Highly selective production of value-added γ-valerolactone from biomass-derived levulinic acid using the robust Pd nanoparticles

Article abstract of DOI:10.1016/j.apcata.2013.08.037

A series of Pd nanoparticles deposited on the SiO₂ support were facilely and successfully synthesized in the presence of the green solvent CO₂, where the uniform distribution of Pd with small particle size was successfully achieved. The resulting Pd/SiO₂ nanoparticles catalysts exhibited excellent catalytic performances in the selective hydrogenation of biomass-derived levulinic acid, showing close to perfect selectivity of biofuel γ-valerolactone with the TON of 884.7 at 97.3% conversion of levulinic acid. The catalytic performance was superior to the activities of the 5 wt% Pd/SiO₂ nanoparticle catalyst prepared by the traditional impregnation method. Besides, the reaction parameters (e.g., the Pd loading, reaction time, reaction temperature, and hydrogen pressure), catalyst stability and reaction mechanism on the hydrogenation performance were studied. The resulting Pd nanoparticles catalysts behaved high stability in the hydrogenation.

Full text of DOI:10.1016/j.apcata.2013.08.037

Applied Catalysis A: General 468 (2013) 52–58
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Highly selective production of value-added ␥-valerolactone from
biomass-derived levulinic acid using the robust Pd nanoparticles
Kai Yana,∗, Todd Lafleura, Guosheng Wua, Jiayou Liaob, Chen Cenga, Xianmei Xiec
a Department of Chemistry, Lakehead University, 955 Oliver Road, Thunder Bay, ON P7B 5E1, Canada
b School of Chemical Engineering and Technology, Tianjin University, 92 Weijin Road, Tianjin 300072, China
c College of Chemical Engineering, Taiyuan University of Technology, 79 Yingze Street, Taiyuan 030024, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 May 2013
Received in revised form 14 August 2013
Accepted 19 August 2013
Available online 28 August 2013
A series of Pd nanoparticles deposited on the SiO2 support were facilely and successfully synthesized
in the presence of the green solvent CO2, where the uniform distribution of Pd with small particle size
was successfully achieved. The resulting Pd/SiO2 nanoparticles catalysts exhibited excellent catalytic
performances in the selective hydrogenation of biomass-derived levulinic acid, showing close to perfect
selectivity of biofuel ␥-valerolactone with the TON of 884.7 at 97.3% conversion of levulinic acid. The
catalytic performance was superior to the activities of the 5 wt% Pd/SiO2 nanoparticle catalyst prepared
by the traditional impregnation method. Besides, the reaction parameters (e.g., the Pd loading, reaction
time, reaction temperature, and hydrogen pressure), catalyst stability and reaction mechanism on the
hydrogenation performance were studied. The resulting Pd nanoparticles catalysts behaved high stability
in the hydrogenation.
Keywords:
Synthesis
Pd Nanoparticle
Hydrogenation
Levulinic acid
␥-Valerolactone
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
hydrolyze under neutral conditions, making it a safe material for
large scale use. A comparative evaluation performed on a mixture
Currently, environmental protection and efficient use of natural
resources are the sustainable alternatives to transform our society
into a greener future [1–3]. The utilization of biomass or biomass-
derived products is becoming vitally crucial, because it is aimed not
only at the development of effective and environmentally benign
technology, but also simultaneously solves the problem of agri-
cultural and forestry waste use [4,5]. In most of the approaches
being investigated, platform chemical levulinic acid (LA) has often
been hydrogenated to produce value-added chemicals and biofu-
els [6–8], which is produced from the renewable cellulose [8,9].
A family of valuable chemicals (e.g., ␥-valerolactone (GVL), pen-
tanoic acid) and attractive biofuels (e.g., 2-methyltetrahydrofuran
(2MTHF) and GVL) can be produced from LA as shown in Scheme 1
[10–12]. It was recently proposed that GVL, a frequently used food
additive, exhibits the most important characteristics of an ideal
liquid for the sustainable production of energy and carbon-based
products [13]. It is renewable, has low melting point of −31 ◦C,
high boiling point of 207 ◦C and flash point of 96 ◦C, a defini-
tive but acceptable smell for easy recognition of leaks and spills,
low toxicity, and high solubility in water to assist biodegradation
[13]. Besides, Horvath et al. [13,14] have shown that GVL does not
of 10 v/v% GVL or EtOH and 90 v/v% 95-octane gasoline, show very
similar properties [15].
The production of GVL was often obtained by the hydrogenation
of LA through the reaction C5H8O3 + H2 → C5H8O2 + H2O (path-
way 1). However, to exert the hydrogenation of LA efficiently
into GVL, its undesired side reaction to MTHF via the reaction
C5H8O2 + 2H2 → C5H10O + H2O (pathway 2) must be avoided. Stud-
ies have shown that a critical dependency of the hydrogenation of
LA via pathway 1, avoiding of the side reaction of GVL via path-
way 2, depended on the catalyst and the catalytic reaction system
[10,12].
Recent studies have shown that LA could be transformed into
GVL (pathway 1) using the Ru-based homogeneous catalyst sys-
tem [10,16,17]. High conversion of LA as well as good selectivity
of GVL has been reported. However, the recycle and separation of
the homogeneous catalyst from reaction residues later on was still
a challenge. Subsequently, more attractive heterogeneous catalyst
was developed due to its easy recycle. The most often employed
catalysts were on the Ru-based nanocatalysts [18,19]. Manzer
reported that 97% yield of GVL was obtained using Ru/C catalyst
at 150 ◦C and 34.5 bar H2 in dioxane solvent [18]. Bourne et al. [19]
reported that 99% yield of GVL was achieved using the Ru/Al2O3
catalyst under 200 ◦C and 20 MPa H2 via water in combination
with supercritical CO2 as the reaction medium. The inherent draw-
back existed in the utilization of Ru-catalysts was the stability and
∗
Corresponding author. Tel.: +1 807 627 3059.
E-mail addresses: kyan@lakeheadu.ca, xxmsxty@sina.com (K. Yan).
0926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.08.037

K. Yan et al. / Applied Catalysis A: General 468 (2013) 52–58
53
sample was pressurized with liquid CO2 and reduced with H2 at
room temperature for 12 h under the stirring condition.
In comparison, 5 wt% Pd/SiO2 nanoparticle was prepared by the
traditional impregnation method in the liquid solution. The typical
procedure was performed as following [20,27]: firstly, the pre-
calculated amount of Pd(OAc)2 was added into 5 mL H2O, then
impregnated with the SiO2 support (∼0.3 g) for 24 h. After this, the
solvent H2O was slowly evaporated under 110 ◦C and then calci-
nated at 450 ◦C for 6 h in the furnace with air environment. Finally,
the sample was reduced by H2–N2 (v/v ratio of 10/90) at 200 ◦C for
3 h.
Scheme 1. Hydrogenation of LA to value-added chemicals and biofuels.
2.3. Characterizations
deactivation [20]. More recently, our previous studies have shown
that the Cu-based catalysts derived from their hydrotalcite pre-
cursors were efficient for the hydrogenation of LA. However, the
Cu-catalysts were fabricated through several steps and required
the regeneration [21,22]. The development of a more stable solid
catalyst with high performance remains challenging.
The resulting nanoparticles catalysts were examined by trans-
mission electron microscopy (TEM) to investigate structural
features with a JEOL-2010 instrument. The mean particle size d
was calculated from the following formula: d = ꢀnidi/ꢀni, where
ni is the number of particles of size di. The detection limit is about
1 nm for the supported Pd particles. The energy-dispersion X-ray
(EDX, Hitachi SU-70) mapping was further employed to verify the
existence and dispersion of elements in the resultant samples. The
powder X-ray diffraction (XRD) patterns for the crystal phase anal-
ysis were collected on a Bruker AXS D8 Advance with Cu K␣1 of
Considering the aforementioned challenges for the production
of biofuel GVL from biomass-derived LA, we conceived of a series
of more stable and robust Pd nanoparticles (Pd/SiO2) catalysts that
were facilely fabricated by the assistance of green solvent CO2. The
approach developed in this work avoids the use of aqueous means,
allows the minimization of liquid waste generation, improves rapid
separation of products [23,24], thus, can be considered as facile and
eco-friendly approach. The resulting Pd nanoparticles catalysts pre-
sented uniform distribution with small particle size, and exhibited
excellent performances in the hydrogenation of biomass-derived
LA, showing close to perfect selectivity of GVL with the TON of
884.7 at 97.3% conversion under mild conditions. The catalytic per-
formance was superior to the values of the 5 wt% Pd/SiO2 catalyst
prepared by the traditional impregnation method. Consequently,
the metal Pd loading, reaction temperature, hydrogen pressure,
reaction time, catalyst stability and reaction mechanism were also
studied in this work.
˚
1.54060 A as a radiation source. The data were collected in the
range of 25◦ to 90◦ with intervals of 0.05◦ and the scanning rate
of 10◦/min. The particle size for each sample has been calculated
from the Scherrer equation (1), where ꢁ corresponds to the Cu K␣
radiation, and ˇ is the full width at half-maximum for a reflection
maximum located at 2ꢂ).
0.9ꢁ
ˇ cos ꢂ
L =
(1)
X-ray photoelectron spectroscopy (XPS) measurements were
performed with an ESCALAB 250 spectrometer with a hemi-
spherical analyzer and a monochromatized Al K␣ X-ray source
(E = 1486.6 eV), operated at 15 kV and 15 mA. For the narrow scans,
analyzer pass energy of 40 eV was applied. The temperature-
programmed reduction was carried out in a U-shaped quartz
reactor placed in a furnace controlled by a temperature program-
mer (Omega Model CN 2000).
2. Experimental
2.1. Chemicals
The reducibility of the calcined samples was determined by
H2-temperature-programmed reduction (TPR). In these measure-
ments, 50 mg of a sample was placed in a quartz reactor and heated
at 10 ◦C/min up to 550 ◦C under a He flow of 20 mL/min, and held
at this temperature for 1 h. The reactor was then cooled down to
0 ◦C and the sample exposed to a stream of 5% H2/Ar at a flow rate
of 20 mL/min. Subsequently, the sample was heated up to 400 ◦C at
a heating rate of 10 ◦C/min. The amount of hydrogen consumed as
a function of temperature was monitored on-line on a TCD detec-
tor. The maximum rate of H2 consumption was used to choose the
reduction temperature for each catalyst to be conducted in situ
before reaction.
SiO2 (surface-area of 220 m2/g), palladium (II) acetate
(Pd(OAc)2, ≥99.9% trace metals basis), levulinic acid (98%),
pentanoic acid (≥99%), 1,4-pentanediol (99%), ␥-valerolactone
(99%), 2-methyltetrahydrofuran (analytical standard) were bought
from Sigma–Aldrich. Pure H2O was obtained from a NANOpure®
Diamond TM UV ultrapure water purification system. All the
chemicals were directly used without further treatment after
purchase.
2.2. Catalyst preparation
A series of Pd nanoparticles were prepared using the Pd precur-
sor (Pd(OAc)2) deposited on the amorphous SiO2 support, which
was prepared in the presence of CO2 similarly reported as previ-
ous studies [25,26]. The general procedure was described in Fig.
S1 and the typical procedure was performed as following: firstly,
a constant amount of SiO2 support (∼0.3 g) and the precalculated
amount of Pd(OAc)2 was added into a 50 mL micro-vessel. After
this, the vessel was sealed and filled with ∼18 g CO2. The ves-
sel was under the stirring condition for another 24 h to ensure
the high dispersion of the Pd(OAc)2 at room temperature. After
this step the vessel was depressurized and the sample was calci-
nated at 450 ◦C for 6 h in the oven with air environment, whereby
organic compounds were burned and removed. Finally, the
2.4. Hydrogenation of LA
The hydrogenation of LA was performed in a 25 mL micro-
vessel (Parr Company) with the external temperature controller
and pressure indicator. Typically, LA (∼5.0 g) was dissolved in
deionized water (5 mL) and then transferred into the vessel. After
this, the catalyst (∼0.1 g) was added, followed by the repeated
procedure of filling argon and slowly making vacuum for three
times. After reactions, the vessel was cooled down by water. The
reaction products were firstly centrifuged for 30 min and then
filtrated to obtain clear solution. The samples were analysed by
GC (Shimadzu 2014, Column: 30 m DBWaxetr, FID Detector; the

54
K. Yan et al. / Applied Catalysis A: General 468 (2013) 52–58
Fig. 1. TEM images of the resulting Pd/SiO2 nanoparticles catalysts: (a) 3 wt%, (b) 5 wt%, (c) 7 wt%.
column temperature was raised from 40 to 250 ◦C with a heat-
ing rate of 3 ◦C/min; the injector temperature was set at 350 ◦C
with a heating speed of 10 ◦C/min and the sampling volume
was 0.2 ␮L). The potential Pd leach was detected by inductively
coupled plasma-mass spectroscopy (ICP-MS, Perkin–Elmer Elan
DRC-e).
3. Results and discussion
3.1. Characterizations of catalysts
Fig. 1 shows the representative TEM images of the as-prepared
Pd/SiO2 nanoparticles catalysts. TEM images (Fig. 1) revealed

K. Yan et al. / Applied Catalysis A: General 468 (2013) 52–58
55
Table 1
Catalytic activities of the resulting Pd/SiO2 nanoparticles catalysts in the hydrogena-
tion of LA.a
No.
Catalyst
Conv. (LA)/%
Sel. (GVL)/%
TON b
1
2
3
1 wt% Pd/SiO2
3 wt% Pd/SiO2
5 wt% Pd/SiO2
7 wt% Pd/SiO2
5 wt% Pd/SiO2
8.6
30.2
57.3
75.6
32.8
95.2
99.1
98.9
97.8
90.7
375.1
457.2
519.4
484.0
272.7
4
5b
a
Hydrogenation conditions: m(LA) = 5.0 g, m(catalyst) = 0.10 g, V(H2O) = 5 mL,
p(H2) = 50 bar, T = 160 ◦C, t = 6 h, stirring speed = 1000 rpm. [b] Catalyst was prepared
by the impregnation method as described in Experimental section.
b
Turnover numbers (TON) = the molar of the GVL/the molar of the metal catalyst
used.
corresponding to the 3d3/2 and 3d5/2 orbital states. The peak regions
of Pd can be fitted with two sets of peaks at 340.6 eV (3d3/2) and
335.3 eV (3d5/2), which indicated that the successful reduction was
achieved. Besides, the very weak peak was observed at 341.7 eV
and 336.5 eV, which was possible due to the presence of very small
amount of PdO on the surface or the interaction between of Pd and
the oxygen from the SiO2 support [29].
Fig. 2. Powder X-ray diffraction patterns of the resulting Pd/SiO2 nanoparticles
catalysts: (a) 3 wt%, (b) 5 wt%, (c) 7 wt%, (d) 5 wt% prepared by the traditional impreg-
nation method.
The TPR profiles of the chosen 3% Pd/SiO2 and 5% Pd/SiO2 cat-
alysts are compared in Fig. S5. For 3% Pd/SiO2 catalyst, the TPR
profile mainly exhibits one peak centered on ∼105 ◦C, indicative
of the H2 consumption [30]. The H2 profile was consistent with
the decomposition of Pd(OAc)2, which was reported to form at
room temperature over large Pd particles (>2 nm) [31]. The other
weak peak at ∼150 ◦C, was ascribed to water evolution as previ-
ously reported in [32,33]. In the case of 5 wt% Pd/SiO2, the main
peak was observed at ∼125 ◦C, which was from the reduction of
Pd(II) to Pd(0). It is also worth noting that the higher amount of Pd
decreases the reducibility of Pd, leading to a shift toward a little
higher temperatures, which indicated that bigger particle size was
existed. Besides, a second broader peak was detected in the range
of 200–350 ◦C, which was due to the water evolution [32,33].
that uniform and small Pd nanoparticles were homogeneously
dispersed on the amorphous SiO2 support. TEM measurement indi-
cated that the mean size of Pd nanoparticle was 4.32 nm in the case
of 3 wt% loading. With the Pd loading increasing to 5 wt%, more
black dots of Pd nanoparticles were observed and the mean par-
ticle size was 6.15 nm. Further increasing the Pd loading to 7 wt%
(Fig. 1c), the bigger size of Pd with a mean particle size of 8.86 nm
was observed.
EDX analysis was used to further confirm the existence of Pd and
the element distribution in the resulting samples, EDX mapping of
the representative 5 wt% Pd/SiO2 nanoparticle catalyst is shown in
Fig. S2. Highly homogeneous dispersion of Pd, Si, O in the 5 wt%
Pd/SiO2 nanoparticle catalyst was observed (Fig. S2). Besides, no
distinct difference of Pd amount was revealed by the EDX analysis
in several different areas, indicating the uniform dispersion of Pd
nanoparticles in the sample.
3.2. Hydrogenation of LA to GVL
3.2.1. Influence of Pd loading
The amorphous SiO2 support and the synthesized nanoparti-
cles catalysts are further studied by XRD tests (Fig. S3 and Fig. 2).
The peak at ∼22◦ in Fig. S3 indicated the amorphous silica. The
diffraction peaks (Fig. 2) indexed as 2ꢂ of (1 1 1) approximately at
40.1◦, (2 0 0) at 46.3◦, (2 2 0) at 67.7◦, (3 1 1) at 81.8◦, and (2 2 2)
at 86.5◦, were identified as a single fcc phase of palladium [28].
The experimental pattern matched well with the standard crys-
tal graph (JCPDS PDF 01-087-0643, Fm-3m, a = b = c = 0.3908 nm)
for the syn-Pd0. Besides, no other phase (e.g. PdO) was observed,
which confirmed that the bulk Pd/SiO2 nanoparticles catalysts were
successfully reduced. For the 1 wt% Pd loading, the diffraction peaks
were too weak, which was possibly due to the detection limit of the
XRD fixture. With the Pd loading increasing (Fig. 2a–c), the peaks
appeared more intense. Furthermore, the crystal size of 4.5 nm,
6.7 nm and 9.0 nm was calculated based on the Scherrer Equa-
tion, respectively, which was close to the mean size measured by
TEM images (Fig. 1). In comparison, the 5 wt% Pd/SiO2 catalyst pre-
pared by the traditional impregnation method was also successfully
reduced as shown in the XRD pattern (Fig. 2d).
For the supported metal nanoparticles, the metal surface was
easy to be oxidized during the preparation process due to the
unstable thermodynamic property of Pd0. A different chemical
environment of the atoms at the interface will generally lead to the
appearance of new spectral features in the XPS spectra. Thus XPS
was employed and the recorded XPS patterns of the representative
3 wt%, 5 wt% and 7 wt% Pd/SiO2 nanoparticle catalysts are shown
in Fig. S4. The presence of two prominent sets of Pd (3d) peaks,
Table 1 reports the catalytic results obtained for the resulting
Pd/SiO2 nanoparticles catalysts in the hydrogenation of biomass-
derived LA. It was interesting to find that high selectivity of GVL
was achieved in the presence of Pd/SiO2 nanoparticles catalysts.
The metal Pd loading presented crucial influence on the observed
100
80
60
40
20
0
100
90
80
70
60
50
Conversion/%
Selectivity/%
TON: 629.7
TON: 576.5
TON: 572.5
TON: 519.4
TON: 179.5
140
160
180
200
220
o
Reaction temperature/ C
Fig. 3. Influence of reaction temperature on the catalytic activities. Note: Hydro-
genation conditions: m(LA) = 5.0 g, m(catalyst) = 0.1 g, V(H2O) = 5 mL, p(H2) = 50 bar,
t = 6 h, stirring speed = 1000 rpm.

56
K. Yan et al. / Applied Catalysis A: General 468 (2013) 52–58
Fig. 4. TEM analysis of the reduced 5 wt% Pd/SiO2 nanoparticles catalyst.
catalytic activities. With the Pd loading increasing from 1 wt% to
7 wt% (No. 1–4), the conversion of LA increased clearly from 8.6% to
75.6% and the TON value ranged between 375.1 and 519.4. These
results indicated that Pd loading can be used to alter the extent of
hydrogenation, where the uniform Pd with small size (Fig. 1) would
play a role. Reviewing the catalytic performance in Table 1, 95.2%
selectivity of GVL with the TON of 375.1 at the 8.6% conversion of LA
was observed when the 1 wt% Pd/SiO2 nanoparticles catalyst was
employed (No. 1). While 98.9% selectivity of GVL with TON of 519.4
at 57.3% conversion of LA as well as was achieved in the case of 5 wt%
Pd/SiO2 nanoparticles catalyst (No. 3). In comparison, 5 wt% Pd/SiO2
nanoparticles catalyst was prepared by the traditional impregna-
tion method, displaying 90.7% selectivity with the TON of 272.7 at
32.8% conversion (No. 5). These results indicated that the crucial
role of Pd in the hydrogenation. Based on the economic view and
the catalytic performances (Table 1), 5 wt% Pd/SiO2 nanoparticles
catalyst was thus chosen for further studies.
of LA. The selectivity of GVL behaved relatively stable in general.
High hydrogen pressure resulted in the rapid conversion of LA and
higher TON values. For example, when 30 bar H2 was employed (No.
5), 25.5% conversion of LA and the TON of 232.6 were achieved,
where the conversion and TON were increased significantly to
97.3% and 884.7 under 90 bar H2 (No. 7), respectively. This was
possibly related to the fact that more H2 was transported to the
reactants, resulting a fast rate and higher conversion. Besides, to
efficiently compare with the previous works, representative works
from literatures [20–22,34–41] were summarized in Table S1. To
our limited knowledge, our study shows the highest value obtained
so far on the supported Pd catalysts. Even in comparison with the
often utilized Ru-, Cu-, and Ir-catalysts, our results are still among
the top-list as shown in Table S1 and the catalysts developed in
this work are more selective for the production of GVL. Besides,
the method to fabricate the metal catalyst developed in this work
avoids the use of aqueous means, allows the minimization of liq-
uid waste generation, improves a higher rate of mass transfer, and
rapid separation of products, thus, can be considered as the simple,
safe and ecofriendly route.
3.2.2. Influence of reaction temperature
Fig. 3 reveals the reaction temperature influence on the catalytic
activities. It was observed that the conversion of LA was increased
from 19.7% to 73.4% and the TON ranged from 179.5 to 629.7 as the
reaction temperature was raised from 140 to 220 ◦C. This was due to
the fact that the high temperature would promote a faster reaction
rate and mass transfer of reactants as well as H2. However, too
high temperature (e.g. 220 ◦C) would promote the side reactions
and clearly decreased the selectivity of GVL. For example, 99.4%
selectivity of GVL was achieved at 140 ◦C, which was reduced to
92.3% at 220 ◦C. Meantime, the GC detected the whole percentage of
the side products (pentanoic acid and MTHF) were ∼6.8% at 220 ◦C.
Even at 200 ◦C, small amount (<5%) of pentanoic acid and MTHF was
found, which indicated that side reactions would be more prone to
occur at high temperature. Based on these results, 180 ◦C was thus
chosen for further studies.
3.3. Recyclability tests
To understand the stirring speed on the catalytic performance,
different stirring speed was tested as shown in Table S2. Increas-
ing the stirring speed has very slide influence on the conversion
and selectivity, which suggested that the mass transportation influ-
ence was weak. Subsequently, we focus on the recyclability tests of
the resulting nanoparticle catalyst, the 5 wt% Pd/SiO2 nanoparticle
catalyst was recycled from reaction residues through centrifuga-
tion for 30 min, dried at 90 ◦C for 12 h, calcinated at 350 ◦C for 3 h
and then reduced by H2 in liquid CO2. TEM is further employed to
detect the potential aggregation of Pd nanoparticles in the process
of reduction. TEM images and particle distribution are shown in
3.2.3. Influence of reaction time and hydrogen pressure
Table 2
Table 2 presents the influence of reaction time and hydrogen
pressure on the catalytic activities. The conversion of LA and the
TON values increased with the prolonging reaction time (No. 1–4).
At the reaction time of 4 h (No. 1), the conversion of LA was 39.4%,
which was increased to 88.1% at the reaction time of 10 h (No. 4).
However, further prolonging reaction time, it would cause the fur-
ther ring-opening of GVL to pentanoic acid or MTHF, as detected
by our GC and indicated by the decrease of the GVL selectivity (No.
1–4). The hydrogen pressure was subsequently studied at the reac-
tion time of 6 h as shown in Table 2. As it turned out, the hydrogen
pressure (No. 5–7) presented crucial influence on the conversion
Reaction time and hydrogen pressure on the catalytic activity.a
Entry
Time/h
Pressure/bar
Conv. (LA)/%
Sel. (GVL)/%
TON
1
2
3
4
5
6
7
4
6
8
10
6
50
50
50
50
30
70
90
39.4
51.7
72.1
88.1
25.5
87.6
97.3
99.6
99.3
98.2
96.7
99.5
98.8
99.2
359.7
470.5
648.9
780.8
232.6
794.0
884.7
6
6
a
Other hydrogenation conditions: T = 180 ◦C, V(H2O) = 5 mL, catalyst was 5 wt%
Pd/SiO2, m(catalyst) = 0.10 g, m(LA) = 5.0 g, stirring speed = 1000 rpm.

K. Yan et al. / Applied Catalysis A: General 468 (2013) 52–58
57
and liquid LA [36,38], whereby the uniform dispersion of Pd0 with
small size (Fig. 2) might play an important role. Molecular H2 was
adsorbed on the surface of Pd0 through the formation of hydrogen
bonds between hydrogen with Pd. LA was also adsorbed on the sur-
face of Pd in the combination of Pd0 with carboxylic C and O atoms
[40,42]. When the first H was added to LA molecule, an intermedi-
ate that was linked by a ꢃ-bond was formed. The intermediate was
stabilized by its interaction with Pd0. If the intermediate acquired
one more H atom, it then formed 4-hydroxypentanoic acid and fur-
ther bonded on the metal surface. Then the hydroxyl group and the
acid group would react to lose one molecule of water through the
self-condensation, rapidly producing GVL [16,42].
100
80
60
40
20
0
100
TON: 870.3
TON: 828.9
TON: 815.4
80
60
40
20
0
Conversion/%
Selectivity/%
TON: 551.6
4. Conclusions
Run 1
Run 2
Run 3
Run 4
A series of Pd nanoparticles supported on the SiO2 support
were successfully synthesized and exhibit efficient catalytic per-
formance in the hydrogenation of biomass-derived LA, showing
the perfect selectivity of biofuel GVL with high TON of 884.7 at
97.3% conversion of on the 5 wt% Pd/SiO2 catalyst under mild con-
ditions of 180 ◦C, 6 h and 90 bar H2. The facile, safe and eco-friendly
approach to fabricate Pd nanoparticles in the presence of liquid CO2,
avoids the use of aqueous means, allows the minimization of liquid
waste generation, and improves rapid separation of nanoparticles.
The catalytic performance of the resulting Pd nanoparticles catalyst
was superior to the values of the 5 wt% Pd/SiO2 catalyst prepared
by the traditional impregnation method. Besides, the parameters of
Pd loading, reaction temperature and hydrogen pressure presented
crucial influence on the catalytic performance.
Fig. 5. Catalytic activities of the recycled 5 wt% Pd/SiO2 nanoparti-
cles catalyst. Note: Other conditions: T = 180 ◦C, V(H2O) = 5 mL, t = 6 h,
p(H2) = 90 bar, m(LA) = 5.0 g, m(catalyst) = 0.10 g, stirring speed = 1000 rpm.
Fig. 4. Uniform dispersion of Pd nanoparticles with a mean size of
6.86 nm was achieved (Fig. 4). No clear increase in the Pd nanopar-
ticle size was found in comparison with the fresh 5 wt% Pd/SiO2
nanoparticle catalyst.
The catalytic performances of the spent and reduced 5 wt%
Pd/SiO2 nanoparticles catalysts are shown in Fig. 5. When the spent
5 wt% Pd/SiO2 nanoparticles catalyst was used for the hydrogena-
tion of LA, 65.2% conversion was obtained with 92.3% selectivity of
GVL and the TON of 551.6 (Run 1), which was a little lower than
the values (99.2% selectivity and the TON of 884.7 at 97.3% conver-
sion of LA) of the fresh 5 wt% Pd/SiO2 nanoparticles catalyst, which
was possible due to carbon deposit from the reactant and/or prod-
ucts [40]. Fortunately, the further hydrogenation using the reduced
5 wt% Pd/SiO2 nanoparticles catalysts (Run 2–4), the selectivity of
GVL behaved rather stable under the standardized conditions. Also
the conversion of LA dropped slightly from 96.7% to 91.9% and the
TON decreased from 870.3 to 815.4. These small decreases were
most likely the result of an incomplete catalyst recovery in the pro-
cedure. Furthermore, analysis of the reaction solution after the third
run revealed no detectable leach of Pd, which confirmed the high
stability of Pd nanoparticles catalysts in this work.
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.apcata.
2013.08.037.
References
[1] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044–4098.
[2] P.M. Mortensen, J.D. Grunwaldt, P.A. Jensen, K.G. Knudsen, A.D. Jensen, Appl.
Catal. A 407 (2011) 1–19.
[3] B. Kamm, Angew. Chem. Int. Ed. 46 (2007) 5056–5058.
[4] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539–554.
[5] J.X. Xi, Y. Zhang, Q. Xia, X.H. Liu, J.W. Ren, G.Z. Lu, Y.Q. Wang, Appl. Catal. A 459
(2013) 52–58.
3.4. Hydrogenation pathway of LA
[6] K. Yan, T. Lafleur, J.Y. Liao, J. Nanopart. Res. 15 (2013) 1906–1912.
[7] D.M. Alonso, J.Q. Bond, J.A. Dumesic, Green Chem. 12 (2010) 1493–1513.
[8] C.H. Zhou, X. Xia, C.X. Lin, D.S. Tong, J. Beltramini, Chem. Soc. Rev. 40 (2011)
5588–5617.
[9] X.L. Tong, Y. Ma, Y.D. Li, Appl. Catal. A 385 (2010) 1–13.
[10] F.M.A. Geilen, B. Engendahl, A. Harwardt, W. Marquardt, J. Klankermayer, W.
Leitner, Angew. Chem. Int. Ed. 49 (2010) 5510–5514.
[11] R. Luque, J.H. Clark, Catal. Commun. 11 (2010) 928–931.
[12] D.M. Alonso, S.G. Wettstein, M.A. Mellmer, E.I. Gurbuz, J.A. Dumesic, Energy
Environ. Sci. 6 (2013) 76–80.
Fig. S5 shows the representative GC spectrum of the products
distribution in the hydrogenation of LA. Under the experimental
conditions, a main reaction network for the formation of GVL was
proposed in Scheme 2. It was proposed that the first step during
the hydrogenation reaction was the chemisorption of molecular H2
[13] I.T. Horvath, H. Mehdi, V. Fabos, L. Boda, L.T. Mika, Green Chem. 10 (2008)
238–242.
[14] V. Fabos, G. Koczo, H. Mehdi, L. Boda, I.T. Horvath, Energy Environ. Sci. 2 (2009)
767–769.
[15] K. Nasirzadeh, D. Zimin, R. Neueder, W. Kunz, J. Chem. Eng. Data 49 (2004)
607–612.
[16] F.M.A. Geilen, B. Engendahl, M. Hölscher, J. Klankermayer, W. Leitner, J. Am.
Chem. Soc. 133 (2011) 14349–14358.
[17] H. Mehdi, V. Fábos, R. Tuba, A. Bodor, L.T. Mika, I.T. Horváth, Top. Catal. 48 (2008)
49–54.
[18] L.E. Manzer, Appl. Catal. A 272 (2004) 249–256.
[19] R.A. Bourne, J.G. Stevens, J. Ke, M. Poliakoff, Chem. Commun. (2007) 4632–4634.
[20] S.G. Wettstein, J.Q. Bond, D.M. Alonso, H.N. Pham, A.K. Datye, J.A. Dumesic, Appl.
Catal. B 117–118 (2012) 321–329.
[21] K. Yan, J.Y. Liao, X. Wu, X.M. Xie, RSC Adv. 3 (2013) 3853–3856.
[22] A.M. Hengne, C.V. Rode, Green Chem. 14 (2012) 1064–1072.
[23] W. Leitner, Acc. Chem. Res. 35 (2002) 746–756.
[24] C.H. Yen, K. Shimizu, Y.Y. Lin, F. Bailey, I.F. Cheng, C.M. Wai, Energy Fuels 21
(2007) 2268–2271.
Scheme 2. Hydrogenation pathway of LA to GVL on Pd/SiO2 nanoparticles catalysts.

58
K. Yan et al. / Applied Catalysis A: General 468 (2013) 52–58
[25] J. Kim, G.W. Roberts, D.J. Kiserow, Chem. Mater. 18 (2006) 4710–4712.
[26] J. Kim, M.J. Kelly, H.H. Lamb, G.W. Roberts, D.J. Kiserow, J. Phys. Chem. C 112
(2008) 10446–10452.
[27] J. Wu, Y.M. Shen, C.H. Liu, H.B. Wang, C. Geng, Z.X. Zhang, Catal. Commun. 6
(2005) 633–637.
[28] K. Yan, T. Lafleur, J.Y. Liao, X.M. Xie, Sci. Adv. Mater. 6 (2014) 1–6.
[29] M. Flytzani-Stephanopoulos, B.C. Gates, Annu. Rev. Chem. Biomol. Eng. 3 (2012)
545–574.
[35] A.M.R. Galletti, C. Antonetti, V. De Luise, M. Martinelli, Green Chem. 14 (2012)
688–694.
[36] P. Azadi, R. Carrasquillo-Flores, Y.J. Pagán-Torres, E.I. Gürbüz, R. Farnood, J.A.
Dumesic, Green Chem. 14 (2012) 1573–1576.
[37] X.L. Du, L. He, S. Zhao, Y.M. Liu, Y. Cao, H.Y. He, K.N. Fan, Angew. Chem. Int. Ed.
50 (2011) 7815–7819.
[38] D.M. Alonso, S.G. Wettstein, J.Q. Bond, T.W. Root, J.A. Dumesic, ChemSusChem
4 (2011) 1078–1081.
[30] A.F. Gusovius, T.C. Watling, R. Prins, Appl. Catal. A 188 (1999) 187.
[31] J.J. Panpranot, O. Tangjitwattakorn, P. Praserthdam, J.G. Goodwin Jr., Appl. Catal.
A 292 (2005) 322–327.
[32] S. Fessi, A.S. Mamede, A. Ghorbel, A. Rives, Catal. Commun. 27 (2012)
109–113.
[39] P.P. Upare, J.M. Lee, D.W. Hwang, S.B. Halligudi, Y.K. Hwang, J.S. Chang, J. Ind.
Eng. Chem. 17 (2011) 287–292.
[40] Z.P. Yan, L. Lin, S. Liu, Energy Fuels 23 (2009) 3853–3858.
[41] P.P. Upare, J.M. Lee, Y.K. Hwang, D.W. Hwang, J.H. Lee, S.B. Halligudi, J.S. Hwang,
J.S. Chang, ChemSusChem 4 (2012) 1749–1752.
[33] F. Sotoodeh, L. Zhao, K.J. Smith, Appl. Catal. A362 (2009) 155–162.
[34] Y. Gong, L. Lin, Z.P. Yan, BioResources 6 (2011) 686–699.
[42] J.C. Serrano-Ruiz, D.J. Braden, R.M. West, J.A. Dumesic, Appl. Catal. B 100 (2010)
184–189.