One-pot, one-step synthesis of 2,5-diformylfuran from carbohydrates over Mo-containing Keggin heteropolyacids

Article abstract of DOI:10.1002/cssc.201402468

In this work, a one-pot strategy for directly converting fructose into 2,5-diformylfuran (DFF) over Mo-containing Keggin heteropolyacids (HPAs) in open air is developed. H₃PMo₁₂O₄₀ HPA is found to show high activity and selectivity to the formation of DFF owing to its higher Bronsted acidity and moderate redox potential. The partial substitution of the H⁺ in H₃PMo₁₂O₄₀ with Cs⁺ leads to the heterogenization of the HPA by forming its cesium salts Cs_xH_3-xPMo₁₂ (x = 0.5, 1.5, and 2.5). A satisfactory yield of 69.3% to DFF is obtained over Cs_0.5H_2.5PMo₁₂ polyoxometalate after deliberate optimization of the reaction conditions. The heterogenized polyoxometalate could be recycled and reused without significant loss of catalytic activity for five runs. The produced DFF could be separated from the resulting mixture by an adsorption-desorption method using activated carbon as the adsorbent and furfural as the desorbent. A highest isolated yield of 58.2% is obtained.

Full text of DOI:10.1002/cssc.201402468

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201402468
One-pot, One-step Synthesis of 2,5-Diformylfuran from
Carbohydrates over Mo-Containing Keggin
Heteropolyacids
Yu Liu, Liangfang Zhu,* Jinqiang Tang, Mingyang Liu, Ruodi Cheng, and Changwei Hu*[a]
In this work, a one-pot strategy for directly converting fructose
into 2,5-diformylfuran (DFF) over Mo-containing Keggin hetero-
polyacids (HPAs) in open air is developed. H3PMo12O40 HPA is
found to show high activity and selectivity to the formation of
DFF owing to its higher Brønsted acidity and moderate redox
potential. The partial substitution of the H+ in H3PMo12O40 with
Cs+ leads to the heterogenization of the HPA by forming its
cesium salts CsxH3ꢀxPMo12 (x=0.5, 1.5, and 2.5). A satisfactory
yield of 69.3% to DFF is obtained over Cs0.5H2.5PMo12 polyoxo-
metalate after deliberate optimization of the reaction condi-
tions. The heterogenized polyoxometalate could be recycled
and reused without significant loss of catalytic activity for five
runs. The produced DFF could be separated from the resulting
mixture by an adsorption–desorption method using activated
carbon as the adsorbent and furfural as the desorbent. A high-
est isolated yield of 58.2% is obtained.
Introduction
The increasing environmental concerns arisen from the use of
fossil resources accelerate the development of biorefinery to
produce fuels and chemicals from biomass.[1] Thus, there is tre-
mendous amount of interest in developing methods for con-
verting readily available biomass-based carbohydrates to
useful and polyfunctional molecules, e.g., 5-hydroxymethylfur-
fural (HMF).[2] As one of the key platform chemicals bridging
biomass chemistry and petrochemistry, HMF has been recently
used as starting material to produce corresponding deriva-
tives.[3] 2,5-diformylfuran (DFF), the partially oxidized product
from HMF, is considered as a potential intermediate for phar-
maceuticals,[4] fungicides,[5] macrocyclic ligands,[6] as well as
a cross-linking agent for poly(vinylalcohol).[7] Because of its
symmetrical and unsaturated structure, DFF is expected to be
applied as a monomer for novel polymeric materials.[8] The
wide application of DFF makes its synthesis attractive and chal-
lenging.
Recently, Yadav and coworkers[20] obtained a DFF yield as high
as 99% from HMF in the presence of Ag-OMS-2 in isopropyl al-
cohol at 1658C under 15 atm of air. However, although the
yield of DFF is high when synthesized from HMF, the energy-
intensive production and the unstable character of HMF
during the purification step limit the application of pure HMF
as raw feedstock for scale synthesis of DFF. Therefore, DFF is
commercially available only in milligram quantities and the
cost of DFF is very high.
The direct synthesis of DFF from readily available carbohy-
drates provides an alternative approach to reduce its produc-
tion cost.[21] Typically, Takagaki et al.[22] demonstrated a one-pot
approach for the three-step conversion of glucose to DFF via
isomerization of glucose to fructose over hydrotalcite, dehydra-
tion of fructose to HMF over Amberlyst-15, and oxidation of
HMF to DFF over Ru/HT catalysts in N,N-dimethylformamide.
The DFF yields were 25% and 49%, respectively, when using
glucose or fructose as starting materials. In our previous
work,[23] we used CrCl3·6H2O/NaBr as co-catalysts for the simul-
taneous isomerization/dehydration step, while NaVO3·2H2O
was used as catalyst for the oxidation step, realizing the two-
step conversion of glucose to DFF with a maximum yield of
55% in N,N-dimethylacetamide. Recently, Fu and coworkers[24]
reported the joint utilization of Fe3O4-SBA-SO3H and K-OMS-2
catalysts for the two-step conversion of fructose to DFF in di-
methyl sulphoxide (DMSO) under oxygen flow (10 mLminꢀ1).
The yield of DFF was as high as 80% after a reaction time of
8 h. In all cases, however, more than one catalyst was used for
the multi-step conversion of carbohydrates. Therefore, the sep-
aration of catalysts prior to the next step was necessary to ach-
ieve higher selectivity to product. From this point of view, it is
expected to integrate the multiple functions originally per-
formed by several catalysts into one single catalyst, turning
the multi-step reactions into one step, to further lower the
So far, the reported routes for DFF synthesis are mostly
based on the selective oxidation of hydroxymethyl group on
pure HMF. Many typical catalysts, such as VOPO4·2H2O,[9] MnIII-
Salen,[10] PVP/Cu(VO),[11] Cu(NO3)2/VOSO4,[12] VOx/TiO2,[13] V2O5/
zeolite,[14] intercalated vanadium phosphate oxides,[15] Ru/g-alu-
mina,[16] V-Cu-sulfonated carbon,[17] RuCo(OH)2CeO2,[18] and
Fe3O4/Mn3O4,[19] had been found to exhibit good performance.
[a] Y. Liu, Dr. L. Zhu, J. Tang, M. Liu, R. Cheng, Prof. C. Hu
Key Laboratory of Green Chemistry and Technology
Ministry of Education, College of Chemistry
Sichuan university, Chengdu, Sichuan 610064 (P.R. China)
Fax: (+86)28-85411105
E-mail: changweihu@scu.edu.cn
zhulf@scu.edu.cn
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201402468.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
1
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
energy consumption. However, there are few examples of such
one-pot, one-step conversion of carbohydrates to DFF.
Heteropoly acids (HPAs) are well known for their well-de-
fined structure, strong Brønsted acidity, fast multi-electron
transfer, high proton motility, and resistance to hydrolytic and
oxidative degradations in solutions.[25] Actually, the HPAs and
their salts have found several applications in acid and oxida-
tion catalysis for the transformation of hydrocarbons.[26] Re-
cently, Keggin-type polyoxometalates, that is, H3PW12O40,
Cs2.5H0.5PW12O40, and FePW12O40, have been reported as efficient
acid catalysts for the dehydration of fructose[27] or degradation
of cellucose[28] into platform molecules. Up to now, only limited
studies[29] have focused on the bifunctional characteristics of
HPAs for biomass conversion.
Table 1. Effect of PMo11V1 HPA amounts used on the one-pot conversion
of fructose.[a]
Entry Amount of
HPA [mg]
FA Yield
[mol%]
LA Yield
[mol%]
HMF Yield
[mol%]
DFF Yield
[mol%]
1
2
3
4
5
6
7
8
–
1.0
2.5
5.0
10.0
20.0
40.0
60.0
6.2
7.0
9.6
12.9
13.3
15.8
22.3
29.3
3.3
4.3
4.8
6.9
6.7
7.2
7.3
8.7
58.4
10.1
3.6
2.9
–
–
–
–
6.1
50.7
52.4
49.0
46.1
39.7
33.6
26.5
[a] Reaction conditions: 45.0 mg of fructose, 2.5 mg of PMo11V1 HPA, 1 mL
of DMSO, 2 h, in air.
Herein, we report a new approach to directly convert carbo-
hydrates to DFF in a single step using Mo-containing Keggin
HPAs as the only catalyst. The carbohydrates undergo acid-cat-
alyzed dehydration and successive aerobic oxidation through
the bifunctional HPAs catalyst in open air without any separa-
tion and purification of HMF. The components of the catalysts,
as well as the reaction conditions, were deliberately optimized
to enhance the product yield. The approach is valorized by
herterogenation of HPA to form its cesium salts. The recycling
use of the heterogenous HPA catalyst and the separation of
product are also discussed.
mum of 52.4% (2.5 mg of catalyst), and then decreased to
49.0% (5.0 mg of catalyst). A larger amount of PMo11V1 catalyst
led to a further decrease of DFF yield to 26.5% (60.0 mg of cat-
alyst). In comparison, the yield of HMF exhibited a declining
tendency, even to trace amounts, when higher catalyst concen-
tration was used. The yields of FA and LA increased monotoni-
cally, illustrating that hydrolysis of HMF was also promoted by
the HPA catalyst. Moreover, the yield of FA was higher than
that of LA, which implied that FA was produced both through
the hydrolysis of HMF and the decomposition of fructose.[33] As
a support, furfural, the by-product from direct decomposition
of fructose, was detected with a yield of about 4.9% when
2.5 mg of PMo11V1 HPA was used. In all cases, no insoluble
humins-like products were observed. Since fructose was com-
pletely converted, the unbalanced carbons may be ascribed to
the formation of oligomerized byproducts that could not be
detected by GC and HPLC.
Results and Discussion
Keggin-type HPAs, including H4PMo11V1O40 (PMo11V1),
H5PMo10V2O40 (PMo10V2), H4PW11V1O40 (PW11V1), H3PMo12O40
(PMo12), and H5SiMo11V1O40 (SiMo11V1), were synthesized accord-
ing to the literature.[30] Their Keggin structures were confirmed
by IR and XRD spectra (Figure S1–S5).[31] All catalytic reactions
were carried out in a 5 mL glass vessel heated in a tempera-
ture-controlled oil bath with magnetic stirring. Briefly, fructose
(45.0 mg, 0.25 mmol), HPA (2.5 mg), and DMSO (1 mL) were
mixed in the glass vessel before being immersed in the oil
bath. The mixture was incubated at 1608C for 2 h in open air
with continuous stirring under reflux. After cooling to room
temperature, the liquid products were analyzed by GC and
HPLC for quantification.
Time course for the one-pot conversion of fructose to DFF
We analyzed the specimens sampled from the reaction mixture
at specified time intervals. The variation of product yield with
time reveals the evolution of products during DFF formation.
As shown in Figure 1, HMF formed quickly, reaching a maxi-
mum yield of 63.1% at 30 min. As the reaction progressed, the
yield of HMF decreased to trace after 6 h’s reaction, accompa-
nying the increase of DFF yield from 7.0% (15 min) to 64.1%
(6 h). With prolonging the reaction time beyond 6 h, the yield
Effect of the amount of catalyst used
In the absence of catalyst, the incubation of fructose in DMSO
at 1608C for 2 h gave a HMF yield of 58.4% and a DFF yield of
6.1% (Table 1), indicating the dominating reaction of dehydra-
tion by DMSO.[2a,32] The formation of a small amount of DFF
was deemed to mainly result from the aerobic oxidation of
HMF in open air. Significantly, when catalytic amount of
PMo11V1 (1.0 mg, 5ꢁ10ꢀ4 mmol) was added as the catalyst, the
yield of DFF reached 50.7% with fructose being 100% convert-
ed, and HMF (10.1%), formic acid (FA, 7.0%) and levulinic acid
(LA, 4.3%) were detected as the main by-products. The color
of the resulting solution turned from yellow before the reac-
tion to be dark-brown after reaction. With increasing the
amount of PMo11V1, the yield of DFF increased, reaching a maxi-
Figure 1. Variation of product yields with reaction time. Reaction conditions:
450.0 mg of fructose, 25.0 mg of PMo11V1, 10 mL of DMSO, 1608C, in air.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
2
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
of DFF declined gradually, which was possibly ascribed to the
further reaction of DFF to form oligomers under the reaction
conditions performed. As such oligomers could not be detect-
ed by GC or HPLC, we measured the viscosity of the resulting
solution with a single-point method by Ubbelohde viscometer
to correlate the molecular weight of the oligomers with the
viscosity of the solution (Table S1).[31] The increasing value of
the viscosity with prolonged reaction time over 6 h indicated
the increase of the molecular weight of the oligomers. Mean-
while, the yields of LA and FA exhibited increasing tendency in
the whole reaction period, which implied the aggravation of
the hydrolysis of HMF. Notably, the consumption of HMF and
the accumulation of DFF with time strongly confirm that DFF
formed via aerobic oxidation of HMF, the primary dehydrated
product from fructose, without addition of a second catalyst.
the acid strength of these HPAs.[35] However, the yields of DFF
over these HPAs displayed different order, that is, PMo12 >
PMo11V1 >PMo10V2 >SiMo11V1. In addition to the acid strength,
the redox potential of the HPAs may also play critical role on
the formation of DFF. As the redox potentials of these HPAs
fall in the order of PMo10V2 >PMo11V1 >PMo12 >SiMo11V1,[36] the
higher DFF yield over PMo10V2 than over SiMo11V1 may be the
result of the higher redox potential of PMo10V2 in comparison
with SiMo11V1.
Putting these together, we can conclude that the bifunction-
al performance (dehydration and oxidation) of the Keggin
HPAs is due to the synergistic effects of their intrinsic Brønsted
acidity and redox potential. Higher Brønsted acidity of HPA
may lead to higher yield of dehydrated product (HMF), while
higher redox potential favors the subsequent oxidation step
for DFF formation. Relative to the redox potential, the acidity
of the HPAs may play a major role because the dehydration of
fructose is the prerequisite procedure for the one-step conver-
sion of fructose to DFF. Among the HPA catalysts investigated,
the PMo12 HPA exhibited the highest activity towards the for-
mation of DFF, which resulted from the synergistic contribution
of its higher Brønsted acidity and moderate redox potential.
Effect of the redox potential and acidity of the HPAs
We were interested in the high selectivity towards DFF forma-
tion in this one-pot reaction. In a control experiment, we con-
ducted the reaction using pure HMF as the starting material
over PMo11V1. A DFF yield of 94% was obtained, accompanying
a high selectivity of 94% towards DFF. The high oxidation per-
formance of PMo11V1 was deemed to be closely related to its
redox potential. In another control experiment, we used
PW11V1, the HPA with far lower redox potential than PMo11V1,[34]
as the catalyst for the conversion of fructose. As expected, the
PW11V1 HPA was more selective towards the formation of HMF
(Figure 2), suggesting that the lower redox potential was un-
favorable for the oxidation of HMF to DFF.
Optimization of the reaction conditions over PMo12 HPA
In the presence of PMo12, the reaction conditions affecting the
catalytic activity of HPA were optimized to further enhance the
DFF yield. As plotted in Figure 3, within the temperature range
Figure 3. One-pot conversion of fructose to DFF over PMo12 HPA at different
reaction temperatures. Reaction conditions: 45.0 mg of fructose, 2.5 mg of
PMo12 HPA, 1 mL of DMSO, 2 h, in air.
Figure 2. Catalytic conversion of fructose over various Keggin-type HPAs. Re-
action conditions: 45.0 mg of fructose, 2.5 mg of HPAs, 1 mL of DMSO,
1608C, 2 h, in air.
investigated, the yield of HMF increased from 36.1% at 808C
to 56.0% at 1208C, then decreased to 4.2% at 1608C. Mean-
while, the yield of DFF increased continuously with increasing
temperature to 1608C (60.3%), and then decreased at 1708C
(57.9%). The viscosity of the resulting solution after reaction at
1708C was higher than that at 1608C (Table S1),[31] suggesting
the aggravated formation of oligomers at much higher reac-
tion temperature. Moreover, the yield variations of HMF and
DFF at various reaction temperatures implied that the oxida-
tion hardly occurred at lower reaction temperature, while oxi-
dation and dehydration were both favored at higher tempera-
ture. Obviously, the product selectivity to HMF or DFF over
PMo12 HPA can be simply tuned by controlling the reaction
Furthermore, the overall yields of DFF and HMF over PW11V1
were higher than those over PMo11V1, indicating a higher dehy-
dration ability of the former HPA. As the Brønsted acidity of
PW11V1 was stronger than that of PMo11V1,[35] it is possible to
correlate the dehydration performance of the HPA catalysts to
their Brønsted acidity. In other control experiments, the catalyt-
ic performances of PMo10V2, PMo12, and SiMo11V1 HPAs on the
conversion of fructose were compared under identical reaction
conditions. As illustrated in Figure 2, the activity based on the
overall yields of DFF and HMF declined in the order PMo12 >
PMo11V1 >SiMo11V1 >PMo10V2, which was well consistent with
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
3
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
of glucose-based carbohydrates may be achieved by endowing
the catalyst with Lewis acidity or base sites for isomerizing glu-
cose to fructose.[37] On the other hand, for the conversion of
inulin, the polymerized form of fructose, a DFF yield of 32.0%
(based on fructose) was obtained. The lower activity for the
conversion of inulin was possibly owing to the use of DMSO
solvent, known to be unfavorable for depolymerizing the poly-
saccharides.[38]
Table 2. Optimization of the reaction conditions for the one-pot conver-
sion of fructose to DFF over PMo12 HPA.[a]
Entry Amount of PMo12 HPA Reaction
DFF Yield
[mol%]
HMF Yield
[mol%]
used [mg]
time [h]
1
2
3
4
5
6
7
8
1.0
2.5
4.0
2.5
2.5
2.5
2.5
2.5
2
2
2
4
5
6
7
9
51.0
60.3
54.0
63.5
16.0
4.2
1.1
6.7
3.0
0.9
–
64.8
68.0(57.1)[b]
66.7
Heterogenization of the Keggin HPA
61.6
–
[a] Reaction conditions: 45.0 mg of fructose, 1 mL of DMSO, 1608C, in air.
[b] Isolated yield in parentheses.
Although the PMo12 HPA showed higher catalytic activity to-
wards the production of DFF from fructose, its homogeneous
nature made practical applications difficult because it could
not be separated from the resulting mixture for recycling. As
a solution, we partially substituted H+ in the PMo12 HPA by
Cs+, obtaining series of CsxH3ꢀxPMo12 (x=0.5, 1.5, and 2.5)
polyoxometalates that were known to be insoluble in water or
any polar solvents.[25a] The XRD and IR spectra of the as-synthe-
sized CsxH3ꢀxPMo12 polyoxometalates are shown in Figure S7a–
c and S8.[31] The partial substitution of H+ with Cs+ was
deemed to lower the Brønsted acidity of the polyoxometa-
lates,[39] that is, the Brønsted acidity of the polyoxometalates
declined as follows: PMo12 >Cs0.5H2.5PMo12 >Cs1.5H1.5PMo12 >
Cs2.5H0.5PMo12. Meanwhile, the redox potential of the Cs-substi-
tuted polyoxometalates showed declining tendency on in-
creasing the number of the Cs+ ions. As shown in Tables 2 and
3, the DFF yield over Cs0.5H2.5PMo12 was lower than that of
PMo12 (50.4% vs. 60.3%), but higher than those of
Cs1.5H1.5PMo12 and Cs2.5H0.5PMo12 (54.8% vs. 51.0% and 46.4%)
under identical reaction conditions. The results further support
our previous deduction that higher Brønsted acidity and redox
potential of the HPAs favor the dehydration of fructose and
successive oxidation of HMF to DFF. Although the heterogeni-
zation of the HPA led to a small decrease in the catalytic per-
formance, a considerable DFF yield of 69.3% was obtained
temperature. A similar temperature-sensitive selectivity for DFF
or HMF formations was observed for the PMo11V1 HPA catalyst
(Figure S6).[31]
The effects of the amount of PMo12 used and the reaction
time on the reaction are shown in Table 2. Similar to PMo11V1,
a moderate amount (2.5 mg) of the PMo12 HPA used resulted
in a higher yield of DFF. Moreover, the catalytic activity for DFF
formation increased on prolonging the reaction time to 6 h,
giving a higher DFF yield of 68.0%. Further reaction beyond
6 h led to slight decrease of both DFF and HMF yields, possibly
owing to the oligomerization of the intermediates and prod-
ucts.
One-pot conversion of disaccharide and polysaccharide to
DFF
To expand the substrate scope of the bifunctional Mo-contain-
ing Keggin HPA catalyst, we applied the PMo12-DMSO system
to convert monosaccharide, disaccharide, and polysaccharide
(Figure 4). However, when using glucose as the starting materi-
al, the yields of both HMF and DFF were low (1.0% for HMF
and 8.4% for DFF). It seemed that the PMo12 HPA lacked the
ability of simultaneous isomerization and dehydration of this
aldose. Similarly, the conversion of sucrose, the disaccharide
composed of glucose and fructose, was only efficient for con-
verting the fructose part, achieving a DFF yield of 28.9%
(based on monosaccharide). Actually, the effective conversion
Table 3. One-pot conversion of fructose to DFF using heterogeneous
CsxH3ꢀxPMo12 (x=0.5, 1.5, and 2.5) catalysts.[a]
Entry Catalyst
Amount of cata- Reaction HMF
lyst used [mg] time [h] Yield
[mol%]
DFF Yield
[mol%]
1
2
3
4
5
6
7
8
Cs0.5H2.5PMo12 15.0
Cs1.5H1.5PMo12 15.0
2
2
2
2
2
2
2
2
3
4
5
6
7
–
54.8
51.0
46.4
50.4
57.2
56.8
54.7
54.1
1.0
11.1
8.0
0.8
0.1
0.1
–
0.5
0.2
0.1
–
Cs2.5H0.5PMo12 15.0
Cs0.5H2.5PMo12 2.5
Cs0.5H2.5PMo12 5.0
Cs0.5H2.5PMo12 10.5
Cs0.5H2.5PMo12 20.0
Cs0.5H2.5PMo12 40.0
Cs0.5H2.5PMo12 5.0
Cs0.5H2.5PMo12 5.0
Cs0.5H2.5PMo12 5.0
Cs0.5H2.5PMo12 5.0
Cs0.5H2.5PMo12 5.0
9
61.8
10
11
12
13
69.3(58.2)[b]
53.1
51.5
48.7
–
Figure 4. Catalytic conversion of carbohydrates with PMo12 HPA. Reaction
conditions: 45.0 mg of carbohydrates, 2.5 mg of PMo12, 1 mL of DMSO,
1608C, 2 h, in air.
[a] Reaction conditions: 45.0 mg of fructose, 1 mL of DMSO, 1608C, in air.
[b] Isolated yield in parentheses.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
4
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
over Cs0.5H2.5PMo12 after careful optimization of the reaction
conditions (Table 3, entry 10).
amount of furfural used on the desorption efficiency were in-
vestigated (Table S3).[31] When the amount of furfural increased
from 5 to 10 mL, the desorption efficiency increased from 73%
to 89%. While further increasing the furfural amount to 15 mL
led to slight decrease of the desorption efficiency to 84%.
When 10 mL of furfural was used to treat 1.8 g of AC, about
90% of the absorbed DFF could be desorbed. Overall, about
84% of the produced DFF was isolated from the resulting reac-
tion mixture by this adsorption–desorption method.
Recycling of the catalyst
We investigated the reusability of the used Cs0.5H2.5PMo12 cata-
lyst under identical reaction conditions. The used catalyst was
simply separated from the resulting reaction mixture by centri-
fugation, followed by washing with ethanol and drying under
vacuum at 508C. As plotted in Figure 5, the yield of DFF de-
Conclusions
In summary, we developed a simple and effective method to
synthesize DFF directly from fructose in open air. The Mo-con-
taining Keggin-type HPAs possessed both Brønsted acidity and
oxidability, which synergistically contributed to the dehydra-
tion and successive aerobic oxidation of the carbohydrates.
The produced DFF could be separated from the resulting mix-
ture with an adsorption–desorption method using AC as ad-
sorbent and furfural as desorbent. Higher DFF yield of 68.0%
(isolated yield of 57.1%) and 69.3% (isolated yield of 58.2%)
were achieved over PMo12 and Cs0.5H2.5PMo12 HPAs, respective-
ly, under optimal reaction conditions. The heterogenized
Cs0.5H2.5PMo12 polyoxometalate could be recycled and reused
without significant loss of the catalytic activity. This one-pot,
one-step conversion represents a new strategy, circumventing
the limitation of expensive raw materials (i.e., HMF) for DFF
production. It is also low energy-consuming because no sepa-
ration of reaction intermediates is required.
Figure 5. Recycle use of the Cs0.5H2.5PMo12 for the one-pot conversion of
fructose to DFF. Reaction conditions: 45.0 mg of fructose, 15.0 mg of
Cs0.5H2.5PMo12 HPA, 1 mL of DMSO, 1608C, 2 h, in air.
clined slightly in the first two runs while remaining stable after
the third cycle. No significant loss of the catalytic activity was
observed after five catalytic cycles. When compared with the
fresh Cs0.5H2.5PMo12 catalyst, no obvious change was observed
on the XRD pattern of the used Cs0.5H2.5PMo12 catalyst after the
fifth catalytic run (Figure S7d),[31] indicating that the structure
of the heterogeneous catalyst remained stable during the cata-
lytic reaction.
Experimental Section
Materials
All of the following chemical reagents were used without further
purification. HMF (98%), inulin (95%), and Cs2CO3 (99.9%) were
purchased from J&K Chemical. FA (98%) was purchased from
Sigma Aldrich. LA (98%) was purchased from Acros Organics. DFF
(98%) was purchased from Tokyo Chemical Industry. d-fructose (BR
grade), glucose (AR grade), sucrose (AR grade), Na2CO3 (AR grade),
and Na2SiO3·9H2O (AR grade) were obtained commercially from
Kolong Chemical Company (Chengdu, China). DMSO was pur-
chased from Fuyu Fine Chemical Company (Tianjin, China).
Na2HPO4·12H2O (Chemically pure) was purchased from Chengdu
Chemical Reagent Company. Na2MoO4·2H2O and Na2WO4·9H2O in
AR grade were purchased from Tianjin Chemical Reagent Compa-
ny. V2O5 (AR grade) was purchased from Fucheng Chemical Re-
Product separation
We tried to separate DFF from the resulting reaction mixture
by an adsorption–desorption method. Activated carbon (AC,
40–80 mesh) was used as the adsorbent to absorb DFF from
the resulting reaction mixture, and furfural was used as the de-
sorbent to desorb the absorbed DFF from the AC. In the ad-
sorption step, certain amount of AC was added to 2 mL of the
resulting reaction mixture, then the mixture was stirring at
room temperature for adsorption. The effects of the source of
AC (coconut shell, nutshell, and coal), the amount of AC used
as well as the adsorption time on the adsorption efficiency of
DFF were studied (Table S2).[31] The AC from coconut shell ex-
hibited good performance for effective adsorption of DFF from
the resulting reaction mixture. The adsorption efficiency in-
creased with increasing the amount of the coconut shell AC
and the adsorption time. About 93% of the produced DFF
could be separated from the resulting reaction mixture by
1.8 g of the coconut shell AC after stirring at room tempera-
ture for 12 h. In the desorption step, the coconut shell AC con-
taining adsorbed DFF was immersed in furfural, followed by
stirring at room temperature for desorption. The effect of the
agent Company (Tianjin, China). AC (BET surface area=960 m2 gꢀ1
)
was purchased from Jiangsu Nantong Activated Carbon Company
(China). Deionized water (conductivity=18.25 MW·cmꢀ1) was used
in all of our experiments.
Catalyst preparation
PMo11V1: Na2HPO4·12H2O (3.6 g) was dissolved in water (50 mL), fol-
lowed by mixing with Na2MoO4·2H2O (26.6 g) that had been dis-
solved in water (160 mL). The mixture was heated to boiling for
30 min by the electric heater, forming solution I. Na2CO3 (0.53 g)
was dissolved in water (50 mL) and then V2O5 (0.9 g) was added.
The color of the mixture gradually changed from colorless to
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
5
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
orange, forming solution II. Then, the two solutions were further
mixed, heated at 908C for 30 min, and then cooled to room tem-
perature gradually. The color of the resulting mixture turned to
light yellow. Finally, concentrated sulfuric acid was added slowly to
the above mixture with vigorous stirring until the pH value of the
mixture was approximately 2. A red mixture containing PMo11V1
HPA was then obtained. The HPA was then extracted with ethyl
ether (150 mL). In this extraction, the heteropoly etherate existed
in the middle layer. After separation, a stream of air was passed
through the heteropoly etherate layer to free it of ether. The
yellow solid remained was dissolved in a small amount of distilled
water, concentrated to the first appearance of crystals in a vacuum
desiccator, and then allowed to be crystallized further. The quantity
of product was 15.1 g.
Briefly, fructose (45.0 mg, 0.25 mmol), HPA (2.5 mg), and DMSO
(1 mL) were mixed in the glass vessel before immersion in the oil
bath. The mixture was incubated at 1608C for 2 h in open air with
continuous stirring under reflux. After the reaction, the mixture
was gradually cooled to room temperature. To extend the sub-
strate scope, glucose (45 mg), sucrose (45 mg), or inulin (45 mg)
were used as the substrates instead of fructose.
To monitor the product evolution with time, a certain amount of
samples was removed from the reaction mixture over the range
from 15 min to 10 h at specified time intervals. The volume of the
catalytic reaction was magnified by 10 times to allow adequate
sampling. Briefly, fructose (450.0 mg, 2.5 mmol), HPAs (25.0 mg),
and DMSO (10 mL) were mixed in the glass vessel before being im-
mersed in the oil bath. The mixture was incubated at 1608C in
open air with continuous stirring under reflux. The samples re-
moved from the reaction mixture at specified time were cooled to
room temperature quickly to quench the reaction.
PMo10V2: The synthesis of PMo10V2 HPA was similar to that of
PMo11V1 except that Na2HPO4·12H2O (3.66 g), Na2MoO4·2H2O
(24.25 g), and V2O5 (2.3 g) were used. The quantity of product was
14.3 g.
SiMo11V1: The synthesis of SiMo11V1 HPA was similar to that of
PMo11V1 except that Na2HPO4·12H2O (3.36 g), Na2SiO3·9H2O
(32.04 g), and NH4VO3 (1.40 g) were used. The quantity of product
was 9.5 g.
Product analysis
The liquid products were qualitatively analyzed by GC–MS (Aglient,
5973 Network 6890N). The quantitative analysis of HMF and DFF
was performed on GC (FILI, GC-9790) with an Innowax capillary
column (30 mꢁ0.25 mm), a hydrogen flame-ionization detector
(FID) and ZB-2020 integrator under the following conditions: injec-
tor temperature 2708C, detector temperature 2708C, phenol was
used as the internal standard substance. The amounts of remaining
fructose, and formed FA and LA were analyzed with HPLC (Dionex,
U-3000) using a dionex PG-3000 pump, an aminex column HPX-87
column (Bio-Rad), and shodex 101 refractive index detector.
0.005m of H2SO4 solution was used as the mobile phase at a flow
rate of 0.6 mLminꢀ1, and the temperatures of the column and the
detector were 508C and 358C, respectively. The yields of products
were calculated based on external standard curves constructed
with authentic samples. The yields of products are defined as fol-
lows:
PW11V1: The synthesis of PW11V1 HPA was similar to that of PMo11V1
except that Na2HPO4·12H2O (3.60 g), Na2WO4·9H2O (36.30 g), and
V2O5 (0.91 g) were used. The quantity of product was 13.35 g.
PMo12: Na2HPO4·12H2O (3.69 g) was dissolved in water (50 mL), fol-
lowed by being mixed with Na2MoO4·2H2O (29.50 g) that had been
dissolved in water (100 mL). The mixture was heated to boiling for
30 min by the electric heater. Then, concentrated sulfuric acid was
added slowly to the mixture with vigorous stirring until the pH
value of the mixture was about 2. The color of the mixture gradu-
ally changed from colorless to yellow. The PMo12 HPA was then ex-
tracted with ethyl ether (150 mL). In this extraction, the heteropoly
etherate existed in the bottom layer. After separation, a stream of
air was passed through the heteropoly etherate layer to free it of
ether. The orange solid remained was dissolved in a small amount
of distilled water, concentrated to the appearance of crystals in
a vacuum desiccator, and then allowed to be crystallized further.
The quantity of product was 14.3 g.
CsxH3ꢀxPMo12O40: The cesium salts CsxH3ꢀxPMo12O40 (x=0.5, 1.5, and
2.5) were prepared by titration method. An example for the prepa-
ration of Cs0.5H2.5PMo12O40 is given as follows: H3PMo12O40 (3.6 g)
was dissolved in deionized water (20 mL) at room temperature.
Then, Cs2CO3 solution (0.5m, 2.2 mL) was added slowly with vigo-
rous stirring, forming precipitate in the solution. The resulting solu-
tion was aged at room temperature overnight. After centrifugation
and drying at 353 K, the obtained precipitate was calcined at 473 K
for 3 h.
Yield of DFF ðmol %Þ ¼
ðmole of DFF producedÞ=ðmole of starting fructoseÞ ꢁ 100 %
Yield of HMF ðmol %Þ ¼
ðmole of HMF producedÞ=ðmole of starting fructoseÞ ꢁ 100 %
Yield of FA ðmol %Þ ¼
ðmole of FA producedÞ=ðmole of starting fructoseÞ ꢁ 100 %
Yield of LA ðmol %Þ ¼
ðmole of LA producedÞ=ðmole of starting fructoseÞ ꢁ 100 %
Catalyst Characterizations
Product separation
X-ray diffraction (XRD) patterns of the HPA catalysts were taken
with a LTD DX-1000 CSC diffraction instrument operated at 40 kV
and 25 mA using nickel-filtered CuKa radiation (l=1.5405 ꢂ). A
scan rate of 28/min with a step size of 0.038 was used for data col-
lection in the 2q range of 5–708. Infrared (IR) spectra of the HPAs
were recorded on a Nicolet 6700 spectrometer in KBr disks at
room temperature.
Adsorption of DFF from the resulting reaction mixture: AC (40–
80 mesh) was used as the adsorbent to absorb DFF from the re-
sulting mixture containing DFF, HMF, HPA, by-products, and DMSO.
Typically, AC (1.4~1.8 g) was added to the resulting mixture (2 mL).
The mixture was stirred at room temperature for 12 h, followed by
suction filtration. The absorption efficiency of DFF by the AC was
calculated by quantifying the filtrate with GC.
Desorption of DFF from the AC: Furfural was used as the desorbent
for desorbing DFF from the AC. The AC contained adsorbed DFF
were immersed in furfural (5–15 mL), followed by stirring at room
temperature for 1–5 h. The desorption efficiency of DFF by furfural
Catalytic reaction
All catalytic reactions were carried out in a 5 mL glass vessel
heated in a temperature-controlled oil bath with magnetic stirring.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
6
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[16] C. A. Antonyraj, J. Jeong, B. Kim, S. Shin, S. Kim, K.-Y. Lee, J. K. Cho, J.
Ind. Eng. Chem. 2013, 19, 1056–1059.
[17] N. T. Le, P. Lakshmanan, K. Cho, Y. Han, H. Kim, Appl. Catal. A 2013, 464,
305–312.
[18] Y. Wang, B. Liu, K. Huang, Z. Zhang, Ind. Eng. Chem. Res. 2014, 53,
1313–1319.
were calculated by quantifying the stripping liquid with GC. The
separation efficiency of this adsorption–desorption method was
evaluated by multiplying the value of the adsorption efficiency of
AC by that of the desorption efficiency of furfural.
[19] B. Liu, Z. Zhang, K. Lv, K. Deng, H. Duan, Appl. Catal. A 2014, 472, 64–
71.
Acknowledgements
[20] G. D. Yadav, R. V. Sharma, Appl. Catal. B 2014, 147, 293–301.
[21] G. A. Halliday, R. J. Young, V. V. Grushin, Org. Lett. 2003, 5, 2003–2005.
[22] A. Takagaki, M. Takahashi, S. Nishimura, K. Ebitani, ACS Catal. 2011, 1,
1562–1565.
[23] X. Xiang, L. He, Y. Yang, B. Guo, D. Tong, C. Hu, Catal. Lett. 2011, 141,
735–741.
This work was supported by the National Basic Research Pro-
gram (973 program, No. 2013CB228103), NNSFC (No. 21072136),
and the Special Research Fund for the Doctoral Program of
Higher Education (No. 20120181130014) of China, as well as the
Science and Technology Program of Sichuan (No. 2013JY0015)
and Chengdu (No. 12DXYB339JH-002).
[24] Z.-Z. Yang, J. Deng, T. Pan, Q.-X. Guo, Y. Fu, Green Chem. 2012, 14,
2986–2989.
[25] a) J. Zhang, F. P. Xiao, J. Hao, Y. G. Wei, Dalton Trans. 2012, 41, 3599–
3615; b) T. Okuhara, Chem. Rev. 2002, 102, 3641; c) D. L. Long, R. Tsuna-
shima, L. Cronin, Angew. Chem. Int. Ed. 2010, 49, 1736–1758; Angew.
Chem. 2010, 122, 1780–1803; d) C. L. Hill, Chem. Rev. 1998, 98, 1–2.
[26] a) F. Cavani, Catal. Today 1998, 41, 73–86; b) I. V. Kozhevnikov, Chem.
Rev. 1998, 98, 171–198; c) N. Mizuno, M. Misono, Chem. Rev. 1998, 98,
199–217.
[27] a) K.-i. Shimizu, R. Uozumi, A. Satsuma, Catal. Commun. 2009, 10, 1849–
1853; b) Q. Zhao, L. Wang, S. Zhao, X. Wang, S. Wang, Fuel 2011, 90,
2289–2293.
[28] a) K.-i. Shimizu, H. Furukawa, N. Kobayashi, Y. Itaya, A. Satsuma, Green
Chem. 2009, 11, 1627–1632; b) J. Tian, J. H. Wang, S. Zhao, C. Jiang, X.
Zhang, X. H. Wang, Cellulose 2010, 17, 587–594; c) W. P. Deng, Q. H.
Zhang, Y. Wang, Dalton Trans. 2012, 41, 9817–9831.
[29] a) A. Onda, T. Ochi, K. Yanagisawa, Catal. Commun. 2011, 12, 421–425;
b) J. Z. Zhang, X. Liu, M. N. Hedhili, Y. H. Zhu, Y. Han, ChemCatChem
2011, 3, 1294–1298; c) D. L. An, A. H. Ye, W. P. Deng, Q. H. Zhang, Y.
Wang, Chem. Eur. J. 2012, 18, 2938–2947.
[30] G. A. Tsigdinos, C. J. Hallada, Inorg. Chem. 1968, 7, 437–441.
[31] See Supporting Information for details.
[32] Q. X. Bao, K. Qiao, D. Tomida, C. Yokoyama, Catal. Commun. 2008, 9,
1383–1388.
[33] J. Zhang, E. Weitz, ACS Catal. 2012, 2, 1211–1218.
[34] M. N. Timofeeva, Appl. Catal. A 2003, 256, 19–35.
[35] A. Martin, U. Armbruster, H. Atia, Eur. J. Lipid Sci. Technol. 2012, 114, 10–
23.
[36] I. V. Kozhevnikov, J. Mol. Catal. A 2007, 262, 86–92.
[37] a) Y. Yang, C. W. Hu, M. A. Mahdi, Green Chem. 2012, 14, 509–513;
b) H. P. Yan, Y. Yang, D. M. Tong, X. Xiang, C. W. Hu, Catal. Commun.
2009, 10, 1558–1563; c) Y. Yang, C. W. Hu, M. A. Mahdi, ChemSusChem
2012, 5, 405–410.
Keywords: 2,5-diformylfuran · carbohydrates · molybdenum ·
one-pot reaction · polyoxometalates
[1] J. R. Rostrup-Nielsen, Science 2005, 308, 1421–1422.
[2] a) Y. Roman-Leshkov, J. N. Chheda, J. A. Dumesic, Science 2006, 312,
1933–1937; b) H. Zhao, J. E. Holladay, H. Brown, Z. C. Zhang, Science
2007, 316, 1597–1600; c) Z. Huang, W. Pan, H. B. Zhou, F. Qin, H. Xu, W.
Shen, ChemSusChem 2013, 6, 1063–1069; d) A. Dutta, D. Gupta, B.
Saha, A. Bhaumik, ChemSusChem 2014, 7, 925–933.
[3] a) O. O. James, S. Maity, L. A. Usman, K. O. Ajanaku, O. O. Ajani, T. O.
Siyanbola, S. Sahu, R. Chaubey, Energy Environ. Sci. 2010, 3, 1833–1850;
b) M. M. Antunes, P. A. Russo, P. V. Wiper, J. M. Veiga, M. Pillinger, L.
Mafra, D. V. Evtuguin, N. Pinna, A. A. Valente, ChemSusChem 2014, 7,
804–812.
[4] K. T. Hopkins, W. D. Wilson, B. C. Bender, D. R. McCurdy, J. E. Hall, R. R.
Tidwell, A. Kumar, M. Bajic, D. W. Boykin, J. Med. Chem. 1998, 41, 3872–
3878.
[5] M. D. Poeta, et al., Antimicrob. Agents Chemother. 1998, 42, 2495–2502,
see Supporting Information.
[6] D. T. Richter, T. D. Lash, Tetrahedron Lett. 1999, 40, 6735–6738.
[7] D. W. Sheibley, M. A. Manzo, O. D. Gonzalez-Sanabria, J. Electrochem.
Soc. 1983, 130, 255–259.
[8] a) A. Gandini, N. M. Belgacem, Polym. Int. 1998, 47, 267–276; b) A. S.
Amarasekara, D. Green, L. D. Williams, Eur. Polym. J. 2009, 45, 595–598.
[9] C. Carlini, P. Patrono, A. M. R. Galletti, G. Sbrana, V. Zima, Appl. Catal. A
2005, 289, 197–204.
[10] A. S. Amarasekara, D. Green, E. McMillan, Catal. Commun. 2008, 9, 286–
288.
[11] O. C. Navarro, A. C. Canꢃs, S. I. Chornet, Top. Catal. 2009, 52, 304–314.
[12] J. P. Ma, Z. T. Du, J. Xu, Q. H. Chu, Y. Pang, ChemSusChem 2011, 4, 51–
54.
[38] A. Dan, S. Ghosh, S. P. Moulik, Biopolymers 2009, 91, 687–699.
[39] M. Sun, J. Zhang, C. Cao, Q. Zhang, Y. Wang, H. Wan, Appl. Catal. A
2008, 349, 212–221.
[13] J. F. Nie, H. C. Liu, Pure Appl. Chem. 2012, 84, 765–777.
[14] I. Sꢄdaba, Y. Y. Gorbanev, S. Kegnaes, S. S. R. Putluru, R. W. Berg, A. Riis-
ager, ChemCatChem 2013, 5, 284–293.
Received: May 26, 2014
[15] F. L. Grasset, B. Katryniok, S. Paul, V. Nardello-Rataj, M. Pera-Titus, J.-M.
Clacens, F. De Campo, F. Dumeignil, RSC Adv. 2013, 3, 9942–9948.
Published online on && &&, 0000
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
7
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
Y. Liu, L. Zhu,* J. Tang, M. Liu, R. Cheng,
C. Hu*
One-stop shop for DFF: 2,5-Diformyl-
furan (DFF) was directly produced from
readily available carbohydrates in a one-
pot, one-step method over Mo-contain-
ing Keggin heteropolyacids (HPAs) with-
out any separation and purification of
5-hydroxymethylfurfural (HMF). A high
DFF yield was obtained and the pro-
duced DFF could be separated from the
resulting mixture by an adsorption–de-
sorption method using activated carbon
as the adsorbent and furfural as the de-
sorbent.
&& – &&
One-pot, One-step Synthesis of 2,5-
Diformylfuran from Carbohydrates
over Mo-Containing Keggin
Heteropolyacids
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 8
&
8
&
ÞÞ
These are not the final page numbers!