Efficient aerobic oxidation of biomass-derived 5-hydroxymethylfurfural to 2,5-diformylfuran catalyzed by magnetic nanoparticle supported manganese oxide

Article abstract of DOI:10.1016/j.apcata.2013.12.014

Magnetic Fe₃O₄ supported Mn₃O₄ nanoparticles (Fe₃O₄/Mn₃O₄) were prepared by the solvent thermal method, and its structure was characterized by XRD, XPS, TEM, and FT-IR technologies. The resulting Fe₃O ₄/Mn₃O₄ nanoparticles could be used as an excellent heterogeneous catalyst for the aerobic oxidation of the biomass-derived model molecule 5-hydroxymethylfurfural into 2,5-diformylfuran (DFF) under benign reaction conditions. Some important reaction parameters such as reaction temperature, catalyst amount, solvent, oxidant, and oxygen pressure were explored and high DFF yield of 82.1% with HMF conversion of 100% were obtained in DMF under optimal reaction conditions. More importantly, the catalyst could be readily separated from the reaction mixture by a permanent magnet, and recycled up to 6 times without the significant loss of its catalytic activity.

Full text of DOI:10.1016/j.apcata.2013.12.014

Applied Catalysis A: General 472 (2014) 64–71
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Efficient aerobic oxidation of biomass-derived
-hydroxymethylfurfural to 2,5-diformylfuran catalyzed by magnetic
nanoparticle supported manganese oxide
5
a
a,∗
a
a
b,∗
Bing Liu , Zehui Zhang , Kangle Lv , Kejian Deng , Hongmin Duan
a
Department of Chemistry, Key Laboratory of Catalysis and Material Sciences of the State Ethnic Affairs Commission & Ministry of Education, South-Central
University for Nationalities, Wuhan 430074, China
Dalian National Laboratory for Clean Energy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023,
b
China
a r t i c l e i n f o
a b s t r a c t
Article history:
Magnetic Fe O supported Mn O nanoparticles (Fe O /Mn O ) were prepared by the solvent thermal
3
4
3
4
3
4
3
4
Received 10 September 2013
Received in revised form 2 December 2013
Accepted 10 December 2013
Available online 17 December 2013
method, and its structure was characterized by XRD, XPS, TEM, and FT-IR technologies. The resulting
Fe3O4/Mn3O4 nanoparticles could be used as an excellent heterogeneous catalyst for the aerobic oxidation
of the biomass-derived model molecule 5-hydroxymethylfurfural into 2,5-diformylfuran (DFF) under
benign reaction conditions. Some important reaction parameters such as reaction temperature, catalyst
amount, solvent, oxidant, and oxygen pressure were explored and high DFF yield of 82.1% with HMF
conversion of 100% were obtained in DMF under optimal reaction conditions. More importantly, the
catalyst could be readily separated from the reaction mixture by a permanent magnet, and recycled up
to 6 times without the significant loss of its catalytic activity.
Keywords:
Biomass
2
,5-Diformylfuran
Aerobic oxidation
-Hydroxymethylfurfural
5
© 2013 Elsevier B.V. All rights reserved.
Manganese oxide
Magnetic nanoparticles
1
. Introduction
At present, we are entering an epoch of diminishing availabil-
versatile platform chemical. It can be used as a precursor for the
synthesis of fine chemicals, pharmaceuticals, polymers and liquids
fuels [8]. In this context, various methods have been developed for
the synthesis of HMF from carbohydrates [9–13].
ity of petrochemical resources, which are mainly used to produce
chemical products and energy for our society. The limited oil
resources render the current petroleum based fuel and chemical
production schemes unsustainable. Thus, much effort has been
devoted to explore the renewable feedstocks to meet the increasing
demand for fuels and chemicals [1]. Biomass as the only carbon-
containing renewable resource has received considerable attention
as an alternative feedstock for the production of fuels and chem-
icals [2–4]. Carbohydrates are the major component of biomass
feedstock. The biopolymers such as cellulose and hemi-cellulose
are readily available from various biomass sources. From those,
monosaccharides are obtained by the depolymerization of the
biopolymers, which can be further converted into various chem-
icals and liquids fuels [5–7].
Recently, 2,5-diformylfuran (DFF), which is one of the oxidation
products of HMF, has attracted much attention. It can not only be
used to as a monomer for the production multifunctional mate-
rials [14,15], but also as the starting material for the synthesis of
drugs, antifungal agents, nematocides, and ligands [16,17]. How-
ever, it is challenging for the selective oxidation of HMF into DFF,
as various oxidation products can be produced such as 5-formyl-2-
furancarboxylic acid (FFCA), 5-hydroxymethyl-2-furancarboxylic
acid (HMFCA), and 2,5-furandicarboxylic acid (FDCA). Synthesis of
DFF has been early reported from the oxidation of HMF using tra-
ditional oxidants such as NaOCl [18], BaMnO [19], and pyridinium
4
chlorochromat [20]. These methods not only required stoichiomet-
ric oxidants, but also released serious toxicity wastes. Therefore, it
is strongly demanded to develop environmentally friendly systems
for the selective oxidation of HMF into DFF.
5
-Hydroxymethylfurfural (HMF), the acid-catalyzed dehydra-
tion product of C6 based carbohydrates, is considered to be a
Recently, there is a growing attention in the synthesis of DFF
from the oxidation of HMF with molecular oxygen catalyzed by
heterogeneous catalysts. The supported Ru based catalysts gener-
ally showed high HMF conversion and DFF selectivity [21–24], but
some required high pressure. The vanadium-based catalysts also
∗ Corresponding authors. Tel.: +86 27 67842752; fax: +86 27 67842752.
E-mail addresses: zehuizh@mail.ustc.edu.cn (Z. Zhang), dhm@dicp.ac.cn
H. Duan).
(
0
926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.12.014

B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
65
have been attracted much attention for the aerobic oxidation of
2.4. Catalyst characterization
HMF into DFF [25,26], but some vanadium-based catalyst easily lost
its catalytic activity. Furthermore, the catalyst recycling procedures
using filtration and centrifugation were tedious, which inevitably
leads to the loss of catalyst in the separation processes. Therefore,
the development of new catalytic systems for the oxidation of HMF
into DFF is strongly demanded. Recently, magnetic materials based
heterogeneous catalysts have received growing attention due to
their unique properties [27]. Magnetic separation is a convenient
method for catalyst recycling, which overcomes the disadvantages
caused by the traditional separation method.
Due to the high price of noble metals, innovating cheap transi-
tion metal-based heterogeneous catalysts for the synthesis of DFF
from the oxidation of HMF is thus highly desirable. In the past
decade, some methods on the aerobic oxidation of alcohols using
relative cheap Mn based catalysts were reported [28]. Inspired by
Transmission electron microscope (TEM) images were obtained
using an FEI Tecnai G -20 instrument. The sample powder were
firstly dispersed in ethanol and dropped onto copper grids for
observation. FT-IR measurements were recorded on a Nicolet
2
−
1
NEXUS-6700 FTIR spectrometer with a spectral resolution of 4 cm
−
1
in the wave number range of 500–4000 cm . X-ray powder
diffraction patterns of samples were determined with a Bruker
advanced D8 powder diffractometer (Cu K␣). The scan ranges
◦
◦
were 10–80 with 0.016 steps, respectively. X-ray photoelec-
tron spectroscopy (XPS) was conducted on a Thermo VG scientific
ESCA MultiLab-2000 spectrometer with a monochromatized Al
K␣ source (1486.6 eV) at constant analyzer pass energy of 25 eV.
The binding energy was estimated to be accurate within 0.2 eV.
All binding energies (BEs) were corrected referencing to the
C1s (284.6 eV) peak of the contamination carbon as an internal
standard.
the excellent catalytic activity of Mn O4 nanoparticles in the aer-
3
obic oxidation of alcohols and the unique property of magnetic
catalysts, herein, we developed a new method for the synthesis of
magnetic Fe O supported Mn O4 nanoparticles (Fe O /Mn O ),
3
4
3
3
4
3
4
2
.5. Aerobic oxidation of HMF under atmospheric pressure
and used as a novel heterogeneous catalyst for the aerobic oxi-
dation of HMF into DFF. To the best of our knowledge, this is the
first report on the aerobic oxidation of HMF into DFF catalyzed by
magnetic Fe O /Mn O nanoparticles.
The aerobic oxidation of HMF under atmospheric pressure was
carried out in a 25 mL round bottom flask, which was coupled
with a reflux condenser and capped with a balloon. Typically,
HMF (1 mmol, 126 mg) was firstly dissolved into DMF (7 mL) with
a magnetic stirrer. Then, the catalyst Fe3O4/Mn3O4 was added
into the reaction mixture and flushed with pure oxygen at a rate
3
4
3
4
2
. Experimental
−
1
◦
2.1. Materials and methods
of 20 mL min . The reaction was carried out at 110 C for the
desired reaction time. Time zero was taken when the oxygen
was flushed into the reaction mixture. After reaction, the cata-
lyst Fe3O4/Mn3O4 was separated from the reaction mixture by
a permanent magnet, and the products were analyzed by HPLC
method.
Fe O nanoparticles were prepared and characterized as
3
4
described in our previous work [29]. Mn(OAc) ·4H O and hex-
2
2
adecyl trimethyl ammonium bromide (CTAB) were purchased
from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China).
5
-Hydroxymethylfurfural (98%) was purchased from Aladdin
Chemicals Co., Ltd. (Beijing, China). 2,5-Diformylfuran (DFF) was
purchased from J&K Co., Ltd. (Beijing, China). Acetonitrile (HPLC
grade) was purchased from Tedia Co. (Fairfield, USA). All other
reagents were provided by local supplies (Wuhan, China). All the
solvents were purchased from Sinopharm Chemical Reagent Co.,
Ltd. (Shanghai, China) and freshly distilled before use.
2
.6. Aerobic oxidation of HMF under high pressure
Aerobic oxidation of HMF under high pressure was carried out
in semi-batch reactor. Briefly, HMF was firstly dissolved in DMF,
and then the catalyst was added into the reactor. The reactor was
subsequently sealed and oxygen was charged to the desired pres-
sure. Then the temperature was adjusted to the chosen value and
the reaction was allowed to proceed with a magnetic stirring at
2
.2. Preparation of Mn O4 nanoparticles
3
6
00 rpm. After a predetermined reaction time, the reaction was
stopped by cooling the reactor with tap water to room tempera-
ture. The gas in the reactor was slowly released and the reaction
liquid was analyzed by HPLC method.
Mn O nanoparticles were prepared according to the known
3
4
procedures [30]. Mn(OAc) ·4H O (0.50 g) was firstly dissolved in
2
2
dimethyl sulfoxide (DMSO) (30 mL) with sonication-assist. This
solution was then transferred into a 100 mL Teflon-lined stainless
◦
steel autoclave and heated at 120 C for 6.5 h. After the slow cooling
2.7. Determination of the products
to room temperature, water (30 mL) was added into the mixture,
and the black Mn O4 nanoparticles were precipitated after keep-
3
HMF and DFF were quantified by HPLC system using external
ing it at room temperature for 2 h. The black Mn O4 nanoparticles
3
standard calibration curve method. Samples were separated by a
reversed-phase C18 column (200 mm × 4.6 mm) at a wavelength
of 280 nm. Acetonitrile and 0.1 wt.% acetic acid aqueous solution,
volume ratio 15:85, were used as the mobile phase, and the sam-
were thoroughly washed with deionized water and ethanol sev-
◦
eral times, separated by centrifugation, then dried at 60 C for 6 h
in vacuum.
−
1
◦
ples were eluted at a rate of 1.0 mL min at 25 C. The content of
HMF and DFF in samples were obtained directly by interpolation
from calibration curves, which were constructed based on the pure
compounds.
2
.3. Preparation of the catalyst Fe O /Mn O
3 4 3 4
Mn(OAc) ·4H O (0.50 g) and CTAB (30 mg) were firstly dissolved
2
2
To calculate the yield of DFF from HMF equation (1) was used
in DMSO (50 mL) in a 100 mL Teflon-lined stainless steel autoclave.
Then Fe O (0.10 g) was dispersed into the above solution with
3
4
Moles of HMF added−Moles of unreacted HMF
HMF conversion =
sonication-assist for 30 min. Then the autoclave was sealed, and
Moles of HMF added
◦
it was heated at 120 C for 6.5 h. Other steps were the same as
×
100%
(1)
described above for the synthesis of Mn O .
3
4

6
6
B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
Fe O /Mn O
3
4
3
4
Fe O
•
3
4
Mn O
3
4
•
♦
♦
•
•
♦
•
♦
♦
•
Fe O @Mn O
4
3
4
3
Mn O4
Mn–O
Mn–O
Fe–O
3
2
0
30
40
50
60
70
1200
1000
800
600
400
2θ
-
1
Wavenumber(cm )
Fig. 1. XRD patterns of the samples. The circle represents Fe3O4, and the rhombus
represents Mn3O4.
Fig. 2. FT-IR spectra of the samples.
◦
One molecule of HMF gave rise to one molecule DFF. DFF yield
can be calculated using Eq. (2)
[32]. Particularly, a strong peak at 2ꢀ = 35.5 was observed in the
XRD pattern of Fe3O4 sample. The XRD pattern of Fe3O4/Mn3O4
was also collected, and a strong peak at 2ꢀ = 35.7 was observed.
◦
Moles of DFF
Compared the XRD patterns of Fe O /Mn O with those of Mn O4
DFF yield =
× 100%
(2)
3
4
3
4
3
Moles of HMF added
.8. Catalyst recycling
After reaction, the catalyst Fe O /Mn O was separated from
◦
and Fe O , the strong peak at 2ꢀ = 35.7 in Fe O /Mn O is differ-
ent from those of Fe O4 at 2ꢀ = 35.5 and Mn O at 2ꢀ = 36.1 . The
3
4
3
4
3
4
◦
◦
3
3
4
2
reason might be that it was the overlap of the peak of Fe O at
3
4
◦ ◦
ꢀ = 35.5 and the peak of Mn O4 at 2ꢀ = 36.1 . In addition, some
3
2
3
4
3
4
peaks with low intensity were also appeared in Fe O /Mn O , and
all the peaks could be assigned to the corresponding peaks either
in Fe O4 or Mn O . These XRD results confirmed that the catalyst
3
4
3
4
the reaction mixture by a permanent magnet, and washed three
◦
times with water and ethanol, dried at 60 C over night in a vacuum
3
3
4
oven. Then the recovered catalyst was used for the second cycle
under the same conditions, which were used for the first time. These
processes were repeated for five times to test the stability of the
catalyst.
Fe O /Mn O was composed of Fe O4 and Mn O .
3
4
3
4
3
3
4
−1
−1
As shown in Fig. 2, two bands at 621 cm and 513 cm were
assigned to the Mn O stretching modes in tetrahedral sites and
the distortion vibration of Mn O in an octahedral environment,
−
1
respectively [33]. A weak band at 410 cm was attributed to the
3
. Results and discussion
vibration of manganese species (Mn3+) in an octahedral site. Com-
pared the FT-IR spectra of Fe O /Mn O with that obtained from
3
4
3
4
−
Mn O , a new peak at 583 cm was appeared in Fe O /Mn O . The
3 4 3 4 3 4
1
3
.1. Synthesis and characterization of catalyst
The schematic illustration for the synthesis of Fe O /Mn O cat-
−
1
peak at 583 cm was the characteristic peak of Fe O , which was
3
4
the vibration of Fe O bond [34]. The FT-IR results also indicated
3
4
3
4
alyst is shown in Fig. S1. The magnetic Fe O4 nanoparticles used
that Mn O4 was successfully supported on the magnetic Fe O4
3
3
3
as the support were prepared via co-precipitation of an aqueous
mixture of ferric and ferrous salts under a strong alkaline solu-
tion [29]. According to the previous results [30], the sulfur atom
in DMSO had so-called lone pair electrons, which captured the
nanoparticles.
Fig. 3 shows the representative TEM images of Mn O4 and
3
Fe O /Mn O . As shown in Fig. 3(a), most of Mn O nanoparticles
3
4
3
4
3
4
showed spherical morphology and uniform size distribution with
the diameter in 5–6 nm. Compared the TEM image of Fe O /Mn O
+
−
in H O. Thus, a lot of OH were released from H O in the
2 2
H
3
4
3
4
presence of lone pair electrons in DMSO, which accelerated the
with that of Mn O , Fe O /Mn O was also observed with spherical
shape, but some slight aggregation was observed in Fe O /Mn O .
3 4 3 4
3 4 3 4 3 4
2−x
formation of coordination compounds [Mn(OH)x(DMSO)
]
6−x
2−x
(
Fig. S1). [Mn(OH)x(DMSO)6
]
species were not stable at high
As the density and the color of Fe O4 are close to Mn O , it
−x
3 3 4
temperature, being decomposed into tiny and large quantities of
Mn(OH)2 crystal nuclei. Then Mn(OH)2 nuclei was further oxi-
dized by O2 in ambient environment into hausmannite Mn O ,
was difficult to discriminate Fe O4 and Mn O4 in the sample of
3 3
Fe O /Mn O . Therefore, the catalyst Fe O /Mn O was further
3
4
3
4
3
4
3
4
characterized by the energy-dispersive analysis of X-ray (EDAX).
The results obtained from EDAX provided the information that the
catalyst Fe O /Mn O was composed of O, Fe, and Mn elements
3
4
which was the most stable phase of manganese oxide. The formed
Mn O nanoparticles were then deposited on the dispersed Fe O4
3
4
3
3
4
3
4
nanoparticles to give rise to the magnetic Fe O /Mn O nanopar-
(Fig. S2). The ratio of Mn wt.% to Fe wt.% from the EDAX elemental
analysis is 1.87.
3
4
3
4
ticles.
Fig.
1
shows the XRD patterns of Fe O , Mn O and
The surface of Mn O4 and Fe O /Mn O was further charac-
3
4
3
4
3
3
4
3
4
Fe O /Mn O . All the peaks of Mn O could be assigned to the
terized by XPS (Fig. 4 and Fig. S3). All the measurements were
carried out with reference to the C1s binding energy (284.6 eV) as
an internal standard. A Mn 2p3/2–2p1/2 doublet at 653.7 eV and
3
4
3
4
3
4
tetragonal cell structure of Mn O , which were consistent with
3
4
those of tetragonal phase (JCPDS Card No. 18-0803) [31]. An intense
◦
peak at 2ꢀ = 36.1 was observed in the XRD pattern of Mn O , which
642.9 eV is observed for Mn O4 in Fig. 4(a), and the splitting width
3
4
3
was corresponded to (2 1 1) plane of Mn O . Other peaks with low
(11.8 eV) is consistent well with the earlier reports [35]. Compared
the XPS spectra of Mn O4 with that of Fe O /Mn O (Fig. S3), two
3
4
intensity were also matched well with those reported in JCPDS Card
18-0803). All the diffraction peaks of Fe O4 should be assigned to
3
3
4
3
4
(
new peaks with the binding energy (BE) values at 711.1 eV and
724.6 eV were only existed in Fe O /Mn O . The peaks at 711.1 eV
3
the planes of inverse cubic spinel structured Fe O , which was con-
3
4
3
4
3
4
sistent with the standard Fe O4 sample (JCPDS File No. 19-0629)
and 724.6 eV were attributed to the binding energy of Fe 2p3/2 and
3

B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
67
Fig. 3. TEM image of the samples. (a) Mn3O4; (b) Fe3O4/Mn3O4.
Fe 2p1/2, respectively, Fig. 4(b), which was the characteristic of
provided mainly the surface elemental analysis. The result of XPS
was much higher than that obtained by EDX, indicating that most
Mn O particle supported on the Fe O4 surface.
2+
Fe in Fe O [36]. From the XPS results, it was clearly confirmed
3
4
that the catalyst Fe O /Mn O was composed of Fe O and Mn O .
3
4
3
4
3
4
3
4
3
4
3
The ratio of Mn wt.% to Fe wt.% determined by XPS analysis was
.82, which was much higher than that obtained by EDX analysis.
As EDAX provided the whole elemental analysis, and XPS analysis
2
3.2. Synthesis of DFF from HMF under various conditions
The stability of HMF was firstly studied. HMF was stable in DMF
without catalyst and the flush of oxygen, and it was recovered by
◦
9
5.7% in DMF for 4 h at 120 C (Table 1, Entry 1). The results indi-
cated that HMF itself was stable in DMF under neutral conditions,
which was consistent with the previous results [37]. Under acidic
conditions, the acid-catalyzed side reactions such as aldol addition
and condensation were present [38]. Under alkaline conditions,
some side reactions such as condensation and cannizarro reactions
occurred [39]. Thus the neutral reaction condition was an important
safeguard to achieve high DFF yield and selectivity.
Mn O nanoparticles showed high catalytic activity on the aer-
3
4
obic oxidation of HMF into DFF, which produced DFF in a high
yield of 83.2% with HMF conversion of 100% (Table 1, Entry 2).
To our pleasure, the magnetic Fe O /Mn O catalyst also showed
3
4
3
4
high catalytic activity, and DFF yield of 82.1% was obtained, which
was close to that obtained with Mn O4 as the catalyst (Table 1,
3
Entries 2 vs 3). As a blank experiment, the support Fe O4 was inef-
3
fective for the aerobic oxidation of HMF (Table 1, Entry 4). These
results indicated that Mn O4 nanoparticles were the active sites
3
of Fe O /Mn O in the aerobic oxidation of HMF. Unlike the pre-
3
4
3
4
pared Mn O4 nanoparticles, the commercial available manganese
3
oxides such as Mn O3 and MnO2 gave poor catalytic performance
2
(Table 1, Entries 5–7). Specifically, the catalytic activity of the com-
mercial Mn O4 was much lower than that of the prepared Mn O4
3
3
nanoparticles (Table 1, Entries 2 vs 7). As the BET surface area of our
prepared Mn O4 nanoparticles and the commercial Mn O were
3
3
4
2
determined to be 86.7 and 7 m /g, respectively, the higher catalytic
activity of the prepared Mn O4 was mainly due to the high surface
3
area, which could provide a large number of active sites. In addition,
poor results were also obtained when homogeneous manganese
salts such as MnCl ·4H O and Mn(OAc) ·4H O were used as the
2
2
2
2
catalysts (Table 1, Entries 8 and 9).
3.3. Effect of the solvents on the aerobic oxidation of HMF
The aerobic oxidation of HMF was further carried out in a variety
of solvents. As shown in Table 2, the solvent showed a remark-
able effect on the oxidation of HMF. DMSO and DMF with strong
polarity and high boiling point were found to be the best solvents.
HMF conversions reached 83.5% and 84.5% with DMSO and DMF,
respectively (Table 2, Entries 1 and 2). However, the selectivity of
Fig. 4. XPS spectra of Fe3O4/Mn3O4. (a) Mn 2p region; (b) Fe 2p.

6
8
B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
Table 1
Catalytic conversion of HMF into DFF under various conditions.
a
Entry
Catalyst
Catalyst amount (mg)
Time (h)
HMF conversion (%)
DFF yield (%)
b
1
2
3
4
5
6
7
8
9
–
–
108
160
108
112
123
108
346
280
4
4
4
12
12
12
12
12
12
4.3
100
99.8
5.9
7.7
6.5
23.4
11.5
7.1
0
c
Mn3O4
Fe3O4/Mn3O4
Fe3O4
Mn2O3
MnO2
83.2
82.1
0.4
2.3
0.6
17.8
9.0
3.5
d
Mn3O4
Mn(OAc)2.4H2O
MnCl2·4H2O
a
◦
Reaction conditions: HMF (126 mg, 1 mmol) and a set amount of catalyst were added into DMF (7 mL), and the reaction was carried out at 120 C with the flush of molecular
−1
oxygen at a rate of 20 mL min
.
b
The reaction was carried out without catalyst and molecular oxygen.
The catalyst was the prepared Mn3O4 nanoparticles.
The catalyst was the commercial Mn3O4.
c
d
DFF in DMF was much higher than that in DMSO, as much more
DFF was further oxidized into FDCA in DMSO. Moderate HMF con-
versions around 40% were obtained in tetrahydrofuran (THF) and
toluene (Table 2, Entries 3 and 4), but the selectivity of DFF in
THF was the lowest of all testing solvents. Reactions carried in
other solvents MIBK (methyl isobutyl ketone), MeCN and CHCl3
produced low HMF conversion (Table 2, Entries 5–7). These results
indicated that DMF was a superior solvent for the aerobic oxida-
tion of HMF into DFF. According to the results reported by Biswal
et al., we speculated that Mn3 was the active site for oxidation
of HMF. According to the general acceptable mechanism of oxida-
of Fe O /Mn O was used (Table 3, Entries 3 and 4). Further increas-
3 4 3 4
ing the amount of Fe O /Mn O to 160 mg, the aerobic oxidation
3
4
3
4
of HMF proceeded fast (Table 3, Entries 5 and 6). It only required
4 h to obtain the maximum DFF yield of 82.1% with HMF conver-
sion of 100% (Table 3, Entry 6). The higher the catalyst amount
was, the higher the HMF conversion and DFF yield were at the
same reaction time. It should be attributed to an increase in the
availability and number of catalytically active sites, and a similar
phenomenon was also observed for the oxidation of benzyl alco-
hol [41]. However, it was noted that the selectivity of DFF was
almost the same around 80–85% with various catalyst loadings.
These results indicated that the selectivity of DFF had no relation-
ship with the catalyst loading when the reactions were carried
out under otherwise the same reaction conditions such as reaction
temperature, and solvent. Compared our method with the aero-
bic oxidation of HMF with K-OMS-2 (KMn O ·nH O) reported by
+
3+
tion of HMF and alcohols [40], the active center Mn was reduced
to Mn2+ by the abstraction of one hydrogen atom connected with
carbon atom in CH OH group, and simultaneously produced the
2
+
carbon cation, which then released H to give rise to DFF. Then the
oxygen was reoxidized Mn2 to Mn , which generated the catalytic
+
3+
8
16
2
cycle.
Yang et al. [42], our method produced a lower HMF yield than that
catalyzed by K-OMS-2. The main reason was mainly caused by the
catalyst itself. Our catalyst showed lower selectivity as 11.5% FDCA
was also formed. However, the great advantage of our newly devel-
oped method was that the catalyst could be readily removed from
the reaction mixture by an external magnet.
Based on the general knowledge of solvent effect, the strong
polar solvents such as DMSO and DMF are beneficial to stabilize
the positive charge of the cations. Thus, high DFF yield was obtained
when the aerobic oxidation of HMF catalyzed by Fe O /Mn O was
3
4
3
4
carried out in DMF.
3.4. Effect of catalyst amount the aerobic oxidation of HMF
3.5. Oxidation of HMF using various oxidants
The aerobic oxidation of HMF was also carried out with vari-
ous amounts of Fe O /Mn O in order to study the effect of the
The influence of oxygen pressure on the aerobic oxidation of
3
4
3
4
HMF was also studied. Reactions were carried out in autoclave reac-
tor with the oxygen pressure at 5 bar, 10 bar, and 15 bar with 40 mg
of the catalyst, respectively. As shown in Table 4, higher pressure
promoted higher HMF conversion and DFF yield (Table 4, Entries
1–4). The reason was that the oxygen concentration in DMF became
higher with the increase of oxygen pressure, thus HMF conversion
became higher with the increase of oxygen pressure at the same
reaction time period. However, the effect of oxygen pressure on
the catalytic activity of Fe O /Mn O was not very remarkable with
catalyst loading. As shown in Table 3, the loading of Fe O /Mn O
3
4
3
4
showed a remarkable effect on the aerobic oxidation of HMF. HMF
conversions were 27.1% and 43.9% after 12 h with 40 mg and 80 mg
of Fe O /Mn O , respectively, and the corresponding DFF yields
3
4
3
4
were 22.2% and 37.5% (Table 3, Entries 1 and 2). HMF conversions
reached 84.5% and 100% for 6 h and 10 h, respectively, when 120 mg
3
4
3
4
Table 2
40 mg of Fe3O4/Mn3O4. Sádaba et al. [26] also found that HMF con-
version in DMSO increased with the increase of oxygen pressure,
but it was also not distinct, in which HMF conversions were 84%
and 91% with 2.5 bar and 10 bar oxygen under otherwise the same
conditions.
a
Effect of the solvents on the aerobic oxidation of HMF into DFF.
Entry
Solvent
HMF
conversion
DFF yield
(%)
DFF
FDCA
yield
(%)
selectivity
(%)
(%)
1
2
3
4
5
6
7
DMSO
DMF
THF
Toluene
MIBK
MeCN
CHCl3
83.5
84.5
39.8
43.2
17.8
9.1
50.7
71.8
17.6
30.1
11.4
5.8
60.7
84.9
44.2
69.7
64.0
63.7
63.4
27.8
7.9
15.0
9.1
4.0
1.9
In addition, the effect of the oxygen pressure on the aerobic
oxidation of HMF was also studied with the use of 160 mg of
Fe O /Mn O . A similar trend was also observed for the use of
3
4
3
4
1
60 mg of Fe O /Mn O (Table 3, Entry 5 and Table 4, Entries 6 and
3 4 3 4
7
). Other oxidants such as H O and t-BuOOH were also used for the
2
2
19.7
12.5
4.1
oxidation of HMF. Among these oxidants, the clean and cheap oxi-
dant O2 was found to be the best oxidant (Table 4, Entry 5). Although
t-BuOOH has been used as a common oxidant in the oxidation of
organic compounds, it was ineffective for the oxidation of HMF.
a
Reaction conditions: HMF (126 mg, 1 mmol) and Fe3O4/Mn3O4 (120 mg) were
◦
added into the solvent (7 mL), and the reaction was carried out at 120 C for 6 h with
the flush of molecular oxygen at a speed of 20 mL min
−1
.

B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
69
Table 3
Effect of catalyst loading on the aerobic oxidation of HMF into DFF.
a
Entry
Catalyst amount (mg)
Time (h)
HMF conversion (%)
DFF yield (%)
DFF selectivity (%)
FDCA yield (%)
1
2
3
4
5
6
40
80
120
120
160
160
12
12
6
10
2
27.1
43.9
84.5
100
81.3
100
23.2
37.5
71.8
80.9
69.7
82.1
85.6
85.4
84.9
80.9
85.7
82.1
3.2
4.9
7.9
12.8
9.4
11.5
4
a
◦
Reaction conditions: HMF (126 mg, 1 mmol) and a set amount of Fe3O4/Mn3O4 were added into DMF (7 mL), and the reaction was carried out at 120 C with the flush of
−1
molecular oxygen at a speed of 20 mL min
.
HMF conversion reached 84.2% using t-BuOOH as oxidant, but the
selectivities of DFF and FDCA were both low (Table 4, Entry 8). Other
oxidation furan compounds such as HMFCA also did not formed.
The reason was that the furan ring was destroyed by the strong
oxidative ability of t-BuOOH. Both the HMF conversion and DFF
by Ru/␥-alumina gradually increased with the increase of reaction
temperature in the range from 80 C to 130 C.
◦
◦
3.7. Catalyst recycling experiments
yield were very low when H O2 was used as the oxidant (Table 4,
The catalyst recycling experiments were investigated with goal
of green chemistry, and the aerobic oxidation of HMF into DFF was
used as a model reaction. Reactions were carried out in DMF at
2
Entry 9). It was reported that the manganese oxide such as MnO2,
Mn5O8 and Mn O could catalyze the decomposition of H O2 [43].
3
4
2
◦
120 C. After reaction, the Fe O /Mn O catalyst was dispersed in
3 4 3 4
It was observed a severe reaction between the catalyst and H O at
2
2
the start of the addition of H O2 that was caused by H O2 decom-
the reaction mixture, and the reaction mixture was muddy (Fig.
S4(a)). But the catalyst Fe O /Mn O could be fast separated from
2
2
position and, consequently, the oxidation of HMF did not occur
sufficiently.
3
4
3
4
the reaction mixture by a permanent magnet (Fig. S4(b)). Then the
recovered catalyst was washed three times with water and ethanol,
◦
3.6. Time course of the aerobic oxidation of HMF into DFF at
dried at 80 C over night in a vacuum oven. Then the recovered
different reaction temperatures
catalyst was used for the second cycle under the same conditions.
These processes were repeated for five times to test the stability
of the catalyst. As shown in Fig. 6, DFF yields remained stable dur-
ing the recycling experiments (82.1% in first cycle vs 78.9% in sixth
cycle). The spent catalyst after the sixth cycle was characterized by
XRD. It was found that there was no difference in the XRD patterns
between the fresh catalyst and the spent catalyst (Fig. S5). Thus the
most possible reason for the slight decrease of the catalytic activ-
ity was due to the inevitable loss of the catalyst mass during the
recycling procedures.
◦
Finally, the aerobic oxidation of HMF was carried out at 80 C,
◦ ◦
1
00 C, and 120 C, respectively, with the aim to study the effect
of reaction temperature. As shown in Fig. 5, HMF conversion and
DFF yield increased with the increase of reaction temperature at
the same reaction time points. Taking the reaction time of 2 h as an
◦
example, HMF conversions reached 19.6%, 30.6% and 81.3% at 80 C,
◦ ◦
1
00 C and 120 C, respectively, and the corresponding DFF yields
were 16.7%, 28.4%, and 69.7%, respectively. It only required 4 h to
obtain the highest DFF yield of 82.1% with HMF conversion of 100%
◦
at 120 C. Interestingly, DFF was stable under the reaction condi-
3.8. Comparison our method with other reported methods for the
oxidation of HMF
tions, as DFF yield slightly decreased from 82.1% at 4 h to 77.0% at
◦
1
2 h. HMF conversion and DFF yield gradually increased at 80 C
◦
and 100 C during the reaction process. HMF conversions reached
In order to give a reasonable evaluation of our developed
method for the synthesis of DFF from HMF, the developed method
was compared with other recently reported methods, and the
results are shown in Table 5. The noble metal Ru based catalysts
generally showed high HMF conversion and DFF selectivity (Table 5,
Entries 1–3), but some cases required high pressure (Table 5, Entries
1 and 2), which was potentially dangerous in practical application.
◦ ◦
4
9.9% and 87.5% after 12 h at 80 C and 100 C, respectively, and
the corresponding DFF yields were 40.6% and 80.2%, respectively.
These results indicated that high reaction temperature promoted
the aerobic oxidation of HMF into DFF. Our results were consis-
tent with the previous results reported by Antonyraj et al. [23], in
which they found that the aerobic conversion of HMF catalyzed
Table 4
a
Catalytic oxidation of HMF using various oxidants.
Entry
Oxidant
Catalyst amount (mg)
Time (h)
HMF conversion (%)
DFF yield (%)
FDCA yield (%)
b
1
2
3
4
5
6
7
8
O2
O2
O2
O2
O2
O2
O2
t-BuOOH
H2O2
40
40
40
40
160
160
160
160
160
12
12
12
12
4
2
2
4
4
27.1
35.5
53.1
60.8
100
95.4
100
84.2
13.2
22.2
29.8
41.9
45.1
82.1
80.3
79.4
7.2
3.2
2.6
6.3
c
d
e
7.8
b
11.5
10.1
12.8
5.6
c
d
f
g
9
5.3
3.6
a
◦
Reaction conditions: HMF (126 mg, 1 mmol) and a set amount of Fe3O4/Mn3O4 were added into DMF (7 mL), and the reaction was carried out at 120 C using various
oxidant.
b
c
d
e
f
−1
.
1
5
1
1
7
7
bar with the flush of oxygen at 20 mL min
bar oxygen.
0 bar oxygen.
5 bar oxygen.
mmol of t-BuOOH was used.
mmol of H2O2 was used.
g

7
0
B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
Table 5
Catalytic oxidation of HMF into DFF using various methods.
Entry
Catalyst
Reaction conditions
HMF conversion (%)
DFF yield (%)
References
1
2
3
4
5
6
7
Ru/C
383 K, 2.0 MPa O2
403 K, 40 pis O2
393 K, 1 bar O2 at 20 ml/min
413 K, 40 bar Ar
398 K, 40 bar O2
100
100
94.8
100
92
63
96.0
97.0
92.0
98.0
81
[21,22]
[23]
[24]
[25]
[26]
Ru/␥-alumina
Ru/Hydrotalcite
VO /Cu2+-immobilized on sulfonated carbon catalyst
2+
V2O5/H-beta
Porphyrin-based polymer-supported iron(III)
Fe3O4/Mn3O4
383 K, 5 bar O2
393 K, 1 bar O2 at 20 mL min
31
82.1
[44]
This work
−
1
100
In addition, the high cost of Ru made those catalysts less attrac-
tive. Recently, the non-noble catalysts have been attracted much
attention. The vanadium based catalysts seemed to be promising
in the aerobic oxidation of HMF. However, these catalytic systems
also had some distinct drawbacks such as the requirement of high
pressure, the need of co-catalyst (Table 5, Entry 4) and the loss of
catalytic activity (Table 5, Entry 5). Recently, the porphyrin-based
polymer-supported iron(III) was also used for the oxidation of HMF,
but low HMF conversion and low DFF selectivity were obtained
80
60
40
20
0
(
Table 5, Entry 6). Albeit our developed Fe O /Mn O catalytic sys-
3 4 3 4
tem showed a little lower DFF selectivity as compared with some of
the mentioned methods, it demonstrated some unique advantages
(
Table 5, Entry 7). Firstly, the recovery of Fe O /Mn O was conve-
3 4 3 4
nient with the assist of an external magnet. Secondly, the cost was
1
2
3
4
5
6
Recycle No.
Fig. 6. Recycle experiments of the catalyst Fe3O4/Mn3O4. Reaction conditions: HMF
(
126 mg, 1 mmol) and Fe3O4/Mn3O4 (160 mg) were added into 7 mL of DMF, then
◦
the reaction was carried out at 120 C with the flush of dioxygen at a speed of
2
−
1
0 mL min
.
much lower in comparison with the tradition noble metal catalysts,
which undoubtedly saved the cost in the practical application.
4. Conclusion
In conclusion, a new method was developed for the synthesis
of DFF from the aerobic oxidation of HMF into DFF. The magnetic
Fe O /Mn O catalyst showed an excellent catalytic performance
3
4
3
4
with DFF yield of 82.1%. The catalyst could be readily separated
from the reaction mixture by a permanent magnet, and could be
reused for several times without the loss of its catalytic activity.
The high activity, stability, magnetic recyclability, combining the
facile preparation of the catalyst make them promising applications
in conversion of biomass derived platform chemicals into valuable
bulk chemicals.
Acknowledgements
This project was supported by National Natural Science Foun-
dation of China (Nos. 21203252, 21206200).
Appendix A. Supplementary data
Supplementary data associated with this article can be
found, in the online version, at http://dx.doi.org/10.1016/j.apcata.
2
013.12.014.
References
[
1] J.D. Murphy, J. Browne, E. Allen, C. Gallagher, Renew. Energy 55 (2013) 474.
Fig. 5. Time course of the aerobic oxidation of HMF into DFF at different reaction
temperatures. (a) HMF conversion; (b) DFF yield. Reaction conditions: HMF (126 mg,
[2] H. Tolvanen, L. Kokko, R. Raiko, Fuel 111 (2013) 148–156.
[3] E.S. Kim, S. Liu, M.M. Abu-Omar, N.S. Mosier, Energy Fuels 26 (2012) 1298–1304.
[
4] H. Kayser, C.R. Müller, C.A. García-González, I. Smirnova, W. Leitner, P.D. de
María, Appl. Catal. A: Gen. 445–446 (2012) 180–186.
1
mmol) and Fe3O4/Mn3O4 (160 mg) were added into 7 mL of DMF, and the reac-
tions were carried out at a set temperature with the flush of oxygen at a speed of
−1
[5] Z.H. Zhang, Z.K. Zhao, Carbohydr. Res. 344 (2009) 2069–2072.
2
0 mL min
.

B. Liu et al. / Applied Catalysis A: General 472 (2014) 64–71
71
[
[
[
[
6] S. Al-Zuhair, M. Al-Hosany, Y. Zooba, A. Al-Hammadi, S. Al-Kaabi, Renew. Energy
6 (2013) 85–89.
7] T. Stauner, I.B. Silva, O.A. El Seoud, E. Frollini, D.F.S. Petri, Cellulose 20 (2013)
109–1119.
8] R.J. van Putten, J.C. van der Waal, E. de Jong, C.B. Rasrendra, H.J. Heeres, J.G. de
Vries, Chem. Rev. 113 (2013) 1499–1597.
[26] I. Sádaba, Y.Y. Gorbanev, S. Kegnæs, S.S.R. Putluru, R.W. Berg, A. Riisager, Chem-
CatChem 5 (2013) 284–293.
[27] M.J. Jin, D.H. Lee, Angew. Chem. Int. Ed. 49 (2010) 1119–1122.
[28] H.Y. Sun, Q. Hua, F.F. Guo, Z.Y. Wang, W.X. Huang, Adv. Synth. Catal. 354 (2012)
569–573.
[29] S.G. Wang, Z.H. Zhang, B. Liu, J.L. Li, Catal. Sci. Technol. 3 (2013) 2104–2112.
[30] W.G. Fan, L. Gao, J. Inorg. Mater. 21 (2006) 789–794.
[31] R. Song, S.H. Feng, H.J. Wang, C.M. Hou, J. Solid State Chem. 202 (2013) 57–60.
[32] J. Lee, Y. Lee, J.K. Youn, H.B. Na, T. Yu, H. Kim, S.M. Lee, Y.M. Koo, J.H. Kwak, H.G.
Park, H.N.M. Hwang, J.G. Park, J. Kim, T. Hyeon, Small 4 (2008) 143–152.
[33] H. Zhang, C. Liang, Z. Tian, G. Wang, W. Cai, J. Phys. Chem. C 114 (2010)
12524–12528.
[34] C. He, T. Sasaki, Y. Shimizu, N. Koshizaki, Appl. Surf. Sci. 254 (2008) 2196–2202.
[35] A.M.E. Raj, S.G. Victoria, V.B. Jothy, C. Ravidhas, J. Wollschlager, M. Suendorf,
M. Neumann, M. Jayachandran, C. Sanjeeviraja, Appl. Surf. Sci. 256 (2010)
2920–2926.
5
1
9] Z.H. Zhang, Z.K. Zhao, Bioresour. Technol. 101 (2010) 1111–1114.
[
[
10] S. De, S. Dutta, A.K. Patra, B.S. Rana, A.K. Sinha, B. Saha, A. Bhaumik, Appl. Catal.
A: Gen. 435–436 (2012) 197–203.
11] Z.H. Zhang, Q. Wang, H.B. Xie, W.J. Liu, Z.K. Zhao, ChemSusChem 4 (2011)
1
31–138.
[
[
[
[
[
12] N. Nikbin, S. Caratzoulas, D.G. Vlachos, ChemCatChem 4 (2012) 504–511.
13] B. Liu, Z.H. Zhang, Z.K. Zhao, Chem. Eng. J. 215–216 (2013) 517–521.
14] X. Tong, Y. Ma, Y. Li, Appl. Catal. A: Gen. 385 (2010) 1–13.
15] J. Ma, Z. Du, J. Xu, Q. Chu, Y. Pang, ChemSusChem 4 (2011) 51–54.
16] K.T. Hopkins, W.D. Wilson, B.C. Bendan, D.R. McCurdy, J.E. Hall, R.R. Tidwell, A.
Kumar, M. Bajic, D.W. Boykin, J. Med. Chem. 41 (1998) 3872–3878.
17] A. Gandini, Green Chem. 13 (2011) 1061–1083.
[36] G. Kataby, A. Ulman, M. Cojocarua, A. Gedanken, J. Mater. Chem. 9 (1999)
1501–1506.
[
[
[
[
18] A.S. Amarasekara, D. Green, E. McMillan, Catal. Commun. 9 (2008) 286–288.
19] F.W. Lichtenthaler, Acc. Chem. Res. 35 (2002) 728–737.
20] L. Cottier, G. Descotes, J. Lewkowski, R. Skowro n´ ski, E. Viollet, J. Heterocycl.
Chem. 32 (1995) 927–930.
[37] K.R. Vuyyuru, P. Strasser, Catal. Today 195 (2012) 144–154.
[38] S.K.R. Patil, C.R.F. Lund, Energy Fuels 25 (2011) 4745–4755.
[39] E. Kang, D.W. Chae, B. Kim, Y.G. Kim, J. Ind. Eng. Chem. Res. 18 (2012) 174–177.
[40] M. Biswal, V.V. Dhas, V.R. Mate, A. Banerjee, P. Pachfule, K.L. Agrawal, S.B. Ogale,
C.V. Rode, J. Phys. Chem. C 115 (2011) 15440–15448.
[41] V.V. Namboodiri, V. Polshettiwar, R.S. Varma, Tetrahedron Lett. 48 (2007)
8839–8842.
[42] Z.Z. Yang, J. Deng, T. Pan, Q.X. Guo, Y. Fu, Green Chem. 14 (2012) 2986–2989.
[43] M.D.D. Vigil, F.P. Delgado, V.C. Martinez, A.L. Ortiz, Int. J. Chem. Reactor Eng. 13
(2008) 1122–1128.
[
[
[
21] J.F. Nie, J.H. Xie, H.C. Liu, J. Catal. 301 (2013) 83–91.
22] J.F. Nie, J.H. Xie, H.C. Liu, Chin. J. Catal. 34 (2013) 871–875.
23] C.A. Antonyraj, J. Jeong, B. Kim, S. Shin, S. Kim, K-Y. Lee, J.K. Cho, J. Ind. Eng.
Chem. 19 (2013) 1056–1059.
[
[
24] A. Takagaki, M. Takahashi, S. Nishimura, K. Ebitani, ACS Catal.
562–1565.
25] N.-T. Le, P. Lakshmanan, K. Cho, Y. Han, H. Kim, Appl. Catal. A: Gen. 464–465
2013) 305–312.
1 (2011)
1
[44] B. Saha, G. Dinesh, M.M. Abu-Omar, A. Modak, A. Bhaumik, J. Catal. 299 (2013)
316–320.
(