Efficient aerobic oxidation of 5-hydroxymethylfurfural to 2,5-diformylfuran on manganese oxide catalysts

Article abstract of DOI:10.1016/j.jcat.2014.05.003

A cryptomelane-type manganese oxide octahedral molecular sieve with a (2 × 2, 4.6 A × 4.6 A) tunnel size (OMS-2) efficiently catalyzed aerobic oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-diformylfuran (DFF) with a high yield of 97.2% at 383 K and 0.5 MPa O ₂ in N,N-dimethylformamide. OMS-2 was superior to other MnO ₂ catalysts with different morphologies, including OMS-1, OMS-6, and OMS-7 with various tunnel sizes, amorphous MnO₂ and birnessite-type MnO₂, apparently due to its (2 × 2) tunnel structure and consequently high reducibility and oxidizability. Kinetic and isotopic studies on OMS-2 showed near half-order dependence of the activities on HMF and O ₂ concentrations and marked kinetic isotope effects for deuterated HMF at its methylene group. These results, together with the similar initial rates under aerobic and anaerobic conditions, suggest that HMF oxidation to DFF on OMS-2 proceeds via a redox mechanism involving kinetically-relevant steps of C-H bond cleavage in adsorbed alcoholate intermediate, derived from quasi-equilibrated dissociation of HMF, using lattice oxygen and reoxidation of Mn³⁺ to Mn⁴⁺ by dissociative chemisorption of O ₂.

Full text of DOI:10.1016/j.jcat.2014.05.003

Journal of Catalysis 316 (2014) 57–66
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Efficient aerobic oxidation of 5-hydroxymethylfurfural
to 2,5-diformylfuran on manganese oxide catalysts
⇑
Junfang Nie, Haichao Liu
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Stable and Unstable Species, College of Chemistry and Molecular
Engineering, Peking University, Beijing 100871, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A cryptomelane-type manganese oxide octahedral molecular sieve with a (2 ꢀ 2, 4.6 Å ꢀ 4.6 Å) tunnel
size (OMS-2) efficiently catalyzed aerobic oxidation of 5-hydroxymethylfurfural (HMF) to 2,5-diformylfu-
ran (DFF) with a high yield of 97.2% at 383 K and 0.5 MPa O2 in N,N-dimethylformamide. OMS-2 was
superior to other MnO2 catalysts with different morphologies, including OMS-1, OMS-6, and OMS-7 with
various tunnel sizes, amorphous MnO2 and birnessite-type MnO2, apparently due to its (2 ꢀ 2) tunnel
structure and consequently high reducibility and oxidizability. Kinetic and isotopic studies on OMS-2
showed near half-order dependence of the activities on HMF and O2 concentrations and marked kinetic
isotope effects for deuterated HMF at its methylene group. These results, together with the similar initial
rates under aerobic and anaerobic conditions, suggest that HMF oxidation to DFF on OMS-2 proceeds via a
redox mechanism involving kinetically-relevant steps of C–H bond cleavage in adsorbed alcoholate inter-
mediate, derived from quasi-equilibrated dissociation of HMF, using lattice oxygen and reoxidation of
Mn3+ to Mn4+ by dissociative chemisorption of O2.
Received 25 January 2014
Revised 1 May 2014
Accepted 3 May 2014
Keywords:
Biomass
5-Hydroxymethylfurfural
2,5-Diformylfuran
Aerobic oxidation
Manganese oxide
Reaction mechanism
Ó 2014 Elsevier Inc. All rights reserved.
1. Introduction
a high catalyst/substrate ratio (2:1 wt/wt) is required. Supported
Ru-based catalysts [20–22], such as Ru(OH)x/HT, Ru/C, and RuCl3/
5-Hydroxymethylfurfural (HMF) is an important platform mol-
ecule, which can be synthesized readily from cellulose-derived
sugars or directly from cellulose [1–4] and applied to produce
many value-added chemicals via different catalytic reactions [5].
One of the noteworthy examples is selective oxidation of HMF to
2,5-diformylfuran (DFF), a versatile chemical intermediate for the
synthesis of pharmaceuticals [6,7], functional polymers [8–10],
and other important products [5].
Al2O3, can afford good DFF yields (92–97%), but the Ru catalysts
are expensive. Therefore, it is highly desirable to develop more
practical and efficient catalysts for the synthesis of DFF.
Recently, a cryptomelane-type manganese oxide octahedral
molecular sieve (KMn8O16ꢁnH2O; OMS-2), which possesses
a
2 ꢀ 2 hollandite structure with one-dimensional pores [23] and
also high efficacy in many oxidation reactions [24–26], has been
reported to be a cheap and efficient catalyst for the HMF oxidation
to DFF [27]. Similar catalytic performance was achieved by Yadav
and Sharma on an OMS-2 supported Ag catalyst [28]. However,
in these preliminary reports, the key fundamental issues including
the relationship between the structure of OMS-2 and its perfor-
mance and the reaction mechanism remain unclear, which are
unequivocally important for further improving the catalyst effi-
ciency in the DFF synthesis. Herein, we present a detailed study
on the HMF oxidation to DFF on OMS-2 and other MnO2 catalysts
with different morphologies. Their structures and redox properties
are characterized by X-ray diffraction (XRD), transmission electron
microscopy (TEM), Brunauer–Emmett–Teller (BET), X-ray photo-
electron spectra (XPS), H2-temperature-programmed reduction
(H2-TPR), and O2-temperature-programmed oxidation (O2-TPO)
and correlated with their HMF oxidation activities. Reaction kinetic
It is known that the presence of the more reactive a,b-unsatu-
rated aldehyde group in HMF leads to the challenge in the aerobic
oxidation of HMF to DFF with a high yield [11]. In this regard, var-
ious catalysts have been explored to date for the aerobic oxidation
of HMF to DFF [11–18]. Quantitative DFF yields can be achieved by
homogeneous catalysts [11,12], but they encounter difficulties,
such as separation between catalysts and products. On the other
hand, heterogeneous catalysts generally suffer low activity or low
DFF selectivity [13–18] or require use of noble metals [19–22].
For example, an 85% DFF yield is obtained on VOx/TiO2 [13], but
⇑
Corresponding author. Fax: +86 10 6275 4031.
E-mail address: hcliu@pku.edu.cn (H. Liu).
http://dx.doi.org/10.1016/j.jcat.2014.05.003
0021-9517/Ó 2014 Elsevier Inc. All rights reserved.

58
J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
and isotopic studies are also carried out to elucidate the reaction
mechanism for the HMF oxidation on OMS-2.
2.3. HMF oxidation reactions
HMF oxidation reactions were carried out in a Teflon-lined
stainless steel autoclave (50 mL). Typically, 1 mmol HMF (98%, Alfa
Aesar) and 50 mg catalysts were introduced to 10 mL DMF (J.T.
Baker, 60.02% H2O) in the autoclave and were stirred at ca.
700 rpm. The reactants and products were analyzed by HPLC (Shi-
madzu LC-20A) using a UV detector and an Alltech OA-1000
organic acid column (0.005 M H2SO4 mobile phase, 0.7 mL/min
flow rate, and 353 K oven temperature). HMF reaction activities
were reported as molar HMF conversion rates per gram (or m2)
of catalysts per hour (i.e., mmol HMF/(gcat h) or mmol HMF/
(m2cat h)) and selectivities on a carbon basis. For catalyst recycling
tests, if not specially stated, the retrieved catalysts were washed
thoroughly with deionized water and then dried in vacuum oven
before being recycled.
2. Experimental
2.1. Catalyst preparation
OMS-2 was prepared by a reflux route [23]. Briefly, 30 mL aque-
ous solution of KMnO4 (2.2 g) was added dropwise to a solution of
MnSO4ꢁH2O (3.0 g) in water (10 mL) and a concentrated HNO3 solu-
tion (1.0 mL) under magnetic stirring. The resulting slurry was
refluxed at 383 K for 24 h. Afterward, the obtained dark brown pre-
cipitates were filtered and washed with deionized water until the
filtrate was neutral, which were then dried at 383 K overnight.
For comparison, other MnO2 catalysts including amorphous
MnO2 (AMO), birnessite-type MnO2 (Na-OL-1), OMS-7,
c-MnO2,
OMS-6, and OMS-1 were also prepared according to the methods
reported in the literature [29–31]. Mn2O3 and Mn3O4 were pre-
pared by calcination of OMS-2 in a N2 flow at 873 and 1173 K,
respectively. MnO was obtained by calcination of MnCO3 at
873 K in a 20% H2/N2 flow. CrO3, V2O5, and MoO3 were purchased
(AR, Sinopharm Chemical) and used directly.
2.4. Kinetic isotope effects and oxygen isotope exchange
Kinetic isotope effects on the HMF oxidation were examined
using
a-deuterated HMF (CHO–C4H2O–CH2OD) and b-deuterated
HMF (CHO–C4H2O–CD2OH), i.e. HMF molecules deuterated at the
hydroxyl (–OH) and methylene (–CH2–) groups, respectively, for
comparison with undeuterated HMF. The deuterated HMF mole-
cules were prepared according to our previous report [21].
Oxygen isotope exchange reactions between 16O2 and 18O2 were
carried out under the same reaction conditions described above for
the HMF oxidation, using a mixture of 0.25 MPa 18O2 (97% isotopic
purity, GAISI) and 0.25 MPa 16O2. The oxygen isotopomers 16O2,
16O18O, and 18O2 were measured by mass spectrometry (Hiden
HPR20).
2.2. Catalyst characterization
BET surface areas of the catalysts were obtained by nitrogen
physisorption at 78.3 K. The measurement was taken on a Microm-
eritics ASAP 2010 analyzer after the samples were treated at 423 K
under dynamic vacuum conditions to desorb the possible adsor-
bates (e.g., H2O and CO2).
The Mn contents in the reaction filtrate were determined by
inductively coupled plasma-atomic emission spectroscopy (ICP-
AES, Profile Spec, Leeman Labs).
3. Results and discussion
XRD patterns of the as-prepared catalysts were taken on a Rig-
aku D/Max-2000 diffractometer in the 2h range of 10–80° using a
3.1. Characterization of MnO2 catalysts
Cu Ka radiation source (k = 1.5406 Å) at a scanning rate of 4°/
min. The measurement was taken at 40 kV and 100 mA.
TEM images were obtained on a Philips Tecnai F30 FEG-TEM
operated at 300 kV. Before the measurement, the catalyst powders
were treated ultrasonically in ethanol for 30 min and then depos-
ited on carbon-coated Cu grids.
Fig. 1 shows the XRD patterns for the MnO2 catalysts. AMO only
shows a broad peak at 2h = 37.2°. The peaks of Na-OL-1 at 2h = 12.5
and 25.0° are characteristic features of birnessite materials with
layered structures (JCPDS 43-1456), corresponding to the (001)
and (002) crystal phases, respectively [32,33]. The patterns of
XPS analyses were performed on an AXIS Ultra (Kratos, UK)
spectrometer using an Al anode (Al Ka, hm = 1486.6 eV). The mea-
suring chamber was operated at ꢂ5 ꢀ 10ꢃ9 Torr. The binding ener-
gies were calibrated by referring to the C1s line at 284.6 eV.
H2-TPR experiments were carried out on a TP-5000 (Tianjin
Xianquan) flow unit. Typically, 20 mg samples were placed in a
quartz cell and then heated in a 5% H2/N2 flow (30 mL/min). The
temperature ramped from 298 to 923 K at a rate of 10 K/min.
The consumption of H2 was calibrated by the reduction in pure
CuO powder. The initial reduction rates for the MnO2 catalysts,
denoted as Rred,497K, were estimated by the reduction rates at
497 K, corresponding to the temperature at which 5% of the lattice
oxygen atoms of OMS-2 were reduced by H2.
O2-TPO experiments were performed on a TP-5080 (Tianjin
Xianquan) flow unit in a way similar to the H2-TPR measurement.
Before the O2-TPO, the samples were reduced in a 5% H2/N2 flow
(30 mL/min) for 1 h at 497 K, a low temperature to avoid destroy-
ing the tunnel structures of the OMS materials. After the samples
were cooled to 298 K, they were heated in a 5% O2/He flow
(20 mL/min) from 298 to 723 K at a rate of 10 K/min. The consump-
tion of O2 was calibrated by the oxidation of Cu powder. The initial
oxidation rates for the MnO2 catalysts, denoted as Roxi,383K, were
estimated by the oxidation rates at 383 K, the reaction temperature
typically used for HMF oxidation in this work.
Fig. 1. XRD patterns for MnO2 catalysts with different morphologies.

J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
59
OMS-7 and
c
-MnO2 are consistent with those of pyrolusite-type
icantly decrease the effective pore sizes of these MnO2 catalysts
[24,25,36]. It appears that the inner pores of the MnO2 catalysts
are not accessible to the reactants, namely HMF (0.82 nm) and oxy-
gen (0.35 nm), and thus, the HMF oxidation reaction mainly occurs
on their external surfaces.
The TEM image of the OMS-2 sample shows a nanorod mor-
phology with typical diameters of ꢂ30 nm and lengths of 100–
500 nm (Fig. 2a). The fringe distances of 0.68 nm correspond to
the lattice spacing of the (101) plane, growing preferentially along
the direction of [001] (Fig. 2b). The other MnO2 catalysts show
similar nanorod morphologies, except amorphous AMO and
layered Na-OL-1 (Fig. S1). These nanorods have similar lengths of
several hundred nanometers, although the diameters of OMS-7
(50–100 nm) are larger than those of the other samples
(20–30 nm). The larger diameters of OMS-7 reflect its higher crys-
tallinities, as characterized by XRD (Fig. 1).
(JCPDS 24-0735) [29,30] and ramsdellite-type MnO2 (JCPDS 14-
0644) [30,34], which possess tunnel sizes of 1 ꢀ 1 (2.3 Å ꢀ 2.3 Å)
and 1 ꢀ 2 (2.3 Å ꢀ 4.6 Å), respectively. The pattern of OMS-2 can
be well indexed to a pure tetragonal phase of cryptomelane-type
MnO2 (JCPDS 44-0141) with a tunnel size of 2 ꢀ 2 (4.6 Å ꢀ 4.6 Å)
[29,30]. OMS-6 shows the XRD pattern of romanechite-type (JCPDS
14-0627) MnO2 with a tunnel size of 2 ꢀ 3 (4.6 Å ꢀ 6.9 Å). The
peaks of OMS-1 appear at 2h = 9.4 (not shown here) and 18.6°,
reflecting that it possesses a todorokite structure with a tunnel size
of 3 ꢀ 3 (6.9 Å ꢀ 6.9 Å) [31,35]. Taken together, these results dem-
onstrate that the MnO2 catalysts with expected structures were
prepared. Moreover, it can be seen that except OMS-7 and AMO,
these MnO2 catalysts possess similar crystallinities, consistent
with the following TEM characterization.
The Mn contents of the MnO2 catalysts were measured by ICP-
AES. As shown in Table 1, OMS-7 and
c
-MnO2 contained 60.0–
O1s XPS spectra were used to probe the surface oxygen species
of the MnO2 catalysts. As shown in Fig. S2 and Scheme S1, three
types of surface oxygen species can be identified from their binding
energies. The peaks at low binding energies (529.3–529.7 eV),
medium binding energies (530.8–531.1 eV), and high binding
energies (532.6–533.6 eV) are ascribed to the lattice oxygen atoms
(OI), oxygen vacancies or surface adsorbed O (OH) groups (OII), and
adsorbed water (OIII), respectively [24,37]. Accordingly, the relative
fractions of these oxygen species were estimated and are listed in
Table 2. Na-OL-1, OMS-7, and OMS-2 possess the most abundant OI
species (71.5–72.8%). In contrast, lower OI species fractions (45.3–
65.2%) are observed for OMS-1 and AMO. On the other side, the
fraction of the OII species is 38.1% for OMS-1, which is lower
60.7 wt% Mn, while slightly lower Mn contents (53.8–56.3 wt%)
were detected in the other MnO2 catalysts, namely AMO, Na-OL-
1, OMS-2, OMS-6, and OMS-1, mostly likely as a result of their larger
sizes of pores that could accommodate other cations (e.g., K, Na, and
Mg) up to ꢂ5 wt%. These MnO2 catalysts possessed BET surface
areas in
a
narrow range of ꢂ40–80 m2/g, except for AMO
(189 m2/g) and OMS-7 (17 m2/g). The BET surface areas of these
MnO2 catalysts may be mainly contributed from their external sur-
faces as nitrogen molecule (0.36 nm) could hardly enter their pores
of smaller sizes (e.g., the pore size of OMS-2 is ꢂ0.26 nm)
[24,25,36]. According to the previous reports, the accommodation
of other cations, such as K, Na, and Mg, inside the pores can signif-
Table 1
Mn contents and surface areas of MnO2 catalysts with different morphologies.
Entry
Catalyst
AMO
Na-OL-1
OMS-7
c
-MnO2
OMS-2
OMS-6
OMS-1
1
2
Mn (%)
53.8
189
55.7
57
60.7
17
60.0
38
56.0
81
56.3
53
54.6
47
SBET (m2/g)
(b)
(a)
0.68 nm
[110]
10 nm
Fig. 2. TEM images of OMS-2.
Table 2
Relative fractions of surface oxygen species for MnO2 catalysts with different morphologies.
Entry
Catalyst
OI
OII
OIII
B.E. (eV)
Fractions (%)
B.E. (eV)
Fractions (%)
B.E. (eV)
Fractions (%)
1
2
3
4
5
6
7
AMO
Na-OL-1
OMS-7
529.6
529.6
529.3
529.6
529.7
529.6
529.6
65.2
71.9
71.5
66.3
72.8
68.3
45.3
531.1
531.1
530.8
531.1
531.1
531.1
530.8
26.6
17.8
21.3
27.4
21.4
20.2
38.1
533.2
533.6
532.6
533.1
533.2
533.4
532.4
8.2
10.3
7.2
6.3
5.8
c
-MnO2
OMS-2
OMS-6
OMS-1
11.5
16.5

60
J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
Table 3
(26.6–27.4%) for AMO and
other samples.
c-MnO2, and only 17.8–21.4% for the
H2-TPR and O2-TPO results for MnO2 catalysts with different morphologies.
a
a
Entry Catalyst Rred,497K
mol H2/(m2cat h))
Roxi,383K
(l
mol O2/(m2cat h)) (TPR)
AOSb
AOSb
(XPS)
Mn3s XPS spectra were also collected to compare the valence of
the surface Mn species for these MnO2 catalysts (Fig. S3). It is
known that the Mn3s multiplet splitting is sensitive to the average
Mn oxidation states (AOS (XPS)) [24,38,39] and thus can be used to
determine the AOS (XPS) by the following equation [24,39]: AOS
(
l
1
2
3
4
5
6
7
AMO
150.6
Na-OL-1 184.0
8.2
3.76
3.76
3.92
3.60
3.91
3.39
3.42
3.72
3.62
3.83
3.56
3.71
3.53
3.64
24.6
20.1
12.4
17.0
24.5
19.5
OMS-7
-MnO2
119.3
56.8
94.2
46.4
26.9
c
(XPS) = 8.95 ꢃ 1.13ꢁ(
nitude of the Mn3s multiplet splitting (Table S1). The results show
that the AOS (XPS) is high (3.83) for OMS-7, low (3.53–3.56) for
DMn3s), where (DMn3s) represents the mag-
OMS-2
OMS-6
OMS-1
c
-
MnO2 and OMS-6, and medium (3.62–3.72) for the others. Such
difference in the fractions of the lattice oxygen atoms and AOS of
the Mn species for these MnO2 catalysts led to the difference in
their redox properties, as discussed below.
a
Rred,497K and Roxi,383K represent the rates of reduction/oxidation at 497 and
383 K, respectively.
b
AOS represents average oxidation state.
The reducibilities of the MnO2 catalysts were probed by H2-TPR.
As shown in Fig. 3a, all the MnO2 catalysts exhibited two or three
reduction peaks which can be ascribed to the sequential reduction
steps of MnO2, i.e. MnO2(?Mn2O3) ? Mn3O4 ? MnO [37,38,40]. It
was reported that the initial reduction stages for many metal oxi-
des are the chemical processes most relevant to the redox cycles
required for the catalytic turnovers in selective oxidation of differ-
ent hydrocarbons and alcohols [16,41–43]. Moreover, only 1.5–
3.0% of the lattice oxygen atoms on OMS-2 were found to be
involved in its redox cycles at 473 K, behaving as the active oxygen
species in NOx–NH3 selective catalytic reduction (SCR) and CO oxi-
dation reactions [25,37]. By referring to these previous studies, in
this work, the initial reduction rates at 497 K (Rred,497K) were used
to estimate and compare the reducibilities of the different MnO2
catalysts. As shown in Table 3, AMO and Na-OL-1 are the two most
reducible MnO2 catalysts, affording Rred,497K of 150.6 and
The oxidizabilities of the MnO2 catalysts were probed by
O2-TPO. Before the O2-TPO measurements, the samples were first
reduced in H2 at 497 K for 1 h. As shown in Fig. 3b, most of the
MnO2 catalysts exhibited oxidation peaks in a narrow temperature
range of ca. 450–620 K. Similar to the aforementioned definition of
Rred,497K in the H2-TPR profiles, the initial oxidation rates (Roxi,383K)
at 383 K can be used to compare the oxidizabilities of the MnO2
catalysts and their oxidation rates at the HMF oxidation tempera-
ture (383 K). As shown in Table 3, the Roxi,383K were as high as
24.5–24.6
and 12.4
O2/(m2cat h) for OMS-7, OMS-2, and OMS-1. Combined with the
red,497K results, it can be found that the Rred,497K of AMO is very high,
but it possesses the lowest Roxi,383K among the MnO2 catalysts. In
contrast, OMS-6 and OMS-1 possess low Rred,497K, but high Roxi,383K
-MnO2 exhibits both low Rred,497K and Roxi,383K, which are, however,
l
mol O2/(m2cat h) for Na-OL-1 and OMS-6, while only 8.2
l
mol O2/(m2cat h) for AMO and
c-MnO2, and 17.0–20.1 lmol
R
184.0
l
mol H2/(m2cat h), respectively. OMS-7 and OMS-2 are less
mol H2/(m2cat h)), while
-MnO2, OMS-
mol
.
c
reducible (119.3 and 94.2
l
c
high for Na-OL-1, OMS-7 and OMS-2. The origin of this difference in
the redox properties for these MnO2 catalysts, although remained
unclear, might be related to the discrepancy of their Mn–O bond
lengths and exposed planes [45,46]. These results suggest that the
structure of the MnO2 catalysts strongly influence their redox prop-
erties, which should subsequently determine their reaction activi-
ties, as indeed observed in the next section.
6, and OMS-1 are the least reducible catalysts (56.8–26.9
l
H2/(m2cat h)). This reducibility that tends for the MnO2 catalysts is
in accordance with the change in their AOS values calculated from
the H2-TPR results (AOS (TPR)). AMO, Na-OL-1, OMS-7, and OMS-2
with higher AOS values (3.76–3.92) were more reducible than
c-
MnO2, OMS-6, and OMS-1 with lower AOS values (3.39–3.60). Such
correlations between AOS and reducibility for these MnO2 catalysts
reflect the involvement of their lattice oxygen atoms in the HMF oxi-
dation, as discussed in the next section. Moreover, it is apparent
from Table 3 that the AOS values estimated by H2-TPR (AOS (TPR))
are in well agreement with those by XPS (AOS (XPS)), indicating
no significant reconstruction on the MnO2 surfaces [44].
3.2. HMF oxidation activity and DFF selectivity on MnO2 catalysts
Table 4 shows the HMF conversions and DFF selectivities in
N,N-dimethylformamide (DMF) at 383 K and 0.5 MPa O2 on
(a)
(b)
AMO
AMO
Na-OL-1
Na-OL-1
OMS-7
OMS-7
γ-MnO2
γ-MnO2
OMS-2
OMS-2
OMS-6
OMS-1
OMS-6
OMS-1
Fig. 3. H2-TPR (a) and O2-TPO (b) profiles for MnO2 catalysts with different morphologies.

J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
61
Table 4
Table 5
Conversions and product selectivities in aerobic oxidation of HMF on MnO2 catalysts,
Activities and product selectivities in aerobic oxidation of HMF on MnO2 catalysts
and for comparison on other metal oxide catalysts.a
with different morphologies.a
Entry
Catalyst
Conversion (%)
Selectivity (%)
Entry Catalyst Activity
Activity
Selectivity
(mmol HMF/(gcat h)) (mmol HMF/(m2cat h)) (%)
DFF
FFCA
DFF FFCA
1
2
3
4
5
6
7
8
9
OMS-2
AMO
KMnO4
100
97.2
94.7
37.6
21.4
88.7
67.3
<1
60.3
49.5
69.1
–
1.7
2.6
59.0
93.4
12.1
37.8
8.2
1
2
3
4
5
6
7
AMO
Na-OL-1 95.7
OMS-7 12.2
-MnO2 14.5
24.1
0.13
1.68
1.11
0.38
0.96
0.34
0.14
95.3 1.3
92.0 1.2
81.5 1.4
90.1 1.3
97.6 1.4
82.7 1.1
92.3 1.3
33.4 (9.1)c
9.8
b
b
MnSO4
b
Mn2O3
5.2
10.8
–
13.7
4.8
2.3
c
b
Mn3O4
OMS-2
OMS-6
OMS-1
77.6
17.7
6.4
MnOb
CrO3
<1
77.6
60.8
6.0
V2O5
MoO3
Blank
383 K, 0.5 MPa O2, 1.0 mmol HMF, ꢂ1.8 m2 catalyst, 10 mL DMF, ꢂ30% HMF
a
10
11
conversion.
No reaction
–
a
b
c
383 K, 0.5 MPa O2, 1.0 mmol HMF, 50 mg catalyst, 10 mL DMF, 1 h.
The same amount of Mn was used as compared with OMS-2.
The value in the parenthesis is the selectivity to FDCA.
(0.34–0.38 mmol HMF/(m2cat h)). The DFF selectivities, however,
were superior on OMS-2 (97.6%) and AMO (95.3%), and slightly lower
on
c-MnO2, Na-OL-1, and OMS-1 (90.1–92.3%), while only 81.5–
different MnO2 catalysts. DMF was chosen as solvent because it
was superior, in terms of both the catalytic activity and DFF selec-
tivity, to the other polar and non-polar solvents, such as dimethyl-
sulfoxide (DMSO), H2O and CH2Cl2, in the HMF oxidation
(Table S2). As shown in Table 4, HMF was completely transformed
in 1 h on OMS-2, and the selectivity to DFF was as high as 97.2%,
corresponding to a 97.2% DFF yield. Only 1.7% FFCA was detected,
which indicates that DFF is stable on OMS-2 in DMF under the
reaction conditions employed in this work, as further verified by
the separate reaction of DFF under the identical conditions. AMO
shows a similar DFF selectivity (94.7%), but with a much lower
HMF conversion (59.0%). The precursors used for the preparation
of OMS-2, KMnO4, and MnSO4 were also tested. KMnO4 was effi-
cient for the HMF oxidation (93.4% conversion), but the DFF selec-
tivity was only 37.6%, as a result of the further oxidation of HMF or
DFF to FFCA (33.4%) and FDCA (9.1%). In contrast, MnSO4 shows
much lower HMF conversion (12.1%) and DFF selectivity (21.4%).
On other manganese oxides with lower valences, the oxidation
rates decreased dramatically. Mn2O3 exhibited only a moderate
HMF conversion (37.8%) with a DFF selectivity of 88.7%, while
Mn3O4 was much less active and MnO was almost inactive for
the HMF oxidation. Similar effect of oxidation state of manganese
oxides on their activities was also reported for SCR of NOx using
NH3 [47]. For comparison with OMS-2, other metal oxides were
examined (Table 4). CrO3 and V2O5 were also active for the HMF
oxidation, but showing lower HMF conversions (77.6% and 60.8%)
and DFF selectivities (60.3% and 49.5%). MoO3 was, however, not
effective under the identical reaction conditions (6.0% HMF conver-
sion and 69.1% DFF selectivity). These results show that only the
Mn(IV) dioxide with tunnel structures is efficient for the aerobic
oxidation of HMF to DFF.
82.7% on OMS-7 and OMS-6. These results clearly demonstrate the
strong effect of the structures of the MnO2 catalysts on their activi-
ties, among which OMS-2 is the most efficient catalyst for the syn-
thesis of DFF in term of its highest DFF selectivity and good activity.
The stability and recyclability of the OMS-2 catalyst were exam-
ined for six successive cycles in the HMF oxidation at 383 K and
0.5 MPa O2. OMS-2 was characterized by XRD (Fig. S4), showing
that its crystalline structure remained unaltered after the sixth
cycle in the HMF oxidation. The ICP-AES analysis of the filtrate
after each reaction cycle showed negligible leaching of the manga-
nese species (<0.1%). Using such filtrates, no further HMF oxidation
was detected after the removal of OMS-2, consistent with the ICP-
AES result and also confirming the heterogeneous catalysis nature
of OMS-2 in the HMF oxidation. Accordingly, as shown in Fig. S5,
after recycling the OMS-2 catalyst for six times, the HMF conver-
sion only decreased slightly to 84.2%, while the DFF selectivity
remained constant (97.0%). Such slight activity loss most likely
arose from the blockage of the active sites by carbonaceous by-
products on OMS-2, as implied by Mizuno et al. in the synthesis
of amides [48]. Consistent with such proposition, the activity of
the OMS-2 catalyst after the six cycles can be largely regenerated
(93.1% HMF conversion) upon calcination at 573 K for 1 h to
remove the adsorbed by-products (Fig. S5).
3.3. Effects of reaction parameters on HMF oxidation to DFF
Fig. 4 shows the effect of HMF concentrations on the activities
and selectivities for HMF oxidation on OMS-2 at 383 K and
0.5 MPa O2. At similar HMF conversions (ꢂ30%), the activities
increased almost linearly from 0.33 to 0.50 mmol HMF/(m2cat h)
with increasing HMF concentrations from 10 to 25 mmol/L, which
then increased gradually to reach a constant value of ꢂ0.60 mmol
HMF/(m2cat h) at above 100 mmol/L. Such effect indicates the satu-
rated adsorption of HMF-derived intermediates (e.g., alcoholate) on
the active OMS-2 surfaces, as also observed on Ru/C and other cata-
lysts in the HMF oxidation [13,16,21,49]. The reaction order was esti-
mated to be 0.44 0.06 with respect to HMF concentrations from the
linear range. The DFF selectivities decreased slightly from 98.3% to
96.8% with increasing the HMF concentrations from 10 to
150 mmol/L, while the FFCA selectivities were almost unchanged
(ꢂ1.0%).
Considering the broad variety of the tunnel structures of MnO2,
several other MnO2 catalysts including Na-OL-1, OMS-7, c-MnO2,
OMS-6, and OMS-1 were also examined for comparison with
OMS-2 and AMO, as shown in Table 5. These catalysts possess mor-
phologies varying from amorphous (AMO), layered-type (Na-OL-1)
to one-dimensional tunnel structures (the other MnO2 samples)
with tunnel sizes ranging from (1 ꢀ 1) to (3 ꢀ 3), as discussed
above (Figs. 2 and S1). When compared at similar HMF conversions
(ꢂ30%) in the kinetic regime, OMS-2 and Na-OL-1 exhibited much
superior mass activities (77.6 and 95.7 mmol HMF/(gcat h),
respectively) than the other MnO2 catalysts (6.4–24.1 mmol
HMF/(gcat h)). After normalized by their surface areas, the most
active catalysts turned to be Na-OL-1, OMS-7 and OMS-2 (1.68,
1.11, and 0.96 mmol HMF/(mc2at h), respectively). AMO and OMS-1
were the least active ones (0.13–0.14 mmol HMF/(m2cat h)), while
Similarly, the product selectivities are not sensitive to the O2
pressures. As shown in Fig. 5, the DFF (ꢂ97%) and FFCA (ꢂ1.0%)
selectivities remained almost unchanged at similar HMF conver-
sions (ꢂ30%) when O2 pressures increased from 0.1 to 2.0 MPa.
However, the effect of O2 pressures on the activities strongly
c-MnO2
and
OMS-6
showed
the
moderate
activities

62
J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
[50,51]. Such requirement for O2 adsorption is consistent with the
Mars–van Krevelen redox mechanism involving the lattice oxygen
atoms on OMS-2, as discussed in the next section. The reaction order
was estimated to be 0.47 0.01 with respect to O2 pressures in the
linear range (<1.0 MPa), similar to the order in the HMF concentra-
tions, indicating the dissociative chemisorption of O2 and HMF on
the OMS-2 surface, as observed on Ru/C [21].
We also studied the effect of water, an inevitable by-product, on
the HMF oxidation on OMS-2 at 383 K and 0.5 MPa O2. Water was
externally introduced into the reaction solutions by controlling its
concentrations in the range of 20–100 mmol/L, corresponding to
the amount of water produced in the HMF oxidation with ꢂ20–
100% conversion. As shown in Fig. 6, both the activities
(ꢂ0.60 mmol HMF/(m2cat h)) and the selectivities to DFF (ꢂ97.5%)
and FFCA (ꢂ0.7%) remained essentially constant with variation in
the H2O concentrations between 20 and 100 mmol/L, indicating
the lack of inhibition by water for the HMF oxidation on OMS-2, as
also observed on the Ru/C catalyst [21]. This result is consistent with
Fig. 4. Dependence of activities and selectivities to DFF and FFCA on HMF
concentrations for aerobic oxidation of HMF on OMS-2 at ca. 30% HMF conversion
(383 K, 0.5 MPa O2, 0.1–1.5 mmol HMF, 1–15 mg OMS-2, 10 mL DMF).
depends on the HMF concentrations employed. At a low HMF con-
centration (10 mmol/L, Fig. 5a), the activities increased slightly
from 0.24 to 0.33 mmol HMF/(m2cat h) with increasing O2 pressures
from 0.1 to 0.5 MPa and then reached
a constant value of
ꢂ0.35 mmol HMF/(m2cat h) up to 4.0 MPa. Such phenomenon is sim-
ilar to that observed on supported VOx catalysts [13,16]. Differently,
at a high HMF concentration (100 mmol/L, Fig. 5b), the activities
increased almost linearly from 0.36 to 0.70 mmol HMF/(m2cat h) with
increasing O2 pressures from 0.1 to 1.0 MPa and then increased more
gradually to 1.17 mmol HMF/(m2cat h) at 4.0 MPa. Such different
behavior indicates that at high HMF concentrations, the active sur-
faces of OMS-2 are saturated by the HMF-derived alcoholate species,
which hinder the adsorption of O2 and thus require higher O2 pres-
sures to improve the surface coverage of oxygen species and HMF
oxidation activities. Fu and coworkers [27] also found that the
change of oxidant from air to oxygen increased the HMF conversion
dramatically from 39% to 99% on OMS-2. Similar phenomenon was
reported in oxidation of volatile and phenol compounds on OMS-2
Fig. 6. Dependence of activities and selectivities to DFF and FFCA on H2O
concentrations for aerobic oxidation of HMF on OMS-2 at ca. 30% HMF conversion
(383 K, 0.5 MPa O2, 1.0 mmol HMF, 0.2–1.0 mmol H2O, 10 mg OMS-2, 10 mL DMF).
Fig. 5. Dependence of activities and selectivities to DFF and FFCA on O2 pressures for aerobic oxidation of HMF on OMS-2 at a low HMF concentration ((a) 0.1 mmol HMF, 0.1–
2.0 MPa O2) and high HMF concentration ((b) 1.0 mmol HMF, 0.1–4.0 MPa O2) at ca. 30% HMF conversion (383 K, 10 mg OMS-2, 10 mL DMF).

J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
63
the aforementioned dominant adsorption of the HMF-derived alco-
holate species on the OMS-2 surface, which thus suppresses the
adsorption of water. Such difference in adsorption of HMF and water
reflects the hydrophobic nature of the OMS-2 surfaces with stronger
affinity for organic species than for water, as reported by Suib and
coworkers [25], which can account for the observed low fraction of
the surface OIII species (from the adsorbed H2O molecules) on
OMS-2 by XPS (Table 2).
Fig. 7 shows the effect of the reaction temperature on the activ-
ities and selectivities in the HMF oxidation on OMS-2. At similar
HMF conversions (ꢂ30%), the activities increased dramatically
from 0.07 to 1.20 mmol HMF/(m2cat h) with increasing the tempera-
ture from 353 to 393 K. The apparent activation energy (Ea) was esti-
mated to be 81 kJ/mol, which is comparable with that of VOx/TiO2
catalyst (67–77 kJ/mol) [13,16]. The DFF selectivities decreased
slightly from 98.9% to 96.2% when reaction temperature increased
from 353 to 393 K while the FFCA selectivities were always below
1% in this temperature range. These results show that OMS-2 is
advantageous for the aerobic oxidation of HMF, and its activity can
be increased at higher temperatures without detriment to its high
DFF selectivity.
3.4. Reaction kinetics and mechanism of HMF oxidation
As discussed above, the activities of the MnO2 catalysts strongly
depend on their tunnel structures (Table 5). Such dependence is
found to be related to the structural effects on their redox
properties as probed by H2-TPR and O2-TPO (Table 3). The activities
(normalized per surface area) of the MnO2 catalysts, except for
AMO, show a good correlation with their reducibilities (Fig. 8a).
The more reducible Na-OL-1, OMS-2, and OMS-7 (94.2–184.0
H2/(m2cat h)) exhibited higher activities (0.96–1.68 mmol HMF/
(m2cat h)) while the less reducible OMS-1, OMS-6, and
-MnO2
(26.9–56.8
mol H2/(m2cat h)) led to lower activities (0.14–0.38 mmol
HMF/(m2cat h)). It is noted that AMO had
high reducibility
(150.6
mol H2/(m2cat h)), but it was the least active MnO2 catalyst
lmol
c
l
a
l
(0.13 mmol HMF/(m2cat h)), most likely as a result of its low oxidiz-
ability, as discussed below, and thus, the limitation of its redox cycle
required for the HMF oxidation. Similarly, the activities of the MnO2
catalysts, except for OMS-6 and OMS-1, increased linearly with their
oxidizabilities (Fig. 8b). OMS-6 and OMS-1 had very high oxidizabil-
ities (19.5–24.5
l
mol O2/(m2cat h)) but very low reducibilities
(26.9–46.4
l
mol H2/(m2cat h)), which may account for their low
activities. These results show that the efficient HMF oxidation on
the MnO2 catalysts requires their both high reducibilities and high
oxidizabilities, as exhibited by Na-OL-1, OMS-2, and OMS-7.
Such requirement indicates the involvement of the lattice oxy-
gen atoms on the MnO2 catalysts in the HMF oxidation, as indeed
confirmed by comparison of the anaerobic and aerobic experi-
ments on OMS-2 under N2 and O2 atmospheres, respectively.
Fig. 9 shows the DFF yields in the HMF oxidation as a function of
reaction time (within 35 min) on OMS-2 at 383 K, and the initial
reaction rates were extracted by using a first-order exponential
decay function to fit the reaction profiles for the first 2 min. The
initial rates were essentially the same (1.18 vs. 1.21 mmol HMF/
(m2cat h)) under N2 and O2, reflecting the reactivity of the lattice oxy-
gen atoms on OMS-2 for the HMF oxidation to DFF by the Mars-van
Krevelen redox mechanism (Scheme 1), which is consistent with the
weak kinetic consequences of O2 pressures on the HMF oxidation
observed at low HMF concentrations (Fig. 5a).
Fig. 7. Dependence of activities and selectivities to DFF and FFCA on reaction
temperatures for aerobic oxidation of HMF on OMS-2 at ca. 30% HMF conversion
(353–393 K, 0.5 MPa O2, 1.0 mmol HMF, 10 mg OMS-2, 10 mL DMF).
Fig. 8. Relationship between activities and reducibilities (a) and oxidizabilities (b) of MnO2 catalysts with different morphologies.

64
J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
(KIE) were studied with HMF molecules deuterated at the hydroxyl
group (R–CH2OD) and at the methylene group (R–CD2OH). Table 6
shows the KIE (kH/kD), defined as the ratio of the HMF oxidation
rates for undeuterated and deuterated HMF at 383 K and 0.5 MPa
O2 in the kinetic regime. R–CH2OD gave a KIE value of 1.03, close
to unity, which indicates that the dissociative adsorption of HMF
to form the adsorbed R–CH2Oꢄ species is quasi-equilibrated and
kinetically inconsequential during the HMF oxidation on OMS-2.
However, the use of R–CD2OH led to a significant KIE value of
4.19 at a low HMF concentration (0.1 mmol HMF). This value is
similar to those observed in the HMF oxidation on Ru/C (3.73)
[21] and in oxidation of other alcohols (e.g., p-methylbenzyl alco-
hol) on Ru(OH)x/Al2O3 (5.0–4.2) at 333–373 K in liquid phases
[49]. At a high HMF concentration (1.0 mmol HMF), a smaller KIE
value of 1.71, but still greater than unity, was found. Notwith-
standing such discrepancy at low and high HMF concentrations,
these normal KIE reveal that the C–H bond cleavage of the
adsorbed R–CH2Oꢄ species is kinetically-relevant in the HMF oxida-
tion on OMS-2. For the dissociative chemisorption of O2 to Oꢄ, its
irreversible nature was confirmed by oxygen isotope exchange
between 16O2 and 18O2, showing no detectable formation of the
mixed 18O16O isotopomer under the identical reaction conditions.
The O2 dissociation also appears to be kinetically-relevant in the
HMF oxidation, as reflected by the combined results at high HMF
concentrations including the observed strong effect of O2 pressures
on the HMF oxidation activities (Fig. 5b), the increased fraction of
the oxygen vacancies on the used OMS-2 surface characterized by
XPS (Fig. S6), and the smaller KIE value (Table 6).
Fig. 9. Initial reaction rates on OMS-2 for HMF oxidation under O2 and N2
atmospheres, respectively (383 K, 0.5 MPa O2, 1.0 mmol HMF, 5 mg OMS-2, 10 mL
DMF).
O2-
Mn4+
HMF
Moreover, the effect of a radical scavenger, e.g. 2,6-di-tert-
butyl-p-cresol, on the HMF oxidation was also examined on
OMS-2. The scavenger was introduced to the reaction solution,
and the DFF yield was 52.1% after reaction for 6 min at 383 K and
0.5 MPa O2 with 50 mg OMS-2, which was close to the value of
53.7% in the absence of the scavenger. This result suggests that rad-
ical intermediates, even present, did not involve in the kinetic rel-
evant steps of the HMF oxidation.
step 2
step 1
0.5 O2
DFF
Mn3+
Based on these kinetic and isotopic results, a sequence of ele-
mentary steps for the HMF oxidation on OMS-2 is proposed in
Scheme 2. In this sequence, the reaction is preceded by the disso-
ciative chemisorption of HMF on a reduced Mn center or oxygen
vacancy (denoted as ꢄ) of the OMS-2 surface via interaction with
its neighboring lattice oxygen Oꢄ to form the adsorbed R–CH2Oꢄ
and OHꢄ species (Step 1). R–CH2Oꢄ then undergoes irreversible
and kinetically-relevant b-H elimination by its reaction with Oꢄ
to form the adsorbed DFF (R–CHOꢄ) and OHꢄ species (Step 2). The
subsequent recombination of these OHꢄ groups forms the adsorbed
water (H2Oꢄ) and Oꢄ species (Step 3), which is assumed to be fast
and irreversible on the basis of the observed lack of inhibition by
H2O (Fig. 6). Afterward, both H2Oꢄ and R–CHOꢄ species readily des-
orb from the OMS-2 surface to liberate the reduced Mn centers (ꢄ)
(Steps 4 and 5). Finally, the reduced Mn centers are reoxidized by
Scheme 1. Proposed reaction mechanism involving Mn4+/Mn3+ redox cycles.
As depicted in Scheme 1, the HMF oxidation to DFF is proposed
to proceed via the Mn4+/Mn3+ redox cycles including the HMF oxi-
dation with the lattice oxygen atoms on OMS-2 and the reduction
of Mn4+ to Mn3+ (Step 1), followed by the replenishment of the con-
sumed lattice oxygen atoms (i.e., oxygen vacancies) by gaseous O2
and the reoxidation of Mn3+ to Mn4+ (Step 2). Characterization of
OMS-2 before and after the HMF oxidation at a high HMF concen-
tration (1.0 mmol HMF, 383 K and 0.5 MPa O2) by XPS showed the
fraction of the surface lattice oxygen atoms decreased from 72.8%
to 65.0%, while the fraction of the oxygen vacancies increased from
21.4% to 27.8% after the reaction (Fig. S6). The XPS result indicates
that the reoxidation of Mn3+ to Mn4+ by O2 is a slower step relative
to the Mn4+ reduction step in the redox cycles at high HMF concen-
trations, which may limit the overall rate of the HMF oxidation and
also account for the strong effect of O2 pressures on the activities of
OMS-2 observed at high HMF concentrations (Fig. 5b).
Table 6
Activities, selectivities, and kinetic isotopic effects (KIE) for aerobic oxidation of HMF
on OMS-2.a
The observed near half-order dependence of the HMF oxidation
activities on HMF and O2 concentrations (Figs. 4 and 5) suggests
that HMF and O2 adsorb dissociatively on OMS-2, similar to the
finding on Ru/C [21], to form the corresponding adsorbed alcohol-
ate (R–CH2Oꢄ, where R hereinafter denotes the furfural CHO–
C4H2O– part of HMF) and hydroxyl (OHꢄ) species and the adsorbed
atomic oxygen (Oꢄ) species, respectively. It has been reported that
alcohol oxidation generally involves the C–H bond cleavage of the
Entry
Reactant
Activity
(mmol HMF/(gcat h))
Selectivity (%)
KIE (kH/kD)
DFF
FFCA
1
2
3
4
5
R–CH2OH
R–CH2OD
R–CD2OH
R–CH2OHb
R–CD2OHb
48.3
46.9
28.2
26.8
6.4
97.5
94.8
85.2
98.3
86.0
1.3
1.4
2.6
1.1
2.3
–
1.03
1.71
–
4.19
adsorbed alcoholate species as
[16,26,43]. To understand the kinetic relevance of the C–H bond
cleavage in the HMF oxidation on OMS-2, kinetic isotope effects
a
kinetically-relevant step
a
383 K, 0.5 MPa O2, 1.0 mmol HMF, 10 mg OMS-2, 10 mL DMF, ꢂ30% HMF
conversion.
b
0.1 mmol HMF.

J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
65
Scheme 2. Proposed reaction pathways for aerobic oxidation of HMF to DFF on
OMS-2.
dissociative chemisorption of O2 to complete the catalytic cycle
(Step 6), which, as mentioned above, is irreversible and assumed
to be a kinetically-relevant step.
These proposed elementary steps in Scheme 2, together with
pseudo-steady state assumptions for all chemisorbed species and
quasi-equilibrium assumptions for Steps 1, 4, and 5, led to the der-
ivation of the HMF oxidation rate equation (Eq. (1)) by referring to
the method reported in the literature [21,52,53]. The derivation
details of this equation are included in the Supplementary Content.
Fig. 10. Parity plot for measured HMF oxidation activities (reaction conditions:
383 K, 0.1–4.0 MPa O2, 0.10–1.5 mmol HMF, 10 mg OMS-2, 10 mL DMF, ꢂ30% HMF
conversion) and calculated activities from Eq. (6).
ð1Þ
In Eq. (1), each term in the denominator represents the surface
coverage of the indicated adsorbed species relative to that of the
free sites on the OMS-2 surfaces. k6 is the rate coefficient for Step
kinetic dependence on HMF, O2 and H2O concentrations and the
isotopic data. Therefore, Eq. (1) can be further simplified to Eq. (5).
2k6 ꢁ ½O2ꢅ
6 while K and K represent the equilibrium constants for DFF
ꢃ4
ꢃ5
r ¼
ꢀ
ꢁ
2
1=4
1=2
3=4
1 þ M1 ꢁ ð½O2ꢅÞ ꢁ ð½R—CH2OHꢅÞ þ M3 ꢁ ð½O2ꢅÞ ꢁ ð½R—CH2OHꢅÞꢃ1=2
and H2O adsorption, respectively. M1, M2 and M3 are given by
the combination of the rate and equilibrium constants for the ele-
mentary steps, as shown below (Eqs. (2)–(4)).
ð5Þ
The parameters in Eq. (5) were derived from the nonlinear
regression analysis of the measured HMF oxidation activities
(Figs. 4 and 5), as shown in Eq. (6) and Table 7.
M1 ¼ ð2k6 ꢁ k3Þ ꢁ ðK1 ꢁ kꢃ1Þ1=2
ð2Þ
1=4
2
M2 ¼ ð2k6 ꢁ kꢃ1Þ1=2
ð3Þ
ð4Þ
0:590 ꢁ ½O2ꢅ
3
r ¼
ꢀ
ꢁ
2
1=4
1=2
3=4
1 þ 1:49 ꢁ ½O2ꢅ ꢁ ½R—CH2OHꢅ þ 0:0508 ꢁ ½O2ꢅ ꢁ ½R—CH2OHꢅꢃ1=2
M3 ¼ ðK1 ꢁ k2Þꢃ1=2 ꢁ k3ꢃ1=4 ꢁ ð2k6Þ
The surface coverages of the OHꢄ, R–CHOꢄ and H2Oꢄ species are
likely much smaller than those of R–CH2Oꢄ, Oꢄ and free Mn surface
sites (ꢄ) under our reaction conditions, according to the observed
3=4
ð6Þ
As shown in Fig. 10, the predicted HMF oxidation rates from Eq. (6)
are found to agree well with the measured rates (Figs. 4 and 5) in a
wide range of HMF concentrations (10–150 mmol/L) and O2 pres-
sures (0.1–4.0 MPa), demonstrating that Eq. (6) can accurately
describe the experimental kinetic results. Alternate elementary
steps of the HMF oxidation were also considered; for example, the
recombination of OHꢄ (Step 3) or the dissociative chemisorption
of O2 (Step 6) was assumed to be reversible. But they did not lead
to rate expressions in agreement with the experimental activities
or led to discrepancy with the isotopic data. The excellent agree-
ment between these data in Fig. 10 indicates that the proposed ele-
Table 7
Kinetic parameters in HMF oxidation obtained from nonlinear regression analysis of
Eq. (5) with kinetic data measured on OMS-2 at 383 K.
Entry
M1
M3
k6
k2
((mol/L)ꢃ1/2 atmꢃ1/4
)
((mol/L)1/2 atmꢃ3/4
)
(hꢃ1 atmꢃ1
)
(hꢃ1
)
1
1.49
5.08 ꢀ 10ꢃ2
2.95 ꢀ 10ꢃ1
7.76

66
J. Nie, H. Liu / Journal of Catalysis 316 (2014) 57–66
[8] C. Moreau, M.N. Belgacem, A. Gandini, Top. Catal. 27 (2004) 11.
[9] A.S. Amarasekara, D. Green, L.D. Williams, Eur. Polym. J. 45 (2009) 595.
[10] J. Ma, M. Wang, Z. Du, C. Chen, J. Gao, J. Xu, Polym. Chem. 3 (2012) 2346.
[11] J. Ma, Z. Du, J. Xu, Q. Chu, Y. Pang, ChemSusChem 4 (2011) 51.
[12] W. Partenheimer, Vladimir V. Grushin, Adv. Synth. Catal. 343 (2001) 102.
[13] C. Moreau, R. Durand, C. Pourcheron, D. Tichit, Stud. Surf. Sci. Catal. 108 (1997)
399.
[14] C. Carlini, P. Patrono, A.M.R. Galletti, G. Sbrana, V. Zima, Appl. Catal. A: Gen.
289 (2005) 197.
[15] O. Navarro, A. Canós, S. Chornet, Top. Catal. 52 (2009) 304.
[16] J. Nie, H. Liu, Pure. Appl. Chem. 84 (2012) 765.
mentary steps in Scheme 2 reflect the reaction mechanism of the
aerobic HMF oxidation to DFF on OMS-2.
4. Conclusions
OMS-2 is an efficient and stable catalyst in the selective aerobic
oxidation of HMF to DFF, affording a DFF yield of as high as 97.2% at
383 K and 0.5 MPa O2 in DMF. It exhibits both high activity and DFF
selectivity, compared to other MnO2 catalysts with different struc-
[17] I. Sádaba, Y.Y. Gorbanev, S. Kegnæs, S.S.R. Putluru, R.W. Berg, A. Riisager,
ChemCatChem 5 (2013) 284.
[18] F.L. Grasset, B. Katryniok, S. Paul, V. Nardello-Rataj, M. Pera-Titus, J.-M. Clacens,
F. De Campo, F. Dumeignil, RSC Adv. 3 (2013) 9942.
tures, including amorphous MnO2, Na-OL-1, OMS-7, c-MnO2, OMS-
6, and OMS-1, most likely as a result of its (2 ꢀ 2) tunnel structure
and consequently its high reducibility and oxidizability. The corre-
lation between the structures and redox properties of the MnO2
catalysts and their HMF oxidation activities, combined with the
similar initial rates of anaerobic and aerobic HMF oxidation on
OMS-2, suggests that the HMF oxidation to DFF proceeds via a
Mars-van Krevelen redox mechanism involving Mn4+/Mn3+ redox
cycles, and the high oxidation activity requires for both high reduc-
ibility and oxidizability of the MnO2 catalysts. The near half-order
dependence of the HMF oxidation activity on the HMF and O2 con-
centrations, together with the measured kinetic isotopic data, indi-
cates that the C–H cleavage of the adsorbed alcoholate
intermediate using lattice oxygen and the reoxidation of Mn3+ to
Mn4+ by dissociative chemisorption of O2 are kinetically-relevant
steps in the HMF oxidation on OMS-2. This mechanism is well con-
firmed by the excellent agreement between the measured HMF
oxidation activities and predicted results by this mechanism-
derived kinetic rate expression in a wide range of HMF and O2 con-
centrations. Such insights into the structural requirements and
reaction mechanism provide a rationale for the design of more effi-
cient MnO2 catalysts for the HMF oxidation to DFF and other valu-
able products.
[19] M. Lilga, R. Hallen, M. Gray, Top. Catal. 53 (2010) 1264.
[20] A. Takagaki, M. Takahashi, S. Nishimura, K. Ebitani, ACS Catal.
1 (2011)
1562.
[21] J. Nie, J. Xie, H. Liu, J. Catal. 301 (2013) 83.
[22] C.A. Antonyraj, J. Jeong, B. Kim, S. Shin, S. Kim, K.-Y. Lee, J.K. Cho, J. Ind. Eng.
Chem. 19 (2013) 1056.
[23] R.N. DeGuzman, Y.-F. Shen, E.J. Neth, S.L. Suib, C.-L. O’Young, S. Levine,
J.M. Newsam, Chem. Mater. 6 (1994) 815.
[24] V.P. Santos, O.S.G.P. Soares, J.J.W. Bakker, M.F.R. Pereira, J.J.M. Órfão, J. Gascon,
F. Kapteijn, J.L. Figueiredo, J. Catal. 293 (2012) 165.
[25] J. Luo, Q. Zhang, J. Garcia-Martinez, S.L. Suib, J. Am. Chem. Soc. 130 (2008)
3198.
[26] V.D. Makwana, Y.-C. Son, A.R. Howell, S.L. Suib, J. Catal. 210 (2002) 46.
[27] Z.-Z. Yang, J. Deng, T. Pan, Q.-X. Guo, Y. Fu, Green Chem. 14 (2012) 2986.
[28] G.D. Yadav, R.V. Sharma, Appl. Catal. B: Environ. 147 (2014) 293.
[29] X.F. Shen, Y.S. Ding, J. Liu, J. Cai, K. Laubernds, R.P. Zerger, A. Vasiliev,
M. Aindow, S.L. Suib, Adv. Mater. 17 (2005) 805.
[30] X. Wang, Y. Li, Chem. Eur. J. 9 (2003) 300.
[31] Z.-R. Tian, Y.-G. Yin, S.L. Suib, C. O’Young, Chem. Mater. 9 (1997) 1126.
[32] Q. Feng, H. Kanoh, Y. Miyai, K. Ooi, Chem. Mater. 7 (1995) 1226.
[33] I. Atribak, A. Bueno-Lopez, A. García-García, P. Navarro, D. Frías, M. Montes,
Appl. Catal. B: Environ. 93 (2010) 267.
[34] F. Cheng, J. Chen, X. Gou, P. Shen, Adv. Mater. 17 (2005) 2753.
[35] H. Huang, C.-H. Chen, L. Xu, H. Genuino, J. Garcia-Martinez, H.F. Garces, L. Jin,
C.K.o. Kithongo, S.L. Suib, Chem. Commun. 46 (2010) 5945.
[36] Z.-M. Wang, S. Tezuka, H. Kanoh, Chem. Mater. 13 (2001) 530.
[37] C. Wang, L. Sun, Q. Cao, B. Hu, Z. Huang, X. Tang, Appl. Catal. B: Environ. 101
(2011) 598.
[38] K. Frey, V. Iablokov, G. Sáfrán, J. Osán, I. Sajó, R. Szukiewicz, S. Chenakin, N.
Kruse, J. Catal. 287 (2012) 30.
Acknowledgments
[39] V. Galakhov, M. Demeter, S. Bartkowski, M. Neumann, N. Ovechkina, E.
Kurmaev, N. Lobachevskaya, Y.M. Mukovskii, J. Mitchell, D. Ederer, Phys. Rev. B
65 (2002) 1131021.
[40] T. Chen, H. Dou, X. Li, X. Tang, J. Li, J. Hao, Microporous Mesoporous Mater. 122
(2009) 270.
[41] K. Chen, A.T. Bell, E. Iglesia, J. Catal. 209 (2002) 35.
[42] K. Chen, S. Xie, A.T. Bell, E. Iglesia, J. Catal. 198 (2001) 232.
[43] H. Liu, E. Iglesia, J. Phys. Chem. B 109 (2005) 2155.
[44] S. Yamamoto, O. Matsuoka, I. Fukada, Y. Ashida, T. Honda, N. Yamamoto,
J. Catal. 159 (1996) 401.
This work was supported by the National Natural Science Foun-
dation of China (Grants 21373019, 21173008 and 51121091) and
National Basic Research Project of China (Grants 2011CB201400
and 2011CB808700).
Appendix A. Supplementary material
[45] S. Liang, F. Teng, G. Bulgan, R. Zong, Y. Zhu, J. Phys. Chem. C 112 (2008)
5307.
[46] R. Radhakrishnan, S.T. Oyama, J.G. Chen, K. Asakura, J. Phys. Chem. B 105
(2001) 4245.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2014.05.003.
[47] F. Kapteijn, L. Singoredjo, A. Andreini, J. Moulijn, Appl. Catal. B: Environ. 3
(1994) 173.
References
[48] K. Yamaguchi, H. Kobayashi, Y. Wang, T. Oishi, Y. Ogasawara, N. Mizuno, Catal.
Sci. Technol. 3 (2013) 318.
[49] K. Yamaguchi, N. Mizuno, Chem. Eur. J. 9 (2003) 4353.
[50] V. Santos, M. Pereira, J. Órfão, J. Figueiredo, Appl. Catal. B: Environ. 99 (2010)
353.
[51] M. Abecassis-Wolfovich, R. Jothiramalingam, M. Landau, M. Herskowitz, B.
Viswanathan, T. Varadarajan, Appl. Catal. B: Environ. 59 (2005) 91.
[52] D.K. Kim, E. Iglesia, J. Phys. Chem. C 112 (2008) 17235.
[53] M. Ojeda, B.-Z. Zhan, E. Iglesia, J. Catal. 285 (2012) 92.
[1] Y. Román-Leshkov, J.N. Chheda, J.A. Dumesic, Science 312 (2006) 1933.
[2] H. Zhao, J.E. Holladay, H. Brown, Z.C. Zhang, Science 316 (2007) 1597.
[3] S. Zhao, M. Cheng, J. Li, J. Tian, X. Wang, Chem. Commun. 47 (2011) 2176.
[4] T. Ståhlberg, W. Fu, J.M. Woodley, A. Riisager, ChemSusChem 4 (2011) 451.
[5] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411.
[6] 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. 41 (1998) 3872.
[7] M. Del Poeta, W.A. Schell, C.C. Dykstra, S. Jones, R.R. Tidwell, A. Czarny, M. Bajic,
M. Bajic, A. Kumar, D. Boykin, J.R. Perfect, Antimicrob. Agents Chemother. 42
(1998) 2495.