Densification of biorefinery schemes by H-transfer with Raney Ni and 2-propanol: A case study of a potential avenue for valorization of alkyl levulinates to alkyl γ-hydroxypentanoates and γ-valerolactone

Article abstract of DOI:10.1016/j.molcata.2013.11.031

Alkyl γ-hydroxypentanoates and γ-valerolactone are promising platform chemicals that can be produced from alkyl levulinates in the lignocellulosic biorefinery. Accordingly, this report aims to provide in-depth insight into the molecular aspects involved in the conversion of alkyl levulinates by H-transfer catalyzed by Raney Ni and using 2-propanol as an H-donor and solvent. We demonstrate this methodology as a highly flexible approach in regard to the high degree of control over the product selectivity. In fact, up to 90% yield of alkyl γ-hydroxypentanoates is obtained at temperatures as low as 298 K. In turn, 94% yield of γ-valerolactone is achieved at 393 K. In order to shed light on the fundamental aspects of this chemical route, we address: (1) the energetics of the transfer vs. conventional hydrogenation of methyl levulinate, (2) the thermal stability of methyl γ-hydroxypentanoate in the absence and in the presence of solid catalysts, and (3) the stability of Raney Ni in the conversion of several alkyl levulinates. Lastly, a process concept based on the current results is also proposed. This concept provides a comprehensive overview of the practical possibilities of this process as part of the lignocellulose-based biorefineries.

Full text of DOI:10.1016/j.molcata.2013.11.031

Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Densification of biorefinery schemes by H-transfer with Raney Ni and
2-propanol: A case study of a potential avenue for valorization of alkyl
levulinates to alkyl ␥-hydroxypentanoates and ␥-valerolactone
Jan Geboers, Xingyu Wang, Alex Bruno de Carvalho, Roberto Rinaldi∗
Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1, D-45470 Mülheim, Ruhr, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Alkyl ␥-hydroxypentanoates and ␥-valerolactone are promising platform chemicals that can be produced
from alkyl levulinates in the lignocellulosic biorefinery. Accordingly, this report aims to provide in-depth
insight into the molecular aspects involved in the conversion of alkyl levulinates by H-transfer catalyzed
by Raney Ni and using 2-propanol as an H-donor and solvent. We demonstrate this methodology as a
highly flexible approach in regard to the high degree of control over the product selectivity. In fact, up
to 90% yield of alkyl ␥-hydroxypentanoates is obtained at temperatures as low as 298 K. In turn, 94%
yield of ␥-valerolactone is achieved at 393 K. In order to shed light on the fundamental aspects of this
chemical route, we address: (1) the energetics of the transfer vs. conventional hydrogenation of methyl
levulinate, (2) the thermal stability of methyl ␥-hydroxypentanoate in the absence and in the presence
of solid catalysts, and (3) the stability of Raney Ni in the conversion of several alkyl levulinates. Lastly,
a process concept based on the current results is also proposed. This concept provides a comprehensive
overview of the practical possibilities of this process as part of the lignocellulose-based biorefineries.
© 2013 Elsevier B.V. All rights reserved.
Received 30 July 2013
Received in revised form
25 November 2013
Accepted 27 November 2013
Available online 7 December 2013
Keywords:
␥-Valerolactone
Alkyl ␥-hydroxypentanoates
Raney Ni
Transfer hydrogenation
One-pot synthesis
1. Introduction
The conventional approach for the production of ␥-
valerolactone comprises the catalytic hydrogenation of
Mounting concerns over diminishing fossil fuel reserves and cli-
mate change have driven academic and industrial research into
new methodologies for conversion of biomass to renewable chem-
icals [1]. Accordingly, the selective defunctionalization of biomass,
yielding platform chemicals, is of great interest [2]. In this con-
text, the conversion of cellulose and hemicellulose into levulinic
acid by acid and metal catalyzed reactions, and its subsequent
transformation into ␥-valerolactone by catalytic hydrogenation,
constitutes an important process chain in lignocellulosic biorefin-
ery [3,4]. In this regard, ␥-valerolactone holds great promise as
a platform chemical with a large number of prospective applica-
tions (e.g. green solvent, polymers, fine chemicals, fuel additive or
fuel precursor) [4–10]. Alternatively, alkyl ␥-hydroxypentanoates
are particularly desirable as monomers for the synthesis of renew-
able polyesters [26]. Nonetheless, the efficient production of alkyl
␥-hydroxypentanoate is still not possible by reported chemical
approaches, because these intermediates easily undergo lactoniza-
tion, forming ␥-valerolactone, under the severe process conditions.
levulinic acid followed by instantaneous lactonization of the
␥-hydroxypentanoic acid to ␥-valerolactone (Scheme 1) [10].
Several supported noble metal catalysts have been explored in
the production of ␥-valerolactone from levulinic acid by hydro-
genation and lactonization [6,8,10–12]. Typically, levulinic acid is
subjected to H2 pressures of 12–50 bar at temperatures between
403 and 443 K. In these studies, Ru/C is consistently found to exhibit
the highest activity and selectivity, enabling ␥-valerolactone yields
higher than 90% [6,8,10–12]. Another strategy receiving increasing
attention is based on the use of formic acid, which is co-produced
with levulinic acid, as a hydrogen source for the hydrogenation step
[13–17].
The aforementioned methodologies face, however, major prob-
lems regarding catalyst stability [10]. Moreover, the need for noble
metal catalysts and the relatively severe process conditions also
pose problems for the large-scale production of ␥-valerolactone at
reasonable prices [10].
In an effort to improve the practicality of the process, the
recycling of sulfuric acid was recently proposed in production
of levulinic acid from cellulose. Phase separation is induced by
converting the levulinic acid stream into water-insoluble alkyl
levulinates. As sulfuric acid remains in the aqueous phase, its
recycling is thus facilitated [16,18]. In addition, and perhaps more
∗
Corresponding author. Tel.: +49 208 306 2344; fax: +49 208 306 2995.
E-mail addresses: rinaldi@mpi-muelheim.mpg.de, rinaldi@kofo.mpg.de
(R. Rinaldi).
1381-1169/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcata.2013.11.031

J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
107
2.2. Synthesis of methyl ꢀ-hydroxypentanoate
Since methyl ␥-hydroxypentanoate is not readily commercially
available, we developed the following procedure for its synthesis
at high purity grade. By providing this procedure, we do not
mean that this synthesis methodology is suitable for industrial
production of methyl ␥-hydroxypentanoate. In addition to the use
of methyl ␥-hydroxypentanoate as a GC standard, this compound
was used in order to investigate the thermal stability of methyl
␥-hydroxypentanoate and elucidate the bifunctional nature of
Raney Ni.
Methyl levulinate (10 mL) was subjected to an initial pressure
of 50 bar H2 (r.t.), in the presence of the Ru/C catalyst, in a stainless
steel autoclave at 353 K for 4 h. This procedure resulted in full con-
version of methyl levulinate into 80% methyl ␥-hydroxypentanoate
and 20% ␥-valerolactone. After separation of the Ru/C catalyst
by filtration and subsequent removal of methanol by distillation
under reduced pressure, methyl ␥-hydroxypentanoate was sep-
arated from the reaction mixture by flash chromatography on a
Biotage Isolera One device using a 50 g silica column. The elu-
tion was performed by a linear gradient starting with 2% MTBE
in dichloromethane for 5 min and increasing the concentration
of MTBE to 100% in 1 min, after which the flow was maintained
for an additional 5 min (flow rate: 50 mL/min). After removal of
the solvent, an isolated yield of 40% was obtained. 1H NMR spec-
trum (300 MHz, CDCl3): ı 3.87–3.76 (m, 1H); 3.16 (s, 3H), 2.43 (t,
J = 7.3 Hz, 2H), 1.85–1.65 (m, 2H), 1.19 (d, J = 5.2 Hz, 3H); 13C NMR
spectrum (75 MHz, CDCl3): ı 174.49, 67.23, 51.60, 33.83, 30.45,
23.45.
Scheme 1. Production of ␥-valerolactone from levulinic acid by hydrogenation and
subsequent lactonization of the ␥-hydroxypentanoic acid.
importantly, the direct production of levulinates from carbo-
hydrates is proposed to be more efficient than production of
levulinic acid itself [19]. The conversion of alkyl levulinates
also sets forth new horizons for the utilization of acid-sensitive
catalysts.
Raney Ni has been demonstrated to be an efficient cat-
alyst for the transfer hydrogenation of aliphatic ketones to
their corresponding alcohols in the presence of 2-propanol
(2-PrOH) under reflux conditions [20]. Moreover, transfer hydro-
genation and transfer hydrogenolysis were proven an effective
methodology for hydrodeoxygenation of bio-based phenols, bio-
oil and several other biogenic molecules in the presence of
Raney Ni and 2-PrOH [21]. More recently, freshly prepared
Raney Ni was reportedly shown to catalyze transfer hydro-
genation of ethyl levulinate into ␥-valerolactone at 298 K using
2-PrOH as an H-donor and solvent under ambient pressure
[22].
In this contribution, we examine Raney Ni-catalyzed conver-
sion of alkyl levulinates by H-transfer reactions, and provide a
critical assessment of this emerging route. This paper is organized
as follows. After a description of the experimental methods, we
address the energetics of the transfer and conventional hydro-
genation of alkyl levulinates into alkyl ␥-hydroxypentanoates and
␥-valerolactone. Next, we discuss the results obtained by the cat-
alytic transfer hydrogenation of alkyl levulinates. Thereafter, we
examine the thermal stability of methyl ␥-hydroxypentanoate in
either the absence or presence of several solid catalysts. In addi-
tion, we study the influence of co-catalysts in the production of
␥-valerolactone. In sequence, we study the stability of Raney Ni in
the H-transfer conversion of different alkyl levulinates. Finally, we
propose a process scheme based on the H-transfer conversion of
alkyl levulinates.
2.3. Synthesis of 2-propyl levulinate
As 2-propyl levulinate is not commercially available, we estab-
lished this procedure for its synthesis at high purity grade for
the purposes of this research work. In a two-neck round bottom
flask (250 mL) equipped with a reflux condenser, levulinic acid
was slowly added under stirring to a solution containing H2SO4
(0.34 g) and 2-PrOH (108 g) at 348 K. The reaction was performed
overnight. After removal of 2-PrOH by distillation under reduced
pressure, the product mixture was dissolved in MTBE to a total
volume of 200 mL and extracted three times with 200 mL of a sat-
urated KHCO3 solution and three times with 100 mL of distilled
water. The solution was dried with anhydrous Na2SO4 and then
passed through an alumina column. After removal of MTBE under
reduced pressure, an isolated yield of 2-propyl levulinate of 53%
was obtained. The absence of residual acid was confirmed by titra-
tion with a 0.01 mol/L NaOH using an automated titrator (Metrohm
848 Titrino Plus). 1H NMR spectrum (300 MHz, CDCl3): ı 4.98 (sep,
J = 6.3 Hz, 1H), 2.72 (t, J = 6.7 Hz, 2H), 2.52 (dt, J = 0.4, 6.7 Hz, 2H), 2.17
(s, 3H), 1.21 (d, J = 6.3 Hz, 6H); 13C NMR spectrum (75 MHz, CDCl3):
ı 206.59; 172.17, 67.92, 37.98, 29.80, 28.37, 26.94, 21.73.
2. Experimental
2.1. Chemicals
Methyl tert-butyl ether was purchased from Biesterfeld and
distilled again (≥99.9%). Sulfuric acid (95–97%) was purchased
from J.T. Baker and used as received. ␥-Al2O3 (Puralox NWa 155)
was purchased from Sasol and calcined for 5 h at 713 K prior
to use. 2-Propanol (99.9%), dichloromethane (≥99.8%), n-heptane
(99%), 2-methyltetrahydrofuran (≥99%, inhibitor-free), di-n-butyl
ether (≥99%), levulinic acid (98%), methyl levulinate (≥98%), ethyl
levulinate (99%), butyl levulinate (98%), ␥-valerolactone (99%),
nickel nitrate (≥98%), activated charcoal (Norit CA1), calcium
oxide (≥96%), hydrotalcite (Mg6Al2(CO3)(OH)16·2H2O, synthetic),
Raney Ni 2800 (aqueous slurry, unpromoted, produced by W.R.
Grace & Co.), 5 wt% Ru/C and Amberlyst 15 were purchased from
Sigma–Aldrich and used as received. In regard to the freshness
of Raney Ni, the supplier guarantees the initial catalytic activ-
ity for at least three months after production. Nevertheless, great
care was taken to ensure that the results of all experiments were
reproducible and not influenced by the catalyst aging. All exper-
iments were performed within a few days and the catalyst age
did not exceed two months. In addition, several experiments were
repeated with a fresh batch of catalyst and compared to earlier data.
These routines verified the reproducibility of the current dataset.
2.4. Preparation of a Ni/C catalyst
Activated charcoal was dried at 393 K under reduced pressure
(7 mbar) for 48 h. The material was loaded with 45 wt% Ni by incip-
ient wetness impregnation with a 7.7 mol/L Ni(NO3)2 solution and
dried at room temperature in open air. Next, the material was
thermally treated under an argon flow (80 mL/min), whereby the
temperature was increased by 2 K/min from 298 to 673 K. The sam-
ple was maintained for 2 h at 673 K, after which the gas flow was
switched to H2 and the reduction performed at 673 K for 1 h. The
sample was left to cool down to room temperature under H2 flow.
Prior to further manipulation of the activated catalyst, the cell was
flushed with argon and introduced into a glove box, where the acti-
vated Ni/C catalyst sample was stored in a closed vial and handled
before the catalytic tests.

108
J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
2.5. Transfer hydrogenation experiments
To verify whether the G2(MP2) method provides a good
estimate of the thermodynamic properties, the computed and
experimental values found for the hydrogenation of acetone (Eq.
(1)) were compared. The predicted values of ꢁrH and ꢁrG are in
very good agreement with the experimental data, that is, under the
desired limit of 8.4 kJ/mol (2 kcal/mol) [25].
Typically, Raney Ni (0.5 g wet, corresponding to 0.3 g of dry cat-
alyst) was placed into a glass vial (20 mL) and washed three times
with 2-PrOH (10 mL each) to remove water. The procedure was per-
formed under argon. In turn, in experiments with Ni/C, the catalyst
(0.3 g) was first suspended in 2-PrOH (7 mL) in a glove box, and then
the vial with the suspension was handled outside the glove box, but
still under an argon atmosphere. To the suspension of the catalyst,
the substrate (1.4 mmol) and di-n-butyl ether (0.76 mmol, internal
standard for GC) were added under argon. Next, a magnetic bar
was placed in the vial. The vial was tightly closed with a septum-
equipped cap and heated under stirring in a heating block at the
indicated temperatures. The samples were taken with a syringe
at the indicated reaction times, and analyzed by GC–MS (identi-
fication) and GC-FID (quantification). The results are reported in
mol%.
␪
␪
The comparison of the reaction energetics (Eqs. (2) and (3))
shows the transfer hydrogenation of methyl levulinate as 50 kJ/mol
less exothermic than the conventional hydrogenation with H2. This
is an important feature because the consecutive lactonization of
methyl ␥-hydroxypentanoate forming ␥-valerolactone (Eq. (4)) is
predicted to be highly endothermic (26.6 kJ/mol) and thus driven
␪
by entropy (ꢁrG = −14.0 kJ/mol). The implication of this finding is
also demonstrated by the energetics of Eqs. (5) and (6). While the
overall conversion of methyl levulinate to ␥-valerolactone by H-
transfer is endothermic (17.5 kJ/mol), the same reaction performed
under H2 pressure is very exothermic (−33.1 kJ/mol).
In concentrated solutions, the reaction energetics of the
conventional hydrogenation of methyl levulinate should thus
favors the formation of ␥-valerolactone. Conversely, the low
exothermicity of the transfer hydrogenation of methyl levulinate
(−9.1 kJ/mol) should favor, in turn, the formation of methyl ␥-
hydroxypentanoate. Unquestionably, the parameters associated
with the reaction kinetics (e.g., reaction temperature, reaction time
and feed concentration) are also essential parameters in controlling
the selectivity to methyl ␥-hydroxypentanoate or ␥-valerolactone.
In addition to the high exothermicity of the conventional
hydrogenation of ketones, the severe reaction conditions, typi-
cally applied to the hydrogenation of levulinic acid or its alkyl
esters, highly favor the selective formation of ␥-valerolactone.
This is probably a reason why often only low to no yield of the
intermediate ␥-hydroxypentanoic acid is reported [8,10,11,13–17].
However, since transfer hydrogenation of ketones is a very favor-
able reaction even at temperatures as low as 298 K, the low
exothermicity of this reaction may enable the selective formation
of alkyl ␥-hydroxypentanoates at low temperatures. This possi-
bility improves the prospect of the current methodology because
alkyl ␥-hydroxypentanoates are well sought-after renewable start-
ing materials for polyester synthesis [26], but their selective
production is currently hindered by lactonization that leads to ␥-
valerolactone instead.
2.6. Recyclability of Raney Ni
Catalyst stability was assessed in experiments with different
alkyl levulinates at 393 K for a reaction time needed to reach
50–60% yield of ␥-valerolactone. Next, the glass vial was placed
into an ice bath and an aliquot of the solution was taken for GC
analysis. In sequence, the catalyst was washed seven times with
2-PrOH (10 mL each) prior to the next experiment.
2.7. GC analysis
Quantification of the products was performed on a gas chro-
matograph (Shimadzu GC-2010 Plus) equipped with a SupelcoWax
10 column (30 m). The temperature program started with an
isothermal step of 2.5 min at 323 K. Next, the temperature was
increased by 50 K/min from 323 K to another isothermal step of
2 min at 493 K. The components were quantified using calibration
curves. Compound identification was performed on GC–MS (Shi-
madzu QP2010 Plus gas chromatograph equipped with a GC-MQ
QP2010 Ultra mass spectrometer) using an Rxi-1ms column (30 m).
The temperature program started with an isothermal step of 6 min
at 313 K. Next, the temperature was increased by 30 K/min from
313 K to another isothermal step of 1 min at 433 K. Again, the tem-
perature was increased by 30 K/min from 313 K to 473 K. From this
point on, the temperature was increased by 10 K/min, until reach-
ing the last isothermal step of 1 min at 513 K. Compounds were
identified by comparing the EI-MS spectrum with the MS libraries
NIST 08, NIST 08s and Wiley 9.
3.2. Catalytic transfer hydrogenation of alkyl levulinates
The thermodynamic predictions suggest that the H-transfer
conversion of methyl levulinate should preferentially form methyl
␥-hydroxypentanoate when the reaction is performed at low tem-
peratures. The thermodynamic predictions starkly contrast the
recently published results by Fu et al. [22] that show the forma-
tion of ␥-valerolactone at very high yields at 298 K. In order to give
a broad analysis of the conversion of alkyl levulinates at different
temperatures, the H-transfer conversion of several alkyl deriva-
tives, namely methyl, ethyl, 2-propyl and butyl levulinate, was
examined. Experimental evidence that these reactions proceeded
through H-transfer from 2-PrOH is provided in Supporting Infor-
mation.
In regard to the substrate reactivity, methyl, ethyl and 2-propyl
levulinates are fully converted within 75 min at 298 K. Butyl lev-
ulinate, the least reactive substrate, needs 300 min to fully convert
into butyl ␥-hydroxypentanoate at room temperature. At 393 K,
however, even butyl levulinate undergoes full conversion within
5 min.
3. Results and discussion
3.1. Energetics of the transfer hydrogenation versus conventional
hydrogenation
In this section, the thermodynamic features distinguishing
the transfer hydrogenation from the conventional hydrogenation
of alkyl levulinates are addressed in detail. To the best of our
knowledge, there is no thermodynamic data available for methyl
levulinate, methyl ␥-hydroxypentanoate and ␥-valerolactone in
gas-phase so far published in the current chemical databases (e.g.
NIST Webbook, CRC Handbook of Physics and Chemistry or Lange’s
␪
Handbook of Chemistry). Consequently, the values of ꢁRH and
␪
ꢁRG were computed by the G2(MP2) method starting with the
minimized geometry (B3LYP/6-31++g(d,p)) of each component of
the chemical equations given in Table 1. The calculations were per-
formed using the software Gaussian 09 [23]. The thermodynamic
predictions for the hypothetical gas-phase reactions at 298.15 K
(1 bar) are listed in Table 1.
Fig. 1 shows the influence of reaction temperature on both
the selectivity to alkyl ␥-hydroxypentanoates and ␥-valerolactone
upon reaching conversion higher than 97%. These results are con-
sistent with the thermodynamic predictions. In our experiments,

J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
109
Table 1
Predicted thermodynamic properties of hypothetical gas-phase reactions at 298.15 K, 1 bar.
␪
␪
Equation
1
Reaction
ꢁrH (kJ/mol)
ꢁrG (kJ/mol)
−50.6 (−55.2)a
−9.1
−9.4 (−15.3)a
2
3
4
5
−10.1
−19.5
−14.0
−24.1
−33.6
−59.7
26.6
17.5
6
−33.1
a
Values between parentheses correspond to those calculated from experimental values [24].
alkyl ␥-hydroxypentanoates always formed as the major prod-
uct at 298 K. The effect of temperature on the reaction selectivity
depends strongly on the size of the alkyl residue, as shown in
Fig. 1. In regarding to a non-catalytic (thermal) lactonization of the
alkyl ␥-hydroxypentanoates taking place in solution, the reactiv-
ity order for the alkyl ␥-hydroxypentanoates would be arranged
as methyl < ethyl < 2-propyl < butyl, since a large alkyl residue is a
better leaving group than a short one for a nucleophilic substitu-
tion. Conversely, our results show the longer the alkyl residue, the
lower the selectivity to ␥-valerolactone. This observation suggests
that the lactonization take place on the surface of Raney Ni, where a
large alkyl residue may pose a major steric hindrance for a suitable
adsorption for the conversion of alkyl ␥-hydroxypentanoate into
␥-valerolactone.
Surprisingly, selectivity to butyl ␥-hydroxypentanoate is limited
to about 78%. At the same time, selectivity to ␥-valerolactone does
not exceed 7%. This result reveals a carbon deficit of about 15%.
Since no byproduct was detected by GC, the loss in carbon is most
likely caused by adsorption of the substrate or the intermediate
on the Raney Ni surface. This observation is also valid for 2-propyl
levulinate (carbon balance around 90%). The conversion of methyl
and ethyl levulinate displayed a carbon balance higher than 95%.
The implications of fouling for the catalyst performance will be
discussed in full detail in Sections 3.3 and 3.4.
As hydrogenation and lactonization are consecutive reactions,
product selectivity also depends upon the experiment duration.
In particular, extended reaction times increase the yield of ␥-
valerolactone when starting with short-alkyl levulinates. As shown
in Fig. 2(a) methyl levulinate can be converted into 94% ␥-
valerolactone in 60 min at 393 K. For longer alkyl levulinates this
strategy is less effective, as shown in Fig. 2(b).
In summary, the current results demonstrate the transfer
hydrogenation of alkyl levulinates as a flexible methodology
for the production of alkyl ␥-hydroxypentanoates, in addition
to ␥-valerolactone. Furthermore, this approach does not utilize
expensive noble metals as catalysts. Strikingly, the product selec-
tivity is effectively tuned simply through the choice of substrate as
well as reaction temperature and time.
3.3. Stability of alkyl ꢀ-hydroxypentanoates and lactonization
over Raney Ni
␥-Hydroxypentanoic acid and alkyl ␥-hydroxypentanoates
are reported as highly prone to undergo lactonization [11,17].
Fig. 1. Selectivity to (A) alkyl ␥-hydroxypentanoates and (B) ␥-valerolactone in function of reaction temperature (at full conversion of substrate). Reaction mixture: substrate
(1.4 mmol), Raney Ni (0.3 g, dry basis), 2-PrOH (7 mL).

110
J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
Fig. 2. Effect of reaction time on product selectivity. Conversion of (a) methyl levulinate and (b) 2-propyl levulinate at 393 K. Reaction mixture: substrate (1.4 mmol), Raney
Ni (0.3 g, dry basis), 2-PrOH (7 mL).
Although lactonization of methyl ␥-hydroxypentanoate is ther-
modynamically favored at 298 K (Table 1, Eq. (4)), it may still be
kinetically hindered. To address the thermal stability of methyl
␥-hydroxypentanoate, two control experiments were performed.
Methyl ␥-hydroxypentanoate in addition to a 0.2 mol/L solu-
tion thereof in 2-PrOH were heated at 353 K for 90 min. In both
experiments, we found that methyl ␥-hydroxypentanoate is very
stable, that is, less than 2% ␥-valerolactone was formed upon
thermal treatment. In addition, at room temperature methyl ␥-
hydroxypentanoate is fully stable, even for months. Thus, the
kinetic stability of methyl ␥-hydroxypentanoate implies that Raney
Ni be involved in the lactonization of alkyl ␥-hydroxypentanoate to
␥-valerolactone. Obviously, since lactonization is an acid- or base-
catalyzed reaction, this leads to a key question: what is the role of
Raney Ni in the lactonization step?
In regard to the preparation of Raney Ni, a NiAl alloy (1:1, wt%)
is leached with concentrated alkali solutions (e.g. 6 mol/L NaOH).
Hence, NaOH residues could then be involved in the catalysis of
the lactonization step. When performing the transfer hydrogena-
tion of ethyl levulinate in the presence of 0.25 mmol NaOH in
2-PrOH, we found that ethyl levulinate readily undergoes trans-
esterification forming 2-propyl levulinate. As a result, 2-propyl
␥-hydroxypentanoate is formed as the major product of the H-
transfer reaction. Overall, these observations clearly show that a
base impurity does not account for the lactonization activity of
Raney Ni.
of O(4), thus activating the alkyl ␥-hydroxypentanoate toward
lactonization.
Recently, transfer hydrogenation of ethyl levulinate in the pres-
ence of freshly prepared Raney Ni was reported to result in
quantitative conversion of ethyl levulinate into ␥-valerolactone
with very low catalyst loading at 298 K for 120 min [22]. In contrast,
in our study such a result is only achieved at higher temperatures
(>353 K). Apparently, the results by Fu et al. may be influenced
by the residual alumina content in the freshly prepared Raney Ni.
Another factor that may account for the conflicting results is per-
haps the analysis of the products by GC. Here, caution concerning
GC analysis of alkyl ␥-hydroxypentanoates must be regarded. We
have found that alkyl ␥-hydroxypentanoates are extremely sensi-
tive to lactonization in a non-passivated or slightly used GC liner.
Indeed, the lactonization may be catalyzed by trace contaminants
eventually present in the GC injection port and liner, thus resulting
in an overestimation of the ␥-valerolactone yield. Hence, great care
must be taken to ensure that both GC injection port and liner are
thoroughly clean. Otherwise, the reliability of the analysis is ques-
tionable. We regularly verified the system reliability by injecting
solutions of methyl ␥-hydroxypentanoate in 2-PrOH. The absence
of a ␥-valerolactone peak in the chromatogram indicates the sys-
tem as suitable for analysis.
Yet, the digestion of the NiAl alloy with alkali solution does
not completely remove the Al-content. Rather, an alumina phase
(e.g. bayerite and amorphous domains) is formed. We previously
reported that the surface of fresh Raney Ni shows a Ni:Al molar
ratio of 54:46 [21]. Since alumina may contain a wide range of acidic
and basic sites [35], these sites could be involved in the catalysis of
the lactonization reaction. To elucidate the role of alumina in the
lactonization of methyl ␥-hydroxypentanoate, the reaction in the
presence of ␥-Al2O3, Ni/C and Raney Ni was studied.
Fig.
3
shows the conversion profile of methyl ␥-
hydroxypentanoate into ␥-valerolactone in the presence
of ␥-Al2O3, Ni/C and Raney Ni. Surprisingly, all the afore-
mentioned materials catalyze the lactonization of methyl
␥-hydroxypentanoate, albeit with very different activities. By
catalyst weight, Raney Ni and ␥-Al2O3 are, respectively, four and
seven times more active than Ni/C. These results indicate that
lactonization should be predominantly catalyzed by the residual
alumina on the Raney Ni surface. In addition, Ni surface may be,
to a lesser extent, involved in the lactonization, as proposed in
Scheme 2. In this case, the interaction of both the C(4) H and
O(4) H with the Ni surface should increase the electron density
Fig. 3. Lactonization of methyl ␥-hydroxypentanoate in the presence of (ꢀ) ␥-Al2O3,
(᭹) Ni/C or (ꢁ) Raney Ni at 353 K. Reaction mixture: methyl ␥-hydroxypentanoate
(1.4 mmol), catalyst (0.3 g, dry basis), 2-PrOH (0.3 g).

J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
111
Scheme 2. Qualitative representation of interactions proposed for Ni-catalyzed lac-
tonization of alkyl ␥-hydroxypentanoates.
Fig. 5. Comparison of transfer hydrogenation of levulinic acid (full symbols) and
butyl levulinate (open symbols) at 353 K. (ꢀ/ꢂ) Conversion, (᭹/ꢀ) yield of ␥-
hydroxypentanoic acid, (ꢁ/ꢁ yield of ␥-valerolactone. Reaction mixture: substrate
(1.4 mmol), Raney Ni (0.3 g), 2-PrOH (7 mL).
3.4. Use of co-catalysts to promote lactonization of alkyl
ꢀ-hydroxypentanoates
The current results show that, at a given temperature and for a
given substrate, the formation of ␥-valerolactone relies upon the
acid–base functionality of an alumina phase in Raney Ni. How-
ever, the quantity of acidic/basic sites on the Raney Ni surface very
likely depends strongly upon the digestion procedure applied to
the NiAl alloy. To diminish the reliance of the selectivity to ␥-
valerolactone on the ‘undefined’ bifunctional nature of Raney Ni,
we exploited the effects of several co-catalysts (CaO, hydrotalcite,
␥-Al2O3 and Amberlyst 15 DRY) in the conversion of 2-propyl lev-
ulinate into ␥-valerolactone. 2-Propyl levulinate was selected as
a substrate to avoid the formation of a mixture of alkyl esters by
transesterification with 2-PrOH, which is catalyzed by acid or basic
catalysts.
Fig. 4 summarizes the results obtained from experiments
with concurrent use of Raney Ni and a co-catalyst in the con-
version of 2-propyl levulinate to ␥-valerolactone at 353 K. CaO
and hydrotalcite are basic catalysts, which have been successfully
applied to the transesterification of triglycerides to fatty acid
methyl esters [27]. However, they do not improve the yield of
␥-valerolactone. Surprisingly, hydrotalcite exerts an inhibitory
effect on the formation of ␥-valerolactone.
Unlike CaO and hydrotalcite, ␥-Al2O3 improves the lactoniza-
tion considerably. Despite this, we found that ␥-Al2O3 decreases
the carbon balance of the reaction by about 5%. No byproduct was
detected by GC-FID or GC–MS. Since a loss in carbon balance was
not detected in any other experiment performed with co-catalysts,
alumina seems to be especially susceptible to fouling. Actually, the
formation of several surface acetates was demonstrated in other
kind of experiments involving ethyl acetate and transitional alu-
minas [36]. This substantiates the formation of surface levulinates
or ␥-hydroxypentanoates on alumina as a plausible reason for the
decrease in the carbon balance.
The most effective co-catalyst is Amberlyst 15, a highly acidic
material [33]. The concurrent use of Raney Ni and Amberlyst 15
results in a ␥-valerolactone yield of 83% in 1 h. In agreement with
this result, Amberlyst 70 was also demonstrated as an efficient
lactonization catalyst in the hydrogenation of levulinic acid [34].
The concurrent use of Raney Ni and Amberlyst 15 at 353 K is an
effective way to improve the yield of ␥-valerolactone. Nonetheless,
in the presence of an acid catalyst, 2-PrOH is prone to undergo dehy-
dration forming propene. Since this reaction in gas-phase shows a
ꢁrG value equals to zero at 347 K under a pressure of 1 bar (ꢁrG
calculated from experimental values taken from Ref. [37]), the one-
pot reaction should preferably be performed at temperatures lower
than 347 K in order to avoid significant depletion of 2-PrOH.
3.5. Stability of the Raney Ni catalyst
Raney Ni is a promising and versatile catalyst for H-transfer con-
version of alkyl levulinates. Nonetheless, two compounds may pose
a problem to its stability: free carboxylic acids and primary alcohols.
Carboxylic acids may cause leaching of the Ni phase when present
at high concentrations in the reaction mixture. However, at low
concentrations, the formation of surface carboxylates may, in turn,
take place blocking the active sites [30]. Thereby, we chose to eval-
uate the performance of Raney Ni in the transfer hydrogenation of
levulinic acid, since this condition is the most deleterious one for
the catalyst.
Fig. 4. Effect of co-catalysts on the yield of ␥-valerolactone by transfer hydrogena-
tion of 2-propyl levulinate at 353 K. (ꢀ) No co-catalyst, (᭹) CaO, (ꢁ) hydrotalcite,
(ꢂ) ␥-Al2O3, (ꢀ) Amberlyst 15. Reaction mixture: 2-propyl levulinate (1.4 mmol,
dry basis), Raney Ni (0.3 g), co-catalyst (0.5 g), 2-PrOH (7 mL).
Fig. 5 compares the conversion of butyl levulinate, the least reac-
tive alkyl levulinate studied here, with the conversion of levulinic

112
J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
Table 2
Recycling of Raney Ni at 393 K at 2-propyl levulinate-to-2-PrOH molar ratio of 1.0.
Run
Conversion (%)
Yield (%) 3-propyl
␥-Valerolactone
␥-hydroxypentanoate
1
2
85
76
70
74
25
38
48
42
51
34
21
31
3
4a
a
Recycled catalyst washed with MTBE after the third run.
Scheme 3. Interactions of a primary (a) and secondary alcohol (b) with a Ni surface.
the substrate-to-2-PrOH molar ratio. After the first run, the catalyst
performance stabilizes at a level that depends on methyl levulinate-
to-2-PrOH molar ratio. The catalytic activity is not recovered by
washing the catalyst with MTBE or hot 2-PrOH, indicating that the
deactivating species are strongly bonded to the catalyst surface.
Secondary alcohols do not form strong alkoxy-Ni bonds
(Scheme 3). Accordingly, the use of 2-propyl levulinate could avoid
catalyst deactivation. To examine whether the use of 2-propyl lev-
ulinate is more promising in regard to preserving the catalyst from
deactivation, recycling experiments with 2-propyl levulinate in a
molar ratio of substrate-to-2-PrOH of 1.0 were performed. The
results are listed in Table 2.
The hydrogenation activity is only moderately affected but
selectivity to ␥-valerolactone decreases dramatically from 60% to
30% after the first reaction run. At the same time, the carbon balance
increases from 89% to 99%. As previously discussed, the alumina
content is responsible for lactonization activity of the Raney Ni
material. Aluminum oxides are also very sensitive to fouling, as
already discussed. Altogether, the increase in the carbon balance in
addition to the decrease in selectivity to ␥-valerolactone suggest
that the alumina phase in Raney Ni should be undergoing fouling
to a considerable extent, decreasing the lactonization activity. By
washing the used catalyst with MTBE after the third reaction run,
a partial recovery of the selectivity to ␥-valerolactone is obtained
(Table 2, fourth run).
acid in 2-PrOH at 353 K. While butyl levulinate is fully converted
within 10 min, only 35% levulinic acid is converted in 150 min. This
result agrees well with earlier reports showing the incompatibil-
ity of Raney Ni as a catalyst for the hydrogenation of levulinic acid
at temperatures below 448 K [29,30]. This observation also reveals
that the content of free carboxylic acids should be as low as possible
in the substrate stream.
A second concern for Raney Ni stability is the formation of stoi-
chiometric quantity of primary alcohols upon lactonization of alkyl
␥-hydroxypentanoates. Primary alcohols may irreversibly adsorb
onto the Ni surface in the absence of 2-PrOH [21]. Thereby, the
active sites become blocked by alkoxide species [21,28,31]. In con-
trast, the adsorption of 2-PrOH breaks the C␣ H and O H bonds
releasing of acetone. Additionally, this process loads the Ni surface
with H-atoms, which are then used in the transfer hydrogenation.
Scheme 3 depicts the processes occurring on the Raney Ni surface
upon adsorption of a primary (a) and a secondary alcohol (b).
To evaluate the effect of primary alcohols on performance of
Raney Ni throughout the recycling experiments, the conversion of
alkyl levulinates (R = methyl, ethyl and butyl) was performed. The
reaction conditions were set for each substrate in order to achieve
60% yield of ␥-valerolactone in the first reaction run. Raney Ni was
then recycled four times under these conditions. In this way, the
catalyst is exposed to the same quantity of primary alcohol in the
first reaction run, and thus the influence of the alcohols can be
properly compared.
3.6. Proposed process based on transfer hydrogenation of alkyl
levulinates
Fig. 6 compares the results obtained by the H-transfer conver-
sion of methyl, ethyl and butyl levulinate and reveals important
features. Of utmost interest, the catalyst is not deactivated by
the primary alcohol released upon lactonization. However, the
inhibitory effect of primary alcohols, when performing H-transfer
reactions in 2-PrOH [21], is clearly substantiated by the results
showed in Fig. 6. Both methanol and ethanol affect the forma-
tion of ␥-valerolactone after the first run, albeit the formation of
␥-valerolactone is affected, to the greatest extent, by the release
of methanol in the experiment performed with methyl levuli-
nate (Fig. 6a and b). A modest recovery of the selectivity to
␥-valerolactone is achieved, revealing that the inhibitory effect
of methanol or ethanol may be partially alleviated by perform-
ing the reaction using 2-PrOH as an H-donor and solvent (note
that the substrate-to-2-PrOH molar ratio is 0.01 in these exper-
iments). Remarkably, the formation of ␥-valerolactone is not
affected throughout the recycling experiments with butyl levuli-
nate (Fig. 6c).
To evaluate Raney Ni stability in a more concentrated feed, the
recycling experiments were performed with molar ratios of methyl
levulinate-to-2-PrOH of 0.1 and 1.0. Fig. 7 shows four recycling runs
under these conditions.
The inhibitory effect of methanol becomes more evident at
ratios of methyl levulinate-to-2-PrOH higher than 0.1. In each
series, a decrease in the conversion and yield of ␥-valerolactone
occurs after the first run. This decrease is related to the increase in
In order to provide an overview of the possibilities of the H-
transfer conversion of alkyl levulinates, a concept for a continuous
process based on H-transfer hydrogenation is proposed in Fig. 8.
An alkyl levulinate feed can be produced directly from biomass
or by esterification of biomass-derived levulinic acid with a sec-
ondary alcohol. This feed is then mixed with 2-PrOH and passed
through a bed of Raney Ni at temperatures lower than 338 K. The
concurrent use of 2-propyl levulinate and 2-PrOH seems to be par-
ticularly attractive for several reasons. First, it allows the use of
feeds rich in substrate (e.g., substrate-to-2-PrOH ratio of 1.0). Sec-
ond, the use of 2-propyl levulinate and 2-PrOH does not generate
a lower alcohol mixture in the product stream. Third, the feed of
2-propyl levulinate and 2-PrOH should enable a stable activity of
the catalyst in hydrogenation of 2-propyl levulinate to 2-propyl
␥-hydroxypentanoate. Forth, since 2-PrOH has a low boiling point
(353 K), the cost associated with the product isolation by distillation
of 2-PrOH and acetone is minimized.
Most importantly, the transfer hydrogenation of 2-propyl
levulinate performed at room temperature enables the pro-
duction of 2-propyl ␥-hydroxypentanoate selectively. Since the
lactonization activity of Raney Ni will decrease due to foul-
ing upon extended use of the catalyst (Table 2), an additional
reactor with Amberlyst 15 should be necessary to maintain a
high productivity to ␥-valerolactone. However, should alkyl
␥-hydroxypentanoates be the desired product, only one reactor
with Raney Ni is necessary in the overall process scheme. Hence,

J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
113
Fig. 6. Performance of Raney Ni in recycling experiments performed to achieve 60% yield of ␥-valerolactone (first run) at 393 K. (A) Methyl levulinate, (B) ethyl levulinate,
and (C) butyl levulinate. Black columns: conversion; gray columns: yield of ␥-valerolactone. Reaction mixture: 1.4 mmol levulinate, 0.6 g Raney Ni, 7 mL 2-PrOH.
the mixture leaving the hydrogenation reactor should contain
2-propyl ␥-hydroxypentanoate, 2-PrOH and acetone.
This product holds promise as a renewable monomer for the syn-
thesis of high value-added polyesters. Alternatively, production
excess could be converted into ␥-valerolactone in the presence of
Amberlyst 15 (second reactor). This versatility should contribute to
the densification of biorefining schemes, and thus to an increase in
the profitability of these process chains.
Considering that alkyl ␥-hydroxypentanoates are stable
compounds, as demonstrated here, a stream of 2-propyl ␥-
hydroxypentanoate could be produced by stripping 2-PrOH and
acetone from the product mixture by distillation under reduced
pressures at temperatures below 353 K. As a result, 2-propyl
␥-hydroxypentanoate at a high purity grade is easily produced.
Finally, 2-PrOH is regenerated by gas-phase hydrogenation of
acetone (e.g., with Raney Ni) [32]. The high exothermicity of acetone
Fig. 7. Recycling of Raney Ni at 393 K with a methyl levulinate-to-2-PrOH molar ratio of (A) 0.1 or (B) 1.0. Black columns: conversion; gray columns: yield of ␥-valerolactone.

114
J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
Acknowledgements
Roberto Rinaldi is grateful to the Alexander von Humboldt
Foundation for the funds provided in the framework of the
Sofja Kovalevskaja Award 2010, endowed by the German Federal
Ministry of Education and Research. Jan Geboers is grateful for a
research fellowship by Alexander von Humboldt Foundation. Alex
Carvalho is thankful to CNPq for the scholarship granted within the
program “Ciência sem Fronteiras” supported by the Brazilian Fed-
eral Government. This work was performed as part of the Cluster
of Excellence “Tailor-Made Fuels from Biomass”, which is funded
by the Excellence Initiative by the German Federal and State Gov-
ernments to promote science and research at German universities.
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.molcata.2013.11.031.
References
[1] A.J. Ragauskas, C.K. Williams, B.H. Davison, G. Britovsek, J. Cairney, C.A. Eckert,
W.J. Frederick, J.P. Hallett, D.J. Leak, C.L. Liotta, J.R. Mielenz, R. Murphy, R. Tem-
pler, T. Tschaplinski, Science 311 (2006) 484–489;
P. Gallezot, Chem. Soc. Rev. 41 (2012) 1538–1558.
[2] D.M. Alonso, J.Q. Bond, J.A. Dumesic, Green Chem. 12 (2010) 1493–1513.
[3] J.J. Bozell, L. Moens, D.C. Elliott, Y. Wang, G.G. Neuenscwander, S.W. Fitzpatrick,
R.J. Bilski, J.L. Jarnefeld, Resourc. Converv. Recycl. 28 (2000) 227–239;
M.J. Climent, A. Corma, S. Iborra, Green Chem. 13 (2011) 520–540.
[4] J.-P. Lange, R. Price, P.M. Ayoub, J. Louis, L. Petrus, L. Clarke, H. Gosselink, Angew.
Chem. Int. Ed. 49 (2010) 4479–4483.
[5] I.T. Horváth, H. Mehdi, V. Fabos, R. Tuba, A. Bodor, L.T. Mika, Green Chem. 10
(2008) 238–242;
R. Palkovits, Angew. Chem. Int. Ed. 49 (2010) 4336–4338;
J.A. Geboers, S. Van de Vyver, R. Ooms, B. Op de Beeck, P.A. Jacobs, B.F. Sels,
Catal. Sci. Technol. 1 (2011) 38–50.
[6] D. Fegyverneki, L. Orha, G. Làng, I.T. Horváth, Tetrahedron 66 (2010) 1078–1081.
[7] J.C. Serrano-Ruiz, D.J. Braden, R.M. West, J.A. Dumesic, Appl. Catal. B: Environ.
100 (2010) 184–189;
J.Q. Bond, D.M. Alonso, D. Wang, R.M. West, J.A. Dumesic, Science 327 (2010)
1110–1114;
D.M. Alonso, J.Q. Bond, J.C. Serrano-Ruiz, J.A. Dumesic, Green Chem. 12 (2010)
992–999.
[8] L.E. Manzer, Appl. Catal. A: Gen. 272 (2004) 249–256.
[9] I. van der Meulen, E. Gubbels, S. Huijser, R. Sablong, C.E. Koning, A. Heise, R.
Duchateau, Macromolecules 44 (2011) 4301–4305.
Fig. 8. Concept for catalytic transfer hydrogenation of levulinate feeds.
[10] W.R.H. Wright, R. Palkovits, ChemSusChem 5 (2012) 1657–1667;
D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Green Chem. 15 (2013) 594–595.
[11] Z.-P. Yan, L. Lin, S. Liu, Energy Fuels 23 (2009) 3853–3858.
[12] M.G. Al-Shaal, W.R.H. Wright, R. Palkovits, Green Chem. 14 (2012) 1260–1263.
[13] L. Deng, Y. Zhao, J. Li, Y. Fu, B. Liao, ChemSusChem 3 (2010) 1172–1175.
[14] X.-L. Du, Q.-Y. Bi, Y.-M. Liu, Y. Cao, K.-N. Fan, Angew. Chem. Int. Ed. 50 (2011)
7815–7819.
hydrogenation (55 kJ/mol) should reduce the energy costs associ-
ated with regeneration of 2-PrOH. For instance, the hydrogenation
heat can be used for preheating the 2-PrOH/acetone feed or in the
reclamation of 2-PrOH by distillation. With proper process integra-
tion, one can also envision that major fraction of the heat required
for the transfer hydrogenation of 2-propyl levulinate can be gener-
ated by the gas-phase hydrogenation of acetone.
[15] D.J. Braden, C.A. Henao, J. Heltzel, C.C. Maravelias, J.A. Dumesic, Green Chem.
13 (2011) 1755–1765.
[16] X.-L. Du, Q.-Y. Bi, Y.-M. Liu, Y. Cao, K.-N. Fan, ChemSusChem
1838–1843.
4 (2011)
[17] A.M. Hengne, C.V. Rode, Green Chem. 14 (2012) 1064–1072;
M. Chia, J.A. Dumesic, Chem. Commun. 47 (2011) 12233–12235.
[18] E.I. Gürbüz, D.M. Alonso, J.Q. Bond, J.A. Dumesic, ChemSusChem 4 (2011)
357–361.
4. Conclusions
The high catalytic activity of Raney Ni in transfer hydrogenation
of ketones was exploited in the valorization of alkyl levulinates.
By choosing a transfer hydrogenation route with low exother-
micity instead of the highly exothermic hydrogenation using H2,
the hydrogenation and lactonization steps involved in the conver-
sion of alkyl levulinates to ␥-valerolactone can be decoupled. This
unique feature enables the control over product selectivity by sim-
ple choice of reaction parameters, such as reaction temperature,
reaction time, and substrate. However, this contribution should be
taken as a fundamental development. We believe that there are yet
many avenues to explore for further improvement of the current
results and product selectivity toward alkyl ␥-hydroxypentanoates
or ␥-valerolactone.
[19] S. Saravanamurugan, A. Riisager, Catal. Commun. 17 (2012) 71–75;
S. Saravanamurugan, O. Nguyen Van Buu, A. Riisager, ChemSusChem 4 (2011)
723–726.
[20] R.C. Mebane, K.L. Holte, B.H. Gross, Synth. Commun. 37 (2007) 2787–2791.
[21] X. Wang, R. Rinaldi, Energy Environ. Sci. 5 (2012) 8244–8260.
[22] Z. Yang, Y.-B. Huang, Q.-X. Guo, Y. Fu, Chem. Commun. 49 (2013) 5328–5330.
[23] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman,
G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato,
X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada,
M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, M.
Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Nor-
mand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N.
Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W.
Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador,
J.J. Dannenberg, S. Dapprich, A.D. Daniels, Ö. Farkas, J.B. Foresman, J.V. Ortiz, J.

J. Geboers et al. / Journal of Molecular Catalysis A: Chemical 388–389 (2014) 106–115
Cioslowski, D.J. Fox, Gaussian 09, Revision A. 1, Gaussian, Inc., Wallingford, CT, [31] Y.-H. Zhou, P.-H. Lv, G.-C. Wang, J. Mol. Catal. A 258 (2006) 203–215;
115
2009.
D. Al-Mawlawi, J.M. Saleh, J. Chem. Soc. Faraday Trans.
2965–2976;
1 77 (1981)
[24] E. Buckley, E.F.G. Herington, Trans. Faraday. Soc. 61 (1965) 1618–1625.
[25] J.B. Foresman, A. Frisch, Exploring Chemistry with Electronic Structure Meth-
ods, 2nd edition, Gaussian Inc., Pittsburgh, 1996, pp. 154.
[26] V. Gorenflo, G. Schmack, R. Vogel, A. Steinbüchel, Biomacromolecules 2 (2001)
45–57.
D. Al-Mawlawi, J.M. Saleh, J. Chem. Soc. Faraday Trans. 1 77 (1981) 2977–2988.
[32] K. Morizane, T. Shirahata, K. Higashi, S. Senoo, K. Doi, 2009, WO 200910459.;
F. Sun, D. Meng, B. Zhang, C. Mao, 2012, CN 102603473.
[33] F.H. Richter, K. Pupovac, R. Palkovits, F. Schüth, ACS Catal. 3 (2013) 123–127.
[34] A.M.R. Galletti, C. Antonetti, V. De Luise, M. Martinelli, Green Chem. 14 (2012)
688–694.
[27] Z. Helwani, M.R. Othman, N. Aziz, J. Kim, W.J.M. Fernano, Appl. Catal. A: Gen.
363 (2009) 1–10;
E.M. Shahid, Y. Jamal, Renew. Sustain. Energy Rev. 15 (2011) 4732–4745.
[28] X. Wang, R. Rinaldi, ChemSusChem 5 (2012) 1455–1466.
[29] L.P. Kyrides, J.K. Craver, Monsanto Chemical Co., US 2,368,366, 1945;
R.V. Christian, H.D. Brown, R.M. Hixon, J. Am. Chem. Soc. 69 (1947) 1961–1963.
[30] P. Gallezot, P.J. Cerino, B. Blanc, G. Flèche, P. Fuertes, J. Catal. 146 (1994) 93–102;
W.M.H. Sachtler, J. Fahrenfort, Proceedings of the 2nd International Congress
on Catalysis, 1960, Technip, Paris, 1961, p. 831.
[35] R. Rinaldi, F.Y. Fujiwara, W. Hölderich, U. Schuchardt, J. Catal. 244 (2006)
92–101.
[36] R. Rinaldi, F.Y. Fujiwara, U. Schuchardt, J. Catal. 245 (2007) 456–465.
[37] J. Speight, Lange’s Handbook of Chemistry, 16th edition, McGraw-Hill, New
York, 2005.