Catalytic Transfer Hydrogenation of Furfural to 2-Methylfuran and 2-Methyltetrahydrofuran over Bimetallic Copper–Palladium Catalysts

Article abstract of DOI:10.1002/cssc.201601122

The catalytic transfer hydrogenation of furfural to the fuel additives 2-methylfuran (2-MF) and 2-methyltetrahydrofuran (2-MTHF) was investigated over various bimetallic catalysts in the presence of the hydrogen donor 2-propanol. Of all the as-prepared catalysts, bimetallic Cu-Pd catalysts showed the highest catalytic activities towards the formation of 2-MF and 2-MTHF with a total yield of up to 83.9 % yield at 220 °C in 4 h. By modifying the Pd ratios in the Cu-Pd catalyst, 2-MF or 2-MTHF could be obtained selectively as the prevailing product. The other reaction conditions also had a great influence on the product distribution. Mechanistic studies by reaction monitoring and intermediate conversion revealed that the reaction proceeded mainly through the hydrogenation of furfural to furfuryl alcohol, which was followed by deoxygenation to 2-MF in parallel to deoxygenation/ring hydrogenation to 2-MTHF. Finally, the catalyst showed a high reactivity and stability in five catalyst recycling runs, which represents a significant step forward toward the catalytic transfer hydrogenation of furfural.

Full text of DOI:10.1002/cssc.201601122

DOI: 10.1002/cssc.201601122
Full Papers
Catalytic Transfer Hydrogenation of Furfural to
2-Methylfuran and 2-Methyltetrahydrofuran over
Bimetallic Copper–Palladium Catalysts
Xin Chang+, An-Feng Liu+, Bo Cai, Jin-Yue Luo, Hui Pan, and Yao-Bing Huang*[a]
The catalytic transfer hydrogenation of furfural to the fuel ad-
ditives 2-methylfuran (2-MF) and 2-methyltetrahydrofuran
(2-MTHF) was investigated over various bimetallic catalysts in
the presence of the hydrogen donor 2-propanol. Of all the
as-prepared catalysts, bimetallic Cu-Pd catalysts showed the
highest catalytic activities towards the formation of 2-MF and
2-MTHF with a total yield of up to 83.9% yield at 2208C in 4 h.
By modifying the Pd ratios in the Cu-Pd catalyst, 2-MF or 2-
MTHF could be obtained selectively as the prevailing product.
The other reaction conditions also had a great influence on
the product distribution. Mechanistic studies by reaction moni-
toring and intermediate conversion revealed that the reaction
proceeded mainly through the hydrogenation of furfural to
furfuryl alcohol, which was followed by deoxygenation to 2-MF
in parallel to deoxygenation/ring hydrogenation to 2-MTHF. Fi-
nally, the catalyst showed a high reactivity and stability in five
catalyst recycling runs, which represents a significant step for-
ward toward the catalytic transfer hydrogenation of furfural.
Introduction
The effective conversion of lignocellusic biomass into sustaina-
ble chemicals and fuels has gained much attention during the
past few decades. As a result of the oxygen-rich nature of bio-
mass-derived molecules, the selective removal of the oxygen-
ated groups has become a primary challenge for the utilization
of these molecules.[1] A variety of chemical reactions, which in-
clude dehydration, dehydrogenation, hydrogenation, hydroly-
sis, and hydrogenolysis, have been reported widely as versatile
tools for the conversion,[2] for example, the conversion of
oxygen-rich resources to platform molecules and their further
upgrade to the desired value-added products. A representative
example is the conversion of hemicellulose into, for example,
2-methylfuran (2-MF) and 2-methyltetrahydrofuran (2-MTHF)
via the platform molecule furfural.[3] 2-MF and 2-MTHF are
good liquid fuel additives with a high energy density, boiling
point, and octane number and can be partly blended into gas-
oline for engines.[3] Additionally, they have numerous applica-
tions as solvents in organic/polymer chemistry and as chemical
intermediates for drugs.[3] Although the dehydration of hemi-
cellulose to furfural has been industrialized over various cata-
lysts,[4] the conversion of furfural to 2-MF or 2-MTHF is still
under development.
tions as shown in the reaction network displayed in Scheme 1.
Therefore, catalytic sites with excellent hydrogenation/hydro-
genolysis abilities would be potential candidates for this con-
version. To date, a variety of catalysts based on Cu,[5] Mo,[6]
Pd,[7] Ru,[8] Pt,[9] Ni,[10] and others[11] have been applied success-
Scheme 1. Reaction network of the hydrogenolysis of furfural.
The hydrodeoxygenation (HDO) of furfural to 2-MF or
2-MTHF occurs mainly through the hydrogenation and hydro-
genolysis of the C=O bonds in parallel to some tandem reac-
fully for this conversion in the presence of molecular hydro-
gen. Recently, the use of a hydrogen donor to replace molecu-
lar hydrogen has emerged as a green and cost-effective
method (i.e., catalytic transfer hydrogenation; CTH) for the re-
ductive upgrading of biomass-derived molecules such as levu-
linic acid/esters[12] and 5-hydroxymethylfurfural.[13] Numerous
studies have been reported on the development of new effi-
cient catalysts for these conversions. However, very few cata-
lysts have been developed for the CTH of furfural. For example,
Hermans et al.[14] used a Pd/Fe2O3 catalyst for the CTH of furfu-
ral in the presence of 2-propanol, however, only 40% yield of
[a] X. Chang,+ A.-F. Liu,+ B. Cai, Dr. J.-Y. Luo, Dr. H. Pan, Dr. Y.-B. Huang
Department of Chemical Engineering
Nanjing Forestry University
Longpan Road, Nanjing (China)
E-mail: hyb123@mail.ustc.edu.cn
[+] These authors equally contributed to this work.
Supporting Information for this article can be found under:
http://dx.doi.org/10.1002/cssc.201601122.
ChemSusChem 2016, 9, 1 – 9
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
2-MF+2-MTHF was obtained. Vlachos et al.[15] reported the use
of a Ru/C catalyst for the CTH of furfural in 2-propanol, which
provided a higher 2-MF yield of 61% at 1808C after 10 h. Later,
they demonstrated that a partially oxidized Ru/C (Ru/RuOx/C)
possessed a high reactivity toward the CTH of furfural because
of the introduction of RuOx Lewis acid sites and yielded 76%
2-MF under the same reaction conditions.[16] Despite these ad-
vances, more efficient systems for the CTH of furfural are still
highly desired.
For 3Pd/ZrO2, no characteristic peaks that correspond to Pd
species were detected in the pattern, probably because of the
superposition with the strong diffraction peaks of the ZrO2
phase and the highly dispersed Pd particles, which was report-
ed previously.[18] For 10Cu/ZrO2, three new peaks were found
at 2q=43.3, 50.5, and 74.28 that correspond to the (111),
(200), and (220) planes of Cu0, respectively. If Pd was coimmo-
bilized on the support, the diffraction peaks became much
weaker. For 10Cu-1Pd/ZrO2, the diffraction peak of Cu0 was still
found but with a low intensity. A new weak peak at 42.98 was
found. This new peak was between the (111) reflections of
pure Cu (2q=43.38) and Pd (2q=40.18), which indicates the
successful distribution of the Pd atoms in the Cu lattice by the
formed Cu-Pd alloys and leads to a change of the original Cuꢀ
Cu lattice spacing and a shift of the diffraction peak.[19] This in-
dicates that both Cu0 and Cu-Pd alloy were present in the 10:1
Cu/Pd ratio catalyst. However, the diffraction peak of Cu disap-
peared in the 10:3 and 10:5 catalysts (magnified images are
presented in the Supporting Information). The diffraction peak
of the Cu-Pd alloy shifted continuously toward lower angles as
the Pd content increased: 2q=42.7 and 42.58 for 10Cu-3Pd/
ZrO2 and 10Cu-5Pd/ZrO2, respectively. This suggests a different
composition of alloys in these two catalysts. Another important
finding was that the diffraction peaks became much weaker
compared to that in the pattern of the monometallic Cu cata-
lyst, which indicates the better dispersion of Cu particles with
the introduction of Pd by the formation of CuPd alloys and
shows the advantages of a bimetallic catalyst over a monome-
tallic catalyst in the dispersion of the metal sites.
In light of previous work, the search for active sites for the
effective HDO of the aldehyde group and/or further ring hy-
drogenation has become the top priority for the CTH of furfu-
ral to 2-MF or 2-MTHF.[3] Generally, Cu-based catalysts are po-
tential candidates toward the HDO of the aldehyde group to
the methyl group, whereas Pd-based catalysts exhibit a good
performance in the hydrogenation of the unsaturated C=O
bond or the furan ring.[17] Therefore, a combination of these
two metal sites in one catalyst may provide a multifunctional
catalyst for the conversion of furfural.
Herein, we report the use of robust and efficient bimetallic
Cu-Pd catalysts for the CTH of furfural to 2-MF and 2-MTHF in
the presence of the hydrogen donor 2-propanol. By modifying
the Cu/Pd ratio, either 2-MF or 2-MTHF could be obtained se-
lectively as the major product under the optimized reaction
conditions. The effect of the Cu/Pd ratio, support material, re-
action temperature, and reaction time on the reaction conver-
sion and product distribution was studied. Furthermore, a de-
tailed investigation on the reaction mechanism and the syner-
gy between Cu and Pd is also presented. Finally, the catalyst
recycling ability was investigated.
A TEM image of the representative 10Cu-3Pd/ZrO2 catalyst is
presented in Figure 2. The metal particles were well dispersed
Results and Discussion
Catalyst characterization
XRD patterns of the as-prepared metal catalysts and the ZrO2
support are shown in Figure 1. The diffraction pattern of the
support reveals the monoclinic phase of ZrO2, which corre-
sponds to JCPDF 37-1484. These peaks are still present in the
patterns of the obtained catalysts without noticeable change.
Figure 2. TEM (left) and HRTEM (right, Cu-Pd alloy domain) analysis of the
10Cu-3Pd/ZrO2 catalyst. The inset shows the particle size distribution.
on the supports with a mean size of 10.8 nm, which was small-
er than that of the monometallic Cu/ZrO2 catalyst (mean size
of 13.2 nm; Figure S1). The introduction of Pd reduced the ag-
gregation of Cu particles significantly during the preparation
procedure, which resulted in a better dispersion and smaller
sizes of the metal particles. This may be attributed to the inter-
actions between the two metals, known as the geometric and
stabilizing effect, and the addition of a second metal promoter
alters the geometry of the particle size and improves the sta-
bility of the active metal sites by inhibiting metal sintering
during the preparation or reaction.[20] A high-resolution trans-
Figure 1. XRD patterns of the as-prepared catalysts.
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 3. XPS spectra of the as-prepared 10Cu-3Pd/ZrO2 catalyst: Cu2p, Cu LMM, and Pd3d level spectra.
mission electron microscopy (HRTEM) image of the 10Cu-3Pd/
ZrO2 catalyst was also recorded. The clear crystalline fringe of
the selected domain was measured to be 0.217 nm, which cor-
responds to the (111) interplanar spacing of the face-centered
cubic structure of the CuPd alloy.[20,21] This result agreed with
the XRD results, which showed that the two metals formed an
alloy particle in the as-prepared catalyst.
and Pd were immobilized successfully and dispersed homoge-
neously on the support ZrO2 in the form of Cu-Pd alloy and
metal oxide states.
The textural properties of the catalysts are presented in
Table S1. The support ZrO2 has a moderate BET surface area of
109.18 m2 gꢀ1, a total pore volume of 0.34 cm3 gꢀ1, and an aver-
age pore diameter of approximately 12.47 nm. After the immo-
bilization of the metals, all three parameters of the as-prepared
catalysts decreased to some extent. With the increase of the
Pd content at a fixed Cu loading, the BET surface area, total
pore volume, and average pore size of the obtained catalysts
decreased gradually (Table S1, entries 4–6). This may be caused
by the slight collapse of the structure during the catalyst prep-
aration and the blockage of the pores as more metal particles
were incorporated into the ZrO2 structure.
The chemical states of the metal particles of the Cu-Pd cata-
lyst were further identified by using X-ray photoelectron spec-
troscopy (XPS). 10Cu-3Pd/ZrO2 was selected as a representative
catalyst for the XPS analysis, and the results are presented in
Figure 3. Apart from the metal and alloy states of Cu particles
as indicated by the XRD results, the oxidative states also exist
in the catalyst. The Cu2p3/2 and Cu2p3/2 levels were fitted to-
gether with a fixed separation. In the Cu2p3/2 spectrum, the
main peak at a binding energy (BE) of 930–935 eV was decon-
voluted into two peaks that correspond to Cu0/Cu+ (BE=
931.9 eV) and Cu2+ (BE=933.8 eV) species, respectively.[19,22]
The satellite peak around BE=941 eV was clear evidence of
the existence of Cu2+ particles in the Cu-Pd catalysts. As the
binding energies of Cu0 and Cu+ are very close, the Cu LMM
spectrum was recorded to distinguish these two species. The
peak was deconvoluted into three peaks at BE=568.1, 569.0,
and 570.1 eV, which were assigned to the Cu0, Cu2+, and Cu+
species, respectively, consistent with values reported previous-
ly.[23] The percentages of the three components were calculat-
ed from their area peak ratios, which were approximately 44.2,
21.0, and 34.8%, respectively. All three different valence Cu
species were present in the as-prepared catalyst, which indicat-
ed that the catalyst was not fully reduced under our reaction
conditions. The metal oxide may also contribute to the CTH re-
action by providing more Lewis acid sites to facilitate the de-
oxygenation of the O atoms, as revealed by Panagiotopoulou
and Vlachos.[16] The Pd2d5/2 and Pd2d3/2 peaks overlapped par-
tially with the strong peak from Zr4+ in the Zr3p3/2 spectrum.
After fitting, it can be seen that Pd existed mainly in the Pd0
state, shown by peaks at BE=335.8 and 341.1 eV, and no high-
valance Pd was observed. This result suggested that most of
the Pd particles were reduced to the metallic state under the
specific catalyst reduction conditions. If we combine the XPS
results with the XRD analysis, it could be found that both Cu
Catalyst screening for the CTH of furfural
The CTH of furfural was performed by using a batch reactor
with 2-propanol as the solvent and the H donor. Initially, three
different bimetallic catalysts (Cu-Ni, Cu-Ru, and Cu-Pd) with
a fixed metal ratio of 10:3 were prepared and subjected to the
CTH reaction to compare their reactivities in the CTH reaction.
For the reactions with Cu-Ni or Cu-Ru, furfural was converted
fully at 2208C in 4 h but with low yields of the desired prod-
ucts 2-MF and 2-MTHF of 28.5 and 9.7%, respectively (Table 1,
entries 1–2). Furfuryl alcohol (FA), from the hydrogenation of
furfural, was detected as the main byproduct of these reac-
tions (33.8 and 36.3%). Apart from that, only a small amount
of byproducts [e.g., g-verolactone (GVL), cyclopentanone, and
2-(isopropoxymethyl)furan (IPMF)] were detected. The mass
balance of the two reactions was relatively low, which may be
caused by the instability of FA at an elevated temperature,
possibly by forming humins or polymers that cannot be de-
tected by using GC.[24] Therefore, it can be inferred that the
Cu-Ni and Cu-Ru catalysts only had limited deoxygenation abil-
ity to catalyze the cleavage of the CꢀOH bond of FA to pro-
duce 2-MF or 2-MTHF. Strikingly, if we used Cu-Pd as the cata-
lyst, the CTH of furfural provided 82.3% yield of 2-MF+2-MTHF,
in which 2-MF was obtained as the major product in 61.9%
yield. THFA (2.9%) and GVL (1.2%) were also obtained as by-
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Table 1. CTH of furfural with different catalysts.[a]
Entry Catalyst
Conversion [%][b] Yield[b] [%]
2-MF 2-MTHF FA
2-MF/2-MTHF Mass balance[d]
THFA GVL others[c] 2-MF+2-MTHF
1
2
3
4
5
6
7
8
10Cu-3Ni/ZrO2
100
100
100
100
100
100
100
27.0
8.6
1.5
1.1
20.4
26.2
24.4
5.6
78.8
11.4
7.0
33.5
48.3
33.8
36.3
0
0
0
0
0
0
38.7
0
0
0
1.5
4.6
1.2
0
1.1
1.9
1.0
4.5
0
7.3
8.9
3.5
6.2
4.1
10.2
1.3
3.8
7.5
15.3
11.4
28.5
9.7
18.6
7.8
3.0
1.23
1.4
10.9
0.06
5.6
4.5
0
71.1
59.5
89.9
65.5
87.4
79.8
93.0
84.8
84.6
48.8
59.7
10Cu-3Ru/ZrO2
10Cu-3Pd/ZrO2
10Cu-3Pd/Al2O3
10Cu-3Pd/SiO2
10Cu-3Pd/ TiO2
10Cu-5Pd/ZrO2
10Cu-1Pd/ZrO2
10Cu/ZrO2
61.9
32.3
35.2
61.2
5.1
63.6
31.4
0
2.9
0.8
22.6
0.9
6.8
1.5
0
82.3
58.5
59.6
66.8
83.9
75.0
38.4
33.5
48.3
98.5
9
10
11[e]
98.9
97.2
97.9
3Pd/ZrO2
10Cu/ZrO2+3Pd/ZrO2
0
0
0
0
0
0
0
[a] Conditions: furfural 1 mmol, catalyst 120 mg, 2-PrOH 14 mL, 2208C. [b] Determined by using GC. [c] Other products include cyclopentanone/cyclopenta-
nol, 2-potanone and 2-(isopropoxymethyl)furan. [d] The mass balance was calculated based on the detected known products. [e] The catalyst loading was
120 mg+120 mg.
products (Table 1, entry 3), and the mass balance was 89.9%.
This indicates that the combination of Cu and Pd in the bimet-
allic catalyst was an appropriate choice for the CTH of furfural
to produce 2-MF and 2-MTHF.
crease of the Pd content led to a slight decrease in the prod-
uct yield as compared to that of 10Cu-3Pd/ZrO2, the catalyst
still maintained a high reactivity in the CTH reaction. Notably,
the 2-MF/2-MTHF ratio increased from 3.0 to 5.6, which indicat-
ed that the ring hydrogenation of 2-MTHF was suppressed at
a lower Pd loading. From the results of the catalysts with vari-
ous Pd loadings, the selectivity to 2-MF increased along with
the Pd loading, which was consistent with the previous discus-
sion that the ring hydrogenation occurred mainly on the Pd
sites.
Other commercially available supports (Al2O3, SiO2, and TiO2)
were employed in the preparation of Cu-Pd catalysts to com-
pare their reactivities in the CTH of furfural. Moderate yields of
the desired products 2-MF and 2-MTHF were obtained with
the three catalysts, among which the TiO2-supported catalyst
showed the highest 2-M+2-MTHF yield (Table 1, entries 4–6),
which was still lower than that of the catalyst with ZrO2 as the
support. Therefore, ZrO2 was the best support for the CTH re-
action. This may be explained by the special properties of zir-
conium oxides as reported previously for the CTH reaction.[25]
Zirconium oxides helped the dissociation of H from the H
donor and transferred it to the unsaturation through its acidic
and basic sites by forming strong hydrogen-bonding or metal–
oxygen interactions. In light of these previous results, we spec-
ulate that the acid–base sites of Zr support of our catalyst also
assists the absorption of alcohols and participates in the disso-
ciation of H from alcohols, which thereby promoted the reac-
tion rate in addition to the Cu and Pd sites. This may be the
key to the excellent performance of the ZrO2-supported cata-
lysts.
To investigate out the role of Cu and Pd in the CTH reaction,
monometallic Cu/ZrO2 and Pd/ZrO2 were prepared and tested
in the CTH of furfural. If 10Cu/ZrO2 was used, only a 31.4%
yield of 2-MF and a 7.0% yield of 2-MTHF was obtained
(Table 1, entry 9) accompanied by a considerable amount of FA
(38.7%). This result shows clearly that the monometallic Cu
catalyst had a HDO ability but cannot catalyze the cleavage of
the CꢀOH bond efficiently under the specific CTH reaction con-
ditions. Cu/ZrO2 showed an inferior catalytic activity to the Cu-
Pd bimetallic catalyst. This comparison further highlighted the
necessity to employ Pd as a promoter to the catalyst. However,
if Pd/ZrO2 was used, only a 33.5% yield of 2-MTHF was ob-
tained without other detectable byproducts (Table 1, entry 10).
Many of the products may underwent the CꢀC cracking reac-
tion and decomposed to small molecules or gaseous products
under the specific reaction conditions. It has been described
previously that monometallic Pd had a strong hydrogenation
ability that caused the cracking of CꢀC bonds easily at elevat-
ed reaction temperatures.[17,20] Therefore, the introduction of
Cu to the Pd catalyst in an appropriate ratio can inhibit the
strong hydrogenation ability of Pd, which thereby increases
the reaction selectivity to the desired product. Based on these
results, we can see that both the properties of the two metals
were modified in the bimetallic catalyst, by the formation of
alloys or mixed phases, to result in a more efficient catalyst
than either of the single monometallic catalysts in the CTH of
furfural. Additionally, the physical mixture of the two catalysts
The influence of Pd on the CTH reaction was also investigat-
ed by varying the Pd loading at a fixed Cu loading. If 10Cu-
5Pd/ZrO2 was used, a higher yield of 83.9% of the 2-MF and
2-MTHF mixture was obtained under the same reaction condi-
tions (Table 1, entry 7). Contrary to the reaction with 10Cu-
3Pd/ZrO2, 2-MTHF was produced selectively as the main prod-
uct (78.8%) with only a 5.1% yield of 2-MF. This may be
caused by the increase of the Pd loading, which created more
active sites for ring hydrogenation to facilitate the conversion
of 2-MF to 2-MTHF. This was confirmed by the result with the
catalyst that had a lower Pd loading: if 10Cu-1Pd/ZrO2 was
used, 98.5% furfural conversion was achieved with a 75.0%
yield of 2-MF+2-MTHF (Table 1, entry 8). Although the de-
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
in the CTH of furfural only yielded 48.3% 2-MF+2-MTHF
(Table 1, entry 11), which is much lower than that of the
as-prepared bimetallic catalysts. Based on these result, we can
infer that the synergism between the two metals is key to the
high efficiency of the catalyst, either in the form of an alloy or
the physical proximity between the two metals. The physical
mixing cannot create the “interactions” between the two
metals, which thereby lack the sufficient reactivity. Overall, the
CTH of furfural to 2-MF and 2-MTHF was realized over the Cu-
Pd bimetallic catalyst supported on ZrO2 with high yields. By
varying the metal loading, 2-MF or 2-MTHF could be obtained
selectively as the major product.
2208C was selected as optimal. Notably, the yield of 2-MTHF
increased gradually over the entire reaction temperature
range, which implies that the ring hydrogenation ability of the
Cu-Pd catalyst increased as the reaction temperature in-
creased.
The influence of the catalyst loading on the product distri-
bution was also investigated to assess its role in the CTH reac-
tion. To reserve the starting material and intermediates,
a milder temperature of 1808C was used for the reaction
(Table S2). The furfural conversion increased from 98.6 to
100%. During this stage, the yield of 2-MF increased continu-
ously to reach the highest yield of 65.6% at 18.9 wt% catalyst
loading, which remained almost unchanged at 21.7 wt% cata-
lyst loading. At the same time, the yield of 2-MF also increased
as the catalyst loading increased. An excessive catalyst loading
did not further improve the yield of 2-MF+2-MTHF and de-
creased the reaction carbon balance (18.9 and 21.7 wt%). This
phenomenon may be caused by side reactions in the presence
of excessive active metal sites. In these catalytic runs, FA was
present as the main intermediate, and its concentration de-
creased as the catalyst loading increased. This indicated that
FA might be the key intermediate to the final product from
furfural. In addition, the carbon balance of the reactions con-
ducted at 1808C was a little lower than that of the reactions
conducted at under the optimized reactions. This may be
caused by the lower reaction rate of furfural and FA at a lower
reaction temperature. These two substrates undergo side reac-
tions easily (to form humins or polymers) if they are exposed
to the reaction medium for longer (Table S6).
Effect of the reaction temperature and catalyst loading
The effect of the reaction temperature on the CTH of furfural
was investigated in the range of 160–2408C, and the other re-
action conditions were maintained (Figure 4). The reaction con-
ducted at 1608C provided a 27.6% yield of 2-MF+2-MTHF at
99.0% furfural conversion, accompanied by a 51.6% yield of
FA. This result suggested the hydrogenation of the aldehyde
group proceeded well under relatively mild conditions, howev-
Reaction pathway and proposed mechanism
FA and the other byproducts were identified by using GC–MS
to give direct evidence of the CTH reaction mechanism. Hence,
the product evolution with respect to reaction time was traced
by using GC at a mild reaction temperature of 1808C
(Figure 5). During the heating of the reaction to the desired re-
action temperature (15 min), 52.5% furfural was converted and
FA was detected in 18.3% yield together with 2.3% yield of
Figure 4. Influence of the reaction temperature on the product distribution.
Conditions: furfural 1 mmol, catalyst 120 mg, 2-PrOH 14 mL, 4 h.
er, the hydrogenolysis of the CꢀO group was difficult. As the
reaction temperature increased, furfural was converted fully, FA
was converted gradually, and the yield of 2-MF+2-MTHF in-
creased continuously until it reached a maximum of 82.3% at
2208C. This implied that the hydrogenolysis ability of the cata-
lyst was promoted as the reaction temperature increased. The
carbon balance also increased even at the full conversion of
furfural, which may be caused by the absorption of furfural or
FA on the surface of the catalyst that cannot be detected
(Table S6). However, an excessively high temperature caused
a decrease in the yield of 2-MF+2-MTHF (66.7%) and more by-
products were detected. The carbon balance decreased to
80.1% as some of the side products may undergo decomposi-
tion (Table S3). Therefore, a moderate reaction temperature of
Figure 5. Time course of CTH with the 10Cu-3Pd/ZrO2 catalyst. Conditions:
furfural 1 mmol, catalyst 120 mg, 2-PrOH 14 mL, 1808C.
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
2-MF and no 2-MTHF. The formation of other byproducts was
negligible. As the reaction proceeded, furfural was converted
quickly in 30 min and, at the same time, FA reached a maxi-
mum yield of 42.0%, which decreased gradually afterwards.
The yields of 2-MF and 2-MTHF increased over the entire reac-
tion time. The yield of 2-MF reached a maximum of 64.8% at
240 min and then it remained almost unchanged at 300 min.
The yield of 2-MTHF increased continuously over the entire re-
action time until it reached 5.8% at 300 min. From these re-
sults, it can be seen that furfural was converted quickly and FA
was the main intermediate, the concentration of which first in-
creased and then decreased in the reaction. During this stage,
the concentration of GVL exhibited similar changes to FA in
the reaction mixture, which indicated that GVL might also be
an intermediate of this reaction (Table S4).
only a 4.6% yield of 2-MTHF (Table 2, entry 2). Accordingly, the
Cu-Pd catalyst showed a limited HDO ability to convert GVL to
2-MTHF under the current reaction conditions. This was further
confirmed by using GVL as the starting material in the CTH re-
action (Table 2, entry 7). Accordingly, the pathway via IPL con-
tributed very little to the production of 2-MTHF. To examine
route 2, THFA was subjected to CTH and there was almost no
conversion within the reaction time and no desired product
was found (Table 2, entry 3). This result shows clearly that
THFA was not able to be converted to 2-MTHF in our system
but was just an undesired byproduct. In the case of IPMF in
route 3, an analogous substrate 2-(ethoxymethyl)furan was
used instead of IPMF for the CTH reaction as IPMF was not
commercially available. A 66.1% conversion was obtained with
11.8 and 3.1% yields of 2-MF and 2-MTHF, respectively (Table 2,
entry 4). A considerable amount of GVL was also produced.
These results showed that the ether may also undergo the hy-
drogenolysis of the CꢀO bond to the desired product but with
a relatively slow reaction rate. Although IPMF was only detect-
ed in a very low amount in our reaction optimization, it may
contribute to the formation of the desired products. Similar
conversions of the ether group to the methyl group have been
reported previously for the CTH of furanics.[26] Finally,
To further investigate the reaction pathways, possible inter-
mediates as listed above were also subjected to the CTH reac-
tion under the optimized reaction conditions. Furfuryl alcohol
was first tested in the CTH reaction. 2-MF and 2-MTHF were
obtained in 64.8 and 20.1% yield, respectively, which is similar
to the results with furfural as the starting material (Table 2,
entry 1). This is direct evidence of the furfural!FA!
2-MF was also tested in the CTH reaction (Table 2,
Table 2. CTH of the possible intermediates and products.[a]
entry 5) to give a 18.6% yield of 2-MTHF under the
reaction conditions, which was similar to the result
Entry Substrate
Conversion Yield[b] [%]
Mass
obtained from the CTH of furfural (20.4% yield of 2-
MTHF). If Pd/ZrO2 was used, the ring hydrogenation
proceeded efficiently with an 84.0% yield of 2-MTHF,
which confirmed the high reactivity of the Pd parti-
cles in the hydrogenation. These results agreed with
and further supported our earlier discussion that 2-
MTHF was produced mainly through the ring hydro-
genation of 2-MF. The CTH of 2-MTHF did not pro-
vide any products, and the carbon balance was 92%,
which further confirmed the thermal decomposition
of the desired products under the reaction conditions
(Table 2, entry 6). Based on the experimental results
outlined above, we found that multiple reaction
pathways existed in the CTH of furfural, among
which the hydrogenation of furfural to FA and further
hydrogenolysis/hydrogenation to 2-MF and 2-MTHF
was the main route. In parallel to this route, the
other two pathways via the intermediates IPL and
[%]
2-MF 2-MTHF
THFA GVL
balance
1
2
98.9
64.8 20.1
3.5
–
–
89.5
98.0
–
4.6
78.5 86.8
3
4
5
6
4.5
–
0
–
–
–
93.0
66.1
11.8
3.2
38.6 89.8
30.5/95.0[c]
9.0
–/–
18.6/84.0[c] –/–
–/–
–
88.1/90.0[c]
–
–
–
–
–
92.3
90.8
7
12.0
2.8
–
[a] Conditions: substrate 1 mmol, 10Cu-3Pd/ZrO2 120 mg, 2-PrOH 14 mL, 2208C, 4 h.
[b] Determined by using GC. [c] Pd/ZrO2 (120 mg) as catalyst.
2-MF/2-MTHF reaction pathway of the CTH reaction, through
hydrogenation, hydrogenolysis, and ring hydrogenation. In ad-
dition to this pathway, three other possible routes might also
existed in the CTH reaction: 1) FA is alcoholyzed to isopropyl
levulinate (IPL) with 2-propanol, then hydrogenated to GVL,
and finally hydrodeoxygenated to 2-MTHF; 2) FA is hydrogen-
ated to THFA and then hydrogenolyzed to 2-MTHF; 3) FA is
converted to IPMF by the etherification between FA and the al-
cohol and then hydrogenolyzed/hydrogenated to 2-MF and 2-
MTHF.
IPMF also contributed partially to the formation of the desired
products.
Catalyst recycling
Stability and recyclability are of great importance for a hetero-
geneous catalyst in practical applications. Therefore, the recy-
cling of the catalyst was also investigated. After the reaction,
the catalyst was separated simply from the reaction mixture by
filtration, washed with 2-propanol, and used directly in the
next run. Five catalytic runs were performed, and the results
are presented in Figure 6. The desired products 2-MF+2-MTHF
were obtained in a total yield of approximately 82% in each
To verify these three hypothesis, the related intermediates
were subjected to the CTH reaction. The reaction with IPL
(route 1) as the reactant provided a 78.5% yield of GVL with
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
experimental data that the ring hydrogenation ability of Pd
species was suppressed after the reaction as more furanics
were retained in the final products. The active Pd sites may be
poisoned slightly during the reaction. However, the catalyst
still showed
a high reactivity toward the formation of
2-MF+2-MTHF. The spent catalyst was also analyzed by using
TEM and XRD, and the results are presented in Figures S2 and
S3. No difference was observed in the XRD patterns of the
fresh and spent catalysts. However, the used catalyst showed
a slight increase in the metal particle size (from 10.8 to
13.9 nm). Overall, 10Cu-3Pd/ZrO2 showed an excellent reactivi-
ty and recycling ability in the CTH of furfural to the desired
products 2-MF and 2-MTHF.
Figure 6. Recycling of the 10Cu-3Pd/ZrO2 catalyst. Conditions: furfural
1 mmol, catalyst 120 mg, 2-PrOH 14 mL, 2208C, 4 h.
Conclusions
We have demonstrated an efficient catalytic system based on
bimetallic Cu-Pd catalysts for the transfer hydrogenation of fur-
fural to 2-methylfuran (2-MF) and 2-methyltetrahydrofuran
(2-MTHF) with high yields of up to 83.9% in the presence of
the hydrogen donor 2-propanol. By changing the ratio of the
two metals of the bimetallic catalyst, 2-MF and 2-MTHF could
also be obtained selectively as the major products in high
yields. Reaction studies on the catalysts revealed that the syn-
ergism between Cu and Pd enhanced the hydrodeoxygenation
ability of the catalyst significantly and suppressed the side re-
actions. Mechanistic studies showed that the reaction proceed-
ed mainly via the intermediate furfuryl alcohol, which was then
hydrogenolyzed to 2-MF or hydrogenated to 2-MTHF. Finally,
the catalyst showed an excellent reactivity and stability in the
recycling experiment to provide approximately 82% yield of
2-MF+2-MTHF in each catalytic run. Characterization of the
spent catalyst revealed that the in situ reduction of the metal
particles helped to maintain its activity. The current reaction
provides a new efficient system for the upgrade of furfural,
which might also find important applications in the reductive
upgrade of other biomass-derived molecules for the produc-
tion of value-added products.
run, and no clear decrease was observed in the fifth run. These
results demonstrate that 10Cu-3Pd/ZrO2 possesses a high reac-
tivity and stability in recycling. Notably, the yields of 2-MF and
2-MTHF in the first run over the fresh catalyst were 61.9 and
20.4%, respectively. In the following four runs, the yield of
2-MF increased to ꢁ70%, whereas that of 2-MTHF decreased
to ꢁ10%. This change in the product distribution may be
caused by the different chemical states of the metal species in
the catalyst. Thus, to investigate out this change, spent 10Cu-
3Pd/ZrO2 was analyzed by using XPS (Figure 7). In the XPS
Experimental Section
Materials
Furfural (98%), furfuryl alcohol (98%), 2-methylfuran (99%),
2-methyltetrahydrofuran (98%), and tetrahydrofurfuryl alcohol
(97%) were purchased from TCI Chemical Reagent Company
(Shanghai, China). GVL (98%), RuCl3·xH2O (99%), and Pd(OAc)2
(98%) were purchased from J&K Chemical company (Shanghai,
China). Ni(NO3)2·6H2O and solvents were purchased from Sino-
pharm Chemical Reagent Company (Shanghai, China). The metal
oxide supports were purchased from Saint-Gobain NorPro Compa-
ny (USA). Other chemicals were purchased from local companies
and used without further purification.
Figure 7. XPS spectra of the fresh and recycled 10Cu-3Pd/ZrO2 catalysts.
Cu2p3/2 spectrum, the peak that corresponds to Cu0/Cu+ spe-
cies became stronger and the peak that corresponds to Cu2+
became weaker than that of fresh 10Cu-3Pd/ZrO2. This change
indicated that the Cu species underwent in situ reduction in
the CTH reaction system, which might cause the catalyst to be
more active toward the HDO of the aldehyde group. However,
no significant change was observed for Pd species in the
Pd3d5/2 spectrum because of the overlap with the stronger dif-
fraction peak of Zr. In spite of that, we can still speculate from
Catalyst preparation
All the catalysts were prepared by an incipient impregnation
method. Typically, to prepare 10Cu-3Pd/ZrO2, ZrO2 (1.74 g) was first
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[1] a) K. A. Rogers, Y. Zheng, ChemSusChem 2016, 9, 1750; b) M. J. Gilkey, B.
Xu, ACS Catal. 2016, 6, 1420.
impregnated with
a
aqueous solution that contained
Cu(NO3)2·3H2O (0.753 g) and Pd(OAc)2 (0.127 g). The mixture was
stirred vigorously for 12 h and dried at 1108C overnight. The result-
ing powder was ground and calcined at 5008C for 4 h. The catalyst
was reduced in a flow of 5% H2/N2 from RT to 5008C at
108Cminꢀ1 then kept at 5008C for another 2 h. After cooling to RT,
the catalyst was collected and weighed. This catalyst was denoted
as 10Cu-3Pd/ZrO2 (10%Cu-3%Pd/ZrO2). The other catalysts were
prepared by the same procedure.
[2] a) C. Chatterjee, F. Pong, A. Sen, Green Chem. 2015, 17, 40; b) K. Wu, Y.
Wu, Y. Chen, H. Chen, J. Wang, M. Yang, ChemSusChem 2016, 9, 1355.
[3] a) R. Mariscal, P. Maireles-Torres, M. Ojeda, I. Sꢂdaba, M. L. Granados,
Energy Environ. Sci. 2016, 9, 1144; b) K. Yan, G. Wu, T. Lafleur, C. Jarvis,
Renewable Sustainable Energy Rev. 2014, 38, 663; c) A. Bohre, S. Dutta, B.
Saha, M. M. Abu-Omar, ACS Sustainable Chem. Eng. 2015, 3, 1263; d) Y.
Nakagawa, M. Tamura, K. Tomishige, ACS Catal. 2013, 3, 2655.
[4] a) R. Xing, W. Qi, G. W. Huber, Energy Environ. Sci. 2011, 4, 2193; b) J.
Lange, E. van der Heide, J. van Buijtenen, R. Price, ChemSusChem 2012,
5, 150; c) S. Peleteiro, S. Rivas, J. L. Alonso, V. Santos, J. C. Parajꢃ, Biore-
sour. Technol. 2016, 202, 181.
Characterization
[5] a) F. Dong, Y. Zhu, H. Zheng, X. Li, Y. Li, J. Mol. Catal. A 2015, 398, 140;
b) H. Zheng, Y. Zhu, B. Teng, Z. Bai, C. Zhang, H. Xiang, J. Mol. Catal. A
2006, 246, 18.
[6] W. S. Lee, Z. Wang, W. Zheng, D. G. Vlachos, A. Bhan, Catal. Sci. Technol.
2014, 4, 2340.
[7] a) A. J. Garcia-Olmo, A. Yepez, A. M. M. Balu, A. A. Romero, Y. Li, R.
Luque, Catal. Sci. Technol. 2016, 6, 4705; b) O. F. Aldosari, S. Iqbal, P. J.
Miedziak, G. L. Brett, D. R. Jones, X. Liu, J. K. Edwards, D. J. Morgan, D. K.
Knight, G. J. Hutchings, Catal. Sci. Technol. 2016, 6, 234.
[8] a) J. Wang, X. Liu, B. Hu, G. Lu, Y. Wang, RSC Adv. 2014, 4, 31101; b) A. S.
Gowda, S. Parkin, F. T. Ladipo, Appl. Organomet. Chem. 2012, 26, 86.
[9] M. Hronec, K. Fulajtarova, Catal. Commun. 2012, 24, 100.
[10] S. Sitthisa, W. An, D. E. Resasco, J. Catal. 2011, 284, 90.
[11] J. Lessard, J. F. Morin, J. F. Wehrung, D. Magnin, E. Chornet, Top. Catal.
2010, 53, 1231.
XRD patterns of the catalysts were recorded by using a Rigaku D
MAX III VC diffraction system. XPS spectra were recorded by using
a Scientific Escalab 250-X-ray photoelectron spectrometer. The BET
surface area, average pore size, and total pore volume of the cata-
lyst were measured by using an Automatic Surface Area and Pore
Analyzer (Tristar II 3020M, Micromeritics). TEM images were ac-
quired by using a JEOL-2010 electron microscope.
The reaction mixture was analyzed by using GC–MS (Agilent
7890A, Agilent 5975C MS and GC Agilent 7890A) with a DB-5 capil-
lary column (30 mꢁ0.25 mmꢁ0.25 mm, Agilent). The temperature
of the column increased from 62 to 2508C at a 158Cminꢀ1 rate.
Naphthalene was used as the internal standard.
[12] a) W. R. H. Wright, R. Palkovits, ChemSusChem 2012, 5, 1657; b) Z.
Zhang, ChemSusChem 2016, 9, 156.
[13] a) L. Hu, L. Lin, S. Liu, Ind. Eng. Chem. Res. 2014, 53, 9969; b) G. Bottari,
A. J. Kumalaputri, K. K. Krawczyk, B. L. Feringa, H. J. Heeres, K. Barta,
ChemSusChem 2015, 8, 1323.
[14] D. Scholz, C. Aellig, I. Hermans, ChemSusChem 2014, 7, 268.
[15] P. Panagiotopoulou, N. Martin, D. G. Vlachos, J. Mol. Catal. A 2014, 392,
223.
[16] P. Panagiotopoulou, D. G. Vlachos, Appl. Catal. A 2014, 480, 17.
[17] F. Dong, Y. Zhu, G. Ding, J. Cui, X. Li, Y. Li, ChemSusChem 2015, 8, 1534.
[18] P. M. de Souza, L. Nie, L. E. P. Borges, F. B. Noronha, D. E. Resasco, Catal.
Lett. 2014, 144, 2005.
[19] M. V. Castegnaro, A. Gorgeski, B. Balke, M. C. M. Alves, J. Morais, Nano-
scale 2016, 8, 641.
[20] D. M. Alonso, S. G. Wettsteina, J. A. Dumesic, Chem. Soc. Rev. 2012, 41,
8075.
[21] F. Yu, W. Zhou, R. M. Bellabarba, R. P. Tooze, Nanoscale 2014, 6, 1093.
[22] S. K. Sengar, B. R. Mehta, Govind, J. Appl. Phys. 2014, 115, 124301.
[23] a) A. Jha, D. Jeong, J. Shim, W. Jang, Y. Lee, C. V. Rodeb, H. Roh, Catal.
Sci. Technol. 2015, 5, 2752; b) J. Batista, A. Pintar, D. Mandrino, M. Jenko,
V. Martin, Appl. Catal. A 2001, 206, 113.
Catalytic reactions
The CTH of furfural was performed by using a Parr reactor. Typical-
ly, furfural (1.0 mmol), 10Cu-3Pd/ZrO2 (120 mg), and 2-propanol
(14 mL) were charged into the reactor with a magnetic stirrer. The
reactor was then sealed and purged with N2 three times to remove
air. The reactor was heated to 2208C and kept at this temperature
for 4 h at a stirring speed of 600 rpm. At the end of the reaction,
the reactor was cooled to RT, and the reaction mixture was sam-
pled and subjected to product analysis.
For the recycling of the catalyst: the catalyst was centrifuged after
the reaction and washed with 2-PrOH three times. All the liquid
was collected, combined, and analyzed by using GC. The solid cat-
alyst was added directly to the reaction tube for the next run.
Acknowledgements
[24] a) E. Ahmad, Md. I. Alam, K. K. Pant, M. Ali Haider, Green Chem. DOI:
10.1039/C6GC01523A; b) I. Delidovich, K. Leonhard, R. Palkovits, Energy
Environ. Sci. 2014, 7, 2803.
[25] X. Tang, L. Hu, Y. Sun, G. Zhao, W. Hao, L. Lin, RSC Adv. 2013, 3, 10277.
[26] J. Jae, W. Zheng, R. F. Lobo, D. G. Vlachos, ChemSusChem 2013, 6, 1158.
The authors are grateful for the financial support by National
Natural Science Foundation of China (NSFC 21502095), Natural
Science Foundation of Jiangsu Province (BK20150872), Jiangsu
student’s platform for innovation training program (SPITP
201510298026Z, 201510298048Z) and Top-notch Academic Pro-
grams Project of Jiangsu Higher Education Institutions (TAPP).
Received: August 15, 2016
Keywords: alloys · biomass · copper · palladium · supported
catalysts
Revised: September 20, 2016
Published online on && &&, 0000
&
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
Identifying the intermediates: The cat-
alytic transfer hydrogenation of bio-
mass-derived furfural to fuel additives
2-methylfuran and 2-methyltetrahydro-
furan is performed over a bimetallic Cu-
Pd catalyst in the presence of
X. Chang, A.-F. Liu, B. Cai, J.-Y. Luo,
H. Pan, Y.-B. Huang*
&& – &&
Catalytic Transfer Hydrogenation of
Furfural to 2-Methylfuran and
2-propanol. The reaction proceeds via
the intermediate furfuryl alcohol, which
is then deoxygenated/hydrogenated to
the desired products.
2-Methyltetrahydrofuran over
Bimetallic Copper–Palladium Catalysts
ChemSusChem 2016, 9, 1 – 9
www.chemsuschem.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ