Fabrication of spinel CoMn2O4 hollow spheres for highly selective aerobic oxidation of 5-hydroxymethylfurfural to 2,5-diformylfuran

Article abstract of DOI:10.1016/j.cattod.2018.04.069

Spinel CoMn₂O₄ hollow spheres were prepared by a solvothermal method, where uniform cobalt-manganese glycerate spheres as the precursor were self-assembled and converted into hollow spheres by calcination. The as-prepared catalysts were characterized by field emission scanning electron microscopy (FESEM), X-ray powder diffraction (XRD), X-ray photoelectron spectroscopy (XPS), nitrogen adsorption and desorption isotherms, thermogravimetric analysis, and H₂ temperature programmed reduction (H₂-TPR). The results show that body-centered tetragonal and face-centered cubic spinel structures of CoMn₂O₄ hollow spheres can be tuned by different Co/Mn molar ratios. Moreover, the spinel CoMn₂O₄ hollow spheres exhibit high catalytic activity for the selective oxidation of 5-hydroxymethylfurfural (HMF) into 2,5-diformyfuran (DFF). Importantly, the spinel CoMn₂O₄ hollow spheres with Co/Mn molar ratio of 2/3 shows the highest catalytic performance with a 41.6percent HMF conversion and a 100percent selectivity to DFF. On one hand, the spinel CoMn₂O₄ hollow spheres with large surface area have sufficient exposed active sites for the oxidation of HMF and oxidized product DFF can fast leave the catalyst surface leading to high selectivity, avoiding further oxidation. On the other hand, the synergistic effect between Co and Mn species in the binary oxides by the formation of heteronuclear cluster complexes CoMn₂O₄, which affects and promotes the electron-transfer process, results in significant improvement in catalyst performance. Additionally, the effects of different Co/Mn molar ratios, catalyst amounts, reaction time and temperatures of the catalyst for HMF oxidation were also investigated. Furthermore, the spinel CoMn₂O₄ hollow spheres can be used for six consecutive runs without significant loss of its catalytic activity.

Full text of DOI:10.1016/j.cattod.2018.04.069

Catalysis Today xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Fabrication of spinel CoMn O hollow spheres for highly selective aerobic
2
4
oxidation of 5-hydroxymethylfurfural to 2,5-diformylfuran
⁎
Lei Ding, Wenyu Yang, Lifang Chen , Hongye Cheng, Zhiwen Qi
State Key Laboratory of Chemical Engineering, School of Chemical Engineering, East China University of Science and Technology, Shanghai 200237, China
A R T I C L E I N F O
A B S T R A C T
2 4
Spinel CoMn O hollow spheres were prepared by a solvothermal method, where uniform cobalt-manganese
Keywords:
5
2
-hydroxymethylfurfural
,5-diformyfuranm
glycerate spheres as the precursor were self-assembled and converted into hollow spheres by calcination. The as-
prepared catalysts were characterized by field emission scanning electron microscopy (FESEM), X-ray powder
diffraction (XRD), X-ray photoelectron spectroscopy (XPS), nitrogen adsorption and desorption isotherms,
Selective oxidation
Spinel CoMn
Hollow spheres
2 4
O
thermogravimetric analysis, and H
centered tetragonal and face-centered cubic spinel structures of CoMn
ferent Co/Mn molar ratios. Moreover, the spinel CoMn hollow spheres exhibit high catalytic activity for the
selective oxidation of 5-hydroxymethylfurfural (HMF) into 2,5-diformyfuran (DFF). Importantly, the spinel
CoMn hollow spheres with Co/Mn molar ratio of 2/3 shows the highest catalytic performance with a 41.6%
HMF conversion and a 100% selectivity to DFF. On one hand, the spinel CoMn hollow spheres with large
2
temperature programmed reduction (H
2
-TPR). The results show that body-
2 4
O
hollow spheres can be tuned by dif-
2 4
O
2 4
O
2 4
O
surface area have sufficient exposed active sites for the oxidation of HMF and oxidized product DFF can fast
leave the catalyst surface leading to high selectivity, avoiding further oxidation. On the other hand, the sy-
nergistic effect between Co and Mn species in the binary oxides by the formation of heteronuclear cluster
complexes CoMn O , which affects and promotes the electron-transfer process, results in significant improve-
2 4
ment in catalyst performance. Additionally, the effects of different Co/Mn molar ratios, catalyst amounts, re-
action time and temperatures of the catalyst for HMF oxidation were also investigated. Furthermore, the spinel
CoMn
2
O
4
hollow spheres can be used for six consecutive runs without significant loss of its catalytic activity.
1. Introduction
ring, which open a promising avenue for chemical reactions. Thus, HMF
with function groups can be oxidized to produce high value-added
The main sources of functional carbon skeletons including fuels and
chemicals of industrial importance, such as 2,5-diformylfuran (DFF), 5-
formylfuroic acid (FFCA), 5-hydroxymethylfuroic acid (HMFCA), 2,5-
furandicarboxylic acid (FDCA), and other useful compounds e.g. levu-
linic acid (LA), formic acid, and fumaric acid (FA) through ring opening
[6]. Typically, DFF is one of the selective oxidative products of HMF
with important commercial value, and can be used for the synthesis of
various poly-schiff bases, pharmaceuticals, organic conductor, and
cross-linking agents of poly (vinyl alcohol) for battery separations [7].
In consideration of the complexity of the HMF reaction pathway shown
in Fig. 1, the development of high selectivity and efficient catalytic
procedure for the synthesis of DFF from HMF is highly desirable.
Compared with homogeneous catalysts, heterogeneous metal
oxides, supported noble metal and metal catalysts with many ad-
vantages of readily separation and reusability, are successfully used in
various organic transformations [8]. Noble metal catalysts with mod-
erate reaction conditions and high catalytic activity become a research
chemicals highly rely on fossil resources such as coal, petroleum, and
natural gas [1]. In view of the fast depletion of fossil resources and
increasing environmental concerns, the development of efficient and
greener processes for the production of chemicals and fuels from re-
newable biomass has attracted growing interests over the last decade.
Biomass platform compounds can be effectively converted into high
value chemicals and fuels, which is an ideal alternative to fossil re-
sources. As the only sustainable source of carbon or organic com-
pounds, biomass has been regarded as equivalent to petroleum for the
production of fuels and chemicals [2–4].
As one of the most valuable platform chemicals between biomass
and specialty chemicals, 5-hydroxymethylfurfural (HMF) can be ob-
tained from lignocellulosic biomass conversion in the dehydration of
fructose and glucose using acid catalysts (see Fig. 1) [5]. It has multiple
functional groups containing hydroxyl, carbonyl, formyl, and furan
⁎
Corresponding author.
E-mail address: lchen@ecust.edu.cn (L. Chen).
https://doi.org/10.1016/j.cattod.2018.04.069
Received 28 February 2018; Received in revised form 24 April 2018; Accepted 30 April 2018
0920-5861/ © 2018 Elsevier B.V. All rights reserved.
Please cite this article as: Ding, l., Catalysis Today (2018), https://doi.org/10.1016/j.cattod.2018.04.069

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 1. Reaction pathway for aqueous HMF oxidation.
hotspot. Au, Ag, Ru, Pt, and alloy catalysts have much effect on the
selective oxidation of HMF into DFF [9–13]. For instance, Au nano-
absorptions [23,24], ion battery electrodes [25,26], and oxygen re-
ductions or evolution reactions [27,28] due to their excellent perfor-
mance with large surface area, reduced mass and charge transport [29].
particles supported on MnO
2
afforded an 82% conversion of HMF and a
99% selectivity of DFF at 120 °C within 8 h reaction time under O
2
2 4
In this work, a series of spinel CoMn O hollow spheres were synthe-
atmosphere [10]. Yadav et al. reported Ag-OMS-2 (OMS: octahedral
molecular sieve) as a catalyst for selective oxidation of HMF with a high
DFF yield of 99% at 165 °C within 4 h reaction time under 15 atm of air
sized using a facile solvothermal approach combined with subsequent
calcination and examined for the selective oxidation of HMF into DFF.
In addition, the as-synthesized spinel CoMn O hollow spheres catalysts
2 4
[
11]. Ruthenium nanoparticles supported on N-containing mesopor-
were characterized and the influencing factors of the selective aerobic
ouspolymer was synthesized and excellent conversion of HMF (99.6%)
and 85% DFF yield were obtained under optimized reaction conditions
oxidation were explored.
[
12]. Our group also prepared bimetallic Au-Pd alloys supported on
zinc hydroxycarbonate with excellent catalytic activity towards HMF
oxidation in the presence of weak base NaHCO solution at benign
2
. Experimental
3
2.1. Preparation of spinel CoMn O hollow spheres
2 4
conditions. Although these noble metal-based catalysts generally give
relatively high yields of DFF, drawbacks for this kind of catalysts are
conspicuous, such as high cost and stability problems. To promote the
conversion of HMF into DFF to a practical application level, it is ne-
cessary to develop efficient non-noble metal catalysts with the merits of
low price, low toxicity and high stability.
Therefore, transition metal based heterogeneous catalysts for the
synthesis of DFF from HMF are widely-studied [14–19]. Yang et al.
synthesized a series of 3D flower-like micro/nano Ce-Mo composite
oxide catalysts, which could reach a 94% DFF yield in 10 h [14]. Het-
erogeneous mesoporous manganese/cobalt oxide catalysts were applied
to the selective oxidation of HMF into DFF with 80% conversion and
All chemicals used in the experiments were of analytical grade, and
used without further purification. Spinel CoMn hollow spheres were
prepared by a hydrothermal process as described in the followings. In a
typical procedure, 0.80 mmol of Co(NO •6H (0.232 g) and
.20 mmol of Mn(NO •4H O (0.301 g) were dissolved in 50.5 mL of
isopropanol (IPA) to form a clear solution. Then 7.5 mL of glycerol
Gly) and 2 mL of deionized (DI) water were added into the solution.
2 4
O
3
)
2
2
O
1
3
)
2
2
(
After stirring for another 10 min, the solution was put into a 100 mL
Teflon-lined stainless steel autoclave and kept in a muffle furnace at
190 °C for 12 h. Cooling down to room temperature naturally, the
precipitate was separated by centrifugation, washed several times with
ethanol and dried in an oven at 70 °C for 10 h. Finally, the precipitate
9
6% selectivity using air as the sole oxidant [15]. Meanwhile, the
catalytic performance of a series of zeolite-supported vanadia catalysts
was examined for the aerobic oxidation of HMF to DFF in organic
solvents such as N,N-dimethylformamide (DMF), methyl isobutyl ke-
tone, toluene, trifluorotoluene, and dimethyl sulfoxide (DMSO). Under
optimized conditions, a high selectivity towards DFF of 99% and an
was calcined at 350 °C for 3 h to obtain product named CoMn
series of spinel hollow spheres were prepared by changing Co-Mn
precursor molar ratios named CoMn -1:2, CoMn -1:1, CoMn
:1. Besides, Mn , Co , and mechanical mixed Mn -Co were
2 4
O -2:3. A
2
O
4
2
O
4
2 4
O -
2
3
O
4
3
O
4
3
O
4
3 4
O
also prepared for comparison using similar method above.
8
4% HMF conversion were obtained [16]. In addition, manganese-
copper layered double hydroxides prepared by co-precipitation method
were investigated for the selective oxidation of HMF into DFF. It has
been found that a composition corresponding to Mn0.70Cu0.05Al0.25
leads to the highest HMF conversion (90%) and DFF selectivity (87%)
2.2. Catalyst characterization
X-ray diffraction (XRD) patterns were recorded on a D8 Advance
diffractometer equipped with radiation source
[
17]. Another interesting finding is that the morphology of transition
a
CuKα
metal catalysts has a certain effect on catalytic activity [20], such as 3D
flower-like micro/nano and mesoporous mentioned above. It is of great
significance to improve the catalytic performance through controllable
synthesis of catalysts with special shapes.
As an important family of functional materials, hollow structures
are widely used in photocatalytic reactions [21,22], microwave
(λ = 1.54056 Å), operating at 40 kV and 40 mA. A continuous mode
was used for collecting data in the 2θ range from 10° to 80° at a
−
1
scanning speed of 10° min . The morphology was examined with a
Nova NanoSEM-450 field emission scanning electron microscopy
(FESEM). Prior to the beginning, the sample was sprayed on platinum-
coated stainless steel substrate. UV–vis absorption of coordination
compound solution was measured by a T6 New-century ultraviolet and
2

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
visible spectrophotometer with DI water as a baseline against back-
ground. Thermogravimetric and differential scanning calorimetry (TG-
DSC) analysis of the synthesized powders was carried out on a TA Q600
thermal gravimetric analyzer. The experiment was performed in at-
mosphere from room temperature to 800 °C with a heating rate of
−
1
1
0 °C min , using alumina as reference and 5–10 mg of sample.
The surface area of spinel CoMn hollow spheres was determined
2 4
O
by nitrogen adsorption and desorption measurements at liquid nitrogen
temperature using 3 Flex multi-purpose gas adsorption instrument.
Prior to adsorption, the sample was degassed in vacuum at 200 °C for
4
h. Specific surface area was calculated from the Brunauer-Emmett-
Teller (BET) method and main pore size was analyzed with Barrett-
Joyner-Halenda (BJH) method. The element contents of Co and Mn
were detected with an inductively coupled plasma atomic emission
spectrometer (ICP-AES, Agilent, 167 nm–785nm/725). X-ray photo-
electron spectroscopy (XPS) was performed on an ESCALAB 250Xi
(
Thermo Fisher) apparatus with Al Ka X-rays radiation to analyze the
surface oxidation state distribution of the elements in the sample. H
temperature programmed reduction (H -TPR) experiments were carried
out on an AutoChem II 2920 V4.05 apparatus. The samples were re-
duced using 10% H /Ar (v/v) mixed gas with a flow of 30 mL/min at a
heating rate of 10 °C min
2
2 4 2 4
Fig. 2. XRD patterns of CoMn O precursor and spinel CoMn O hollow
spheres.
2
2
purity of the product.
−
1
.
The final morphology of the spinel CoMn
characterized by FESEM. As shown in Fig. 3, the spinel CoMn
2
O
4
hollow spheres was
is
2 4
O
2
.3. The selective oxidation of HMF
composed of uniform microspheres with diameters of 3.5 ± 0.5 μm
and the morphology could be perfectly preserved after calcination
without severe cracking. The surface of the sphere in Fig. 3b is a bit
rough and cracking due to the loss of glycerol ligand during calcination,
while there is a cavity inside with relative small diameters of the par-
ticles (Fig. 3c), demonstrating the formation of hollow structure. Si-
As a general procedure, the selective oxidation of HMF was per-
formed in a stainless-steel autoclave equipped with an internal thermo-
controller under a magnetic stirring. Typically, 50 mg of catalyst, 63 mg
of HMF and 5 g of DMF as solvent were added into the reactor. The air
2 2
in the reactor was replaced with O by purged 3 times and kept the O
3 4
milar Ni/Co oxides and Fe O hollow spheres have been reported pre-
pressure at 0.75 Mpa at room temperature. When the mixture was he-
ated to the desired temperature, the pressure was controlled at 0.8 MPa
and maintained a specific reaction time. All the changes of reaction
parameters: temperature, amount of the catalysts or reaction time were
notified. Finally, the autoclave was cooled down to room temperature
by quenching with cold water.
The reaction supernatant solution was analyzed using LC-2010AHT
SHIMADZU liquid chromatograph equipped with a UV–vis detector and
2 4
Benson polymeric BP-OA column. The mobile phase consisted of H SO
with pH = 3, at a flow rate of 0.8 mL min , and the analysis was
carried out at 45 °C, using an injection volume of 1 μL. HMF conversion
and DFF selectivity were calculated based on external calibration
curves using five individual samples.
viously [30,31]. However, there is no spheres formed without addition
of glycerol in the preparation (see Fig. 3d). The material is composed of
granules with diameter of 1 μm accumulated by the irregular, amor-
phous, and rough particles. The results demonstrate the important and
indispensable role of glycerol for the formation of hollow sphere
structure during solvothermal synthesis.
In order to explore the possible fabrication process of spinel
CoMn O hollow spheres, UV–vis adsorption spectra of different pre-
2 4
cursor solutions are shown in Fig. 4. It is acceptable that d-d ligand field
transition happens for the d orbitals of transition metal ions under the
action of ligand, which can be used for the qualitative analysis of
complex structure by the migration of UV–vis absorption peaks. To
achieve a better understanding of offset, the solution was diluted with
different proportions of IPA. There is no absorption peak of IPA and
−1
3
. Results and discussion
.1. Formation of spinel CoMn
The preparation of spinel CoMn
2+
2+
2+
Gly, while an absorption peak of M
(M = Co
and Mn ) in IPA
appears at 526 nm, indicating the formation of M-IPA coordination
compound. After the addition of glycerol, the absorption peak leading
to a blue shift to the wavelength of 518 nm results from the formation of
new M-Gly coordination compound, where -IPA is replaced by -Gly.
3
2
O
4
hollow spheres
hollow spheres involves the
2 4
O
solution synthesis of uniform metal glycerate solid spheres and sub-
sequent thermal annealing in air to remove coordinated organic li-
gands. In the first step, uniform cobalt-manganese glycerate (Co/Mn-
Gly) spheres as the precursor (Fig. 2) was prepared by a facile sol-
vothermal method. There is a strong peak at around 10° ascribing to the
formation of metal alkoxide by alcoholysis and coordination of glycerol
with the metal ions. The crystalline of the precursor obtained from
solvothermal reaction is particularly similar to reported Ni/Co-glyce-
rate [30] and Fe-glycerate [31]. The metal ions form coordination
compounds with organic ligands, which crystallizes in Co/Mn-Gly
under the solvothermal conditions. The Co/Mn-Gly solid spheres can be
When IPA and H
shifts to 511 nm, which stems from a red shift of M
an absorption wavelength at 508 nm ascribes to hydration effect be-
2
O coexist in system, the absorption peak continually
2+
in water. While
2+
tween M
and water, which has been confirmed by water structure
2+
2+
around metal ions [32]. The red shift of hydrated M and blue-shifted
M-Gly is that H
2
O strongly interacts with IPA and M
via hydrogen
bond interaction and it finally forms -M-Gly-H
2
O- coordination com-
pounds.
Based on the discussion above, the preparation process of spinel
CoMn hollow spheres is as following (Fig. 5). Firstly, M-IPA co-
ordination complexes were generated due to complexation reaction
2 4
O
readily converted to spinel CoMn
2
O
4
hollow spheres by a simple non
2+
2+
between Co , Mn , and IPA, where -IPA was formed by isopropanol
dehydrogenation. Then, -IPA was replaced by -Gly because of M-IPA
equilibrium heat treatment via removal of coordinated organic ligands.
All diffraction peaks of the sample after calcination can be indexed to
poor dynamical stability [31]. Simultaneously, Co/Mn-Gly-H
ordinated complex was formed owing to the added H O also combined
through hydrogen bond. Under solvothermal
2
O co-
2 4
CoMn O
(JCPDS 01-1126), a typical spinel structure, where Co2+ and
3
+
2
Mn
separately occupy tetrahedral A and octahedral B sites. Besides,
2+
with Gly and M
no other impurity diffraction peaks are observed, implying the high
3

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
2 4
Fig. 3. FESEM images of spinel CoMn O hollow spheres (a–c) with and (d) without glycerol addition.
Fig. 4. UV–vis adsorption spectra of different precursor solutions.
Fig. 5. Schematic illustration of the formation process of spinel CoMn
hollow spheres.
2 4
O
2
environment, -Co/Mn-Gly-H O- connecting each other will gradually
aggregate into hierarchical spheres through dehydration and alcohol-
oxygen groups driven by the minimization of interfacial energy, leading
to formation of Co/Mn-Gly by the heterogeneous nucleation process.
Finally, the release of glycerol ligand was caused by calcination in air
precursor. There are two-stage mass loss from room temperature to
00 °C in air through thermogravimetric analysis (TGA) of CoMn
precursor as shown in Fig. 6. The weight loss at stage I below 200 °C is
.1%, which is attributed to the release of physically adsorbed water on
the surface, trapped in the pores and interlayer space. The stage Ⅱ
appeared from 200 °C to 300 °C with a 37.6% weight loss corresponding
to a sharp DSC exothermic peak can be ascribed to the decomposition of
glycerol ligand in the precursor, as confirmed by XRD recorded before
and after calcination. It demonstrates the decomposition of the
8
2 4
O
5
and Co/Mn-Gly spheres converted into high crystalline spinel CoMn
hollow spheres by a simple nonequilibrium treatment process.
2 4
O
2 4
3.2. Characterization of spinel CoMn O hollow spheres
The spinel CoMn
2
O
4
hollow spheres were successfully prepared by
CoMn
2 4
O precursor starts from around 200 °C and completes at around
solvothermal method combined with calcination of Co/Mn-Gly
300 °C. According to the TGA and DSC curves, it is reasonable that a
4

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
Fig. 8. XRD patterns of spinel CoMn
2
O
4
hollow spheres with different Co/Mn
molar ratios.
Fig. 6. TGA and DSC curves of as-prepared Co/Mn-glycerate precursor.
crystallizes in
< x < 1.6 and a face-centered cubic form between 1.6 < x < 3.0
34].
For spinel CoMn
occupy tetrahedral and octahedral sites of the spinel structure, which
ascribes to body-centered tetragonal spinel (space group I41/adm,
JCPDS No. 01-1126). As Co/Mn radio in precursor increases, the peak
a
body-centered tetragonal structure between
0
[
calcined temperature of 350 °C was chosen for thermal treatment of the
precursor to ensure its complete decomposition.
The spinel CoMn O hollow spheres were further characterized by
2 4
nitrogen adsorption/desorption at −196 °C (see Fig. 7). The isotherms
are of type IV isotherms with a distinct H2-type hysteresis loop in the P/
0
P region from 0.45 to 0.99, which does not close until the saturation
2 4
O
hollow spheres, Co2+ and Mn3+ are separately
2 4
positions of CoMn O obviously shifts to high 2θ values. The shift of 2θ
pressure is reached. The Brunauer-Emmett-Teller (BET) method was
used to calculate surface area and the corresponding Barrett-Joyner-
Halenda (BJH) formula was employed to evaluate the pore distribution
using the desorption branch of the isotherm curve. The BET surface area
values indicates that a part of Mn atom locations in lattice is occupied
by smaller Co atom (rMn3+ = 0.645 Å, rCo = 0.545 Å), which results
in decrease in the lattice parameter. The decreased lattice constants of
2+
2 4
tetrahedral crystallites are 5.7567, 5.7283, and 5.6953 Å for CoMn O
2
−1
is estimated to be 82.6 m g
pore size distribution ranging from 3 to 10 nm (inset of Fig. 7). Thus,
the porous spinel CoMn hollow spheres are beneficial for oxidation
and the BJH calculation gives a main
with Co/Mn radios of 1:2, 2:3, and 1:1, respectively, based on measured
XRD patterns. On the other hand, with increased Co/Mn ratio to 2, the
prepared material undergo characteristic structural changes, which are
a consequence of the Co and Mn cations exchanging their lattice sites
under the bottom-up solvlthermal process. Thus, it is identical to face-
2 4
O
reaction during the charging/discharging products and expect to im-
prove catalytic activity [33].
The spinel structure of CoMn
Mn molar ratios were also taken into consideration. The XRD patterns
of different spinel CoMn hollow spheres, sole Mn and Co are
shown in Fig. 8. The diffraction peaks of both Mn and Co pre-
pared are perfectly associated to JCPDS No. 24-0734 and JCPDS No. 43-
003, respectively. Sharp peaks observed are indicative of highly
crystallized grains. Mn containing both di- and tri- valent manga-
nese belongs to body-centered tetragonal spinel (space group I41/adm),
while Co similar to Mn is indexable to the face-centered cubic
spinel (space group Fd-3 m). Compared with Mn and Co , the
diffraction peaks of spinel CoMn with different Co/Mn molar ratios
can be assigned to different spinel structures. There is a phase transi-
tions for Co (0 ≤ x ≤ 3.0) depending on the x values. It
2 4
O hollow spheres with different Co/
3 4
centered cubic Co O phase, where the octahedral and tetrahedral sites
are occupied exclusively by Co and Mn cations. Therefore, the prepared
spinel Co-Mn-O phases correspond to a class of configuration rather
than a single crystallographic structure, which matches well with XRD
results and previous reports [34,35].
2
O
4
3
O
4
3 4
O
3
O
4
3 4
O
1
2 4
XPS was used to monitor the oxidation states of spinel CoMn O -2:3
3 4
O
and the results are shown in Fig. 9. All of the binding energies were
corrected for specimen charging by referencing them to the C 1 s peak
at 284.4 eV. The peaks of Co, Mn, O and C elements have been observed
in the scan survey spectrum (Fig. 9a). Due to the Co 2p3/2 binding
3
O
4
3 4
O
3
O
4
3 4
O
2
O
4
2+
3+
energy of Co
is close to that of Co , the Co 2p3/2 spectrum of the
2+
3+
sample is fitted into two peaks of Co , Co , and a shakeup satellite
above the main 2p3/2 line by a Gaussian fitting method (Fig. 9b). The
Mn3−xO
x 4
2
+
peak at 780.3 eV is assigned to Co
and the peak at 781.6 eV is at-
tributed to Co3+ [36]. In the Mn 2p3/2 spectrum (Fig. 9c), three peaks
2
+
3+
observed at 641.3, 642.8, and 645.1 eV are assigned to Mn , Mn
,
and Mn4+, respectively [37]. It indicates the CoMn
2 4
O -2:3 definitely
coexists mixed phases, which is in good agreement with the molar ratio
of Co and Mn during the prepared process. As shown in Fig. 9d, the O
1
s spectrum is fitted into two peaks at 530.1 and 531.3 eV, which are
attributed to metal-oxygen bonds and defect sites with low oxygen
coordination [38]. The low oxygen coordination with vacancies plays
an important role in the improving mobility of oxygen and favors
oxygen transfer on the surface of catalysts [38,39], which makes spinel
CoMn
2 4
O hollow spheres improving catalytic activity for oxidation re-
actions.
3.3. Catalytic activities of spinel CoMn
2 4
O hollow spheres
It is generally accepted that the conversion of HMF and selectivity of
DFF were extremely sensitive to the reaction temperature [40]. In order
to obtain high HMF conversion and DFF selectivity, the reaction
Fig. 7. Nitrogen adsorption/desorption isotherms and pore size distribution
plots (inset) of the spinel CoMn hollow spheres.
2 4
O
5

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
2 4
Fig. 9. (a) Wide survey XPS spectrum, high-resolution XPS spectra (b) Co 2p3/2, (c) Mn 2p3/2, and (d) O 1 s of as-synthesized CoMn O -2:3.
reactants. On the other hand, ring-open reaction occurs and produces
levulinic acid and insoluble humin [42,43]. Therefore, it can be de-
duced that 100 °C should be the optimized reaction temperature in this
process.
For the sake of comparison, the catalytic activities of different cat-
alysts are shown in Table 1. A moderate HMF conversion of 30.6%
along with ca. 100% selectivity to DFF is obtained for prepared Mn
hollow spheres, while commercial Mn with BET surface area of
only gives HMF conversion of 1.0% (Table 1, entries 1–2).
Previous reports on 3D NiO hollow spheres and CeO /Al hollow
3 4
O
3 4
O
2
−1
5
–7 m g
2
2 3
O
fibers demonstrate that hollow structures with more BET surface areas
can provide more active sites and diffusion pathways for absorbing
more reactants, resulting in enhanced bio and adsorption properties
[
3 4
44,45]. Herein, the hollow Mn O spheres with exposed more active
sites than commercial one significantly improves its catalytic perfor-
mance. Additionally, oxidized product DFF can fast leave the catalyst
surface leading to high selectivity, avoiding further oxidation.
Fig. 10. Influence of reaction temperature on HMF oxidation.
Single Co
3
O
4
with poor catalytic performance (Table 1, entry 7)
temperature of HMF oxidation was optimized, while other reaction
conditions maintained constant such as catalyst amount, reaction
pressure and time, etc. (Fig. 10). The higher the reaction temperature is,
the higher HMF conversion can be obtained. The HMF conversion is
only 29.6% at 90 °C, while greatly increases to 84.9% with reaction
temperature at 160 °C. According to molecular dynamics theory, more
molecules are in contact with catalyst active sites at higher reaction
temperatures due to the more intense molecular thermal motion [41].
Besides HMF conversion, the selectivity to DFF is also greatly affected
by the reaction temperature and decreases with the increased reaction
temperature. It is close to 100% at reaction temperature less than
Table 1
a
Different Co-Mn catalysts for HMF oxidation .
Entry
Catalysts
HMF conversion (%)
Selectivity(%)
DFF
FFCA
Others
1
2
3
4
5
6
7
8
Mn
3
O
4
4
2
2
2
2
-commercial
0.9
–
–
100
–
–
–
–
2.6
9.2
–
Mn O
3
30.6
34.2
41.6
33.7
36.0
6.2
> 99
> 99
> 99
> 99
92.1
69.9
> 99
–
–
–
–
5.2
20.9
–
CoMn
O
4
O
4
O
4
O
4
-1:2
-2:3
-1:1
-2:1
CoMn
CoMn
CoMn
100 °C. However, DFF selectivity decreases to 24.0% further raising the
Co
Co
3
O
O
4
temperature up to 160 °C. On the one hand, the instability of HMF at
high temperature results in by-products such as ether, ketone, ester, and
water generated by aldol condensation reactions between parts of
3
4
-Mn
3
O
4
-2:3
33.7
a
2
Reaction conditions: 0.5 mmol HMF, 5 g DMF, 50 mg catalyst, 0.8 MPa O ,
100 °C, 2 h.
6

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
demonstrates it is not active in this reaction, but it can efficiently
promote Mn
binary oxides are used, the catalytic activities improve obviously
Table 1, entries 3–6). Furthermore, the catalytic activities initially
3 4
O catalyst with enhanced catalytic activity. When Co/Mn
(
increase to the maximum HMF conversion (41.6%) and 100% se-
lectivity to DFF, and then decrease with the increasing of Co/Mn ratios.
The HMF conversion decreases to 36.0% and DFF selectivity is down to
9
2
2.1% with continually increased Co/Mn ratio to 2:1. As seen in entries
and 7, Mn hollow spheres is active, while Co can just promote
3
O
4
3 4
O
the reaction. Therefore, excessive Co replaces Mn lattice sites in the
spinel CoMn hollow spheres, which may reduce the amounts of
2 4
O
active Mn species and result in the decrease of catalytic performance.
The similar results have been reported by previous works for the se-
lective oxidation of HMF into DFF [46,47].
It is reported manganese oxides exhibit moderate catalytic perfor-
mance for the selective oxidation into DFF [47]. After the addition of Co
(
3 2 2
NO ) •6H O as Co source during the preparation, there is a synergistic
effect between Co and Mn species in binary oxides, which leads to
significant improvement in catalyst performance. When the Co/Mn
ratio is 1:2 forming spinel CoMn
2
O
4
hollow spheres, HMF conversion is
34.2% and DFF selectivity is 100%, indicating synergistic effect of Co
2 4 3 4
and Mn species of spinel CoMn O hollow spheres. When Co O and
Mn
Mn ratio of 2:3 (Co
Co -Mn -2:3 shows very similar catalytic performance to Mn
Table 1, entry 2), and significantly lower than that of CoMn -2:3.
These results indicate that the high catalytic performance of CoMn
:3 maybe due to the synergistic effect between Co and Mn species in
the spinel CoMn hollow spheres, rather than physical contact. In
brief, the selective oxidation of HMF to DFF catalyzed by Co-Mn-O is a
single-electron-transfer process. The CoMn system exhibits high
catalytic activity maybe due to the formation of heteronuclear cluster
3
O
4
were mixed through conventional mechanical mean with a Co/
Fig. 11. H
2
-TPR profiles of spinel CoMn
2
O
4
with different Co-Mn ratios.
3
O
4
-Mn -2:3, Table 1, entry 8), it is found that
3 4
O
3
O
4
3
O
4
3 4
O
(
2 4
O
2 4
O -
2
2 4
O
2 4
O
complexes CoMn
process [48,49].
2 4
O , which affects and promotes the electron-transfer
ICP-AES results make it clear that Co/Mn ratios are similar to those
added in the preparation. CoMn -2:3 remaining spinel phase de-
monstrate excessive Co species highly dispersed or incorporated into
CoMn lattice exhibiting the strongest synergistic effect, resulting in
the highest catalytic performance. However, further increase in Co
content to Co/Mn ratio of 1, CoMn -1:1 shows a decreased HMF
2 4
O
2 4
O
2 4
O
conversion (33.7%) and high selectivity (> 99%). The high con-
centrated Co species means low amounts of active Mn species, which
leads to decreased catalytic activity. When Co/Mn ratio continuingly
Fig. 12. Influence of catalyst amount for selective aerobic oxidation of HMF.
increases to 2, CoMn
decreased selectivity to DFF (92.1%). Here, CoMn
centered cubic Co phase, which should affect the original active
sites of Mn, results in decreased DFF selectivity. It is also supported by
the poor catalytic activity of Co with low selectivity (Table 1, entry
).
2
O
4
-2:1 reveals 36.0% HMF conversion as well as a
2 4
O
-2:1 with face-
3 4
O
3 4
O
7
To further substantiate the effects of Co/Mn molar ratios on the
catalytic activity, the H -TPR profiles of different spinel CoMn
hollow spheres, Mn , and Co are shown in Fig. 11. Two obvious
reduction peaks in the reduction profile of Mn are observed. The
low-temperature peak at 220 °C is attributed to the reduction of Mn
species dispersed on Mn and the high-temperature peak at 362 °C is
due to the typical reduction of Mn
has been confirmed by the Mn 2p3/2 spectrum. In the case of Co
reduction peak at 225 °C is ascribed to the reduction of Co
and the peak around 240–255 °C is due to the transformation of CoO to
Co° [51]. Obviously, the reduction temperatures of all pinel CoMn
hollow spheres are lower than Mn and Co , implying higher
oxygen mobility in the spinel CoMn hollow spheres. The reduction
peak around 200 °C is attributed to the reduction of Mn
Mn in spinel CoMn and the appearance of peak around 290 °C is
due to the reduction of Mn to MnO and Co to CoO [36].
In particular, the reduction peak of CoO to Co° for CoMn
2
2 4
O
3
O
4
3 4
O
3 4
O
4+
3 4
O
4
+
3
O
4
to MnO [50]. The Mn species
, the
to CoO,
3 4
O
3 4
O
2 4
O
2 4
Fig. 13. Time course study of HMF oxidation with spinel CoMn O hollow
spheres.
3
O
4
3 4
O
2 4
O
4
+
species to
obtained around 450 °C, which shifts to higher temperature attributing
to a synergistic effect between Co and Mn in the spinel CoMn hollow
3
O
4
2 4
O
3
O
4
3
O
4
2 4
O
4
+
2
O
4
was
3 4,
spheres [36,51]. On the contrary, the reduction peaks of Mn , Mn O
7

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
polymerization or cracking [43]. Simutaneously, a small amount of DFF
is further oxidized into FFCA and other products, resulting in the de-
creased selectivity of DFF.
Catalyst recycling is of great importance in the field of hetero-
2 4
geneous catalysis. The reusability and stability of spinel CoMn O
hollow spheres was investigated in six successive consecutive runs for
the selective oxidation of HMF into DFF with the same conditions. After
each reaction, the catalyst was dried at 70 °C for 10 h after being wa-
shed with ethanol more than three times, and then calcined at 350 °C
for 3 h before reused. For comparison, another catalyst without thermal
treatment (1*) was also performed for the reaction. As shown in Fig. 14,
the HMF conversion of 1* seriously declines to 14.8%, probably owing
to the substrate accumulated on the catalyst surface. On the contrary,
both HMF conversion and DFF selectivity of the catalysts after calcined
only decrease about 2% after six runs, demonstrating that the spinel
2 4
CoMn O hollow spheres are highly stable during the selective oxida-
tion of HMF to DFF and the catalytic activity can be recovered well by
calcination.
2 4
Fig. 14. Recyclability experiments with spinel CoMn O hollow spheres for the
selective oxidation of HMF into DFF.
4. Conclusion
and CoO have obvious shifts to lower temperatures with the addition of
Co species, implying that synergistic effect between Co and Mn im-
2 4
In summary, cost-effective spinel CoMn O hollow spheres have
been prepared by a solvothermal method and the self-assembly for-
mation processes were detailed studied. The as-prepared CoMn O
2 4
hollow spheres system exhibited high catalytic activity for the selective
oxidation of HMF into DFF in comparison with single metal oxide.
2 4
Particllarly, CoMn O hollow spheres with Co/Mn molar ratio of 2/3
showed the highest catalytic performance with 41.6% HMF conversion
and 100% DFF selectivity due to the strong synergistic effect of metal
oxides between Mn and Co species. In addition, the catalyst could be
readily reused without significant loss of its catalytic activity. Thus, the
2 4
spinel CoMn O hollow spheres with superior catalytic activity for the
selective oxidation of HMF are expected to be useful in many other
applications.
proves the oxygen mobility and reducibility in the spinel CoMn
52]. Therefore, spinel CoMn catalysts have better catalytic activ-
ities than those of Mn and Co , which are well agreed with the
2 4
O
[
2 4
O
3
O
4
3 4
O
results in Table 1. In addition, the peak areas indicate hydrogen con-
sumption amounts and the lowest reduction peak area of different
spinel CoMn
process [53]. It is found that the peak area ratios of the lowest reduction
peak obtained by integration are 2:2:1.4:1 for spinel CoMn with Co/
Mn radios of 1:2, 2:3, 1:1, and 2:1. According to the reducibility tested
by H -TPR profiles, spinel CoMn -2:3 shows both the maximum re-
2 4
O illustrates the reduction capacity of the same reduction
2 4
O
2
2 4
O
duction peak area and the lowest reduction temperature. Therefore, it
possess the maximum amount of mobile oxygen species to be adsorbed
and further generate active oxygen species, which would be involved in
the reaction. As a result, the CoMn
2
O
4
-2:3 with the highest reduction
References
4
+
amount of Mn
and the lowest reduction peak shows the best per-
formance, as shown in Table 1.
[
1] A.A. Rosatella, S.P. Simeonov, R.F.M. Frade, C.A.M. Afonso, Green Chem. 13 (2011)
54–793.
[2] M.J. Climent, A. Corma, S. Iborra, Green Chem. 16 (2014) 516–547.
Aerobic oxidation of HMF to DFF was also carried out to study the
7
effects of amounts of spinel CoMn
Fig. 12, there is only about 8.0% HMF conversion with 14.9% se-
lectivity to DFF in the absence of spinel CoMn hollow spheres due to
self-polymerization of HMF. With the addition of 25 mg of the
CoMn -2:3 catalysts, the catalytic performance significantly im-
2 4
O hollow spheres. As shown in
[
[
[
3] X. Li, Y. Zhang, Green Chem. 18 (2016) 643–647.
4] D.M. Alonso, J.Q. Bond, J.A. Dumesic, Green Chem. 12 (2010) 1493–1513.
5] G. Yong, Y. Zhang, J.Y. Ying, Angew. Chem. Int. Ed. 47 (2008) 9345–9348.
2 4
O
[6] C.T. Chen, C.V. Nguyen, Z.Y. Wang, Y. Bando, Y. Yamauchi, M.T.S. Bazziz,
A. Fatehmulla, W.A. Farooq, T. Yoshikawa, T. Masuda, K.C.W. Wu, ChemCatChem
1
2 4
O
0 (2018) 361–365.
7] J. Ma, Z. Du, J. Xu, Q.H. Chu, Y. Pang, ChemSusChem 4 (2011) 51–54.
[8] M. Shaikh, S.K. Singh, S. Khilari, M. Sahua, K.V.S. Ranganatha, Catal. Commun. 106
2018) 64–67.
9] K.R. Vuyyuru, P. Strasser, Catal. Today 195 (2012) 144–154.
10] Y.Q. Zhu, M.N. Shen, Y.G. Xia, M. Lu, Catal. Commun. 64 (2015) 37–43.
proves, however the active sites are not sufficient to activate the sub-
strate. When the catalyst amount increases to 50 mg, 41.6% conversion
of HMF and ca. 100% selectivity of DFF are obtained. The more the
catalytic amount, the more active sites are provided, the more HMF is
converted in the reaction. Thus the HMF conversion increases and DFF
selectivity is up to 100%. HMF conversion maintains constant, but DFF
selectivity shows a downward trend with a continual increase in cata-
lyst amount. It is the fact that excessive catalyst amount provides too
many active sites, which not only catalyzes the selective aerobic oxi-
dation of HMF to DFF, but further oxidizes DFF into other products,
such as 5-formylfuroic acid, 2,5-furandicarboxylic acid, and so on.
The influence of reaction time on the conversion of HMF and se-
[
(
[
[
[11] G.D. Yadav, R.V. Sharma, Appl. Catal. B Environ. 147 (2014) 293–301.
[12] K. Ghosh, R.A. Molla, M.A. Iqubal, S.S. Islama, S.M. Islam, Appl. Catal. A Gen. 520
(2016) 44–52.
[
13] Z.Y. Gui, W.R. Cao, S. Saravanamurugan, A. Riisager, L.F. Chen, Z.W. Qi,
ChemCatChem 8 (2016) 3636–3643.
[14] Z.Z. Yang, W. Qi, R.X. Su, Z.M. He, ACS Sustain. Chem. Eng. 5 (2017) 4179–4187.
[
[
[
[
15] S. Biswas, B. Dutta, A. Mannodi-Kanakkithodi, R. Clarke, W.Q. Song,
R. Ramprasadc, S.L. Suib, Chem. Commun. 53 (2017) 11751–11754.
16] I. Sádaba, Y.Y. Gorbanev, S. Kegnæs, S.S.R. Putluru, R.W. Berg, A. Riisager,
ChemCatChem 5 (2013) 284–293.
17] F. Neatu, N. Petrea, R. Petre, V. Somoghi, M. Florea, V.I. Parvulescu, Catal. Today
lectivity to DFF is studied with the CoMn
Fig. 13. The HMF conversion is about 24.1%, 28.6%, and 41.6% in 0.5,
, and 2 h of reaction time together with selectivity to DFF about 100%,
2 4
O -2:3 catalyst as shown in
2
78 (2016) 66–73.
18] D.X. Martínez-Vargas, J.R.D.L. Rosa, L. Sandoval-Rangel, J.L. Guzmán-Mara,
M.A. Garza-Navarrob, C.J. Lucio-Ortiza, D.A.D.H. Ríoa, Appl. Catal. A Gen. 547
1
(
2017) 132–145.
19] L.M. Ning, S.Y. Liao, Y.F. Sun, L.H. Yu, X.L. Tong, Waste Biomass Valori. 9 (2018)
5–101.
[20] B.Y. Lin, Y. Liu, L. Heng, J. Ni, J.X. Lin, L.L. Jiang, Catal. Commun. 101 (2017)
5–19.
respectively. The longer the reaction time is, the more HMF is con-
verted in the reaction and the conversion increases. With prolonged
reaction time, the conversion of HMF increases gradually to 47.2% (4 h)
and 58.9% (8 h) while selectivities to DFF decreases sharply to 88.9%
and 67.3%, respectively. It needs to be pointed out that the yield of DFF
is retained by improving reaction time from 2 to 8 h. The main reason
for the reduction of DFF selectivity is side reaction occurs such as HMF
[
9
1
[
21] Y. Cao, Z. Xing, M. Hu, Z.Z. Li, X.Y. Wu, T.Y. Zhao, Z.Y. Xiu, S.L. Yang, W. Zhou, J.
Catal. 356 (2017) 246–254.
[22] Z. You, Y. Su, Y. Yu, H. Wang, T. Qin, F. Zhang, Q.H. Shen, H. Yang, Appl. Catal. B
Environ. 213 (2017) 127–135.
8

L. Ding et al.
Catalysis Today xxx (xxxx) xxx–xxx
[38] Y.R. Liu, B.C. Zhang, J.K. Feng, S.L. Xiong, RSC Adv. 5 (2015) 26863–26871.
[39] H.Y. Chen, A. Sayari, A. Adnot, F. Larachi, Appl. Catal. B Environ. 32 (2001)
195–204.
[
[
[
23] Y. Cheng, Z. Li, Y. Li, S.S. Dai, G.B. Ji, H.Q. Zhao, J.M. Cao, Y.W. Du, Carbon 127
(2018) 643–652.
24] Y. Cheng, J.M. Cao, Y. Li, Z.Y. Li, H.Q. Zhao, G.B. Ji, Y.W. Du, ACS Sustain. Chem.
Eng. 6 (2018) 1427–1435.
25] L. Wu, Y. Hu, X. Zhang, J.Q. Liu, X. Zhu, S.K. Zhong, J. Power Sources 374 (2018)
[40] D.X. Yan, J.Y. Xin, C.Y. Shi, X.M. Lu, L.L. Ni, G.Y. Wang, S.J. Zhang, Chem. Eng. J.
323 (2017) 473–482.
4
0–47.
[41] L.F. Liao, Y. Liu, Z.Y. Li, J.P. Zhuang, Y.B. Zhou, S.Z. Chen, RSC Adv. 6 (2016)
94976–94988.
[
26] Y. Yi, G.H. Lee, J.C. Kim, H.W. Shim, D.W. Kim, Chem. Eng. J. 327 (2017) 297–306.
[
27] S.C. Cai, Z.H. Meng, H.L. Tang, Y. Wang, P. Tsiakarasc, Appl. Catal. B Environ. 217
[42] K.W. Omari, J.E. Besaw, F.M. Kerton, Green Chem. 14 (2012) 1480–1487.
[43] H. Choudhary, S. Nishimura, K. Ebitani, Appl. Catal. A Gen. 458 (2013) 55–62.
[44] W. Huang, Sj. Ding, Y. Chen, W.J. Hao, X.Y. Lai, J. Peng, J.C. Tu, Y. Cao, X.T. Li, Sci.
Rep. 7 (2017) 5220–5231.
(2017) 477–484.
[
28] Z. Huang, H.Y. Pan, W.J. Yang, H.H. Zhou, N. Gao, C.P. Fu, S.C. Li, H.X. Li,
Y.F. Kuang, ACS Nano 12 (2018) 208–216.
[
[
29] L. Yu, H. Hu, H.B. Wu, X.W. Lou, Adv. Mater. 29 (2017) 1604563–1604602.
30] L.F. Shen, L. Yu, X.Y. Yu, X.G. Zhang, X.W. Lou, Angew. Chem. Int. Ed. 54 (2015)
[45] X.H. Zhou, W.M. Kang, W. Xu, B.W. Cheng, RSC Adv. 5 (2015) 84535–84542.
[46] E. Hayashi, T. Komanoya, K. Kamata, M. Hara, ChemSusChem 10 (2016) 654–658.
[47] Z.Y. Gui, S. Saravanamurugan, W.R. Cao, L. Schill, L.F. Chen, Z.W. Qi, A. Riisager,
ChemistrySelect 2 (2017) 6632–6639.
1
868–1872.
31] F.X. Ma, H. Hu, H.B. Wu, C.Y. Xu, Z.C. Xu, L. Zhen, X.W. Lou, Adv. Mater. 27 (2015)
097–4101.
[
4
[48] Y.W. Cheng, X. Li, L.J. Wang, Q.B. Wang, Ind. Eng. Chem. Res. 45 (2006)
[32] Y. Lee, D. Thirumalai, C. Hyeon, J. Am. Chem. Soc. 139 (2017) 12334–12337.
4156–4162.
[
33] Q.H. Wei, Q.X. Ma, P.P. Zuo, H.L. Fan, S.J. Qu, W.Z. Shen, ChemCatChem 10 (2018)
[49] S.A. Chavan, S.B. Halligudi, D. Srinivas, P. Ratnasamy, J. Mol. Catal. A Chem. 161
(2000) 49–64.
[50] S. Todorova, H. Kolev, J.P. Holgado, G. Kadinov, Ch. Bonev, R. Pereguez,
A. Caballero, Appl. Catal. B Environ. 94 (2010) 46–54.
[51] H. Chen, M. Yang, S. Tao, G. Chen, Appl. Catal. B Environ. 209 (2017) 648–656.
[52] A.C. Zhang, W.W. Zheng, J. Song, S. Hu, Z.C. Liu, J. Xiang, Chem. Eng. J. 236
(2014) 29–38.
[53] J.H. Zhang, Y.B. Li, L. Wang, C.B. Zhang, H. He, Catal. Sci. Technol. 5 (2015)
2305–2313.
1
019–1026.
34] H. Lin, Z. Hao, X.R. Zheng, Y.M. Huang, P. Zhang, Q.W. Chen, Sci. Rep. 2 (2012)
86–994.
35] V. Stevanović, M. D’Avezac, A. Zunger, J. Am. Chem. Soc. 133 (2011)
1649–11654.
[
[
9
1
[
[
36] J. Ajay, K.R. Patil, C.V. Rode, ChemPlusChem 78 (2013) 1384–1392.
37] Z.M. Huang, W. Zhou, C. Ouyang, J. Wu, F. Zhang, J.G. Huang, Y.Q. Gao, J.H. Chu,
Sci. Rep. 5 (2015) 10899–10907.
9