Investigations on heterogeneously catalysed condensations of glycerol to cyclic acetals

Article abstract of DOI:10.1016/j.jcat.2006.11.006

The acid-catalysed condensation of glycerol, a chemical from renewable materials, with benzaldehyde, formaldehyde, acetone (acetalisation), and their dimethyl acetals (transacetalisation) to mixtures of [1,3]dioxan-5-ols and [1,3]dioxolan-4-yl-methanols was investigated. Various solid acids were evaluated as heterogeneous catalysts for the desired glycerol conversion into these potential novel platform chemicals. [1,3]dioxan-5-ols are of particular interest as precursors for 1,3-propanediol derivatives. Therefore, the reported investigations were focused on the identification of reaction conditions that promote the formation of [1,3]dioxan-5-ols and suppress the formation of [1,3]dioxolan-4-yl-methanols.

Full text of DOI:10.1016/j.jcat.2006.11.006

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of Catalysis 245 (2007) 428–435
www.elsevier.com/locate/jcat
Investigations on heterogeneously catalysed condensations of glycerol
to cyclic acetals
J. Deutsch ∗, A. Martin, H. Lieske
Leibniz Institute for Catalysis at University Rostock, Branch Berlin, Richard-Willstätter-Strasse 12, 12489 Berlin, Germany
Received 21 August 2006; revised 2 November 2006; accepted 2 November 2006
Available online 5 December 2006
Dedicated to Prof. Bernhard Lücke on the occasion of his 70th birthday
Abstract
The acid-catalysed condensation of glycerol, a chemical from renewable materials, with benzaldehyde, formaldehyde, acetone (acetalisation),
and their dimethyl acetals (transacetalisation) to mixtures of [1,3]dioxan-5-ols and [1,3]dioxolan-4-yl-methanols was investigated. Various solid
acids were evaluated as heterogeneous catalysts for the desired glycerol conversion into these potential novel platform chemicals. [1,3]dioxan-5-
ols are of particular interest as precursors for 1,3-propanediol derivatives. Therefore, the reported investigations were focused on the identification
of reaction conditions that promote the formation of [1,3]dioxan-5-ols and suppress the formation of [1,3]dioxolan-4-yl-methanols.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Solid acids; Glycerol; Acetalisation; Transacetalisation; Cyclic acetals
1. Introduction
primary and secondary hydroxyl group in glycerol and to simi-
lar tendencies for the formation of the 5- and 6-membered rings
in 2 and 1. The direct separation of an isomer mixture 1 + 2
into their components obtained by the condensation of glycerol
with acetaldehyde was carried out on a small scale with prepar-
ative gas chromatography [2]. An isolation of one or both of
the compounds on a large scale is feasible by converting the
mixed isomers 1 and 2 into derivatives with sufficiently differ-
ent physical properties (i.e., boiling point, polarity, solubility).
However, this procedure is disadvantageous because of expen-
sive purification steps and results in considerable loss of the
target product [3,4].
Up to now, the condensation of glycerol with aldehydes or
ketones, and especially possible influences on the reaction prod-
uct ratio 1:2, have not yet been investigated systematically.
Within the framework of our studies, we intended to identify
the reaction conditions that influence the product composition.
The main aspects of these investigations were monitoring (i) the
reaction product formation and (ii) the variation of the reaction
partner for glycerol. In a continuation of our earlier studies on
heterogeneously catalysed aromatic acylations [5,6], we evalu-
ated various solid acids as environmentally benign catalysts for
the reaction.
Glycerol is obtained from the hydrolysis of fats into fatty
acids and, recently, in large quantities in the production of
biodiesel from rape oil. Due to this development, suitable uses
for cheap glycerol from natural sources need to be found.
A possible solution to the problem is the condensation of
glycerol with aldehydes, ketones, or their dimethyl acetals to
[1,3]dioxan-5-ols 1 and [1,3]dioxolan-4-yl-methanols 2 (iso-
meric 6- and 5-membered cyclic acetals) as novel fine chemical
intermediates (Scheme 1).
According to their structure, the 6-membered ring acetals 1
derived from glycerol are potential precursors for the produc-
tion of the “green” platform chemicals 1,3-dihydroxyacetone
and 1,3-propanediol that might be obtained by subsequent se-
lective oxidation or hyrogenation followed by acetal cleavage.
But, unfortunately, the condensation of glycerol with aldehy-
des or ketones gives product mixtures with an excess of 2 or
approximately equimolar amounts of 1 and 2 [1]. This phenom-
enon seems to be related to the nearly identical reactivity of the
*
Corresponding author.
E-mail address: deutsch@aca-berlin.de (J. Deutsch).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.11.006

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
J. Deutsch et al. / Journal of Catalysis 245 (2007) 428–435
429
Scheme 1. The cyclic acetals formed in the condensation of glycerol with aldehydes or ketones.
2. Experimental
and concentrated in vacuum to remove the solvent. The yields
and isomer ratios of the formed cyclic acetals were determined
by 1H-NMR spectroscopy (solvent, CDCl3). In addition, a sam-
ple of the pure product mixture (without the 1H-NMR standard
hexamethylbenzene) was prepared.
2.1. Chemicals
Glycerol (99+%), benzaldehyde (99+%), paraformalde-
hyde (95%), acetone (99+%), benzaldehyde dimethyl ac-
etal (99%), formaldehyde dimethyl acetal (99%), acetone
dimethyl acetal (98%), p-toluenesulphonic acid (98.5+%),
toluene (99.8%), benzene (99.9+%), chloroform (99.9+%),
and dichloromethane (99.9%) were purchased from Aldrich.
The chemicals were used without further purification. The
following solid acids were investigated as potential catalysts:
Amberlyst-36 (an arenesulphonic acid polymer), Nafion-H NR-
50 (a perfluoroalkanesulphonic acid polymer), and Montmo-
rillonite K-10 (a clay mineral), purchased from Aldrich, and
H-BEA (a zeolite, Si/Al = 25), purchased from Süd-Chemie.
The catalysts were pretreated before the experiments. Am-
berlyst-36 and Nafion-H NR-50 (beads) were dried at room
temperature in an evacuated exsiccator over concentrated sul-
phuric acid instead of silica gel and swollen after the drying
procedure overnight in the solvent used as a reaction medium.
H-BEA and Montmorillonite K-10 were calcined at 500 and
200 ◦C, respectively, in air for 3 h.
2.3.2. Condensation of glycerol with formaldehyde (source:
paraformaldehyde) and acetone (acetalisation)
The reactions were carried out as described above, but with-
out monitoring the product formation and thus in the absence
of the 1H-NMR-standard. After water separation was com-
plete, the reaction mixtures were cooled to room temperature,
filtered through CELITE® 521, and concentrated in vacuum.
The remainders were purified by Kugelrohr distillation (boiling
ranges: 75–100 ◦C at 7–8 mbar and 70–95 ◦C at 5–6 mbar). The
yields of the acetal mixtures thus obtained were calculated from
their weight, and the isomer ratios were analyzed by 1H-NMR
spectroscopy.
2.3.3. Condensation of glycerol with benzaldehyde dimethyl
acetal, formaldehyde dimethyl acetal, and acetone dimethyl
acetal (transacetalisation)
The reaction mixture (glycerol, dimethyl acetal of the
respective carbonyl compound, dichloromethane/methanol as
a solvent system, and catalyst) was magnetically stirred (1000
rpm) at room temperature in a closed 100-ml flask. To calcu-
late the product yield, 0.45 g of hexamethylbenzene was added
2.2. Characterization of the solid acids used as catalysts
Specific surface areas and pore diameters were measured
with nitrogen adsorption at −196 ◦C (ASAP 2000 system,
Micromeritics). The acidities of H-BEA and Montmorillonite
K-10 were characterised by temperature-programmed desorp-
tion (TPD) of ammonia (heat conductance detection), which
was preadsorbed at 100 ◦C. The ammonia desorbed was quan-
tified by reaction with 0.1 N sulphuric acid and back-titration.
1
as an inert internal H-NMR standard. Samples (0.5 ml) were
obtained periodically from the dichloromethane phase, filtered
through CELITE® 521, and concentrated in vacuum to remove
the solvent. The yields and isomer ratios of the formed cyclic
acetals were determined with 1H-NMR spectroscopy (solvent,
CDCl3).
2.3.4. 1H-NMR spectroscopic identification of the reaction
products
2.3. Catalytic experiments
Selected characteristic proton signals for the identification
of the synthesised cyclic acetals 1a–2c and 2a–2c are compiled
in Table 1. The measured chemical shifts are in agreement with
previously published data [4,7–13].
2.3.1. Condensation of glycerol with benzaldehyde
(acetalisation)
The magnetically stirred (1000 rpm) reaction mixture (glyc-
erol, aldehyde, solvent, and catalyst) was refluxed in a 100-mL
flask with a condenser. A Dean–Stark trap was used to remove
the formed water continuously. For monitoring the product for-
mation and calculating the product yield, 0.45 g of hexam-
ethylbenzene was added as an inert internal 1H-NMR standard
(CH3:2.21 ppm in CDCl3). Reaction samples (0.5 ml) were
obtained periodically, filtered immediately through CELITE®
521 (catalyst separation from the hot mixture), cooled to room
temperature, extracted with 0.5 mL of water to remove unre-
acted glycerol, dried with potassium carbonate, filtered again,
2.3.5. 1H-NMR spectroscopic analysis of the product
composition with varying temperature in presence of an acid
To 0.5 mmol of the acetal mixture (1a + 2a: a prepared
61:39 mixture; 1b + 2b: a commercially available 58:42 mix-
ture purchased from ACROS) 0.7 ml of toluene-d8 was added
as a solvent, followed by the addition of 0.01 mmol of p-
toluenesulphonic acid. The samples were analyzed at 25, 40,
60, and 80 ◦C. The equilibrium in the samples was reached im-
mediately after arriving the target temperature.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
430
J. Deutsch et al. / Journal of Catalysis 245 (2007) 428–435
Table 1
The cyclic acetals 1a–1c and 2a–2c synthesised in the condensation of glycerol with benzaldehyde, formaldehyde, acetone or their dimethyl acetals and characteristic
1
proton signals used for their identification and quantification with H-NMR spectroscopy
Reactants
Formed cyclic acetals as reaction products
Compound no
Proton signal
OCH(C H )O
Chemical shift (ppm)
Glycerol, benzaldehyde
or benzaldehyde
dimethyl acetal
(E)- and (Z)-2-phenyl-[1,3]dioxan-5-ol
(E)- and (Z)-(2-Phenyl-[1,3]dioxolan-4-yl)-
methanol
(E)-1a, (Z)-1a
(E)-2a, (Z)-2a
5.53(s), 5.37(s)
5.80(s), 5.93(s)
6
5
a
b
c
Glycerol, formaldehyde
or formaldehyde dimethyl acetal
[1,3]Dioxan-5-ol
[1,3]Dioxolan-4-yl-methanol
1b
2b
OCH O
4.78(d) , 4.91(d)
2
4.88(s), 5.06(s)
c
Glycerol, acetone
or acetone dimethyl acetal
2,2-Dimethyl-[1,3]dioxan-5-ol
(2,2-Dimethyl-[1,3]dioxolan-4-yl)-methanol
1c
2c
OC(CH ) O
1.44(d) , 1.46(d)
3 2
1.37(s), 1.44(s)
a
J = 6.2 Hz.
b
J = 6.3 Hz.
c
J = 0.6 Hz.
Scheme 2. Heterogeneously catalysed condensation of glycerol with benzaldehyde, formaldehyde, acetone (method A, acetalisation) or their dimethyl acetals
(method B, transacetalisation) to mixtures of [1,3]dioxan-5-ols 1 and [1,3]dioxolan-4-yl-methanols 2.
3. Results and discussion
an azeotropic mixture with the solvent) are strong driving
forces for the reaction. All solid acids were catalytically ac-
tive and produced the cyclic acetals 1a (six-membered cyclic
acetal) + 2a (five-membered cyclic acetal) in yields >90%
within few hours (Table 3) without side products. The required
catalyst loading to achieve comparable reaction rates was sig-
nificantly higher for H-BEA and Montmorillonite K-10 than
for Amberlyst-36 and Nafion-H NR-50. The synthesised prod-
uct mixture 1a + 2a contained the four conformational isomers
(E)-1a, (Z)-1a, (E)-2a, and (Z)-2a shown in Scheme 3, which
were identified with 1H-NMR-spectroscopy (see Table 1).
Cooling the reaction mixture and their storage at room tem-
perature over the solid catalysts gave a final product ratio 1a:2a
of about 60:40. Over Amberlyst-36 and H-BEA, this ratio was
already achieved after standing overnight. Under the same con-
ditions, the ratio was only 52:48 on Nafion-H NR-50 and 48:52
for Montmorillonite K-10. These ratios were conserved by re-
moval of the solid acids with filtration. A final product ratio of
about 60:40 was obtained on Nafion-H NR-50 and Montmoril-
lonite K-10 only after few days (Table 3). The slow product
3.1. Textural properties, acidities, and catalytic activity of
various solid acids
The four catalytically investigated solid acids and their
textural and acidic properties are listed in Table 2. The ion-
exchanger resins Amberlyst-36 and Nafion-H NR-50 have
small BET surface areas and sites of strong acidity (Brønsted
sites: protons of sulphonic acid groups). In comparison, the
inorganic solid acids H-BEA and Montmorillonite K-10 have
considerable higher BET surface areas, but weaker acid sites
(Lewis sites: Al ions, Brønsted sites: protons of Al ion coor-
dinated hydroxyl groups). Amberlyst-36 contains the largest
pores and the highest number of acid sites among these cata-
lysts.
The solid acids were tested in the condensation of glyc-
erol with benzaldehyde (Scheme 2, method A). The formation
of water as a thermodynamically stable condensation prod-
uct and its continuous removal from the reaction mixture (as

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
J. Deutsch et al. / Journal of Catalysis 245 (2007) 428–435
431
Scheme 3. Conformational isomers of 1a and 2a.
Table 2
Table 4
Properties of various solid acids used as catalysts in the condensation of glyc-
erol with benzaldehyde, formaldehyde, acetone, and their dimethyl acetals
Catalytic condensation of glycerol with benzaldehyde dimethyl acetal at room
temperature on various solid acids according to Scheme 2, method B
a
Solid acid
BET surface
area (m /g)
Average pore
diameter (Å)
Number of acid
sites (mmol/g)
Solid acid
Reaction
time (h)
Yield of
Ratio 1a:2a
2
1a + 2a (%)
81
82
a,b
Amberlyst-36
8
24
59:41
60:40
Amberlyst-36
19
200
5.4
a
Nafion-H NR-50 (beads) <1
No pores
0.8
Nafion-H NR-50
H-BEA
8
240
7
83
Not determined
61:39
c
d,e
H-BEA
500
233
7.6 × 6.4
60
0.87
d,e
Montmorillonite K-10
a
0.28
4
24
79
80
60:40
61:39
Ion exchange capacity according to product information.
b
c
d
e
mmol/ml.
Channels.
From ammonia TPD.
Sum of Lewis and Brønsted sites.
Montmorillonite K-10
a
4
24
79
79
59:41
60:40
Reaction mixture: 0.1 mol glycerol, 0.11 mol benzaldehyde dimethyl ac-
etal, 17.5 ml dichloromethane/1.5 ml methanol as solvent mixture, 0.5 g
(Amberlyst-36, Nafion-H NR-50) or 2 g (H-BEA, Montmorillonite K-10) cata-
lyst. The given yields are related to glycerol.
Table 3
Catalytic condensation of glycerol with benzaldehyde on various solid acids
according to Scheme 2, method A
a
Solid acid
Reaction Yield of Ratio 1a:2a after cooling the
the miscibility of the liquid components. With the exception of
Nafion-H NR-50, all catalysts gave the product mixture 1a+2a
within a few hours in high yield (Table 4). The achieved ratios
1a:2a of about 60:40 were the same, as for the condensation ac-
cording to method A (analysed for the product mixtures cooled
over the catalysts; see Table 3).
Due to the high activity and the low amounts to be used for
the glycerol condensations, Amberlyst-36 was selected as the
catalyst for the investigations reported in the following sections.
time
(h)
1a + 2a reaction mixture and
(%)
94
standing at room temperature
b
Amberlyst-36
Nafion-H NR-50
H-BEA
4
4
6
6
61:39
b
c
94
94
95
52:48 , 59:41
b
59:41
b
c
d
e
Montmorillonite K-10
a
48:52 , 54:46 , 59:41 , 60:40
Reaction mixture: 0.11 mol glycerol, 0.1 mol benzaldehyde, 17.5 ml reflux-
ing chloroform as the solvent, 0.1 g (Amberlyst-36, Nafion-H NR 50) or 1 g
(H-BEA, Montmorillonite K-10) catalyst. The given yields are related to benz-
aldehyde.
3.2. Catalytic condensation of glycerol with benzaldehyde
b
After standing overnight.
After standing for 3 days.
After standing for 7 days.
After standing for 10 days.
The condensation of glycerol with benzaldehyde to 1a + 2a
was performed on the catalyst Amberlyst-36 at varying reaction
temperatures in different solvents (Table 5). The reaction tem-
perature increased in the final phase of the experiments (com-
pleted water formation) from the boiling point of the azeotropic
mixture solvent/water to the boiling point of the pure solvent.
An extended reaction time was necessary to achieve a high
educt conversion in the low boiling solvent dichloromethane.
The catalytic experiments showed a significant dependence
of the final product ratio 1a:2a on the reaction temperature. It
was found that the formation of the six-membered cyclic acetal
c
d
e
equilibration may be related to a hindered access (Nafion-H
NR-50: low BET-area, no pores) and to a low number (Mont-
morillonite K-10) of acid surface sites (see Table 2).
The condensation of glycerol with benzaldehyde dimethyl
acetal at room temperature (Scheme 2, method B) was used as a
further probe reaction for the solid catalysts. A small amount of
methanol was added to the solvent dichloromethane to increase