Role of Water on the Activity of Magnesium Silicate for Transesterification Reactions

Article abstract of DOI:10.1002/cctc.201700139

This study aims to reveal how water adsorbed on to a solid can influence its catalytic properties for transesterification in liquid phase. A commercial magnesium silicate was subjected to a range of thermal pretreatments and used for the transesterification of ethyl acetate with methanol. Conversion of ethyl acetate decreased with increasing pretreatment temperature, in direct relation to the release of the water content of the magnesium silicate. Thermogravimetric analysis, diffuse reflectance infrared Fourier transform (DRIFT), and ¹H NMR spectroscopies revealed that physically adsorbed water had little influence on the reactivity. The water incorporated within the catalysts, however, which desorbs at higher temperatures, played a key role on the conversion. Calorimetry, in situ DRIFT spectroscopy, and ¹H NMR characterization indicate that two kinds of active site exist. These are created from the water coordinated to magnesium located on the edge of the clay-like particles or in the defects present in the silicate layer, respectively. Their role could be to stabilize methanol deprotonated by basic Mg-OH groups, activate the ester, or help the departure of the alkoxyl moiety.

Full text of DOI:10.1002/cctc.201700139

Accepted Article
Title: Role of water on the activity of magnesium silicate for
transesterification reaction
Authors: Longfei Lin, Damien Cornu, Maya Mounir Daou, Cyril
Domingos, Virginie Herledan, Jean-Marc Krafft, Guillaume
Laugel, Yannick Millot, and Helene Lauron-Pernot
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201700139
Link to VoR: http://dx.doi.org/10.1002/cctc.201700139
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
Role of water on the activity of magnesium silicate for
transesterification reaction
#
Longfei Lin, Damien Cornu , Maya Mounir Daou, Cyril Domingos, Virginie Herledan, Jean-Marc Krafft,
[
a]
Guillaume Laugel, Yannick Millot, and Hélène Lauron-Pernot*
[
34]
Abstract: This study aims to reveal how water adsorbed on a solid
influences its catalytic properties for transesterification reaction in
liquid phase. The reactivity of a commercial magnesium silicate was
studied after various thermal pretreatments for the reaction of
transesterification of ethyl acetate with methanol. It was observed
that the conversion decreases with the increase of the pretreatment
publication , we worked on a model transesterification reaction,
the conversion of ethyl acetate with methanol in order to
compare the activities of a series of magnesium based catalysts
in both liquid and gas phases. We found that magnesium silicate
shows a very good activity, in agreement with the results
[
35]
obtained by Narasimharao et al. . We also found that they are
more active than the strongly basic material MgO in the
transesterification in liquid phase and exhibit Magnesium Silicate
Hydrates (MSH) surface species that we proposed to play a key
temperature, therefore with the release of the water content of the
1
magnesium silicate. Thermogravimetric analysis, DRIFT and
H
NMR spectroscopies show that physically adsorbed water has little
influence on the reactivity, while water incorporated within the
[
34]
role in the reactivity
.
catalysts thus desorbing at higher temperatures plays a key role on
Magnesium silicate can be synthesized by different
1
[36–40]
the conversion. Calorimetry, in situ DRIFT spectroscopy and
H
methods such as mechano-chemical dehydration
,
[
35]
[41]
NMR indicate that two kinds of active sites exist, that are created
from the water coordinated to magnesium located on the edge of the
clay-like particles or in the defects present in the silicate layer,
respectively. Their role could be to stabilise methanol deprotonated
by basic Mg-OH groups, activate ester or help the departure of the
alkoxyl moiety.
hydrothermal synthesis
and precipitation . Sol-gel method
[
42]
was also proposed by some authors . Magnesium silicate
exists under different phases, such as chondrodite, clinohumite,
[
43]
forsterite, phase B, superhydrous B . The very hydroxylated
magnesium silicate (MSH) structure has been described by
[
40]
Roosz et al.
based on results from X-ray diffraction, modelling
2
9
XRD patterns and Si MAS NMR spectroscopy. XRD patterns
of MSH showed some broad peaks at 2θ = 20, 35, 60 and 72°.
Very recently it has been modelled by DFT calculations and
Introduction
[44]
showed to be very similar to defective talc . This kind of phase
thus consists in magnesium surrounded by an octahedron
In the field of biomass derivatives valorisation,
formed by oxygen, sandwiched between defective silicate
transesterification reactions are used in
a
wide range of
[40]
layers
.
applications, for instance in liquid phase for the production of
It is well-known that water inhibits base-catalysed
transesterification reactions since it competes with the alcohol
[
1]
[2–4]
biodiesel and lubricant
and in gas phase for the generation
[
5]
of monomers for polymerization . Because of global warming
issues and depletion of the fossil fuels, the production of
biodiesel, used as an alternative to fossil sources has attracted a
(
reactant), thus lowering the conversion as observed for MgO in
[15]
gas phase
or transforming the required transesterification
reaction into ester hydrolysis, leading to the formation of free
[
6–11]
great deal of interest
. Homogeneous catalysts, such as
[45,46]
fatty acids
. This latter effect is however less pronounced in
sodium and potassium hydroxides, could achieve a high yield for
[45]
heterogeneous catalysis than in the homogeneous one
.
[
12]
the conversion of triglycerides . However, the high cost for the
purification of products make heterogeneous catalysts more
Nevertheless, water has also been shown to play a positive role
[47]
in different heterogeneous catalytic reactions . For example, in
[
13,14]
attractive
. Basic solids have gained more attention than
[8]
the case of biodiesel production, Zabeti et al. noted that the
[
7]
acidic materials due to their higher activity . Catalysts with
strong basic properties, such as calcium and magnesium
presence of little amounts of water in the reaction medium can
enhance the catalytic activity of calcium oxide (CaO). Their
conclusion is that water is necessary to form hydroxide ions,
which will be an homogeneous catalyst in the reaction. However,
all these publications study the influence of water in the reaction
mixture, as biomass products often contain water. Few works
have studied specifically the influence of adsorbed water on the
catalyst.
[
15–20]
[21–25]
oxides
or zeolites
were reported to catalyse
transesterification reactions. In addition, some other solids with
weaker surface basicity are also able to catalyse these reactions,
[
26–29]
[30]
i.e. hydrotalcites
, potassium carbonate or phosphate
,
[
31]
[32,33]
barium hydroxide , or zinc based materials
. In a previous
[
34]
In previous work , we reported that magnesium silicate is
a bifunctional catalyst and offers on the same surface acidic and
basic active sites. This bifunctional character may be a key
parameter to explain the high activity but it is well known that
water and other adsorbates may modify the reactivity of a
[
a]
L, Lin, Dr. D. Cornu, M. M. Daou, C. Domingos, Dr. V. Herledan, Dr.
J. Krafft, Dr. G. Laugel, Dr. Y. Millot, Prof. H. Lauron-Pernot
Sorbonne Universités, UPMC Univ Paris 06, UMR CNRS 7197,
Laboratoire de Réactivité de Surface
ème
Tour 43-33, 3
France
étage, Case 178, 4 Place Jussieu, F-75252, Paris,
[48]
surface . Such a phenomenon has been evidenced, in the
*
E-mail: helene.pernot@upmc.fr
Present address:
[
28,49]
#
case of hydrotalcites, in transesterification reactions
.
Foundation ICIQ Avda. Països Catalans, 16 43007 Tarragona,
Spain
The aim of this work is to study the role of adsorbed water
on the catalytic behaviour of a commercial magnesium silicate
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
(
labelled MS) in the model reaction of transesterification of ethyl
acetate by methanol. In order to link the yield obtained in
catalysis to surface properties, the characterization of the
1
catalysts is achieved by infrared and H Nuclear Magnetic
Resonance spectroscopies.
Results and Discussion
Catalysis performances and TGA results
Table 1 indicates the conversion of ethyl acetate (AcOEt)
for magnesium silicate with (MS-X, X being the temperature of
the pretreatment in °C) and without (MS) pretreatment.
Conversions of ethyl acetate decreases for increasing
pretreatment temperatures, especially for pretreatment
temperatures higher than 70°C. Figure 1 shows the results from
thermogravimetric
analysis
(TGA)
and
differential
thermogravimetric analysis (DTA) of MS and MS-X. DTA curves
(
Figure 1B) show a main weight loss feature, occurring below
[
38]
2
00 °C that can be attributed to the removal of water . Water
adsorbed on catalysts is released step by step (Figure 1A) and
the maximum for DTA curve shifts towards higher temperatures
(
Figure 1B) for increasing pretreatment temperatures. The data
are listed in Table 1 to compare with catalytic performance. The
quite significant decrease of conversion does not occur for the
important loss weight under 70°C, attributed to water weakly
bonded to the surface, but when more strongly adsorbed water
molecules are desorbed. Similar results were also found by Xi et
[
28,29]
al.
for the activity of rehydrated hydrotalcites in the
transesterification reaction of tributyrin with methanol. This effect
-
was ascribed to the loss of OH base sites that were produced
[
28,29,50]
by dissociating interlayer water
.
Figure 1. (A) Thermogravimetric analysis (TGA), (B) differential
thermogravimetric analysis (DTA) for MS depending on pretreatment
temperature. Little rehydration in the oven is responsible for the slight weight
increase at low temperature.
Table 1. Conversion and TGA/DTG measurements for magnesium silicate
without (MS) and with (MS-X) pretreatment (X being the pretreatment
temperature in °C).
Rehydration experiments were performed in water saturated
atmosphere on the sample MS-140. As shown in Table 2 and
Figure 2, 0.4 h exposure is enough to fully recover the global
water content of the sample MS, with similar desorption pattern.
Longer exposure times increase slowly the water content of the
sample, reaching 51.8% after 425 h. Catalytic tests of
rehydrated samples show that the activity can be fully recovered
after 0.4 h rehydration period (Table 2). It indicates that the
catalyst is not irreversibly modified upon drying and that the
decrease of the reactivity cannot be explained by a structure
modification. However, excessive rehydration is harmful to the
reactivity, leading to some decrease of the conversion (Table 2).
The effect of this excessive water in the catalyst is similar to the
one observed with direct introduction in the reactant mixture of
liquid water: when 0.1 mL of water is added into the schlenk
together with the reactants, the conversion decreases to 4.7%,
compared to 22.8% initially. This may be due to the deactivation
caused by the hydrolysis of the ester molecules over the catalyst
Sample
AcOEt
Conv./Mass
TGA/DTA Measurement
-
1 b
Conv.
(% g )
a
(
%)
Weight
loss
Temp. maximum
desorp. (°C)
c
(
%)
MS
22.8
22.0
15.2
14.5
13.8
1.52
1.45
1.03
1.00
0.88
16.5
5.4
4.5
3.4
1.8
71
MS-70
MS-90
MS-110
MS-140
128
136
145
148
[
a] The mass of catalysts added in schlenk flask is around 15 -18 mg, and thus
pure mass of catalyst (excluding water) is around 15 mg.
b] Conversion of ethyl acetate divided by pure mass of catalyst (excluding
water).
c] Weight lost below 200 °C.
[
[
[
29]
as mentioned by Xi et al. on hydrotalcites catalyst
.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
DRIFTS analysis
Figure 3 gathers DRIFT spectra of MS and MS-X. Only
Table 2. Results of catalytic tests and TGA/DTG measurement as a function
of rehydration time for MS-140.
-
1
wavenumbers from 4000 to 1500 cm are shown because they
provide information on hydroxyl groups and water stretching and
bending. In each spectrum four bands can be observed, at 3740
Sample
AcOEt
Conv./Mass
TGA/DTA Measurement
a
-1
b
Conv. (%)
(% g )
Weight
loss
Temp.
-
1
-
3720, 3672, 3620 - 3200, and 1636 cm . The νOH band is
maximum
desorp. (°C)
c
-1
[51]
(
%)
located at 3737 cm for the stretching of silanols over silica
-
1
[52,53]
and at 3720 cm on sepiolite, a magnesium silicate clay
.
MS
22.8
13.8
22.0
16.1
16.1
11.2
1.52
0.88
1.47
1.21
1.26
1.29
16.5
1.8
69
148
69
64
61
70
-1
The band observed here at 3740 to 3720 cm is thus assigned
to OH stretching of silanols species. The band located at 3672
MS-140
-
1
cm has been ascribed to hydroxyls coordinated to the
[
38]
MS-140-R0.4
MS-140-R2.5
MS-140-R24
MS-140-R425
17.1
26.2
29.0
51.8
magnesium (Mg-OH stretching) , as what is observed in
[
52,54]
sepiolite, antigorites and talc
. The broad band from 3620 to
-1
3
200 cm was attributed to OH stretching of adsorbed
[
35,51]
water
either coordinated to magnesium or adsorbed on
[
55]
-1
silanols . The contribution at 1636 cm was assigned to H-O-H
[35,38,39,51]
bending vibrations of molecular bound water
. The
suggested assignments found in the literature are listed in Table
[
a] The mass of catalysts added in schlenk flask is 18 mg (except for MS-140
3
.
that is 16 mg).
It can be seen in Figure 3 that, when increasing the
[
b] Conversion of ethyl acetate divided by pure mass of catalyst (excluding
water).
c] Weight lost below 200 °C.
-1
pretreatment temperature, the intensity of the band at 1636 cm
-
1
and that of the broad band at 3620 to 3200 cm decrease
significantly, which is well consistent with the water loss
observed with TGA results. As the catalytic performances (Table
[
1
), show a significant decrease in conversion by pretreating the
sample from 70 to 140 °C, the spectrum of MS-140 is compared
to the one of MS-70, with the curve log(MS-140/MS-70) (Figure
-
1
4
). A sharp band appears at 3732 cm , which means that the
intensity of this band increases for MS-140 compared to MS-70.
-
1
However, the band at 3672 cm is quasi not altered.
Thus, accompanying the release of adsorbed water, the
-
1
intensity of silanol bands (3740 to 3720 cm ) arises, whereas
the intensity of Mg-OH stretching band is not changed, although
the bands are revealed and become more apparent.
Table 3. Assignments of Infrared absorption bands of MS comparing with that
-
1
of sepiolite given in the literature (wavelength in cm ).
[
52,53,55,56]
MS
Sepiolite
Suggested assignments
Si-OH hydroxyl stretch
Internal Mg-OH stretch
Mg-OH stretch
3
3
740 - 3720
3720
3680
672
3
3
3
3
620
565
430
230
OH-stretch of water coordinated to Mg
Water OH-stretch
3
1
620 - 3200
Water OH-stretch
717
1710
Eventually a bending mode of water with
strong hydrogen bonding
1
1
684
651
1660 - 1670
1656
Bending of water coordinated to Mg
Bending of zeolite-like water
(2 H-bonded)
Figure 2. (A) Thermogravimetric analysis (TGA), (B) differential
thermogravimetric analysis (DTA) for MS-140 depending on time of
rehydration (Rt, with t being the rehydration time in h).
1632
1630
Bending of zeolite-like water
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
(
1 H-bonded)
layered silicate. The brucite-like sheet is thus sandwiched
between two silicate sheets. The talc is therefore
centrosymmetric and exhibits only internal OH with a stretching
1
613
1616
Bending of water weakly H-bonded
-
1[59]
mode at 3677 cm
. For our MS sample, the intensity of the
-
1
band at 3672 cm is not altered while releasing water, this band
should therefore be ascribed to internal Mg-OH.
Figure 5. Two dominant modes of adsorption of water on magnesium: (a)
adsorption of water on magnesium located at the edge of layer, (b) adsorption
of water on magnesium located in defects.
Sepiolite is another kind of layered silicate, with ordered
cavities filled with water molecules. For this solid, the bands at
-
1
Figure 3. DRIFT spectra of MS after each pretreatment step. They are taken
3
620 - 3200 cm in mid-IR spectrum were deconvoluted into
at the pretreatment temperature.
-1[52]
four bands at 3620, 3565, 3430 and 3230 cm
near-infrared with mid-IR spectroscopy, the authors assigned
. By comparing
-
1
the band at 3620 cm to OH stretching for hydroxyl group
-
1
adsorbed on Mg, the one at 3565 cm to the OH stretching of
water also coordinated to Mg cation, and bands at 3430 and
-
1
3
230 cm were assigned to water OH-stretching weakly
coordinated to the surface. For the same magnesium silicate,
[
55]
-1
Cannings et al. ascribed the bands at 3599 and 3532 cm to
antisymmetric and symmetric stretching modes of molecular
water coordinated to Mg cation at the edges of the channels.
2
+
2-
Water molecule could be adsorbed and dissociated on Mg -O
[
60,61]
pairs to form two OH groups associated by H-bonding
, and
these OH groups contribute in a stretching band around 3583
-
1[62,63]
and 3515 cm
. As MSH is a defective structure (Figure 5) in
silicate groups, Mg located at the edge or close to the defects of
the silicate layer are probably able to adsorb water or exhibit OH
-
1
groups. The broad bands at 3620 - 3200 cm may result from
different contributions such as the stretching of Mg-OH at edge
or defects of layer, water OH-stretch linked to magnesium, water
OH-stretch adsorbed on Mg-OH or silanols.
Figure 4. DRIFT spectra of MS-70 and MS-140, and their subtraction curve.
A
more precise assignment can be obtained from
comparison with the IR spectra of crystalline magnesium
silicates: chysolite, talc, sepiolite. Chrysotile is a 1:1 layered
silicate, in which an octahedral brucite-like sheet is covered up
[
57]
by a SiO
OH, one is at the top of the brucite-like sheet (external Mg-OH)
4
tetrahedral sheet . Hence, there are two kinds of Mg-
-
1
with a stretching frequency of 3695 cm , and the other one is at
the bottom of the brucite-like sheet next to the silicate one
-
1[58]
(
internal Mg-OH) with a frequency of 3649 cm
. Only external
Mg-OH is accessible to water. On the opposite, talc, is a 2:1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
-
1
by water molecule. The band at 1660 - 1670 cm is assigned to
water strongly bonded to magnesium cation in sepiolite. Famer
[
64]
-1
et al.
distinct water molecules in sepiolite: the former has two OH
assigned the bands at 1656 and 1630 cm to two
[
56]
groups involved in hydrogen bonding, the later only one . Prost
[
56]
et al. reported that when sepiolite was dehydrated, a band at
-
1
1
616 cm appeared. They attributed this band to water
molecules connected with weak hydrogen bonds in agreement
with physical considerations showing that the wavenumber of
water bending vibration should decrease with a decrease in the
[
65]
strength of hydrogen bonding
.
Nevertheless an inverse
low
relationship was observed by Kolesov et al. and
a
-
1
wavenumber of around 1590 cm was ascribed to the bending
of water adsorbed on cation and strongly H-bond to structural
[
66]
oxide
.
As can be shown from Figure 7, up to 90°C, the bands at
-
1
1
684, 1651, 1632 and 1613 cm decrease substantially in the
-
1
same way. Then, beyond 90°C, the band located at 1684 cm
-
1
does not seem to evolve. The band at 1613 cm is not modified
for pretreatment temperatures higher than 110°C. It is possible
-
1
that the bands at 1651, 1632 and 1613 cm characterize the
water interacting with other water molecules or silanols and that
-
1
the band at 1684 cm characterizes the water molecules
interacting with magnesium, therefore more rigid than the one in
the H bond network. The evolution of the intensities of this band
with pretreatment temperature (Figure 7) can be understood
considering two kinds of magnesium presented on Figure 5: the
one located on the edge of the clay-like particles and the other
one in the defects of the silicate layer, respectively a and b on
the scheme proposed Figure 5. Please note that in each site,
water molecule can also interact through H-bond with Mg-OH or
Si-OH groups but for sake of clarity, these H-bonds are not
shown on Figure 5. One of these species may be released up to
Figure 6. Deconvoluted DRIFT spectra of MS and MS-X at the region of 1800
-
1
–
1500 cm .
9
1
0°C while the other one, more strongly bounded is stable up to
40°C.
-
1
The evolution of the band located at 1613 cm is more
difficult to comment due to controversial assignment in literature.
Again two different species must be involved in this contribution,
both adsorbed on the H-bond network formed on sites a or b
Figure 5 (not represented). The first one is weakly bonded in
[
56]
agreement with the assignment of Prost et al. and desorbed at
low temperature. The second one is stable above 110 °C and
must be linked to water more strongly bonded.
Further work is necessary to confirm these hypotheses and to
1
determine water adsorption energy on both sites. H NMR
analysis
Figure 7. The areas of deconvoluted DRIFT spectra at the region of 1800 –
-
1
1
1
500 cm evolved with pretreatment temperature.
H NMR spectra are obtained for MS and MS-X and shown
in Figure 8. For sake of comparison, the spectrum of fully
dehydrated sample, thus pretreated at 350°C, is also given.
The experimental bands of water bending modes around
Three resonance regions are observed. The broad resonance at
-
1
[38,43,51]
1
1
636 cm are decomposed and fitted with five bands at 1717,
5
ppm can be assigned to adsorbed water
. The
-
1
[51]
684, 1651, 1632, and 1613 cm using Gaussian line shape
resonance at 1.8 ppm can be attributed to silanol . And the
[
38]
functions (Figure 6). The areas for each band evolved with
pretreatment temperature (in Figure 7). Sepiolite is a structure
with several water molecules incorporated in the channel. For
resonance in the range 0 - 1 ppm can be assigned to Mg-OH
.
When increasing pretreatment temperature, the peak at 5
ppm decreases dramatically and the maximum shifts to lower
-
1
1
this structure, the band at 1710 cm is difficult to attribute, as
Frost et al, states that it is not sure that this absorption is done
value. Yesinowski et al. found that, for water in mineral,
H
isotropic chemical shifts of oxygen-bound hydrogen depend
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
linearly on the O-H··O distance. For instance, two structurally
distinct water molecules were observed in nepheline (which is
natural aluminosilicate), located at 4.6 and 3.4 ppm,
corresponding to O-H··O distances of 2.92 and 2.97 Å,
respectively. Furthermore, the peak at 5 ppm is assigned to
highly mobile water molecules found in fluid inclusions and the
broad peak at around 4 ppm is ascribed to a structurally isolated
[
67]
water group . In our experiments, the highly mobile water
molecules are released at low pretreatment temperatures. This
desorption leads to higher O-H··O distances, thus the peak is
shifted to lower values. Confronting these results to catalytic
performances, we can note that the activity of catalyst is not
affected by removal of the highly mobile water below 70°C, but
the activity decreases significantly for pretreatment temperatures
higher than 90 °C when the structurally isolated water molecules
at 4 ppm are released.
The silanol resonance at 1.8 ppm appears after the
pretreatment at 70 °C and does not shift upon further
dehydratation.
On the reverse, the Mg-OH contribution at 0-1 ppm is
shifted upon the pretreatements. In the fully hydrated sample,
the centre of Mg-OH resonance is located at 0.4 ppm. A
moderate thermal treatment releases the highly mobile water
and reveals the structurally isolated water molecules that are
adsorbed on Mg-OH, shifting the resonance to 0.6 ppm. Then
further thermal treatment at 350 °C removes the isolated water
molecules and leaves bare Mg-OH, causing a further shift to 0.8
ppm. This is also observed by Nied et al. for the position of the
Mg-OH peak which shifts from 0.57 to 0.74 ppm as the amount
[
38]
of water in MSH decreases . In addition, on MS-350, a new
contribution appears at 3.6 ppm that can be related to bridging
OH groups formed upon dehydroxylation, as the case of Si-
[
68,69]
(
OH)-Al in zeolite
The appearance of the silanol resonance at 1.8 ppm upon
drying at 70 °C is in agreement with DRIFTS experiments
.
1
Figure 8. H NMR spectra of MS and MS-X.
-
1
showing that the silanol contribution at 3732 cm grows for
higher pretreatment temperatures. Concerning the Mg-OH
contribution, DRIFT spectroscopy shows clearly that internal Mg-
Calorimetry analysis
Figure 9 shows the adsorption of ethyl acetate on MS and
MS-X by calorimetry measurement. The experiments were
conducted in solid-liquid phase at 30 °C, measuring the heat
released by contacting 3.8 mmol of ethyl acetate with 0.15 g of
the sample. The overall heat of adsorption is thus obtained by
integration after reaching equilibrium (8 h). Heats of adsorption
increase with the pretreatment temperature up to 90 °C. A
further increase of pretreatment temperature no longer improves
it. There are two dominant modes of adsorption of ester on the
surface of MSH, either via its electronegative oxygen in carbonyl
-
1
OH (3672 cm ) are not affected by water adsorption but cannot
conclude about the interaction between external Mg-OH and
-
1
water as the contribution is located in the 3620 – 3200 cm large
1
band. By
H NMR, the shift of the Mg-OH band upon
dehydration clearly evidences that water desorbs from sites
implying Mg-OH groups. Two successive shifts are evidenced at
moderated temperature and at 350 °C, that can be related to the
successive desorption of water from sites a and b (Figure 5) in
which it may interact with Mg-OH groups by H-bonding. This
hypothesis is in line with the evolution of water adsorbed IR
[
70]
group to surface cation
or through a hydrogen bond to OH
-
1
[49]
bands located 1684 and 1613 cm that is discussed above.
group . Similarly, water can also adsorb on surface cation
[
70]
through its nucleophilic properties
or on OH group via
[
71]
hydrogen bonding . Clausen et al. studied the adsorption of
ethyl acetate and water on the dry surface of a smectite clay
through density functional theory and they showed that
adsorption energy of ethyl acetate on sodium cation is close to
[
70]
the one of water . Even if our system is different, a competition
between the adsorption of ethyl acetate and water exists. This
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
may explain the increase of heat of adsorption of ethyl acetate
when releasing water.
measured by calorimetry as a function of pretreatment
temperature exhibit a comparable shape: it increases up to
90 °C and is stable up to 140°C, in agreement with a two
steps water releasing.
The fact that the mean heat of ester adsorption is constant
for pretreatment temperature over 90 °C can be understood
considering the two sites proposed from IR results for water
coordination on Mg cations (a and b on Figure 5). At room
temperature, both sites are hydrated leading to a low value of
heat of adsorption. Up to 90 °C, water is desorbed from one of
the sites leading to the decrease of the IR band located at 1684
Water is known to inhibit the activity of basic
[
15,45,46]
catalysts
. Nevertheless, the above results show that water
molecules are beneficial to this reaction. From literature analysis
we can formulated three hypothesis to explain the specific role
of water in our systems.
-
1
cm and to a more efficient adsorption of the ester. Over 90 °C
no more water is released from these sites as seen by DRIFTS,
in agreement with a constant value of the mean heat adsorption
measured by calorimetry.
-
It was reported that in amorphous silica-alumina, water
molecules adsorbed on aluminium atoms and interacting
with neighbouring silanols can easily be deprotonated and
[
72,73]
therefore are Brønsted acidic sites
. Thus in MSH,
magnesium may polarize adsorbed water and form
Brønsted acidic sites Figure 5 a and b), which could
stabilise the methoxide molecule, or activate ester, directly
by bonding to the carbonyl group, or indirectly, favouring
the departure of the alkoxyl group as reported by Greenwell
et al. for adsorbed water on MgAl-layered double
[
49]
hydroxides
.
-
-
Besides, water may also act as an inhibitor of too strong
acidic sites that retain too strongly the reactants, especially
the ester as shown by calorimetry.
Moreover, water contained in catalyst can facilitate the
migration of methanol due to the high miscibility between
water and methanol, as mentioned by several
[
28,35]
researchers
. The removal of this positive water may
also affect the conversion of ethyl acetate.
The precise role of this water molecule on the mechanism
needs further studies to be elucidated.
-
1
Figure 9. Adsorption heats of ethyl acetate on MS and MS-X. The unit J g is
Joule per gramme of pure mass of catalyst (excluding water).
Conclusions
Discussion of the structural characterizations and role of water
Magnesium silicate was thermally treated to remove water
adsorbed, and the catalytic performance of this solid before and
after pretreatments was tested in the transesterification of ethyl
1
Combining the analyses of DRIFTS,
H NMR and
calorimetry, various types of adsorption of water molecules over
MSH surface are evidenced.
1
acetate with methanol. Based on DRIFT spectra and H NMR
-
-
Water molecules adsorbed on silanol groups via hydrogen
bonding can easily be desorbed, and the uncovered silanol
results, the effect of water adsorbed on MS on the conversion in
transesterification reaction is the following: highly mobile water
molecules released for pretreatment temperature lower than
70 °C have no effect on the reactivity. Then, for pretreatment
temperatures between 70 to 90 °C, water molecules that are
believed to be bonded to magnesium from their DRIFT spectra
-
1
groups are evidenced both by DRIFTS (band at 3732 cm )
1
and H NMR (peak at 1.8 ppm).
Although internal Mg-OH cannot be accessed directly and
are therefore not perturbed during dehydratation, water
molecules can interact with Mg-OH and/or Mg cations
located on the edge of the clay-like particles or in the
defects present in the silicate layer. Specifically, the two
modes of the water molecules linked to Mg cations (Figure
1
and H NMR characteristics are lost, and the conversion for the
transesterification reaction decreases markedly. Further
increase of the pretreatment temperature only led to a small
decrease of the conversion as the number of water molecules
bonded to magnesium is maintained. Two kinds of active sites
are proposed to be water molecules bonded to magnesium
located on the edge or in the defects of the clay-like particles.
They role may be to stabilise methanol deprotonated by basic
Mg-OH groups, activate ester or help the departure of the
alkoxyl moiety. Upon their removal, the adsorption energy of
ethyl acetate over the catalyst increases, which mean that the
ester can directly be connected to the magnesium ion. Too
5
a and b) can be released at different temperatures. The
-
1
band at 1684 cm in DRIFTS show that one species can
be released up to 90°C while the other one, is stable up to
1
40°C. This observation is consistent with the evolution of
1
the peak at 4 ppm in H NMR (ascribed to isolated water)
that decreases up to 90°C and is stable up to 140°C. The
adsorption of these two kinds of water molecules
competes with that of ethyl acetate on the surface of MSH.
The evolution of the mean heat adsorption of ethyl acetate
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
strong adsorption energy of the ester is therefore detrimental for
its reactivity.
recording conditions and using the same empty or filled rotor, the probe
and rotor signals were subtracted from the total FID.
Heat of adsorption measurements were conducted on SETARAM
Calvet Calorimeter C80 at 30 °C. 0.15 g of sample was loaded into the
sample cell (Membrane Mixing Cell), then a tube with a telfon membrane
at the bottom was set up at the top of sample cell, and then 0.38 mmol.
of ethyl acetate was added into the tube. After the temperature of sample
and the heat energy are stable, the measurement was started by
breaking the telfon membrane with a stirrer bar so the sample and ethyl
acetate mixed. The stirring rate was kept at 1500 rpm.
Experimental Section
Catalysts pretreatment
Magnesol® catasorb was kindly provided by the Dallas group of
America®. It contains Cl (<0.5%) and Na (<2%) impurities. Here it is
denoted as MS. The MS were thermally treated at 70, 90, 110, 140 °C
-
1
-1
respectively for 2 h after heating at a rate of 5 °C min in 20 mL min of
flowing N . Then they cooled down to room temperature. These samples
are denoted as MS-70, MS-90, MS-110 and MS-140, respectively. MS-
40 was put in a desiccator with water at bottom for rehydrating. Then
2
Acknowledgements
1
the samples were noted as MS-140-Rt, t is the time of rehydration (hour),
The authors acknowledge China Scholarship Council (File
No.201406140144) for Longfei Lin PhD grant. The cluster of
excellence Matisse is thanked for Maya M. Daou Master’s
degree grant. The NMR experiments have been performed on
IMPC (Institut des Matériaux de Paris Centre, FR2482) platform.
Transesterification reaction
Around 18 mg of the catalyst were introduced in a Schlenk flask.
Vacuum was performed in the Schlenk flask through a vacuum manifold
-
3
(
(
10 torr). Nitrogen (Air liquide 99.99%) was introduced into the Schlenk
1 bar) and then methanol (10 mL, Sigma Aldrich anhydrous 99.8%) as
Keywords: Transesterification • magnesium silicate • adsorbed
well as 1,4 dioxane (1 mL, Sigma Aldrich anhydrous 99.8%), as standard
reference, were introduced through needles into the Schlenk. The
temperature of the liquid phase was maintained at 60 °C. The
introduction of ethyl acetate (AcOEt, 1 mL, Sigma Aldrich anhydrous
water • infrared spectroscopy • NMR spectroscopy
[
1]
L. Bournay, G. Hillion, P. Boucot, J.-A. Chodorge, C. Bronner, A.
Forestiere, Process for Producing Alkyl Esters from a Vegetable or
Animal Oil and an Aliphatic Monoalcohol, 2005, US 6878837 B2.
S. Gryglewicz, Appl. Catal. Gen. 2000, 192, 23–28.
J. Barrault, F. Jerome, Eur. J. Lipid Sci. Technol. 2008, 110, 825–830.
F. Jérôme, Y. Pouilloux, J. Barrault, ChemSusChem 2008, 1, 586–613.
J.-P. Lange, J. Z. Vestering, R. J. Haan, Chem. Commun. 2007, 3488.
R. Jothiramalingam, M. K. Wang, Ind. Eng. Chem. Res. 2009, 48, 6162–
9
9.8%) determines the initial time of the reaction (t = 0). The molar ratio
between the reactants was kept at 24.25. The ratios AcOEt/dioxane was
checked at the beginning (t = 0 h) and after 5 hours of reaction (t = 5 h)
[
[
[
2]
3]
4]
by gas chromatography (Perichrom PR2100 equipped with
a FID
detector and a CP WAX 57 CB column). To perform these analyses a
few drops of reaction mixture were drawn from the Schlenk and diluted in
[5]
[
6]
6
172.
2
mL of 1-propanol (Sigma Aldrich, ACS reagent). The only detected
[
[
7]
8]
D.-W. Lee, Y.-M. Park, K.-Y. Lee, Catal. Surv. Asia 2009, 13, 63–77.
M. Zabeti, W. M. A. Wan Daud, M. K. Aroua, Fuel Process. Technol.
2009, 90, 770–777.
products were methyl acetate and ethanol; therefore the conversion was
calculated through equation 1:
[
[
9]
S. Semwal, A. K. Arora, R. P. Badoni, D. K. Tuli, Bioresour. Technol.
2
Conv. = [(AcOEt/dioxane)t=0 - (AcOEt/dioxane)t=5] / (AcOEt/dioxane)t=0
Pure mass of catalyst that was used to calculate Conv./Mass in Table 1
and 2, was given by equation 2:
011, 102, 2151–2161.
10] Y. M. Sani, W. M. A. W. Daud, A. R. Abdul Aziz, Appl. Catal. Gen. 2014,
70, 140–161.
[11] A. L. de Lima, C. M. Ronconi, C. J. A. Mota, Catal Sci Technol 2016, 6,
877–2891.
12] G. Vicente, M. Martın
05.
4
Pure mass = mass of catalyst x (100% – weight loss below 200 °C)
2
[
́
ez, J. Aracil, Bioresour. Technol. 2004, 92, 297–
Characterizations of the catalysts
3
Thermogravimetric analysis were carried out on
a
STD Q600
[13] M. Di Serio, R. Tesser, L. Pengmei, E. Santacesaria, Energy Fuels 2008,
22, 207–217.
[14] Z. Helwani, M. R. Othman, N. Aziz, J. Kim, W. J. N. Fernando, Appl.
Catal. Gen. 2009, 363, 1–10.
-
1
apparatus from TA® instruments under 20 mL min of N
2
with a ramp
-
1
rate of 5 °C min up to 400 °C.
Diffuse Reflectance Infrared Fourier Transform Spectroscopy
DRIFTS) was performed using a Vertex70 Bruker spectrometer with a
[
15] D. Cornu, H. Guesmi, G. Laugel, J.-M. Krafft, H. Lauron-Pernot, Phys
Chem Chem Phys 2015, 17, 14168–14176.
(
DRIFTS cell (collector from Spectratech). The sample compartment of
[16] M. Kouzu, T. Kasuno, M. Tajika, Y. Sugimoto, S. Yamanaka, J. Hidaka,
Fuel 2008, 87, 2798–2806.
[17] M. Verziu, B. Cojocaru, J. Hu, R. Richards, C. Ciuculescu, P. Filip, V. I.
Parvulescu, Green Chem 2008, 10, 373–381.
the cell was filled with the as-prepared sample (� 20 mg), which was first
3
−1
-1
heated in situ under Ar flow (20 cm .min ) up to 50 °C (5 °C min ) for 2
h. A first spectrum was taken at this temperature. The sample was then
[18] J. M. Montero, P. Gai, K. Wilson, A. F. Lee, Green Chem 2009, 11, 265–
268.
-
1
pretreated to 70 °C (5 °C min ) for 2 h and then the second spectrum
was taken at 70 °C. The spectra of samples pretreatment at 90, 110 and
[
19] H. Jeon, D. J. Kim, S. J. Kim, J. H. Kim, Fuel Process. Technol. 2013,
16, 325–331.
20] M. Kouzu, J. Hidaka, Fuel 2012, 93, 1–12.
1
1
40 °C respectively, were then taken in the same way. All spectra were
[
recorded in situ and were converted into Kubelka–Munk units after
[21] G. Suppes, Appl. Catal. Gen. 2004, 257, 213–223.
subtraction of the spectrum recorded on the dehydrated KBr sample
[22] A. Brito, M. E. Borges, N. Otero, Energy Fuels 2007, 21, 3280–3283.
[
23] W. Xie, X. Huang, H. Li, Bioresour. Technol. 2007, 98, 936–939.
[24] O. Babajide, N. Musyoka, L. Petrik, F. Ameer, Catal. Today 2012, 190,
4–60.
(
Fluka, purity > 99.5%).
1
All the H MAS NMR spectra were recorded at room temperature
5
using a Bruker Advance spectrometer operating in a static field of 11.7 T.
[25] H. Wu, J. Zhang, Q. Wei, J. Zheng, J. Zhang, Fuel Process. Technol.
2013, 109, 13–18.
[26] C. S. Castro, D. Cardoso, P. A. P. Nascente, J. M. Assaf, Catal. Lett.
1
1
The resonance frequency of H were 500.16 MHz. The H chemical shifts
1
were referenced to external standards of tetra-methylsilane (TMS).
H
2
011, 141, 1316–1323.
MAS NMR spectra are performed with a 90° pulse duration of 2.8 µs, a
recycle delay of 5 s, and a scan number of 32. The MAS equipment for
rotation was carefully cleaned with ethanol then dried to avoid spurious
proton signals. From two successive experiments performed in the same
[
27] J. F. P. Gomes, J. F. B. Puna, L. M. Gonçalves, J. C. M. Bordado,
Energy 2011, 36, 6770–6778.
[28] Y. Xi, R. Davis, J. Catal. 2008, 254, 190–197.
[
[
29] Y. Xi, R. J. Davis, J. Catal. 2009, 268, 307–317.
30] S. L. Britton, J. Q. Bond, T. W. Root, Energy Fuels 2010, 24, 4095–4096.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
[
31] H. Hattori, M. Shima, H. Kabashima, Stud. Surf. Sci. Catal. 2000, 130,
507–3512.
[52] R. L. Frost, O. B. Locos, H. Ruan, J. T. Kloprogge, Vib. Spectrosc. 2001,
27, 1–13.
3
[
[
32] A. Ziȩba, A. Pacuła, A. Drelinkiewicz, Energy Fuels 2010, 24, 634–645.
33] D. M. Reinoso, D. E. Damiani, G. M. Tonetto, Appl. Catal. B Environ.
[53] R. L. Frost, E. Mendelovici, J. Colloid Interface Sci. 2006, 294, 47–52.
[54] V. C. (Victor C. Farmer, Infrared Spectra of Minerals, Mineralogical
Society, 1974.
2
014, 144, 308–316.
[
[
[
[
[
34] D. Cornu, L. Lin, M. Daou, M. Jaber, J.-M. Krafft, V. Herledan, G. Laugel,
Y. Millot, H. Lauron-Pernot, Catal. Sci. Technol. 2017, submitted.
35] K. Narasimharao, T. T. Ali, S. Bawaked, S. Basahel, Appl. Catal. Gen.
[55] F. R. Cannings, J. Phys. Chem. 1968, 72, 1072–1074.
[56] R. Prost, Spectrochim. Acta Part Mol. Spectrosc. 1975, 31, 1497–1499.
[57] J. T. Kloprogge, R. L. Frost, L. Rintoul, Phys. Chem. Chem. Phys. 1999,
1, 2559–2564.
[58] M. K. Titulaer, Porous Structure and Particle Size of Silica and
Hydrotalcite Catalyst Precursors: A Thermoporometric Study, Faculteit
Aardwetenschappen Der Rijksuniversiteit Te Utrecht], [Utrecht, 1993.
[59] G. J. Rosasco, J. J. Blaha, Appl. Spectrosc. 1980, 34, 140–144.
[60] E. Knözinger, K.-H. Jacob, S. Singh, P. Hofmann, Surf. Sci. 1993, 290,
388–402.
2
014, 488, 208–218.
36] J. Temuujin, K. Okada, K. J. D. MacKenzie, J. Solid State Chem. 1998,
38, 169–177.
37] J. Temuujin, K. Okada, K. J. MacKenzie, J. Am. Ceram. Soc. 1998, 81,
54–756.
1
7
38] D. Nied, K. Enemark-Rasmussen, E. L’Hopital, J. Skibsted, B.
Lothenbach, Cem. Concr. Res. 2016, 79, 323–332.
[
[
39] F. Jin, A. Al-Tabbaa, Cem. Concr. Compos. 2014, 52, 27–33.
40] C. Roosz, S. Grangeon, P. Blanc, V. Montouillout, B. Lothenbach, P.
Henocq, E. Giffaut, P. Vieillard, S. Gaboreau, Cem. Concr. Res. 2015,
[61] C. Chizallet, G. Costentin, M. Che, F. Delbecq, P. Sautet, J. Am. Chem.
Soc. 2007, 129, 6442–6452.
[62] M. Bailly, C. Chizallet, G. Costentin, J. Krafft, H. Lauronpernot, M. Che,
J. Catal. 2005, 235, 413–422.
7
3, 228–237.
[
[
41] D. R. M. Brew, F. P. Glasser, Cem. Concr. Res. 2005, 35, 85–98.
42] A. G. Kalampounias, N. Bouropoulos, K. Katerinopoulou, S. N.
Yannopoulos, J. Non-Cryst. Solids 2008, 354, 749–754.
43] B. L. Phillips, P. C. Burnley, K. Worminghaus, A. Navrotsky, Phys.
Chem. Miner. 1997, 24, 179–190.
[63] S. Coluccia, L. Marchese, S. Lavagnino, M. Anpo, Spectrochim. Acta
Part Mol. Spectrosc. 1987, 43, 1573–1576.
[64] V. C. Farmer, Infrared Spectra of Minerals, Mineralogical Society,
London, 1974.
[65] M. Falk, Spectrochim. Acta Part Mol. Spectrosc. 1984, 40, 43–48.
[66] B. A. Kolesov, Am. Mineral. 2006, 91, 1039–1048.
[67] J. P. Yesinowski, H. Eckert, G. R. Rossman, J. Am. Chem. Soc. 1988,
110, 1367–1375.
[68] J. Jiao, S. Altwasser, W. Wang, J. Weitkamp, M. Hunger, J. Phys. Chem.
B 2004, 108, 14305–14310.
[69] Y. Jiang, J. Huang, W. Dai, M. Hunger, Solid State Nucl. Magn. Reson.
2011, 39, 116–141.
[70] P. Clausen, W. Andreoni, A. Curioni, E. Hughes, C. J. G. Plummer, J.
Phys. Chem. C 2009, 113, 12293–12300.
[71] A. Krysztafkiewicz, R. Werner, L. K. Lipska, T. Jesionowski, Colloids
Surf. Physicochem. Eng. Asp. 2001, 182, 65–81.
[72] F. Leydier, C. Chizallet, A. Chaumonnot, M. Digne, E. Soyer, A.-A.
Quoineaud, D. Costa, P. Raybaud, J. Catal. 2011, 284, 215–229.
[73] J. Blanchard, J.-M. Krafft, C. Dupont, C. Sayag, T. Takahashi, H.
Yasuda, Catal. Today 2014, 226, 89–96.
[
[
[
[
[
[
[
[
[
44] A. Pedone, F. Palazzetti, V. Barone, J. Phys. Chem. C 2017, DOI
1
0.1021/acs.jpcc.7b00708.
45] I. M. Atadashi, M. K. Aroua, A. R. Abdul Aziz, N. M. N. Sulaiman, Renew.
Sustain. Energy Rev. 2012, 16, 3456–3470.
46] A. F. Lee, J. A. Bennett, J. C. Manayil, K. Wilson, Chem Soc Rev 2014,
4
3, 7887–7916.
47] C.-R. Chang, Z.-Q. Huang, J. Li, Wiley Interdiscip. Rev. Comput. Mol.
Sci. 2016, 6, 679–693.
48] H. Petitjean, H. Guesmi, H. Lauron-Pernot, G. Costentin, D. Loffreda, P.
Sautet, F. Delbecq, ACS Catal. 2014, 4, 4004–4014.
49] H. C. Greenwell, S. Stackhouse, P. V. Coveney, W. Jones, J. Phys.
Chem. B 2003, 107, 3476–3485.
50] M. J. Kim, S. M. Park, D. R. Chang, G. Seo, Fuel Process. Technol.
2
010, 91, 618–624.
51] M. Tonelli, F. Martini, L. Calucci, E. Fratini, M. Geppi, F. Ridi, S.
Borsacchi, P. Baglioni, Dalton Trans 2016, 45, 3294–3304.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201700139
FULL PAPER
Entry for the Table of Contents
FULL PAPER
#
Two kinds of active sites exist, that
Longfei Lin, Damien Cornu , Maya
are
created
from
the
water
Mounir Daou, Cyril Domingos, Virginie
Herledan, Jean-Marc Krafft, Guillaume
Laugel, Yannick Millot, and Hélène
Lauron-Pernot*
coordinated to magnesium located on
the edge of the clay-like particles or in
the defects present in the silicate
layer, respectively, which could
stabilise methanol deprotonated by
basic Mg-OH groups, activate ester or
help the departure of the alkoxyl
moiety .
Page No. – Page No.
Role of water on the activity of
magnesium silicate for
transesterification reaction
This article is protected by copyright. All rights reserved.