Hydrogenolysis through catalytic transfer hydrogenation: Glycerol conversion to 1,2-propanediol

Article abstract of DOI:10.1016/j.cattod.2012.03.067

Important biorefinery processes imply hydrogenolysis reactions where high hydrogen pressures are required. As most of the nowadays available hydrogen gas is produced from fossil fuels there are great incentives to develop alternative technologies able to both substitute non-renewable reactants and operate at lower severity conditions. The use of hydrogen donor molecules from renewable origin can be a promising alternative to tackle simultaneously with both objectives. In the present study the use of methanol, 2-propanol and formic acid in the glycerol hydrogenolysis process to obtain 1,2-propanediol was investigated using a Ni-Cu/Al₂O₃ catalyst, prepared by sol-gel method, and under N₂ atmosphere. A semi-continuous set-up was designed in which the donor solution was continuously fed into the autoclave reactor containing the glycerol aqueous phase. The best results in terms of glycerol conversion and 1,2-propanediol selectivity were obtained with formic acid.

Full text of DOI:10.1016/j.cattod.2012.03.067

Catalysis Today 195 (2012) 22–31
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Hydrogenolysis through catalytic transfer hydrogenation: Glycerol conversion to
1
,2-propanediol
∗
Inaki Gandarias , Pedro Luis Arias, Sara G. Fernández, Jesús Requies, Mohammed El Doukkali,
María Belén Güemez
University of the Basque Country (UPV/EHU), calle Alameda Urquijo s/n, 48013 Bilbao, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Important biorefinery processes imply hydrogenolysis reactions where high hydrogen pressures are
required. As most of the nowadays available hydrogen gas is produced from fossil fuels there are great
incentives to develop alternative technologies able to both substitute non-renewable reactants and oper-
ate at lower severity conditions. The use of hydrogen donor molecules from renewable origin can be a
promising alternative to tackle simultaneously with both objectives. In the present study the use of
methanol, 2-propanol and formic acid in the glycerol hydrogenolysis process to obtain 1,2-propanediol
was investigated using a Ni–Cu/Al2O3 catalyst, prepared by sol–gel method, and under N2 atmosphere.
A semi-continuous set-up was designed in which the donor solution was continuously fed into the auto-
clave reactor containing the glycerol aqueous phase. The best results in terms of glycerol conversion and
1,2-propanediol selectivity were obtained with formic acid.
Received 1 February 2012
Received in revised form 19 March 2012
Accepted 21 March 2012
Available online 3 May 2012
Keywords:
Glycerol
1
,2-Propanediol
Hydrogenolysis
In situ hydrogen
Heterogeneous catalysis
© 2012 Elsevier B.V. All rights reserved.
1
. Introduction
published studies related to glycerol hydrogenolysis have been per-
formed under hydrogen pressure, both in gas [8–11] or liquid phase
Biodisel production from vegetable oils through transesterifi-
cation processes cogenerates glycerol at an approximate rate of
kg for every 10 kg of biodiesel. The rapid growing production
[12–25]. Nevertheless, the use of molecular hydrogen presents
some drawbacks. First, because nowadays molecular hydrogen is
still derived from fossil fuels through energy intensive processes.
Second, because hydrogen has high diffusivity is easily ignited and
presents considerable hazards on a large scale [26]. Finally, because
the low solubility of hydrogen in aqueous solutions requires to
operate at elevated pressures. In situ generation of the hydrogen
required obviates most of these difficulties [27].
Hydrogenolysis of glycerol combined with in situ production of
hydrogen by the simultaneous reforming of glycerol has already
been studied [19,28,29]. It was recently reported that catalytic
transfer hydrogenation (CTH) using 2-propanol (2-PO) as a hydro-
gen donor molecule gives higher yields to 1,2-PDO as compared
1
of biodiesel has generated an oversupply of glycerol in the mar-
ket, leading to a significant decrease in the price of glycerol [1].
Excess glycerol disposal is an expensive and environmentally prob-
lematic alternative. Effective valorization of glycerol will therefore
contribute to (i) improve the cost-competitiveness of biodiesel pro-
cesses and (ii) gradually replace fossil fuels by biomass as the source
of organic carbon [2–4]. One promising and valuable compound
that can be obtained from glycerol is 1,2-propanediol (1,2-PDO).
1
,2-PDO is widely used as an anti-freeze agent, as a monomer for
polyester resins and in paints, cosmetic, food, etc. [5]. Currently,
,2-PDO is commercially produced from propylene oxide. The con-
1
to aqueous reforming of glycerol on Ni–Cu/Al O3 based catalysts
2
version of glycerol into 1,2-PDO is a promising process that might
help in the gradual replacement of petroleum-derived liquid fuels
and plastics by biomass derived products [6].
[30]. It was also observed that 2-PO was not able to hydro-
genate acetol, and that therefore when 2-PO was used as hydrogen
donor molecule glycerol hydrogenolysis to 1,2-PDO did not occur
through the widely accepted dehydration–hydrogenation mech-
anism (glycerol → acetol → 1,2-PDO). Based on characterization
results and reactivity trends, a reaction mechanism through inter-
mediate alkoxide formation was proposed (see Scheme 1).
Glycerol hydrogenolysis process to obtain 1,2-PDO involves
C
O bond dissociation and simultaneous addition of hydrogen [7].
Since hydrogenolysis uses hydrogen as a reactant, most of the
In this previous work it was also observed that the amount of
hydrogen donor affects both glycerol conversion and selectivities to
main reaction products, as there is a competition between the OH
groups of glycerol and of 2-PO for the active sites of the catalyst to
∗ Corresponding author at: Escuela Técnica Superior de Ingeniería, Alameda
Urquijo s/n, P.C. 48013 Bilbao, Spain. Tel.: +34 946 017 297; fax: +34 946 014 179.
E-mail address: inaki gandarias@ehu.es (I. Gandarias).
0
920-5861/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2012.03.067

I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
23
Scheme 1. Proposed reaction mechanism for direct glycerol hydrogenolysis to 1,2-PDO using 2-PO as a hydrogen donor molecule [30].
form the intermediate alkoxides. High initial 2-PO concentrations
reduced glycerol reaction rate while increasing 1,2-PDO selectivity.
For low initial 2-PO concentrations, there was not enough hydrogen
available to convert all the glycerol. Based on these findings, the use
of a semi-continuous batch system seems to be interesting as the
hydrogen donor could be continuously pumped into the glycerol
water solution, lowering active site competition and, at the same
time, avoiding hydrogen supply problems.
Fluka) or methanol (99.9%; Alfa Aesar), was pumped at a constant
rate of 0.02 mL/min. In order to investigate the effect of the amount
of hydrogen donor added on glycerol hydrogenolysis process, the
concentration of hydrogen donor in the aqueous solution pumped
was different in each test.
Four liquid samples (1.2–1.5 mL) were taken throughout the
reaction time in order to obtain the evolution of reactant and
product concentrations. These samples were analyzed using a gas
chromatograph (Agilent Technologies, 7890 A) equipped with a
flame ionization detector (FID) and a thermal conductivity detec-
tor (TCD). A Meta-Wax capillary column (diameter 0.53 mm, length
30 m) was used for product separation. For formic acid detection all
the liquid samples were also analyzed by a GC–MS (Agilent Tech-
nologies, 5973). After reaction, the gas phase was collected in a gas
bag and analyzed with another GC–TCD–FID (Agilent Technologies,
In this article, the use of methanol, 2-PO and formic acid as
hydrogen donor molecules in glycerol hydrogenolysis process over
Ni–Cu/Al O catalyst is investigated in a semi-continuous batch
2
3
system. The effect of the type of hydrogen donor and donor feed
rate on glycerol conversion and selectivity to 1,2-PDO is discussed.
2
. Experimental
7
890 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).
2.1. Catalyst
From the GC–FID analysis of the liquid samples, the concentra-
Ni–Cu/Al O was prepared by the sol–gel method. Aluminum
isopropoxide (99.99%; Sigma Aldrich) was dissolved in deionized
2
3
−
1
tion (mol L ) of each reactant and product was obtained. In order
to estimate the moles of each reactant and product in the reactor,
it was necessary to know the volume inside the reactor at the time
the sample was taken. 1 mL of the withdrawn sample (taken with
an automatic pipette) was weighted to know the density of the liq-
uid phase. Next, the grams of the liquid solution in the reactor at
the given time (gt) were calculated using the mass balance equation
below. By multiplying the grams inside the reactor in the time the
sample was taken by the density of the liquid phase, the volume of
the liquid phase inside the reactor was obtained.
water (9 mL of H O per gram of aluminum isopropoxide) by vig-
2
orous stirring of the solution at 313 K. The pH was measured and
kept between 3.8 and 4.2 adding the required amounts of HNO3
0.5 M). Simultaneously, nickel (II) nitrate hexahydrate (99.999%;
(
Sigma Aldrich) and copper (II) nitrate hemi pentahydrate (98.0%;
Alfa Aesar) were dissolved in ethanol. The precursor solution was
slowly added to the aluminum isopropoxide solution. The mixture
was stirred for 30 min at 313 K and then introduced into an ultra-
sonic apparatus for another 30 min. The mixture was then rested
for 24 h at 313 K and subsequently for another 12 h at 375 K. The
product obtained was crushed and calcined from room tempera-
ture to 723 K at a heating rate of 2 K/min. This temperature was
maintained for 4 h. Catalyst samples for activity tests were used in
powdered form with a granule size between 320 and 500 ␮m.
gt = g + gpt − gst − ggt
i
where g are the total grams of the glycerol aqueous solution in the
i
reactor at the beginning of the test; gpt are the grams pumped at a
given time t; gst are the grams of the reaction mixture withdrawn
during liquid sampling at a given time t; ggt are the total grams that
has gone to the gas phase at a given time t.
2
.2. Activity test
The gas phase was only analyzed at the end of the activity
tests. To calculate ggt, it was assumed that gas products were
formed homogenously during the reacting time. Moreover, for sim-
plification it was considered that there were not liquid products
saturating the gas phase of the reactor. Therefore:
The hydrogenolysis of glycerol was carried out in a 50 mL stain-
less steel autoclave with a magnetic stirrer. The catalyst powder
0.5 g) was introduced into the autoclave and the reactor was then
purged with N . After purging, the catalyst was pre-treated during
(
2
4
h under a 300 mL/min flow of 50 vol% H /N2 at 1 bar and 723 K.
2
t
The pre-treatment of the catalysts was done in situ in the autoclave
in order to avoid the contact of reduced samples with air when
the catalyst is transferred to the reactor. After the pre-treatment,
ggt = ggt=600
6
00
where ggt=600 is the amount of grams converted to gas products at
the end of the activity test (after 600 min) and t is any given time
the reactor was purged with N2 or H , the temperature was set
2
to 493 K, and the pressure was increased to 45 bar. 20 mL of the
aqueous solution having 20 wt% of glycerol (99.0%; Sigma Aldrich)
was placed in a feed cylinder, heated to the reaction temperature
and pressurized to 50 bar. Next, the stirring speed was set constant
at 550 rpm. The reaction starting time was established when the
line connecting the feed cylinder and the reactor was opened. N2
or H2 was directed through the feed cylinder for 3 min to stabilize
the pressure of the reactor and to guarantee that all the feed went
into the reactor. During the 10 h reaction time, hydrogen donor
aqueous solution, 2-PO (99.9%; Sigma Aldrich), formic acid (98.0%;
in minutes.
The conversion and selectivity values were calculated on carbon
bases:
ꢀ
ꢀ
C-based mol prod. t = t
Conv. of glyc.% = ꢀ
× 100
C-based mol glyc. t = 0
C-based mol of prod.i t = t
C-based mol of all prod. t = t
Select. of prod.% = ꢀ
× 100

2
4
I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
For mechanistic interpretation, acetol (90.0%; Sigma Aldrich)
was also utilized as reactant in some activity tests.
Being n and m the stoichiometric coefficients, of reactants and
products, respectively, and ꢀG f0 the Gibbs energy of formation at
2
98 K, 1 bar and assuming that all reactants and products behave
0
2.3. Catalyst characterization
as an ideal gas (ꢀGf values were obtained from the Chemical
Engineers’ Handbook [40]). Even though the activity tests were car-
ried out at higher temperatures and pressures, the calculated ꢀGR0
values allow a qualitative discussion about the dehydrogenation
reaction of each selected donor being or not thermodynami-
cally favored. As observed in Scheme 2, methanol conversion to
formaldehyde and H2 is the thermodynamically most disfavored
Ni–Cu/Al O fresh catalyst was characterized by ICP-AES, N2-
2
3
physisorption, H -chemisorption, TPD of NH , TPR and XPS.
2
3
Characterization methods and results are reported elsewhere [30].
A summary of the main characteristics of Ni–Cu/Al O3 catalyst is
2
presented in Table 1.
0
Temperature-programmed oxidation analyses of fresh and used
reduced catalysts were carried out using a thermo-gravimetric
analyzer (Mettler Toledo TGA/SDTA 851e). The standard pro-
tocol involved the pre-treatment of the sample (45–50 mg) in
reaction, as it has the higher positive value for ꢀGR while formic
acid dehydrogenation is a thermodynamically favored reaction due
0
R
to the negative value of the ꢀG .
1
1
25 mL/min of N2 flow from 297 K to 673 K at a heating rate of
0 K/min. The sample was then cooled to 323 K and the weight
3.2. Glycerol hydrogenolysis under N2 atmosphere
change of the sample was continuously monitored during its heat-
ing in 75 mL/min of 10 vol% O2 in He as reactive gas from 323 to
2
-Propanol, methanol and formic acid were studied as hydrogen
donor molecules for the hydrogenolysis of glycerol to 1,2-PDO. The
hydrogen donor water solution was continuously pumped, at a con-
stant rate of 0.02 mL/min, to the 20 mL glycerol aqueous solution
initially placed in the reactor. In order to investigate the effect of the
hydrogen donor pumping rate on glycerol conversion and selectiv-
ity to main reaction products, the concentration of the donor in the
pumping solution was kept constant in each test but modified from
test to test. In the activity tests without hydrogen donor addition,
distilled water was not pumped to the system.
1
173 K at a heating rate of 5 K/min.
For TEM characterization,
equipped with a LaB6 filament and a super win lent operating at
00 kV was used. Bright field images were acquired using a CCD
a
Phillips CM200 microscope
2
camera (TVIPS GmbH). Powders of the catalysts samples were dis-
persed in ethanol with ultrasound and drops were deposited on
copper grids coated with amorphous carbon films.
3
. Results and discussion
3
.2.1. N + 2-propanol
As it can be observed in Table 2, glycerol conversion and selectiv-
2
3.1. Hydrogen donor selection
ity to 1,2-PDO values were higher in all tests where 2-PO was added
as compared to the test without donor addition. This indicates
that glycerol hydrogenolysis to 1,2-PDO by transfer hydrogenation
using 2-PO is more effective than glycerol hydrogenolysis using
the hydrogen produced by simultaneous reforming of glycerol. At
the same time, the positive effect of hydrogen donor addition was
much more significant than the possible negative effect of glycerol
dilution with the pumped water. Indeed, the low reaction order on
glycerol concentration has already been reported [41,42].
Three different hydrogen donor molecules were used in the
activity tests: 2-PO, methanol and formic acid. 2-PO has already
been reported to be active in glycerol hydrogenolysis process to
1
,2-PDO [30,31]. The main reactions taking place with the selected
hydrogen donors are represented in Scheme 2. Under the exper-
imental conditions, 2-PO can be dehydrogenated to acetone or
dehydrated to propene. Propene can be further hydrogenated to
propane using the hydrogen available from 2-PO dehydrogena-
tion. It was observed that, under the experimental conditions used,
acetone is also dehydrated to propene [30]. Formic acid together
with levulinic acid and furfural are formed from non-food biomass
feedstock by acid catalyzed hydrolysis [32]. Formic acid can also
be used as efficient storage material for H2 [33,34], as it can
release hydrogen at low temperatures using iron catalysts [35,36].
Other potential uses of formic acid are currently being investigated
and catalytic transfer hydrogenation with formic acid as hydro-
gen source could be an important application. Formic acid can be
Regarding the effect of 2-PO pumping rate, it can be observed
in Table 2 that both glycerol conversion and 1,2-PDO selectivity
increased with increasing pumping rate of 2-PO, until a maximum
−
1
was achieved at 7.2 mmol h of 2-PO fed. For higher feeding rates
of 2-PO (runs 4 and 5), similar glycerol conversions but lower selec-
tivities to 1,2-PDO were observed. Therefore, there is an optimum
2
-PO feeding rate that maximizes 1,2-PDO formation. At low feed-
ing rates, increasing 2-PO feeding rate increased 1,2-PDO selectivity
and glycerol conversion, because more 2-PO molecules reacted
providing a higher number of active hydrogen atoms. Neverthe-
catalytically decomposed to give hydrogen and CO . A parallel reac-
2
tion gives water and CO, which can be further converted to CO2
and hydrogen through water gas shift reaction [37]. Methanol is a
widely used molecule in the catalytic hydrogenation of unsaturated
compounds [38,39]. CTH using methanol is promoted in water solu-
tions as formaldehyde readily reacts with water to form formic acid,
which further reacts to yield H and CO [35]. Therefore, from each
−1
less, above 7.2 mmol h , further increments in 2-PO feeding rate
did not increase glycerol conversion and had a negative effect on
1
,2-PDO selectivity. It seems that there are negative factors asso-
ciated to high 2-PO availability in the system that affect glycerol
hydrogenolysis to 1,2-PDO.
2
2
Focusing on hydrogen donor behavior, it can be noticed in
Table 2 that the pumping rate of 2-PO did not have significant
influence in the conversion of 2-PO, but affected the selectivity to
dehydrogenation and dehydration products. For low 2-PO feeding
rates (runs 2 and 3), low selectivities to acetone were detected,
indicating a further dehydration of the formed acetone to propene.
In all the tests propane was detected in the gas phase, indicating
that part of the formed propene was hydrogenated to propane.
molecule of methanol, three hydrogen molecules can be formed.
An additional advantage is gained when the products of the
decomposing donor have large negative Gibbs energy of formation.
Thus, hydrogen donor molecules that produce gaseous products
after their dehydrogenation, like formic acid giving CO , have added
2
driving force to their reactivity. In Scheme 2 the standard Gibbs
energy of the dehydrogenation reactions of the three hydrogen
donors are represented. These values were calculated from the
following equation:
−
1
In the light of these results, the reason why above 7.2 mmol h
further increments in 2-PO pumping rate did not increment glyc-
erol conversion and had a negative effect on 1,2-PDO selectivity
seems to be a combination of various factors. It was previously
ꢁ
ꢁ
0
R
nꢀG0
mꢀG0
f (reactants)
ꢀ
G
=
−
f (products)

I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
25
Table 1
Characteristics of the catalyst.
Ni (wt%)a
7.7
Cu (wt%)a
28.0
SBET (m /g Al2O3)
2
Dispersion (%)
1.6
MSAb (m /g)
2
Particle size (nm)c
65.6
Des. NH3 × 10 (mol g )d
4
−1
Sample
Ni–Cu/Al2O3
208
3.7
3.4
a
Chemical composition determined with ICP.
Metal surface area.
Determined from H2 chemisorption analysis.
b
c
d
Acidity of the fresh samples from temperature-programmed desorption of ammonia.
Scheme 2. Main reactions taking place with the selected hydrogen donors.
Table 2
Glycerol conversion and selectivity to 1,2-PDO and acetol, and 2-PO conversion and selectivity to acetone, propene and propane as a function of 2-PO feed rate. 45 bar N2
pressure, 10 h, 493 K, 20 mL (20 wt%) glycerol aqueous solution, 166 mg catalyst/g glycerol, 0.02 mL/min donor solution feed rate.
Run
2-PO conc. pumped
solution (wt%)
2-PO feed-rate
Glyc. conv. (%)
Selectivity (%)a
(from glycerol)
C-balance
2-PO conv. (%)
Selectivity (%) (from 2-PO)
−
1
(mmol h
)
1,2-PDO
Acetol
Acetone
Propene
Propane
1b
2
3
4
5
–
20.0
40.0
60.0
80.0
–
15.8
17.8
28.2
27.8
27.6
49.3
53.1
77.4
53.0
56.3
38.8
44.5
15.4
42.3
40.7
97.2
103.4
97.5
93.6
94.1
–
–
1.2
7.7
18.0
19.7
–
–
8.2
33.4
21.5
19.8
3.6
7.2
9.8
12.5
51.8
46.4
50.4
49.7
90.6
58.8
60.4
60.5
a
The other products detected coming from glycerol were ethylene glycol, ethanol, 1-propanol and methane.
No solution was pumped to the reactor.
b
reported that glycerol and 2-PO compete for the same active sites
considered is that propene, formed from 2-PO and acetone dehy-
dration, hydrogenates to propane competing with glycerol for the
active hydrogen atoms. At high 2-PO feeding rates, there is higher
propene formation and therefore higher competition.
Carbon dioxide was only detected in run 1, indicating that
aqueous phase reforming of glycerol did only take place when
no hydrogen donor was introduced in the system. It has already
been observed that glycerol APR did not take place when exter-
nal hydrogen is introduced in the system [19], due to a decrease in
the surface concentrations of reactive intermediates formed from
glycerol dehydrogenation reactions [43].
for the formation of alkoxides [30]. In fact, the donor and the accep-
tor need to be adsorbed on contiguous active sites in order to make
possible the transfer hydrogenation [27]. If the active hydrogen
atoms formed from 2-PO are not adsorbed near a glycerol molecule,
they combine to yield molecular hydrogen, while if glycerol is not
adsorbed near an active hydrogen atom it does not react to 1,2-
PDO. Hence, it seems that there should be an optimum proportion
between adsorbed donor and acceptor molecules that maximizes
the production of 1,2-PDO. For 2-PO, this optimum was obtained
−1
for a feeding rate near 7.2 mmol h . Another factor that needs to be

2
6
I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
Table 3
Glycerol conversion and selectivity to 1,2-PDO and acetol, and formic acid conversion and selectivity to CO and CO2 as a function of formic acid feed rate. 45 bar N2 pressure,
1
0 h, 493 K, 20 mL (20 wt%) glycerol aqueous solution, 0.5 g of catalyst, 0.02 mL/min donor solution feed rate.
Run
Formic acid conc.
pumped solution
Formic acid
feed-rate
Glyc. conv. (%)
Selectivity (%)a
(from glycerol)
C-balance
Formic acid
conv. (%)
Selectivity (%)
(from formic acid)
(wt%)
(mmol h−1)
1,2-PDO
Acetol
CO
CO2
1b
2
3
4
5
0
3.5
7.0
14.0
28.0
42.0
7.0
0
15.8
22.4
33.5
28.0
20.2
4.5
49.3
63.5
85.9
78.7
77.6
73.0
0.0
38.8
28.0
6.7
11.5
12.3
17.9
35.5
32.4
97.2
94.3
96.8
95.1
93.9
101.7
104.5
98.5
–
100
100
100
100
100
164.5
100
–
0
0
0
0
0
0
0
–
100
100
100
100
100
100
100
0.9
1.8
3.6
7.2
10.8
1.8
–
6
7
8
c
1.2
16.2
d
–
55.8
a
The other products detected coming from glycerol were ethylene glycol, ethanol, 1-propanol and methane.
No solution was pumped to the reactor.
This test was carried out without catalyst.
b
c
d
1
8 mmol of formic acid were placed from the beginning of the reaction in the reactor and no solution was pumped to the reactor.
3
.2.2. N + formic acid
As observed in Table 3, glycerol conversion and selectivity to 1,2-
acid pumping rate was offset by the increase in active site compe-
tition. Indeed, the occupation of active sites by formic acid was so
significant that in the test with the higher pumping rate (run 6 of
Table 3) the achieved glycerol conversion was even lower than the
one obtained under N2 pressure (run 1 of Table 3) without donor
addition.
2
PDO increased with formic acid feeding rate until a maximum was
−1
obtained in the test with 1.8 mmol h feeding rate. With this opti-
mized hydrogen donor supply, a selectivity value to 1,2-PDO of 86%
was achieved, which is a considerable high value and comparable
to the best values previously reported in processes under H2 pres-
sure [8,10,23]. For higher pumping rates, both glycerol conversion
and the selectivity to 1,2-PDO decreased.
Regarding formic acid behavior, the most interesting aspect is
that in all the activity tests, and therefore for all the range of pump-
ing rates studied, there was 100% conversion of formic acid. No
formic acid was detected in any of the tests and in any of the five
liquid samples taken during each test, which means that formic acid
readily reacts under the experimental conditions. As it was pointed
Finally, a further test was carried out in order to compare the
semi-batch system, in which the donor is continuously pumped
into the reactor, with the batch system in which the donor is placed
directly in the feed cylinder with glycerol at the beginning of the
test. In this test, run 8 Table 3, the mmol of formic acid (18) placed
in the feed cylinder at the beginning of the test were the same
as the mmol of formic acid pumped into the reactor during run
3. As it can be observed, even though the total mmol of formic
acid that reacted, and therefore the total active hydrogen atoms
formed, were the same in both tests, glycerol conversion was dou-
bled in the test performed with the semi-batch system. It needs to
be considered that in the first liquid sample of run 8, after 2 h, no
formic acid was detected. This means that the 18 mmol of formic
acid initially placed in the batch reactor readily reacted at the begin-
ning of the test, without transferring most of the active hydrogen
atoms to glycerol or intermediates, but forming molecular hydro-
gen that was released to the gas phase. This result demonstrates
the superiority of the semi-batch reactor, as it reduces the compe-
tition between formic acid and glycerol for active sites, ensuring
a continuous supply of active hydrogen atoms at an appropriate
rate.
out before, the formation of gaseous CO , with a large negative
2
Gibbs energy of formation provides an important driving force for
the reactivity of formic acid. Negligible CO amounts were detected
in the analysis of the gas phase, indicating that the possible CO
formed through formic acid decomposition was readily converted
to CO2 by water gas shift reaction The mol of CO2 detected in the
gas phase quite precisely coincided with the mol of formic acid fed,
which indicates that glycerol reforming did not significantly take
place when formic acid was fed.
In order to ensure that formic acid decomposition occurred in
the active sites of the catalyst, a further activity test was performed
(
run 7 of Table 3) without the addition of the catalyst. As it can be
observed, without catalyst, formic acid conversion after 10 h was
only 14.5%. This proves that under the operating conditions used,
formic acid dehydrogenation mainly took place in the active sites
3
.2.3. N + methanol
The results of the activity tests carried out under nitrogen atmo-
2
of the Ni–Cu/Al O3 catalyst. It is likely that formic acid adsorbs
2
sphere and methanol as hydrogen donor molecule are shown in
Table 4. Again, glycerol conversion presented a maximum as a func-
tion of methanol pumping rate. Nevertheless, the effect of methanol
pumping rate on glycerol conversion was not so marked as com-
pared to formic acid pumping rate, indicating a lower affinity for
active sites of methanol than of formic acid.
It can be observed in Table 4 that methanol conversion
decreased with increasing pumped methanol. As compared to 2-PO
and formic acid conversions (see Tables 2 and 3), methanol conver-
sions were considerably lower. Moreover, neither formaldehyde
nor formic acid was detected in any of the liquid samples ana-
lyzed during the experiments. This means that the controlling step
is the thermodynamically disfavored methanol dehydrogenation
to formaldehyde, while the further reaction of formaldehyde with
water to yield formic acid and the subsequent formic acid decom-
position to hydrogen and CO2 occurred almost instantaneously.
This behavior has already been observed for heterogeneous transfer
dissociatively to give a formate species and an adsorbed hydrogen
atom, and that the formate species then dissociate to give gaseous
CO2 and another adsorbed hydrogen atom [37].
As it was previously stated, when formic acid pumping rate
−1
increased above 1.8 mmol h , glycerol conversion significantly
decreased. In fact, high formic acid pumping rates affected in
greater extent glycerol conversion than high 2-PO pumping rates. It
seems that glycerol and formic acid compete for the active sites of
the catalyst, and that formic acid has higher affinity for adsorption
sites than glycerol and 2-PO. The high activity of formic acid under
the operating conditions used supports this idea. For the pumping
−1
rate interval between 0.0 and 1.8 mmol h , increments in formic
acid feeding rate increased both glycerol conversion and selectiv-
ity to 1,2-PDO, as higher number of active hydrogen atoms were
−
1
available in the system. However, above 1.8 mmol h the positive
effect of higher hydrogen atoms availability with increasing formic

I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
27
Table 4
Glycerol conversion and selectivity to 1,2-PDO and acetol, and methanol conversion and selectivity to formaldehyde, formic acid and CO2 as a function of methanol feed rate.
5 bar N2 pressure, 10 h, 493 K, 20 mL (20 wt%) glycerol aqueous solution, 0.5 g of catalyst, 0.02 mL/min donor solution feed rate.
4
Run
Methanol conc.
pumped solution
Methanol feed-rate
Glyc. conv. (%)
Selectivity (%)a
(from glycerol)
C-balance
Methanol
conv. (%)
Selectivity (%)
(from methanol)
−1
(mmol h
)
(wt%)
1,2-PDO
Acetol
Formald.
Form. ac.
CO2
1b
2
3
4
5
–
3.5
7.0
10.0
21.0
30.0
–
15.8
21.0
24.4
26.2
21.5
21.3
49.3
47.0
49.0
51.2
49.7
50.7
38.8
49.7
48.1
44.3
46.8
45.9
97.2
92.9
100.8
95.3
96.7
96.5
–
–
0
0
0
0
0
–
0
0
0
0
0
–
100
100
100
100
100
1.2
2.4
4.0
7.2
9.8
43.0
40.8
39.1
38.2
27.6
6
a
The other products detected coming from glycerol were ethylene glycol, ethanol, 1-propanol and methane.
No solution was pumped to the reactor.
b
hydrogenation using methanol as hydrogen donor over Pd-based
catalysts [38].
using Pt–NaY [28] and Roy et al. using a mixture of Ru/Al O3 and
2
Pt/Al O3 [29] obtained higher glycerol conversions. Nevertheless,
2
reported selectivities were significantly lower (64.0%) and (47.2%)
respectively. The lower C C bond cleavage activity of Ni–Cu metal
sites as compared to Pt sites, and the fact that the addition of
formic acid suppresses glycerol reforming might be the main fac-
tors behind the higher measured selectivities.
3
.2.4. Comparison between hydrogen donors
In order to get a better understanding of the different perfor-
mance of the three hydrogen donors tested, it seems interesting to
relate the total moles of hydrogen atoms formed from each donor
to the total moles of 1,2-PDO obtained (see Fig. 1A). For each donor,
the test with higher 1,2-PDO yield was selected. It has to be con-
sidered that for every molecule of 2-PO or formic acid that reacted,
two hydrogen atoms were obtained, while for every molecule of
methanol that reacted, 6 hydrogen atoms were obtained. Taking
into account that two hydrogen atoms are needed to convert glyc-
erol to 1,2-PDO, the efficiency of each donor was calculated by
dividing the moles of 1,2-PDO obtained by the half of the moles
of hydrogen atoms formed (see Fig. 1B).
3.3. Glycerol hydrogenolysis under H2 pressure
It is well known that there are differences between heteroge-
neous catalytic hydrogenation using molecular hydrogen as the
hydrogen source, and heterogeneous catalytic hydrogenation using
hydrogen donor molecules as the source of hydrogen [27]. Indeed,
for glycerol hydrogenolysis process, different mechanisms are
involved as a function of the hydrogen source [30]. Consequently,
it was decided not only to compare, the activity and selectivity
regarding the hydrogen source, but also to perform activity tests
under H2 pressure and adding a hydrogen donor. The aim was to
determine if there is a useful synergy for glycerol hydrogenolysis
through the simultaneous use of both types of hydrogen sources.
As formic acid was the most appropriate hydrogen donor
molecule under the experimental conditions used, activity tests
under H2 pressure and with the addition of formic acid were car-
ried out. Again, in order to study the effect of the amount of formic
acid added on glycerol conversion and selectivity to main reaction
products, the concentration of formic acid in the pumping solution
was kept constant in each run but modified from run to run. In the
test without donor addition (run 1 of Table 5), no water solution
was pumped to the system.
As it can be observed in Fig. 1A, the greater number of moles of
,2-PDO was obtained when formic acid was used as a hydrogen
1
donor molecule, while the smallest with methanol. Even though,
much more hydrogen atoms were formed from methanol. The effi-
ciency of each donor can be better compared in Fig. 1B. For formic
acid, around 72% of the hydrogen atoms formed were transferred
to 1,2-PDO, being the efficiency in the case of 2-PO and methanol
much lower (30 and 13%, respectively). As it was addressed before,
in the case of 2-PO one reason for the lower efficiency is that part of
the propene coming from acetone dehydration was hydrogenated
to propane, consuming hydrogen atoms. In the case of methanol,
it is not clear why so low efficiency was obtained. Maybe the fact
that hydrogen is formed in three consecutive steps is not appropri-
ate for the glycerol hydrogenolysis reaction. In fact, it was observed
that under the operating conditions used the first dehydrogenation
to formaldehyde was the rate controlling step, while the other two
dehydrogenations occur readily. Therefore, an adsorbed glycerol
molecule might interact with two active hydrogen atoms, com-
ing from methanol dehydrogenation to formaldehyde, and yield
In Table 5 the main results achieved in the activity tests under H2
pressure are displayed. It is interesting to recognize that formic acid
was not detected in any of the samples of the different activity tests,
which proves that formic acid readily reacts also in the presence of
dissolved molecular hydrogen. Again, CO2 mol detected in the gas
phase quite well coincided with the mol of formic acid fed; suggest-
ing that also under hydrogen pressure formic acid was completely
converted to CO2 and two hydrogen atoms.
1
,2-PDO. Nevertheless, the subsequent 1,2-PDO desorption and
occupation of the active site by another glycerol molecule might
not be as fast as formaldehyde reaction with water and consecu-
tive formic acid dehydrogenation. As a consequence of this, the four
additional hydrogen atoms formed might not find a near adsorbed
glycerol molecule, and hence, are combined to form two hydrogen
molecules that are released to the gas phase. Indeed, the highest H2
concentration in the gas phase was detected for the activity tests
performed with methanol.
Formic acid feeding rate significantly affected glycerol con-
version, with little effect on the selectivity to 1,2-PDO. For low
pumping rates, the addition of formic acid had a positive effect.
−
1
Indeed, in the test with a formic acid pumping rate of 1.8 mmol h
(run 3 of Table 5) glycerol conversion and yield to 1,2-PDO were
26% higher than the one obtained under H pressure without donor
2
In the light of these results it is clear that formic acid is the best
hydrogen donor molecule for the glycerol hydrogenolysis process
under the studied conditions, as the highest yield to 1,2-PDO was
obtained with the smallest amount of donor used. If we compare the
obtained results with other reference works in liquid phase glyc-
erol hydrogenolysis with in situ generated hydrogen, D ˇı Hont et al.
addition (run 1 of Table 5). It might appear that the amount of
hydrogen supplied by formic acid is negligible as compared to
all the hydrogen that is in the system. Nonetheless, two impor-
tant aspects should be taken into account. First, the activity tests
were carried out in liquid phase, and the solubility of hydro-
gen in water solutions is very low. Hence, even though the gas

2
8
I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
Fig. 1. (A) Total mmol of 1,2-PDO obtained as a function of the mmol of hydrogen atoms formed from each donor (2-PO, formic acid or methanol) in the tests with the higher
yield to 1,2-PDO. (B) Efficiency of each hydrogen donor in the test with higher yield to 1,2-PDO: hydrogen donor efficiency (%) = 100 × (mol 1,2-PDO formed/(1/2 × mol H
atoms formed).
Table 5
Glycerol conversion and selectivity to 1,2-PDO and acetol, and formic acid conversion and selectivity to CO and CO2 as a function of formic acid feed rate. 45 bar H2 pressure,
1
0 h, 493 K, 20 mL (20 wt%) glycerol aqueous solution, 0.5 g of catalyst, 0.02 mL/min donor solution feed rate.
Run
Formic acid conc.
pump solution
Formic acid
feed-rate
Glyc. conv. (%)
Selectivity (%)a
(from glycerol)
C-balance
Formic acid
conv. (%)
Selectivity (%)
(from formic acid)
(wt%)
(mmol h−1)
1,2-PDO
Acetol
CO
CO2
1b
2
3
4
5
–
3.5
7.0
14.0
28.0
–
34.8
35.4
43.9
27.4
24.0
90.9
89.8
89.4
89.5
89.5
1.6
1.2
1.8
3.0
2.8
96.3
97.7
95.2
104.4
103.5
–
100
100
100
100
–
0
0
0
0
–
100
100
100
100
0.9
1.8
3.6
7.2
a
The other products detected coming from glycerol were ethylene glycol, ethanol, 1-propanol and methane.
No solution was pumped to the reactor.
b
phase is mainly formed by H , in the liquid phase the amount of
dissolved hydrogen molecules is much lower. For a 20 wt% glyc-
erol water solution, 45 bar of hydrogen pressure and 493 K, the
the catalyst. In other words, formic acid does not provide molecular
hydrogen; it provides adsorbed hydrogen atoms ready to interact
with a glycerol molecule adsorbed in a contiguous site.
2
concentration of H in the liquid phase was calculated to be around
However, for high pumping rates, the effect of adding formic
acid is negative as lower glycerol conversions were obtained (runs
4 and 5 of Table 5) as compared to the test without formic acid addi-
tion (run 1 of Table 5). It seems that when the supply of formic acid
is too high, and due to its high affinity for active sites, it reduces
the active sites available for glycerol and as a result also glyc-
erol activity. It should be noticed that the formic acid feeding rate
2
−1
0
.029 mol kg [19], which means that in our liquid phase there
−4
−2
were around 6 × 10 mol of H2 dissolved. The 3.6 × 10 mol of
hydrogen atoms supplied by formic acid are not therefore negligi-
ble. Second, the hydrogen molecules dissolved in the liquid phase
have to dissociate in the active sites of the catalysts before interact-
ing with adsorbed glycerol molecules. In the case of formic acid, it
provides those active hydrogen atoms directly in the active sites of
−
1
that maximizes glycerol conversion was the same (1.8 mmol h
regardless the test was conducted under H2 or N2 pressure.
)
Fig. 2. Time evolution of acetol conversion and selectivity to 1,2-PDO under H2
−
1
pressure and under N2 pressure with the continuous pumping of 1.8 mmol h of
formic acid. 45 bar H2 or N2 pressure, 493 K, 20 mL (20 wt%) acetol aqueous solution,
Fig. 3. Relative weight change profile corresponding to the TGA-TPO of the reduced
fresh Ni-Cu/Al O and Ni-Cu/Al O samples used under H2 or N2 pressure, with or
without the addition of donor.
2
3
2
3
0
.5 g of catalyst.

I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
29
Fig. 4. TEM images of reduced fresh and used Ni–Cu/Al2O3 samples: (a) reduced fresh, (b) used under N2 pressure, (c) used under H2 pressure + formic acid.
Fig. 5. Particle size distribution of reduced fresh and spent Ni–Cu/Al2O3 samples used with different sources of hydrogen.

3
0
I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
Concerning the selectivity to 1,2-PDO, there is not a signif-
In Fig. 4 TEM images corresponding to reduced fresh and spent
Ni–Cu/Al O samples are displayed, while the results from the size
icant influence of the amount of formic acid added. However,
higher selectivities to 1,2-PDO were obtained under H2 pressure
as compared to the same tests but under N2 atmosphere (see
Tables 3 and 5). The main difference is the significant decrease in
acetol selectivity in the tests under H2 pressure.
In order to check if formic acid is able to hydrogenate acetol to
,2-PDO, another two tests were performed with acetol as reactant.
In Fig. 2 the evolution of acetol conversion with reaction time for the
2
3
distribution analysis of the metallic particles in the alumina sur-
face are presented in Fig. 5. For all the samples, the size intervals
30–40 and 40–50 nm were the ones with the highest amount of
particles. In the case of the reduced fresh sample, the calculated
average particle size was 42.2 nm, which is lower than the particle
size obtained from the chemisorption analysis result (65.6 nm) pre-
sented in Table 1. In the case of the reduced fresh sample, particles
with a size larger than 80 nm were not detected, however, for spent
samples, some particles in the range of 80–110 nm were observed.
Even though the proportion of these particles was small as com-
pared to the ones in the range of 30–50, this result indicates that
there is a small degree of sintering under the operating conditions
used. Concerning coke formation, no coke deposits were observed
in any of the images of the samples analyzed, which is consistent
with the TGA–TPO results presented above.
1
test performed under H pressure and for the test performed under
2
N2 pressure and with the addition of formic acid is displayed. Ace-
tol quickly reacted in both activity tests. Nevertheless, in the test
under H pressure, most of the acetol was hydrogenated to 1,2-PDO,
2
while under N2 pressure and formic acid really low selectivity to
,2-PDO was achieved. In the experiments under N2 pressure, ace-
1
tol reacted to many different products. Identification by GC–MS
revealed the formation of a wide range of C5–C6 compounds,
such as 3-hexanol-5-methyl, 3-hexanone, 3,5-hexadien-2-ol. Ace-
tol could not be hydrogenated to 1,2-PDO when the hydrogen
supply came from formic acid. This result (identical to the one
obtained for 2-PO [26]) confirms that also for formic acid glycerol
conversion to 1,2-PDO occurs through the direct hydrogenolysis
of the intermediate alkoxide, 1,3-dihydroxy isopropoxide, and not
through the dehydration to acetol and subsequent hydrogenation.
Both sources of hydrogen are therefore complementary. At low
feeding rates, 1,2-PDO is formed not only from glycerol through
intermediate alkoxide formation using the hydrogen species com-
ing from formic acid, but also from acetol hydrogenation in the
presence of dissolved molecular hydrogen. Consequently, for opti-
mized amounts of the donor, the effect of adding formic acid was
positive. However, at high donor feeding rates, the really active
formic acid occupies a high proportion of active sites, reducing glyc-
erol accessibility to catalytic sites and consequently also glycerol
reaction rate.
4. Conclusions
2
-PO, formic acid and methanol were used as hydrogen donor
molecules in the catalytic transfer hydrogenation process to con-
vert glycerol into 1,2-PDO. As glycerol and the hydrogen donors
compete for the same active sites, a semi-continuous process in
which the donor was continuously pumped into the autoclave reac-
tor containing the aqueous glycerol was developed. Activity test
results showed that there must be a balance between the posi-
tive effect of the higher supply of active hydrogen and the negative
effect of the higher competition for active sites when increasing
the pumping rate of the hydrogen donor molecule. Therefore, for
each hydrogen donor tested there was an optimum in the feeding
rate that maximized the yield to 1,2-PDO. Formic acid proved to
be the most effective hydrogen donor molecule, as higher glycerol
conversions and selectivities to 1,2-PDO were obtained with the
lowest amount of hydrogen donor used.
3
.4. Used catalyst characterization
Acknowledgements
Reduced fresh and spent Ni–Cu/Al O3 samples were character-
2
ized by TGA–TPO and also analyzed by TEM images in order to
determine if changes occurred in the catalyst during the activity
test, and if these changes were different as a function of the reacting
atmosphere.
As the thermogravimetric equipment used was not coupled
to mass spectrometry, it was not possible to quantify the exact
amount of coke formed in the spent samples. Nevertheless, it was
possible to carry out a qualitative discussion comparing the final
weight loss between reduced fresh and spent samples after the
TGA–TPO analysis. Fig. 3 illustrates the relative weight change pro-
file for each reduced sample. In all the samples, increments in
sample weight were measured, assigned to the oxidation of pre-
viously reduced Cu and Ni metal sites. The weight loss suffered by
This work was supported by funds from the Spanish Min-
istry of Science and Innovation ENE2009-12743-C04-04, and from
the Basque Government (Researcher Training Programme of the
Department of Education, Universities and Research). The authors
also gratefully acknowledge the University of the Basque Country
for their technical support.
References
[
[
1] G.J. Suppes, W.R. Sutterling, M.A. Dasari WO2005095536 (2005).
2] A. Corma, S. Iborra, A. Velty, Chemical Reviews 107 (2007) 2411.
[3] A.J. Ragauskas, C.K. Williams, B.H. Davison, G. Britovsek, J. Cairney, C.A. Eck-
ert, W.J. Frederick, J.P. Hallett, D.J. Leak, C.L. Liotta, J.R. Mielenz, R. Murphy, R.
Templer, T. Tschaplinski, Science 311 (2006) 484.
[4] Y. Nakagawa, K. Tomishige, Catalaysis Science and Technology 1 (2011) 179.
[5] M. Pagliaro, M. Rossi, The Future of Glycerol: New Usages for a Versatile Raw
Material, first ed., RSC Publishing, Cambridge, 2008.
the fresh Ni–Cu/Al O sample at high temperature range (T > 700 K)
2
3
might be related to the decomposition of the nitrates remaining
in the catalyst after calcination at 723 K [44]. The relative differ-
ences in the final weights between the reduced fresh and the used
samples can be ascribed to the formation of coke deposits during
the activity tests. These relative weight variations as a function
[
[
6] J.J.J. Bozell, G.R. Petersen, Greem Chemistry 12 (2010) 539.
7] M. Schlaf, Dalton Transactions (2006) 4645.
[8] M.A. Dasari, P. Kiatsimkul, W.R. Sutterlin, G.J. Suppes, Applied Catalysis A: Gen-
eral 281 (2005) 225.
[
9] S. Sato, M. Akiyama, R. Takahashi, T. Hara, K. Inui, M. Yokota, Applied Catalysis
A: General 347 (2008) 186.
of the hydrogen source follow the order: H + formic (0.8%) < N2
2
(
3.5%) ≈ N + formic acid (3.8%) < H (5.8%). As it can be observed,
[10] M. Akiyama, S. Sato, R. Takahashi, K. Inui, M. Yokota, Applied Catalysis A: Gen-
eral 371 (2009) 60.
2
2
the presence of formic acid reduced coke formation under H pres-
2
[
11] L. Huang, Y. Zhu, H. Zheng, Y. Li, Z. Zeng, Journal of Chemical Technology and
Biotechnology 83 (2008) 1670.
sure, and had not significant effect under N2 pressure. The relative
differences are quite small, so it can be said that there were low
coke formation rates during the tests under the operating condi-
tions used. Nevertheless, long term activity tests and tests using
recycled catalyst should be performed in order to check if the coke
formation affects catalyst stability.
[
12] E.S. Vasiliadou, E. Heracleous, I.A. Vasalos, A.A. Lemonidou, Applied Catalysis B:
Environmental 92 (2009) 90.
[
[
13] E.S. Vasiliadou, A.A. Lemonidou, Applied Catalysis A: General 396 (2011) 177.
14] C. Montassier, J.C. Ménézo, L.C. Hoang, C. Renaud, J. Barbier, Journal of Molecular
Catalysis 70 (1991) 99.
[15] J. Wang, S. Shen, B. Li, H. Lin, Y. Yuan, Chemistry Letters 38 (2009) 572.

I. Gandarias et al. / Catalysis Today 195 (2012) 22–31
31
[
[
[
[
[
[
16] Y. Kusunoki, T. Miyazawa, K. Kunimori, K. Tomishige, Catalysis Communica-
[30] I. Gandarias, P.L. Arias, J. Requies, M. El Doukkali, M.B. Güemez, Journal of
Catalysis 282 (2011) 237.
[31] M.G. Musolino, L.A. Scarpino, F. Mauriello, R. Pietropaolo, Greem Chemistry 11
(2009) 1511.
[32] D.J. Hayes, S. Fitzpatrick, M.H.B. Hayes, J.R.H. Ross, Biorefineries-Industrial Pro-
cesses and Products, Wiley-VCH, Weinheim, 2006.
[33] B. Loges, A. Boddien, F. Gaertner, H. Junge, M. Beller, Topics in Catalysis 53
(2010) 902.
tions 6 (2005) 645.
17] T. Miyazawa, Y. Kusunoki, K. Kunimori, K. Tomishige, Journal of Catalysis 240
(
2006) 213.
18] T. Miyazawa, S. Koso, K. Kunimori, K. Tomishige, Applied Catalysis A: General
29 (2007) 30.
3
19] I. Gandarias, P.L. Arias, J. Requies, M.B. Güemez, J.L.G. Fierro, Applied Catalysis
B: Environmental 97 (2010) 248.
20] J. Feng, H. Fu, J. Wang, R. Li, H. Chen, X. Li, Catalysis Communications 9 (2008)
[34] A. Boddien, F. Gaertner, D. Mellmann, P. Sponholz, H. Junge, G. Laurenczy, M.
Beller, Chimia 65 (2011) 214.
1
458.
21] E.P. Maris, W.C. Ketchie, M. Murayama, R.J. Davis, Journal of Catalysis 251 (2007)
81.
[35] A. Boddien, B. Loges, F. Gaertner, C. Torborg, K. Fumino, H. Junge, R. Ludwig, M.
Beller, Journal of the American Chemical Society 132 (2010) 8924.
[36] A. Boddien, D. Mellmann, F. Gaertner, R. Jackstell, H. Junge, P.J. Dyson, G. Lau-
renczy, R. Ludwig, M. Beller, Science 333 (2011) 1733.
[37] D.A. Bulushev, J.R.H. Ross, Catalysis Today 163 (2011) 42.
[38] Y.Y. Xiang, X. Li, C. Lu, L. Ma, Q. Zhang, Applied Catalysis A: General 375 (2010)
289.
[39] K. Tani, A. Iseki, T. Yamagata, Chemical Communications (1999) 1821.
[40] R.H. Perry, D.W. Green, Chemical Engineers’ Handbook, seventh ed., Mc-Graw
Hill, USA, 1999, p. 187, Section 2.
[41] Y. Amada, Y. Shinmi, S. Koso, T. Kubota, Y. Nakagawa, K. Tomishige, Applied
Catalysis B: Environmental 105 (2011) 117.
[42] D.G. Lahr, B.H. Shanks, Journal of Catalysis 232 (2005) 386.
[43] J.W. Shabaker, R.R. Davda, G.W. Huber, R.D. Cortright, J.A. Dumesic, Journal of
Catalysis 215 (2003) 344.
2
[
[
22] E.P. Maris, R.J. Davis, Journal of Catalysis 249 (2007) 328.
23] Z. Yuan, L. Wang, J. Wang, S. Xia, P. Chen, Z. Hou, X. Zheng, Applied Catalysis B:
Environmental 101 (2011) 431.
[
[
[
[
[
[
24] Z. Huang, F. Cui, H. Kang, J. Chen, C. Xia, Applied Catalysis A: General 366 (2009)
2
88.
25] L. Guo, J. Zhou, J. Mao, X. Guo, S. Zhang, Applied Catalysis A: General 367 (2009)
3.
26] A. Wolfson, C. Dlugy, Y. Shotland, D. Tavor, Tetrahedron Letters 50 (2009)
951.
27] R.A.W. Johnstone, A.H. Wilby, I.D. Entwistle, Chemical Reviews 85 (1985)
29.
28] E. D ˇı Hont, S. Van de Vyver, B.F. Sels, P.A. Jacobs, Chemical Communications
2008) 6011.
29] D. Roy, B. Subramaniam, R.V. Chaudhari, Catalysis Today 156 (2010)
1.
9
5
1
(
[44] M. Korac, Z. Andjic, M. Tasic, Z. Kamberovic, Journal of the Serbian Chemical
Society 72 (2007) 1115.
3