CO2 role on the glycerol conversion over catalyst containing CaO-SiO2 doped with Ag and Pt

Article abstract of DOI:10.1016/j.cattod.2019.02.009

Solids with strong base sites of calcium-silica mixed oxides (CaO-SiO₂), and zirconia (ZrO₂), both doped with platinum (Pt) or argentum (Ag), were synthesized and catalytically evaluated in the dehydration of glycerol for valuable chemical intermediates production. The samples were characterized by X-ray diffraction (XRD), Scanning Electron Microscopy and Energy Dispersive X-ray Spectroscopy (SEM-EDS), Basicity measurements by CO₂ temperature programmed desorption (CO₂-TPD), thermal analyses (TG/DTA) and N₂ adsorption–desorption isotherms. The effect of two different carrier gas, inert (N₂) or a soft oxidant (CO₂) over the activity and selectivity were evaluated in glycerol dehydration for all the samples. The high affinity between Ca and glycerol has effect over the product selectivity, which shows CO₂ effect. The promotion with Ag or Pt (0.2 wt%) increase the glycerol conversion and change the product selectivity. The data points out to the hydrogen transfer improving due to the Pt presence. Both, Pt or Ag doping or CO₂ addition as the carrier gas may inhibit the acetic acid production, over calcium containing catalyst. The promotion of ZrO₂ with Ag or Pt play a less marked effect over the glycerol conversion, under CO₂ or N₂.

Full text of DOI:10.1016/j.cattod.2019.02.009

Accepted Manuscript
Title: CO2 role on the glycerol conversion over catalyst
containing CaO-SiO2 doped with Ag and Pt
´
Authors: Jose de Ribamar M. Costa, Regina C.R. Santos,
Lucas P. Coutinho, Oderlando R. Silva, Helenilson O. Barros,
Valder N. Freire, Antoninho Valentini
PII:
S0920-5861(18)30343-2
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2019.02.009
CATTOD 11961
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
28 April 2018
31 December 2018
10 February 2019
Please cite this article as: Costa JdRM, Santos RCR, Coutinho LP, Silva OR,
Barros HO, Freire VN, Valentini A, CO2 role on the glycerol conversion over
catalyst containing CaO-SiO2 doped with Ag and Pt, Catalysis Today (2019),
https://doi.org/10.1016/j.cattod.2019.02.009
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

CO2 role on the glycerol conversion over catalyst containing CaO-SiO2 doped
with Ag and Pt
a
José de Ribamar M. Costa a,b, Regina C. R. Santos a,c, Lucas P. Coutinho , Oderlando R.
Silva a, Helenilson O. Barros a, Valder N. Freire c, Antoninho Valentini a,*
a
Department of Analytical Chemistry and Physical Chemistry, Federal University of Ceará,
Campus of Pici, Fortaleza, CE, CEP:60440-554, Brazil.
b Federal University of Maranhão, Imperatriz, MA, CEP:65915-240, Brazil.
c
Department of Physical, Federal University of Ceará, Campus of Pici, Fortaleza, CE,
CEP:60440-554. Brazil.
* Corresponding author. Tel.: +55 85 3366 9951. valent@.ufc.br
Graphical abstract
Highlights
1

 CaO-SiO2 based catalysts produce useful chemicals from glycerol in the gas – phase
reaction.
 Ca(OH)2 sites play an important role for the acetic acid formation.
 CO2 has effect over the catalytic performance by forming CaCO3.
 Pt addition improve the hydrogen transfer modifying the selectivity.
Abstract
Solids with strong base sites of calcium-silica mixed oxides (CaO-SiO2), and zirconia (ZrO2), both
doped with platinum (Pt) or argentum (Ag), were synthesized and catalytically evaluated in the
dehydration of glycerol for valuable chemical intermediates production. The samples were
characterized by X-ray diffraction (XRD), Scanning Electron Microscopy and Energy Dispersive
X-ray Spectroscopy (SEM-EDS), Basicity measurements by CO2 temperature programmed
desorption (CO2-TPD), thermal analyses (TG/DTA) and N2 adsorption–desorption isotherms. The
effect of two different carrier gas, inert (N2) or a soft oxidant (CO2) over the activity and selectivity
were evaluated in glycerol dehydration for all the samples. The high affinity between Ca and
glycerol has effect over the product selectivity, which shows CO2 effect. The promotion with Ag
or Pt (0.2wt%) increase the glycerol conversion and change the product selectivity. The data points
out to the hydrogen transfer improving due to the Pt presence. Both, Pt or Ag doping or CO2
addition as the carrier gas may inhibit the acetic acid production, over calcium containing catalyst.
The promotion of ZrO2 with Ag or Pt play a less marked effect over the glycerol conversion, under
CO2 or N2.
Keywords: glycerol, CaO, CO2, dehydration.
2

1. Introduction
The change of the world energy matrix toward renewable sources, which is being
implemented, is necessary and very important to improve the quality of life of the world population
[1,2]. The research in this sense is motived by the expectation of a decline in oil reserves [3,4],
and the need to minimizing the environmental impacts inherent to the transfer and use of energy
processes [5,6]. The biomass using, or the biomass derived materials, to produce chemicals or
fuels is an excellent alternative to oil reserve [7,8], since it is considered a neutral carbon dioxide
source. Worldwide, the biodiesel play a crucial role in this context [9,10]. However, despite the
actual development stage of the biodiesel using, it is not economically feasible [11,12]. Research
toward the use of the by-product (glycerol), to synthesize chemicals with high added value can
change the perspective of the biodiesel economy [13–15]. Actually, the glycerol is applied in
various industrial process [2,16,17]; however, due to its chemical structure it is considered an
excellent raw material for a wide range of chemicals with high added-value [18,19].
The glycerol, similarly to ethanol, may be used for hydrogen production, which is
interesting from energetic and environmental point of view [12,20–24]. It is known that a high
hydrogen to carbon (H/C) ratio make the feedstock more attractive for the H2 production. The
glycerol has a lower H/C ratio, if compared with ethanol and methane, which are 8/3, 3 and 4,
respectively. Consequently, the CO2 amount produced for the glycerol steam reforming is high
[25]. However, it have been shown that a high-purity H2 generation by steam reforming is achieved
by using a catalyst with simultaneous CO2 removal [20,21,24,26]. Therefore, despite the low H/C
ratio, the glycerol may play an important role on the H2 market.
The synthesis of liquid chemical products from glycerol generally is carried out under
heterogeneous catalytic process, such as dehydration and/or dehydrogenation reaction, which
3

make use of solids with acidic properties [19,27,28]. Catalysts such as Al2O3, V2O5, WO3, Nb2O5
and zeolites, are active to produce acrolein by glycerol dehydration [29]. Recent studies have
shown that the use of a catalyst comprising of bifunctional metals, i.e., containing acid/base and
redox sites have high activity and selectivity compared to other monometallic systems for glycerol
conversion to fuels and chemicals products [13,30,31].
On the other hand, copper based catalysts, such Cu/SiO2 or Cu/Al2O3 and Cu/Cr2O3
[13,19,32] present a high selectivity to acetol, a product of glycerol dehydration or
dehydrogenation process. The acetol, is an interesting compound, since it present various
application in industrial processes [33]. Due to its high reactivity, acetol plays a key role as an
intermediate substance in organic synthesis and also as a starting material in some chemical
reactions, including dehydration, dehydrogenation, hydrogenation, oxidation and polymerization
[34–37].
As previously mentioned, the glycerol structure (C3H8O3), a C3 molecule with 3 O-H
group, make it a high reactivity raw material. Therefore, the catalytic mechanism for the glycerol
transformation may be more complex than the suggested until now. The proposed reaction
mechanism for the dehydration process suggest the acrolein as the main product for the Brönsted
acid sites [1,38]. However, the Lewis acid/base pair can produce mainly acetol [14,19].
Additionally, the action of redox catalytic sites is also suggested to favour the acetol production
[13,39].
A base catalytic sites can play an important role in the process, since the mechanism
proposed consider the action of the Lewis acid/base pair [39–42]. On the other hand, few works
make use of solids with basic properties for the glycerol conversion. Using solid catalytic with
various composition such, hydrocalcite containing sodium aluminate, calcium oxide or
magnesium hydroxide, researches have been carried out the glycerol conversion in a liquid
medium, aiming the synthesis of glycerol carbonate, 1,2 propanediol, glyceric acid or lactic acid
4

[43–47]. The solid base catalyst play an important role in the glycerol oxidation, since it can
replace the sodium hydroxide, making the process more practical for industrial application.
Despite the rapidly adsorption of H2O and CO2 from the air at CaO surface [48], the high affinity
between CaO and glycerol observed by various researchers [47,49,50], may be explored for the
glycerol conversion process. However, in order to decrease the residence time and consequently
decreasing the selectivity for condensation products, it is interesting to carry out the process in the
gas phase.
Therefore, considering the previous reports, in this work we present the investigation of
the catalytic reactivity and selectivity of the calcium-silica mixed oxide and zirconia, both doped
with Pt or Ag, on the glycerol dehydrogenation/oxidation reaction. The process was carried out in
the gas phase under inert (N2) or a soft oxidant (CO2) flow gas, to evaluate the effect over the
products selectivity. These data may help in the elucidation of the mechanism process of the
glycerol transformation over the base sites on catalyst containing calcium oxide. In addition,
results from the amphoteric solid (ZrO2) performance are shown, in order to evidence the CO2
effect over the catalytic process. Then, under the light of the results presented by various
researchers previously and the data presented in this paper, it is proposing a reaction pathway for
the glycerol transformation with participation of solid base catalyst.
2. Experimental
2.1. Catalyst preparation
The syntheses of the supports of calcium-silica (CaO-SiO2) mixed oxides and zirconia was
performed based on the polymeric precursor method previously described [51,52]. The method
consists on the chelation of cations (metals) by citric acid (CA) in a water solution with subsequent
polyesterification reaction of the remaining carboxylic acid groups with ethylene glycol (EG).
5

Calcium carbonate, tetraethylorthosilicate (TEOS, Si(OC2H5)4), zirconium acetate, citric acid
monohydrate (CA) {C6H8O7.H2O} and ethylene glycol (EG) {C2H6O2} were used as starting
chemicals.
The CaO-SiO2 sample was prepared as follows: CA (0.369 mol) was dissolved in 100 ml
of water. Separately, the calcium carbonate (0.738 mol) was dissolved in 50 ml of HNO3 aqueous
solution. Always under stirring, the calcium solution and TEOS (0.369 mol) were added in the CA
solution and kept for 15 min at 50 °C. EG (1.476 mol) was subsequently added to the solution.
After polymerisation at 100 °C under stirring, the viscous resin was heat-treated at 350 °C for 3 h
in a muffle furnace. The resulting precursor compound was ground and heat-treated at 700 °C for
2 h under airflow. Similar synthesis conditions were used to prepare the ZrO2 sample, but using a
EG/CA/metal molar ratio of 1:1:1.
The Pt or Ag incorporation over CaO-SiO2 mixed oxide and ZrO2, was carried out by wet
impregnation method. An aqueous solutions of potassium hexachloroplatinate, (K2 [Pt(Cl)6]) and
silver nitrate (AgNO3) containing appropriate amount of the precursors salts to obtain 0.2 wt.% of
Pt or Ag was added for supports impregnation. The mixture, support oxide and the metal solution
(Pt or Ag), was placed in a rotary evaporator and, under constant vacuum at 70 °C the water was
extracted. After the sample drying at 100 °C for 8 h, the calcination was carried out in a muffle
furnace at 700 °C for 2 h under airflow. The samples were prepared to achieve a composition of
0.2 wt.% (percentage by mass of Pt or Ag) loaded on the CaO-SiO2 or ZrO2 oxides and, Ca/Si
molar ratio of 2.
2.2. Catalyst characterization
The crystalline structure of the solids was characterized by X-ray diffraction (XRD) using
XPert – Panalytical diffractometer operated at 40 kV and 40 mA employing Cu-Kα radiation
source (λ = 1.5406 Å). The experiments were performed at high diffraction angles (from 10 to
6

90°). Using the software X-Pert HighScore Plus® (PANalytical), the crystalline phases were
identified, and the crystallographic data for all phases were determined using the Inorganic Crystal
Structure Database (ICSD).
Textural properties of the catalysts were determined by adsorption/desorption isotherms
at −196 ºC on a Quantachrome instrument, Autosorb-1B model. The samples were degassed at
200 °C prior to the measurement. The specific surface area of the samples was determined using
the BET method. The pore size distribution was derived from the adsorption branches of the
isotherms using the Barrett–Joyner–Halenda (BJH) method. Scanning electron micrographs
(SEM) were recorded using a TESCAN VEGA XMU electron microscope equipped with an EDS
Bruker QUANTAX system coupled to the SEM microscope. The acceleration voltage of 20 kV
was used. Prior the analysis, the powder samples were dispersed on conductive carbon strip placed
on an aluminium sample holder and sputtered with gold.
The presence of base sites were determined by temperature-programmed desorption of CO2
(TPD - CO2) experiments. Prior to TPD, the samples were heat treated under He flow (20 mL
min−1) at 700 °C for 30 min and cooled under the same gas flow to 25 °C. CO2 adsorption was
carried out by using a flow of pure CO2 through the reactor for 30 min at 25 °C. After baseline
stabilization under He flow (20 mL min−1) at room temperature, the CO2-TPD run was carried out
with a heating rate of 10 °C min−1 from room temperature up to 700 °C. The CO2 desorption was
monitored using a thermal conductivity detector (TCD).
The thermogravimetric analysis (TGA) and differential thermal analysis (DTA) of the
spent catalyst samples were performed in a Shimadzu DTA-60H equipment to examine the
amounts of carbon deposits. The measurements were made under airflow (40 ml/min) from room
temperature to 800 °C, with a heating rate of 10 °C/min.
7

2.3. Catalytic activity
Catalytic tests were carried out in a fixed bed flow reactor at atmospheric pressure using
0.100 g of catalyst. Previously, the samples were activated at 700 °C under H2 flow of 20 mL/min.
The catalytic reaction was carried out at 280 °C under different carrier gas, inert (N2) or a soft
oxidant (CO2), for all the samples. The aqueous solution of 5wt% glycerol was fed to the reactor
by means of an injection pump at a flow rate of 3.6 cm3 h-1. This experimental condition was used
since the CaO in H2O presence form Ca(OH)2, while the Ca(OH)2 with CO2 presence form CaCO3.
Possibly the Ca(OH)2 and CaCO3 phases should present different properties, which may have
effect over the catalytic performance.
The reaction mixture was condensed in a cold trap maintained in a bath containing an ice-
water-salt trap (-15 °C). The main products were identified by GC–MS in a QP 2010 Plus,
Shimadzu model, and the quantification was performed using a Simple Chrom GC gas
chromatograph equipped with flame ionization detector and DB-5 column (20m x 0.18mm x
0.4µm), by using 1,4-butanediol as standard. The glycerol conversion and the main products
selectivity were calculated as follow:
퐶 −퐶
0
푡 푥100
(1)
( )
퐶표푛푣푒푟푠푖표푛 % =
퐶
0
퐶
퐶푆
( )
푃푟표푑푢푐푡푠 푆푒푙푒푐푡푖푣푖푡푦 % =
푥100
푡
(2)
퐶 −퐶
0
Where, C0 and Ct are the initial and the final amount of glycerol in the reaction mixture,
respectively. The CCS is the amount of chemical substance formed.
8

3. Results and discussion
3.1. Structural features of the solids by X-ray diffraction (XRD)
The Fig. 1a shows the XRD patterns of the samples containing calcium (Ca2+). The
calcination temperature of 700 °C is enough to decompose the Ca(OH)2 and CaCO3 phases and
produce CaO [53]. However, unlike the other samples, the peaks with highest intensity for the
sample CaO-SiO2 correspond to Ca(OH)2 phase. It is known that the formation of calcium
hydroxide from CaO and H2O vapour (Eq. 3), is thermodynamically possible up to 400 °C [53].
∆퐻2098 = −109푘퐽/푚표푙
(3)
( ) ) ( )
퐶푎푂 푠 + 퐻2푂 푔 → 퐶푎 푂퐻 푠
2
( )
(
Therefore, possibly this high ratio of Ca(OH)2 phase is due to the higher storage time and
the exposition to the atmospheric air humidity [48], before the XRD measurement. In the
experimental procedure, the sample was heat treated at 700 °C and cooled to room temperature
under atmospheric air. Therefore, despite the high calcination temperature (700 °C), as the
formation of CaCO3 from CaO and CO2 or Ca(OH)2 from CaO and H2O at atmospheric pressure
and ambient temperature are thermodynamically possible [48,54,55], the diffraction peaks are
consistent with the experimental conditions used.
On the other hand, the diffractogram of the samples Pt/CaO-SiO2 and Ag/CaO-SiO2 shows
a low intensity for the peaks corresponding to the Ca(OH)2 phase, which suggest the formation of
CaCO3 from Ca(OH)2 after the Ag or Pt impregnation and the new calcination. None of the
samples (Pt/CaO-SiO2, Ag/CaO-SiO2, Pt/ZrO2 and Ag/ZrO2), showed characteristic diffraction
peaks of Ag or Pt oxide phases (Fig. 1a and b), suggesting the high dispersion; however, the low
weight amount of Pt or Ag in the samples may be the main reason.
Fig. 1b shows the XRD patterns of pure ZrO2 support and the samples doped with Pt or Ag.
Diffraction peaks ascribed to the tetragonal and monoclinic crystalline phases of zirconia are
observed in all the samples, which are consistent with the calcination temperature of 700 °C [56].
9

As cited previously, the absence of Pt or Ag diffraction peaks is due to low content of Pt and Ag
in the sample composition (0.2 wt.%).
Fig. 1. XRD patterns of the samples calcined at 700 °C on different supports. (a) CaO-SiO2 mixed
oxides and, (b) ZrO2.
3.2. Textural properties and SEM–EDS analysis
The N2 adsorption-desorption isotherms and the corresponding pore size distribution of
fresh catalysts are shown in Fig. 2 and 3. The isotherms profile point out to the predominance of
macroporous or a non-porous characteristic of the solids [57]. The samples containing calcium
oxide displayed isotherms of type III, according to the IUPAC classification (Fig. 2a), which
results in a pore size distribution with high volume of N2 adsorption for pores with larger diameter.
The isotherm profile of the sample CaO-SiO2 present the H3 hysteresis loop, which may be due to
the formation of non-rigid particle aggregates, giving rise to slit-shaped pores [57,58].
10

Fig. 2. N2 adsorption/desorption isotherms. (a) CaO-SiO2 mixed oxides and, (b) ZrO2, supports.
The silver or platinum addition (Ag/CaO-SiO2, Pt/CaO-SiO2) lowered the inexpressive
hysteresis loop of the CaO-SiO2 sample (Fig. 2a), which not means the absence of pores. These
features of isotherms are typical from macroporous material (type III according to the IUPAC
classification).
The synthesis methodology affect directly on the morphological properties of the material.
Generally, the using of the sol-gel methods leads to the formation of mesoporous zirconia [59].
However, the N2 adsorption/desorption isotherms analysis for the zirconia containing samples
indicates a process of stacking crystalline plates, giving rise to narrow mesoporous in the form of
slits, as indicated by type V isotherms with H4 hysteresis loop obtained for ZrO2, Ag/ZrO2 and
Pt/ZrO2 (Fig. 2b).
11

Noticeably, the BJH pore size distribution for all the catalysts, reveals a relatively wide
pore size distribution, with a significant fraction of pore diameter in the range of 100–300Å,
confirming the existence of large mesopores, resulting in a broad pore size distribution (Fig. 3).
The addition of Pt or Ag over ZrO2 support resulted in a decrease of the pore volume, which is
perceptible in the pore diameter range of 100–300Å, pointing to the partial filling or obstruction
of zirconia pores (Fig. 3b).
Fig. 3. Pore size distribution of the samples determined by the BJH model. (a) CaO-SiO2 and, (b)
ZrO2, supports.
The specific surface area and the pore volume of the catalysts are summarize in Table 1,
and it shows that all the samples present a not meaningful surface area. However, the calcination
of the pure CaO-SiO2 support at 700 °C shows a surface area higher than the value observed
normally [60]. This surface area value observed may be due to the SiO2 effect, which contributed
positively for the morphology property of the sample.
12

Table 1. Textural properties of the supports and prepared catalysts.
Samples
BET surface area *Pore volume
(m2g-1)
(cm3g-1)
CaO-SiO2
Ag/CaO-SiO2
Pt/CaO-SiO2
ZrO2
41
48
50
14
9
0.47
0.51
0.43
0.13
Ag/ZrO2
0.07
Pt/ZrO2
8
0.09
* Pore volume at P/P0 = 0.99.
The samples containing ZrO2, despite of the low fraction of mesoporous presence
suggested by the low hysteresis loop (Fig. 2), shows a lower surface area and pore volume (Table
1), than the values usually observed [61].
The silver or platinum addition in the sample composition has little effect over the textural
properties (Fig. 2, Fig. 3 and Table 1). However, the data suggest that the deposition of Ag or Pt
on the CaO-SiO2 reduces the sintering process of the support, which is emphasized by a small
increase of the surface area.
On the other hand, as pointed out by the XRD (Fig. 1a), the samples Ag/CaO-SiO2 and
Pt/CaO-SiO2 present a low relative intensity for the peaks from the Ca(OH)2 phase, if compared
with CaO-SiO2 sample. Therefore, the higher surface area for the samples Ag/CaO-SiO2 and
Pt/CaO-SiO2 may be due to the lower Ca(OH)2 presence, which may have effect over the porous
availability, since the Ca(OH)2 phase present a lower density than CaCO3 phase.
The opposing effect is observed for the samples containing ZrO2. Although, the real role
of silver and platinum over the samples morphologic properties is not clear, all the catalysts
13

containing ZrO2 present a hysteresis suppression due to the Ag or Pt addition. Such behaviour has
been ascribed to the porous blocking by the Ag or Pt particles [62]. However, the Ag or Pt presence
in the composition may also have effect over the crystal grow or particle sintering, which results
in a textural property change.
The Fig. 4 shows the SEM micrographs. The samples containing calcium oxide shows a
spongy like feature, which is characteristic for the materials synthesized by the Pechini method
[63].
Fig. 4. Scanning electron micrographs of the samples after calcination at 700 °C.
14

This feature may suggests a sample with high surface area. Of course, the porous observed
with the SEM magnification of 8000x is far above the diameter of the pores detected by N2
adsorption isotherms. Therefore, the porous observed in Fig. 4 do not have effect over the
hysteresis loop of N2 adsorption/desorption isotherms.
Silver or platinum addition over the solid sample containing CaO, despite the low
magnification does not produce a significant change in the morphologic appearance of the sample,
which is consistent with the data presented in Table 1. Having as reference the samples that contain
CaO, the SEM micrography of the samples containing ZrO2 present a material feature with high
density, characteristic of the ZrO2 phase; which suggests a solid with low surface area, being
consistent with the Table 1 values and pore diameter distribution (Fig. 3). The low amount of citric
acid used in the synthesis procedure (CA:Zr of 1:1), contributes to this feature of the samples
containing ZrO2, since the organic precursor decomposition is the main process which promotes
the porous structure formation [63].
Figs. 5 and 6 shows EDS mapping imagens and spectrum to evaluate the distribution of
metals over both supports studied. The EDS mapping imagens of the pure CaO-SiO2 support and
doped with Pt or Ag (Fig. 5), shows a non-homogeneous distribution of the Ca and Si elements on
surface. Elemental mapping (EDS) of the materials can provide clear imagens of element
distributions, to further clarity the presence of Pt or Ag particles well dispersed. In contrast, the
signal is much weaker, which show that the metallic Pt or Ag species are at doping level on surface
of the particles. Additionally, during the Pt or Ag impregnation process the Ca(OH)2 and a fraction
of the CaCO3 may be dissolved, which with the solvent extraction precipitate again. This process
should promote a new dispersion of the calcium, which may also partially cover the Pt or Ag
particles. However, the EDS spectrum analysis also suggest a decrease of the Ca/Si ratio on the
surface samples due to the Pt or Ag impregnation process. Therefore, considering the XRD and
15

the MEV-EDS, the results point out that a ratio of the calcium is leached out during the
impregnation process, which have effect over textural properties, and is consistent with the data
presented in Table 1.
Fig. 5. EDS mapping images and respective spectrum of the pure CaO-SiO2 support and doped
catalysts with Pt or Ag.
The EDS mapping images and spectrum of the samples ZrO2 and Ag/ZrO2 (Fig. 6),
similarly to Fig. 4, reveal the formation of large aggregates of zirconium oxide particles. The
chemical element Hf present in both samples is due to the contamination of the zirconium acetate
solution used as precursor. Although, Zr and Au signals overlapped, and silver can be detected on
the Ag/ZrO2 catalyst the majority component is Zr. Unlike the samples containing calcium and
siliceous mixed oxide, the EDS image of the sample Ag/ZrO2 suggest a lower dispersion of the
16

silver over the zirconia surface, since it is possible to visualize small agglomerations, which are
highlighted with an arrow.
Fig. 6. EDS mapping images and respective spectrum for ZrO2 and Ag/ZrO2 catalysts,
respectively.
3.4. Basic properties by CO2-TPD measurements
The base properties of CaO-SiO2 and ZrO2 samples were investigated by CO2-TPD, a
commonly used experiment for measure the strength of base sites. It is well known, the affinity of
CaO or Ca(OH)2 to interact with CO2 [64,65]. Therefore, it is not surprising that CO2 addition to
the reaction mixture results in a catalytic effect under low temperature reaction. The chemisorption
of CO2 also occurs over ZrO2 [66], therefore the CO2 presence also may have effect over ZrO2
catalytic performance. The CO2-TPD profiles (Fig. 7), point out to the different base site strength
distribution in the CaO-SiO2 sample with evident two desorption peaks of CO2 at around 100 and
650 °C, implying the presence of weak and strong base sites [67]. This CO2 desorption may result
from the SiO2 contribution or effect.
17

On the other hand, the CO2-TPD profile for the sample ZrO2 shows only a weak CO2
desorption peak, and at low temperature, which mean a low CO2 adsorption ability and weak
interaction. This CO2 profile observed for ZrO2 sample, possibly is a consequence of the high
calcination temperature (700 °C), which can decrease drastically the surface area (Table 1) and
modify the surface acid-base properties of the ZrO2 [66]. Therefore, the results from CO2-TPD can
better to elucidate the catalytic effect produced by CO2 addition in the reaction mixture, which
results are presented below.
Fig. 7. Carbon dioxide temperature-programmed desorption (CO2-TPD) profiles of the samples.
The curves were divided for weak, medium and strong base sites, according to temperature range.
3.5. Catalytic activity
3.5.1. Effect of atmosphere reaction on glycerol conversion
The catalytic activity of all samples was evaluated for the gas phase glycerol conversion at
N2 or CO2 gas flow (Fig. 8 and 9). The Fig. 8 (a and b) shows a superior initial catalytic conversion
18

for the samples containing CaO in both reaction atmosphere, which is the result of the interaction
of glycerol with the weak and strong basic sites on surface of calcium-silica support. The
pronounced deactivation may be ascribed to the Ca(OH)2 formation, since the reaction of CaO
with H2O occurs promptly at 280 °C [43]. However, the compounds formation between Ca and
glycerol [49] also may contribute to deactivation process, due to a strong interaction.
Fig. 8. Catalytic conversion of glycerol for the samples containing CaO-SiO2, at different reaction
atmosphere. (a) Reaction carried out under N2 and (b) CO2 flow.
It has been reported that acid sites are responsible for the glycerol dehydration activity,
while the dehydrogenation reaction of alcohol requires both, strong base sites and weak acid sites
[14,19,38,39]. Therefore, the glycerol conversion observed for the sample CaO-SiO2 under N2
flow suggests that base solid can catalyse both, dehydration and/or dehydrogenation reactions of
glycerol. Additionally, the initial deactivation suggest that the CaO is more active than Ca(OH)2.
The addition of Ag and Pt on the CaO-SiO2 catalyst improved the glycerol conversion for the first
19

hour of time on stream, under N2 flow. This behaviour is consistent with the superior surface area
(Table 1) and redox properties of the catalysts doped with Ag or Pt. Therefore, there is an
improvement of the catalytic conversion due to the Pt or Ag addition; but it have no effect over
the catalytic stability.
On the other hand, the reaction carried out under CO2 flow shows a low initial catalytic
conversion for all the samples containing calcium oxide (Fig. 8b). It is known that in the
experimental condition used here, such as high vapour pressure and CO2 flow, the CaO undergoes
to the CaCO3 phase. Therefore, the CaCO3 phase formation is prejudicial to the catalytic activity,
and considering the similarity of the deactivation rate of the samples, the CaCO3 formation,
likewise to the Ca(OH)2, should result in a partial covering process of the Ag or Pt metals.
The profiles of Fig. 8b, shows that the Ag or Pt addition in the CaO-SiO2 sample do not
produced a meaningful change for the glycerol conversion along the time on stream. These results
indicate that, under the experimental conditions used, the Pt or Ag metallic sites are not the main
active sites for the glycerol conversion, therefore it is the basicity of the catalysts the key parameter
to perform the reaction. By using CO2 flow, the data points out that the change from CaO to
Ca(OH)2 and CaCO3 should be the main responsible by the catalytic deactivation observed.
However, despite the lower glycerol conversion on time on stream for all the samples, the data
points to the Ca(OH)2 activity for the glycerol dehydration process.
On the other hand, there are also two factors to be considered for the observed improving
of the catalytic activity for the samples containing Pt and Ag sites, under N2 flow. One is the
possible electronic effect from Ca to Pt or Ag. This behaviour is suggested due to the low Ag or
Pt amount (nominal value of 0.2 wt.%) and good dispersion of particles as evidenced by EDS
analysis (Fig. 5). A second effect is the superior surface area of the samples Pt/CaO-SiO2 and
20

Ag/CaO-SiO2 (Tale 1), which suggest to be the result from the lower Ca(OH)2 phase ratio in these
samples composition (Fig. 1).
Textural and chemical properties as well as the loading or dispersion of the active metals
have been reported to play an important role in the catalytic performance [68–70]. However, the
Fig. 9 points out no effect over the glycerol conversion due to the Ag or Pt addition on the
amphoteric ZrO2 oxide. The samples containing ZrO2, for the reaction condition applied, in
addition to the low glycerol conversion also shows deactivation process (Fig. 9). Despite the lower
initial conversion, compared to the samples containing calcium, the Fig. 9a suggests that under N2
flow the deactivation process is less pronounced. Therefore, although less expressive, the Fig. 9b
shows that the CO2 also has effect over the catalytic performance of the samples containing ZrO2,
despite the initial glycerol conversion remain near 10% (under N2) or under CO2, differently from
the samples containing CaO.
The data presented in Fig. 8 and 9, therefore, clearly points out to a strong CO2 effect over
the glycerol conversion only for the samples containing CaO. On the other hand, it is important to
emphasize that the impregnation process and a high temperature of the heat treatment under H2
flow has a considerable effect of the ZrO2 over the Pt active sites [71–73]. Therefore, the low effect
observed due to the Pt or Ag addition for the ZrO2 samples, over the catalytic performance, may
be a consequence of the metal-support interaction effect, which may results in the Pt or Ag partial
covering by the zirconium oxide. This process is possible considering the high temperature of
activation used (700 °C under H2 flow).
21

Fig. 9. Catalytic conversion of glycerol for the samples containing ZrO2, at different reaction
atmosphere. (a) Reaction carried out under N2 and (b) CO2 flow.
3.5.2. Effect of atmosphere reaction on the products selectivity
The product selectivity analysis may provide interesting information about the solid
property change and CO2 effects over the catalytic process. The Fig. 10 shows the selectivity to
the main products identified for the catalysts studied, based on the different supports.
The catalysts of CaO-SiO2 support presented acetol, methanol, acetic acid and ethylene
glycol as the main product. However, the acetic acid production is affected by the Pt or Ag addition
and by reaction conditions under CO2 atmosphere. Nevertheless, under N2 flow, the Ag addition
improved significantly the methanol selectivity, while Pt presence improve mainly the acetol
selectivity.
22

Fig. 10. Catalytic selectivity for the samples. Reaction carried out under (a) N2, (b) CO2 flow. (c)
Selectivity to acetol at different gas flow for samples containing ZrO2.
It is well accept that the glycerol dehydration produce acetol, and such reaction is an acid
catalysed process [13,74]. Therefore, the data suggest a modified reaction pathway, in which the
–OH group play a role, despite the lower basicity of the Ca(OH)2 relatively to CaO. It is known
that CaO reacts with glycerol forming calcium glycerides [49,50], additionally it is reported that
Ca(OH)2 solubility in glycerol/water mixture is higher than in water only [75], possibly due to the
formation of calcium glycerides. Lu Chen et all [47], by using CaO as solid base catalyst in
glycerol conversion to lactic acid, showed that the water content on reaction mixture has effect
over the lactic acid yields from glycerol; however, the decrease of the glycerol conversion is not
significantly affected. Therefore, the water presence produce Ca(OH)2, which should change the
reaction pathway, modifying the product selectivity.
The Scheme 1 present a proposed reaction pathway for acetol formation from glycerol with
Ca(OH)2 catalysed process. Due to the high water content in the reaction mixture (95 wt.%), it is
reasonable to propose the water action in the process. Considering the high affinity between
glycerol and CaO or Ca(OH)2 reported early [49,50,75], firstly is proposed an alkoxy formation
under CaO or Ca(OH)2 action. Following, with water participation (as OH group from Ca(OH)2),
the acidic hydrogen from C2 is abstracted, which may make easer the C-O bond rupture with
23

Ca(OH)2 and enol formation. The acetol is formed from the keto-enol tautomerization (Scheme 1)
[13,47]. The acetol may be desorbed or undergoes to posterior reactions.
Scheme 1. Reaction pathway proposed for acetol formation from glycerol over basic sites with
terminal OH abstraction.
It is accepted that the allyl alcohol may be produced from acetol by the hydrogen transfer
process [13], likewise it is reasonable to accept that the adsorbed species like the enol is subjected
to the hydrogen transfer. Therefore, the hydrogen transfer may occurs from enol to enol or from
an alkoxy to enol, producing allyl alcohol and pyruvaldehyde (Scheme 2a), or glyceraldehyde and
1,2-propanediol (Scheme 2 b). The glyceraldehyde may undergoes to a new dehydration under a
base action to form 2-hydroxypropenal [47], which by the keto-enol tautomerization produce
pyruvaldehyde.
24

Scheme 2. Reaction pathway proposed for the formations of (a) allyl alcohol and pyruvaldehyde
and, (b) glyceraldehyde and 1,2-propanediol.
The pyruvaldehyde, which present two carbon atom with low electron density, in water or
Ca(OH)2 presence may undergoes to nucleophilic addition reaction producing formaldehyde and
acetic acid or, formic acid and acetaldehyde (Scheme 3a). Subsequently, the formaldehyde and/or
the acetaldehyde are transformed to methanol and ethanol (Scheme 3b), by hydrogen transfer,
which may be from the glycerol adsorbed. The glyceraldehyde in the base presence (Ca(OH)2),
also may undergoes, by the retro-aldol process, to the production of 2-hydroxyacetaldehyde and
formaldehyde, which by hydrogen transfer will produce ethylene glycol and methanol [76].
25

Unfortunately, the pyruvaldehyde may undergoes to the condensation process with various alkoxy
structure present in the catalyst surface [47], which may neutralise the active sites.
Scheme 3. Reaction pathway proposed for the formations of (a and b) acetic acid and methanol
from pyruvaldehyde and hydrogen transfer, (c) methanol and ethylene glycol from glyceraldehyde,
with the metal site participation, adapted from [76].
26

Despite the low catalytic conversion (Fig. 8 and 9), the catalysts presented a relatively large
distributions of products. The Fig. 10 display the main products only; however other products
identified were: acetaldehyde, acetone, isopropanol, allyl alcohol, 1,3-propanediol.
The data presented, suggests that CO2 has effect over the catalytic activity and selectivity,
by forming calcium carbonate (CaCO3), which decreases the Ca or Ca(OH)2 ratio available to
interact with glycerol. The main change is observed for the acetic acid production, which is
affected by Ag or Pt addition and by the CO2 presence. The low effect over the glycerol conversion
observed after the Pt or Ag addition is consistent with a low amount and with a partial Pt or Ag
covering during the reaction process, by the in situ formed Ca(OH)2 or CaCO3 species, which
neutralise the metals catalytic activity. In addition, the partial covering of Pt or Ag particles may
occurs also during the impregnation procedure, since it was carried out using water solution of Pt
or Ag salts. However, the Pt presence clearly has effect over the product selectivity, by suppressing
the acetic acid production and improving the acetol, methanol and glyceraldehyde. Therefore, Pt
may act mainly on the hydrogen transfer process, favouring the acetol desorption, which have
effect over the production of acetic acid, since the pyruvaldehyde production may be affected by
the acetol desorption, according to Scheme 3a.
On the other hand, the product selectivity is significantly changed by altering the catalytic
system from CaO-SiO2 to ZrO2, an amphoteric support [66], which must have a superior stability
in the vapour presence. The selectivity observed for the samples containing ZrO2 (Fig. 10c), shows
acetol as the main product. However, it is observed the CO2 effect, despite the CO2-TPD profile
(Fig. 7) do not show CO2 desorption at the reaction temperature.
The samples Pt/CaO-SiO2 and Pt/ZrO2 were tested for the benzene hydrogenation reaction,
in order to detect the Pt particle or atom on the catalyst surface. The test was carried out with 0.100
27

g of sample with activation at 450 °C under H2 flow. The benzene vapour pressure in the reactional
mixture (H2 and benzene), was maintained at 35 mmHg, the total flow used was 15 cc/min at
reaction temperature of 60 °C. Both samples showed a very low hydrogenation product, lower
than 1% of benzene conversion. The low conversion points out to the Pt partial covering by the
support. It may explain the low effect over the glycerol conversion.
3.6. Characterization of spent catalysts by thermogravimetric and differential thermal analysis
(TGA/DTA)
Many catalytic process are affected by coke or carbonaceous material deposition [77–79],
which results in catalyst deactivation or selectivity change. The thermal analysis (TG-DTA) results
of the used catalysts are presented in Fig. 11.
Fig. 11. Thermal analysis (TGA-DTA) of the spent catalysts after 5 h of reaction at 280 °C. (a, b,
c) Catalysts CaO-SiO2 based, from reaction carried out under N2 flow, (d, e) ZrO2 catalysts at
different reaction atmosphere and, (f) Pt/CaO-SiO2 catalyst under CO2 flow.
28

The catalysts containing CaO-SiO2 shows an interesting profile of weight loss (Fig. 11a, b
and c). The samples subjected to the catalytic test carried out under N2 flow, clearly shows two
exothermic peaks in the range of 250 to 450 °C. These events are not perceptible for the sample
after the catalytic test under CO2 flow. This occurrence can be attributed to the combustion of
complex organic species of Ca with glycerol products deposited at the surface of the catalyst
[47,75,80]. This weight loss is of about 4.9, 3.1 and 16.3% for the pure calcium-silica and Ag or
Pt doped catalysts, respectively. The high amount of deposited organic products over Pt/CaO-SiO2
catalyst, which corresponds to a weight loss of 16.3% is resulted of higher glycerol conversion
with formation of stable carbon species. The TG data point out that CO2 act in a competitive way
with glycerol, producing CaCO3, which is no active in the glycerol conversion. Therefore, the
weight loss observed above at 600°C in all calcium-silica based samples is associated to the
endothermic CaCO3 decomposition [65].
The sample ZrO2 shows a low mass elimination, near 1wt%, in a large temperature range,
independently of the carrier gas applied (Fig. 11d and e). This profile suggest a partial rehydration
of the ZrO2 during the catalytic test, which is gradually eliminated. The carbonaceous material
presence can not be ruled out, however.
The Fig. 11f also shows that the events of weight loss between 250 and 450 °C are not
perceptible for the Pt/CaO-SiO2 sample after the catalytic test under CO2 flow. The glycerol
conversion over CaO based catalyst containing a large amount of water in the reaction mixture,
should produce the reactive Ca(OH)2. However, for the reaction carried out under CO2, the water
produce a low amount of Ca(OH)2, which can react with glycerol. Therefore, the TG curve analysis
show no accumulation of the Ca-glycerol complex products on the Pt/CaO-SiO2 surface (Fig. 11f).
The endothermic event in the range of 600 to 700 °C is characteristic of the CaCO3 decomposition,
consequently, the mass elimination due to the CaCO3 decomposition is less intense for the samples
29