Conversion of glycerol to hydroxyacetone over Cu and Ni catalysts

Article abstract of DOI:10.1016/j.cattod.2020.11.025

The conversion of glycerol to hydroxyacetone, in gas phase and atmospheric pressure, was investigated over γ-Al₂O₃ and TiO₂ supported copper and nickel catalysts. The catalysts were prepared by co-impregnation and pre-treated with hydrogen at 450 °C. It was studied eight different catalysts at 280 °C under nitrogen atmosphere. The higher glycerol conversion and selectivity to hydroxyacetone were obtained with the monometallic Cu/γ-Al₂O₃ and Cu/TiO₂ catalysts. Furthermore, high selectivity to lactic acid was obtained over Cu/TiO₂. These catalysts were characterized by XRD, SEM, EDS, TEM, N₂-physisoption, and NH₃-TPD. Furthermore, the Cu/γ-Al₂O₃ catalyst was studied on different reaction temperature and gas flow conditions. It was obtained that the selectivity to hydroxyacetone is favors at lower temperatures. However, the catalyst suffer deactivation. Carbonaceous deposits over the used sample were not revealed by Raman and FTIR. Finally, a regeneration treatment by oxidation–reduction reactivated the Cu/γ-Al₂O₃ catalyst successfully.

Full text of DOI:10.1016/j.cattod.2020.11.025

Catalysis Today xxx (xxxx) xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Conversion of glycerol to hydroxyacetone over Cu and Ni catalysts
´
´
Barbara C. Miranda Morales*, Braulio Andres Ortega Quesada
´
Escuela de Ingeniería Química, Universidad de Costa Rica, San Jose, 2060, Costa Rica
A R T I C L E I N F O
A B S T R A C T
Keywords:
Glycerol
The conversion of glycerol to hydroxyacetone, in gas phase and atmospheric pressure, was investigated over
γ-Al2O3 and TiO2 supported copper and nickel catalysts. The catalysts were prepared by co-impregnation and
pre-treated with hydrogen at 450 ◦C. It was studied eight different catalysts at 280 ◦C under nitrogen atmo-
sphere. The higher glycerol conversion and selectivity to hydroxyacetone were obtained with the monometallic
Cu/γ-Al2O3 and Cu/TiO2 catalysts. Furthermore, high selectivity to lactic acid was obtained over Cu/TiO2. These
catalysts were characterized by XRD, SEM, EDS, TEM, N2-physisoption, and NH3-TPD. Furthermore, the Cu/
γ-Al2O3 catalyst was studied on different reaction temperature and gas flow conditions. It was obtained that the
selectivity to hydroxyacetone is favors at lower temperatures. However, the catalyst suffer deactivation.
Carbonaceous deposits over the used sample were not revealed by Raman and FTIR. Finally, a regeneration
treatment by oxidation–reduction reactivated the Cu/γ-Al2O3 catalyst successfully.
Hydroxyacetone
Dehydration
Alumina
Titania
Copper
1. Introduction
copper are ideal candidates to replace noble metal catalysts with cheap
and abundant transition metals. According to Nakagawa et al. [10], the
Glycerol is produced as a byproduct from biodiesel manufacturing
process at a ratio of 1:10 [1]. According to reports from the Organisation
for Economic Co-operation and Development (OECD) together with the
Food and Agriculture Organization of the United Nations, the global
biodiesel production should increase from 37 blnL in 2016 to 40.5 blnL
by 2026 [2]. This situation has led to consistently low glycerol prices,
and new uses of glycerol as a raw material for the production of chemical
products of added value have been researched [3].
Ni/γ-Al2O3 ability for hydrogenation–dehydrogenation along with its
acid properties constitute a determinant bifunctional property which is
needed to obtain important secondary chemicals like hydroxyacetone,
pyruvaldehyde and acetaldehyde. Furthermore, Sun et al. [11] reported
that transition metals such as Pt, Ru, and Cu work as hydrogenation
catalysts, and also are involved in the glycerol dehydration into
hydroxyacetone. It is known that the introduction of a second metal in a
catalyst system may provide significant changes in the catalytic activity,
selectivity and in the formation of coke as compared with the mono-
metallic ones [12]. Miranda et al. [13] studied a NiCu/γ-Al2O3 catalyst
and concluded that exist electronic and geometric effects between both
metals which affect their catalytic activity in the conversion of glycerol.
In the present research, the study of the NiCu catalyst is focused in
obtaining hydroxyacetone.
Hydrogenolysis, dehydration and dehydrogenation of glycerol pro-
vides different routes for the production of commodity chemicals such as
hydroxyacetone, pyruvaldehyde, lactic acid, and others compounds [4].
Hydroxyacetone is widely used in the textile manufacturing, cosmetic
industry, and food industry. Furthermore, hydroxyacetone can be used
to obtain products such as propylene glycol by hydrogenation, or
acrolein by dehydration. It is also used to synthesize propionaldehyde,
acetone and derivatives of furan [5].
As was mentioned before, the nature of the support is also important
in the hydroxyacetone production [14]. Mane et al. [8] reported a study
on the mechanism of liquid-phase glycerol dehydration into hydrox-
yacetone over Cu supported catalysts. They concluded that glycerol
dehydration to hydroxyacetone is not only catalyzed by metal but also
by acid sites. Different supports have been studied in the glycerol
dehydration [15–17]. Previous works have reported the production of
hydroxyacetone by dehydration of glycerol over solid acid catalysts [18,
19]. It has been reported that on Lewis acid sites glycerol is selectively
Hydroxyacetone can be produced by a direct dehydration of glycerol
[6,7] or by dehydrogenation followed by dehydration [8]. In different
catalysts, the metal phase has both hydrogenating and dehydrogenating
functions, and the acid sites of the support has both hydration and
dehydration functions, both phases forming together a bifunctional
catalyst. Many noble metal catalysts have been studied in the glycerol
conversion with good results [9]. However, metals like nickel and
* Corresponding author.
E-mail address: barbara.mirandamorales@ucr.ac.cr (B.C.M. Morales).
https://doi.org/10.1016/j.cattod.2020.11.025
Received 24 May 2020; Received in revised form 17 October 2020; Accepted 23 November 2020
Available online 24 December 2020
0920-5861/© 2020 Elsevier B.V. All rights reserved.
Please cite this article as:
Bárbara C. Miranda Morales,
Braulio Andrés Ortega Quesada,
Catalysis Today,
https://doi.org/10.1016/j.cattod.2020.11.025

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
dehydrated to hydroxyacetone, while in Bronsted acid sites glycerol is
converted to acrolein [20,21]. In this research, it was decided to
continue the use of the γ-Al2O3 as support which has Lewis acid sites,
and its excellent dehydrating activity has been proved. Moreover, ac-
cording to Bagheri et al. [22], TiO2 is an alternative support material for
heterogeneous catalyst due to characteristics such as its high surface
area, the chemical stability and acid-base property. Therefore, TiO2 was
also selected as support in this research to study its catalytic activity in
the glycerol conversion.
standard 3 mm holey carbon-coated copper grid and letting the alcohol
evaporate at RT.
The specific surface areas, cumulative pore volumes and average
pore diameters of the samples were measured by the BET method using
N2 adsorption/desorption at 77 K on an Autosorb iQ C XR Quantach-
rome Instrument. First, each sample was placed in a 9 mm cell and
degassed at 80 ◦C for 30 min, at 120 ◦C for 720 min and at 120 ◦C for 120
min. It was used a ramp of 5 ◦C/min. Then the nitrogen adsorption/
desorption analysis was programmed for the following relative pres-
sures: surface area determination (0.025 to 0.3 torr), adsorption (0.35 to
0.99 torr), and desorption (0.98 to 0.2 torr).
Given the commercial importance of hydroxyacetone and abundance
of glycerol, the present research is devoted to study the catalytic activity
of cheap and abundant metals such as Cu and Ni catalysts over titania
and alumina supports to obtain hydroxyacetone from glycerol. The
physical and chemical properties of the samples were examined by N2-
physisoption, XRD, SEM, EDS, TEM, and NH3-TPD.
The acidity of the catalysts was determined by temperature pro-
grammed desorption of NH3 (NH3-TPD) on an Autosorb iQ C XR
Quantachrome Instrument equipped with a TCD detector. Each sample
(0.15 g) was placed in a 9 mm cell and heated at a temperature of 120 ◦C
at a rate of 5 ◦C/min for 30 min in He flow. The sample was then treated
with 5%H2/N2 at 450 ◦C at a rate of 10 ◦C/min for 1 h. The gas was
subsequently changed to He and the oven was cooled to 100 ◦C. Then a
10.02 % NH3/He mixture was introduced for 40 min. Finally, NH3
desorption was measured by heating the sample from 100 ◦C to 700 ◦C at
a rate of 10 ◦C/min in He flow.
2. Experimental
2.1. Materials
Glycerol (Sigma-Aldrich, 99.5 %) was used as reagent in the dehy-
dration reaction, and Nitrogen UAP (99.99 %) was used as carrier gas.
For the analysis of the reaction products hydroxyacetone (90 %), pyr-
uvaldehyde (40 %), lactic acid (85 %), lactide (99 %), ethylene glycol
(96 %), acetic acid (99 %), 1,2-propanediol (99 %), 1,3-propanediol (97
%), methanol (99.90 %), ethanol (99 %), pyruvic acid (98 %), and
acrolein analytical standard were purchased from Sigma-Aldrich and
used as received. Catalysts were prepared by using the corresponding
metallic precursors Al(NO3)39H2O (98 %), Cu(NO3)22.5H2O (98 %), and
Ni(NO3)26H2O (98.5 %), all bought from Sigma-Aldrich. In addition,
NaOH (98.0 %) from Sigma-Aldrich was also used for their synthesis.
TiO2 (99.5 %) was bought from Sigma-Aldrich and used as received.
Hydrogen (5% balance in Argon) was used to reduce the catalysts.
The FTIR analysis of the spent catalyst was performed using a
Spectrum 1000 spectrometer from Perkin-Elmer in a range of 4400 cmꢀ 1
to 500 cmꢀ 1, with a resolution of 4 cmꢀ 1. Raman spectra of the spent
catalyst was obtained by using an ACC250 G600 spectrometer.
Approximately, 10 mg of the sample was excited using a green lightning
laser operating at 532 nm and a power of 0.13 mW.
2.4. Conversion of glycerol
The catalytic conversion of glycerol was carried out in gas phase in a
quartz fixed bed down flow reactor at 280 ◦C under atmospheric pres-
sure for 5 h. The reaction temperature was chosen based on the ther-
modynamic analysis. An aqueous solution of glycerol (5 v/v %) was
pumped by a syringe pump with a flow of 5 mL/h, and N2 was used as
carrier gas (150 mL/min). Typically, 200 mg of sample in the form of
pellets, with size ranging 2–3 mm, were loaded in the quartz reactor and
the reactor was heated to the desired temperature. It was screened eight
catalysts.
2.2. Catalyst preparation
The γ-Al2O3 support was prepared by precipitation of Al(NO3)39H2O
and NaOH. The precipitate was dried in an oven at 100 ◦C for 12 h, and
then calcined at 500 ◦C for 4 h. The catalysts containing 10 wt% of metal
were prepared by incipient wetness impregnation of Cu and/or Ni over
γ-Al2O3 and TiO2 supports. After impregnation, the catalysts were dried
at 120 ◦C for 12 h and then reduced at 450 ◦C in a flow of 150 ml/min
with 5% H2/Ar for 3 h. It was decided to study a Cu:Ni atomic ratio of 4
because according to a previous research [13], a higher amount of Ni
could lead to produce a high amount of hydrogenolysis products such as
methane, and a high amount of Cu could lead to cover completely the
effect of Ni.
The condensed products were trapped in an ice bath condenser every
20 or 30 min of reaction and analyzed by HPLC (Agilent LC1220 In-
finity) equipped with an Aminex HPX-87H column (30 m x 7.8 mm), an
ultraviolet detector (UV) and a refractive index (RID) detector. The
mobile phase was deionized and filtered water with pH controlled to 2.2
by addition of sulfuric acid with flow of 0.6 ml/min and a pressure of 50
bar. The temperature of the HPLC column was 65 ◦C and 30 min of
analysis for each chromatogram. Calibration curves of these analytes
were constructed to analyze the product concentrations.
2.3. Catalyst characterization
The gaseous products were analyzed by GC (Shimadzu GC-2014)
equipped with an FID detector and a Carboxen 1010 Plot column (30
m x 0.53 mm). Helium was used as the carrier gas at a flow of 117.4 ml/
min. The operational temperature of the GC for the detector and column
oven was 230 ◦C, and an injection pressure of 300 kPa.
X-ray diffraction (XRD) analysis of the fresh catalysts were recorded
using an X-ray Panalytical Empyrean diffractometer with an angular 2θ
diffraction range between 15◦ and 92◦. The samples were placed in a
sample holder type zero-background. Cu Kα radiation (λ =1.54 Å) was
The conversion of the glycerol was defined as follows:
obtained from a copper X-ray tube operated at 45 kV and 40 mA.
The catalysts were analyzed by scanning electron microscopy (SEM)
using a S-3700 N Hitachi microscope operating at an acceleration
voltage of 15.0 kV and an emission current of 85,000 nA. The catalyst
powder was spread on a graphite tape and coated with a layer of gold to
make it conductive. Furthermore, the samples were analyzed by energy-
dispersive X-ray spectroscopy (EDS) using a working distance (WD)
between 10 mm–12 mm and an analysis time of 100 s.
Moles of glycerol reacted
Initial moles of glycerol
Conversion of glycerol(%) =
x100
The selectivity to each product was defined as follows:
Moles in specific product
Moles in all products
Selectivity(%) =
x100
Transmission electron microscopy (TEM) analysis was performed in
a JEOL JEM-2100 electron microscope operating at an accelerating
voltage of 200 kV. The samples were prepared by dispersing the as-
prepared catalysts in alcohol and then dropping the suspension on a
2.5. Statistical analysis
It was decided to evaluate the influence of two variables (reaction
temperature and gas flow) on the average selectivity of hydroxyacetone
2

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
Table 1
Variables and levels evaluated using a factorial design 22.
Variable
Level 1
Level 2
Temperature (◦C)
280
150
250
75
Gas flow (mL/min)
Fig. 2. Conversion of glycerol over different catalysts.
fraction of hydroxyacetone is sensitive to temperature. It decreases as
the temperature increases. A hydroxyacetone molar fraction of 0.4 to 0.5
is obtained at temperatures between 100ꢀ 280 ◦C. However, its molar
fraction decreases a about 0 from 300 ◦C to 350 ◦C. In contrast, the
acrolein molar fraction increases until 0.33 at 350 ◦C. Furthermore, the
molar fraction of water is higher than the values obtained for hydrox-
yacetone and acrolein. In the reaction of glycerol to acrolein are formed
two molecules of water, and in the case of hydroxyacetone one molecule
of water is released per molecule of glycerol [23]. In this case, the molar
fraction of water increases from 0.53 to 0.66 as the temperature in-
creases. The methanal was not quantified as it was irrelevant. It is
important to observe the discontinuity in the trend of the curves at
around 275 ◦C which is due to the change of state of the glycerol solution
from the liquid to the vapor state. In this research, a reaction in gas
phase is studied, so it was decided to run the tests at 280 ◦C.
Fig. 1. Mole fraction of products as function of reaction temperature of glycerol
dehydration obtained by the thermodynamic analysis.
by performing a factorial design 22. The levels studied are shown in
Table 1. The response was the average selectivity of hydroxyacetone.
The reactions were carried out as mentioned in section 2.4.
2.6. Catalyst regeneration
The Cu/γ-Al2O3 catalyst was chosen for a regeneration test. After 5 h
of reaction the catalyst was regenerated in air flow (150 ml/min) from
RT to 450 ◦C at 5 ◦C/min and kept at 450 ◦C for 3 h. The catalyst was
then reduced under 5% H2/Ar in a flow of 150 ml/min at 450 ◦C for 3 h.
Finally, the conversion of glycerol was performed again for 5 h.
4.2. Conversion of glycerol
In the first part of this research it was studied different catalysts in
the conversion of glycerol to hydroxyacetone. The glycerol conversion
versus time of reaction for the catalysts is shown in Fig. 2. It can be seen
that the catalysts containing γ-Al2O3 as support present higher conver-
sion values than the catalysts containing TiO2. It is evident that the Cu/
γ-Al2O3 catalyst presents the highest conversion ranging from 86 % to 40
% at 5 h of reaction, while the Cu/TiO2 catalyst shows a conversion
between 60 % and 20 %. For the monometallic catalysts containing Ni
(Ni/γ-Al2O3, Ni/TiO2), it is observed that this metal on the two supports
generates a decrease in the conversion from around 40 % to 10 % since
the first hour of reaction. Nickel tends to generate ruptures between
3. Thermodynamic analysis
A thermodynamic analysis of the reaction was carried out in order to
obtain valuable information about the effect of temperature on the
progress of reaction and on the equilibrium composition. According to
Pala et al. [23], acrolein and hydroxyacetone may be formed as the
primary reaction products from glycerol dehydration under
a
non-reactive atmosphere (N2), while methanal and ethanal may be
formed at minor concentration. Then it was decided to the use acrolein,
hydroxyacetone and methanol for the thermodynamic study. The anal-
ysis was performed in the UniSim Design R390.1 software using a Gibbs
reactor. The Peng Robinson Stryjeck-Vera (PRSV) thermodynamic
package was used for the determination of the properties of the sub-
stances involved. For the analysis, a reagent concentration of 1 mol/L
was used, and the reaction temperature was varied from 100 ◦C to 360
◦C in order to obtaining the molar fraction of the products.
–
C
C bonds, and it also tends to cause poisoning of the catalyst as it
generates early formation of carbon which reduces the catalytic activity
[24]. This could explain the low conversion observed in Fig. 2 for
monometallic nickel catalysts.
It was analyzed the effect of combining the metals Cu:Ni in a molar
ratio of 4:1 over both supports. The behavior of conversion for the CuNi/
TiO2 catalyst is remarkably similar to that of the Cu/TiO2 catalyst
(Fig. 2). However, in the case of the CuNi/γ-Al2O3 catalyst, the incor-
poration of Ni seems to cause a negative effect on the conversion of
glycerol which decreases from 53 % to 30 % during the 5 h of reaction.
For the reaction under study, the metal impregnated on the support
4. Results and discussion
4.1. Thermodynamic analysis
–
–
C
should have more affinity for the cleavage of C O bonds than C
A thermodynamic analysis of the reaction was carried out in order to
obtain valuable information about the effect of reaction temperature on
the progress of reaction. The production of hydroxyacetone from glyc-
erol is an exothermic reaction [23]. In Fig. 1 is shown that the molar
bonds, because the glycerol and hydroxyacetone molecules both have
–
three moles of carbon, so it is not convenient to break C C bonds in this
reaction.
–
Moreover, the cleavage of C C bonds can provoke the deposition of
3

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
Fig. 3. Time-on stream product selectivity for a) Cu/γ-Al2O3, b) Cu/ TiO2, c) CuNi/γ-Al2O3, d) CuNi/TiO2, e) Ni/γ-Al2O3, f) Ni/TiO2, g) γ-Al2O3, h) TiO2.
4

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
carbon on the catalysts and Ni causes it, then it could be a reason for the
deactivation observed in the bimetallic catalysts. However, according to
Table 2
Hydroxyacetone concentrations (g/L) obtained for catalysts tested at different
reaction time.
–
a previous work [13], the active Ni sites for rupture of C C bond of
glycerol may require a large ensemble of adjacent Ni atoms on the
catalyst surface for the hydrogenolysis route. The probability of finding
a suitable array of active Ni atoms to accommodate the reactive mole-
cule may greatly decreases when Cu is dispersed on the Ni catalyst
surface. In this way, the deactivation observed (Fig. 2) for the CuNi/-
γ-Al2O3 catalyst is lower than that observed for the Ni/γ-Al2O3.
Furthermore, the activity of γ-Al2O3 and TiO2 supports without metal
was studied. In the case of the γ-Al2O3, the conversion decreases from 51
% to 21 % in the first hour of reaction, and the conversion of glycerol
over TiO2 decreases from 46 % to 7 %. According to Braga et al. [25], the
main role of this reaction is the catalyst metal, and the support has a
secondary role, but equally important in obtaining the product of in-
terest, creating a synergistic effect between them. Hirunsit et al. [26]
indicate that the alumina support facilitates copper to be more active
toward interacting with glycerol and hydroxyacetone intermediate
species which results in an improved catalytic activity. According to the
results obtained in this research (Fig. 2), it is not possible to obtain a
high glycerol conversion with only the presence of the support. In
addition, the reaction was carried out without the presence of a catalyst
and the conversion was as low as that obtained with the TiO2 support at
5 h of reaction.
Time, h
Catalyst
0.70
2.50
5.00
Cu/γ-Al2O3
Cu/TiO2
12.35
7.14
11.38
7.49
0.63
2.34
1.92
1.68
10.50
5.90
6.43
4.68
0.45
0.70
1.33
1.25
6.18
4.68
5.72
2.92
0.62
0.64
1.15
1.32
CuNi/γ-Al2O3
CuNi/TiO2
Ni/γ-Al2O3
Ni/TiO2
γ-Al2O3
TiO2
selectivity to hydroxyacetone ranging from 55 % to 45 %. However, the
selectivity to lactic acid increases to around 28 % in 5 h of reaction, and
it was accompanied by a decrease in pyruvaldehyde selectivity to 20 %.
According to Auneau et al. [29], the lactic acid may be formed from
pyruvaldehyde via a Cannizzaro reaction under basic conditions.
Furthermore, Albuquerque et al. [30] indicate that hydroxyacetone can
be converted into lactic acid by oxidation/intramolecular rearrange-
ment, which may occur on solid basic catalysts. Tanaka et al. [31]
showed that both acidic and basic active sites are present on TiO2 sur-
faces and reduction of TiO2 increases its basic character. In addition,
pyruvic acid can be converted into lactic acid by hydrogenation [27].
The introduction of Ni in this catalyst (CuNi/TiO2) provoked an increase
in the selectivity of ethanol (10 %) and acetic acid (5 %), and the
selectivity to hydroxyacetone decreased from 40 % until 15 % in 5 h
(Fig. 3d). This fact could be explained by the hydrogenolytic activity of
Ni. It can also be observed for the monometallic Ni/γ-Al2O3 and Ni/TiO2
catalysts in Fig. 3e and f. The hydroxyacetone decreases from 30 % to 10
% and from 50 % to 15 %, respectively. As it is observed, nickel catalysts
are not selective towards hydroxyacetone in this reaction system.
However, the lactic acid selectivity was favored (40 %) for the
Ni/γ-Al2O3, and lactic acid (30 %) and acetic acid (15 %) experiment an
increase for the Ni/TiO2 catalyst. It can be remarked that the incorpo-
ration of Ni increases the C-C breaking. From a previous work [27], the
conversion of hydroxyacetone with the Ni/γ-Al2O3 catalyst was about 16
% and the major product was lactic acid with a selectivity of 61 %.
It was shown in Fig. 2 that the glycerol conversion obtained with the
supports γ-Al2O3 and TiO2 was very low. In Fig. 3g and h is shown that
the selectivity to dehydration-dehydrogenation products decreases with
respect to the other catalysts. Akizuki and Oshima [32] obtained low
yields of hydroxyacetone on the dehydration of glycerol using TiO2 as
catalyst. Moreover, the selectivity to hydrogenolysis products (ethanol
and acetic acid) increases for these supports. Finally, the reaction was
carried out without catalyst during 2 h. It was obtained a selectivity of
100 % to ethanol at the temperature and pressure conditions of the
system which may be produced from the direct hydrogenolysis of
glycerol.
The selectivity profiles obtained for the catalysts are shown in Fig. 3.
A test was carried out with the Ni/γ-Al2O3 catalyst to verify the for-
mation of methane. However, the methane formation detected by GC
analysis was very small and it would be difficult to quantify it. There-
fore, it was decided to continue analyzing only the liquid phase
obtained.
The main products in the condensable phase during conversion of
glycerol at 280 ◦C under atmospheric pressure were: hydroxyacetone,
pyruvaldehyde, lactic acid, acetic acid, ethylene glycol, ethanol, and
traces of pyruvic acid. It can be seen that the selectivity to dehy-
dration–dehydrogenation products such as hydroxyacetone and pyr-
uvaldehyde is higher for the catalysts containing Cu. In the case of the
Cu/γ-Al2O3 catalyst, it is observed in Fig. 3a that there is a preference
towards the formation of hydroxyacetone with a selectivity between 70
% and 40 %. It is also evident that there is a balance between hydrox-
yacetone and pyruvaldehyde. Miranda et al. [13] showed in their study
that hydroxyacetone may be formed not only from direct dehydration of
glycerol but also from the hydrogenation of pyruvaldehyde. Also, pyr-
uvaldehyde may be formed from the dehydrogenation of hydrox-
yacetone. Furthermore, traces of pyruvic acid were detected, this
product may be formed via oxidation from pyruvaldehyde.
In addition, hydrogenolysis products such as ethanol, ethylene glycol
(EG) and acetic acid were detected with values lower than 5 % of
–
selectivity (Fig. 3a) which could be due to copper tends to cleavage C
O
–
bonds instead of C C bonds. According to Miranda et al. [27], EG was
not detected in the experiments when hydroxyacetone, pyruvaldehyde
and lactic acid were used as reactants, suggesting that EG is formed by
direct hydrogenolysis of glycerol. However, the acetic acid may be ob-
According to these results (Fig. 3), it is possible to attribute the
higher formation of hydroxyacetone to the presence of copper in the
catalysts which favors the glycerol dehydration in a terminal hydroxyl
group [33]. Moreover, the presence of acid sites in the support is
required in order to give way to a dehydration reaction [34]. According
to Alhanash et al. [21], Bronsted acid sites are required for the formation
of acrolein, and hydroxyacetone is obtained over Lewis acid sites.
Acrolein was not observed in the products of the reaction in the present
research. It is known that the γ-Al2O3 support have strong Lewis acid
sites [35]. On the other hand, Watanabe et al. [36] indicate that TiO2 has
both acid and basic sites. The increase in the formation of lactic acid for
the Cu/TiO2 catalyst (Fig. 3b) may be due to the presence of basic sites.
Akizuki and Oshima [32] mentioned that one possible production route
for lactic acid is direct production from glycerol catalyzed by basic sites
of metal oxide catalysts. According to Sato et al. [33], the addition of
acidic oxide support such as Al2O3 to copper effectively promoted
–
tained from the C C cleavage of the hydroxyacetone [28].
The CuNi/γ-Al2O3 catalyst shows a similar trend in products selec-
tivity to the Cu/γ-Al2O3 catalyst according to the results showed in
Fig. 3c. It could be due to the low proportion of nickel used in this
catalyst (Cu/Ni = 4). However, it is observed an increase to about 7 % in
the selectivity to lactic acid. This could be because the Ni metal catalyzes
the hydrogenation of pyruvic acid. Miranda et al. [13] studied CuNi
catalysts in different molar ratios. They determined that the dehydration
and dehydrogenation products such as hydroxyacetone, pyruvaldehyde,
lactide and pyruvic acid started to be produced when the Cu content
increased, and in the monometallic Cu sample the hydrogenolysis
products such as CH4, EG, acetaldehyde and acetic acid were totally
suppressed.
In Fig. 3b is observed that the Cu/TiO2 catalyst also shows a high
5

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
Fig. 5. SEM micrographs and EDS content for catalysts a) Cu/γ-Al2O3 and b)
Cu/TiO2.
In the Fig. 4a is observed the XRD pattern of the Cu/γ-Al2O3 catalyst.
Two distinct crystallographic phases are observed, which correspond to
metallic Cu, and the corresponding γ-Al2O3 phase of the support. The
peaks at around 2θ = 19.6◦, 37.5◦, 45.7◦ and 66.7◦ were assigned to
γ-Al2O3 phase [27,37]. The broad and poorly defined peaks of the sup-
port are an indication of low crystallinity in this catalyst [37]. The peaks
at around 2θ = 43.3◦, 50.4◦, 74.1◦ and 90◦ (Fig. 4a) are associated with
the metal Cu phase [38–40]. The peaks of the Cu are identified to cor-
responding (1 1 1), (2 0 0), (2 0 2) and (3 1 1) crystal planes. Peaks
attributable to copper oxides were not detectable in these samples [41,
42].
Fig. 4. X-ray diffraction patterns for the fresh catalysts (a) Cu/γ-Al2O3 (b)
Cu/TiO2.
hydroxyacetone selectivity while basic support shows low selectivity
towards hydroxyacetone. Furthermore, Albuquerque et al. [30] indi-
cated glycerol may be dehydrated to hydroxyacetone on solid acid cat-
alysts and this ketone can be converted into lactic acid on solid basic
catalysts.
In Table 2 is shown the concentration of hydroxyacetone versus the
reaction time for all the catalysts studied. It can be seen that the Cu/
γ-Al2O3 and CuNi/γ-Al2O3 catalysts reach the higher hydroxyacetone
concentration in the first hour of reaction. However, all these catalysts
suffer deactivation (Fig. 2) and their selectivity as well as the concen-
tration of hydroxyacetone decrease during the 5 h of reaction.
In the Fig. 4b is observed the XRD pattern for the Cu/TiO2 catalyst. It
can be seen the crystallographic phases corresponding to anatase and
rutile from the TiO2 support. The peaks at around 2θ = 25.4◦, 36.95◦,
37.77◦, 38.0◦, 48.0◦, 54.7◦, 55.1◦, 62.7◦, 68.5◦, 70.2◦, 75.1◦, and 82.8◦
were assigned to anatase [43–45]. Moreover, the peaks at around 2θ =
27.4◦, 36.1◦, 41.3◦, and 56.6◦ were assigned to rutile [43]. In addition,
seeing the peaks in Fig. 4b, it can be interpreted that this catalyst pre-
sents a higher crystallinity than the Cu/γ-Al2O3, since the peaks are thin
and well defined [46]. In Fig. 4b are observed the same peaks corre-
sponding to metallic Cu as in the Cu/γ-Al2O3. However, these peaks are
4.3. Catalyst characterization
It was decided to characterize the Cu/γ-Al2O3 and Cu/TiO2 catalysts
because of their higher selectivity and concentration of hydroxyacetone.
XRD was used to identify the crystalline phases present in the samples.
Table 3
Textural properties and average crystallite size of the fresh Cu monometallic catalysts.
Catalyst
Surface area BET (m2/g)
Pore volume (cm3/g)
Pore size (nm)
Average crystallite size (nm)
Cu/γ-Al2O3
160
45
0.28
0.25
7.26
22.4
41.08
46.97
Cu/TiO2
6

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
Fig. 6. TEM images for (a, c) Cu/γ-Al2O3 and (b, d) Cu/TiO2.
of lower intensity. This could indicate that the Cu phase is much higher
dispersed on TiO2 than γ-Al2O3.
The Cu/γ-Al2O3 catalyst presents a surface area of 160 m2/g against 45
m2/g of the Cu/TiO2 catalyst. Bagheri et al. [22] indicate that one
drawback in using TiO2 as a heterogeneous catalyst is its small specific
surface area. Akizuki and Oshima [32] determined BET surface areas
between 15.2 m2/g and 44.4 m2/g for WO3/TiO2 catalysts. Moreover,
Furthermore, the corresponding average crystallite size was esti-
mated using the Scherrer equation regarding the (111) plane associated
with the angle of 43.35◦ for metallic copper. The Cu/γ-Al2O3 catalyst
presents an average crystallite size of 41.08 nm while the Cu/TiO2
catalyst presents an average crystallite size of 46.97 nm (Table 3).
For both samples, SEM y EDS were performed by analyzing different
regions on the catalysts. In the SEM images (Fig. 5) can be observed that
the surface morphology is rough and presents particles of different sizes
and shapes distributed randomly in both catalysts. It can be also
observed different bright color particles which are associated to copper
deposited over the support surface. In the EDS analysis, it was detected
6.727 wt% of copper for Cu/γ-Al2O3 and 6.582 wt% for Cu/TiO2.
The fresh Cu/γ-Al2O3 and Cu/TiO2 catalysts were characterized by
TEM (Fig. 6). It can be seen the shape and density of Cu particles
distributed over γ- Al2O3 and TiO2 supports. In Fig. 6a is showed a low
magnification TEM image of the monometallic Cu/γ- Al2O3 catalyst. It
shows a dense particle of Cu with sphere-like shape over the γ-Al2O3
support. In Fig. 6b is observed Cu nanoparticles with higher electron
contrast corresponding to Cu/TiO2 catalysts. Fig. 6c shows Cu particles
well distributed over the γ- Al2O3 support. Finally, in Fig. 6d is observed
particles with higher electron contrast corresponding to metal particles
well-dispersed along the TiO2 support.
´
Gutierrez-Alejandre et al. [47] determined that Al-containing samples
have much higher surface areas than the pure titania sample, confirming
the ability of alumina to act as a morphology promoter for anatase. The
larger surface area of the Cu/γ-Al2O3 catalyst is an advantage for its
catalytic activity.
The N2 adsorption-desorption isotherms of the monometallic Cu
catalysts are shown in Fig. 7a. The Cu/γ-Al2O3 catalyst exhibit the type
IV isotherm according to the Brunauer-Deming-Deming Teller (BDDT)
classification, which exhibits the condensation and evaporation step
characteristic of mesoporous materials. The isotherm exhibits a hyster-
esis loop of the H4 type. On the other hand, the Cu/TiO2 catalyst exhibit
the type V isotherm according to the BDDT classification (Fig. 7a), and
its isotherm exhibits a hysteresis loop of the H5 type. It has a distinctive
form associated with certain pore structures containing both open and
partially blocked mesopores [48].
The pore diameter distribution curves of the samples are shown in
Fig. 7b. Both catalysts present mesopores. The Cu/γ-Al2O3 catalyst
exhibit a pore size of 7.26 nm against 22.4 nm of the Cu/TiO2 sample.
The pore size distribution curves of all the samples are bimodal.
TPD profiles of ammonia for Cu/γ-Al2O3 and Cu/TiO2 catalysts are
The N2-physisorption results of the samples are depicted in Table 3.
7

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
Fig. 9. Pareto chart of the Standardized Effects (response is hydroxyacetone
selectivity,
α = 0.05).
acidic sites), while the higher temperature region to desorption of
terminally bound NH3 on tri-coordinated aluminium (strong Lewis acid
sites) [35]. The NH3-TPD analysis for the supports without copper was
not performed, however, a contribution of copper to the moderate acid
sites cannot be discarded [49]. Hirunsit et al. [26] indicate that the acid
Fig. 7. (a) Adsorption-desorption isotherms, and b) Pore size distribution for
the Cu/γ-Al2O3 and Cu/TiO2 catalysts.
Fig. 8. Ammonia TPD profile for the catalysts a) Cu/γ-Al2O3 and b) Cu-TiO2.
shown in Fig. 8. The strength of the acid sites can be classified in weak
(<200 ◦C), moderate (200ꢀ 450 ◦C) and strong (>450 ◦C) [35]. The
NH3-TPD for the fresh Cu/γ-Al2O3 catalyst (Fig. 8a) showed two main
desorption regions at 100ꢀ 450 ◦C and at 450ꢀ 650 ◦C, indicating sites
with different acid strengths. It is also observed a shoulder at around 300
◦C. The lower desorption temperature region is attributed to desorption
of NH3 bridge-bound to penta-coordinated aluminium (weak Lewis
Fig. 10. (a) Average selectivity and (b) Glycerol conversion for the Cu/γ-Al2O3
catalysts at different reaction conditions.
8

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
Fig. 11. (a) Catalytic behavior, and (b) Time-on stream product selectivity in glycerol conversion of the Cu/γ-Al2O3 catalyst for two cycles of regeneration.
function was mainly from Al2O3 surface since the NH3 desorption curves
appearing from 150 to 700 ◦C from the catalysts before and after copper
loading were nearly comparable. As shown in Fig. 8b, the Cu/TiO2
presents a main desorption region at 100ꢀ 450 ◦C and a very small re-
gion at 500ꢀ 650 ◦C, indicating the presence of mainly weak and mod-
erate acid strengths sites [42]. The acidity was quantified, and it was
Cu/TiO2. These basic sites may contribute to the formation of lactic acid
as was mentioned before [32].
As was observed in Fig. 3b, hydroxyacetone was also produced for
the Cu-TiO2 catalyst which presents weak and moderate acid sites. Based
on a research reported by Suprun et al. [17], hydroxyacetone can be
obtained in the presence of moderate acidic support catalyst. However,
the high surface area and high acidic strength of the Cu/γ-Al2O3 catalyst
could explain its higher hydroxyacetone selectivity compared to the
Cu-TiO2 sample.
obtained that Cu/γ-Al2O3 presented 330.5
μ
mol gꢀ 1 of weak-moderate
acidic sites and 254.7
μ
mol g-1 of strong acidic sites, while Cu/TiO2
presented 427.49
μ
mol gꢀ 1 of weak-moderate acidic sites and 70.6
μmol
g-1 of strong acidic sites. Liu et al. [50] determined that 10 %Cu/Al2O3
presented 14.3
μ
mol gꢀ 1 of weak acidic sites and 182.5
μ
mol g-1 of
mol gꢀ 1 of
4.4. Statistical analysis
strong acidic sites, while 10 %Cu/TiO2 presented 157.1
μ
weak acidic sites and 62.7
μ
mol g-1 of strong acidic sites which present a
It was decided to choose the Cu/γ-Al2O3 catalyst to perform the
second phase of this research because it presented the highest glycerol
conversion and hydroxyacetone selectivity in the first phase. The in-
fluence of reaction temperature and gas flow on the average selectivity
of hydroxyacetone was studied. The tests were carried out according to
the order dictated by a factorial design 22 (Table 1).
similar trend with respect to our data.
As can be seen in Fig. 8, the Cu-TiO2 catalyst presents mainly weak
and moderate acid sites while the Cu/γ-Al2O3 also presents strong acid
´
sites. Gutierrez-Alejandre et al. [51] studied alumina-rich Al2O3–TiO2
mixed oxides and they determined that the titania-modified samples
present two desorption peaks with maxima at 85 ◦C (physisorbed NH3)
and 125 ◦C (weak acid sites) in the low-temperature region. However,
According to the conditions of this analysis, a t-student equal to 2.78
was obtained. In the Pareto chart (Fig. 9) can be observed that the re-
action temperature has a significant influence on hydroxyacetone
selectivity at 95 % confidence. In this case the gas flow and the inter-
action reaction temperature × gas flow had no significant influence on
hydroxyacetone selectivity. However, in Fig. 10a can be observed that
the pure alumina sample showed the clear presence of
a
high-temperature peak. The presence of basic sites was not determined
in this research. Liu et al. [50] determined that 10 %Cu/Al2O3 presented
68.1
μ
mol gꢀ 1 of weak basic sites against 80.2
μ
mol gꢀ 1 of the 10 %
9

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
acid catalyst surface sites and side reactions such as aldol condensation
˜
of carbonyl compounds. Furthermore, Chimentao et al. [54] indicate
that dehydration which takes place on the acid sites is the mainly route
related to the generation of carbon deposition and to the observed
catalyst deactivation.
As catalytic sites blocked by chemisorbed carbon species is generally
reversibly in hydrogen or oxygen atmosphere [55], then a regeneration
test was performed on the Cu/γ-Al2O3 catalyst. The conversion and
selectivity profiles obtained for two reaction cycles are shown in Fig. 11.
In general, it can be seen that the catalyst can regenerate and start
producing again. The product distribution and glycerol conversion just
before the deactivation process and after regeneration were similar in
trend. The catalyst regeneration by oxidation–reduction processes at
450 ◦C was sufficient to regain the catalytic activity.
IR and Raman analysis were done to determine the presence of
carbonaceous deposits on the surface of the used catalyst. In Fig. 12a is
observed the IR spectrum for the Cu/γ-Al2O3 catalyst. It can be seen the
bands at 1640 cmꢀ 1 and at 1405 cmꢀ 1 which are characteristics of
amorphous alumina [56]. Furthermore, there is a band at 3498 cm-1
corresponding to the stretching vibration of OH groups of the water
impregnated on the sample [57]. The characteristic bands of carbon
were not identified [58].
Through Raman, the bands corresponding to deposited carbon usu-
ally appear centered at about 1347 cmꢀ 1 and 1590 cmꢀ 1, which are
associated with the defect and graphite modes, e.g., D-band and G-band,
respectively [27,58,59]. In Fig. 12b can be observed that these bands
were not detected for the Cu/γ-Al2O3 catalyst. Miranda et al. [13] barely
detected the D-band and G-band for a monometallic Cu sample. How-
ever, 0.34 wt.% of carbon was determined by TPO in the used sample.
There could be other causes of the Cu/γ-Al2O3 catalyst deactivation.
Bienholz et al. [60] reported a decrease in glycerol conversion on
CuO/ZnO and concluded that the water loaded into the reactor or the
water formed during the hydrogenolysis deactivated the catalyst. Ac-
˜
cording to Chimentao et al. [54], the partial oxidation of the metal
species due to the aqueous medium may also contribute to some deac-
tivation observed. The water in aqueous solution of glycerol may oxidize
copper and nickel in the first hours of reaction contributing to the
observed deactivation as shown in Fig. 2. Miranda et al. [13] obtained a
lower glycerol conversion (60.8 %) when the monometallic Cu catalyst
was previously calcined at 450 ◦C instead of reduced (67.4 %). It must be
also considered that the support surface may also suffer hydroxylation
by both the water produced during dehydration step in the reaction and
that already present in the reaction medium [26].
Fig. 12. (a) FTIR and (b) Raman spectra for the used Cu/γ-Al2O3 catalyst.
the average selectivity of hydroxyacetone is the highest at 250 ◦C and 75
ml/min. Then the average selectivity increases as the temperature and
the gas flow decrease. As was evidenced by the thermodynamic analysis
(Fig. 1), a decrease in temperature favors the selectivity of hydrox-
yacetone. It is proven that the formation of hydroxyacetone is an
exothermic reaction, and it is favored at low temperatures [23,33].
Moreover, in Fig. 10b is shown the glycerol conversion obtained at
different reaction temperature-gas flow combinations. It can be seen
that the conversion decreases when the reaction temperature decreases
from 280 ◦C to 250 ◦C. On the other hand, the conversion increases as
the gas flow decreases from 150 ml/min to 75 ml/min. Therefore,
hydroxyacetone selectivity and conversion are favor at low gas flows
and higher residence times.
Furthermore, according to Liu et al. [50], it is likely that the Lewis
acidity of the catalyst was enhanced by the presence of Cu oxides.
Excessive basicity or acidity of the catalyst might block the catalytic sites
by strongly adsorbed reactants which likely deteriorate the conversion
and product selectivity.
5. Conclusions
According to the thermodynamic analysis, the molar fraction of
hydroxyacetone decreases as the temperature increases. The higher
glycerol conversion and selectivity to hydroxyacetone were obtained
with the monometallic Cu/γ-Al2O3 and Cu/TiO2 catalysts. From the
statistical analysis is observed that the reaction temperature has a sig-
nificant influence on hydroxyacetone selectivity at 95 % confidence. In
this case the gas flow and the interaction reaction temperature × gas
flow had no significant influence on hydroxyacetone selectivity. The
catalysts suffer deactivation, but there is no presence of carbon depos-
ited in the used catalyst according to IR and Raman analysis, then
deactivation of the catalyst could be associated to oxidation of the
copper and strongly adsorbed reactants. The catalyst regeneration by
oxidation–reduction processes at 450 ◦C was sufficient to regain the
catalytic activity.
4.5. Catalyst regeneration
As can be seen in Fig. 2, the Cu/γ-Al2O3 catalyst suffers deactivation
during the reaction. The deactivation of the Cu/γ-Al2O3 catalyst could
be attributed to carbonaceous deposits. As reported by Corma et al. [28],
the dehydration route markedly contributes to the carbon deposition on
the catalysts. Vasiliadou and Lemonidou [52] observed deactivation by
carbon during glycerol hydrogenolysis on copper catalysts. Moreover,
Atia et al. [53] observed catalyst deactivation during dehydration of
glycerol on heteropolyacid catalysts and they attributed it to the
blocking of the catalyst surface by carbon, which was related to the
catalyst acidity and pore size. According to Suprun et al. [17], the for-
mation of carbon deposits during dehydration reaction of glycerol may
result from consecutive reactions of glycerol, like oligomerization on
10

B.C.M. Morales and B.A.O. Quesada
Catalysis Today xxx (xxxx) xxx
[22] S. Bagheri, N. Muhd Julkapli, S.B.A. Hamid, Sci. World J. 2014 (2014), 727496.
[23] I. Pala Rosas, J.L. Contreras, J. Salmones, C. Tapia, B. Zeifert, J. Navarrete,
CRediT authorship contribution statement
´
T. Vasquez, D.C. García, Catalysts 7 (2017).
´
Barbara C. Miranda Morales: Conceptualization, Methodology,
[24] M.D. Argyle, C.H. Bartholomew, Catalysts 5 (2015) 145–269.
[25] T.P. Braga, N. Essayem, S. Prakash, A. Valentini, J. Braz. Chem. Soc. 27 (2016)
2361–2371.
Investigation, Project administration, Supervision, Data curation,
´
Formal analysis, Writing - review & editing. Braulio Andres Ortega
[26] P. Hirunsit, C. Luadthonga, K. Faungnawakija, RSC Adv. 5 (2015) 11188–11197.
Quesada: Data curation, Software, Investigation, Formal analysis,
˜
[27] B.C. Miranda, R.J. Chimentao, J.B.O. Santos, F. Gispert-Guirado, J. Llorca,
´
Writing - original draft.
F. Medina, F. Lopez Bonillo, J.E. Sueiras, Appl. Catal. B Environ. 147 (2014)
464–480.
[28] A. Corma, P.J. Miguel, A.V. Orchilles, J. Catal. 145 (1994) 171–180.
[29] F. Auneau, C. Michel, F. Delbecq, C. Pinel, P. Sautet, Chem. Eur. J. 17 (2011)
14288–14299.
Declaration of Competing Interest
[30] E.M. Albuquerque, L.E.P. Borges, M.A. Fraga, C. Sievers, ChemCatChem. 9 (2017)
2675–2683.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
[31] T. Tanaka, H. Kumagai, H. Hattori, M. Kudo, S. Hasegawa, J. Catal. 127 (1991)
221–226.
[32] M. Akizuki, Y. Oshima, Ind. Eng. Chem. Res. 51 (2012) 12253–12257.
[33] S. Sato, M. Akiyama, R. Takahashi, T. Hara, K. Inui, M. Yokota, Appl. Catal. A Gen.
347 (2008) 186–191.
Acknowledgments
[34] M. Velasquez, A. Santamaria, C. Batiot-Dupeyrat, Appl. Catal. B Environ. 160-161
(2014) 606–613.
´
´
Braulio Andres Ortega Quesada and Barbara C. Miranda Morales
gratefully acknowledges the Universidad de Costa Rica for the financial
support (Project 325-B8-299). We would also like to acknowledge
CELEQ, LANAMME and CICIMA for the collaboration with character-
ization data acquisition.
[35] J. Lif, I. Odenbrand, M. Skoglundh, Appl. Catal. A Gen. 317 (2007) 62–69.
[36] M. Watanabe, T. Iida, Y. Aizawa, H. Ura, H. Inomata, K. Arai, Green Chem. 5
(2003) 539–544.
[37] V. Thyssen, T.A. Maia, E.M. Assaf, J. Braz. Chem. Soc. 26 (2015) 22–31.
[38] A.R. Naghash, T.H. Etsell, S. Xu, Chem. Mater. 18 (2006) 2480–2488.
ˇ
´
´
´
´
´
[39] Z. Andic, M. Korac, M. Tasic, K. Raic, Z. Kamberovic, Kovove Mater. 44 (2006)
145–150.
References
[40] T. Le Minh, H. Than Quoc An, T.P. Huu, Chem. Cent. J. 7 (2013) 143–150.
[41] H. Li, Y. Huang, S. Cui, Chem. Cent J. 7 (2013) 143–150.
[42] S. He, C. Zhang, M. Yang, Y. Zhang, W. Xu, N. Cao, H. He, Sep. Purif. Technol. 58
(2007) 173–178.
[1] M. Ayoub, A.Z. Abdullah, Renew. Sustain. Energy Rev. 16 (2012) 2671–2686.
[2] OECD/FAO, OECD-FAO Agricultural Outlook 2017-2026, OECD Publishing, Paris,
2017.
[43] J. Wang, J. Yu, X. Zhu, X.Z. Kong, Nanoscale Res. Lett. 7 (2012) 646.
[44] X. Wei, G. Zhu, J. Fang, J. Chen, Int. J. Photoenergy 2013 (2013) 726–872.
[45] P. Srinivasu, S.P. Singh, A. Islam, L. Han, Adv. Optoelectron. 2011 (2011) 539.
[3] R. Ciriminna, C. Della Pina, M. Rossi, M. Pagliaro, Eur. J. Lipid Sci. Technol. 116
(2014) 1432–1439.
¨
[46] P. Ahvenainen, I. Kontro, K. Svedstrom, Cellulose 23 (2016) 1073–1086.
[4] M. Akiyama, S. Sato, R. Takahashi, K. Inui, M. Yokota, Appl. Catal. A Gen. 371
(2009) 60–66.
´
[47] A. Gutierrez-Alejandre, M. Trombetta, G. Busca, J. Ramírez, Microporous
Mesoporous Mater. 12 (1997) 79–91.
[5] M.H. Mohamad, R. Awang, W.M. Yunus, Am. J. Appl. Sci. 8 (2011) 1135–1139.
[6] M.A. Dasari, P.P. Kiatsimkul, W.R. Sutterlin, G.J. Suppes, Appl. Catal. A Gen. 281
(2005) 225–231.
[48] M. Thommes, K. Kaneko, A.V. Neimark, J.P. Olivier, F. Rodriguez-Reinoso,
J. Rouquerol, K.S.W. Sing, Pure Appl. Chem. 87 (2015) 1051–1069.
´
[49] J. Mazarío, P. Concepcion, M. Ventura, M.E. Domine, J. Catal. 385 (2020)
[7] B. Katryniok, S. Paul, F. Dumeignil, ACS Catal. 3 (2013) 1819–1834.
[8] R.B. Mane, A. Yamaguchi, A. Malawadkar, M. Shirai, C.V. Rode, RSC Adv. 3 (2013)
16499–16508.
160–175.
[50] H. Liu, Z. Huang, Z. Han, K. Ding, H. Liu, C. Xia, J. Chen, Green Chem. 17 (2015)
4281–4290.
[9] S. Wang, K. Ying, Y. Zhang, H. Liu, ACS Catal. 3 (2013) 2112–2121.
[10] Y. Nakagawa, K. Tomishige, Catal. Sci. Technol. 1 (2011) 179–190.
[11] D. Sun, Y. Yamada, S. Sato, W. Ueda, Appl. Catal. B Environ. 193 (2016) 75–92.
[12] J.H. Sinfelt, J.L. Carter, D.J.C. Yates, J. Catal. 24 (1972) 283–296.
´
´
[51] A. Gutierrez-Alejandre, M. Gonzalez-Cruz, M. Trombetta, G. Busca, J. Ramıŕ ez,
Microporous Mesoporous Mater. 23 (1998) 265–275.
[52] E.S. Vasiliadou, A.A. Lemonidou, Appl. Catal. A Gen. 396 (2011) 177–185.
[53] H. Atia, U. Armbruster, A. Martin, J. Catal. 258 (2008) 71–82.
˜
˜
[13] B.C. Miranda, R.J. Chimentao, J. Szanyi, A.H. Braga, J.B.O. Santos, F. Gispert-
[54] R.J. Chimentao, B.C. Miranda, J. Szanyi, C. Sepulveda, J.B.O. Santos, J.V.S. Correa,
J. Llorca, F. Medina, Mol. Catal. 435 (2017) 49–57.
Guirado, J. Llorca, F. Medina, Appl. Catal. B Environ. 166-167 (2015) 166–180.
[14] W. Oberhauser, C. Evangelisti, R.P. Jumde, R. Psaro, F. Vizza, M. Bevilacqua,
J. Fillipi, B.F. Machado, P. Serp, J. Catal. 325 (2015) 111–117.
[15] G.J. Highfield, B.J. Moffat, J. Catal. 95 (1985) 108–119.
[16] Y.T. Kim, K.D. Jung, E.D. Park, Microporous Mesoporous Mater. 131 (2010) 28–36.
[17] W. Suprun, M. Lutecki, T. Haber, H. Papp, J. Mol. Catal. A Chem. 309 (2009)
71–78.
[55] C.H. Bartholomew, Appl. Catal. A Gen. 212 (2001) 17–60.
[56] S. Ram, Infrared Phys. Technol. 42 (2001) 547–560.
[57] T. Ishikawa, K. Matsumoto, K. Kandori, T. Nakayama, J. Solid State Chem. 179
(2006) 1110–1118.
[58] Y.G. Cao, X.L. Chen, Y.C. Lan, J.Y. Li, Y.P. Xu, T. Xu, Appl. Phys. A. 71 (2000)
465–467.
[18] E. Yoda, A. Ootawa, Appl. Catal. A Gen. 360 (2009) 66–70.
[19] A. Corma, G.W. Huber, L. Sauvanaud, P. O’Connor, J. Catal. 257 (2008) 163–171.
[20] Z. Wu, Y. Mao, M. Song, X. Yin, M. Zhang, Catal. Commun. 32 (2013) 52–57.
[21] A. Alhanash, E.F. Kozhevnikova, I.V. Kozhevnikov, Appl. Catal. A Gen. 378 (2010)
11–18.
[59] F.F. de Sousa, H.S. A de Sousa, A.C. Oliveira, M.C. Junior, A.P. Ayala, E.B. Barros,
B.C. Viana, J.M. Filho, A.C. Oliveira, Int. J. Hydrogen Energy 37 (2012)
3201–3212.
[60] A. Bienholz, F. Schwab, P. Claus, Green Chem. 12 (2010) 290–295.
11