds and is highlighted as a bigger dot.
Scheme 1 Mechanism of acetals/ketals hydrolysis in acidic medium.
2
.3. Hydrolysis in acidic medium
In order to study the structural effects on the rate of hydrolysis
of cyclic acetals/ketals, four petroleum-based cyclic acetals/
ketals 5–8 have been studied together with the glycerol deriva-
tives 1–4. The mechanism of acetal or ketal hydrolysis into
aldehydes or ketones and alcohols in acidic medium is well
5
7
known. It implies two C–O bond cleavages, addition of a
water molecule and proton transfer (Scheme 1). The rate-deter-
+
mining step is the carbocation (B ) formation and hydrolysis is
favoured when the carbocation is stabilised.
1
Hydrolysis of compounds 1–4 was monitored by H NMR at
various pH values ranging from 1.1 to 12.7 at 25 °C. Fig. 6
shows the percentage of hydrolysed solketal 2 versus time at
four pH values (1.7; 2.5; 3.5 and 4.7). Below pH 1.3, hydrolysis
is too rapid (less than 10 min) whereas above pH 6, no hydro-
lysis takes place over a two-month period. Similar behaviours
Fig. 6 pH effect on the rate of hydrolysis of solketal 2 determined by
H NMR at 25 ± 1 °C. The plotted lines correspond to hydrolysis kinetic
1
were obtained for compounds 5–8. Glycerol formal 1 and modelling (pH 1.7▲; pH 2.5 ●; pH 3.5 ■; pH 4.7 ◆).
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
1
,3-dioxolane 5 are very stable even below pH 1.7. 3 and 4
are poorly water-soluble and have been considered separately
(see below).
Acetal and ketal hydrolysis is modelled by eqn (1), where
+
−1
[
A], [E] and [H O ] are the molar concentrations (mol L ) of
3
Scheme 2 Cleavage of 1,3-dioxolanes under acidic conditions.
acetal/ketal, aldehyde/ketone, and hydronium respectively. k is
the rate constant of hydrolysis and x and y are the partial
+
orders of the reaction with respect to A and H O respectively.
3
d½Aꢁ
dt
d½Eꢁ
dt
x
þ y
v ¼ ꢀ
¼ þ
¼ k½Aꢁ ½H3O ꢁ
ð1Þ
kapp is the apparent rate constant when the pH is kept constant
over the reaction time, which is the case here. It is defined
by eqn (2).
kapp ¼ k½H3Oþꢁy
ð2Þ
For all compounds, fitting of the experimental points show
that x and y equal 1 and the integration of eqn (1) gives
the evolution of the concentration of acetal/ketal over time
(eqn (3)).
þ
½
Aꢁ ¼ ½Aꢁ expðꢀk½H3O ꢁtÞ ¼ ½Aꢁ expðꢀkapptÞ
ð3Þ
t
0
0
Fig. 7 Intramolecular hydrogen bond can be established in the case of
The half-life time t1/2 is defined as the time at which half of solketal 2.
the acetal/ketal is hydrolysed (eqn (4)).
ln 2
kapp
ln 2
k ꢂ 10ꢀpH
t1=2
¼
¼
ð4Þ
57
the ratio generally varying from 2.3 to 3.4. This kinetic isoto-
pic effect has been confirmed for solketal 2 at pH 1.8.
Let us now first consider the substituent effect on position
(R /R groups in Scheme 1). Formaldehyde-based acetals
3 4
The hydrolysis rate constants k and kapp and the half-life time
t1/2 are reported in Table 3.
2
+
As the dissociation constants of protonated acetal/ketal A
4 3
(R = R = H, compounds 1 and 5) were found to be stable
+
and hemiacetal/hemiketal C (see Scheme 1) are approximately
under our conditions, whereas acetone-based ketals (com-
pounds 2, 7 and 8) are hydrolysed much faster. Replacing a
methyl group by a hydrogen atom (compound 6) significantly
increases the stability under these acidic conditions since the
three times greater in D O than in H O, hydrolysis is expected
to be three times faster in heavy water than in normal water,
2
2
constant is divided by 16. The more electron-donating R /R
4
CN at various are, the less stable the compound is. This is consistent
Table 3 Half-life times t1/2 and hydrolysis rate constants kapp and k (=
3
+
k
app/[H
3
O ]) for ketals/acetals 1–8 in D
2
O or in D
2
O–CD
3
pH values
with the fact that the rate determining step (r.d.s.) is the carbo-
cation formation in position 2 (step 2 in Scheme 2).
However, even though ketals 2, 7 and 8 bear the same R /R
3 4
5
7
Solvent
O/CD
t
1/2
k
app (min−1
(M min−1
−1
No.
D
2
3
CN
pH
k
)
(min)
)
groups, they exhibit different hydrolysis stabilities which
1
1
2
2
2
2
2
3
4
4
5
6
6
6
7
7
7
8
8
8
a
b
100/0
100/0
100/0
100/0
100/0
83/17
67/33
83/17
67/33
67/33
100/0
100/0
100/0
100/0
100/0
100/0
100/0
100/0
100/0
100/0
1.8–3.5
1.8–3.5
1.8
2.5
3.5
1.9
2.3
1.9
2.3
2.3
1.8–3.5
1.8
2.5
3.5
1.8
2.5
3.5
1.8
≈0
—
—
25
98
1281
34
—
17
30
239
—
59
238
2876
5
22
338
6
27
301
≈0
≈0
1.8
2.2
1.7
1.6
1.4
3.2
4.6
0.6
≈0
0.7
0.9
0.8
8.5
9.9
6.5
7.4
8.1
7.3
implies that the substituents in position 4 or 5 (R /R ) also
1
2
≈0
−
2
3
4
2
3
2
2
3
have an impact on hydrolysis, although much weaker than the
2.81 × 10
7.07 × 10−
R /R substituents. Ketals 7 and 8 exhibit hydrolysis rate con-
3
4
−
5.41 × 10
stants of the same order of magnitude. This result is rather
surprising because one could expect that the methyl group in
position 4 (ketal 8) would favour the protonation of the oxygen
in position 3 (step 1, Scheme 2) and thus decrease its stability
towards hydrolysis. Solketal 2 is more stable than ketals 7 and
8, which could be explained by an intramolecular hydrogen
bond (see Fig. 7) or conformational effects that could make the
oxygen atom less electron rich and thus disfavour the proto-
nation step (step 1, Scheme 2). Even if one can argue that this
intramolecular hydrogen bond established in the pure state is
likely to be disrupted or even annihilated in water, in the case
of sugars (sucrose for instance), NMR and molecular dynamic
2.03 × 10−
−
7.20 × 10
3.98 × 10−
−
a
b
2.30 × 10
2.90 × 10−
≈0
1.17 × 10−
2.91 × 10−
2.41 × 10−
2
3
4
1
2
3
1
2
3
−
1.35 × 10
3.12 × 10−
−
2.05 × 10
1.17 × 10−
−
2.5
3.5
2.55 × 10
2.30 × 10−
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
•
•
studies showed that such bonds are maintained in polar
aprotic solvents and are also encountered in water. In this
case, the direct intramolecular hydrogen bonds are replaced by
indirect ones involving a water molecule. This kind of intra-
molecular contribution has already been put forward to rationa-
lize the rate of hydrolysis of comparable acetals bearing a
ROO þ RH ! ROOH þ R
ð7Þ
Termination
5
8
•
•
R þ R ! R ꢀ R
ð8Þ
ð9Þ
•
•
R þ ROO ! ROOR
2
COOH function at the C4 instead of the CH OH group present
5
9
•
•
in the glycerol acetals.
ROO þ ROO ! non radical products
ð10Þ
As compounds 3 and 4 are weakly water-soluble, their
hydrolysis has been studied in water–acetonitrile mixtures.
D O–CD CN (83 : 17) has been used for compound 3 and
For ethers, the initiation step involves the abstraction of
one of the α-hydrogen atoms giving a radical R which reacts
very rapidly with O (eqn (6)). The rate determining step corres-
ponds to the abstraction of a hydrogen atom from the sub-
strate by the peroxyl radical ROO in the propagation step
eqn (7)). The hydroperoxides formed as primary oxidation pro-
ducts can further decompose into other radicals or stable sec-
ondary oxidation products such as ketones or esters. In the
termination step, two radicals R or ROO may react providing
no-radical products (eqn (8)–(10)). Various hydrogen atoms of
a given acetal/ketal can be abstracted. In particular, the hydro-
gen atom located between the two ether functions of acetals is
•
2
3
hydrolysis was monitored at pH 1.9. Again, hydrolysis rates
have been modelled by eqn (1) and partial order x appears to
be 1. Rate constants k and kapp and half-life time t1/2 are given
in Table 2. Ketal 3 is hydrolysed three times faster than solke-
tal 2 under the same conditions, in agreement with the con-
clusions drawn above. Finally, hydrolysis of acetal 4 and ketal
2
•
(
6
2
•
•
2
4
has been monitored in D
is composed of the two isomers with five- and six-membered
2 3
O–CD CN (67 : 33) at pH 2.3. Acetal
rings (4a and 4b respectively) and the hydrolysis of these two
1
compounds could be monitored separately by H NMR. The
6
1
more reactive than those of common ethers. To evaluate the
susceptibility of the studied ketals and acetals towards oxi-
dation, we compared them with two relevant reference sol-
vents. The first one, d-limonene 20, is chemically different
from solvents 1–4 but it is a readily available bio-solvent widely
used in cleaning products and it is known to be air-sensitive
giving a mixture of highly allergenic hydroperoxides. The
second one, dimethyl isosorbide 19 (DMI), is a non-toxic cyclic
ether derived from sorbitol which is approved for use in cos-
metics and pharmaceuticals. The stability of the acetals/ketals
hydrolysis of the aromatic dioxolane 4a is 3.2 times faster than
that of solketal 2 (see Table 2). This behaviour is consistent
with the fact that the intermediate carbocation B (see
+
Scheme 1) is much more stabilised by the mesomeric effect of
the phenyl group than it is by the inductive effect of the
methyl group. More unexpectedly, dioxolane 4b hydrolyses 8
times faster than dioxane 4a. This reactivity difference can be
explained by a ring strain that is more important in the case of
a five-membered ring than in the case of a six-membered ring.
Similar conclusion has been drawn by Watts who compared
the hydrolysis of 2-methyl-1,3-dioxane and 2-methyl-1,3-
6
3
1
–4 towards autoxidation was thus assessed and compared
6
0
with that of the two reference solvents using the PetroOxy
apparatus which allows to monitor the consumption of oxygen
dioxolane.
This comparative study allows drawing the following con-
clusions on the structural effects with regard to the hydrolysis
of acetals/ketals in acidic medium. In the case of alkyl deriva-
tives (i) acetals are much more stable than ketal homologues
and formaldehyde-based acetals are so stable that no detect-
able hydrolysis is observed at pH 1.8 within two months, (ii)
the more electron-donating the substituent in position 2 is,
the less stable the ketal (iii) glycerol-based dioxolanes are
more stable than glycol-based ones.
5
2
as a function of time at constant temperature (80 °C). The
evolution of oxygen pressure, initially fixed at 400 kPa, is
reported in Fig. 8. The loss of oxygen pressure is directly
related to the consumed oxygen concentration (Δ[O
ing to the following relation:
2
]) accord-
5
2
ðPmax ꢀ PtÞ ðVtot ꢀ VÞ
3
Δ½O2ꢁt ¼ 10
ꢂ
ð11Þ
RT
V
2 t
where Δ[O ] is the consumed oxygen concentration, Pmax and
2
.4. Stability towards autoxidation
Pt are respectively the maximal oxygen pressure and the oxygen
pressure at time t measured by the PetroOxy, Vtot and V are
respectively the volume of the cell and the volume of the
solvent under study in mL.
Simple ethers are known to be prone to autoxidation upon pro-
longed exposure to air giving potentially explosive peroxides.
Like linear or cyclic ethers, acetals and ketals (R) are oxidized
with dioxygen according to a radical chain mechanism follow-
ing three main steps described as initiation, propagation and
6
1
Fig. 8 highlights different behaviours of the solvents
towards oxidation. First, glycerol formal 1 and solketal 2 are
stable under these conditions since there is no oxygen con-
sumption within 4 hours. In contrast, significant oxygen con-
sumption is observed for ketal 3 and aromatic acetal 4.
ð5Þ Surprisingly, DMI 19 is more reactive with oxygen than acetals/
6
2
termination, as depicted by eqn (5)–(10).
Initiation
•
RH ! R
ketals whereas, as expected, the electron-rich terpene d-limo-
nene 20 is oxidized much faster than any other compound
ð6Þ under study.
Propagation
•
•
R þ O2 ! ROO
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
Fig. 8 Monitoring of oxygen consumption with the PetroOxy apparatus during the autoxidation of acetals and ketals 1 to 4, as well as dimethyl
isosorbide 19 and d-limonene 20. Conditions: 2 mL pure compound, T = 80 °C, starting pressure P(O ) = 400 kPa.
2
The oxidation rate of ketals and acetals is mainly governed
by the kinetically determining step (eqn (7)). The homolytic
bond dissociation energy (BDE) of the weakest C–H bonds of
the substrate is one of the key thermodynamic parameters to
6
4,65
rationalize the oxidability of a compound.
Generally speak-
5
2
ing, the lower the BDE, the higher the propagation rate is.
Calculations of the BDEs of the C–H bonds using the DFT
method B3LYP 6-311G(2d,2p) have been performed for the
different solvents. Fig. 9 shows the most labile hydrogen(s) of
each solvent ranked on a scale of increasing BDEs. A limited
number of general rules are known to govern the ease of
6
1,66
abstracting the hydrogen atom from a C–H bond:
i) The C–H bond energy decreases with increasing
(
substitution at the primary > secondary > tertiary positions
since the resulting radicals are stabilised by the electro-
donating inductive effect of the alkyl groups.
(
ii) An α-phenyl group, an α-oxygen and an α-double bond
considerably decrease the BDE.
iii) The abstraction of an α-hydrogen from 5-membered
(
cyclic ethers (tetrahydrofurans) is easier than from the less
strained 6-membered cyclic ethers (tetrahydropyrans) which in
turn is easier than from the non-strained linear ether.
The order of the lowest BDEs of each solvent is in agree-
ment with the above rules: 20 (allylic positions of an electron-
rich cyclohexene) < 4a (dioxolane with one α-phenyl and two
α-oxygens) < 4b (dioxane with one α-phenyl and two α-oxygens)
<
19 (two strained 5-membered rings with two secondary C–H
and four tertiary C–H at the α position of an ether group)
3 (one tertiary C–H) < 1a (dioxolane with two α-oxygens) ≈ 2
electron-rich dioxolane with one α-oxygen) < 1b (dioxane with
two α-oxygens). Moreover, as expected, this scale based on
<
(
Fig. 9 Scale of oxidability of solvents 1–4 showing the weakest BDEs of
each compound (in red) in comparison with reference solvents: limo-
BDE values is also in fair agreement with their reactivity nene 20 and dimethyl isosorbide 19.
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
towards oxygen confirming that limonene 20 is by far the most 3.2. Preparation of benzaldehyde glycerol acetal 4a + 4b
oxidizable solvent. However, dimethyl isosorbide 19 appears to
be more reactive than 4 despite their respective BDE values.
Indeed, DMI has six times as many abstractable hydrogen
Benzaldehyde glycerol acetal 4a + 4b was prepared by acetalisa-
tion of glycerol with benzaldehyde using an acidic Amberlyst
15 resin. Details can be found in ESI.† The ratio between diox-
atoms as 4 per molecule. Furthermore once a first hydrogen is
olane and dioxane isomers varies a little with temperature and
reaction time. The final composition is always slightly in
favour of the dioxolane isomer.
•
abstracted, it forms a peroxyl radical ROO which reacts very
quickly with a second hydrogen in the same molecule of DMI
while the propagation reaction (eqn (7)) of the peroxyl radical
derived from 4 will be necessarily intermolecular and therefore
slower than an intramolecular reaction.
3.3. Hansen parameters
Hansen parameters of the four glycerol acetals/ketals and
other relevant solvents have been calculated using the Yama-
moto Molecular Breaking (Y-MB) method implemented in the
HSPiP software (HSPiP, version 4.0.05). The parameters of clas-
3
. Experimental
3
.1. Material
Glycerol formal 1 was obtained from Glacon Chemie (99%– sical solvents positioned for comparison in Fig. 2 are the ones
5
0
0
.013% wt/wt BHT), 2-isopropylidene glycerol (solketal) 2 was provided by Hansen. The distance DHansen in the Hansen’s
obtained from Sigma-Aldrich (98%) and 1,2-methylisobutyli- space is calculated according to the recommended method
5
0
dene glycerol 3 was a gift from Solvay (98%). Glycerol formal 1 (eqn (12)).
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
2
2
DHansen ¼ 4ðδ
ꢀ δ
Þ þ ðδ
ꢀ δ
Þ þ ðδ
ꢀ δhðcompoundÞÞ
ð12Þ
dðsolventÞ
dðcompoundÞ
pðsolventÞ
pðcompoundÞ
hðsolventÞ
is a mixture of dioxolane (1a) : dioxane (1b) isomers 34 : 66 3.4. COSMO-RS calculations
1
(
3
molar ratio) as determined by H NMR (CDCl ). The sample of
benzaldehyde glycerol acetal (BGA) 4 used for hydrolysis and ArgusLab (ArgusLab 4.0.1, Planaria Software LLC, Seattle, USA)
autoxidation studies is a mixture of dioxolane (4a) : dioxane (4b) was used to sketch molecules. Conformational analysis was
1
isomers 66 : 34 (molar ratio) as determined by H NMR (CDCl
3
). performed by the COSMOconf script. Molecular geometries
For volatility, hydrolysis or autoxidation comparative were optimized thanks to full DFT/COSMO calculations,
studies, 1,3-dioxolane 5, 2-methyl-1,3-dioxolane 6, 2,2- carried out using the quantum chemical program TURBO-
dimethyl-1,3-dioxolane 7, 2,2,4-trimethyl-1,3-dioxolane 8, 1,4- MOLE (TURBOMOLE GmbH, Karlsruhe, Germany) with Ahl-
dioxane 9, methylal 10, ethylal 11, n-butyl acetate 12, ethylene richs’ triple zeta valence polarization basis set and the Becke–
glycol n-butyl ether (C
) 14, diethylene glycol ethyl ether (C
glycol butyl ether (C E ) 16, diethylene glycol hexyl ether (C E ) thermodynamics calculations, including generation of σ-
4
E
1
) 13, propylene glycol n-butyl ether Perdew density functional. COSMOtherm (C31_1301 version,
(C
4
P
1
2 2
E
) 15, diethylene COSMOlogic, Leverkusen, Germany) was used for statistical
4
2
6 2
1
7, tetrahydrofurfuryl alcohol 18, dimethyl isosorbide 19 and potentials. These σ-potentials were then used as described
7
,51
limonene 20 were obtained from Sigma-Aldrich, Alfa Aesar and before
to position the solvents in the COSMO-RS space. The
TCI Europe (>98%). Commercial samples were distilled under closest neighbours were identified by calculating a distance
reduced pressure to remove possible antioxidants and were
kept under argon afterwards.
D
COSMO that takes into account the inertia of each PCA axis
(see ref. 7) (eqn (13)):
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
2
2
2
DCOSMO ¼
0:395ðF1ðsolventÞ ꢀ F1ðcompoundÞÞ þ 0:344ðF2ðsolventÞ ꢀ F2ðcompoundÞÞ þ 0:123ðF3ðsolventÞ ꢀ F3ðcompoundÞÞ þ 0:102ðF4ðsolventÞ ꢀ F4ðcompoundÞÞ
ð13Þ
2 3
For hydrolysis studies, D O (Eurisotop, >99.9%), CD CN 3.5. TGA measurements
(Eurisotop, >99.9%), acetonitrile (HPLC grade, Fluka),
dimethylformamide (SDS, 99.9%), phosphoric acid (Aldrich Volatility was evaluated by thermogravimetric analysis (TGA)
Chemie, 85%) and sodium hydroxide (Sigma-Aldrich, 97%) by recording the weight loss when imposing a temperature
were used.
ramp from room temperature to 300 °C at a rate of 10 °C
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
min− . Thermogravimetric analyses were performed on a TGA Table 4 Comparison of the relative stabilities of the glycerol acetals
1
and ketals towards hydrolysis and autoxidation. The third column indi-
cates the time required to oxidize 10% of the solvent at 80 °C (t10%). The
last column contains the hazardous solvent having the closest solvency
properties to the ones of the glycerol acetals/ketals
Q50 from TA Instruments. The atmosphere of the measuring
chamber was pure nitrogen to avoid possible oxidation of the
solvents at high temperatures. About 20 mg of the compound
was deposited on a platinum sample holder. The temperature
at which half of the sample was vaporised (T50%) was recorded. Ketal/
acetal
Hydrolysis
Autoxidation t10%
at 80 °C (h)
Closest harmful
solvents
The reproducibility was checked against the reference sample
dimethyl isosorbide 19) and was within ±3 °C.
k (M min−1)
−
1
solvent
(
1
2
3
4
a/1b
Stable
1.9
3.2
Stable
Stable
6
Aniline/glycol ethers
Glycol ethers
Nitroethane
3
.6. Hydrolysis
In a glass NMR tube, 0.1 to 0.3 mol L−1 of acetal/ketal (10 µL)
was added to 500 µL of a D O solution, containing 10 µL of
acetonitrile or 10 µL of DMF as internal standards. For ketal 3
and acetal 4, CD CN was also added to ensure complete misci-
a/4b
4.6/0.6
2.5
Aniline/nitroethane
2
3
lities towards hydrolysis and autoxidation. The evaluation of
the solvent properties through Hansen and COSMO-RS
approaches indicate that this family of solvents could rep-
resent interesting alternatives to some harmful solvents
including glycol ethers and aniline. One possible limitation of
glycerol acetals/ketals is their sensitivity towards hydrolysis
under acidic conditions. In-depth kinetic studies on a series of
acetals/ketals showed that structural modifications have a
great impact on the hydrolysis rate. Acetals are less sensitive
than ketals and dioxolane acetals are hydrolysed faster than
their dioxane counterparts. Dioxolane derivatives are also
more sensitive to autoxidation than their dioxane counterparts
in agreement with the bond dissociation energy calculations.
Controlling the dioxolane/dioxane ratio during synthesis is
thus of utmost importance. It is worth noting that glycerol
formal 1 is the most stable towards both hydrolysis and autoxi-
dation. Table 4 summarizes the main results of the present
paper.
This work shows that acetals/ketals of polyols may represent
an interesting class of solvents to be developed from the
biomass feedstock. Glycerol formal and solketal are already
presented as low-toxicity and environmentally-friendly bio-
based solvents. The good choice of aldehyde/ketone used for
the synthesis should allow controlling the stability towards
hydrolysis and autoxidation and tune the solvent properties of
the final product.
bility. The pH of the solutions was adjusted with phosphoric
acid or sodium hydroxide. Acetal/ketal hydrolysis was moni-
1
tored by recording the H NMR spectra on a Bruker AC spectro-
1
meter at 300.13 MHz for H at 25 ± 1 °C as a function of time.
The pH was measured using a pH meter (pH 330i/SET, WTW)
equipped with a combined pH electrode Sentix 81 (Glass elec-
trode, WTW) including a temperature probe. Electrode cali-
bration was carried out with pH 4, 7 and 10 buffers
beforehand.
3
.7. Autoxidation
2
mL of acetal/ketal free antioxidant was oxidized at 80 °C
under an atmosphere of pure dioxygen (400 kPa) using the Pet-
roOxy apparatus (Petrotest, Anton Parr). The compound was
introduced at room temperature into an inox cell (20 × 40 ×
2
6 mm) that was then purged and filled with pure dioxygen at
the desired pressure. The cell was then heated up to 80 °C and
the oxygen pressure was monitored every minute. Further
details about the PetroOxy apparatus are given in ref. 52.
3
.8. Calculation of Bond Dissociation Energies (BDE)
Calculation of Bond Dissociation Energies (BDE) was per-
formed by using Gaussian 03 packages. The geometries of all
the parent molecules were firstly optimised by using the
PM3 method and then the DFT by using the B3LYP/6-311G(d,
p) basis set. Orbitals from this method were used as inputs to
the final RB3LYP/6-311G++(2d,2p) calculation. The conformer
with the lowest electronic energy was used in this work. For
radicals, the optimisation was by the PM3 step plus the final Acknowledgements
UB3LYP/6-311G(d,p) method. The orbitals were then used as
This work was performed within the framework of the InBio-
SynSolv project supported by the French ANR CP2D program
no. ANR-CP2D-2009-08). Solvay is acknowledged for providing
samples of 2-isopropylidene glycerol and 1,2-methylisobutyl-
idene glycerol. We thank Dr Frédéric Cazaux (UMET, Lille) for
the TGA measurements, Dr Zéphirin Mouloungui and
Dr Manon Bergez-Lacoste from Université de Toulouse, INP,
LCA, France and INRA, UMR 1010 CAI, France, for their contri-
butions to the InBioSynSolv project, and Romain Guitard from
inputs in the final UB3LYP/6-311G++(2d,2p) calculation.
Table 1 shows the chemical structure of the oxygenated sol-
vents studied in this work. Relevant properties (boiling points,
thermogravimetric data, Hansen solubility parameters and
COSMO-RS coordinates) in the PCA space can be found in ESI.†
(
4
. Conclusion
In this work, four recently launched acetals/ketals of glycerol Université Lille 1, ENSCL, E.A. 4478 Chimie Moléculaire et
–4 have been investigated as bio-based solvents, with special Formulation, France, for his contribution to the calculations
1
emphasis on their solubilising properties and on their stabi- of BDE.
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
2
7 J. Bigorra Llosas, S. Sato, C. Bueno Ramiro and
E. Graupera, Cognis, Use of Glycerol Acetals, EP 2 0050 784
A1, 2007.
References
1
F. M. Kerton, Alternative Solvents for Green Chemistry, RSC,
Cambridge, UK, 2009.
28 D. H. Bonifacio, J. D. Vicentim and P. R. Garbelotto,
Rhodia, Aqueous Formulations Including Dioxolanes as
Coupling Agents, WO 2011/048481 A1, 2011.
2
P. De Caro and S. Thiébaud-Roux, Biosolvants, Ed.
Techniques Ingénieur, IN 102, 2008.
3
4
P. G. Jessop, Green Chem., 2011, 13, 1391–1398.
B. Zhou, J. Yang, M. Li and Y. Gu, Green Chem., 2011, 13,
29 C. Estévez Company, N. Bayarri Ferrer and J. Castells
Boliart, Institut Univ. De Ciència i Tecnologia, Proceso Para
La Preparacion De Loperamida, ES 2387793T3, 2008.
2
204–2211.
5
6
J. Yang, J. N. Tan and Y. Gu, Green Chem., 2012, 14, 3304– 30 A. Pirotte, Agriphar, Improved Method for Producing
317. Emulsifiable Pesticide Solutions, WO 201311108, 2013.
M. Sutter, L. Pehlivan, R. Lafon, W. Dayoub, Y. Raoul, 31 G. Lambiotte, Composition Insecticide, CH 690259 A5,
3
E. Métay and M. Lemaire, Green Chem., 2013, 15, 3020–
026.
L. Moity, M. Durand, A. Benazzouz, C. Pierlot, V. Molinier
1995.
3
32 R. P. V. Faria, C. S. M. Pereira, V. M. T. M. Silva,
J. M. Loureiro and A. E. Rodrigues, Chem. Eng. J., 2013,
233, 159–167.
7
8
9
and J. M. Aubry, Green Chem., 2012, 14, 1132–1145.
L. Moity, V. Molinier, A. Benazzouz, R. Barone, P. Marion 33 P. Sudarsanam, B. Mallesham, A. N. Prasad, P. S. Reddy
and J. M. Aubry, Green Chem., 2014, 16, 146–160. and M. Reddy, Fuel Process. Technol., 2013, 106, 539–545.
T. Werpy and G. Petersen, Top Value Added Chemicals 34 N. Rahmat, A. Z. Abdullah and A. R. Mohamed, Renewable
from Biomass: Volume I—Results of Screening for Potential & Sustainable Energy Reviews, 2010, 14, 987–1000.
Candidates from Sugars and Synthesis Gas, Report DOE/ 35 G. S. Nair, E. Adrijanto, A. Alsalme, I. V. Kozhevnikov,
GO-102004-1992, 2004.
0 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
411–2502.
1 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–
54.
2 R. A. Sheldon, Green Chem., 2014, 16, 950–963.
D. J. Cooke, D. R. Brown and N. R. Shiju, Catal. Sci.
Technol., 2012, 2, 1173–1179.
36 P. H. R. Silva, V. L. C. Gonçalves and C. J. A. Mota, Biore-
sour. Technol., 2010, 101, 6229–6255.
37 C. Crotti, E. Farnetti and N. Guidolin, Green Chem., 2010,
12, 2225–2231.
1
1
2
5
1
1
3 M. Pagliaro, R. Ciriminna, H. Kimura, M. Rossi and 38 V. R. Ruiz, A. Velty, L. L. Santos, A. Leyva-Pérez,
C. Della Pina, Angew. Chem., Int. Ed., 2007, 46, 4434–4440.
4 D. T. Johnson and K. A. Taconi, Environ. Prog., 2007, 26,
M. J. Sabater, S. Iborra and A. Corma, J. Catal., 2010, 271,
351–357.
39 B. L. Wegenhart and M. M. Abu-Omar, Inorg. Chem., 2010,
49, 4741–4743.
40 P. S. Reddy, P. Sudarsanam, G. Raju and B. M. Reddy,
J. Ind. Eng. Chem., 2012, 18, 648–654.
41 H. Serafim, I. M. Fonseca, A. M. Ramos, J. Vital and
J. E. Castanheiro, Chem. Eng. J., 2011, 178, 291–296.
42 C. X. A. Desilva, V. L. C. Gonçalves and C. J. A. Mota, Green
Chem., 2009, 11, 38–41.
1
1
1
1
1
1
2
3
38–348.
5 A. Behr, J. Eilting, K. Irawadi, J. Leschinski and F. Lindner,
Green Chem., 2008, 10, 13–30.
6 C. H. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. Lu, Chem.
Soc. Rev., 2008, 37, 527–549.
7 B. Katryniok, S. Paul, V. Belliere-Baca, P. Rey and
F. Dumeignil, Green Chem., 2010, 12, 2079–2098.
8 A. Martin and M. Richter, Eur. J. Lipid Sci. Technol., 2011,
1
13, 100–117.
9 F. Jérôme and J. Barrault, Eur. J. Lipid Sci. Technol., 2011,
13, 118–134.
43 J. Deutsch, A. Martin and H. Lieske, J. Catal., 2007, 245,
428–435.
1
44 M. B. Güemez, J. Requies, I. Agirre, P. L. Arias, V. L. Barrio
and J. F. Cambra, Chem. Eng. J., 2013, 228, 300–307.
45 S. B. Umbarkar, T. V. Kotbagi, A. V. Biradar, R. Pasricha,
J. Chanale, M. K. Dongare, A.-S. Mamede, C. Lancelot and
E. Payen, J. Mol. Catal. A: Chem., 2009, 310, 150–158.
46 B. Thomas, V. G. Ramu, S. Gopinath, J. George, M. Kurian,
G. Laurent, G. L. Drisko and S. Sugunan, Appl. Clay Sci.,
2011, 53, 227–235.
0 A. E. Diaz-Alvarez, J. Francos, B. Lastra-Barreira, P. Crochet
and V. Cadierno, Chem. Commun., 2011, 47, 6208–6227.
1 Y. Gu and F. Jérôme, Green Chem., 2010, 12, 1127–1138.
2 J. I. Garcia, H. Garcia-Marin, J. A. Mayoral and P. Perez,
Green Chem., 2010, 12, 426–434.
2
2
2
2
2
2
3 J. I. Garcia, H. Garcia-Marin and E. Pires, Green Chem.,
2
014, 16, 1007–1033.
4 Y. Gu and F. Jérôme, Chem. Soc. Rev., 2013, 42, 9550– 47 G. Vincente, J. A. Melero, G. Morales, M. Paniagua and
570. E. Martin, Green Chem., 2010, 12, 899–907.
5 H. Garcia, O. I. Garcia and J. M. Fraile, Chim. Oggi, 2008, 48 F. Adam, M. S. Batagarawa, K. M. Hello and S. S. Al-juaid,
6, 10–12. Chem. Pap., 2012, 66, 1048–1058.
6 J. L. Couturier, J. P. Lallier and J. L. Dubois, Arkema 49 C. Gonzalez-Arellano, R. A. D. Arancon and R. Luque, Green
France, Use of Glycerol Acetals as Coalescing Agents for Chem., 2014, 16, 4985–4993.
Lacquers, Paints or Varnishes and Composition Thereof, 50 C. M. Hansen, Hansen Solubility Parameters: A User’s
WO 2011/042652 A1, 2011. Handbook, CRC Press, Boca Raton, 2nd edn, 2007.
9
2
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
5
1 M. Durand, V. Molinier, W. Kunz and J. M. Aubry, Chem. – 59 T. C. Bruice and D. Piszkiewicz, J. Am. Chem. Soc., 1967, 89,
Eur. J., 2011, 17, 5155–5164. 3568–3576.
5
2 C. Marteau, F. Ruyffelaere, J. M. Aubry, C. Penverne, 60 P. Watts, J. Chem. Soc. B, 1968, 543–545.
D. Favier and V. Nardello-Rataj, Tetrahedron, 2013, 69, 61 E. T. Denisov and I. B. Afanaśev, Oxidation and Antioxidants
2
268–2275.
in Organic Chemistry and Biology, Taylor & Francis,
5
5
3 A. Benazzouz, L. Moity, C. Pierlot, M. Sergent, V. Molinier
2005.
and J. M. Aubry, Ind. Eng. Chem. Res., 2013, 52, 16585– 62 D. L. Rakhmankulov, S. S. Zlotsky, E. M. Kuramshin and
6597. L. G. Kulak, Int. J. Polym. Mater., 1990, 13, 113–122.
4 A. Benazzouz, L. Moity, C. Pierlot, V. Molinier and 63 A.-T. Karlberg, L. P. Shao, U. Nilsson, E. Gäfvert and
1
J. M. Aubry, Colloids Surf., A, 2014, 458, 101–109.
J. L. Nilsson, Arch. Dermatol. Res., 1994, 286, 97–
5
5
5 P. Rahimi and C. A. Ward, Int. J. Thermodyn., 2005, 8, 1–14.
6 M. Durand, Propriétés physico-chimiques, fonctionnelles 64 D. A. Pratt, J. H. Mills and N. A. Porter, J. Am. Chem. Soc.,
103.
et applicatives des éthers courts d’isosorbide, PhD, Université
de Lille 1, 2010.
7 E. H. Cordes and H. G. Bull, Chem. Rev., 1974, 74, 581–603.
2003, 125, 5801–5810.
65 S. Korcek, J. H. B. Chenier, J. A. Howard and K. U. Ingold,
Can. J. Chem., 1972, 50, 2285–2297.
5
5
8 S. Immel and F. W. Lichtenthaler, Liebigs Ann., 1995, 1925– 66 J. A. Howard and K. U. Ingold, Can. J. Chem., 1969, 47,
935. 3809–3815.
1
Green Chem.
This journal is © The Royal Society of Chemistry 2014