One-step propylene formation from bio-glycerol over molybdena-based catalysts

Article abstract of DOI:10.1039/c4gc01307g

This work presents a novel, one-step catalytic process, enabling highly selective propylene formation via glycerol hydro-deoxygenation (HDO) reactions. Fe-Mo catalysts, supported on black and activated carbons, are selective towards C-O bond cleavage, thus converting glycerol to propylene with high yields. BET, XRD, TPD-NH₃ and TPD-He methods have been employed for the characterization of the samples. Molybdenum oxide, at its reduced state, is essential for driving selectively the reaction towards complete deoxygenation. The only product of glycerol HDO is propene, in the gas phase, while 2-propenol, propanols and propylene glycol have been detected, among others, in the liquid phase. Under the standard reaction conditions (300°C temperature, 8.0 MPa hydrogen pressure), glycerol conversion exceeds 88% and selectivity to propene reaches 76% after 6 hours of reaction. This study includes the investigation of the operating conditions effect (i.e. reaction time, reaction temperature, catalyst loading and H₂ pressure) regarding glycerol HDO towards propene formation. This journal is

Full text of DOI:10.1039/c4gc01307g

Green Chemistry
View Article Online
View Journal
PAPER
One-step propylene formation from bio-glycerol
over molybdena-based catalysts†
Cite this: DOI: 10.1039/c4gc01307g
Vasiliki Zacharopoulou,a Efterpi S. Vasiliadoua and Angeliki A. Lemonidou*a,b
This work presents a novel, one-step catalytic process, enabling highly selective propylene formation via
glycerol hydro-deoxygenation (HDO) reactions. Fe–Mo catalysts, supported on black and activated
carbons, are selective towards C–O bond cleavage, thus converting glycerol to propylene with high
yields. BET, XRD, TPD-NH3 and TPD-He methods have been employed for the characterization of the
samples. Molybdenum oxide, at its reduced state, is essential for driving selectively the reaction towards
complete deoxygenation. The only product of glycerol HDO is propene, in the gas phase, while 2-prope-
nol, propanols and propylene glycol have been detected, among others, in the liquid phase. Under the
standard reaction conditions (300 °C temperature, 8.0 MPa hydrogen pressure), glycerol conversion
exceeds 88% and selectivity to propene reaches 76% after 6 hours of reaction. This study includes the
investigation of the operating conditions effect (i.e. reaction time, reaction temperature, catalyst loading
and H2 pressure) regarding glycerol HDO towards propene formation.
Received 11th July 2014,
Accepted 19th September 2014
DOI: 10.1039/c4gc01307g
www.rsc.org/greenchem
fermentation or enzymatic catalysis, as well as by ligno-
cellulose via catalytic decomposition and through catalytic
1. Introduction
hydrogenolysis of sorbitol.6–11 Increased biodiesel production
worldwide has led to massive glycerol supply in the market,
thus affecting its price. As a result, glycerol is considered a
low cost, alternative feedstock that can also be used as a
model compound, hence serving as a building block.1 Glycerol
can be converted into various valuable products, i.e. ethylene
and propylene glycols,12–16 acetol,17 acrolein,3,18,19 and fuel
additives.20
The partial or complete removal of oxygen from the glycerol
molecule, via a hydro-deoxygenation (HDO) reaction, leads
to 1,2- and 1,3-propanediol,12 propanols21 and propane22 for-
mation. The latter requires non-active to C–C bond scission
catalysts, because during HDO reactions oxygen is removed in
the form of H2O molecules and not as CO2. Complete removal
of oxygen, along with formation of an unsaturated CvC bond,
thus enabling glycerol conversion to propylene, is a desirable
yet challenging reaction path (Scheme 1). Despite the signifi-
cant progress in glycerol transformation technologies, glycerol
selective conversion to propylene has not yet been thoroughly
explored in the open literature. Nevertheless, bio-propylene
production has received considerable attention over the past
decade.23,24 Propylene is one of the most important olefins
used in the petrochemical industry. It is used as a feedstock
for the production of polypropylene, acrylonitrile, propylene
oxide, cumene, isopropyl alcohol, acrylic acid and acrolein,
among others. Propylene is mainly produced as a co-product
of naphtha cracking and as a by-product of fluid catalytic
cracking.25
The finite stocks of fossil resources, as well as the rise of
environmental issues concerning their increased use, have
triggered the exploration and development of alternative feed-
stocks for the production of transportation fuels and chemical
products.1 Biomass (e.g. wood, agricultural crops, algae, as
well as their residues and wastes)2 and biomass-derived com-
pounds (e.g. glucose, glycerol, sorbitol, and lactic acid) can be
effectively used in the production of liquid fuels and platform
chemicals,3 apart from electricity and heat generation, as they
are a valuable renewable source of carbon.4 The above-
mentioned feedstocks contain C–O and C–H bonds; therefore,
in order for them to be converted into value-added products,
the selective cleavage of C–O over C–C bonds is essential. The
removal of oxygen from these compounds can be realized
through promising hydrogenolysis or hydro-deoxygenation
processes.1,4
Glycerol is one of the top twelve building block chemicals
derived from biomass that can be upgraded to high value-
added products.5 It is typically obtained as a by-product of pro-
cesses such as transesterification of oils and fats, soap manu-
facture, fatty acid production, fatty ester production, microbial
aDepartment of Chemical Engineering, Aristotle University of Thessaloniki, University
Campus, Thessaloniki 54124, Greece
bChemical Process Engineering Research Institute (CERTH/CPERI), P.O. Box 60361
Thermi, Thessaloniki 57001, Greece. E-mail: alemonidou@cheng.auth.gr;
Fax: +30 2310 996184; Tel: +30 2310 996273
†Electronic supplementary information (ESI) available. See DOI: 10.1039/
c4gc01307g
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
Table 1 Synthesized catalytic samples
Fe
(wt%)
Mo
(wt%)
Calcination
atmosphere
Catalyst
Support
Mo/BC_A
Fe/BC_A
Fe–Mo/BC_A
Fe–Mo/AC_A
Fe–Mo/BC_B
Fe–Mo/AC_B
Black carbon
Black carbon
Black carbon
Activated carbon
Black carbon
Activated carbon
—
20.8
—
19.3
19.3
19.3
19.3
Air
Air
Air
Air
N2
3.7
2.7
2.7
2.7
2.7
N2
molybdenum (NH4)6Mo7O24·4H2O (Fischer) at 500 °C for 2 h
under synthetic air flow (100 cm3 min−1).
Scheme 1 Glycerol hydro-deoxygenation pathways.
Catalyst synthesis
Fe–Mo catalysts were synthesized by combining wet impreg-
nation and co-precipitation methods, according to the pro-
cedure described by Fadigas et al.27
Recently, the production of lower hydrocarbons from gly-
cerol has been explored in a patent by Hulteberg and Brandin,
describing the catalytic conversion of glycerol to alkanes and
alkenes, such as ethane, propane and propene, over WO3 on
ZrO2 and Pt on CeO2 catalysts.26 The research study of Cao
et al. also focuses on the production of propene from glycerol,
mainly in a two-step reaction, i.e. glycerol hydrogenolysis to
1-propanol, followed by the well-known process of 1-propanol
dehydration to produce propene, using Ir/ZrO2 and HZSM5
catalysts. Moreover, the same group conducted preliminary
experiments of the direct glycerol conversion to propene
with very interesting results, over a combination of the above-
mentioned catalysts.27
The obtained catalysts were calcined at 200 °C, under
synthetic air or nitrogen flow (100 cm3 min−1), for 2 h. The
specific calcination temperature was selected as carbon sup-
ports may burn or smolder at temperatures above 300 °C. For
the same reason catalysts were calcined not only under air
flow, but also under an inert atmosphere (nitrogen flow). The
atomic Mo/Fe ratio of Fe–Mo/C samples is 4 : 1, while their
loading of Fe and Mo (2.7 and 19.3 wt%, respectively) was not
diverged. Catalysts were pre-reduced, at 500 °C for 0.5 h, under
H2 flow (40 cm3 min−1).
In order to study the contribution of each metal to the cata-
lytic performance, catalysts containing only Fe or Mo were also
prepared. These samples were dried, calcined and pre-reduced
following the same procedure as that with the Fe–Mo catalysts.
All the catalytic samples synthesized in this study are pre-
sented in Table 1. BC stands for Black Carbon while AC for
Activated Carbon. “A” and “B” signify calcination under air
and nitrogen flow, respectively.
In a patent, Fadigas et al. explored the catalytic process for
the production of propene via hydro-deoxygenation of glycerol,
over a Fe–Mo/C catalyst, in the gas phase.28
Triggered by the promising performance of the Fe–Mo/C
catalyst described in the above-mentioned patent, this work
aims at the investigation of this new route of one-step propyl-
ene production, in the liquid phase. The effect of the reaction
parameters is of great importance because they possibly affect
the yield of the desired product. The present study includes
the synthesis, detailed characterization and performance
evaluation of Fe–Mo catalysts, supported on different carbons
(i.e. black and activated carbon). More specifically, the effect of
reaction parameters, i.e. reaction time, reaction temperature,
H2 pressure and catalyst loading, on the reaction performance
has been studied. Product distribution with glycerol conver-
sion was used as a guide in order to explore the various reac-
tion routes and principally the propene formation pathway.
Catalyst characterization
BET surface areas, pore volumes and average pore diameters
were determined by N2 physisorption at 77 K, using an Auto-
sorb-1 Quantachrome flow apparatus. Prior to measurements,
samples were dried under vacuum overnight at 250 °C. The
total pore volume was defined as the single point pore volume
at a relative pressure of p/po = 0.95.
XRD patterns of the catalysts were obtained using a D500
diffractometer (Siemens) X-ray with Cu Kα radiation (λ =
1.54 Å) at a scan rate of 0.02° s−1 (2θ) from 5.0° to 80.0°.
TGA experiments were carried out on an SDT Q600 (TA
Instruments) apparatus. SDT Q600 works in conjunction with
a controller and associated software to constitute a thermal
analysis system. A small amount of the sample was placed in
2. Experimental section
Support materials and precursors
Fe–Mo catalysts were supported on commercially available an aluminum sample cup and then heated from room temp-
Black Carbon Vulcan XC72 (Cabot) and on Activated Carbon erature to 800 °C at 10 °C min−1, under air or nitrogen flow
D10 (Norit). Iron nitrate – Fe(NO3)3·9H2O (Fluka) and MoO3 (100 cm3 min−1).
were used as precursor compounds for Fe and Mo, respectively.
Temperature-programmed desorption of ammonia and
MoO3 was synthesized through calcination of the ammonium temperature-programmed desorption, under an inert (helium)
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
atmosphere, as well as temperature-programmed reduction
Conversion and selectivity values were calculated using the
and thermogravimetric analysis, have also been employed for equations described below:
the sample characterization (see ESI†).
Glycerol conversionð%Þ ¼
ðmoles of Glycerol In ꢀ moles of Glycerol OutÞ ꢁ 100
ðmoles of Glycerol InÞ
Catalyst activity test
Selectivity to product i ¼
Glycerol hydro-deoxygenation experiments were conducted in
a 450 ml stainless-steel batch reactor (Parr Instruments),
equipped with an electronic temperature controller, a heating
mantle, a mechanical stirrer, a pressure gage as well as gas
inlet and outlet lines. The typical reaction conditions were:
300 °C temperature, 8.0 MPa hydrogen pressure, 1.50 g catalyst
weight, 90.0 ml of 2.0 w/w % aqueous glycerol solution, 2.0 h
reaction time and 800 rpm stirring speed. The experimental
Carbon moles of product i ꢁ 100
Total Carbon moles of Gas and Liquid Products
3. Results and discussion
Catalyst characterization
procedure consisted of the following steps: feed solution The BET surface areas of all samples, including the catalyst
preparation, the reactor vessel loaded with the catalyst and the supports’, are demonstrated in Table 2. BET areas of all cata-
feed solution, N2 flushing for 10 min at 0.2 MPa and leak lysts are significantly lower than those of the supports, as
check at 8.0 MPa for 20 min, H2 flushing for 5 min at 0.5 MPa, metals fill or block part of the catalytic pores, thus reducing
the pressurized vessel with 8.0 MPa of H2, the temperature the surface area. As expected, the BET surface of the Fe/BC_A
controller was set at 300 °C and then the heater and the agitator catalyst is close to that of the support due to the low Fe
were turned on. The time at which the temperature reaches the loading. The calcination atmosphere has no significant impact
set point is considered the zero point, t = 0. After 2 h of reac- on the surface area of the catalyst. Upon reduction, BET areas
tion time, the heater and the agitator were turned off and the of the catalysts supported on black carbon are further reduced.
vessel was forced to immediately cool to room temperature.
This decrease is not considered important and suggests that
The reaction time effect has been explored up to 6 h of reac- the applied pre-treatment conditions do not affect the surface
tion. In order to study the effect of temperature, experiments area.29 On the other hand, the BET surface of the catalysts
have been conducted at 260, 280 and 300 °C. The effect of H2 supported on activated carbon is increased upon reduction, as
initial pressure has been tested at 3.0, 5.0 and 8.0 MPa, while previously reported,30 because during treatment molybdenum
that of catalyst loading with 0.75, 1.5 and 3.0 g of samples.
oxide crystals contract and fracture, providing higher surface
Gas and liquid samples were analyzed offline. Liquid area. This difference between the samples supported on
samples were analyzed by GC (Agilent 7890A, FID, DB-Wax different carbons is probably due to the porosity of each
30 m × 0.53 mm × 1.0 μm). Acetonitrile was used as a solvent support. The micropore volume measured for both of the
for the GC analysis. The liquid products detected were: supports validates the presence of microporosity in activated
2-propenol, 1,2-propanediol (propylene glycol), 1,3-propanediol, carbon (0.157 cm3
g
−1), in comparison with black carbon
ethylene glycol, hydroxyacetone (acetol), 1-propanol, 2-propa- (0.0293 cm3 g−1). After reduction, the microporosity of Fe–Mo/
nol, ethanol and methanol. Gas samples were analyzed in an AC catalysts is further increased compared to calcined samples
Agilent GC (7890A, Molecular Sieve and Porapak Q) equipped (Table 2), thus explaining the observed increase in the BET
with a thermal conductivity detector (TCD). The only gas- surface area.
phase product detected under the typical experimental con-
The physisorption isotherms prior to and post reduction
ditions was propylene. The mass balance of all experiments treatment are illustrated in Fig. 1, where the letter R stands for
was 90 5%.
reduced samples. Physisorption isotherms of catalysts, sup-
Table 2 Physicochemical catalyst characterization
BET surface area
TPD-NH3
(m2 g−1
)
Pore volume (cm3 g−1
)
(μmol NH3 m−2 BET)
CO/CO2 ratiob
Reduced
Catalyst
Fresh
Reduced
Fresh
Reduced
Reduced
Activated carbon (AC)
Black carbon (BC)
Mo/BC_A
600
250
127
244
128
244
110
218
—
—
0.39 (0.157)a
1.2 (0.0293)a
0.49
—
—
0.44
1.03
0.010
0.012
0.541
0.034
0.294
0.233
0.315
0.331
—
—
110
239
109
282
105
277
13.9
—
16.9
22.8
16.1
16.4
Fe/BC_A
0.72
Fe–Mo/BC_A
Fe–Mo/AC_A
Fe–Mo/BC_B
Fe–Mo/AC_B
0.59
0.60
0.21 (0.034)a
0.62
0.23 (0.063)a
0.52
0.18 (0.044)a
0.23 (0.061)a
a Micropore volume (cm3 g−1). b Results of TPD-He analysis.
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
Fig. 2 XRD diffractograms of fresh catalysts.
low calcination temperature (i.e. 200 °C). Nonetheless,
(NH4)4Mo8O26 species were not detected in the reduced
samples due to their decomposition and further reduction.
All reduced samples show solely peaks characteristic of
MoO2 (Fig. 3). According to the TPR profiles (Fig. S1†), the
formation of this molybdenum sub-oxide was expected, as the
samples were pre-reduced at 500 °C; this temperature is
slightly lower than the Tmax ∼ 520 °C of the first H2 consump-
tion peak. Compared to literature studies on the reduction
patterns of unsupported MoO3,34 the MoO3 reduction step
(MoO3→MoO2) appears at lower temperatures in supported
Mo and Fe–Mo/C catalysts. However, a closer look at the angle
2θ = 25.87° (Fig. 3b) reveals a shift of the MoO2 peak upon Fe
addition, which was correspondingly observed for the rest of
the peaks attributed to this phase, evidencing an interaction
between molybdenum and iron oxides. Despite the fact that
only assumptions can be made based on the present character-
ization data, we suggest that the formation of a solid solution
is possible, where Mo4+ ions of the MoO2-based phase are sub-
stituted for Fe ions.35
The increased CO/CO2 ratio, suggesting that more phenolic/
quinone groups are present on the surface compared to carb-
oxylic groups, is related to lower acidity (Table 2). These
groups (e.g. phenolic, quinone) have weakly acidic, neutral or
even basic character that could affect catalyst activity.36 Treat-
ment under air flow seems to selectively remove oxygen groups
(Fig. S4, S5†).
According to TPD-NH3 data (Fig. S2†), high catalyst acidity
can be primarily attributed to Mo phases on the surface. The
noticeably increased width of MoO2 XRD peaks in the Mo/
BC_A catalyst (Fig. 3) indicates high dispersion of MoO2 that
enhances sample acidity, compared to Fe–Mo catalysts. The
modification of molybdenum with iron results in surface
acidity reduction.
Fig. 1 Adsorption isotherms: (a) catalysts supported on BC and (b) cata-
lysts supported on AC.
ported on black carbon, have a gradual curvature that can be
associated with macroporous materials. The absence of hyster-
esis leads to the conclusion that catalysts contain blind cylind-
rical, wedge-shaped or cone-shaped pores. The isotherms of
activated carbon catalysts are typical of microporous materials.
Their hysteresis loop is due to capillary condensation in
the catalyst mesopores and this specific type corresponds to
narrow spit-like pores.31,32
The X-ray diffractograms of all fresh samples after calcina-
tion are presented in Fig. 2. All catalysts show diffraction
peaks that are characteristic of hexagonal MoO3. There are no
peaks attributed to iron oxides, indicating that the oxide is
highly dispersed. The broad diffraction peaks of the Fe/BC_A
catalyst, at 2θ angles of 25 and 44°, are characteristic of black
carbon33 and of the sample holder used for the XRD analysis,
respectively. The XRD profiles of the Fe–Mo/AC catalysts differ
from the above-mentioned patterns, as the presence of SiO2 is
evident. This is not considered unusual, as Si is one of
the main impurities of activated carbons. The presence of
Catalytic results
(NH4)4Mo8O26 in Fe–Mo/BC_A and Fe–Mo/BC_B samples is The catalytic results of all samples evaluated in glycerol hydro-
probably due to the excess of NH4OH solution used for co- deoxygenation reactions, at 300 °C, are presented in Table 3.
precipitation in the catalyst synthesis procedure and the relatively Propylene was the only product in the gas phase, while
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
slightly converted (i.e. 0.7 and 6.5%) and only liquid products
were formed. In order to examine the role of each metal oxide
(i.e. iron and molybdenum), the performance of these samples
has also been evaluated. In experiments with the Fe/BC_A cata-
lyst, glycerol was also slightly converted to some extent (3.3%),
yet only to liquid products. The use of carbon supports and
Fe/BC_A as catalysts for the hydro-deoxygenation of glycerol
results in the production of liquid products, primarily 2-prope-
nol or propylene glycol (Table 3). Nonetheless, no propene was
formed.
The conversion of glycerol over Mo/BC_A was the highest
among all catalysts (68.2%), indicating that molybdena species
(i.e. MoO2 on the reduced catalysts) drive glycerol hydro-
deoxygenation.37 The higher activity of this catalyst can be
attributed to the increased acidity (Table 2), as acidity is con-
sidered a key factor, promoting hydro-deoxygenation reactions
of glycerol.38–40 Of great importance is the fact that over the
Mo/BC_A catalyst, selectivity to gas-phase products (propylene)
is mostly favored, compared to selectivity to liquid products:
ca. 70% and 30%, respectively. Glycerol conversion over Fe–Mo
catalysts is lower compared to the molybdenum catalyst,
ranging from 48.9 to 56.0%; a lower acidity of Fe–Mo catalysts
might be the reason for this. A closer look at the results of
the various Fe–Mo catalysts shows that the type of support
(activated vs. black carbon) and the calcination atmosphere
(N2 vs. air) only marginally affect conversion. Selectivity to gas
phase products (propylene) does not seem to be affected by
the presence of Fe, as for all Fe–Mo catalysts it remains high
and comparable to that of Mo/BC_A. Minor differences
are observed for catalysts supported on activated carbon (i.e.
Fe–Mo/AC_A and Fe–Mo/AC_B), exhibiting slightly lower
selectivity (i.e. 67.9 and 64.8% respectively) compared to the
catalysts supported on black carbon (selectivity to propylene
ca. 71.5%). Different porous characteristics of the catalysts
(Table 2) and, more specifically, the presence of microporosity
in the activated carbon-based catalysts may induce shape
selectivity in glycerol transformation, due to the existence of
longer diffusion pathways that affect reaction (consumption
and formation) rates. Shape selectivity and diffusivity effects
on the reactant and product rates have been previously
reported for carbon materials with microporosity.41,42 The fact
Fig. 3 (a) XRD diffractograms of reduced catalysts and (b) angle 2θ,
20–30° region shift of the MoO2 peak.
2-propenol, 1-propanol and propylene glycol were primarily detected
in the liquid phase. Other liquid products formed were
methanol, 2-propanol, acetol, ethylene glycol, ethanol and
1,3-propanediol.
Carbon supports were also tested as catalysts in the same
reaction; nevertheless, in these experiments, glycerol was
Table 3 Catalytic performance of all samples in glycerol HDO reactions (reaction time: 2.0 h, H2 pressure: 8.0 MPa, temperature: 300 °C, catalyst
weight: 1.5 g)
Selectivity (%)
Total
conversion (%)
Yield (%)
Propene
Sample
Propene
2-Propenol
1-Propanol
Propylene glycol
Othersa
Activated carbon
Black carbon
Fe/BC_A
0.7
6.5
3.3
68.2
48.9
56.0
51.7
53.7
0.0
0.0
0.0
69.9
71.7
67.9
71.5
64.8
3.1
11.1
47.7
4.4
8.0
10.8
6.3
0.0
0.0
0.0
7.1
1.2
2.7
4.1
7.1
53.3
0.0
0.0
5.4
7.9
8.3
7.8
8.4
43.6
88.9
52.3
13.3
11.2
10.3
10.3
12.2
0.0
0.0
0.0
47.6
35.0
38.0
37.0
34.8
Mo/BC_A
Fe–Mo/BC_A
Fe–Mo/AC_A
Fe–Mo/BC_B
Fe–Mo/AC_B
7.5
a Others refer to methanol, 2-propanol, ethanol, ethylene glycol, hydroxyacetone and 1,3PD.
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
that, under the specific experimental conditions, it appears
plausible that propylene readily desorbs from the catalytic
active sites and then dissolves in the gas phase,43 due to its
low solubility in polar solvents, is of major importance. The
latter explains the absence of propane formation via hydrogen-
ation of propylene.
Glycerol deoxygenation is not complete, leading to the for-
mation of partially deoxygenated products, mostly detected in
the liquid phase. 2-Propenol and propylene glycol are the
main products in the liquid phase over molybdena-containing
catalysts, although selectivity to these products was sig-
nificantly lower (Table 3).
As mentioned above, the presence of molybdena sub-oxides
on the catalyst surface is essential for the deoxygenation. The
reduction procedure and the presence of the H2 atmosphere
are key factors for the formation of this active molybdenum
phase, along with oxygen vacancies,44 resulting in the pro-
duction of metallic/acidic bi-functional catalysts.45–47
Effect of reaction parameters
The above-mentioned encouraging results regarding glycerol
conversion and selectivity to propene led to our subsequent
investigation of the effects of various parameters on the
glycerol hydro-deoxygenation reaction. The effects of operating
conditions (i.e. reaction time, reaction temperature, catalyst
loading and H2 pressure) have been examined and the results
are presented below.
Activity and product distribution with time
Fig. 5 Effect of reaction time on selectivity over the Fe–Mo/BC_A cata-
lyst (H2 pressure: 8.0 MPa, temperature: 300 °C).
Experiments to determine the effect of reaction time were per-
formed using the Fe–Mo/BC_A catalyst for 0.5 to 6 h of reac-
tion time. Fig. 4 illustrates the overall glycerol conversion
versus reaction time. In general, the overall glycerol conversion
gradually increases with time, reaching 88.8% after 6 hours. It
should be underlined that at zero time point, glycerol has
already been partially converted (4.2%) during the time
needed for the temperature to reach the set point. Neverthe-
less, this value is considered significantly low.
Fig. 5 illustrates selectivity versus glycerol conversion for the
various liquid and gas products formed. The distribution of
the reaction products is quite interesting and points out the
fact that propene is the main product formed, with 76.1%
selectivity after a six-hour reaction. It must be highlighted that
selectivity to propene increases with reaction time (and with
glycerol conversion), while a corresponding reduction in 2-pro-
penol selectivity is observed, stressing that 2-propenol is the
major intermediate product in propene formation. Experi-
ments, using 2-propenol as feedstock, have been conducted in
order to further investigate this hypothesis. The Fe–Mo/BC_A
catalyst has been employed for this purpose under typical reac-
tion conditions, using 2 wt% aqueous 2-propenol initial solu-
tion. 2-Propenol conversion was significantly high (86.4%),
while propene was selectively formed (89.2%); 1- and 2-propa-
nols were also detected, yet in considerably low concen-
trations. The high reactivity of 2-propenol, along with the
selective formation of propene, stresses that this compound
actually is an intermediate product of glycerol hydro-deoxy-
genation to propene. Glycerol dehydration can lead to the
Fig. 4 Effect of reaction time on glycerol conversion over the Fe–Mo/
BC_A catalyst (H2 pressure: 8.0 MPa, temperature: 300 °C).
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
formation of acrolein,48,49 which can be hydrogenated to
produce 2-propenol. Despite the fact that acrolein was not
detected in the product mixture, it can be inferred that 2-prope-
nol is formed after acrolein hydrogenation. We suggest that
2-propenol is subsequently hydro-deoxygenated to propene
(loss of oxygen while preserving the CvC double bond).
Apart from 2-propenol, at low conversion levels, propylene
glycol and acetol are the main liquid products formed from
glycerol, in 28 and 10% selectivities, respectively, suggesting
that a parallel pathway of hydro-deoxygenation is possible
via acetol formation and subsequent hydrogenation to pro-
pylene glycol and propanols. However, as glycerol conversion
increases, selectivity to these products decreases because pro-
pylene glycol converts to 1- and 2-propanols. Selectivity to the
rest of the liquid products remains almost constant with gly-
cerol conversion, without showing any considerable change.
The effect of the reaction time on product distribution pro-
vides strong evidence regarding propene formation intermedi-
ate and reaction pathways. The proposed mechanism follows
bi-functional dehydration/hydrogenation pathways. It has been
reported that Brønsted acidity directs the reaction towards
acrolein–propene formation while Lewis acidity towards acetol
production.50 Taking into account the experimental results of
this study, as well as relative reports,30,43,51,52 it could be pro-
posed that molybdenum oxide, at a reduced state, and surface
hydroxyls formed due to water presence and reduction pro-
cedure,30,53,54 acting as Brønsted acids, provide the active sites
for the reaction to occur. It is of great importance that selec-
tivity to propene increases with glycerol conversion while the
overall yield to propene reaches the value of 67.6%, at 6 h of
reaction time.
Fig. 6 (a) Effect of temperature on glycerol conversion and gaseous
product selectivity and (b) on liquid product selectivity over the Fe–Mo/
BC_A catalyst (reaction time: 2.0 h, H2 pressure: 8.0 MPa).
Temperature
amount (Fig. 7). The results specify that incremental addition
of a catalyst, providing more active sites, enhances glycerol
conversion.56 However, it is observed that by doubling the cata-
lyst weight, from 1.5 to 3 g, the conversion does not increase
proportionally, indicating possible mass transfer effects at
high catalyst concentrations. Variations of the catalyst weight
also affect the product distribution. An increase of the catalyst
amount, up to 1.5 g, leads to higher selectivity to propene,
mostly because of higher conversion. In contrast, a further
increase of conversion, from 50 to 65%, is not followed
by enhanced propene selectivity, probably due to sequential
propene hydration to propanols.57 Indeed, selectivity to 2- and
1-propanol increases considerably.
In order to determine the effect of temperature on the catalytic
results, experiments were performed at 260, 280 and 300 °C
using the Fe–Mo/BC_A catalyst. Overall glycerol conversion
clearly increases with temperature.
Fig. 6 shows product distribution with temperature, where
selectivity to propene noticeably increases. At low tempera-
tures, glycerol dehydration to acetol and subsequent formation
of propylene glycol and propanols is favored. Additionally, pro-
pylene glycol can be converted to 1- and 2-propanols, explain-
ing the relatively high selectivity of the latter, at 280 °C. It has
been reported that above 200 °C, propylene glycol reacts
further to produce mono-alcohols.55 Selectivity to 2-propenol
presents an increase at higher temperature, while no signifi-
cant change is observed for the rest of the products. As
2-propenol is considered an intermediate product for propene
H2 pressure
production, its increased formation at 300 °C shows that Experiments were performed by differentiating the H2 initial
higher selectivity to propene could potentially be achieved at pressure (i.e. 3.0, 5.0, and 8.0 MPa), over the Fe–Mo/BC_A cata-
the highest temperature used, after 2 h of reaction.
lyst. Increased H2 pressure has a positive effect on glycerol con-
version (Fig. 8). High pressure favors H2 solubility in water,
thus increasing its concentration in the liquid phase.16 At
Catalyst loading
Experiments were performed with different catalyst loadings lower pressure, selectivity to alcohols is increased while after
(i.e. 0.75 g, 1.5 g and 3.0 g) using the Fe–Mo/BC_A catalyst. As 5.0 MPa, selectivity to most of the products of the liquid phase
expected, glycerol conversion increased with the catalyst decreases. Low selectivity to propene, especially at 3 MPa,
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
Fig. 7 (a) Effect of catalyst loading on glycerol conversion and gaseous
product selectivity and (b) on liquid product selectivity over the Fe–Mo/
BC_A catalyst (reaction time: 2.0 h, H2 pressure: 8.0 MPa, temperature:
300 °C).
Fig. 8 (a) Effect of H2 pressure on glycerol conversion and gaseous
product selectivity and (b) on liquid product selectivity over the Fe–Mo/
BC_A catalyst (reaction time: 2.0 h, temperature: 300 °C).
temperature: 300 °C) for 4 h; prior to each test, the catalyst was
reduced following the typical procedure. According to the
experimental data, glycerol conversion is decreased by 10%
after the fourth recycle experiment. It must be highlighted that
deactivation affects only catalyst activity and not selectivity to
propene, as the latter follows the glycerol conversion decrease
(Fig. 5a). The above results highlight the structural integrity of
the Fe–Mo/BC_A catalyst.
cannot be ascribed solely to low conversion levels but primar-
ily to the low hydrogenation extent due to limited H2 avail-
ability. The critical role of H2 availability for propene
selectivity explains the relatively lower values obtained in the
present study, compared to those mentioned by Fadigas
et al.28
In this study, the H2/glycerol molar ratio is ∼53 : 1(∼88%
conversion and ∼76% selectivity to propene), while the
optimum ratio used by Fadigas et al.28 is 120 : 1 (100% conver-
sion and ∼90% selectivity to propene; the gas phase products
observed were propane, ethane and methane, apart from
propene). Moreover, the optimum H2/glycerol molar ratio used
by Cao et al. is 100 : 1 (100% conversion and 85% selectivity to
propene), while the main by-product formed was propane.27
Nevertheless, according to the present catalytic results, under
the specific reaction conditions, propene is the only gas-phase
product formed (100% selectivity in the gas phase).
4. Conclusions
Propene can be effectively produced via a one-step HDO reac-
tion of glycerol derived from biomass, in the liquid phase, over
Fe–Mo/C catalysts. After 6 h of reaction time, glycerol can
be converted up to 88.8% with 76.1% overall selectivity
to propene (100% in the gas phase). Fe–Mo catalysts were syn-
thesized, characterized and their activity has been evaluated.
Characterization results indicate that molybdenum sub-oxide
(i.e. MoO2) is the active phase of these reactions. Under reac-
Catalyst re-use experiments
The Fe–Mo/BC_A catalyst has been re-used under the standard tion conditions of this research (temperature and pressure:
reaction conditions in order for its stability to be evaluated. 300 °C and 8.0 MPa, respectively), Fe–Mo catalysts are active to
Four consecutive experiments have been conducted under the selective conversion of glycerol to propylene. Increased
the standard reaction conditions (i.e. H2 pressure: 8.0 MPa, temperature and H2 pressure (i.e. H2/glycerol ratio) favor
Green Chem.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Green Chemistry
Paper
glycerol conversion towards propene, but suppress formation 15 E. S. Vasiliadou, T. M. Eggenhuisen, P. Munnik, P. E. de
of acetol, propanols and propane-diols. Higher catalyst loading
also directs the reaction to the formation of the desired
Jongh, K. P. de Jong and A. A. Lemonidou, Appl. Catal., B,
2014, 145, 108–119.
product, in the gas phase. The evolution of reaction products 16 E. S. Vasiliadou and A. A. Lemonidou, Chem. Eng. J., 2013,
as a function of reaction time suggests that propene is mainly 231, 103–112.
formed via the acrolein–2-propenol route while the contri- 17 A. K. Kinage, P. P. Upare, P. Kasinathan, Y. K. Hwang and
bution of the acetol-propylene glycol-2-propanol route is rather J. S. Chang, Catal. Commun., 2010, 11, 620–623.
limited. Stability tests reveal that selectivity to propene is not 18 S. Chai, H. Wang, Y. Liang and B. Xu, J. Catal., 2007, 250,
affected by the moderate activity loss (10%) observed after the 342–349.
fourth re-use test, as it clearly follows the reduction of glycerol 19 E. Tsukuda, S. Sato, R. Takahashi and T. Sodesawa, Catal.
conversion.
Commun., 2007, 8, 1349–1353.
20 H. Serafim, I. M. Fonseca, A. M. Ramos, J. Vital and
J. E. Castanheiro, Chem. Eng. J., 2011, 178, 291–296.
21 Y. Amada, S. Koso, Y. Nakagawa and K. Tomishige,
ChemSusChem, 2010, 3, 728–736.
22 K. Murata, I. Takahara and M. Inaba, React. Kinet. Catal.
Lett., 2008, 93, 59–66.
23 J. van Haveren, E. L. Scott and J. Sanders, Biofuels, Bioprod.
Biorefin., 2008, 2, 41–57.
24 B. A. Rodriguez, C. C. Stowers, V. Pham and B. M. Cox,
Green Chem., 2014, 16, 1066–1076.
25 G. M. Wells, Handbook of Petrochemicals and Processes,
Gower Publishing Company Limited, Great Britain, 1991.
26 J. C. Souza Fadigas, R. Gambetta, C. J. de Araujo Mota and
V. L. da Conceicao Goncalves, US Pat, 0 224 470, 2011.
27 L. Yu, J. Yuan, Q. Zhang, Y.-M. Liu, H.-Y. He, K.-N. Fan and
Y. Cao, ChemSusChem, 2014, 7, 743–747.
Acknowledgements
This research has been co-financed by the European Union
(European Social Fund – ESF) and Greek national funds
through the Operational Program “Education and Lifelong
Learning” of the National Strategic Reference Framework
(NSRF) – Research Funding Program: THALES. Investing in
knowledge society through the European Social Fund.
References
1 A. M. Ruppert, K. Weinberg and R. Palkovits, Angew. Chem.,
Int. Ed., 2012, 51, 2564–2601.
28 C. Hultenberg and J. Brandin, WO Pat, 155 674, 2009.
2 P. Basu, Biomass Gasification and Pyrolysis, Elsevier, USA, 29 O. Bayraktar, Ph.D. Thesis, West Virginia University, 2001.
2010, pp. 325–326.
3 G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006,
106, 4044–4098.
30 T. Matsuda, Y. Hirata, H. Sakagami and N. Takahashi,
Chem. Lett., 1997, 26(12), 1261–1262.
31 G. Leofanti, M. Padovan, G. Tozzola and B. Venturelli,
Catal. Today, 1998, 41, 207–219.
4 P. N. R. Vennestrøm, C. M. Osmundsen, C. H. Christensen
and E. Taarning, Angew. Chem., Int. Ed., 2011, 50, 10502– 32 K. S. W. Sing, D. H. Everett, R. H. W. Haul, L. Moscou,
10509.
R. A. Pierotti, J. Rouquerol and T. Siemieniewska, Pure
Appl. Chem., 1985, 57, 603–619.
5 T. Werpy and G. Petersen, Top Value Added Chemicals from
Biomass, N.R.E.L., U.S. Department of Energy, USA, 2004, 33 J. Wang, G. Yin, Y. Shao, S. Zhang, Z. Wang and Y. Gao,
available online at http:/www.osti.gov/bridge. J. Power Sources, 2007, 171, 331–339.
6 L. R. Morrison, in Kirk-Othmer Encyclopedia of Chemical 34 K. V. R. Chary, K. R. Reddy, G. Kishan,
Technology, John Wiley and Son Inc., 2000, , DOI: 10.1002/
0471238961.0712250313151818.a01.
J. W. Niemantsverdriet and G. Mestl, J. Catal., 2004, 226,
283–291.
7 Z. Wang, J. Zhuge, H. Fang and B. A. Prior, Biotechnol. Adv., 35 Y. Huang, L. Cong, J. Yu, P. Eloy and P. Ruiz, J. Mol. Catal.
2001, 19, 201–223. A: Chem., 2009, 302, 48–53.
8 F. N. Arroyo-López, R. Pérez-Torrado, A. Querol and 36 J. Vissers, S. Bouwens, V. de Beer and R. Prins, Carbon,
E. Barrio, Food Microbiol., 2010, 27, 628–637. 1987, 25, 485–493.
9 T. Modig, K. Granath, L. Adler and G. Lidén, Appl. Micro- 37 H. Ren, W. Yu, M. Salciccioli, Y. Chen, Y. Huang, K. Xiong,
biol. Biotechnol., 2007, 75, 289–296.
D. G. Vlachos and J. G. Chen, ChemSusChem, 2013, 6, 798–
801.
38 C. H. Zhou, J. N. Beltramini, Y. X. Fan and G. Q. Lu, Chem.
Soc. Rev., 2008, 37, 527–549.
39 Y. Zheng, X. Chen and Y. Shen, Chem. Rev., 2008, 108,
5253–5277.
40 B. Katryniok, S. Paul, M. Capron and F. Dumeignil,
ChemSusChem, 2009, 2, 719–730.
10 H. Zhao, C. H. Zhou, L. M. Wu, J. Y. Lou, N. Li, H. M. Yang,
D. S. Tong and W. H. Yu, Appl. Clay Sci., 2013, 74, 154–162.
11 F. Bauer and C. Hulteberg, Biofuels, Bioprod. Biorefin., 2013,
7, 43–51.
12 E. S. Vasiliadou and A. A. Lemonidou, Appl. Catal., A, 2011,
396, 177–185.
13 D. Lahr and B. Shanks, J. Catal., 2005, 232, 386–394.
14 E. S. Vasiliadou, E. Heracleous, I. A. Vasalos and 41 S. M. Csicsery, Zeolites, 1984, 4, 202–213.
A. A. Lemonidou, Appl. Catal., B, 2009, 92, 90–99.
42 P. B. Weisz, Pure Appl. Chem., 1980, 52, 2091–2103.
This journal is © The Royal Society of Chemistry 2014
Green Chem.

View Article Online
Paper
Green Chemistry
43 D. R. Moberg, T. J. Thibodeau, F. G. Amar and 50 A. Alhanash, E. F. Kozhevnikova and I. V. Kozhevnikov,
B. G. Frederick, J. Phys. Chem. C, 2010, 114, 13782–13795. Appl. Catal., A, 2010, 378, 11–18.
44 T. Prasomsri, T. Nimmanwudipong and Y. Román-Leshkov, 51 H. Sakagami, T. Ohno, H. Itoh, Z. Li, N. Takahashi and
Energy Environ. Sci., 2013, 6, 1732. T. Matsuda, Appl. Catal., A, 2014, 470, 8–14.
45 F. C. Meunier, F. Cavallaro, T. Le Goaziou, A. Goguet and 52 C. Hoang-Van and O. Zegaoui, Appl. Catal., A, 1995, 130,
C. Rioche, Catal. Today, 2006, 112, 64–67. 89–103.
46 A. Katrib, P. Leflaive, L. Hilaire and G. Maire, Catal. Lett., 53 H. Al-Kandari, A. M. Mohamed, F. Al-Kharafi, M. I. Zaki
1996, 38, 95–99. and A. Katrib, Appl. Catal., A, 2012, 417–418, 298–305.
47 E. Torres-García, G. Rodríguez-Gattorno, J. A. Ascencio, 54 H. Belatel, H. Al-Kandari, F. Al-Khorafi, A. Katrib and
L. O. Aleman-Vazquez, J. L. Cano-Domínguez, A. Martínez- F. Garin, Appl. Catal., A, 2004, 275, 141–147.
Hernandez and P. Santiago-Jacinto, J. Phys. Chem. B, 2005, 55 M. A. Dasari, P. P. Kiatsimkul, W. R. Sutterlin and
109, 17518–17525. G. J. Suppes, Appl. Catal., A, 2005, 281, 225–231.
48 A. Corma, G. Huber, L. Sauvanaud and P. O’Connor, 56 N. Dimitratos, A. Villa and L. Prati, Catal. Lett., 2009, 133,
J. Catal., 2008, 257, 163–171. 334–340.
49 L. Shen, H. Yin, A. Wang, X. Lu, C. Zhang, F. Chen, 57 M. E. Thibault, D. V. DiMondo, M. Jennings,
Y. Wang and H. Chen, J. Ind. Eng. Chem., 2014, 20, 759–
766.
P. V. Abdelnur, M. N. Eberlin and M. Schlaf, Green Chem.,
2011, 13, 357–366.
Green Chem.
This journal is © The Royal Society of Chemistry 2014