Direct one-pot conversion of monosaccharides into high-yield 2,5-dimethylfuran over a multifunctional Pd/Zr-based metal-organic framework@sulfonated graphene oxide catalyst

Article abstract of DOI:10.1039/c7gc00269f

A one-pot conversion of monosaccharides (fructose and glucose) into high-yield 2,5-dimethylfuran (2,5-DMF) is demonstrated over a multifunctional catalyst obtained by loading Pd on a Zr-based metal-organic framework (UiO-66) that is deposited on sulfonated graphene oxide (Pd/UiO-66@SGO). The Br?nsted acidity associated with UiO-66@SGO activates the fructose dehydration to form 5-hydroxymethylfurfural (5-HMF), while the Pd nanoparticles further convert 5-HMF to 2,5-DMF by hydrogenolysis and hydrogenation. The results show that under the optimized reaction conditions of 160 °C and 1 MPa H₂ in tetrahydrofuran for 3 h, the yield of 2,5-DMF is as high as 70.5 mol%. This value is higher than the previously reported values, and the direct conversion of fructose can be achieved without additional purification of 5-HMF from the reaction mixture. In addition, for the first time, glucose is converted to 2,5-DMF with a high yield of 45.3 mol%. A recyclability test suggests that the 4.8 wt% Pd loaded on the UiO-66@SGO catalyst can be re-used up to five times.

Full text of DOI:10.1039/c7gc00269f

Green Chemistry
View Article Online
View Journal
COMMUNICATION
Direct one-pot conversion of monosaccharides
into high-yield 2,5-dimethylfuran over a
Cite this: DOI: 10.1039/c7gc00269f
multifunctional Pd/Zr-based metal–organic
framework@sulfonated graphene oxide catalyst†
Received 22nd January 2017,
Accepted 24th March 2017
DOI: 10.1039/c7gc00269f
a
a,c
b
a,c
rsc.li/greenchem
Rizki Insyani, Deepak Verma, Seung Min Kim
and Jaehoon Kim
*
−1
A one-pot conversion of monosaccharides (fructose and glucose) boiling point (93 °C), and calorific value (34 MJ kg ) together
into high-yield 2,5-dimethylfuran (2,5-DMF) is demonstrated over a with a very low water solubility (2.3 g L− ). In addition, 2,5-
multifunctional catalyst obtained by loading Pd on a Zr-based DMF can be used as a building-block for producing highly
metal–organic framework (UiO-66) that is deposited on sulfonated valuable aromatic chemical compounds (e.g., p-xylene) via
1 1,4
5
,6
graphene oxide (Pd/UiO-66@SGO). The Brønsted acidity associ- Diels–Alder reactions.
ated with UiO-66@SGO activates the fructose dehydration to form Despite the many advantages of 2,5-DMF as a renewable
-hydroxymethylfurfural (5-HMF), while the Pd nanoparticles fuel and chemical, the direct one-pot conversion of mono-
5
further convert 5-HMF to 2,5-DMF by hydrogenolysis and hydro- saccharides (glucose and fructose) into high-yield 2,5-DMF is still
genation. The results show that under the optimized reaction con- a great challenge. So far, the main approach for the production
ditions of 160 °C and 1 MPa H in tetrahydrofuran for 3 h, the yield of 2,5-DMF from monosaccharides involved two-steps: (1) the
2
of 2,5-DMF is as high as 70.5 mol%. This value is higher than the dehydration of monosaccharides to 5-hydroxymethylfurfural
previously reported values, and the direct conversion of fructose (5-HMF) in the presence of a Lewis or Brønsted acid, (2) the
can be achieved without additional purification of 5-HMF from the hydrogenolysis of 5-HMF to 2,5-DMF over a metal supported
1
,7–10
reaction mixture. In addition, for the first time, glucose is con- catalysts, as listed in Table S1.†
In the case of the pro-
verted to 2,5-DMF with a high yield of 45.3 mol%. A recyclability duction of 5-HMF from fructose, various types of homogenous
1
1
test suggests that the 4.8 wt% Pd loaded on the UiO-66@SGO and heterogeneous acid catalysts (e.g., HCl + CrCl
3
,
HCOOH
3
2−
12
13
catalyst can be re-used up to five times.
+ H SO , SO4 /ZrO –TiO , C–SO H, and Amberlyst-15
2
4
2
2
3
(
ref. 7)) have been tested. For example, Choudhary et al.
reported that the use of HCl as a Brønsted acid for the de-
hydration of fructose in the presence of CrCl3 in a biphasic
solvent system composed of THF and water could result in
Introduction
1
1
Owing to the current global warming issue and rapid depletion 59 mol% 5-HMF yield. Recently, a significantly high 5-HMF
of crude oil, renewable biofuels and value-added chemicals yield of 98 mol% was achieved using a modified Zr-based
1
–3
derived from biomass are of great interest.
various types of bio-derived fuels and chemicals, 2,5-dimethyl- of [Hf
furan (2,5-DMF), a derivative of furan, is considered as one of sulfoxide (DMSO) for 1 h.
the most promising biofuels because of its clear advantages
The production of 2,5-DMF in high yield was demonstrated
Among the metal–organic framework (NUS (Hf)) with a chemical formula
6
O
4
(OH) L]3.5·xH O as the catalyst at 100 °C in dimethyl
8
2
1
4
over the first generation bioethanol, such as a high energy by using high purity 5-HMF as the starting feed. The seminal
density (30 kJ cm− ), research octane number (RON = 119), work by the Dumesic group in 2007 described the conversion
3
of fructose to 2,5-DMF with a 71 mol% yield using a Cu–Ru/C
1
2
catalyst in 1-butanol at 0.68 MPa of H and 220 °C for 10 h.
a
SKKU Advanced Institute of Nanotechnology (SAINT), Sungkyunkwan University,
066 Seobu-Ro, Jangan-Gu, Suwon, Gyeong Gi-Do, 16419, Republic of Korea.
Afterwards, Binder et al. reported the hydrogenolysis of crude
-HMF, which was produced from corn stover, to 2,5-DMF with
2
5
E-mail: jaehoonkim@skku.edu
b
Institute of Advanced Composite Materials, Korea Institute of Science and
a 49 mol% yield over the same type of catalyst and under
2
Technology, Chudong-ro 92, Bongdong-eup, Wanju-gun, Jeonranbuk-do, Republic of
Korea
similar reaction conditions. Wang et al. determined that
Pt–Co nanoparticles on a hollow carbon sphere could provide
c
School of Mechanical Engineering, Sungkyunkwan University, 2066 Seobu-Ro,
2
a 2,5-DMF yield of 98 mol% at 180 °C and 1 MPa of H for 2 h
Jangan-Gu, Suwon, Gyeong Gi-Do, 16419, Republic of Korea
1
5
in 1-butanol. The use of formic acid as a source of H as well
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
2
c7gc00269f
as an acid catalyst over Pd/C led to the conversion of 5-HMF to
This journal is © The Royal Society of Chemistry 2017
Green Chem.

View Article Online
Communication
Green Chemistry
3
2
,5-DMF with a high yield of 95 mol%. Chidambaram et al. still as high as 45.3 mol%. The role of the support and the
found that the low H2 solubility in a solution of 1-ethyl-3- effect of the reaction parameters (Pd loading, temperature,
methylimidazolium chloride ([EMIM]Cl) in acetonitrile over time) are discussed.
Pd/C resulted in a low 2,5-DMF yield of 0.14 mol% with 32%
1
6
selectivity at 120 °C and 0.62 MPa of H for 1 h. In order to
2
find a suitable reaction medium for producing high-yield 2,5-
DMF, Nagpure et al. tested the hydrogenolysis of 5-HMF in
various solvents with a different chemical nature, including
Results and discussion
Catalyst characterization
nonpolar (toluene), polar protic (2-propanol), and polar Fig. S1a† shows the X-ray diffraction (XRD) patterns of the sup-
aprotic (THF, DMSO, and CH CN) solvents, over a Ru–Na–Y ports and the Pd-loaded catalysts, e.g., sulfonated graphene
3
catalyst, and found that THF was the most promising solvent, oxide (SGO), the Zr-based metal–organic framework (UiO-66),
producing 70 mol% of 2,5-DMF owing to its reduced inter- UiO-66@SGO, 2.4 wt% Pd loaded on UiO-66@SGO (2.4Pd/
1
7
action with the Ru active sites. Recently, Upare et al. reported UiO@SGO), and 4.8 wt% Pd loaded on UiO-66@SGO (4.8Pd/
a 92 mol% overall yield of 2,5-DMF starting from fructose in UiO-66@SGO). The Pd loading was measured using inductively
1
-butanol over Amberlyst-15 (for dehydration of fructose)/Ru– coupled plasma-atomic emission spectroscopy (ICP-OES)
7
Sn/ZnO (for hydrogenolysis of 5-HMF) catalysts.
(Table S2†). The XRD pattern of the as-synthesized UiO-66
2
3
Although the high-yield production of 2,5-DMF from fruc- agrees well with that reported in the literature. The XRD pat-
tose has been well-demonstrated using the two-step approach, terns of the 2.4Pd/UiO-66@SGO and 4.8Pd/UiO-66@SGO cata-
most of these methods involve a highly energy intensive and lysts exhibit three clear peaks at 2θ of 40.1, 46.7, and 68.1°,
costly purification of 5-HMF from the dehydration products which correspond to the (111), (200), and (220) planes of the
1
8,19
prior to its hydrogenolysis,
which increases the production face-centered cubic Pd structure (JCPDS no. 05-0681), respect-
cost of 2,5-DMF. Several studies have been dedicated to the ively. The intensity of the peaks associated with Pd increases
one-pot, direct conversion of fructose to 2,5-DMF without the with increasing Pd loadings. The crystalline structure of the
purification of 5-HMF from the dehydration mixture, but the host UiO-66 was maintained after the incorporation of SGO
2
,8–10
yield of 2,5-DMF was very low.
For example, the use of and impregnation of Pd.
The thermal stabilities of the as-synthesized UiO-66, 2.4Pd/
CrCl for the dehydration of corn stover, followed by the hydro-
3
genolysis of the crude 5-HMF solution, resulted in only 9 mol% UiO-66@SGO, and 4.8Pd/UiO-66@SGO catalysts were exam-
of 2,5-DMF using Cu–Ru/C in butanol at 220 °C for 10 h in ined by thermogravimetric analysis (TGA) under air flow con-
2
a batch reactor. The one-pot conversion of fructose over ditions, as shown in Fig. S1b.† The marginal weight loss
ZnCl
DMF in THF at 150 °C and 0.8 MPa H
used a one-flow, two-fixed bed reactor with HY zeolite and Cu/ 250–300 °C can be attributed to the reversible dehydration of
2
–Pd/C as the catalyst resulted in only 22 mol% of 2,5- (5.6 wt%) of the samples below 100 °C was due to the
8
2
for 8 h. Xiang et al. adsorbed solvent in the framework, while the weight loss at
2
4,25
ZnO/Al
2
O
3
as dehydration and hydrogenolysis catalysts, Zr
6
O
4
(OH)
4
to Zr
6
O
6
.
The significant weight loss above
respectively, to convert fructose into 2,5-DMF. The resulting 450 °C was caused by the decomposition of the organic linker
9
2
,5-DMF yield was 40.6 mol% at 240 °C. Very recently, Wei (1,4-benzenedicarboxylic acid) used for the synthesis of
et al. have proposed a direct conversion of fructose to 2,5-DMF UiO-66. The early weight loss of Pd/UiO-66@SGO was probably
with a 66 mol% yield using a combination of homogeneous due to the adsorption of water or volatile species on SGO. A
(
AlCl /H SO /H PO ) and heterogeneous (Ru/C) catalysts at major decomposition of the Pd/UiO-66@SGO catalysts
3 2 4 3 4
2
2
00 °C and 1.5 MPa H for 12 h in N,N-dimethylformamide (N, occurred at 425 °C, and it was caused by the decomposition of
2
0
N-DMF). However, the potential reactor corrosion caused by thermally labile oxygenated species in SGO. Therefore, the
the strong acid, the treatment of the waste acid, and the TGA profiles of UiO-66 and Pd/UiO-66@SGO indicate the high
difficulty in product separation caused by using a homo- thermal stability of this support at temperatures below 250 °C.
geneous-type Brønsted acid should be addressed. In addition,
To investigate the chemical functionality of SGO and the
the use of glucose as feed for the direct production of 2,5-DMF valence state of the Pd atoms in the catalysts, X-ray spectro-
in place of fructose has never been reported. A typical method scopy (XPS) analysis was employed, as shown in Fig. S1c.† The
to produce fructose is the enzymatic isomerization of glucose, high-resolution C 1s spectrum of SGO indicates at least four types
but the use of huge amounts of enzymes to ensure the high- of carbons bonded to different oxygen functional groups: C–O–C
throughput of fructose and the narrow processing window of (286.8 eV), C–OH (285.8 eV), CvO (287.8 eV), and OvC–O
2
1,22
isomerization might increase the overall production cost.
(289.3 eV) along with a highly intense peak of C–C (285.6 eV).
Herein, we demonstrate a one-pot, direct conversion of The O 1s spectrum was deconvoluted to three peaks at 531.2,
monosaccharides (fructose and glucose) to 2,5-DMF with an 532.6, and 533.7 eV, which were attributed to O–CvO, C–O, and
−
26,27
unprecedented high yield up to 70.5 mol% over a multifunc-
tional heterogeneous Pd/Zr-based metal–organic framework@- peaks at the binding energies of 168.1 and 169.2 eV, corres-
sulfonated graphene oxide catalyst (Pd/UiO-66@SGO) without ponding to the S 2p3/2 and S 2p1/2 of the sulfonic acid group (SO
O
2
, respectively.
The deconvoluted S 2p spectrum with two
3
)
27–29
involving the purification of 5-HMF. In addition, when glucose on SGO,
indicates that the sulfonic group has been success-
was used in this one-pot conversion, the 2,5-DMF yield was fully incorporated on the GO sheets. The Pd 3d spectra of the
Green Chem.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Green Chemistry
Communication
Table 1 The textural properties of UiO-66, SGO, and Pd/UiO-66@SGO
a
catalysts
BET surface
area (m g
Pore volume
Mean pore
diameter (nm)
2
−1
3
−1
Sample
)
(cm g
)
b
UiO-66
SGO
1318
51
765
750
715
635
0.67
11.8
0.41
0.44
0.40
0.25
2.17
9.57
2.20
2.65
2.74
2.82
b
b
UiO-66@SGO
.4Pd/UiO-66@SGO
.8Pd/UiO-66@SGO
.8PdUiO-66@SGO
b
b
2
4
4
c
a
BET surface area, pore volume, and pore diameter determined by N
2
b
c
adsorption at 77 K. Fresh synthesized catalyst. Fifth cycle reused
catalyst.
2.4Pd/UiO-66@SGO and 4.8Pd/UiO-66@SGO catalysts clearly
show the characteristic peaks of Pd 3d3/2 and Pd 3d5/2 at 341.0
and 335.6 eV, respectively, which demonstrate the presence of
metallic Pd in the catalysts.
Fig. 1 4.8Pd/UiO-66@SGO catalyst surface morphology: (a) SEM image
and (b–e) the corresponding HR-TEM images. The inset in (b) shows the
size distribution of Pd.
2
Fig. S1d† and Table 1 represent the N adsorption–desorp-
tion isotherms of the support and the catalysts. UiO-66 exhibi-
2
−1
ted a type I isotherm with a high surface area of 1318 m g
,
2
3
which is similar to the previously reported values. The pore
3
−1
volume of UiO-66 was 0.67 cm g . The pore size distribution
of UiO-66 was very narrow (0.6–1.2 nm), as shown in Fig. S1e,†
indicating a rich microporous structure. The surface area of
tion transmission electron microscopy (HR-TEM). The parent
1
4
UiO-66 was comprised of octahedron particles with a mean
size of 200 nm (Fig. S3†). As shown in Fig. 1a, the UiO-66
particles were randomly dispersed on the few layered SGO
sheets in the 4.8Pd/UiO-66@SGO catalyst. The HR-TEM image
of 4.8Pd/UiO-66@SGO illustrates a uniform distribution of the
Pd nanoparticles on the surface of UiO-66 with a mean size of
2
−1
UiO-66@SGO decreased to 765 m g because of the contri-
bution of the low surface area of SGO to the support. After the
Pd loading, the surface area of the catalysts decreased slightly
2
−1
to 715–750 m
g , but the pore size distribution did not
change much as compared to their support. This suggests that
the Pd deposition extensively occurred on the exterior surface
of UiO-66 and the textural properties of the support did not
change significantly during the Pd deposition. The Raman
spectra of GO and SGO (Fig. S1f†) show the characteristic
3
.5 ± 0.5 nm. As shown in Fig. S4,† a high population of
Pd nanoparticles with sizes of 2–6 nm was deposited on the
UiO-66@SGO support when the Pd loading increased to
4.8 wt%. The elemental mapping of 4.8Pd/UiO-66@SGO
indicates a uniform distribution of Pd on the UiO-66@SGO
surface (Fig. S5†).
−
1
D and G bands at 1350 and 1597 cm , respectively, which
correspond to the structural defects and in plane vibration of
2
30
the sp carbons, respectively. The slight increase in the
Fructose and glucose conversion to 2,5-DMF
D-band to G-band ratio (I /I ) of SGO (1.06) as compared to
D
G
2
GO (0.85) indicates the decreasing of the sp carbon domain By using the 4.8Pd/UiO-66@SGO catalyst, monosaccharides
during the sulfonation of GO.
can directly be converted into high-yield 2,5-DMF. As listed in
The acid strengths of the catalysts were also analyzed Table 2, extraordinary high yields of 2,5-DMF were obtained
through NH temperature-programmed desorption (TPD); the from fructose and glucose (70.5 and 45.3 mol%, respectively)
3
TPD profiles are shown in Fig. S2† and the corresponding at temperatures of 160–180 °C and 1 MPa H in THF for 3 h.
2
amount of acid sites are listed in Table S2.† The SGO support The use of 5-HMF as feedstock resulted in an almost complete
−
1
had a total of 2.28 mmol g acid sites, of which 0.35 mmol 2,5-DMF yield of 99.2% at 160 °C for 3 h.
−
1
g
1
was attributed to weak-medium acid sites (150–300 °C) and
The reaction mixtures were analyzed using gas chromato-
.93 mmol g− to strong acid sites (300–600 °C). The acid sites graphy-time-of-flight mass spectroscopy (GC-TOF/MS) to
1
of Pd/UiO-66@SGO were shifted to strong acid regions understand the reaction pathway from fructose to 2,5-DMF
Fig. S2†), suggesting an increase in the Brønsted acidity of the over the Pd/UiO-66@SGO catalyst. The reaction intermediates
catalyst. The total acidities of 2.4Pd/UiO-66@SGO and 4.8Pd/ (5-hydroxymethylfurfural, 5-HMF; 5-methylfurfural, 5-MFA;
(
−
1
UiO-66@SGO were 3.16 and 3.21 mmol g , respectively. The 5-methyl-furanmethanol, 5-MFM), the by-product (furfural,
strong acid sites contributed to over 90% of the total acid FA), the final product (2,5-DMF), and the over-hydrogenated
sites. As listed in Table S1,† a similar trend can be observed in product (2,5-dimethyltetrahydrofuran, 2,5-DMTHF) were
the titration results.
identified in the reaction mixtures. Plausible reaction pathways
The morphology of the Pd/UiO-66@SGO catalysts was exam- for the direct conversion of glucose and fructose over Pd/
ined by scanning electron microscopy (SEM) and high-resolu- UiO-66@SGO are shown in Scheme 1.
This journal is © The Royal Society of Chemistry 2017
Green Chem.

View Article Online
Communication
Green Chemistry
a
Table 2 Monosaccharide conversion to 2,5-dimethylfuran
f
Reaction conditions
Conversion Yield (mol%)
Selectivity
Entry Reactant Catalyst
(%)
(%)
T (°C) Time (h) P (MPa)
5-HMF FA
5-MFA 2,5-DMF 2,5-DMTHF
e
1
2
3
4
4
5
6
7
8
9
1
1
1
1
1
Fructose No catalyst
GO
200
200
200
200
200
200
160
3
3
3
3
3
3
3
1
3
5
3
3
3
3
3
1
1
1
1
1
1
1
1
1
1
0.3
1
1
1
1
33.3
80.8
83.0
87.4
55.3
90.5
88.1
77.5
85.2
89.6
78.6
91.8
87.3
99.9
99.2
—
8.6
Trace
35.3
29.4
32.1
38.3
19.3
Trace 9.6
Trace 36.1
Trace 8.6
Trace 9.7
Trace 16.7
Trace 7.8
Trace 18.0
—
—
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
—
—
—
—
—
—
43.6
33.4
83.8
62.0
59.3
76.8
63.8
99.3
67.5
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
38.5
25.9
68.6
55.3
46.4
70.5
45.3
99.2
68.0
SGO
23.6
35.2
10.7
55.9
Trace
18.5
Trace
Trace
21.4
Trace
5.10
n.d.
n.d.
a
b
UiO-66b
UiO-66
UiO-66/SGO
4.8Pd/SGO
2.4Pd/UiO-66@SGO 160
24.0
n.d.
n.d.
n.d.
n.d.
Trace
Trace
n.d.
31.0
160
160
160
0
1
2
3
4
4.8Pd/UiO-66@SGO 160
Glucose 4.8Pd/UiO-66@SGO 180
c
d
5-HMF
5-HMF
4.8Pd/UiO-66@SGO 160
4.8Pd/SGO 160
n.d.
n.d.
Trace
Trace
a
b
c
Standard reaction conditions: 0.2 g (1.1 mmol) feed, 0.2 g catalyst, and 40 mL THF. Synthesized using the method in the ref. 34. 0.18 g
d
e
f
(1 mmol) glucose. 0.12 g (1 mmol) 5-HMF. Not detectable. 2.5-DMF selectivity = mol of 2,5-DMF/(mol of reactant − mol of unreacted reactant)
×
100.
Scheme 1 Plausible reaction mechanism pathways for glucose conversion to 2,5-dimethylfuran. (1) Fructose; (2) 5-hydroxymethylfurfural (5-HMF);
(
(
3) 5-methylfurfural (5-MFA); (4) 2,5-bis(hydroxymethyl)-furan (2,5-BHMF); (5) 5-methyl-furanmethanol (5-MFM); (6) 2,5-dimethylfuran (2,5-DMF);
7) 2,5-dimethyltetrahydrofuran (2,5-DMTHF); (8) furfural (FA). The H3 process indicates hydrogenolysis/hydrogenation/hydrogenolysis.
3
14
To investigate the roles of the Pd/UiO-66@SGO catalyst and which the use of polar aprotic solvents (THF, DMSO, and
2
0
the reaction medium in the direct conversion of fructose, N,N-DMF ) resulted in high-yield 2,5-DMF from 5-HMF by
several control experiments were conducted. First, in order to avoiding the ring-opening of 5-HMF to levulinic acid
find a preferable fructose dehydration medium, blank experi- derivatives.
ments (in the absence of a catalyst) were conducted in polar
protic (methanol, ethanol, and isopropyl alcohol) and aprotic THF was 33.3 mol%, but 5-HMF was not observed (entry 1,
tetrahydrofuran) solvents, and the production of furanic com- Table 2 and Fig. S6d†). Instead, hydroxypropanone, furfural,
In the absence of a catalyst, the conversion of fructose in
(
pounds was monitored, as shown in Fig. S6.† Based on the butyrolactone, 1,6-anhydrofructofuranoside, and butylated
area% of the detected compound in the GC-TOF/MS chromato- hydroxytoluene were produced. In the presence of GO, the
gram, the conversion of fructose at 200 °C and 1 MPa H in fructose conversion increased to 80.8 mol%, but only a small
2
THF after 3 h resulted in a much higher area% of furanic com- amount of 5-HMF was produced (8.6 mol%, entry 2, see
pounds (e.g., furfural, 13.3%) as compared to those obtained Fig. S7a† for the chromatogram). On the other hand, a high
in alcoholic media under identical conditions (only traces). amount of furfural was detected, which can be formed by the
This is in good agreement with previous reports according to cleavage of the C–C bonds of acyclic hexoses followed by the
Green Chem.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Green Chemistry
Communication
3
1,32
dehydration of pentoses.
5
5
The use of SGO increased the the adsorption of both the components reached an equili-
3
7
-HMF yield to 23.6 mol%, indicating the activation of the brium point (which was achieved after 60 min of adsorption),
-HMF pathway by the presence of the Brønsted acid sites as shown in Fig. S11.† The Pd/SGO catalyst has an almost two
associated with the sulfonated functional group on GO; the times higher adsorption selectivity coefficient of 2,5-DMF over
total acid density of SGO, originated from –COOH, –OH, and 2,5-DMTHF (2.32) than that of the Pd/UiO-66@SGO catalyst
3
3
−
1
–
SO
acidity deriving from the –SO
these values are higher than those of the bare GO (Table S2†). Pd/UiO66@SGO.
3
H groups, was estimated to be 1.95 mmol g , while the (1.22), indicating that 2,5-DMF adsorbs more preferentially
H group was 1.17 mmol g−
1
;
on Pd/SGO with respect to 2,5-DMTHF as compared to
3
The fructose conversion using UiO-66 (entry 4a) in the absence
As illustrated in Scheme 2, the high 2,5-DMF adsorption
of an external Brønsted acid source resulted in a high fructose uptake on the SGO surface can be caused by the π–π inter-
2
conversion of 87.4% with a high HMF yield of 35.2 mol%. It is action between the furan unsaturated ring and the CvC sp of
noted that these values are much higher than the fructose con- the graphitic carbon. A previous report indicates that the CvC
34
2
version over the UiO-66 synthesized without using HCl (entry 4b). sp in the GO structure can allow a strong π–π interaction with
This exceptional phenomenon can be explained by the aromatic compounds, and in the presence of Pd nanoparticles,
nature of the crystal structure of UiO-66. In this work, UiO-66 the unsaturated CvC bond of the aromatic substrate can be
2
3
37
was synthesized according to Katz et al. with additional HCl, readily hydrogenated to its corresponding saturated form.
and it was proved to exhibit the natural defects of UiO-66 On the other hand, in the case of 4.8Pd/UiO-66@SGO, the π–π
(
Fig. S9†). The defected UiO-66 exhibited higher total acidity interaction could be suppressed by the presence of high
−
1
(
1.85 mmol g ) than that of the ideally closed-packed surface area microporous UiO-66, which could inhibit the
UiO-66 (1.31 mmol g−1), as listed in Table S2.† Due to the excessive hydrogenation favoring a high 2,5-DMF selectivity. In
missing linkers (1,4-benzenedicarboxylic acid) in the UiO-66 order to evaluate our hypothesis, the 5-HMF conversion was
framework, this imperfection leads to a large surface area of performed using the same reaction conditions (entry 13,
34
2
−1
1
(
318 m g as compared to the ideally closed-packed UiO-66 Fig. S8d†); an excellent catalyst performance with >99.9% of
1187 m g−1).34 The large surface area of the defected UiO-66 5-HMF conversion and 99.2 mol% of 2,5-DMF yield was
2
with a high degree of open framework (which is caused by observed. In contrast, the use of 4.8Pd/SGO resulted in a
3
5
uncoordinated Zr -cluster nodes ) can provide a synergetic substantial amount of 2,5-DMTHF (31.0 mol%, entry 14).
6
4
+
effect of Lewis (the Zr metal center) and Brønsted (hydroxyl This result confirms that the ring saturation of 2,5-DMF to
(
2
Zr–OH)) acids or aquo (Zr–OH terminal groups) sites in a
3
6
single material.
As listed in entry 5 of Table 2 and shown in Fig. S7d,† the
UiO-66@SGO support (with a 1 : 1 weight ratio) was also able
to produce a high conversion of fructose (90.5%) into the
medium yield of 5-HMF (55.9 mol%), since the combinatorial
benefits of SGO (Brønsted acid, hydrophilic surface) and
UiO-66 (high porosity, high surface area, high acidity) can
enhance the fructose adsorption onto the surface of the cata-
lyst (Fig. S10†). Yet, in the absence of a noble metal phase, the
further hydrogenation of 5-HMF to 5-MFA was not observed.
The use of 4.8Pd/UiO-66@SGO provided a considerably
high yield of 2,5-DMF from fructose (70.5 mol%, entry 11,
Fig. S8c†). The presence of 5-MFA in the reaction mixture and
the trace amount of 5-MFM detected in the chromatogram
(
less than 0.1 area%) may imply that the 5-MFA pathway is
dominant over 4.8Pd/UiO-66@SGO. In contrast, the use of
.8Pd/SGO (without UiO-66) resulted in a lower fructose
4
conversion (88.1%) and 2,5-DMF yield (38.5 mol%, entry 6,
Fig. S8b†) as compared to those resulting from the use of
4
4
.8Pd/UiO66@SGO. The low 2,5-DMF selectivity over
.8Pd/SGO was due to the extensive hydrogenation of 2,5-DMF
over the Pd surface, leading to the saturation of the double
bond in 2,5-DMF to produce 2,5-DMTHF with a relatively high
yield (24 mol%). This phenomenon is possibly caused by the
high 2,5-DMF adsorption uptake over SGO. To investigate the
adsorption selectivity of 2,5-DMF over 2,5-DMTHF on
Scheme 2 Schematic illustration of the 2,5-DMF interaction on the
catalyst surface, the conversion route through a hydrogenolysis–hydro-
genation–hydrogenolysis (H3) process, and the corresponding GC-TOF/
4
.8Pd/SGO and 4.8Pd/UiO-66@SGO, the adsorption selectivity
coefficients of 2,5-DMF to 2,5-DMTHF were calculated when MS chromatograms. (a) Pd@SGO catalyst and (b) Pd/UiO-66@SGO.
This journal is © The Royal Society of Chemistry 2017
Green Chem.

View Article Online
Communication
Green Chemistry
2
5
,5-DMTHF was suppressed over 4.8Pd/UiO66@SGO even with
-HMF, whose potential adsorption on the SGO surface might
be higher than that of fructose because of its smaller size.
Therefore, the high yield production of 2,5-DMF directly from
5
-HMF is due to the coordination of the reactant to the well-
dispersed metallic Pd sites deposited on the UiO-66 framework
3
8,39
(Scheme 2). Recent density functional theory studies
suggest the preferable coordination of reactants on the Pd
surface. Three types of preferable conformations were
suggested; flat (substrate bound via its furan ring and the car-
bonyl group), tilted (substrate bound via either its furan ring
or the carbonyl group) and vertical (substrate bound via
oxygen in the furan ring) conformations of furanic compounds
3
8
on the Pd (111) surface. These binding conformations were
affected by the presence of a hydrogen co-feed, which assisted
the hydrogenolysis reaction. In the ultrahigh vacuum experi-
ment, furanic substrates preferred the flat conformation on
3
8
the surface of Pd (111), while the conformation was tilted on
3
9
the crowded surface of the hydrogen-covered Pd (111).
Finally, under moderate pressure and temperature conditions,
the furfural conformational changes on the Pd (111) surface
were preferred to produce highly selective decarbonylated
3
9
species.
As listed in entries 7–9, as the reaction time increased from
1
to 3 h over the 2.4Pd/UiO-66@SGO catalyst, the 2,5-DMF
yield also increased from 25.9 to 68.6 mol%; however further
increases in the reaction time led to a decrease in the 2,5-DMF
yield to 55.3 mol% owing to side reactions. Nonetheless, even
after 5 h, the formation of 2,5-DMTHF was not observed,
indicating the suppression of excess hydrogenation. When the
H
2
pressure was reduced from 1 to 0.3 MPa, the fructose con-
version and the 2,5-DMF yield decreased to 78.6% and
4
6.4 mol%, respectively.
Since a large quantity of the enzyme (e.g., D-glucose/xylose
Fig. 2 The conversion of (a) glucose and (b) fructose and yields of the
main reaction intermediates and products with varying reaction times.
Reaction conditions: 0.2 g feed, 0.2 g 2.4Pd/UiO-66@SGO, 40 mL THF,
isomerase) and careful control of the process parameters (e.g.,
temperature, pH) are required to meet the high-throughput 180 °C, 1 MPa H . 5-HMF, 5-hydroxymethylfurfural; 5-MFA, 5-methyl-
2
production of fructose from glucose by reversible isomeriza- furfural; and 2,5-DMF, 2,5-dimethylfuran.
2
1,22
tion,
the direct conversion of glucose to 2.5-DMF would be
more beneficial. As compared to the fructose conversion
−1
(
entries 11 and 12), the use of glucose resulted in a lower con- can be more readily converted into 5-HMF (80.47 kJ mol ) as
version (87.3%) and 2,5-DMF yield (45.3 mol%), while compared to glucose (100.99 kJ mol− ). At low glucose and
affording a higher 5-MFA yield (18.0 mol%). Apparently, the fructose conversion at a short reaction time of 10 min, the
additional isomerization step would require longer reaction major products were 5-HMF and unconverted monosacchar-
1
times to achieve a high 2,5-DMF yield.
ides, as listed in Table S5.†
The conversion of glucose and fructose and yields of main
To find out the possible side reactions in fructose conver-
reaction intermediates and products with varying reaction sion at extended reaction times, the conversion reaction was
times from 10 to 300 min are shown in Fig. 2. The main prolonged for 7 h; the final product contained a noticeable
chemical species were fructose (when glucose was used), amount of ketones (e.g., 1-hydroxy-2-propanone, 2,5-hexane-
5
-HMF, and 5-MFA. Only a trace amount of 5-MFM was dione, 2-methylcyclopentanedione). These components could
detected in the chromatograms, and thus the 5-MFM yield is be produced by the C–C cleavage of C6 to C3, ring-opening
not shown in the figure. With increasing reaction time, the of 2,5-DMF, and/or intramolecular aldol reaction of
4
0
5
-HMF yield decreased, while the yields of 5-MFA and 2.5-DMF 2,5-hexanedione.
increased. This confirms the reaction mechanism proposed in
According to the GC-TOF/MS results, the possible pathways
Scheme 1. In Fig. 2, the other compounds include tetrahydro- of glucose and fructose conversion to produce 2,5-DMF over
furans, ketones and alcohols, as listed in Table S4 in the ESI.† the Pd/UiO66@SGO catalyst are presented in Scheme 1. In the
The activation energies (Fig. S12a and b†) reveal that fructose presence of Lewis acid sites (e.g., CrCl ), glucose can be firstly
3
Green Chem.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Green Chemistry
Communication
isomerized to fructose and subsequently dehydrated over a in Fig. S13a and b,† 2-methyltetrahydrofuran, 1-hydroxy-propa-
1
1
Brønsted acid medium (e.g., aqueous HCl) to 5-HMF. In this none, and 2,5-hexanedione were detected by GC-TOF/MS, and
work, the dehydration step is efficiently activated by SGO, could be produced by side reactions including excess hydro-
which contains a high density of hydrophilic groups such as genation, C–C cleavage, and furan ring-opening at a high
–
COOH and –OH as well as the Brønsted acidic sites of reaction temperature regime. When glucose was used over the
4
1
3
–SO H. The fructose molecule can be readily adsorbed onto 2.4Pd/UiO-66@SGO catalyst, a maximum 2,5-DMF yield of
3
0
the SGO surface or in between the SGO layers trapped inside 43.2 mol% was achieved at 180 °C, and a further increase of
the hydrophobic cage and then dehydrated to form 5-HMF.
4
2
the temperature to 200 °C resulted in a significant drop of the
Two plausible reaction pathways from 5-HMF to 2,5-DMF 2,5-DMF yield to 11.5 mol%. Again, the side products detected
involving two different types of intermediates have been in the GC-TOF/MS chromatograms (Fig. S13c,† 3-hydroxy-
2
0
reported and well-summarized in previous work;
,5-bis(hydroxymethyl)furan (2,5-BHMF) can be formed by 3-methylcyclopent-2-enone) can be caused by side reactions.
hydrogenation of the aldehyde group in 5-HMF, or 5-MFA can To investigate the reusability of the Pd loaded on
be produced by the hydrogenolysis of the hydroxyl group in UiO-66@SGO, the 4.8Pd/UiO-66@SGO catalyst was tested over
-HMF. Subsequently, in the presence of a metal catalyst five cycles of fructose and glucose conversion at 160 °C for 3 h
under a H atmosphere, either 2,5-BHMF can be hydrogeno- under 1 MPa H in THF. After each cycle, the recovered catalyst
either 2-butanone, hydroxy-propanone, 3-methylcyclopentanone, and
2
5
2
2
lyzed or 5-MFA can be hydrogenated to form 5-MFM. As shown was washed several times with THF, acetone, methanol, and
in Fig. S6,† 2,5-BHMF was not observed in the product finally water to remove the unreacted substrate, dried at 70 °C
mixture, thus the 5-MFA pathway is believed to be the domi- for 6 h, and then reused in the next cycle. As shown in Fig. 4,
nant reaction over the Pd/UiO-66@SGO catalyst. A further the fructose conversion activity could be well maintained for
hydrogenolysis reaction can occur to cleave the hydroxyl group up to five cycles of the reusability test. The 2,5-DMF yield
in 5-MFM to form 2,5-DMF.
decreased somewhat rapidly from 70.5 to 58.3 mol% in the
The effect of the reaction temperature on the conversion second run, while decreasing only slightly to 52.1 mol% in the
and product yields over the 2.4Pd/UiO-66@SGO catalyst is subsequent cycles. On the other hand, the 5-HMF yield
shown in Fig. 3. The fructose and glucose conversions increased somewhat rapidly from 5.6 to 14.2 mol% in the
increased up to 90% upon a temperature increase from 140 to second run, and slowly to 19.8 mol% up to the fifth run. This
2
00 °C. When using fructose as the reactant, the 2,5-DMF yield suggests that as the number of cycles increased, the fructose
increased significantly from 35.4 to 68.6 mol% with increases dehydration activity was maintained, but some of the formed
in temperature from 140 to 160 °C, while further increases of 5-HMF was not converted to 2,5-DMF. In the case of glucose
temperature to 180 and 200 °C resulted in the reduction of the conversion, the 2,5-DMF yield slowly decreased from 45.3 to
2
,5-DMF yield to 47.1 and 30.4 mol%, respectively. As shown 35.0 mol%, while the 5-HMF yield slightly increased from 6.5
to 7.5 mol% during five-cycle reactions. After the fifth run, the
catalyst was recovered and analyzed; as shown in Fig. S14a,†
Fig. 3 Effect of reaction temperature on fructose and glucose conver-
sions and product yields. Reaction conditions: 0.2 g feed; 0.2 g 2.4Pd/
Fig. 4 Catalyst recyclability test. Reaction conditions: 0.2 g feed, 0.2 g
UiO-66@SGO; 40 mL THF, at 160 °C 3 h, 1 MPa H
2
. 5-HMF, 5-hydroxy-
4.8Pd/UiO-66@SGO; 40 mL THF, 160 °C (fructose) and 180 °C
methylfurfural;
dimethylfuran.
5-MFA,
5-methylfurfural;
and 2,5-DMF, 2,5-
(glucose), 3 h, 1 MPa H
dimethylfuran.
2
. 5-HMF, 5-hydroxymethylfurfural; 2,5-DMF, 2,5-
This journal is © The Royal Society of Chemistry 2017
Green Chem.

View Article Online
Communication
Green Chemistry
the high crystallinity of the spent catalyst indicates that the Renewable
Energy
Core Technology Program
(no.
UiO-66 framework remained almost unchanged during the 20143030090940) of the Korea Institute of Energy Technology
reaction, which can be the reason for the high fructose conver- Evaluation and Planning (KETEP) financed by the Ministry of
sion and 2,5-DMF yield. The oxygen and sulfur contents in the Trade, Industry, and Energy, Republic of Korea is appreciated.
spent catalyst, measured using elemental analysis, were very
similar to those of the fresh catalyst (Table S6†), suggesting a
good stability of the acid sites in SGO. As shown in Fig. S14b,
c† and Table 1, the BET surface of the spent catalyst decreased
References
2
−1
slightly from 715 to 635 m
g
, while the pore volume
1 Y. Román-Leshkov, C. J. Barrett, Z. Y. Liu and
J. A. Dumesic, Nature, 2007, 447, 982–985.
2 J. B. Binder and R. T. Raines, J. Am. Chem. Soc., 2009, 131,
1979–1985.
3 T. Thananatthanachon and T. B. Rauchfuss, Angew. Chem.,
Int. Ed., 2010, 49, 6616–6618.
3
−1
decreased from 0.40 to 0.25 cm g , indicating that the reac-
tion lowered the exposed surface area. In addition, the Pd
content in the spent catalyst decreased to 4.3 wt%, indicating
that some degree of Pd leaching occurred during the reaction.
nd
This indicates that the loss of catalytic activity after the 2 run
can be caused by Pd leaching because of the lack of proper
binding sites on the surface of the UiO-66 framework. To sup-
press the Pd nanoparticle leaching, it would be necessary to
modify the structure of the UiO-66 framework using electron-
4 S. Zhong, R. Daniel, H. Xu, J. Zhang, D. Turner,
M. L. Wyszynski and P. Richards, Energy Fuels, 2010, 24,
2891–2899.
5 I. F. Teixeira, B. T. W. Lo, P. Kostetskyy, M. Stamatakis,
L. Ye, C. C. Tang, G. Mpourmpakis and S. C. E. Tsang,
Angew. Chem., Int. Ed., 2016, 55, 13061–13066.
6 C. L. Williams, C.-C. Chang, P. Do, N. Nikbin,
S. Caratzoulas, D. G. Vlachos, R. F. Lobo, W. Fan and
P. J. Dauenhauer, ACS Catal., 2012, 2, 935–939.
4
3,44
donating functionalized terephthalate linkers.
The
decrease in the surface area and the Pd leaching can also be
responsible for the slight reduction of the hydrogenolysis/
hydrogenation.
7
P. P. Upare, D. W. Hwang, Y. K. Hwang, U. H. Lee, D.-Y. Hong
and J.-S. Chang, Green Chem., 2015, 17, 3310–3313.
B. Saha, C. M. Bohn and M. M. Abu-Omar, ChemSusChem,
2014, 7, 3095–3101.
Conclusion
8
In summary, we have developed a one-pot strategy for produ-
cing high-yield 2,5-DMF directly from monosaccharides (fruc-
tose and glucose) over a highly efficient Pd loaded integrated
9 X. Xiang, J. Cui, G. Ding, H. Zheng, Y. Zhu and Y. Li, ACS
Sustainable Chem. Eng., 2016, 4, 4506–4510.
Zr–metal organic framework@sulfonated graphene oxide cata- 10 G.-Y. Jeong, A. K. Singh, S. Sharma, K. W. Gyak,
lyst. Along with SGO, the defected UiO-66 could provide R. A. Maurya and D.-P. Kim, NPG Asia Mater., 2015, 7, e173.
additional Brønsted acid sites, which resulted in high-yield 11 V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko,
5
-HMF by the dehydration of fructose. The produced 5-HMF in
V. Nikolakis, N. S. Marinkovic, A. I. Frenkel, S. I. Sandler
the reaction mixture was further converted to 2.5-DMF over Pd
and D. G. Vlachos, J. Am. Chem. Soc., 2013, 135, 3997–4006.
nanoparticles by the hydrogenation/hydrogenolysis of 5-HMF. 12 L. Hu, X. Tang, J. Xu, Z. Wu, L. Lin and S. Liu, Ind. Eng.
From the fructose feed, a maximum 2,5-DMF yield of 70.5 mol% Chem. Res., 2014, 53, 3056–3064.
was achieved at 160 °C in tetrahydrofuran in 3 h without 13 J. M. R. Gallo, D. M. Alonso, M. A. Mellmer and
purification of the intermediate 5-HMF. When glucose was J. A. Dumesic, Green Chem., 2013, 15, 85–90.
used, the 2,5-DMF yield was 45.3 mol% under identical reac- 14 Z. Hu, Y. Peng, Y. Gao, Y. Qian, S. Ying, D. Yuan, S. Horike,
tion conditions because of the additional isomerization step. A
reusability test of the Pd/UiO-66@SGO catalyst indicated that
N. Ogiwara, R. Babarao and Y. Wang, Chem. Mater., 2016,
28, 2659–2667.
the 2,5-DMF yield over the first two runs decreased slightly to 15 G.-H. Wang, J. Hilgert, F. H. Richter, F. Wang,
5
8.3 mol%, but remained at 52.1 mol% over the last three
H.-J. Bongard, B. Spliethoff, C. Weidenthaler and
F. Schüth, Nat. Mater., 2014, 13, 293–300.
runs. It was suggested that the direct conversion of fructose to
2
,5-DMF over Pd/UiO-66@SGO is highly efficient and simple, 16 M. Chidambaram and A. T. Bell, Green Chem., 2010, 12,
and could promote the use of 2,5-DMF as a renewable alterna- 1253–1262.
tive to gasoline in the transportation sector and as a building 17 A. S. Nagpure, N. Lucas and S. V. Chilukuri, ACS
block for value-added chemicals.
Sustainable Chem. Eng., 2015, 3, 2909–2916.
1
1
8 Y. Román-Leshkov, J. N. Chheda and J. A. Dumesic, Science,
2
006, 312, 1933–1937.
9 J. Wang, J. Ren, X. Liu, G. Lu and Y. Wang, AIChE J., 2013,
9, 2558–2566.
Acknowledgements
5
This study was supported by a National Research Foundation 20 Z. Wei, J. Lou, Z. Li and Y. Liu, Catal. Sci. Technol., 2016, 6,
of Korea (NRF) grant funded by the Korean Government 6217–6225.
MSIP) (no. 2016R1A2B3008800, 2015H1D3A1066544, and 21 Y. Wang, Y. Pan, Z. Zhang, R. Sun, X. Fang and D. Yu,
(
2
016R1D1A1B03934060). Additional support by the New &
Process Biochem., 2012, 47, 976–982.
Green Chem.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Green Chemistry
Communication
2
2
2 S. H. Bhosale, M. B. Rao and V. V. Deshpande, Microbiol. 34 J. H. Cavka, S. Jakobsen, U. Olsbye, N. Guillou,
Rev., 1996, 60, 280–300.
C. Lamberti, S. Bordiga and K. P. Lillerud, J. Am. Chem.
Soc., 2008, 130, 13850–13851.
3 M. J. Katz, Z. J. Brown, Y. J. Colón, P. W. Siu, K. A. Scheidt,
R. Q. Snurr, J. T. Hupp and O. K. Farha, Chem. Commun., 35 F. Vermoortele, B. Bueken, G. L. Le Bars, B. Van de Voorde,
2
013, 49, 9449–9451.
M. Vandichel, K. Houthoofd, A. Vimont, M. Daturi,
M. Waroquier and V. Van Speybroeck, J. Am. Chem. Soc.,
2013, 135, 11465–11468.
2
2
4 L. Valenzano, B. Civalleri, S. Chavan, S. Bordiga,
M. H. Nilsen, S. Jakobsen, K. P. Lillerud and C. Lamberti,
Chem. Mater., 2011, 23, 1700–1718.
5 H. Wu, Y. S. Chua, V. Krungleviciute, M. Tyagi, P. Chen,
T. Yildirim and W. Zhou, J. Am. Chem. Soc., 2013, 135,
36 M. Rimoldi, A. J. Howarth, M. R. DeStefano, L. Lin,
S. Goswami, P. Li, J. T. Hupp and O. K. Farha, ACS Catal.,
2016, 7, 997–1014.
1
0525–10532.
37 Z. Wei, R. Pan, Y. Hou, Y. Yang and Y. Liu, Sci. Rep., 2015,
5, 15664.
38 V. Vorotnikov, G. Mpourmpakis and D. G. Vlachos, ACS
Catal., 2012, 2, 2496–2504.
2
2
2
2
6 S. Biniak, G. Szymański, J. Siedlewski and A. Światkowski,
Carbon, 1997, 35, 1799–1810.
7 C. Petit, M. Seredych and T. J. Bandosz, J. Mater. Chem.,
2
009, 19, 9176–9185.
39 S. Wang, V. Vorotnikov and D. G. Vlachos, ACS Catal., 2014,
5, 104–112.
40 E. R. Sacia, M. H. Deaner and A. T. Bell, Green Chem., 2015,
17, 2393–2397.
8 A. J. Crisci, M. H. Tucker, M.-Y. Lee, S. G. Jang,
J. A. Dumesic and S. L. Scott, ACS Catal., 2011, 1, 719–728.
9 P. P. Upare, J.-W. Yoon, M. Y. Kim, H.-Y. Kang,
D. W. Hwang, Y. K. Hwang, H. H. Kung and J.-S. Chang, 41 S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi,
Green Chem., 2013, 15, 2935–2943.
0 D. Verma, R. Tiwari and A. K. Sinha, RSC Adv., 2013, 3,
H. Kato and S. Hayashi, J. Am. Chem. Soc., 2008, 130,
12787–12793.
3
3
3
3
1
3265–13272.
42 Z. Wei, Y. Yang, Y. Hou, Y. Liu, X. He and S. Deng,
ChemCatChem, 2014, 6, 2354–2363.
43 V. Pascanu, Q. Yao, A. Bermejo Gómez, M. Gustafsson,
Y. Yun, W. Wan, L. Samain, X. Zou and B. Martín-Matute,
Chem. – Eur. J., 2013, 19, 17483–17493.
1 J. Cui, J. Tan, T. Deng, X. Cui, Y. Zhu and Y. Li, Green
Chem., 2016, 18, 1619–1624.
2 F. Jin and H. Enomoto, Energy Environ. Sci., 2011, 4, 382–
3
97.
3 Q. Hou, W. Li, M. Ju, L. Liu, Y. Chen and Q. Yang, RSC 44 L. Chen, H. Chen, R. Luque and Y. Li, Chem. Sci., 2014, 5,
Adv., 2016, 6, 104016–104024. 3708–3714.
This journal is © The Royal Society of Chemistry 2017
Green Chem.