Hydrogenolysis vs. aqueous phase reforming (APR) of glycerol promoted by a heterogeneous Pd/Fe catalyst

Article abstract of DOI:10.1039/c5cy00656b

The hydrogenolysis and the aqueous phase reforming of glycerol have been investigated under mild reaction conditions, using water as the reaction medium and Pd/Fe as the catalyst. The experiments, in the presence of added H₂ or under inert atmosphere, clearly show that the dehydration/hydrogenation route is the key step in the case of C-O bond cleavage (hydrogenolysis) while dehydrogenation is a prerequisite for C-C bond breaking (APR), with the latter favoured at higher reaction temperatures. The temperature dependence of the C-C and C-O bond rupture is discussed by taking into account the bond energies involved in the competitive hydrogenolysis and APR reactions. Finally, the Pd/Fe catalyst was also tested in the hydrogenolysis and APR of ethylene glycol in the same temperature range, with the aim of clarifying the selective cleavage of C-O and C-C bonds in biomass derived C₂-C₃ polyols.

Full text of DOI:10.1039/c5cy00656b

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
Hydrogenolysis vs. aqueous phase reforming (APR)
of glycerol promoted by a heterogeneous Pd/Fe
Cite this: DOI: 10.1039/c5cy00656b
catalyst†
a
Francesco Mauriello, Alessandro Vinci,a Claudia Espro,b Bianca Gumina,b
*
Maria Grazia Musolinoa and Rosario Pietropaoloa
The hydrogenolysis and the aqueous phase reforming of glycerol have been investigated under mild
reaction conditions, using water as the reaction medium and Pd/Fe as the catalyst. The experiments, in the
presence of added H2 or under inert atmosphere, clearly show that the dehydration/hydrogenation route is
the key step in the case of C–O bond cleavage (hydrogenolysis) while dehydrogenation is a prerequisite for
C–C bond breaking (APR), with the latter favoured at higher reaction temperatures. The temperature
dependence of the C–C and C–O bond rupture is discussed by taking into account the bond energies
involved in the competitive hydrogenolysis and APR reactions. Finally, the Pd/Fe catalyst was also tested in
the hydrogenolysis and APR of ethylene glycol in the same temperature range, with the aim of clarifying
the selective cleavage of C–O and C–C bonds in biomass derived C2–C3 polyols.
Received 5th May 2015,
Accepted 5th July 2015
DOI: 10.1039/c5cy00656b
www.rsc.org/catalysis
oxide hydration. Therefore, glycerol hydrogenolysis to 1,2-
propanediol represents a sustainable alternative process that
Introduction
Growing attention has been gained on the use of renewable
biomass for bulk chemical production.1,2
would assist both the environmental benefits and economic
viability of biodiesel manufacture.
Glycerol, the main byproduct from the production of bio-
diesel, is one of the top 12 building block chemicals largely
used as bio-feedstocks for the production of chemical prod-
ucts.3,4 Due to the worldwide expansion of biofuels, at pres-
ent the production of glycerol has reached over 2 million
tonnes that consistently enter the market every year.5 This
unique situation has made glycerol attractive as a renewable
starting material for the synthesis of valuable chemicals.
The conversion of biomass into chemicals and fuels com-
monly implies oxygen removal. In this context, the selective
hydrogenolysis of glycerol has gained more and more atten-
tion, leading to the formation of propylene diols, ethylene
glycol and propanols.6,7 Considerable research effort has
been directed towards the conversion of glycerol into 1,2-
propanediol (1,2-PDO), which is a high demand chemical,
largely used in the production of unsaturated polyester
resins, functional fluids, food products, cosmetics and phar-
maceuticals.8 Currently, the industrial production of 1,2-
propanediol mainly occurs through the route of propylene
At the same time, the production of hydrogen through the
aqueous-phase reforming (APR) of biomass-derived polyols is
considered as a promising catalytic process.9–14 Glycerol is an
interesting starting substrate for “bio”-hydrogen production
since the related process becomes spontaneous at much
lower temperature than that generally used today for H2 man-
ufacture from CH4. An ideal catalyst for H2 production from
glycerol needs to be very active both for C–C bond breaking
as well as in promoting the water gas shift (WGS) reaction
thus reducing the CO content.
Furthermore, H2 – generated in situ by the aqueous phase
reforming of glycerol – has been successfully used for the
conversion of glycerol to 1,2-PDO, representing an interesting
alternative to the direct use of molecular hydrogen.15–18
Hydrogenolysis and APR reactions involve the breaking of
C–C, C–H, C–O and O–H bonds, therefore glycerol can also
be considered as a model compound for fundamental studies
involving the selective cleavage of chemical bonds of biomass
derived substrates.
Heterogeneous palladium catalysts were found to be
poorly efficient for the hydrogenolysis of biomass derived
polyols.6,7 However, in past years, the coprecipitated Pd/Fe
catalyst has been deeply investigated19–26 for its superior per-
formance in several catalytic reactions including the hydro-
genolysis20,21 and the transfer hydrogenolysis (CTH) of glyc-
erol using 2-propanol as the reaction solvent,22,23 as well as
a Dipartimento DICEAM, Università Mediterranea di Reggio Calabria, Loc. Feo di
Vito, I-89122 Reggio Calabria, Italy. E-mail: francesco.mauriello@unirc.it;
Tel: (+39) 09651692278
b Dipartimento DIECII, Università di Messina, Contrada Di Dio, 1 - Vill. S. Agata,
I-98166 Messina, Italy
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
c5cy00656b
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
in the deoxygenation and aqueous-phase reforming of ethyl-
ene glycol (EG).24,25 The enhanced catalytic performance
shown by the Pd/Fe catalyst can be related to the preparation
method used (coprecipitation technique) that ensures a
strong interaction between palladium and the iron oxide sup-
port, leading to the formation of bimetallic Pd–Fe ensembles,
which positively promote C–O and C–C bond cleavage.19–26
In the present work, we extend the studies to the hydro-
genolysis and APR of glycerol promoted by the Pd/Fe catalyst
using water as the reaction solvent with two different strate-
gic approaches: by using molecular hydrogen as the readily
available H-source and by using the APR condition producing
in situ H2.
By using molecular H2 as the direct hydrogen source, at
180 °C, 1,2-PDO and EG are the main reaction products from
glycerol hydrogenolysis, the Pd/Fe catalyst being more active
in C–O bond cleavage. Upon increasing the reaction tempera-
ture, a major tendency of the Pd/Fe catalyst to promote dehy-
drogenation/decarbonylation reactions is found. Under inert
atmosphere, at 180 °C, the H2 produced by APR is partly con-
sumed in the in situ hydrogenolysis process leading to
1,2-PDO. At higher reaction temperatures, glycerol is fully
converted with major production of gas phase products and
high H2 selectivity.
(TEM) measurements using a JEM-2100F (JEOL, Japan) micro-
scope operating at an acceleration voltage of 200 kV and
directly interfaced with a computer controlled-CCD for real-
time image processing. The specimens were prepared by
grinding the reduced catalyst powder in an agate mortar and
then suspending it in 2-propanol. A drop of the suspension,
previously dispersed in an ultrasonic bath for 1 hour, was
deposited on a copper grid coated by a holey carbon film.
After evaporation of the solvent, the specimens were intro-
duced into the microscope column. The particle size distribu-
tion was obtained by counting a hundred particles visible on
the micrographs. From the size distribution, the average
diameter was calculated by using the expression: dn =
P
nidi/ni where ni is the number of particles of diameter di.
The H2-TPR measurement was performed using a con-
ventional TPR apparatus. The dried sample (50 mg) was
heated at a linear rate of 10 °C min−1 from 0 to 1000 °C in
a 5 vol% H2/Ar mixture at a flow rate of 20 cm3 min−1. The
H2 consumption was monitored with a thermal conductivity
detector (TCD). A molecular sieve cold trap (maintained at
−80 °C) and a tube filled with KOH, placed before the TCD,
were used to block water and CO2, respectively. Signal cali-
bration was done by injecting a known amount of H2 into
the carrier.
Therefore, the dependence of the selectivity in glycerol
C–C or C–O bond breaking on the reaction temperature is
observed.
Finally, the conversion of ethylene glycol under the same
reaction conditions has been also investigated to clarify the
competitive occurrence of the dehydrogenation and dehydra-
tion processes in C2–C3 biomass derived polyols.
XPS measurements were performed on a JPS-9010MC
photoelectron spectrometer using an Al Kα (1486.6 eV) radia-
tion source. After the reduction treatment, the sample was
introduced into the XPS chamber, avoiding exposure to air.
In order to obtain the XPS spectra, the pressure in the analy-
sis chamber was maintained at 5 × 10−9 mbar. The binding
energies (BE) were set taking the C 1s peak at 284.6 eV as the
reference. Peak deconvolution and fitting analyses were
performed using the peak-fitting software “SPECSURF, JEOL”
including the spin–orbit splitting and relative intensities of
the spin–orbit components fixed.
2 Experimental
2.1 Catalyst preparation
All chemicals were purchased from Sigma-Aldrich and used
without further purification. The Pd/Fe catalyst with a nomi-
nal palladium loading of 5 wt% was obtained by using the
coprecipitation technique. Anhydrous palladium chloride was
dissolved in HCl and ironIJIII) nitrate nonahydrate was added.
The obtained aqueous metal salt solution was added
dropwise into a 1 M aqueous solution of Na2CO3. After filtra-
tion, the sample was washed until complete removal of chlo-
ride ions and dried for 1 day at 80 °C under vacuum. Before
the reaction, the Pd/Fe catalyst was reduced at 200 °C for 2 h
under a hydrogen flow.
2.3 Catalytic tests
The reactions were carried out in a stainless steel autoclave
reactor (100 mL). In each reaction, 40 ml of an aqueous glyc-
erol solution (4 wt% glycerol solution) and 0.25 g of the Pd/
Fe catalyst (previously reduced at 200 °C for 2 h under a flow
of hydrogen) were loaded in the autoclave. The reactor was
purged with He (99.99%), the system was then pressurized
with the desired gas (H2 or He) and, finally, heated up to the
reaction temperature at a linear rate of 2.5 °C min−1 until the
constant value was reached.
The reactant and products in the liquid phase were ana-
lyzed using an off-line HPLC (LC Agilent 1290 Infinity)
equipped with a ZORBAX Carbohydrate Analysis Column.
After the reaction, the gas phase products were collected
in a gas sample cylinder and analyzed with a GC-TCD-FID
(Agilent Technologies, 7890 A) equipped with a molecular
sieve column (HP-MOLESIEVE, diameter = 0.535 mm, length
= 30 m) and a capillary column (HP-PLOT/Q, diameter =
0.320 mm, length = 30 m).
2.2 Catalyst characterization
XRD data were acquired at room temperature on a Philips
X-Pert diffractometer by using the Ni β-filtered CuKα radia-
tion (λ = 0.15418 nm). Analyses were registered in the 2θ
range of 10–80° at a scan speed of 0.5° min−1. Diffraction
peaks were compared with those of standard compounds
reported in the JPCDS data file.
The particle size and the relative morphology were
analysed by performing transmission electron microscopy
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
The conversion, product selectivity in the liquid and gas
phases and 1,2-PDO yield were calculated on the basis of the
following equations:
→ Fe3O4 reductions suggesting that Pd nanoparticles are
effective in promoting the reduction of the support from
hematite to magnetite. Furthermore, the conclusion that the
palladium particles promote the reduction of Fe3+ to Fe3O4 is
also confirmed by the XRD analysis on the Pd/Fe catalyst
reduced at 100 °C (Fig. S1.3†) which shows signals only
related to the magnetite structure.
(1)
The presence of partially positively charged metal species
is observed by X-ray photoelectron spectroscopy (Table S1.5
and Fig. S1.5†) (the Pd 3d5/2 binding energy is 0.4 eV higher
than that of metallic palladium) indicating that, in the
reduced Pd/Fe catalyst, intimate interactions of palladium
nanoparticles with the support are present.23,24
Furthermore, the analysis of the extended X-ray absorption
fine structure, carried out on the same catalyst,24 shows a
Pd–Fe distance of 2.59 Å and a shorter, distorted Pd–Pd dis-
tance of 2.66 Å, confirming the formation of Pd–Fe bimetallic
ensembles.
(2)
(3)
(4)
3.2 Hydrogen promoted reactions
Table 1 shows that 1,2-PDO is the main reaction product in
the reaction carried out at 180 °C using glycerol as the
starting substrate, clearly indicating that the Pd/Fe catalyst is
more active for C–O bond cleavage and less efficient for C–C
bond breaking.
3 Results and discussion
3.1 Catalyst preparation and characterization
The Pd/Fe catalyst was prepared by co-precipitation of palla-
dium chloride (dissolved in HCl) and ironIJIII) nitrate
nonahydrate poured together dropwise into a 1 M aqueous
solution of Na2CO3 followed by calcination and reduction
with H2 at 200 °C.
The main characteristics of the Pd/Fe catalyst are reported
in Table S1.1.†
The high-resolution transmission electron microscopy
reveals a particle size distribution with a majority of particles
of 1.2 nm diameter (Fig. 1 and S1.2†).
Accordingly, Pd nanoparticles are well dispersed on the
catalyst surface as revealed also by the absence of the (111)
diffraction line of metallic palladium in the XRD spectrum
(Fig. S1.3†).
The H2-TPR measurement (Fig. S1.4†) shows only one
intense peak at about 80 °C including, on the basis of H2
consumption calculations, both the PdIJII) → Pd(0) and FeIJIII)
The increase in the reaction temperature produces higher
glycerol conversion accompanied by
a lower selectivity
towards 1,2-PDO together with an increased amount of etha-
nol (EtOH), methanol (MeOH), 1-propanol and 2-propanol
(POs). It is worth noting that the amount of EtOH reaches a
maximum of 70.5% in the liquid phase products at 240 °C.
In the reference experiments, 1,2-PDO was also tested in
the presence of H2 (Table S2.1†): the reactions provide EtOH
as the main product in high yield at all temperatures investi-
gated. These results, coupled with the catalytic tests on EG
(see below) suggest that EtOH, found at higher reaction tem-
peratures, arises mainly from the dehydrogenation/
decarbonylation of 1,2-PDO.
Having identified 180 °C as the best reaction temperature
for the production of 1,2-PDO, we investigated the hydro-
genolysis of glycerol at different reaction times (Fig. 2 and
Table S2.2†). As expected, the glycerol conversion and
1,2-PDO yield increase when passing from 6 to 24 hours
(Fig. 2).
Upon increasing the initial hydrogen pressure from 5 to
20 bar (180 °C), a decrease in glycerol conversion as well as a
lower 1,2-PDO yield is noticed (Fig. 2 and Table S2.3†). Such
a conversion decrease by increasing the hydrogen pressure,
in other hydrogenolysis reactions over heterogeneous palla-
dium systems, was ascribed to the competitive adsorption of
the substrate and H2 on the catalyst surface.27
Further recycling of the Pd/Fe catalyst at 180 °C in the
presence of H2, performed after washing the sample
(remaining in the reactor) with pure water, gave, as expected,
a lower conversion (Fig. S2.4†). However, the selectivity did
Fig. 1 High resolution transmission electron micrograph of the Pd/Fe
catalyst (left) and particle size distribution of the palladium
nanoparticles (right).
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Table 1 Conversion of glycerol and ethylene glycol promoted by the Pd/Fe catalyst in the presence of added hydrogen. Conditions: 0.25 g catalyst; 40
ml entry solution (4 wt%); 5 bar H2 pressure, 24 ha
Liquid
phase
(%)
Product selectivity in liquid phase (%)
Temperature
(°C)
Conversion
(%)
Entry
1,2-PDO
POs
EG
MeOH
EtOH
Glycerol
Glycerol
Glycerol
EG
EG
EG
180
210
240
180
210
240
92.4
100
100
11.0
31.7
72.1
78.6
67.6
57.2
91.6
72.7
51.7
77.1
57.7
9.3
—
—
—
0.2
1.7
10.1
—
—
—
15
9.7
0.5
—
—
—
—
3.8
9.6
49.9
56.5
71.8
7.7
27.2
70.5
50.1
43.4
28.2
a
1,2-PDO = 1,2-propanediol; POs = 1-propanol + 2-propanol; EG = ethylene glycol; MeOH = methanol; EtOH = ethanol.
not change appreciably, as already reported in the past, upon
using 2-propanol as solvent.20
equally cleaves the C–C and the C–O bonds of EG, allowing
the production of methanol and ethanol in comparable
yields. Upon raising the reaction temperature, an enhanced
performance of Pd/Fe for C–C bond breaking is noticed.
The presence of added H2 in the catalytic test conducted
by using both glycerol and EG as starting substrates does not
allow a truthful quantitative analysis of gas phase products.
However, at the temperature range investigated (180–240 °C)
a qualitative analysis reveals, together with CO2, a very low
amount of alkanes.
The Pd/Fe catalyst was also tested at higher glycerol con-
centrations (20 wt% of glycerol solution, 1 g of reduced Pd/Fe
catalyst, 20 bar initial pressure of H2, 24 hours of reaction
time): the selectivity towards 1,2-PDO (83,9%), EG (11,2%),
EtOH (4,2%) and POs (0,7%) was detected with a glycerol
conversion of 83,9%.
The performance of the Pd/Fe catalyst was also tested in
the hydrogenolysis of EG in the same temperature range
(Table 1). At all temperatures, the reactions proceed more
slowly with respect to glycerol. At 180 °C, the Pd/Fe catalyst
3.3 Reactions under inert atmosphere
The results for the reactions carried out under inert atmo-
sphere (APR conditions) are summarized in Table 2.
In the case of glycerol, at 180 °C, the liquid products
include mainly 1,2-PDO, EG, and MeOH. In the gas phase, H2
is the main product observed (73.4%) together with CO2
(26.1%) and a negligible amount (0.5%) of light alkanes
(methane, ethane and propane). The hydrogen necessary for
the hydrogenolysis reaction derives from the aqueous phase
reforming of glycerol itself as well as from the WGS process.
Passing from 180 to 210 and 240 °C, glycerol is fully
converted with a higher production of gas phase products. At
240 °C a H2 selectivity of 80.5% in gas phase products is
reached.
Significant changes in the liquid phase product composi-
tion are also detected. As expected, the amount of 1,2-PDO
decreases while a higher selectivity to EtOH and propanols is
observed. Similarly, the ethylene glycol selectivity drops from
12.2% at 180 °C to 6.3% at 210 °C and to 0.2% at 240 °C as a
result of the further dehydrogenation and hydrogenolysis
processes.
It is noteworthy that CO is never detected, which is indica-
tive of a worthy activity of the bimetallic Pd/Fe catalyst in the
water gas shift (WGS) reaction.
The pressure profiles of glycerol conversion promoted by
the Pd/Fe catalyst under inert atmosphere at 180, 210 and
240 °C are shown in Fig. S2.5.† It has to be noted that the cat-
alytic tests under inert atmosphere reported in this contribu-
tion are at variance with standard APR conditions where
higher pressures are used in order to maintain water in the
Fig. 2 Reaction time (top) and initial H2 pressure (bottom) effect on
the glycerol conversion (%) and 1,2-PDO yield (%) in the reaction car-
ried out in presence of added H2 at 180 °C.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
Table 2 Conversion of glycerol and ethylene glycol promoted by the Pd/Fe catalyst under inert atmosphere. Conditions: 0.25 g catalyst; 40 ml entry
solution (4 wt%); 5 bar He pressure, 24 ha
Liquid
phase
(%)
Liquid phase product selectivity (%)
Gas phase selectivity (%)
Temperature
(°C)
Conversion
(%)
Entry
1,2-PDO
POs
EG
MeOH
EtOH
H2
CO2
Alkanes
Glycerol
Glycerol
Glycerol
EG
EG
EG
180
210
240
180
210
240
9.8
76.8
64.7
41.9
89.8
73.5
50.9
72.8
42.9
9.4
—
—
—
3.5
5.2
8.7
—
—
—
12.2
6.3
0.2
—
—
—
10.5
9.0
8.1
60.3
73.3
76.6
1.0
36.6
73.0
39.7
26.3
23.4
73.4
77.8
80.5
81.7
81.6
81.2
26.1
20.8
17.7
18.3
18.4
18.8
0.5
1.4
2.4
—
—
—
100
100
13.3
34.5
66.2
a
1,2-PDO = 1,2-propanediol; POs = 1-propanol + 2-propanol; EG = ethylene glycol; MeOH = methanol; EtOH = ethanol.
liquid state. However, in the temperature range investigated,
the quantity of water in the vapor state should be very low
(below 3% at 240 °C) allowing us to investigate C–C bond
breaking under reaction conditions very close to those usu-
ally adopted for APR reactions.
A combination of the aqueous phase reforming and in situ
hydrogenolysis reactions of ethylene glycol is also observed.
In line with the reactions carried out in the presence of
added H2, a lower conversion of EG with respect to glycerol is
achieved within the temperature range of 180–240 °C. The
liquid phase analysis shows MeOH (C–C bond breaking) as
the main reaction product and a selectivity to EtOH (C–O
bond cleavage) getting lower as the temperature increases.
Gas phase products are characterized by a good H2 produc-
tion and an insignificant amount of alkanes.
promoted by palladium. The unstable enol is then rearranged
to AC which is finally hydrogenated into 1,2-PDO.23
On the other hand, glycerol conversion to EG proceeds via
a previous dissociation of primary alcoholic groups followed
by C–H bond breaking. The ensuing glyceraldehyde undergoes
a final C–CO breaking leading to EG and CO formation.
Other possible alternative processes, such as the one
affording dihydroxyacetone (derived from the dehydrogena-
tion of the secondary alcoholic group), are discarded since
this quite stable compound has never been detected.
Data reported in Table 2 concerning the reactions carried
out under inert conditions show that, together with H2, CO2
and a small amount of alkanes, liquid phase products are
formed. The presence of hydrogenolysis products implies
that the in situ produced H2 directly participates in the hydro-
genation of the intermediates produced from the dehydration
route.
Therefore, it can be stated that hydrogenolysis and APR
are competitive processes (hydrogenolysis vs. APR) and that
the dehydration and dehydrogenation reactions are key ele-
mentary steps in the conversion of glycerol both in the pres-
ence and absence of added H2.
At 180 °C, in the presence of H2 or under inert atmo-
sphere, the Pd/Fe catalyst allows the conversion of glycerol to
1,2-propanediol in a good yield showing an enhanced ability
in C–O bond cleavage. Conversely, at higher reaction temper-
atures, dehydrogenation reactions prevail and the dehydra-
tion/hydrogenation steps become less important, even when
3.4 Hydrogenolysis vs. APR of glycerol: understanding the
relationship between C–O and C–C bond breaking
The results of glycerol conversion at 180 °C in the presence
of H2 over the Pd/Fe catalyst are substantially in agreement
with those obtained, before, on using 2-propanol as the sol-
vent which indicates hydroxyacetone (AC) as a reaction inter-
mediate for 1,2-PDO formation.20,23 The main difference lies
in the higher fraction of EG and EtOH detected in water and
not observed in isopropanol where a good selectivity for 1-PO
was found (Fig. S2.6†). Therefore, it can be stated that in glyc-
erol hydrogenolysis, it is undesirable to use water as a reac-
tion medium in order to drive the equilibrium to propanol
formation.
Cross-check experiments, under both reaction conditions
used in this study (H2 or He) and within the temperature
range investigated (180–240 °C), show that pure Fe3O4 does
not give glycerol conversion, suggesting that the presence of
palladium is an essential prerequisite for C–O and C–C bond
breaking.
Consistent with the catalytic results, the conversion of
glycerol on the bimetallic Pd/Fe catalyst can follow two dis-
tinctive routes for the 1,2-PDO and EG formation (Scheme 1).
In the reaction pathway leading to 1,2-PDO, the first step
is the adsorption of two adjacent alcoholic groups of glycerol
on Pd and Fe sites with a subsequent C–OH breaking
Scheme 1 First steps in glycerol conversion in the presence and
absence of added H2: the dehydration/hydrogenation and
dehydrogenation/decarbonylation routes.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
1,2-PDO is used as the starting substrate (Table S2.1†). To
this regard, the higher ability of the coprecipitated Pd on the
Fe3O4 catalyst (prepared by an analogous synthetic proce-
dure) in the selective C–C bond breaking of biomass derived
polyols at a high reaction temperature (250 °C) has also been
recently observed by the Tsang research group.21 Accordingly,
a recent report clearly shows that reaction temperatures lower
than 180 °C are more favourable to the production of
propanols from 1,2-PDO.28
The temperature dependence of the reaction pathway’s
selectivity can be related to the different bond energies
involved in the alternative processes.29 The dehydration/
hydrogenation route implies the initial C–OH cleavage of the
primary alcoholic group (EC−OH ≅ 80 kcal mol−1)29 followed
by the rearrangement of the unstable enol to hydroxyacetone
which is finally hydrogenated to 1,2-PDO.23 Conversely, the
dehydrogenation/decarbonylation route involves the prior
O–H breaking (EO–H ≅ 104 kcal mol−1)28 followed by the β
C–H (with respect to oxygen) elimination and by the conse-
quent C–CO rupture.30,31
Moreover, since a negligible quantity of methanol was
detected in the reactions carried out by using 1,2-PDO as the
starting substrate (Table S2.1†) it can be confirmed that the
rupture of the O–H bond is a fundamental prerequisite for
C–C cleavage.
Similar conclusions have been also proposed by Lercher
and co-workers who suggested a reaction network consisting
of parallel routes that include dehydrogenation and dehydra-
tion as the primary steps in the hydrodeoxygenation and APR
of C3 alcohols and polyols.32,33
of an internal C–C bond (EC–C ≅ 85 kcal mol−1).29 Therefore,
C–CO is cleaved preferentially with respect to the direct
breaking of the C–C bond since (i) the terminal CO group
does not undergo steric hindrance and easily interacts with
the catalyst metal surface and (ii) its removal is helped by the
simultaneous aldehydic C–H (≅95 kcal mol−1) bond breaking
promoted by the metal particles. Accordingly, on the basis of
theoretical studies, it has been reported that the C–C bond
breaking of glycerol on the Pt (111) surface has a very high
barrier with respect to that of the terminal C–CO bond
cleavage.34
An analogous competition of the dehydration/hydrogena-
tion and dehydrogenation processes (hydrogenolysis vs. APR)
occurs starting from EG. The catalytic results at 180 and 210
°C in the presence of H2 show a competition between the
CO–H bond cleavage and C–O bond breaking reactions
(MeOH and EtOH are almost equally formed) while, at 240
°C, the CO–H bond scission reaction becomes more
favourable with a higher production of MeOH confirming the
temperature dependence of the dehydration and dehydroge-
nation routes. Under inert conditions, methanol is the main
reaction product obtained in the liquid phase as a result of
the prevalence of the dehydrogenation/decarbonylation route.
Recent density functional theory (DFT) calculations,
complemented with temperature-programmed desorption
(TPD) experiments, demonstrate that the decomposition of
ethylene glycol on Pt and Pt/Ni surfaces proceeds first
through dehydrogenation to reach one of the species with a
terminal CO followed by C–C bond cleavage.35
All these results allow the overall pathways for describing
the hydrogenolysis vs. APR reactions starting from glycerol to
be drawn, as represented in Scheme 2.
In this regard, it is worth underlining that the C–CO bond
energy value is about 83 kcal mol−1, quite proximate to that
Scheme 2 Hydrogenolysis vs. APR of glycerol: overall reaction pathways.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
The gas phase composition of the reactions carried out
under inert atmosphere are characterized by a H2/CO2 ratio
higher than the maximum value theoretically expected for
the glycerol APR (1):
significant increase in dehydrogenation products was
observed at higher temperatures.
Therefore, a change in selectivity in the C–O and C–C
bond cleavage of biomass derived C2–C3 polyols, upon using
the Pd/Fe catalyst, was observed by varying the reaction
temperature.
C3H8O3 + 3H2O → 7H2 + 3CO2
(1)
In all experiments, CO was never detected in the gas-
phase products thus revealing an excellent performance of
the Pd/Fe catalyst in promoting the water gas shift (WGS)
reaction.
The explanation for the “apparent” anomaly is that the
catalytic tests are conducted in batch mode with the gas
phase products collected in a gas sampler cylinder directly
connected to the autoclave reactor. The sampling system
adopted allows the collection of gas-phase products at about
2/3 of the pressure measured at 25 °C (Fig. S2.5†). Therefore,
under our sampling conditions, it is necessary to take into
consideration that CO2 is much more soluble in water than
H2 IJCO2IJsol)/H2IJsol) ≅ 55). Therefore the H2/CO2 ratio mea-
sured is expected to be higher than the predicted one on the
basis of reaction (1) and even higher as the final pressure
(related to the reaction temperature) increases.
Acknowledgements
The financial support by POR Calabria – FSE 2007/2013 –
(“Backup” Project) and Mediterranea University of Reggio
Calabria is gratefully acknowledged. The authors thank Prof.
Pierluigi Antonucci and Ing. Patrizia Frontera for their assis-
tance with GC analysis, Mr. Gaetano Bevacqua and Mrs.
Noemi Arcadi for the catalytic experiments related to ethylene
glycol and 1,2-propanediol.
In accordance with all these considerations, the measured
pH values of the liquid phase are 6.10, 5.90 and 5.75 for the
reactions carried out respectively at 180, 210 and 240 °C as
expected from the weak acidity of CO2 in water.
Notes and references
Analogous considerations can be done for the APR reac-
tions of EG.
1 C. O. Tuck, E. Pérez, I. T. Horváth, R. A. Sheldon and M.
Poliakoff, Science, 2000, 337, 695.
Another very important peculiarity of the Pd/Fe catalyst is
the absence of CO (as a gas phase product) in all experi-
ments. The obvious implication is that the WGS equilibrium
is completely shifted towards H2 and CO2 formation. Conse-
quently, the Pd/Fe is very efficient in the WGS reaction but
poorly effective as a Fisher–Tropsch or metanation catalyst.
This is easily understandable considering that both palla-
dium and Fe3O4 are known to be efficient catalysts for
affording WGS products. Analogous results were also
obtained by Dumesic's research group which showed a supe-
rior activity of the impregnated Pd/Fe3O4 catalyst with respect
to the analogous Pd/Fe2O3 in the APR of EG as a result of the
enhanced WGS reaction promoted by the synergistic effect of
Pd nanoparticles and magnetite.12
2 P. Gallezot, Chem. Soc. Rev., 2012, 41, 1538.
3 M. Pagliaro and M. Rossi, The Future of Glycerol, 2nd edn,
Royal Society of Chemistry, Cambridge, 2010.
4 C. H. Zhou, H. Zhao, D. S. Tong, L. M. Wu and W. H. Yu,
Catal. Rev.: Sci. Eng., 2013, 55, 369.
5 R. Ciriminna, C. Della Pina, M. Rossi and M. Pagliaro, Eur. J.
Lipid Sci. Technol., 2014, 116, 1432.
6 Y. Nakagawa and K. Tomishige, Catal. Sci. Technol., 2011, 1, 179.
7 A. M. Ruppert, K. Weinberg and R. Palkovits, Angew. Chem.,
Int. Ed., 2012, 51, 2564.
8 F. Mauriello, M. G. Musolino and R. Pietropaolo, in Glycerol:
Production, Structure and Applications, ed. M. De Santos Silva
and P. C. Ferreira, Nova Science Publishers, Inc, New York,
2012, p. 45.
9 R. D. Cortright, R. R. Davda and J. A. Dumesic, Nature,
2002, 418, 964.
10 G. W. Huber and J. A. Dumesic, Catal. Today, 2006, 111, 119.
11 N. H. Tran and G. S. K. Kannangara, Chem. Soc. Rev.,
2013, 42, 9454.
12 G. W. Huber, J. W. Shabaker, S. T. Evans and J. A. Dumesic,
Appl. Catal., B, 2006, 62, 226.
13 A. V. Kirilin, A. V. Tokarev, E. V. Murzina, L. M. Kustov, J.-P.
Mikkola and D. Yu. Murzin, ChemSusChem, 2010, 3, 708.
14 G. Wen, Y. Xu, H. Ma, Z. Xu and Z. Tian, Int. J. Hydrogen
Energy, 2008, 33, 6657.
15 A. Martin, U. Armbruster, I. Gandarias and P. L. Arias, Eur. J.
Lipid Sci. Technol., 2013, 115, 9.
Conclusions
The catalytic conversion of biomass-derived glycerol was
investigated under both hydrogenolysis and aqueous phase
reforming conditions over the bimetallic Pd/Fe catalyst. The
starting alternative processes were found to be the dehydra-
tion/hydrogenation and the dehydrogenation/decarbonylation
with the latter controlled by initial O–H breaking followed by
β C–H elimination rather than direct C–C bond cleavage. At
180 °C, the C–O bond is preferentially cleaved allowing the
production of 1,2-PDO, while C–C bond breaking becomes
favourable at higher reaction temperatures leading to the for-
mation of H2 in a good yield. In analogy, by using ethylene
glycol (C2 polyol) as the starting substrate, the C–O and C–C
bonds are equally cleaved at 180 and 210 °C, whereas a
16 A. Ciftci, D. A. J. M. Ligthart and E. J. M. Hensen, Green
Chem., 2014, 16, 853.
17 D. Roy, B. Subramaniam and R. V. Chaudhari, Catal. Today,
2010, 156, 31.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
18 B. Li, J. Wang, Y. Yuan, H. Ariga, S. Takakusagi and K.
Asakura, ACS Catal., 2011, 1, 1521.
26 D. Scholz, C. Aellig and I. Hermans, ChemSusChem, 2014, 7,
268.
19 M. G. Musolino, C. Busacca, F. Mauriello and R. Pietropaolo,
Appl. Catal., A, 2010, 279, 77.
27 J. H. Bernas, A. Taskinen, J. Wärnå and D. Y. Murzin, J. Mol.
Catal. A: Chem., 2009, 306, 33.
20 M. G. Musolino, L. A. Scarpino, F. Mauriello and R.
Pietropaolo, ChemSusChem, 2011, 4, 1143.
28 Y. Amada, S. Koso, Y. Nakagawa and K. Tomishige,
ChemSusChem, 2010, 3, 728.
21 F. Liao, T. W. B. Lo, D. Sexton, J. Qu, C. T. Wu and S. C. E.
Tsang, Catal. Sci. Technol., 2015, 5, 887.
29 Y. R. Luo, Comprehensive Handbook of Chemical Bond
Energies, CRC Press, Boca Raton, USA, 2007.
22 M. G. Musolino, L. A. Scarpino, F. Mauriello and R.
Pietropaolo, Green Chem., 2009, 11, 1511.
23 F. Mauriello, H. Ariga, M. G. Musolino, R. Pietropaolo, S.
Takakusagi and K. Asakura, Appl. Catal., B., 2015, 166–167,
121.
24 C.-T. Wu, K. M. K. Yu, F. Liao, N. Young, P. Nellist, A.
Dent, A. Kroner and S. C. E. Tsang, Nat. Commun.,
2012, 3, 1050.
30 M. K. Weldon and C. M. Friend, Chem. Rev., 1996, 96, 1391.
31 R. M. Rioux and M. A. Vannice, J. Catal., 2003, 216, 362.
32 A. Wawrzetz, B. Peng, A. Hrabar, A. Jentys, A. A. Lemonidou
and J. A. Lercher, J. Catal., 2010, 269, 411.
33 B. Peng, C. Zhao, I. Mejía-Centeno, G. A. Fuentes, A. Jentys
and J. A. Lercher, Catal. Today, 2012, 183, 3.
34 Y. Chen, M. Salciccioli and D. G. Vlachos, J. Phys. Chem. C,
2011, 115, 18707.
25 J. Liu, B. Sun, J. Hu, Y. Pei, H. Li and M. Qiao, J. Catal.,
2010, 274, 287.
35 M. Salciccioli, W. Yu, M. A. Barteau, J. G. Chen and D. G.
Vlachos, J. Am. Chem. Soc., 2011, 133, 7996.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2015