Fe4(SiW12O40)3-catalyzed glycerol acetylation: Synthesis of bioadditives by using highly active Lewis acid catalyst

Article abstract of DOI:10.1016/j.molcata.2016.03.003

In order to investigate the effect of Lewis acidic metals on the acid properties of Keggin heteropolyacids catalysts, a set of metal-exchanged HPAs (i.e. M3/nPW12O40, M3/nPMo12O40 and M4/nSiW12O40 (M?=?Cu, Co and Mn, n?=?2; M?=?Fe, n?=?3) were synthesized and their activity on glycerol esterification with acetic acid was evaluated. This procedure avoids the use of corrosive Br?nsted acid catalysts. Regardless of the heteropoly anion, the activity of catalysts in terms of metal cations followed the sequence: Fe³⁺>?Cu²⁺>?Mn²⁺>?Co²⁺. It has been found that among the HPAs, only H4SiW12O40 has the acidic protons that can be successfully exchanged with Fe³⁺ cations, resulting in Fe4(SiW12O40)3 salt, the most active and selective catalyst. The highest conversion (ca.99.9%) was achieved in Fe4(SiW12O40)3-catalyzed esterification reactions, along with the highest selectivity for di and triacetyl glycerol (ca.55 and 42%, respectively). Effects of temperature, stoichiometry of reactants, concentration and nature of catalysts were assessed. A kinetic study of Fe4(SiW12O40)3 or H4SiW12O40-catalyzed reactions was conducted and the activation energy determined.

Full text of DOI:10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Fe4(SiW12O40)3-catalyzed glycerol acetylation: Synthesis of
bioadditives by using highly active Lewis acid catalyst
Márcio José Da Silva∗, Natalia Aparecida Liberto, Lorena Cristina De Andrade Leles,
Ulisses Alves Pereira
Chemistry Department, Federal University of Vic¸ osa, Vic¸ osa, Minas Gerais State 36570-000, Brazil
a r t i c l e i n f o
a b s t r a c t
Article history:
In order to investigate the effect of Lewis acidic metals on the acid properties of Keggin heteropoly-
acids catalysts, a set of metal-exchanged HPAs (i.e. M3/nPW12O40, M3/nPMo12O40 and M4/nSiW12O40
(M = Cu, Co and Mn, n = 2; M = Fe, n = 3) were synthesized and their activity on glycerol esterification with
acetic acid was evaluated. This procedure avoids the use of corrosive Brønsted acid catalysts. Regard-
less of the heteropoly anion, the activity of catalysts in terms of metal cations followed the sequence:
Fe3+> Cu2+> Mn2+> Co2+. It has been found that among the HPAs, only H4SiW12O40 has the acidic protons
that can be successfully exchanged with Fe3+ cations, resulting in Fe4(SiW12O40)3 salt, the most active
and selective catalyst. The highest conversion (ca.99.9%) was achieved in Fe4(SiW12O40)3-catalyzed
esterification reactions, along with the highest selectivity for di and triacetyl glycerol (ca.55 and 42%,
respectively). Effects of temperature, stoichiometry of reactants, concentration and nature of catalysts
were assessed. A kinetic study of Fe4(SiW12O40)3 or H4SiW12O40-catalyzed reactions was conducted
and the activation energy determined.
Received 2 October 2015
Received in revised form 27 February 2016
Accepted 1 March 2016
Available online xxx
Keywords:
heteropolyacid salts
glycerol
bioadditives
glyceryl esters
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
sively studied, as a result of their ready availability, high thermal
stability, strong acidity and ability to undergo fast reversible multi-
Due to economic and environmental reasons, the replacement
of conventional mineral acid catalysts with acid catalysts has stim-
ulated global research efforts. The great flaws of homogenous
Brønsted acid catalysts include difficulty in handling, corrosive
nature, high toxicity and contamination of products, that require
expensive purification methods, and large investment for the neu-
tralization of residues and treatment of effluents.
Heteropolyacids (HPAs) are attractive alternative, because they
may act as eco-friendly and effective solid acid catalysts [1]. There
are two pivotal facts that justify this proposition: the first one is
their unique chemical-physical properties, coupled with their rela-
tively easy preparation; the second one, their versatile composition,
which allows a suitable tailoring of the catalyst constitution, by
introducing elements needed to achieve desired properties [2].
Since Berzelius described HPAs for the first time, a large amount
of research work has been carried out aiming at exploring their
countless usefulness in different fields of chemistry [3–6]. The
Keggin-type HPAs class of polyoxometalates has been comprehen-
electron transfers, making them active catalysts in redox processes
[7–9]. These HPAs contain a central atom (i.e., P or Si) in a basic octa-
hedral structural unit surrounded by metal-oxygen tetrahedral (W
or Mo). The heteropolyanions form an extremely stable and firm
framework, typically with delocalized superacidic protons [10].
The well-defined and adjustable HPAs properties allow the
building of both homogeneous and heterogeneous acid catalysts
[11–13]. Although supporting HPAs on high surface area solids has
been a method applied to synthesize heterogeneous catalysts, the
leaching of catalyst remains a great challenge. On the other hand,
the conversion of HPAs to insoluble salts has been revealed to be
an effective strategy to obtain efficient catalysts [14,15].
Highlighted examples are cesium HPA salts (i.e.,
Cs2.75H0.25PW12O40), which are solids with a high surface area and
have successfully catalyzed reactions in the gas or liquid phase
[16,17]. HPA salt catalysts possess inert anions that do not cause
side- reactions, which always occur with chloride, sulfate or nitrate
anions.
Among various acid-catalyzed organic reactions of industrial
interest, glycerol esterification has become attractive as a result
of its availability and economic viability, as the biodiesel industry
generates glycerol in excess as a by-product, making glycerol an
∗
Corresponding author.
E-mail address: silvamj2003@ufv.br (M.J. Da Silva).
http://dx.doi.org/10.1016/j.molcata.2016.03.003
1381-1169/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
2
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
100
HO
OAc
OH
90
80
HOAc
HOAc
OAc
70
e
d
AcO
HO
OAc
60
50
40
30
20
(a) Co2SiW12O40
(b) Mn2SiW12O40
(c) H4SiW12O40
(d) Cu2SiW12O40
(e) Fe4(SiW12O40)3
OH
OAc
c
b
a
HOAc
HOAc
AcO
OAc
OAc
500
1000
1500
2000
2500
3000
3500
4000
Wavenumber (cm-1)
Fig. 1. Glycerol esterification reaction with acetic acid.
Fig. 3. FT-IR/ATR profiles of metal-exchanged H4SiW12O40 catalysts.
100
90
80
70
60
50
40
30
20
e
d
c
b
(a) Cu3(PW12O40)2
(b) Co3(PW12O40)2
(c) H3PW12O40
(d) Mn3(PW12O40)2
(e) FePW12O40
a
500
1000
1500
2000
2500
3000
3500
4000
Wavenumber (cm-1)
Fig. 2. FT-IR/ATR profiles of metal-exchanged H3PW12O40 catalysts.
Fig. 4. FT-IR/ATR profiles of metal-exchanged H3PMo12O40 catalysts.
2. Experimental section
abundant and low-cost raw material, which has assumed pivotal
importance as key-feedstock for biorefineries [18]. Among the dif-
ferent approaches proposed to valorize glycerol, its conversion to
bio-additives has occupied prominent place [19,20].
2.1. Chemicals
All the received chemicals were used without any further
treatment. Acetic acid (99.9 wt. %) and glycerol (99.5 wt. %)
were obtained from Sigma-Aldrich. The hydrated heteropolyacid
(i.e., H3PW12O40·n H2O; H3PMo12O40·n H2O and H4SiW12O40·n
H2O; (99.9 wt. %)), were purchased from Sigma-Aldrich. Metal
nitrates Co(NO3)2·6 H2O (98 wt. %), Mn(NO3)2·4 H2O (98 wt. %) and
Cu(NO3)2·3H2O (98 wt. %) were supplied by Dinamica Ltda (Brazil).
Fe(NO3)3·9 H2O (98 wt. %) was obtained from Vetec(Brazil).
Glycerol esterification has been carried out in the presence of
solid acid including acid resins, zeolites and nanotubes [21,22].
However, this always requires a laborious synthesis to obtain
efficient solid catalysts. Moreover, the unavoidable leaching under-
gone by the HPAs solid supported provoked by a highly polar
reaction medium is an aspect that still compromises the application
of some heterogeneous catalysts to glycerol esterification [23]. An
alternative developed by us was to apply soluble HPA catalysts [24].
Although soluble HPA can be recovered from the reaction medium
by a distillation process stripping all liquid components from the
reaction mixture. While distillation is an energy-intensive step,
it is always required to separate products and reactants in these
processes.
In this study, a HPA salt-catalyzed glycerol esterification with
HOAc was described. A screening of metal cations and precur-
sor heteropolyacids that resulted in more efficient catalysts was
performed. In addition, various reaction parameters aiming to opti-
mize the reaction conditions to obtain the maximum activity and
selectivity at a minimum catalyst load were investigated.
2.2. Synthesis of M3(PW12O40) x, M3(PMo12O40) x and
M4(SiW12O40) x catalysts(M = Co, Cu or Mn, x = 2; Fe, x = 3)
All heteropolyacid salts catalysts were synthesized by total sub-
stitution of acidic hydrogens in the H3PW12O40, H3PMo12O40 and
H4SiW12O40 precursors by the metal cations of nitrate hydrated
salts (i.e., Fe(NO3)3, Co(NO3)2, Mn(NO3)2, Cu(NO3)2). In order to
synthesize the catalysts, an aqueous solution containing a stoi-
chiometric amount of metal salt was slowly added to an aqueous
solution of heteropolyacid. This process was performed with stir-
ring and heating to 333 K temperature for 3 h. Subsequently, there
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
2.4. Catalytic tests
3
Typically, a 25 mL three-necked glass flask, equipped with a
sampling system and a reflux condenser was charged with glycerol
(ca. 46.72 mmol), different volumes of acetic acid and an adequate
amount of HPA salt catalyst (ca.0.06 mmol). The reaction was per-
formed using magnetic stirring at a temperature of 333 K for over
2 h.
The effects of main reaction parameters (i.e. reactants stoi-
chiometry, temperature, catalyst concentration) were assessed.
The HOAc: glycerol molar ratio was varied over four different pro-
portions (i.e.3:1; 6:1; 9:1 and 12:1 respectively). Control reactions
were performed for each molar proportion in the absence of cata-
lyst. The activity of the most active HPA salt to the activity of metal
nitrate and heteropolyacid precursor alone was also compared.
2.5. Reaction monitoring and products identification
The reaction progress was periodically monitored by
taking aliquots and analyzing in
matograph fitted with Carbowax 20 M capillary column
flame ionization detector.
a Varian 450 gas chro-
a
(0.25 ␮m × 0.25 mm × 30 m) and
a
Matching the GC peak areas of glycerol and products (i.e., mono,
di and triacetyl glycerol) to the corresponding calibration curve
provided the conversion and selectivity.
The analysis in a Shimadzu GC2010 gas chromatograph coupled
with a MS-QP5050A mass spectrometer identified the products.
The MS detector was operated in the EI mode at 70 eV, with a
scanning range of m/z 50–450.
3. Results and Discussion
Fig. 5. XRD patterns of bulk HPW, HPMo, HSiW and their respective Fe3+ salts cat-
alysts.
3.1. General discussion
Glycerol esterification is a reversible Lewis or Brønsted acid-
catalyzed reaction, where mono, di or tri acetyl glycerol and water
are the products obtained (Fig. 1).
Generally, in acid-catalyzed glycerol acetylation; the acetic acid
carbonyl group is protonated by H + cations (i.e., Brønsted acid cat-
alysts) or polarized by interacting with Mn + cations (i.e., Lewis acid
catalysts), favoring the attack by the less hindered hydroxyl group
of glycerol. This results in the formation of water and terminal
ester. Subsequently, similar steps occur consecutively to generate
diacetyl and triacetyl glycerol.
was slow evaporation to dryness of the corresponding HPA salts in
stoichiometric yield [25].
The literature assumes that this preparation method assures
the composition of the metal (Mn+
)
salts of PW12O40
3−
,
3−
PMo12O40 and SiW12O404−, as M3/nPW12O40, M3/nPMo12O40 or
M4/nSiW12O40, respectively (i.e. M = Fe3+, Co3+, Cu2+ and Mn2+) [26].
In this study, we sought the combination of these two acid fea-
tures in a single catalyst. For this reason, catalyst containing Lewis
acid site (i.e., Fe3+, Co2+, Cu2+ or Mn2+) and Brønsted acid sites (i.e.
hydrated PW12O403- anion) was synthesized.
2.3. Characterization of catalysts
Powder X-ray diffraction patterns (XRD) were recorded on a
X-ray Diffraction System model X’ PertPRO (PA Analytical) diffrac-
tometer using Ni filtered Cu-k␣ (␭ = 1.78890 Å) radiation, operating
at 40 kV and 30 mA. The measurements were carried out in step of
0.04◦ with a counting time of 0.5 s in the 2␪ range of 10–80.
FT-IR/ATR spectra of the HPA catalysts were recorded on an
FT-IR Varian 660 spectrometer with reflectance accessory utiliz-
ing KBr plates under ambient conditions. TG-DTA measurements
were carried out in a Simultaneous Thermal Analyzer (STA) Perkin
Elmer, model 6000. The experiments were performed under nitro-
3.2. Characterization of M3(PW12O40) x, M3(PMo12O40) x and
M4(SiW12O40) x catalysts (M = Co, Cu or Mn, x = 2; M = Fe, x = 3)
Salts of HPAs were prepared by the addition of stoichio-
metric quantities of the metal nitrate aqueous solutions to
solutions of the heteropolyacids (i.e. H3PW12O40, H3PMo12O40
and H4SiW12O40), which were stirred and heated (ca.333 K) for
3 h, and then evaporated to dryness [26].
The FT-IR spectra of metal-exchanged HPA catalysts were
similar to those presented by the precursor heteropolyacids
(Figs. 2–4). The typical bands were placed at the same wave num-
bers reported for Keggin anions of unsubstituted heteropolyacids.
The main absorption bands corresponding to stretching vibra-
tion of molybdenum (or tungsten)-oxygen and phosphorus (or
silicon)-oxygen bond sextant in the H3PMo12O40, H3PMo12O40
and H3PMo12O40 were also detected in the respective salts. For
instance, the molybdenum heteropolyacid catalysts displayed the
gen, utilizing10–50 mg samples and a heating rate of 10 Kmin−1
The studied temperature range was 303–873 K, with an error devi-
ation of 0.1 K.
Catalysts acidity was measured by potentiometric titration, in
accordance with the procedure reported by Pizzio et al. [27]. The
electrode potential variation was measured with a potentiometer
(i.e. Bel, model W3B). Typically, a solution containing an ade-
quate amount of HPA dissolved in acetonitrile was titrated with
n-butylamine solution in toluene (ca. 0.025 molL−1).
.
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
4
Table 1
GC-MS analyses of gas-phase headspace samples collected after heating to a temperature of 150 degrees (ca. 40 min).a
Compound
GC-MS chromatogram
FeNO3.·9 H2O
FePMo12O40
FePW12O40
Fe4(SiW12O40
)
3
Diffractogram of peak at 8.2 min time retention
a
Analysis conditions: ultra GC-2010 Shimadzu GC-MS equipment, carbowax capillary column, splitless injector temperature (ca.100 ◦C), GC temperature program
(40 ◦C/3 min, heating rate 5 ◦C/ min−1 until 60 ◦C), MS detector (i.e. 70 eV).
following bands: ((␯ (P-O) = 1055 cm−1, ␯ (Mo = O) = 955 cm−1
,
Therefore, these bands suggest that the Keggin anion structure
is present in the HPA salt catalysts, which remains intact even after
a “possible” partial or total replacement of the acidic protons by
metal cations.
The analysis of FT-IR spectra provided another crucial piece
of information. The IR spectra of H3PW12O40, H3PMo12O40 and
as
as
␯
as
(Mo-O-Mo) = 890 cm−1, ␯ (Mo-O-Mo) = 760 cm−1)). The sil-
as
icon heteropolyacids FT-IR spectra displayed the bands with
wave numbers: ((␦ (Si-O) = 539, (␯ (W-O-W) = 779 cm−1, (␯
as
(W-O-W) = 879 cm−1, ␯as (Si-O) = 925 cm−1, ␯ as (W-O) = 979 cm−a1s
,
␯
as
(W-O) = 1018 cm−1)). Finally, tungsten heteropolyacids have
H4SiW12O40 heteropolyacid precursors showed
a
strong and
characteristic bands at wave numbers: ((␯ (P-O) = 1075,
␯
as
well defined band around 1610–1720 cm−1 due to H5O2 cation
+
(W = O) = 983 cm−1
W) = 790 cm−1)).
,
␯
(W-O-W) = 898 cm−1 and
␯
(W-O-
as
as
(H2O/H+/OH2)[28], however, for the iron (III) HPA salts in the
Fe4(SiW12O40
) catalyst, this band has the lowest intensity. When
3
Fig. 6. Potentiometric titration with n-butylamine of the HPA catalysts.
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
5
Table 2
Amount of H + cations present in the heteropolyacids and their respective salts estimated through potentiometric titration.
HPA catalyst
Theoretical amount of H+ cations
Amount of H+ cations experimentally determined
Heteropolyacid salt composition
H3PMo12O40
H4SiW12O40
H3PW12O40
FePW12O40
Fe4SiW12O40
FePMo12O40
3
4
3
0
0
0
2.7
4.2
3.1
1.5
0.0
2.9
–
–
–
Fe0.5H1.5PW12O40
Fe4.0(SiW12O40
Fe0.0H3PMo12O40
)
3
(c)
(a)
(b)
Fig. 7. TGA curves of silicon (a), tungsten (b) and molybdenum (c) heteropolyacids and their respective iron (III) salts.
Izumi et al. synthesized heteropolyacid salts exchanged with Sn4+
cations, they attributed the decrease in the intensity of this band to
the partial or total removal of protons [25]. Therefore, because the
same observation was noticed in this study, it could mean that Fe3+
cations fully or partially substituted the acidic protons of HPAs. In
the catalyst salt was carried out; upon heating there was no NOx.
After heating the salt samples (ca. 20 mg) in sealed glass flask to a
temperature of 150 degrees for 40 min, a sample (ca. 1.0 ␮L) of gas
phase headspace was collected using a gas tight syringe and was
analyzed by GC-MS.
Firstly, to check the efficiency of this method, the thermal
decomposition procedure with the Fe(NO3)3·9H2O salt was carried
out, a gas-phase headspace sample was collected and analyzed in
GC-MS mass spectrometer. The GC-MS chromatogram exhibited a
peak corresponding to the HNO2 (47 m/z) (Table 1). Afterwards,
GC-MS analyses of gas-phase headspace samples of all Fe3+ het-
eropolyacid salts were submitted to the same procedure. In all the
samples, a peak with 8.2 min retention time was found. It pro-
vided 47 m/z peak, attributed to the HNO2, a product generated
by thermal decomposition of iron nitrate [30].
+
this case, some or all of the protons present in the form of H5O2
could have been exchanged by Fe3+ cations. However, more exper-
imental data are required to ensure that this fact really occurred in
this study.
Another important observation is that the several FT-IR spectra
showed in Figs. 2–4 displayed band(s) at ∼1350 cm−1, which could
also be attributed to nitrates anions [29]. Indeed, the thermal stabil-
ity of nitrate anions is limited to a temperature of 130 degrees. TG
and DSC analyses of the hydrated Fe(NO3)3 reported in literature
emphasis that the most significant loss of mass occurs between 109
and 124 degrees (i.e., the loss of HNO3 molecules) [30]. Thus, the
simple evaporation of solution carried out during the synthesis of
the HPA catalysts (ca. 100 degrees) could enable the decomposition
of nitrate anions present in the Fe(NO3)3 precursor.
If we considered that GC-MS peak area of HNO2 corresponded
to total decomposition of 20 mg Fe(NO3)3, (ca. 0.05 mmol of
nitrate anions), we can estimate the amount of nitrate anions
remaining in the HPA salts; 0.015, 0.0035 and 0.002 mmol, for
−
FePW12O40, Fe4(SiW12O40
) and FePMo12O40, respectively. Ana-
3
To confirm this hypothesis (i.e., the presence of NO3 anions
lyzing the infrared spectra of these salts once more, we verified
in the HPA salts), a GC-MS analysis of the gas-phase headspace of
Table 3
HPA-catalyzed glycerol esterification with acetic acid.a
Run
Selectivity (%)
Catalyst
Conversion
Mono-acetylglycerol
di-acetylglycerol
Tri-acetylglycerol
1
2
3
4
5
6
7
8
H3PW12O40
Co3(PW12O40
FePW12O40
62
50
82
66
68
52
23
57
25
36
100
28
100
55
61
68
80
48
69
63
74
89
28
88
83
42
80
24
50
33
32
21
48
32
38
26
11
72
12
17
53
20
69
48
63
0
0
3
0
0
0
0
0
0
0
5
0
7
2
4
)
2
Mn3(PW12O40
)
2
Cu3(PW12O40
H3PMo12O40
Co3(PMo12O40
FePMo12O40
Mn3(PMo12O40
Cu3(PMo12O40
H4SiW12O40
Co4(SiW12O40
Fe4(SiW12O40
Mn4(SiW12O40
Cu4(SiW12O40
)
2
)
2
9
)
2
10
11
12
13
14
15
)
2
)
2
)
3
)
2
)
2
a
Reaction conditions: glycerol (47.92 mmol); HOAc (143.76 mmol); catalyst (0.06 mol %), temperature (333 K); time (8 h).
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
6
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
(a)
H3PW12O40
Co3(PW12O40
)
2
80
90
75
60
45
30
15
0
FePW12O40
Mn2(PW12O40)2
Cu2(PW12O40
(b)
70
H3PMo12O40
)
2
Co3(PMo12O40
)
2
60
50
40
30
20
10
0
FePMo12O40
Mn3(PMo12O40
)
2
Cu3(PMo12O40
)
2
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8
Time (h)
time (min)
(c)
100
80
60
40
20
0
H4SiW12O40
Co2SiW12O40
Fe4(SiW12O40)3
Mn2SiW12O40
Cu2SiW12O40
0
1
2
3
4
5
6
7
8
Time (h)
Fig. 8. Kinetic curves of glycerol esterification with HOAc in the presence of HPAs or its transition metal salts.
Reaction conditions: glycerol (47.92 mmol); HOAc (143.76 mmol); HPA catalyst (0.03 mol% of H+ cations), HPA salt catalyst (0.03 mmol); temperature (333 K); time (8 h).
that indeed, only the FePW12O40 FT-IR spectrum shows a strong
The FeSiW catalyst (Fe4(SiW12O40) 3) show the same peaks
with similar crystallinity. However, an additional peak at 2␪ = 31.7◦
was observed. All their HPA salts displayed this same peak, which is
a characteristic diffraction peak of Fe3 + cations and Villabrille et al.
found the same characteristic peak when incorporating iron to the
H3PMo12O40 heteropolyacid [32].
Dias et al. stated that substituting H + cations with metal cations
(i.e. Ag2 + ) may improve the crystallinity of HPAs [33]. They sug-
gested that exchange of protons present in the secondary Keggin
structure in the form of H5O2+ dihydronium ions for hydrated
metal cation may provoke a contraction of unit cell. Borghese
et al. also observed the same behavior when they synthesized Ag-
exchanged H4SiW12O40 catalysts [34,35].
adsorption band at a wave number of ∼1350 cm−1
.
The literature describes that Fe(NO3)3 undergoes water loss
within a temperature range varying from 60 to 100 degrees [30].
The loss of HNO3 starts at 109 degrees. Therefore, it can be con-
cluded that the temperature control played a crucial role on the
synthesis of HPA salts. When the temperature did not exceed ca.
109 degrees, it can be assumed that the precursor Fe(NO3)3 was
intact to react with HPA and provide the iron-exchanged salt. In
this case, the nitric acid formed probably remained impregnated in
the solid salt. This fact was apparently more pronounced on syn-
thesis of FePW12O40. However, if it was equal or higher than 109
degrees, the nitrate anions could have been partially decomposed,
which possibly happened during the synthesis of FePMo12O40 and
Fe4(SiW12O40)3. Nevertheless, as will be shown in the later part of
To measure the acidity of the catalysts, potentiometric titration
of HPAs with n-butylamine (Fig. 6) was conducted. The measure-
ment of initial electrode potential (Ei) allows the classification
of the acidity strength of acid sites as; Ei> 100 mV (very strong
sites), 0< Ei< 100 mV (strong sites), −100< Ei< 0 (weak sites) and
Ei< −100 mV (very weak sites) [27].
−
the study, the NO3 anions did not affect the catalytic activity of
HPA salts.
Fig. 5 shows the XRD patterns of the heteropolyacids and their
respective iron salts. The parent HSiW (i.e., H4SiW12O40) exhibited
all markers of a typical X-ray diffractogram of a body-centered cubic
secondary structure of Keggin anions, with characteristic diffrac-
tion peaks at angles 8.0, 10.0, 10.2 and 32.2 [31].
As described in literature, Keggin heteropolyacids typically
exhibit high acidity strength. Misono et al. reported that the change
in chemical composition (i.e., change of the addenda atom, Si by P)
did not significantly affect their acidity [36]. Although solvents may
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
7
100
80
60
40
20
0
100
HOAc:Glicerol = 3:1
HOAc:Glicerol = 6:1
HOAc:Glicerol = 9:1
HOAc:Glicerol = 12:1
HOAc-glicerol = 3:1
HOAc-glicerol = 6:1
HOAc-glicerol = 9:1
HOAc-glicerol = 12:1
80
60
40
20
0
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8
Time (h)
Time (h)
(b) FePMo12O40-catalyzed reactions
(a) FePW12O40-catalyzed reactions
40
35
100
80
1:12
1:9
30
25
20
15
10
5
1:6
1:3
60
HOAc-glicerol = 3:1
HOAc-glicerol = 6:1
40
HOAc-glicerol = 9:1
HOAc-glicerol = 12:1
20
0
0
0
1
2
3
4
5
6
7
8
9
0
50
100
150
200
250
time (min)
Time (min)
(d) blank-reactions
(c) Fe4(SiW12O40)3-catalyzed reactions
Fig. 9. Effect of reactants molar ratio on reaction rates of FePW12O40, FePMo12O40 or Fe4(SiW12O40) 3- catalyzed glycerol esterification with HOAc.
Reaction conditions: glycerol (47.92 mmol); catalyst concentration, FePW12O40 (0.06 mmol); FePMo12O40 (0.06 mmol); Fe4(SiW12O40) 3 (0.015 mmol), temperature
(333 K).
affect pka values, the following acidity trend has been observed
H3PW12O40> H3PMo12O40> H4SiW12O40. The values of elec-
monoacetylglycerol
diacetylglycerol
trode initial potential (Ei) measured showed herein that in our case,
10
8
∼
the acidity order was H3PW12O40 H3PMo12O40> H4SiW12O40
=
(Fig. 6a–c).
The potentiometric titration curves revealed that the efficiency
in the replacement of acidic protons was different for each cata-
lyst. Comparing the titration curves of heteropolyacid salts against
it parents, it can be concluded that only the Fe3+ salt derivative
from silicon heteropolyacid (i.e., H4SiW12O40) seems to have been
successfully obtained (Fig. 6c–f). A comparison of Fig. 6b with 6e
shows that no protons were removed from H3PMo12O40. Differ-
ently, the analysis of Fig. 6a and 6d suggest that the protons of the
heteropolyacid have been partially removed.
6
4
2
Although the measurements of initial electrode potential (Ei)
of solutions of the heteropolyacid salts continue to indicate that
they can be classified as stronger acids (i.e. Ei> 100 mV suggest the
presence of very strong sites), the total or partial removal of protons
slightly affects its acidity strength. When comparing the Ei values of
two heteropolyacids that apparently have Fe3 + cations introduced
in their structure (i.e. H3PW12O40 and mainly H4SiW12O40), it was
noticed that there was a slight decrease in Ei values (ca. Ei = 660 and
Ei = 500; H3PW12O40 and FePW12O40, respectively; ca. Ei = 580 and
Ei = 480; H4SiW12O40 and Fe4(PW12O40)3, respectively).
0
1:3
1:6
1:9
1:12
glycerol to HOAc molar ratio
Fig. 10. Proportion of mono to diacetyl glycerol obtained in blank-reactions.
Reaction conditions: glycerol (47.92 mmol); temperature (333 K), 4 h.
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
8
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
triacetylglycerol
diacetylglycerol
monoacetylglycerol
triacetylglycerol
diacetylglycerol
monoacetylglycerol
105
(a)
(b)
100
90
75
60
45
30
15
80
60
40
20
0
0
1:3
1:6
1:9
1:12
1:3
1:6
1:9
1:12
glycerol: HOAc molar: ratio
FePMo12O40-catalyzed reactions
glycerol: HOAc molarratio
FePW12O40-catalyzed reactions
triacetylglycerol
diacetylglycerol
monoacetylglycerol
triacetylglyceride
diacetylglyceride
(d)
(c)
monoacetylglyceride
100
100
90
80
70
60
50
40
30
20
10
0
80
60
40
20
0
1:3
1:6
1:9
1:12
1:3
1:6
1:9
1:12
glycerol: HOAc molar ratio
H4SiW12O40-catalyzed reactions
glycerol: HOAc molar ratio
Fe4(SiW12O40)3-catalyzed reactions
Fig. 11. Effect of molar ratios on HPA-catalyzed glycerol esterification with HOAc. aReaction conditions: glycerol (47.92 mmol); catalysts concentration, FePW12O40
(0.06 mmol); FePMo12O40 (0.06 mmol); H4SiW12O40 (0.015 mmol); Fe4(SiW12O40)3, (0.015 mmol)), temperature (333 K); reaction time (8 h).
An essential data can be obtained by numerically determining
the first derivative of the titration curves; an indicative of the level
of H+ cations removal (Table 2). Table 2 shows that the procedure
used to incorporate Fe3+ cations into the heteropolyacid catalysts
was well accomplished only for the H4SiW12O40. This result is in
agreement with data obtained from FT-IR analyses, which showed
The TGA curves obtained by thermal analysis of the iron HPA
catalysts demonstrates that the FePMo12O40 possess a completely
different behavior from the other iron salts (i.e., FePW12O40 and
Fe4(SiW12O40) 3). In fact, the profile of the TGA curve almost
equals the H3PMo12O40 precursor heteropolyacid. It suggests that
iron cations were not included in molybdenum HPA structure. Con-
versely, the TGA curves obtained for the FePW12O40 and mainly
Fe4(SiW12O40) 3 salts were noticeably different from those of their
precursor HPA, which means that the presence of Fe(III) probably
changed the TGA curves of HPAs.
For these precursors, any significant loss of weight occurred
within the temperature range of 30 to 200 ◦C. In contrast, in the
metal substituted catalysts, weight loss occurred continuously
throughout the whole duration of heating. Southward et al. inves-
tigated the behavior of tungsten heteropolyacids when heated
to different temperatures [39]. These authors verified that the
replacement of protons by the metal cations improved the ther-
mal stability of HPA catalysts. In agreement with these authors,
besides the dehydration steps, the heating to high temperatures of
heteropolyacids collapses its structure resulting in the WO3 and
POx oxides.
+
that the band corresponding to the H2O5 cations was very small
only for the Fe4(SiW12O40
) salt (Fig. 3). In addition, the titration
3
curves reiterate that no proton was removed from H3PMo12O40
heteropolyacid (i.e. comparison of 3b against 3e, Fig. 6).
The TGA curves show that the precursor heteropolyacids
undergo a significant weight loss in the temperature range from
30 to 200 ◦C (Fig. 7a and b) and between 30 and 100 ◦C (Fig. 7c).
This loss is as a result of physisorbed water. The DSC curves omitted
herein show an endothermic peak at these temperatures. The slight
decrease in weight which occurred at 150 to 600 ◦C may be as a
result of the loss of water molecules of the hydrated heteropolyacid
catalyst [37]. The thermal decomposition of Keggin anion, which
normally occurs after 400 ◦C was not observed by the TGA analysis
[38].
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
9
100
100
80
60
40
20
0
monoacetylglycerol
diacetylglycerol
triacetylglycerol
monoacetylglycerol
diacetylglycerol
triacetylglycerol
80
60
40
20
0
0
30
60
90
120
150
180 210 240
0
30
60
90
120
150
180
210
240
time (min)
time (min)
(a) 3:1 HOAc to glycerol proportion
(b) 6:1 HOAc to glycerol proportion
100
100
monoacetylglycerol
monoacetylglycerol
diacetylglycerol
triacetylglycerol
diacetylglycerol
triacetylglycerol
80
80
60
40
20
0
60
40
20
0
0
30
60
90
120 150
180 210 240
0
30
60
90
120
150
180
210
240
time (min)
time (min)
(d) 12:1 HOAc to glycerol proportion
(c) 9:1 HOAc to glycerol proportion
Fig. 12. Effect of molar ratios on the HPA-catalyzed glycerol esterification with HOAc.
Reaction conditions: glycerol (47.92 mmol); concentration of catalysts (FePW12O40 (0.06 mmol); FePMo12O40 (0.06 mmol); H4SiW12O40 (0.015 mmol); Fe4(SiW12O40) 3
(0.015 mmol)), temperature (333 K); reaction time (4 h).
As demonstrated by the potentiometric titration curves, FT-IR
and thermal analysis, only the Fe4(SiW12O40) 3 salt was suc-
cessfully synthesized in this study. Therefore, the higher stability
observed in the TGA curves of this catalyst if compared to their
precursor can be attributed to the presence of Fe3+ cations.
and Fe4(SiW12O40) 3 were more active than their heteropolyacids
precursor. It is important to note that the reactions with these cat-
alysts had the same amount of Fe3+ cations.
The activity observed in esterification reactions in the pres-
ence of protonated heteropolyacid catalysts followed the trend
H4SiW12O40> H3PW12O40> H3PMo12O40. This is also the trend
of softness of the heteropolyanions. Izumi et al. estimated the soft-
ness of anions of various acids in aqueous solution as follows:
3.3. Screening of most active transition metal-exchanged
heteropoly acid catalyst
2−
SiW12O404-> NO3−> SO4 [40].
Regardless of the heteropoly anion, the activity of cata-
lysts in terms of metal cations always had the tendency:
Fe3+> Cu2+> Mn2+> Co2+ (Fig. 8). This result is partially in agree-
ment with the Lewis acidity reported for these metals by Osaka and
Tanaka, which found the tendency Fe3+> Cu2+> Co2+> Mn2+ [41].
Based on these results, the iron catalysts were selected to proceed
with the assessment.
To select the most active catalyst, we carried out reactions of
glycerol esterification in the presence of (i) Keggin heteropoly-
acids and (ii) transition metal- exchanged Keggin heteropolyacids.
Table 3 summarizes the main results.
Regardless of the addenda atom (i.e., Si or P) or metal present
in the tetrahedral center of HPA (i.e.,W or Mo), the iron salt het-
eropolyacid catalysts were the most active. Notably, FePW12O40
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
10
2000
When we carried out reactions with Fe4(SiW12O40) 3 catalyst
at the same concentration as the other HPA (ca.0.06 mmol), it was
impossible to verify the effect of excess acetic acid, because all reac-
tions were completed within the first hour of reaction. Therefore,
this catalyst was employed in four times lower concentration than
the HPAs catalysts (ca. 0.015 mmol).
The reaction selectivity is also a key aspect in these processes.
In glycerol esterification processes, triacetyl glycerol (i.e., triacetin)
is sometimes the more desirable product, because it is used as a
bio-additive in diesel or gasoline. However, mono or diacetates are
also useful as ingredient in surfactants.
Remarkably, H4SiW12O40 and mainly Fe4(SiW12O40)3were the
most selective catalysts in the formation of di or triacetyl glycerol
products (Fig. 11). In addition, monitoring of the reaction selectivity
throughout the Fe4(SiW12O40)3-catalyzed glycerol esterification
revealed an important aspect of these processes (Fig. 12).
Regardless of the excess of HOAc used, in all cases, the
monoacetyl glycerol selectivity reaches a maximum value within
first 30 min of reaction; subsequently, the rate of formation of
diacetyl glycerol increased substantially. This was more pro-
nounced when the initial HOAc: glycerol ratio was also increased.
Consequently, while monoacetyl glycerol accumulates, it is consec-
utively converted to diacetyl glycerol, which in turn is transformed
to triacetyl glycerol.
1800
1600
1400
1200
1000
800
FePW12O40
FePMo12O40
Fe4(PW12O40)3
600
1:3
1:6
1:9
1:12
glycerol: HOAc molar ratio
Fig. 13. Effect of reactants stoichiometry on TON of Fe-HPA-catalyzed glycerol ester-
ification reactions.
Reaction conditions: glycerol (47.92 mmol); iron (III) concentration (0.06 mmol);
temperature (333 K); reaction time (8 h).
TON calculate: mmol of glycerol converted + 2(mmol of mono acetate converted) + 3
(mmol of diacetyl glycerol converted)/mmol of Fe(III) present in the HPA catalyst.
It can be concluded that nitrate anions have little or no influence
on catalytic activity of HPA salts assessed; the FePW12O40, which
had the highest amount of nitrate anions, and FePMo12O40, which
had the lowest amount of nitrate anions, were much less efficient
than Fe4(SiW12O40) 3, the HPA salt that also had low amount of
nitrate anions.
3.4. Fe3PW12O40, Fe3PW12O40 or Fe4(SiW12O40) 3- catalyzed
glycerol esterification reactions with HOAc: Effect of the molar
ratio of reactants
As can be seen from Fig. 13 and Table 4, that compared to the
other heteropolyacids, Fe4(SiW12O40) 3 was by far the most active
catalyst. It achieved the highest TON (ca.1500 to 1950, molar ratio
of reactants of 1:3 to 1:12, respectively), while the FePW12O40 and
FePMo12O40 reached TON ranging from 750 to 800. The results of
these reactions in the presence of Fe4(SiW12O40) 3 are superior to
those describe in literature (Table 4)[42].
The effect of reactants stoichiometry was investigated at pro-
portions of 1:3, 1:6, 1:9 and 1:12 glycerol to acetic acid. The
conversion vs. time curves obtained are depicted in Fig. 9. Again the
Fe4(SiW12O40) 3 catalyst was the most active in all the reactions
even when utilizing different proportions of HOAc to glycerol.
In general, an increase in the acetic acid to glycerol ratio resulted
in a higher reaction rate. However, this effect was less noticeable
in the FePMo12O40-catalyzed reactions. Conversely, when in pres-
ence of Fe4(SiW12O40) 3, this effect was remarkable. The highest
conversions (ca.100%) were achieved within the first 2 h of reaction,
regardless of the excess HOAc used.
The use of Fe4(SiW12O40) thus represents
a substantial
improvement if compared to H2SO4 and other catalyst and may
quickly become a viable alternative to produce glycerol esters [43].
It was found that in the control-reactions (i.e. without catalyst),
even a large excess of HOAc resulted only in a low glycerol con-
version (ca.34%; 1:12 glycerol to HOAc). Fig. 10 shows selectivity
data.
When glycerol and HOAc were used at 1:3 molar ratio, the major
product was always monoacetyl glycerol. Conversely, in the reac-
tions that were conducted using glycerol to HOAc at molar ratios
of 1:6 and 1:9, the proportion of mono:diacetyl glycerol was 10:1.
An increase in the molar ratio to 1:12 resulted in the formation of
mono: diacetyl glycerol in a 10:2 proportion. No triacetyl glycerol
was detected in these reactions.
3.5. Effect of the catalyst concentration on
FeSiW12O40-catalyzed glycerol esterification
In order to investigate the influence of the Fe4(SiW12O40) 3
concentration on the rate of the reaction, catalytic runs were car-
ried out using concentrations in the range of 0.005 to 0.015mol%
(Fig. 14). The reactions were performed with a HOAc:glycerol:
molar ratio (ca. 3: 1) and at 308 K temperature.
To determine the order in relation with the catalyst concentra-
tion, only the initial rates of reaction were taken into consideration.
Hence, the reactions were carried out at lower temperature
Table 4
Glycerol esterification with HOAc catalyzed by different Brønsted or Lewis acids.a
Catalyst
Conversion (%)
TON
TOF valuesg TONh−1
Products distribution (%)
valuesb,c
monoacetyl-glycerol
diacetyl-glycerol
triacetylglycerol
cni.
PTSAa,b
H2SO4
Fe4(SiW12O40
Ag3PW12O40
85
98
100
75.7
1465
1988
1541
–
183
248
771
–
86
54
28
59.7
8
0
0
8
3.2
6
19
0
a,b
27
64
37.1
c,d
)
e
0
a
Reaction conditions: HOAc (143.76 mmol), glycerol (47.92 mmol), catalyst concentration (H + ) = 0.030 mmol; temperature = 333 K; 8 h reaction.
TON = calculated from number mmol of glycerol reacted + number mmol of monoacetyl glycerol reacted/1 mmol H + of catalyst [42].
TON = mmol of glycerol converted + 2(mmol of monoacetate converted) + 3(mmol of diacetyl glycerol converted)/mmol of heteropolyacid catalyst (i.e.0.015 mmol).
2 h of reaction.
b
c
d
e
HOAc: glycerol = 10:1; 393 K,1 wt.%solid catalyst (ca.0.017 mmol) loading relative to glycerol (ca. 5 g, 54.34 mmol), 15 min [42].
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
11
(a)
B
0.015 mol%
(b)
Linear Fit of B
Linear Fit of B
100
0.0125 mol%
Equation
Weight
Residual Su
m of Squares
y = a + b*x
No Weightin
0.02603
0.01 mol%
0.0075 mol%
-9.1
80
0.005 mol%
Adj. R-Squar
0.97517
0.00125 mol%
Value
-5.8885
0.76354
Standard Err
0.26725
0.05435
B
B
Intercept
Slope
60
-9.8
40
20
0
-10.5
-6.4
-5.6
-4.8
-4.0
0
50
100
150
200
250
ln [Fe4(SiW12O40)3]
time (min)
Fig. 14. Effect of Fe4(SiW12O40) 3 concentration on glycerol esterification (a) and order of reaction in respect to catalyst concentration(b).
Reaction conditions: glycerol (47.92 mmol); HOAc (143.76 mmol); time (4 h), temperature (308 K).
(ca.308 K) than in the previous experiments and only the conver-
sions achieved within the first reaction hour were considered.
From the curve showed in Fig. 14a, reaction rates were obtained,
allowing the construction of a linear plot of ln(reaction rate) versus
ln(catalyst concentration) (Fig. 14b). A high linearity coefficient
was obtained (R2 = 0.98) and thus a linear dependence on the
Fe4(SiW12O40) 3 concentration can be considered. The slope of
curve (ca.1.2) suggests a first-order dependence in respect to cata-
lyst concentration.
The same procedure allowed the determination of the order
related to the H4SiW12O40 concentration, which was equal to ca.
0.7 (Fig. 15).
Therefore, even if it is very difficult to directly compare two dif-
ferent catalysts (i.e. Brønsted and Lewis acids), we can conclude that
rate of the reaction is much sensitive to increase the concentration
The slope of the curves presented in Fig. 17 provided the rate
constants (k) for each reaction. The resulting linear equations were
found to present high linear correlation coefficients (R2) (Table 1,
supplementary material), thus indicating that the reaction is of
first-order in relation to the glycerol concentration at any temper-
ature.
From the slope of the curve (-E/R) obtained by plotting lnkx1/T
(Fig. 18), the activation energy of H4SiW12O40 or Fe4(SiW12O40)
3-catalyzed glycerol esterification with HOAc was determined.
The activation energy of the H4SiW12O40-catalyzed reactions was
found to be 27.7 kJmol-1, while the activation energy obtained in
Fe4(SiW12O40) 3-catalyzed reactions was equal to 36.14 kJmol-
1Initially, this result seems surprising because Fe4(SiW12O40)
3was more active than H4SiW12O40. However, it is important to
note that the concentration of Fe4(SiW12O40) 3 catalyst was four
times lower than that of H4SiW12O40 heteropolyacid.
Indeed, in the reaction conditions studied in this research, it
was impossible to perform a direct comparison of activation energy
of these catalysts at the same concentration, because the effect of
temperature in reactions with Fe4(SiW12O40) 3 at 0.06 mmol was
almost imperceptible.
As suggested by Firouzmanesh and Azar, we normalized the
k values by the catalyst concentration, considering its respective
reaction order [44]. Fig. 19 shows the Arrhenius plot with normal-
ized K values.
The activation energy values were equal to those obtained in
the Arrhenius plots without normalization. It means that this is an
inadequate strategy to compare the efficiency of catalysts when
they are used at different concentrations.
of Fe4(SiW12O40
) than H4SiW12O40 catalyst.
3
3.6. Kinetic studies
3.6.1. Effect of temperature on H4SiW12O40 or Fe4(SiW12O40)
3-catalyzed glycerol esterification reactions with HOAc
To assess the effects of changing temperature on the reaction
rate, plots of glycerol concentration versus time were constructed
for each H4SiW12O40 or Fe4(SiW12O40) 3-catalyzed glycerol
esterification reactions at temperatures of 298 to 333 K (Fig. 16).
An increase in temperature from 298 to 308 K drastically
increased the initial reaction rate (Fig. 16). Generally, an increase
in temperature enhanced the reactions conversion. However, the
temperature effect was more noticeable in Fe4(SiW12O40) 3-
catalyzed reactions. Except for the reaction at 298 K, all runs
attained total conversion within the time period studied (8 h).
Whereas, among the H4SiW12O40-catalyzed reactions, only the
one performed at 333 K temperature achieved total conversion.
Nevertheless, we should be reminded that the rate of the
reaction is more sensitive to increase in the concentration of
Fe4(SiW12O40)3 than H4SiW12O40(i.e., 1.2 against 0.7 order in rela-
tion to the concentration of Fe4(SiW12O40
respectively).
)
3
and H4SiW12O40,
Another strategy was found to compare these two catalysts at
different concentrations. Eyring plot, based on transition state the-
ory, provided the activation parameters (i.e., activation ꢀS, and
activation ꢀH), determined from the temperature dependence of
the rate constant [45]. The intercept values can be obtained from
an extrapolation. It is a more valid approximation for reactions that
occur in liquid phase.
3.6.2. Dependence of reaction rate in relation to glycerol
concentration
Kinetic measurements were performed within the initial period
of reaction, while they are far from their equilibrium positions. For
these reasons, we considered only three first points (Fig. 17). From
the data measured at each temperature, it was possible to obtain
the reaction rate constants. Based on these data, we constructed the
plots shown in Fig. 18. The reaction order of glycerol concentration
was determined from plots of ln(glycerol) versus reaction time built
for each reaction at different temperatures (Fig. 18).
The Eyring equation can be written as:
ln(k/T) = −1/T(ꢁH∗/R) + lnk’/h + ꢁS∗/R’
(1)
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
12
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
0.015 mol%
0.0125 mol%
0.01 mol%
0.0075 mol%
0.005 mol%
0.00125 mol%
B
(a)
Linear Fit of B
Linear Fit of B
(b)
100
80
60
40
20
0
-7.8
Equation
Weight
y = a + b*x
No Weighti
0.00745
Residual Su
m of Square
s
Adj. R-Squa
0.99117
Value Standard Er
B
B
Intercept
Slope
-5.147
0.6906
0.14362
0.02912
-8.4
-9.0
-6.4
-5.6
-4.8
-4.0
0
50
100
150
200
250
ln [H4SiW12O40
]
time (min)
Fig. 15. Effect of H4SiW12O40 concentration on glycerol esterification (a) and order of reaction in respect to H4SiW12O40 catalyst concentration (b).
Reaction conditions: glycerol (47.92 mmol); HOAc (143.76 mmol); time (4 h), range of catalyst concentration (0.015 to 0.0025 mol%), temperature 308 K.
100
100
80
80
298 K
308 K
298 K
308 K
318 K
328 K
333 K
318 K
328 K
333 K
60
40
20
0
60
40
20
0
0
25
50
75
100
125
150
175
200
225
250
0
50
100
150
200
250
Time (min)
Time (min)
(b) Fe4(SiW12O40)3-catalyzed reactions
(a) H4SiW12O40-catalyzed reactions
Fig. 16. Effects of temperature on conversion of H4SiW12O40 (a) or H4SiW12O40 (b)- catalyzed glycerol esterification with HOAc at different temperatures.
Reaction conditions: glycerol (47.92 mmol); HOAc (143.76 mmol), H4SiW12O40 (0.06 mmol); Fe4(SiW12O40
)
3
(0.015 mmol); temperature (333 K); reaction time (8 h).
0.0
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
0,0
-0,2
-0,4
-0,6
-0,8
-1,0
-1,2
-1,4
-1,6
-1,8
298 K
308 K
298 K
308 K
318 K
318 K
328 K
328 K
338 K
333 K
500
1000
1500
2000
2500
3000
3500
4000
1000
1500
2000
2500
Time (s)
Time (s)
(b) Fe4(SiW12O40)3- catalyzed reactions
Fig. 17. Plots of ln(glycerol concentration) t = ti/(glycerol concentration) t = 0 versus reaction time obtained from H4SiW12O40(a) or H4SiW12O40 (b)-catalyzed glycerol
esterification with HOAc at different temperatures.
(a) H4SiW12O40-catalyzed reactions
Reaction conditions: glycerol (47.92 mmol); H4SiW12O40 concentration (0.06 mmol); Fe4(SiW12O40) 3 concentration (0.015 mmol); reaction time.
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
13
Equation
Adj. R-Square
y = a + b*
1
-7,5
(a)
Equation
Adj. R-Squar
y = a + b*
1
(b)
-7,8
-8,0
-8,2
-8,4
-8,6
-8,8
-9,0
-9,2
-9,4
-9,6
Value
2,30378
-3,33411 2,50074E-15
Standard Error
7,95669E-15
Value
Standard Erro
Linear Fit of B Intercept
Linear Fit of B Slope
Linear Fit of Intercept 5,30992 1,08782E-14
Linear Fit of Slope
-4,4031 3,41895E-15
-8,0
-8,5
-9,0
3,0
3,2
3,4
3,0
3,2
3,4
1/T (10-3)
1/T (x10-3)
Fig. 18. Arrhenius plot of H4SiW12O40 (a) or Fe4(SiW12O40
)
3
(b)-catalyzed glycerol esterification with HOAc.
-5,6
(b)
(a)
-2,8
Equation
Adj. R-Square
y = a + b*x
0,99714
-5,8
Equation
Adj. R-Square
y = a + b*x
0,99056
-3,0
-3,2
-3,4
-3,6
-3,8
-4,0
-4,2
-4,4
-4,6
Value
4,27378
-3,33411
Standard Error
0,28212
-6,0
-6,2
-6,4
-6,6
-6,8
-7,0
B
B
Intercept
Slope
Value
Intercept 10,41005
Slope -4,42343
Standard Error
0,682
0,0892
B
B
0,21564
2,95
3,00
3,05
3,10
3,15
3,20
3,25
3,30
3,35
3,40
2,95
3,00
3,05
3,10
3,15
3,20
3,25
3,30
3,35
3,40
1/T (x 10-3)
1/T (x10-3)
H4SiW12O40
Fe4(SiW12O40)3
Fig. 19. Arrhenius plot of H4SiW12O40 (a) or Fe4(SiW12O40
) (b)-catalyzed glycerol esterification with HOAc employing normalized k values.
3
Table 5
Thermodynamic parameters of activation obtained on glycerol esterification with
HOAc in presence of H4SiW12O40 or Fe4(SiW12O40 catalysts.
tion, regardless of temperature, if the thermodynamic equation
(ꢀG = ꢀH − T ꢀS) is applied, it can be verified that Fe4(SiW12O40)3-
catalyzed reactions are always more spontaneous (i.e., more
negative ꢀG values).
)
3
Catalyst
Activation ꢀH∗
Activation ꢀS∗
Intercept (s−1
K−1
.
(kJ mol−1
)
(J.K−1 mol−1
)
)
H4SiW12O40
Fe4(SiW12O40
−26.4
−34.1
−308.1
−309.8
−13.3
−13.5
3.7. Insights on reaction mechanism of Fe4(SiW12O40)
3-catalyzed glycerol esterification
)
3
Table 6 shows the conversion and selectivity achieved in reac-
tions carried out in the presence and absence of Keggin anion and
Fe3 + cation. In the absence of Fe3 + cations, the reaction reached
almost the same conversion and selectivity of catalyst-free reaction
(runs 1 and 2, Table 6). Conversely, in the presence of Fe3 + cations,
the conversion slightly improved (ca. 28to 42%), as well as the
diacetyl glycerol selectivity (runs 2 and 3, Table 6).
The combination of Fe3+ cations and SiW12O403- anions had
a notable effect on both conversion and selectivity. Moreover, the
initial rate of reaction was the highest. Although the reactions in
the presence of H4SiW12O40 Brønsted acid catalyst also achieved
complete conversion, it required a longer time reaction (ca.45 min
for Fe4(SiW12O40) 3 against 180 min for H4SiW12O40; Fig. 21).
The Fe4(SiW12O40) 3 catalyst is partially soluble. On the
contrary, H4SiW12O40 heteropolyacid is totally dissolved in the
reaction medium; even though, the first was most active and selec-
where k = rate constant; T = temperature (K); ꢀH∗ = enthalpy of
activation (J mol−1); ꢀS∗ = entropy of activation (J K−1 mol−1);
R = molar gas constant (8.314 J. K−1 mol−1); k’ = Boltzmann constant
(1.38 × 10−23 J.K−1); h = Planck constant (6.63 × 10−34 J.s).
Thus, we built the Eyring plots for reactions catalyzed by the
silicon heteropoly acids which are displayed in Fig. 20.
The slope of curves corresponds to (ꢀH∗/R) value while
the intercept provides the value of ln k’/h + ꢀS∗/R’ (i.e., ln
k’/h = 23.76 K−1s−1). Therefore, we can easily calculate the values
of activation enthalpy (ꢀH∗) and activation entropy (ꢀS∗) corre-
sponding to two catalyzed reactions. Table 5 summarizes the data
obtained.
The activation ꢀH∗ values suggested that reactions in presence
of Fe4(SiW12O40
)
3
were more favorable (i.e. higher exothermic
character) than that in presence of H4SiW12O40 catalyst. In addi-
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
14
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
(a)
(b)
Fig. 20. Eyring plot of H4SiW12O40 (a) or Fe4(SiW12O40
) (b)-catalyzed glycerol esterification with HOAc.
3
Table 6
Comparison of silicon heteropolyacid catalysts and their precursors.a
Run
Catalyst
Conversion (%)
Products selectivity (%)
monoacetyl-glycerol
diacetyl-glycerol
triacetyl-glycerol
1
1
2
3
4
–
22
28
42
100
100
90
89
79
24
42
10
11
21
69
53
0
0
0
7
5
Na4SiW12O40
Fe(NO3)3
Fe4(SiW12O40
H4SiW12O40
)
3
a
Reaction conditions: glycerol (47.92 mmol), HOAc (143.76 mmol), catalyst (0.06 mol%), temperature (333 K); time (240 min).
100
80
Na4SiW12O40
Fe(NO3)3
Fe4(SiW12O40)3
60
H4SiW12O40
40
20
0
0
30
60
90
120
150
180
210
240
Time (min)
Fig. 21. Comparison of catalytic activity of the Fe4(SiW12O40
) and its precursor heteropolyacid and iron nitrate.
3
Reaction conditions: glycerol (47.92 mmol), HOAc (143.76 mmol), HPAs catalyst (0.06 mol %), Fe(NO3)3 (0.06 mol %), temperature (333 K); time (240 min).
tive. The main advantage is that Fe4(SiW12O40) 3 is a much less
corrosive Lewis acid. Potentially, after an adequate thermal treat-
ment, this catalyst could be used under heterogeneous conditions.
Therefore, because it does not contain protons, we assume that
Fe4(SiW12O40) 3 acts as a Lewis acid catalyst. In this case, we sup-
pose that in solution, the carbonyl group of acetic acid is polarized
by interacting with iron present in the Fe4(SiW12O40) 3 making a
nucleophilic attack by a hydroxyl group of glycerol more favorable.
Subsequently, the reaction proceeds through conventional way,
with the loss of water and formation of the glyceryl esters.
4. Conclusion
The esterification of glycerol in the presence of transi-
tion metal-exchanged Keggin heteropolyacids was investigated.
Metal-exchanged HPAs i.e. M3/nPW12O40, M3/nPMo12O40 and
M4/nSiW12O40 (M = Fe3+, Cu2+, Co2+ and Mn2+ (II); n = nox of
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003

G Model
MOLCAA-9802; No. of Pages15
ARTICLE IN PRESS
M.J. Da Silva et al. / Journal of Molecular Catalysis A: Chemical xxx (2016) xxx–xxx
15
metal) were synthesized. The activity of HPA salt containing metal
cations was in the following order; Fe3+> Cu2+> Mn2+ > Co2+. It was
found that only the acidic protons of H4SiW12O40 heteropoly-
acid were successfully exchanged with Fe3+ cations, resulting in
the Fe4(SiW12O40) 3 salt, the most active and selective catalyst.
The highest conversion (ca.99.9%) was achieved in Fe4(SiW12O40)
3-catalyzed esterification reactions, along with the highest selec-
tivity for di and triacetyl glycerol (ca.55 and 42%, respectively).
A kinetic study allowed the determination of the effects of tem-
perature, HOAc:glycerol molar ratio, concentration and nature of
catalysts on the conversion and selectivity of glycerol esterifica-
tion. The Fe4(SiW12O40) 3 catalyst is an alternative to the corrosive
Brønsted acid catalysts commonly used in esterification reactions.
[16] T. Hibi, K. Takahashi, T. Okuhara, M. Misono, Y. Yoneda, Appl. Catal 24 (1986)
69–83.
[17] T. Nishimura, T. Okuhara, M. Misono, Applied Catalysis 73 (1991) L7–L11.
[18] S. Fernando, S. Adhikari, K. Kota, R. Bandi, Fuel 86 (2007) 2806–2809.
[19] E. Garcia, M. Laca, E. Perez, A. Garrido, J. Peinado, Energ. Fuels 22 (2008)
4274–4280.
[20] V.L.C. Gonc¸ alves, B.P. Pinto, J.C. Silva, C.J.A. Mota, Catal. Today 133 (2008)
673–677.
[21] T. Fukumura, Ind. Eng. Chem. Res. 48 (2009) 1816–1823.
[22] D.E. Lopez, J.G. Goodwin Jr., D.A. Bruce, E. Lotero, Appl. Catal. A 295 (2005)
97–105.
[23] J.A. Melero, R.V. Grieken, G. Morales, M. Paniagua, Energ. Fuels 21 (2007)
1782–1791.
[24] M.J. da Silva, A.L. Cardoso, R. Augusti, J. Am. Oil Chem. Soc. 85 (2008) 555–560.
[25] K.-I. Shimizu, K. Niimi, A. Satsuma, Appl. Catal. A 349 (2008) 1–5.
[26] K.-I. Shimizu, H. Furukawa, N. Kobayashi, Y. Itaya, A. Satsuma, Green Chem. 11
(2009) 1627–1632.
[27] L.R. Pizzio, P.G. Vásquez, C.V. Cáceres, M.N. Blanco, Appl. Catal. A: Gen. 256
(2003) 125–139.
Acknowledgments
[28] B. Hamad, R.O.L. de Souza, G. Sapaly, M.G.C. Rocha, P.G.P. de Oliveira, W.A.
Gonzalez, E.A. Sales, N. Essayem, Catal. Commun. 10 (2008) 92–97.
[29] F.A. Miller, C.H. Wilkins, Anal. Chem. 8 (1952) 1253–1294.
[30] P. Melnikov, V.A. Nascimento, I.V. Arkhangelsky, L.Z.Z. Consolo, L.C.S. de
Oliveira, J Therm Anal Calorim 115 (2014) 145–151.
[31] J. Moffat, Metal-oxygen clusters: the surface and catalytic properties of
heteropoly oxometalates, Kluwer Academic Publishers, New York, 2001.
[32] P. Villabrille, G. Romanelli, L. Gassa, P. Vazquez, C. Caceres, Appl. Catal. A 324
(2007) 69–76.
The authors would like to acknowledge the financial support
from the Brazilian research agencies FUNARBE, CAPEs, FAPEMIG
and CNPq.
References
[33] C.F. Oliveira, L.M. Dezaneti, F.A.C. Garcia, J.L. de Macedo, J.A. Dias, S.C.L. Dias,
K.S.P. Alvim, Appl. Catal. A: Gen. 372 (2010) 153–161.
[1] M. Misono, I. Ono, G. Koyano, A. Aoshima, Pure Appl. Chem. 72 (2000)
1305–1311.
[2] F. Cavani, Catal. Today 41 (1998) 73–86.
[34] S. Borghese, B. Louis, A. Blanc, P. Pale, Catal. Sci. Technol. 1 (2011) 981–986.
[35] S. Borghese, A. Blanc, P. Pale, B. Louis, Dalton Trans. 40 (2011) 1220–1223.
[36] C. Hu, M. Hashimoto, T. Okuhara, M. Misono, J. Catal. 143 (1993) 437–448.
[37] E.F. Kozhevnikova, I.V. Kozhevnikov, J. Catal. 224 (2004) 164–169.
[38] L. Marosi, E.E. Platero, J. Cifre, C.O. Areán, J. Mater. Chem. 10 (2000)
1949–1955.
[39] B.W.L. Southward, J.S. Vaughan, C.T. O’Connor, J. Catal. 153 (1995) 293–303.
[40] Y. Izumi, K. Matsuo, K. Urabe, J. Mol. Catal. A 18 (1983) 299–310.
[41] K.-I. Tanaka, A. Ozaki, J. Catal 8 (1967) 1–7.
[42] C.E. Gonc¸ alves, L.O. Laier, A.L. Cardoso, M.J. da Silva, Fuel Process. Technol. 102
(2012) 46–52.
[43] S. Zhu, X. Gao, F. Dong, Y. Zhu, H. Zheng, Y. Li, J. Catal. 306 (2013) 155–163.
[44] M.R. Firouzmanesh, A.A. Azar, Polym. Int. 52 (2003) 932–937.
[45] J.C. Polanyi, A.H. Zewail, Acc. Chem. Res. 28 (1995) 119–132.
[3] I. Kozhevnikov, J. Mol. Catal. A 262 (2007) 86–92.
[4] I. Kozhevnikov, Catalysts for Fine Chemicals, vol.2: Catalysis by
Polyoxometalates, Wiley Chichester, 2002.
[5] J.J. Berzelius, Poggendrof Ann. Phys. Chem. 6 (1826) 369–392.
[6] M.N. Timofeeva, Appl. Catal. A 256 (2003) 19–35.
[7] J.H. Clark, Accounts Chem. Res. 35 (2002) 791–797.
[8] C.L. Hill, C.M. Prosse-McCartha, Coordin. Chem. Rev. 143 (1995) 407–455.
[9] A. Corma, Chem. Rev. 95 (1995) 69–70.
[10] T. Okuhara, N. Mizuno, M. Misono, Appl. Catal. A 222 (2001) 63–77.
[11] I.V. Kozhevnikov, Chem. Rev. 98 (1998) 171–198.
[12] W. Qi, L. Wu, Polym. Int. 58 (2009) 1217–1225.
[13] J. Macht, R. Carr, E. Iglesia, J. Am. Chem. Soc. 131 (2009) 6554–6565.
[14] K.M. Reddy, N.S. Babu, I. Suryanarayana, P.S.S. Prasad, N. Lingaiah,
Tetrahedron Lett. 47 (2006) 7563–7566.
[15] M.T. Pope, A. Mullar, Angew. Chem. Int. Ed. Engl. 30 (1991) 34–38.
Please cite this article in press as: M.J. Da Silva, et al., J. Mol. Catal. A: Chem. (2016), http://dx.doi.org/10.1016/j.molcata.2016.03.003