Selective hydrogenation of biomass-derived 5-hydroxymethylfurfural (HMF) to 2,5-dimethylfuran (DMF) under atmospheric hydrogen pressure over carbon supported PdAu bimetallic catalyst

Article abstract of DOI:10.1016/j.cattod.2013.10.012

Hydrogenation of 5-hydroxymethylfurfural (HMF) to 2,5-dimethylfuran (DMF) was examined by Pd_xAu_y/C catalysts prepared with various Pd/Au molar ratio (x/y) in the presence of hydrochloric acid (HCl) under an atmospheric hydrogen pressure. Bimetallic Pd_xAu_y/C catalysts had a significant activity for a selective hydrogenation of HMF toward DMF comparing to monometallic Pd/C and Au/C catalysts. To clarify the novelty of Pd_xAu_y/C catalysts, characterizations by using X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), X-ray absorption spectra (XAFS), a transmission electron microscopy (TEM) and other analytical techniques were studied. XPS and X-ray absorption near-edge structure (XANES) analyses indicated that there was the charge transfer phenomenon from Pd to Au atoms in Pd_xAu_y/C. Existence of PdAu alloy structures in Pd_xAu_y/C was expected by XRD, TEM and extended X-ray absorption fine structure (EXAFS) analyses. Accordingly, we concluded that PdAu alloys supported carbon exhibited a good catalytic performance for a selective hydrogenation of HMF to DMF using an atmospheric hydrogen pressure.

Full text of DOI:10.1016/j.cattod.2013.10.012

Catalysis Today 232 (2014) 89–98
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Selective hydrogenation of biomass-derived 5-hydroxymethylfurfural
(HMF) to 2,5-dimethylfuran (DMF) under atmospheric hydrogen
pressure over carbon supported PdAu bimetallic catalyst
Shun Nishimura, Naoya Ikeda, Kohki Ebitani∗
School of Materials Science, Japan Advanced Institute of Science and Technology, 1-1, Asahidai, Nomi, Ishikawa 923-1292, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Hydrogenation of 5-hydroxymethylfurfural (HMF) to 2,5-dimethylfuran (DMF) was examined by
PdxAuy/C catalysts prepared with various Pd/Au molar ratio (x/y) in the presence of hydrochloric acid
(HCl) under an atmospheric hydrogen pressure. Bimetallic PdxAuy/C catalysts had a significant activity
for a selective hydrogenation of HMF toward DMF comparing to monometallic Pd/C and Au/C catalysts. To
clarify the novelty of PdxAuy/C catalysts, characterizations by using X-ray diffraction (XRD), X-ray photo-
electron spectroscopy (XPS), X-ray absorption spectra (XAFS), a transmission electron microscopy (TEM)
and other analytical techniques were studied. XPS and X-ray absorption near-edge structure (XANES)
analyses indicated that there was the charge transfer phenomenon from Pd to Au atoms in PdxAuy/C.
Existence of PdAu alloy structures in PdxAuy/C was expected by XRD, TEM and extended X-ray absorption
fine structure (EXAFS) analyses. Accordingly, we concluded that PdAu alloys supported carbon exhibited a
good catalytic performance for a selective hydrogenation of HMF to DMF using an atmospheric hydrogen
pressure.
Received 16 July 2013
Received in revised form
22 September 2013
Accepted 2 October 2013
Available online 29 October 2013
Keywords:
Hydrogenation
PdAu bimetallic catalyst
Bio-fuel
2,5-Dimethylfuran
© 2013 Elsevier B.V. All rights reserved.
respectively. Moreover, a low solubility in water (2.3 g L−1) is supe-
rior to miscible ethanol, and it enables use as the blend fuel.
1. Introduction
Fossil energies such as natural gas, coal and petroleum dominate
approximately 80% in primary energy consumption estimated by
sources [1]. This excessive reliance on fossil fuels in the world has
serious reservations about their depletion and green house emis-
sion. Now, with increasing the demand for environmental concerns
about global warming, the development of renewable and environ-
mentally friendly energy sources have been a hot and an important
topic in these decades [2,3]. The energy-efficient processes for the
sustainable production of fuels derived from biomass resources
such as bio-ethanol, biodiesel fuel (BDF) and liquid alkanes have
also attracted attention over the year [4].
5-Hydroxymethylfurfural (HMF), obtained by dehydration of
monosaccharides, is considered to be the most promising impor-
tant intermediate for the synthesis of a wide variety of chemicals
and alternative fuels based on bio-refinery [5]. To upgrade furanic
compounds toward bio-fuels, hydrogenation is the most versatile
reaction. 2,5-Dimethylfuran (DMF) is known as one of the potential
transportation fuels due to its high energy density (30 kJ cm−3) and
a high research octane number (RON = 119), these are similar val-
ues to that of gasoline which possesses 34 kJ cm−3 and RON = 89–96,
For the gas-phase converter system affording to DMF, Dumesic
et al. designed the copper/ruthenium (CuRu/C) catalyzed biphasic
reactor with hydrochloric acid (HCl) and sodium chloride (NaCl)
[6]. They also reported that the Cu3Ru1/C catalyst exhibited 71%
yield of DMF from 5 wt% HMF at 220 ◦C under pressured hydro-
gen (6.8 atm H2) in batch reactor. Chidambaram and Bell proposed
the catalytic conversion of glucose to DMF with a following two-
step reaction; dehydration of glucose to HMF using heteropoly
acid (12-molybdophosphoric acid) in ionic liquid and successive
hydrogenation of HMF to DMF with Pd/C catalyst [7]. In their
report, the Pd/C showed 15% yield of DMF via hydrogenation reac-
tion of HMF (1 mmol) in ionic liquid/acetonitrile at 120 ◦C under
62 atm of H2. Hydrogenation of HMF toward BHMF over subnano-
ordered Au cluster-supported Al2O3 (140 ◦C, 38 atm H2) [8] or
toward 2,5-bis(hydroxymethyl)tetrahydrofuran (DHTHF) over Ru,
Pd or Pt-supported catalysts (130 ◦C, 27 atm H2) [9], and syntheses
of DHTHF and 2,5-dimethyltetrahydrofuran (DMTHF) from various
saccharides over RhCl3 catalyst (80–160 ◦C, 20 atm H2) [10] were
also examined under pressured H2 condition, previously.
Under an atmospheric H2 condition, hydrogenation of HMF
toward DMF has mainly catalyzed by Pd-supported carbon in the
presence of strong acidic media. van Bekkum et al. investigated
the hydrogenation of HMF to DMF by using a combination of Pd/C
catalyst and HCl as an activator in alcohol solvents [11]. They
∗
Corresponding author. Tel.: +81 761 51 1610; fax: +81 761 51 1149.
E-mail address: ebitani@jaist.ac.jp (K. Ebitani).
0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.10.012

90
S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
achieved 36% yield of DMF from HMF at 60 ◦C under 1 atm of H2.
In the presence of alcohol, the reaction between alcohol solvent
and furanyl ketone to ether, ex. the monopropyl ether of 2,5-
bis(hydroxymethyl)furan (BHMF) formation with 1-propanol and
BHMF, is an expected intermediary reaction step. Notably, Rauch-
fuss and co-workers succeeded >95% yield of DMF through the
hydrogenation of HMF in the presence of Pd/C, sulfuric acid (H2SO4)
and formic acid (FA) in refluxing tetrahydrofuran (THF) [12,13].
Their key strategy is the multirole of FA as an acid catalyst, solvent,
a source of hydrogen (H2) and a reagent for the dehydrogenation
of furanylmethanols to produce the formate ester intermediate of
BHMF and 2-hydroxymethyl-5-methylfuran (HMMF). Saha and co-
workers also tried the synthesis of DMF from different biomass
sources by using FA with H2SO4 and Ru/C catalyst in refluxing THF,
and they afforded 30% yield of DMF from fructose [14]. Follow-
ing to these approaches, production of the ester intermediate such
as propyl ether and formate ester is one of the crucial points for
enhancement of hydrogenolysis step.
Herein, we studied the hydrogenation of HMF toward DMF over
PdAu bimetallic catalyst under an atmospheric H2 pressure. The
PdAu types of bimetallic catalyst has been well-known as an active
bimetallic nanocatalyst for aerobic oxidation of alcohols [15–18],
direct synthesis of hydrogen peroxide (H2O2) from H2 and O2
[19,20], Suzuki–Miyaura reactions [21,22], dehalogenation [23],
and so on, these achievements have made tremendous perform-
ances comparing to that over the monometallic Au or Pd catalytic
system. We found that the PdAu alloy particles supported on car-
bon promoted the hydrogenolysis step, and it served good yield of
DMF by the selective hydrogenation of HMF under mild reaction
conditions.
room temperature, the NaBH4 reductant was gradually dropped
into the above mixture of Pd2+ and Au3+ ions to reduce, deposit and
stabilize bimetallic ions onto the carbon surface. The produced pre-
cipitate was filtered, washed with 1 L of deionized water, and dried
in oven at 100 ◦C for overnight. Concentration of the sum of Pd and
Au elements was kept constant as 0.94 mmol g−1 for PdxAuy/C cata-
lysts (x + y = 100). Pd50Ru50/C and Pd50Pt50/C were also prepared by
the same method with RuCl3·3H2O and H2PtCl6·6H2O, respectively,
instead of the mixture of PdCl2 and KCl. On the other hand, the
concentration of 0.47 mmol g−1 was applied for the preparation of
monometallic catalysts such as Pd50/various supports, Ru50/C and
Pt50/C.
2.3. Hydrogenation of HMF to DMF with atmospheric hydrogen
(H2)
Hydrogenation reactions were performed in 60 mL volume
of a Schlenk flask. Briefly, HMF (1.0 mmol) was dissolved into
THF (10 mL) solvent, then PdxAuy/C (31.3 mg) catalyst and HCl
(0.17 mmol) acidic media were added into the mixture. A balloon
(4 L) fulfilled with H2 gas was attached with the Schlenk flask during
the reaction heated in the oil bath at 60 ◦C with a vigorous stirring
(550 rpm). The products were analyzed by a GC chromatography
(Shimadzu GC-2014) equipped with a capillary column (Agilent
J&W GC DB-1; 0.320 mm × 50 m) and a flame ionization detector
(FID) under ramping temperature from 40 to 280 ◦C. The naphtha-
lene was used as an internal standard. Retention times in GC-FID
determined by injections of the commercially available chemi-
cals were following orders; 2-MF (3.44 min), THF (3.70 min), DMF
(4.89 min), DMTHF (4.8 min and 5.01 min) [26], furfural (7.94 min),
furfuryl alcohol (9.38 min), 5-MF (14.55 min), MFA (14.81 min),
naphthalene (19.47 min), DHTHF (19.58 min), HMF (19.70 min),
BHMF (20.77 min).
2. Experimental
2.1. Materials
2.4. Characterization of PdxAuy/C catalysts
Fructose, hydrochloric acid (HCl, 35.0 wt%), sulfuric acid (H2SO4,
96.0 wt%), ruthenium(III) chloride trihydrate (RuCl3·3H2O), potas-
sium chloride (KCl) were obtained by KANTO Chemical Co.,
Inc. Glucose, carbon (Norite “SX plus”, surface area is about
X-ray diffraction (XRD) patterns were obtained by a Smart-Labo
X-ray diffractometer (Shimadzu) using Cu-K␣ radiation at 40 kV
and 30 mA in the range of 2Â from 30◦ to 90◦ in 0.01◦ resolu-
tion. X-ray photoelectron spectroscopy (XPS) was measured by a
AXIS-ULTRA DLD spectrometer (Shimadzu–Kratos) equipped with
a Al target accelerated at 15 kV and 10 mA. The binding ener-
gies of obtained spectra were calibrated with the C 1s speak at
284.5 eV. X-ray absorption spectra (XAFS) in Au L3-edge and Pd
K-edge were recorded at a BL5S1 of Aichi Synchrotron Radiation
Research Center (2nd period, 2013) and a BL01B1 of SPring-8 in the
Japan Synchrotron Radiation Research Institute (JASRI) (Proposal
No. 2012B1610) with a quick XAFS scanning mode. All samples
were grained and pressed to a pellet (Ø = 10 mm). The energy for
Au L3-edge and Pd K-edge XAFS spectra were adjusted at the
edges of Cu foil (8.9800 keV) and Pd foil (24.347 keV), respec-
tively. Double Si(1 1 1) single crystals and double Si(3 1 1) single
crystals were used for division of energy in Au L3-edge and Pd
K-edge, respectively. The obtained spectra were analyzed with
the REX2000 software (ver. 2.5.92, Rigaku). Fourier transforma-
800 m2 g−1
) [24], palladium chloride (PdCl2), tetrachloroau-
ric(III) acid tetrahydrate (HAuCl4·4H2O), hexachloroplatinic(IV)
acid hexahydrate (H2PtCl6·6H2O), sodium borohydride (NaBH4),
dehydrated tetrahydrofuran (THF), formic acid (FA, 99 wt%),
2,5-dimethylfuran (DMF) and 2-methylfuran (2-MF) were pur-
chased from WAKO Pure Chemical Industries, Ltd. Tagatose,
5-hydroxymethylfurfural (HMF), 2,5-dimethyltetrahydrofuran
(DMTHF), furfural and 5-methylfurfural (5-MF) were received
from Sigma–Aldrich Co. Llc. Galactose and naphthalene
were served by Tokyo chemical industry (TCI) Co. Ltd. 5-
Methylfurfurylalcohol
(MFA),
2,5-bis(hydroxymethyl)furan
(BHMF) and 2,5-bis(hydroxymethyl)tetrahydrofuran (DHTHF)
were bought from Toronto Research Chemicals (TRC) Inc.
2.2. Preparation of the PdxAuy bimetallic nanoparticles supported
carbon catalyst
−1
˚
tion (FT) was applied in the range of k = 3–13 A in each EXAFS
spectrum. The curve-fitting analyses were performed in the range
−1
˚
using the inverse FTs in R = 1.841–3.283 A for Au
˚
Monometallic or bimetallic nanoparticles-supported catalysts
were prepared by NaBH4 reduction method referred to the previous
report [25] with some modifications. For instance, the prepa-
ration method of palladium and gold-supported carbon catalyst
(PdxAuy/C) was as follows. First, the carbon support (0.47 g) was
dispersed into the mixed solution of deionized water (55 ml) and
2-propanol (55 ml). Subsequently, PdCl2 (a mmol), KCl (1.0 g) and
HAuCl4·4H2O (b mmol) were added into the solution. After an ultra-
sonic treatment for 5 min and a further vigorous stirring for 1 h at
of k = 4–12 A
˚
L3-edge and R = 1.811–3.161 A for Pd K-edge. Morphologies of the
catalysts were observed by a transmission electron microscopy
(TEM) with Hitachi H-7650 at 100 kV. The size distribution was
produced by randomly counting up of 500 particles. Scanning TEM-
high angle annular dark field (STEM-HAADF) image with an energy
dispersive X-ray spectroscopy (EDS) elemental mapping was car-
ried out by a JEOL JEM-ARM200F instrument operated at 200 kV
with the support of Nanotechnology Platform Program in JAIST of

S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
91
Table 1
Synthesis of DMF via hydrogenation of HMF over the precious metal-supported catalysts.a
Entry
Catalystc
Conversion of HMF/%
Yield/%
5-MF
MFA
DMF
1
2
3
4
5
6
7
8
Pd50/C
>99, 91b
>99, >99b
>99, 67b
98
0, 4b
0, 2b
0, 4b
0
0, 11b
>99, 66b
Pd50Au50/C
Pd50Pt50/C
Pd50Ru50/C
Ru50/C
Pt50/C
Au50/C
0, 2b
>99, 96b
0, 5b
93, 31b
62
4
Trace
0
0
0
0
0
0
0
65
61
53
47
0
0
0
0
Blank
aReaction conditions: HMF (0.5 mmol, b1.0 mmol), THF (10 ml), HCl (0.17 mmol), catalyst (62.5 mg, b31.3 mg), H2 balloon (purge), temp. (60 ◦C), time (6 h, b12 h). cConcentration
of the monometallic X50/support is 0.47 mmol g−1 whereas that of the X50Y50/support is totally 0.94 mmol g−1 with a molar ratio of X/Y = 50/50.
the Ministry of Education, Culture, Sports, Science and Technology
(MEXT), Japan.
activities (entries 5 and 6), however, their activities were not supe-
rior to that of Pd50/C (entry 1). Several peaks in GC chromatograph
were appeared in the range of retention time at 20–26 min in cases
of Pd50/SiO2 and Pd50/␤-zeolite. Pd50/Amberlyst-15 and Pd50/␤-
zeolite also indicated a new single peak at 18.1 min due to the
2,5-hexanedione formation which was produced by the hydrol-
ysis reaction of DMF, though this product was not observed in
Pd50/C. Therefore, acidic supports may cause the side pathways,
ex. a ring-opening of DMF. It was reported that the ring open-
ing by hydrolytic cleavage of the furanic C O bond is typically
promoted with Brønsted acid sites of aluminosilicates and ion-
exchange resins [27,28]. As shown in Table 2, the differences of
acidity seemed to influence the catalytic activities. It indicated
that a co-existence of strong acidic media such as HCl and H2SO4
was required for a selective hydrogenation of HMF toward DMF,
and the HCl was found to be the best acid for a high efficiency
(entries 1–4). Lange et al. reported that not only the strong acid-
ity (pKa < −5) but also chloride ions were needed for a selective
and an effective hydrogenation of furanic compounds because the
HCl acted as the co-catalyst which prevented an undesired ring-
hydrogenation and enhanced the hydrogenolysis step by formation
of an active chlorinated intermediate through nucleophilic substi-
tution on the alcohol groups [29]. Combined with these results
discussed in Tables 1 and 2, the PdAu/C catalyst in the presence
of HCl acidic media promoted an efficient hydrogenation of HMF
to DMF even under mild reaction conditions (60 ◦C, atmospheric
H2).
3. Results and discussion
3.1. Screening of Pd-based heterogeneous catalysts for selective
hydrogenation of HMF to DMF
Table 1 shows the activities for hydrogenation of HMF to DMF
over Pd, Au, Pt and/or Ru-supported carbon catalysts. Ru50/C, Pt50/C
and Au50/C were inactive for DMF synthesis (entries 5–7) whereas
Pd50/C gave a good activity affording to DMF (entry 1) among
monometallic catalysts. To accelerate the catalytic activity of Pd50/C
catalyst, effect of second elements such as Au, Pt or Ru into Pd cata-
lyst was also examined. Additional Pt and Ru decreased the activity
of Pd50/C (>99% yield) to 93% and 62%, respectively (entries 3 and
4). While the Au combined with Pd atoms, Pd50Au50/C, showed a
positive effect for the activity, it was observed that 66% yield over
Pd50/C changed to 96% over Pd50Au50/C in 1.0 mmol scale reaction
(entries 1 and 2).
Effects of acidic co-catalyst and support for the synthesis of DMF
from HMF hydrogenation reaction by Pd-supported catalysts were
investigated (Table 2). Instead of carbon support, acidic compounds
such as SiO2, ␤-zeolite, Amberlyst-15 and ␣-Al2O3 were also com-
pared as a support for Pd (Each morphologies were shown in Fig.
S1). Pd50/Amberlyst-15 and Pd50/␣-Al2O3 showed small activities
(entries 7 and 8) whereas Pd50/SiO2 and Pd50/␤-zeolite gave mild
Table 2
Effect of support and acid for the synthesis of DMF via hydrogenation of HMF over the Pd-supported catalysts.a
Entry
Catalystc
Conversion of HMF/%
91
Yield/%
5-MF
MFA
DMF
1
2
3
4
5
6
7
8
9
Pd50/C
4
0
0
11
0
6
66
50
6
Pd50/Cb
>99
Pd50/Cc
66
41
83
91
60
63
38
Pd50/Cd
<1
5
3
Pd50/SiO2
Pd50/␤-zeolite
Pd50/Amberlyst-15
Pd50/␣-Al2O3
Blank
6
5
7
7
0
12
0
0
40
29
8
6
6
0
0
aReaction conditions: HMF (1.0 mmol), THF (10 ml), HCl (0.17 mmol), catalyst (31.3 mg, Pd = 0.47 mmol g−1), H2 balloon (purge), temp. (60 ◦C), time (12 h). Instead of HCl,
bH2SO4 (0.085 mmol), cFA (0.17 mmol) and dwithout acid were used for each runs, respectively.

92
S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
Table 3
Synthesis of DMF via hydrogenation of HMF over the PdxAuy/C catalysts with various Pd/Au molar ratios.a
Entry
Catalyst
Conversion of HMF/%
Yield/%
5-MF
MFA
DMF
1
2
3
4
5
6
7
8
9
Pd100/C
89
96
95
4
4
5
12
7
5
60
79
Pd70Au30/C
Pd60Au40/C
Pd50Au50/C
Pd40Au60/C
Pd30Au70/C
Au100/C
88
>99, 81b
>99, 77b
96
2, 4b
2, 4b
2, 3b
3
96, 55b
4, 4b
94, 41b
6
0
4
0
6
0
84
0
66
0
51
0
32
91
51
81
0
11
0
8
0
Pd50/Cc
Au50/Cc
10
11
Pd50/C + Au50/Cd
Blank
38
aReaction conditions: HMF (1.0 mmol, b3.0 mmol), THF (10 ml), HCl (0.17 mmol), catalyst (31.3 mg, Pd + Au = 0.94 mmol g−1
temp. (60 ◦C), time (12 h, b36 h), dphysical mixture of Pd50/C (15.5 mg) and Au50/C (15.5 mg).
,
cPd or Au = 0.47 mmol g−1), H2 balloon (purge),
3.2. Hydrogenation of HMF to DMF over PdxAuy/C catalysts
intermediate species. BHMF was also detected but it was trace
amount in all steps (not plotted in Fig. 1). Additionally, the
over-hydrogenation products such as DHTHF, DMTHF and the 2,5-
hexanedione were not detected (0% yield in all step). Therefore,
the hydrogenation of HMF to DMF under our experimental con-
ditions mainly involves two-step; the reduction of the aldehyde
group to alcohol group, and the hydrogenolysis of the hydroxyl
group (Scheme 1). It was indicated that the rate of DMF production
over PdAu/C was faster than Pd/C.
In order to get more insights of catalytic enhancement
phenomena over PdAu/C catalyst, time-courses of hydrogenation
of 5-MF and MFA toward DMF were also monitored (Figs. 2 and 3).
The transformation of 5-MF to DMF occurred through the MFA
formation in both cases. Interestingly, the life-time of MFA over
PdAu/C seemed to be shorter than that over Pd/C (Fig. 2). Follow-
ing to the result in Fig. 3, the reaction from MFA to DMF became
faster in the case of PdAu/C. Therefore, the hydrogenolysis step,
converting MFA to DMF, during hydrogenation of HMF toward DMF
seems to be enhanced by the addition of Au into Pd atoms. That is
because the yield of MFA was smaller than 5-MF during the reaction
over PdAu/C though Pd/C remained MFA in Fig. 1. In comparison
To study the contribution of Pd/Al molar ratio, the hydrogena-
tion of HMF to DMF was examined over PdxAuy/C catalysts prepared
with various Pd/Au molar ratio (x/y) (Table 3). It was observed that
a gradual increase of Au contents in PdxAuy/C induced a gradual
increase of activity at y = 0–50 (entries 1–4), thereafter, the activi-
ties were decreasing at y = 50–100 (entries 5–7). Interestingly, the
physical mixture of Pd50/C and Au50/C showed a lower activity (51%
yield, entry 10) than that of bimetallic Pd50Au50/C (96% yield, entry
4). The Pd50Au50/C also presented 55% yield of DMF in 3 mmol scale
(entry 4b), and it was very notable in comparison with previous
reports because the significant decrease of metal amount (Pd and/or
Au) was achieved with a good activity (see Table S1). These results
indicated that the as-prepared PdAu/C catalysts had some interac-
tions between Pd and Au atoms which produced different effects
on active sites from monometallic Pd/C and Au/C.
Time-courses of the hydrogenation of HMF toward DMF over
monometallic Pd/C and bimetallic PdAu/C catalysts were shown
in Fig. 1. In both cases, the production of DMF gradually pro-
ceeded with consuming HMF. 5-MF and MFA were obtained as the
Fig. 1. Time course of the hydrogenation of HMF toward HMF over (A) Pd50/C and (B) Pd40Au60/C catalysts. Reaction conditions; HMF (1.0 mmol), THF (10 ml), HCl (0.17 mmol),
catalyst (31.3 mg), H2 balloon (purge), 60 ◦C.

S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
93
Scheme 1. Reaction pathway for the hydrogenation of HMF toward DMF.
Fig. 2. Time course of the hydrogenation of 5-MF into DMF over (A) Pd50/C and (B) Pd40Au60/C catalysts. Reaction conditions; substrate (0.5 mmol), THF (10 ml), HCl (0.17 mmol),
catalyst (31.3 mg), H2 balloon (purge), 60 ◦C.
between Figs. 2 and 3, it was also ascertained that the consump-
tion rate of 5-MF was slower than MFA in both cases of Pd/C and
PdAu/C. Furthermore, on the other hand, focusing on the produc-
tion rate of DMF gave the information that DMF production from
5-MF was faster than that from MFA; i.e. the production of DMF
through the hydrogenation of MFA seems to be saturated in Fig. 3.
We believed that excess number of MFA adsorbed on the surface of
Pd may inhibit the formation of DMF. [30]
Fig. 3. Time course of the hydrogenation of MFA into DMF over (A) Pd50/C and (B) Pd40Au60/C catalysts. Reaction conditions; substrate (0.5 mmol), THF (10 ml), HCl (0.17 mmol),
catalyst (31.3 mg), H2 balloon (purge), 60 ◦C.

94
S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
Fig. 4. Analyses of XRD patterns of the PdxAuy/C catalysts. (A) XRD patterns in 30–90◦ and (B) emphasized detail in 36–44◦ with the Gaussian fitting (red) of the diffraction
derived from PdAu alloy, and (C) estimated values of lattice parameter for PdxAuy/C catalyst as a function of Au contents (y). (For interpretation of the references to color in
this legend, the reader is referred to the web version of the article.)
3.3. XRD patterns of PdxAuy/C catalysts
Pd70Au30/C, Pd60Au40/C, Pd50Au50/C, Pd40Au60/C and Pd30Au70/C,
respectively.
XRD patterns of PdxAuy/C are shown in Fig. 4(A). Both Pd100/C
and Au100/C showed the characteristic patterns derived from the
face-centered cubic (fcc) structure, the reflection peaks were corre-
sponding to the phases for (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2)
from lower to larger angle. The Pd(1 1 1) plane and the Au(1 1 1)
plane were located at 2Â = 40.06◦ and 2Â = 38.19◦, respectively, and
3.4. XPS spectra of PdxAuy/C catalysts
Fig. 5 illustrated the XPS spectra of PdxAuy/C catalysts in the
range of 81–90 eV ascribed to Au 4f regions. From a viewpoint of
the spectrum of Au100/C, both peaks assigned to Au 4f5/2 around at
88 eV and 4f7/2 around at 84 eV showed a gradual shifts to lower
side in binding energy with increasing the contents of Pd atoms
in PdxAuy/C (Fig. 5(B)). This tendency indicated that the Au atoms
gained electrons from Pd atoms by alloying, which is consistent
with the higher Pauling’s electronegativity of Au (2.54) relative
to Pd (2.20), and that was typical in Pd–Au alloying interaction
[16,23,31,32]. On the other hands, the XPS spectra derived from Pd
3d5/2 around at 335 eV and 3d3/2 around at 340 eV were hard to dis-
tinguish there alterations against Au/Pd molar ratio (x/y) because
some of them had low intensities and overlapping with the peak
of Au 4d5/2 (Fig. S2). Based on the charge compensation concept,
Pd atoms in the PdxAuy/C may have electronic poor states which
strongly contributed to the enhancement of activity for hydrogena-
tion reaction of HMF. Previously, importance of Pd0/Pd2+ ratio for a
direct synthesis of H2O2 from H2 and O2 through the H H bond dis-
sociation and the O O bond retaining steps over Au@Pd core–shell
nanocatalyst have been reported [33,34]. Thus, formation of the
electronically poor states in Pd atoms in PdxAuy/C will play a cru-
cial role in the activation of H2 molecules and/or the enhancement
of reaction pathway of hydrogenation, leading to the high activity
[30].
˚
the lattice parameters were calculated as 3.90 A from Pd(1 1 1) and
˚
4.08 A from Au(1 1 1) planes. On the other hand, PdxAuy/C exhib-
ited two patterns on diffractions, one appeared the same positions
with Au100/C (2Â = 38.19◦) and another located at between that of
Au100/C and Pd100/C (2Â = 39.86–39.20◦). Fig. 4(B) is the emphasized
area in 2Â = 36–44◦ of Fig. 4(A) attributing the (1 1 1) diffraction of
fcc structure. It was reported that the Au@Pd core–shell particles
indicated such a humped patterns in this area though the former
peak was broader [23]. Thus, the former strong peaks observed
at 38.19◦ supposedly suggested the presence of isolated Au parti-
cles in as-prepared PdxAuy/C though which seems to be inactive
for selective hydrogenation of HMF to DMF. Moreover, it was
observed that the shoulder peaks were gradual shifts to higher
angle position with increasing Au contents in PdxAuy/C, though all
peaks still sited between the (1 1 1) peaks of Au(1 1 1) plane and
Pd(1 1 1) plane. These predicted the presence of PdAu alloy struc-
ture in PdxAuy/C. The lattice parameters estimated by the shifting
(1 1 1) diffraction peaks of PdxAuy/C were plotted as a function
of the introduction amount of Au(y) shown in Fig. 4(c). The cor-
relations between lattice parameter and Au content in PdxAuy/C
seems to be linear (R2 = 0.98) but not directly followed with the
Vegard’s law (the dashed line in Fig. 4(C)), the lattice param-
eters scale linear between those values for neat Pd and Au by
the ratio of the PdAu alloy composition. That is due to contri-
butions of the isolated Au particles in PdxAuy/C, which expected
by the strong diffractions attributed to Au particle detected in
XRD patterns, included in the horizontal axis in Fig. 4(C). On the
basis of the hypothesis that the PdxAuy/C were composed by the
isolated Au particles and the homogeneous PdAu alloy particles
belonging to the Vegard’s law, the compositions of PdAu alloy in
PdxAuy/C were estimated by the dash line and the lattice param-
eter as follows; Pd/Au = 90/10, 81/19, 73/27, 68/32 and 54/46 for
3.5. XAFS analyses of PdxAuy/C catalysts
XAFS spectra were measured to further clarify the existence of
PdAu alloy on the PdxAuy/C. Fig. 6(A) shows Au L3-edge XANES spec-
tra of PdxAuy/C. The white-line (WL) feature in Au L3-edge XANES
related to the electron transition from 2p to 5d electronic states, i.e.
a higher density in 5d state makes a lower intensity in the WL fea-
ture. As shown in Fig. 5(A), the WL intensities of PdxAuy/C observed
around at 11.92 keV were lower than that of Au foil regardless of
Au/Pd molar ratio (x/y), it indicated that the negatively charged Au

S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
95
Fig. 5. (A) XPS spectra of the PdxAuy/C catalysts around at Au 4f states and (B) plots of the peak positions in Au 4f5/2 and 4f7/2 as function of Pd contents (x).
˚
5d states were appeared in PdxAuy/C via charge transfer from Pd
to Au. Thus, it strongly suggested that there were electronic inter-
actions between Au and Pd atoms in PdxAuy/C. Pd K-edge XANES
spectra of PdxAuy/C showed the two-humped features similar to
the Pd foil (Fig. S3).
in Fig. 7 [35]. The peak in the range of 2–3 A gradually splitted into
the double with increasing the Au contents in PdxAuy/C [36]. This
behavior proposed that the frequency of Pd–Au coordination and/or
nanostructure of PdAu alloy will be influenced by Au/Pd molar ratio.
With comparing of |FT|s in both Au L3-edge and Pd K-edge, the local
structure around Au atoms were much influenced by Pd addition,
however the inference of Au progressively increased into Pd atoms.
In the previous literatures, it has been reported that the PdAu
bimetallic particle prepared by a simultaneous reduction method
at low temperature easily formed Au@Pd core–shell structure due
to differences in the reduction potential, E0(Pd2+/Pd0) = +0.63 V,
E0(Au3+/Au0) = +1.0 V v.s. NHE, and the formation energy [39,40].
Especially, Toshima et al. proposed that the possible nano-
structures were cluster-in-cluster, Au@Pd core–shell or random
models in the case of Au/Pd = 1/1 with CNAu–Pd = CNPd–Au, while
Au@Pd core–shell or random models in the case of Au/Pd = 1/4
with CNAu–Pd > CNPd–Au [41,42]. Thus, a simple simultaneous EXAFS
curve-fitting procedure using a McKale method [43] was conducted
The |FT|s of Au L3-edge EXAFS spectra were also investigated
in Fig. 6(B) [35]. The Au foil and Au2O3 exhibited a single peak
due to the Au–Au and Au–O coordination with a fitting parame-
˚
ter of 2.84 and 2.01 A, respectively [36]. While, PdxAuy/C showed
˚
intense doublet peaks at 2.11 and 2.95 A in |FT|s. This splitting arises
from the interference between Au–Au and Au–Pd in EXAFS oscil-
lations because the phase shift of Au–Au and Au–Pd are ␲ radian
different from each other, though their corresponding distances
are expected to be the same. In other word, the appearance of this
feature is an obvious support for the presence of Au–Pd interac-
tions in PdxAuy/C. These double peaks were similar to the previous
reports about Au@Pd core–shell and random AuPd alloy structure
[18,23,37,38]. The |FT|s of Pd K-edge EXAFS spectra also described
Fig. 6. (A) Au L3-edge XANES spectra and (B) |FT|s of Au L3-edge EXAFS spectra of the PdxAuy/C catalysts, Au foil and Au2O3 references (black, green and blue solid lines) with
the McKale fitting (red solid lines). (For interpretation of the references to color in this legend, the reader is referred to the web version of the article.)

96
S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
random whereas Pd70Au30/C and Pd30Au70/C were core–shell
and/or random. Though further EXAFS investigations are required
to determine the detailed information about their nanostructures,
we strongly confirmed that the interaction between Pd and Au
atoms derived from alloying assigned to a significant catalysis over
PdxAuy/C for hydrogenation of HMF toward DMF.
3.6. TEM analyses of Pd50Au50/C catalyst
To further investigate the nanoscale morphology in PdxAuy/C,
TEM and STEM-HAADF image with an EDS elemental analysis were
measured for Pd50Au50/C. Fig. 8(A) indicated that a lot of nanopar-
ticles with average diameter around at 7.5 nm were formed on
carbon support [44]. The STEM-HAADF image a particle of 7.1 nm
shown in Fig. 8(B) had no contrast. When the AuPd/C was composed
by cluster-in-cluster and/or core–shell, the STEM-HAADF image
showed bright dark contrast since the heavier element (Au; atomic
number, Z = 79) gave a brighter image than the lighter (Pd; Z = 46)
element. Thus, as-synthesized PdAu/C was composed by homo-
geneous and/or random PdAu alloy. This suggestion is strongly
supported by EDS elemental analysis of a particle shown in Fig. 8(C),
which exhibited the presence of homogeneously-mixed Pd and Au
atoms on the Pd50Au50/C. Moreover, the chemical and physical
properties in Pd50Au50/C were scarcely changed before and after
reaction analyzed by XRD, XPS and TEM techniques (see Figs. S7
and S8).
Fig. 7. |FT|s of Pd K-edge EXAFS spectra of the PdxAuy/C catalysts, Pd foil and PdO
references (black solid lines) with the McKale fitting (red solid lines). (For interpre-
tation of the references to color in this legend, the reader is referred to the web
version of the article.)
3.7. Direct synthesis of DMF from fructose and a further
application for furfural hydrogenation reaction
We further investigated the synthesis of DMF from fructose and
tagatose, a dehydration reaction of these ketoses proceeded over
acid catalyst and it served HMF [45]. In our catalytic system, the
co-catalyst of HCl will act as the acidic catalyst for dehydration pro-
cess. In fact, 40% and 10% yields of DMF were produced over PdAu/C
catalyst from fructose and tagatose, respectively, in the same reac-
tion conditions (Scheme 2) [46]. Dehydration of fructose was easier
than tagatose became the difference position of hydroxyl func-
tional group in C-4 (epimer), thus the yield of DMF from tagatose
seems to be lower than that from fructose [47]. Moreover, the
hydrogenation of furfural, one of the related reactions [48–50], was
also tested over PdAu/C catalyst, it produced 42% yield of 2-MF at
60 ◦C [51]. These advanced and derivative reactions of furanic com-
pounds are currently ongoing to reveal the novel catalysis of the
hydrogenation reaction over the present bimetallic PdAu/C catalyst
(Scheme 3).
for bimetallic PdxAuy/C to further challenge for distinguishing
their nanostructures. The estimated |FT| patterns were shown in
Figs. 6(B) and 7 by red lines, and each fitting parameters were listed
in Table S2.
Focusing on the coordination numbers (CN), three bimetal-
lic catalysts of Pd60Au40/C, Pd50Au50/C and Pd40Au60/C showed
almost equal (2.3–2.9) between CNAu–Pd and CNPd–Au. For a homo-
geneous Pd–Au alloy, the ratio of CNPd–Au/CNPd–Pd should be equal
to the Pd/Au molar ratio, however, the CNPd–Au/CNPd–Pd values
among these were approximately 0.5 which were smaller than
Pd/Au molar ratios estimated by XRD. On the other hand, in
the cases of Pd70Au30/C and Pd30Au70/C, CNAu–Pd (4.1) > CNPd–Au
(1.6) and CNPd–Au (4.1) > CNAu–Pd (1.5) were estimated. Conse-
quently, we supposed that Pd60Au40/C, Pd50Au50/C and Pd40Au60/C
were composed by cluster-in-cluster, Au@Pd core–shell and/or
Fig. 8. (A) TEM and (B, C) STEM-HAADF images with an EDS elemental mapping analysis of Pd50Au50/C.

S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
97
Scheme 2. Direct synthesis of DMF from ketoses.
References
[1] Annual Energy Outlook 2013 (AEO2013), the U.S. Energy Information Admin-
istration (EIA), Washington, DC, 2013.
[2] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044.
[3] D.M. Alonso, J.Q. Bond, J.A. Dumesic, Green Chem. 12 (2010) 1493.
[4] K. Wilson, A.F. Lee (Eds.), Handbook of Heterogeneous Catalysts for Clean
Technology: Design, Analysis and Application, first ed., Wiley-VCH, Weinheim,
Germany, 2013.
Scheme 3. Hydrogenation of furfural to 2-MF.
[5] R.-J. van Putten, J.C. van der Waal, E. de Jong, C.B. Rasrendra, H.J. Heeres, J.G. de
Vries, Chem. Rev. 113 (2013) 1499.
[6] Y. Roman-Leshkov, C.J. Barrett, Z.Y. Liu, J.A. Dumesic, Nature 447 (2009) 982.
[7] M. Chidambaram, A.T. Bell, Green Chem. 12 (2010) 1253.
[8] J. Ohyama, A. Esaki, Y. Yamamoto, S. Arai, A. Satsuma, RSC Adv. 3 (2013) 1033.
[9] R. Alamillo, M. Tucker, M. Chia, Y. Pagan-Torresa, J. Dumesic, Green Chem. 14
(2012) 1413.
[10] W. Yang, A. Sen, ChemSusChem 3 (2010) 597.
[11] G.C.A. Luijkx, N.P.M. Huck, F. van Rantwijk, L. Maat, H. van Bekkum, Heterocy-
cles 77 (2009) 1037.
[12] T. Thananatthanachon, T.B. Rauchfuss, Angew. Chem. Int. Ed. 49 (2010) 6616.
[13] T.B. Rauchfuss, T. Thananatthanachon, US Patent Application, US
2011/02638801 A1.
[14] S. De, S. Dutta, B. Saha, ChemSusChem 5 (2012) 1826.
[15] D.I. Enache, J.K. Edwards, P. Landon, B. Solsona-Espriu, A.A. Herzing, M. Watan-
abe, C.J. Kiely, D.W. Knight, G.J. Hutchings, Science 311 (2006) 362.
[16] S. Nishimura, Y. Yakita, M. Katayama, K. Higashimine, K. Ebitani, Catal. Sci.
Technol. 3 (2013) 351.
[17] A. Villa, M. Schiavoni, S. Campisi, G.M. Veith, L. Prati, ChemSusChem 6 (2013)
609.
[18] T. Balcha, J.R. Strobl, C. Fowler, P. Dash, R.W.J. Scott, ACS Catal. 1 (2011)
425.
4. Conclusions
The PdAu/C catalyst prepared with a simultaneous reduction
method exhibited a higher activity than that of Pd/C, Au/C and
physical mixture of Au/C and Pd/C for hydrogenation of HMF to
DMF under ambient conditions. The XRD patterns indicated that
the catalysts were composed by isolated Au particles and PdAu
alloy particles. XPS peaks attributed to Au 4f states and Au L3-edge
XANES spectra suggested the negatively charged Au atoms were
produced with the co-existence of Pd atoms in PdAu/C. EXAFS anal-
yses in both Au L3-edge and Pd K-edge also supported the direct
Au–Pd interaction in the PdAu/C. According to these results, we
concluded that the AuPd alloy particles stabilized on carbon sup-
port possessed a significant activity for the hydrogenation of HMF
toward DMF using an atmospheric H2 gas. Additionally, the PdAu/C
exhibited good activities for a direct synthesis of DMF from ketoses
in the presence of HCl acidic media and hydrogenation of furfural
toward 2-MF.
[19] J.K. Edwards, B. Solsona, E. Ntainjua, A.F. Carley, A.A. Herzing, C.J. Kiely, G.J.
Hutchings, Science 323 (2009) 1037.
[20] J.K. Edwards, E. Ntainjua, A.F. Carley, A.A. Herzing, C.J. Kiely, G.J. Hutchings,
Angew. Chem. Int. Ed. 48 (2009) 8512.
[21] P.-P. Fang, A. Jutand, Z.-Q. Tian, C. Amatore, Angew. Chem. Int. Ed. 50 (2011)
12184.
[22] J. Xu, A.R. Wilson, A.R. Rathmell, J. Howe, M. Chi, B.J. Wiley, Nano Lett. 5 (2011)
6119.
Acknowledgements
[23] X. Yuan, G. Sun, H. Asakura, T. Tanaka, X. Chen, Y. Yuan, G. Laurenczy, Y. Kou,
P.J. Dyson, N. Yan, Chem. Eur. J. 19 (2013) 1227.
[24] A surface area was analyzed by N2 adsorption at 77 K with BELSORP-max.
[25] M. Brandalise, M.M. Tusi, R.M. Piasentin, M.C. dos Santos, E.V. Spinace, A.O.
Neto, Int. J. Electrochem. Sci. 7 (2012) 9609.
[26] Supposed to be decomposed during GC-FID analysis.
[27] N. Nikbin, S. Caratzoulas, D.G. Vlachos, ChemSusChem (2013) (in press).
[28] L. Bui, H. Luo, W.R. Gunther, Y. Roman-Leshkov, Angew. Chem. Int. Ed. 52 (2013)
8022.
[29] J.-P. Lange, E. van der Heido, J. van Buijtenen, R. Price, ChemSusChem 5 (2012)
150.
[30] Reactivities of the adsorbed aldehyde and hydroxylalkyl groups in furanic com-
pounds and/or H2 species on Pd surface were supposedly influenced by the Pd
condition. Further investigations of the electronic state of Pd atoms and their
contribution onto the reaction mechanism are now on going.
[31] F. Liu, D. Wechsler, P. Zhang, Chem. Phys. Lett. 461 (2008) 254.
[32] P.A.P. Nascente, S.G.C. de Castro, R. Landers, G.G. Kleiman, Phys. Rev. B 43 (1991)
4659.
[33] M. Piccinini, E. Ntainjua, J.K. Edwards, A.F. Carley, J.A. Moulijn, G.J. Hutchings,
Phys. Chem. Chem. Phys. 12 (2010) 2488.
[34] J. Pritchard, M. Piccinini, R. Tiruvalam, Q. He, N. Dimitratos, J.A. Lopez-Sanchez,
D.J. Morgan, A.F. Carley, J.K. Edwards, C.J. Kiely, G.J. Hutchings, Catal. Sci. Tech-
nol. 3 (2013) 308.
The work was supported by a Grant-in-Aid for Young Sci-
entists (B) (No. 25820392) and Scientific Research (C) (No.
25420825) of the Ministry of Education, Culture, Sports, Sci-
ence and Technology (MEXT), Japan. Au L3-edge and Pd K-edge
XAFS measurements were performed at a BL01B1 of SPring-8
by the approval of the Japan Synchrotron Radiation Research
Institute (JASRI) with proposal No. 2012B1610 and a BL5S1
of Aichi Synchrotron Radiation Research Center (2nd period,
2013). The STEM-HAADF image with a EDS elemental mapping
analysis was conducted in JAIST supported by Nanotechnology
Platform Program of the Ministry of Education, Culture, Sports,
Science and Technology (MEXT), Japan. We appreciate Dr. Kohichi
Higashimine (Center for Nano Materials and Technology, JAIST) for
his great help of STEM-HAADF and EDS elemental mapping analy-
sis.
Appendix A. Supplementary data
[35] The corresponding EXAFS spectra were shown in Fig. S4.
[36] Monometalic Au/C or Pd/C indicated a similar peak to Au foil or Pd foil in |FT|
shown in Fig. S5.
[37] P. Dash, T. Bond, C. Fowler, W. Hou, N. Coombs, R.W.J. Scott, J. Phys. Chem. C
113 (2009) 12719.
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.cattod.
2013.10.012.

98
S. Nishimura et al. / Catalysis Today 232 (2014) 89–98
[38] M.R. Knecht, M.G. Weir, A.I. Frenkel, R.M. Crooks, Chem. Mater. 20 (2008) 1019.
[39] S. Okada, K. Fujiwara, T. Kamegawa, K. Mori, H. Yamashita, Bull. Chem. Soc. Jpn.
85 (2012) 1057.
[40] H.B. Liu, U. Pal, A. Medina, C. Maldonado, J.A. Ascencio, Phys. Rev. B 71 (2005)
075403.
[41] N. Toshima, M. Harada, Y. Yamazaki, K. Asakura, J. Phys. Chem. 96 (1992) 9927.
[42] N. Toshima, T. Yonezawa, New J. Chem. 22 (1998) 1179.
[43] A.G. McKale, B.W. Veal, A.P. Paulikas, S.-K. Chan, G.S. Knapp, J. Am. Chem. Soc.
110 (1988) 3763.
[46] Reaction conditions; ketose (0.5 mmol), THF (10 ml), HCl (0.17 mmol),
Pd40Au60/C (31.3 mg), H2 balloon (purge), 60 ◦C, 12 h.
[47] Direct synthesis of DMF from aldoses such as glucose and galactose using the
same reaction conditions did not proceed (0% yields), because the dehydration
reaction of aldoses needs a higher temperature than that of ketoses, normally.
[48] Y. Nakagawa, H. Nakazawa, H. Watanabe, K. Tomishige, ChemCatChem 4 (2012)
1791.
[49] S. Sitthisa, T. Pham, T. Prasomsri, T. Sooknoi, R.G. Mallinson, D.E. Resasco, J.
Catal. 280 (2011) 17.
[44] In some cases, the agglomerates shown in Fig. S6 were also observed but we
ignored these particles.
[45] J. Tutejya, S. Nishimura, K. Ebitani, Bull. Chem. Soc. Jpn. 85 (2012) 275.
[50] S. Sitthisa, D.E. Resasco, Catal. Lett. 141 (2011) 784.
[51] Reaction conditions; furfural (1.0 mmol), THF (10 ml), HCl (0.17 mmol),
Pd40Au60/C (31.3 mg), H2 balloon (purge), 60 ◦C, 12 h.