Aerobic oxidation of hydroxymethylfurfural and furfural by using heterogeneous CoxOy-N@C catalysts

Article abstract of DOI:10.1002/cssc.201402843

2,5-Furandicarboxylic acid (FDCA) is considered to be a promising replacement for terephthalic acid since they share similar structures and properties. In contrast to FDCA, 2,5-furandicarboxylic acid methyl (FDCAM) has properties that allow it to be easily purified. In this work, we reported an oxidative esterification of 5-hydroxymethylfurfural (HMF) and furfural to prepare corresponding esters over Co_xO_y-N@C catalysts using O₂ as benign oxidant. High yield and selectivity of FDCAM and methyl 2-furoate were obtained under optimized conditions. Factors which influenced the product distribution were examined thoroughly. The Co_xO_y-N@C catalysts were recycled five times and no significant loss of activity was detected. Characterization of the catalysts could explain such phenomena. Using XPS and TGA, we made a thorough investigation of the effects of ligand and pyrolysis temperature on catalyst activity.

Full text of DOI:10.1002/cssc.201402843

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201402843
Aerobic Oxidation of Hydroxymethylfurfural and Furfural
by Using Heterogeneous CoxOy–N@C Catalysts
Jin Deng,[a] Hai-Jie Song,[a] Min-Shu Cui,[a, b] Yi-Ping Du,[a] and Yao Fu*[a]
2,5-Furandicarboxylic acid (FDCA) is considered to be a promis-
ing replacement for terephthalic acid since they share similar
structures and properties. In contrast to FDCA, 2,5-furandicar-
boxylic acid methyl (FDCAM) has properties that allow it to be
easily purified. In this work, we reported an oxidative esterifica-
tion of 5-hydroxymethylfurfural (HMF) and furfural to prepare
corresponding esters over CoxOy–N@C catalysts using O2 as
benign oxidant. High yield and selectivity of FDCAM and
methyl 2-furoate were obtained under optimized conditions.
Factors which influenced the product distribution were exam-
ined thoroughly. The CoxOy–N@C catalysts were recycled five
times and no significant loss of activity was detected. Charac-
terization of the catalysts could explain such phenomena.
Using XPS and TGA, we made a thorough investigation of the
effects of ligand and pyrolysis temperature on catalyst activity.
Introduction
With the diminishing reserves of fossil resources and the in-
creasing demand for petroleum-based chemicals, utilization of
renewable replacements for petroleum-derived products has
grown in popularity. Of particular appeal is biomass, which can
be used as an alternative carbon source and is readily available
worldwide.[1] 5-Hydroxymethylfurfural (HMF) can be obtained
from biomass-based carbohydrates, and has the potential to
be upgraded to several valuable compounds.[2] Via oxidation,
HMF can be transformed to 2,5-furandicarboxylic acid (FDCA).
FDCA is considered to be a promising replacement for tereph-
thalic acid (TPA), which is used in large quantities for polymers
and fine chemicals production.[3]
However, FDCA is a solid which has large polarity, no precise
melting point or boiling point, and low solubility in most sol-
vents. Conventional purification methods such as vacuum dis-
tillation and recrystallization have no feasibility on FDCA.
Hence, regarding FDCA purification, there is still a dearth of
straightforward and eco-friendly methods. A possible way to
overcome this inconvenience is to produce the corresponding
ester, 2,5-furandicarboxylicacid dimethyl ester (FDCDM), which
can be easily purified through vacuum distillation and trans-
formed to FDCA through a simple hydrolysis reaction. More-
over, instead of FDCA, FDCDM can be used directly to synthe-
sis polymers through transesterification reaction. Christensen
et al.[10] reported a method for HMF oxidative esterification
using Au/TiO2 catalyst in MeOH with an addition of MeONa
base. They obtained 60% yield of FDCDM under 4 bar O2 at
1308C for 3 h. Furthermore, Corma et al.[11] reported the con-
version of HMF to FDCDM over Au/CeO2 catalyst. The reaction
was also performed in methanol and a yield of 99 mol% was
achieved. Taking into consideration cost and ease of purifica-
tion, a noble-metal-free approach for the oxidative esterifica-
tion of HMF to FDCDM is worth exploring.
Recently, Beller et al.[12] have applied pyrolyzed molecularly
defined complexes in catalytic oxidation and reduction of or-
ganic chemicals. Of most attractive is the application of cobalt-
based catalysts in the direct conversion of benzylic alcohols to
the corresponding methyl ester in methanol.[12b] As reported,
the whole procedure consists of the following steps: first is the
oxidation of alcohol to aldehyde; then in methanol, aldehyde
converts to the corresponding hemiacetal; finally, via dehydro-
genation, hemiacetal transforms to ester. Under 0.1 MPa
oxygen at 808C for 24 h, the esterification of various benzylic
alcohols and heterocyclic alcohols was performed over the
cobalt catalysts in a good yield (85–97%). Furthermore, die-
sters and triesters were obtained directly with a yield up to
91%.
Various methods of heterogeneous catalysis have been stud-
ied for the oxidation of HMF to FDCA. Typically, supported
noble metals such as Pd,[4] Pt,[5] Au,[6] Ru[7] and bimetallic cata-
lysts[8] were utilized. Despite the increasing utilization of these
metals, previous work on catalytic oxidation revealed a relative-
ly wide variation in performance. Yields of FDCA ranging from
48% to nearly quantitative were achieved in different reaction
systems. Considering the high cost of catalysts, a base metal,
cobalt, was used to convert fructose into FDCA.[9] With a yield
of 71%, FDCA was converted directly from fructose using
Co(acac)3 encapsulated in a sol–gel silica matrix as catalysts.
[a] Dr. J. Deng, H.-J. Song, M.-S. Cui, Y.-P. Du, Prof. Dr. Y. Fu
Anhui Province Key laboratory of Biomass Clean Energy
Department of Chemistry
University of Science and Technology of China
Hefei 230026 (PR China)
Fax: (+86)-551-6360-6689
fuyao@ustc,edu.cn
[b] M.-S. Cui
Nano Science and Technology Institute
University of Science and Technology of China
Suzhou 215123 (PR China)
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201402843.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
1
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Herein, we attempted to apply pyrolyzed CoxOy–N@C cata-
lysts to the oxidative esterification of biomass-derived furans
including furfural and HMF. As expected, good results were ob-
tained on both chemicals. For furfural, full conversion and
a 95% yield of methyl 2-furoate (MF) was obtained under
0.1 MPa O2 at 608C for 12 h. Furthermore, we achieved a 100%
conversion of HMF and a 96% yield of target products at
1008C under 1 MPa O2 for 6 h requiring an addition of K-OMS-
2. Both reactions were performed in methanol over heteroge-
neous CoxOy–N@C catalysts with the aid of 0.2 equivalent
amounts of K2CO3.
diPy) as ligand. This result is consistent with the work of Beller
et al.[12d]
From the view of structure, difference between the two li-
gands is reflected in the connection of pyridines. In 2,2-diPy,
a s-bond acts as bridge, while in 1,10-phen a benzene that
acts as bridge. As a result, 2,2-diPy shows more flexibility
which contributes to a difficulty in the formation of graphene-
type layers by nitrogen ligand during the pyrolysis process.
Previous work[13] concluded that in metal–N@C catalysts, active
site mainly consists of pyridine-type nitrogen and graphene-
type nitrogen. Therefore a lack of graphene-type nitrogen
leads to a low catalytic activity. XPS data is shown in Figure 1.
In the N1s spectra, distinct peaks with electron-binding ener-
gies of 403.6 eV and 400.8 eV have been observed and can be
attributed to graphene-type nitrogen and pyrrole-type nitro-
gen, respectively. Peaks with binding energies of 399.6 eV and
398.4 eV represent pyridine-type nitrogen (bond with cobalt)
and nitrogen of ligand (pyridine-type nitrogen) in CoxOy-2,2-
diPy/C respectively.[14] And 399.7 eV and 398.6 eV are character-
ized as pyridine-type nitrogen and as nitrogen of ligand in
CoxOy–1,10-phen/C respectively.[14] From the comparison of
CoxOy-2,2-diPy/C with CoxOy-1,10-phen/C in Figure 1a), no indi-
cation of graphene-type nitrogen and a lower amount of pyri-
dine-type nitrogen have been detected in CoxOy–N@C using
2,2-diPy as ligand. In addition, the chelate of Co(OAc)2·4H2O
and 2,2-diPy has a lower activation energy which contributes
to easier cleavage and gasification (TGA data shown in Fig-
ure S8, Supporting Information). The loss of acetate reduces
the amount of oxygen in the catalyst. As a result, under the
same pyrolysis temperature, Co-2,2-diPy has less potential to
form cobalt oxides, resulting in a low activity of Co-2,2-diPy
catalysts. XPS data of Co2p3/2 for support of the conclusions is
shown in Figure 1b. Peaks with binding energy of 785.6 eV to
786.6 eV, 781.0 eV, 779.3 eV to 780.0 eV, and 778.5 eV can be
attributed to a satellite peak of cobalt, bonding of cobalt and
nitrogen, bonding of cobalt and oxygen and Co0 respective-
ly.[15] Obviously, the content of oxidized cobalt is higher in the
Results and Discussion
To study the feasibility of the cobalt-based catalysts for the cat-
alysis of biomass-derived compounds, we tested initially the
catalytic esterification of furfural in CH3OH. Gratifyingly, the
CoxOy–N@C catalysts performed well in the reaction. Later, to
investigate the optimum catalytic efficiency, exploratory experi-
ments with different metals and ligands were performed under
the same reaction conditions (Table 1, entries 1–3). For a 12 h
reaction time, under 0.1 MPa O2 at 608C, Co with ligand 1,10-
phenanthroline (1,10-phen) showed the best activity in com-
parison with Pd(1,10-phen) and Co with 2,2-dipyridine (2,2-
Table 1. Oxidative esterification of furfural to MF.[a]
Entry
Metal[b]
Ligand
Conversion[c] [%]
Yield[c] [%]
1
2
3
4[d]
5[d]
6[d]
Co
Pd
Co
Co
Mn
Pd
1,10-phen
1,10-phen
2,2-diPy
1,10-phen
1,10-phen
1,10-phen
100
12
58
54
9
95
4
46
42
-
15
1
[a] Conditions: 0.1 MPa O2, 608C, 12 h, 0.2 equiv K2CO3. [b] Metal loading:
3%; 25 mg catalyst. [c] Determined by GC. [d] Reactant: furfuryl alcohol.
Figure 1. XPS of the a) N1s and b) Co2p3/2 for Co with ligands diPy and 1,10-phen.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
2
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
pyrolyzed CoxOy–N@C using ligand 1,10-phen with a certain
amount of Co0 present existence.
the aim of improving the conversion of HMFM, according to
our previous work,[16] we introduced K-OMS-2, an efficient het-
erogeneous catalyst for HMF oxidation to DFF, into the HMF
oxidative esterification reaction system. To our delight, the
yield of HMFM declined to 2% under the optimal conditions.
Meanwhile, we obtained the highest yield of FDCAM of 95%.
To study the influence of reaction conditions on the distribu-
tion of HMF oxidation products, a comprehensive investigation
was performed. Variables including catalyst pyrolysis tempera-
ture, Co loading, amount of catalysts, alkaline intensity, reac-
tion time, oxygen pressure, and reaction temperature were ex-
amined thoroughly.
A similar result was observed in the oxidative esterification
of furfuryl alcohol (FA) over the carbon-supported catalysts
(Table 1, entries 4–6). Co(1,10-phen) catalysts showed the best
performance while Mn(1,10-phen) did not show any activity
toward ester formation, only a 5% yield of furfural was ob-
tained.
As expect, the cobalt-based carbon-supported catalysts were
effective for the oxidation of HMF in methanol as well
(Table 2). Under 0.1 MPa O2 at 608C for 12 h, we obtained
a 95% conversion of HMF and obtained FDCDM, 2,5-furandi-
carboxylic acid monomethyl ester (FDCMM), FDCA, and 5-hy-
droxymethyl-2-furoic acid methyl ester (HMFM) in 28%, 30%,
Catalyst activity
In order to explore how pyrolysis
temperature affected the cata-
Table 2. Products distribution of HMF oxidation.[a]
lyst activity, we tested tempera-
Entry Catalyst 1[b]
Catalyst 2[b]
T
[8C]
P
t [h] Conversion[d] [%]
Yield[d] [%]
tures ranging from 2008C to
HMFM FDCDM FDCMM FDCA FDCAM
10008C with a gradient of 200
1
2
3[c]
4
5
6
CoxOy–N@C-800
CoxOy–N@C-800
CoxOy–N@C-800
CoxOy–N@C-800 K-OMS-2
CoxOy–N@C-800 K-OMS-2
CoxOy–N@C-800 K-OMS-2
–
–
–
60 0.1 12
100 12
60 0.1 12
60 0.1 12
95
96
98
98
36
12
1
24
1
28
38
52
39
48
53
30
44
45
35
45
41
1
1
0
0
1
2
59
83
97
75
95
96
(Figure 2). All the experiments
were performed under 1 MPa O2
at 808C for 12 h over 25 mg
3wt% CoxOy–N@C with an equal
amount of K-OMS-2. As shown
in the figure, catalysts pyrolyzed
at 2008C exhibited little activity
toward FDCAM formation, even
though a 95% conversion of
1
100
100
1
1
12 >99
99
6
1
[a] Conditions: 0.2 equiv K2CO3. [b] Co loading: 3%; 25 mg catalysts. [c] Reactant: DFF. [d] Determined by HPLC.
A typical spectrum was shown in Figure S3.
1%, and 36% yield, respectively. Given the situation that the
major product was HMFM, a change of reaction condition was
made in an effort to improve the yield of 2,5-furandicarboxylic
acid methyl (FDCAM, which corresponds to the sum of
FDCDM, FDCMM, and FDCA) . With an increase in O2 pressure
and reaction temperature, the yield of FDCDM had been great-
ly improved while a decrease of HMFM was clearly detected.
The optimized conditions, 1 MPa O2 and 1008C, led to FDCDM,
FDCMM, FDCA, and HMFM in 38%, 44%, 1%, and 12% yield
respectively. Taking the result of the oxidative reaction of FA
into account, we deduced that the rate-determining step was
the alcohol oxidation to aldehyde. Typically, there are two
pathways for HMF to FDCDM (Scheme 1). One is through alde-
hyde oxidation to HMFM (path A), the other is through alcohol
oxidation to 2,5-diformylfuran (DFF, path B). Under the same
mild reaction conditions with furfural, we examined the cata-
lytic efficiency for DFF. The yield of FDCDM, FDCMM, FDCA,
and HMFM was 53%, 46%, 0 and 1%, respectively. Hence, with
HMF was obtained. The main product was DFF categorized in
the series “others”. Along with the increase of pyrolysis tem-
Figure 2. Effects of pyrolysis temperature on catalyst activity.
perature, catalytic activity was
significantly
improved.
The
carbon-supported catalysts that
were pyrolyzed at 8008C led to
the best performance of 93%
yield of FDCAM. For 10008C py-
rolysis temperature, the activity
declined slightly. Thus, 8008C
Scheme 1. Possible pathways of HMF transformation.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
3
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 3. a) TGA data of nonpyrolyzed CoxOy–N@C; b) XPS of the N1s and c) Co2p3/2 for CoxOy–N@C prepared at different pyrolysis temperature.
was chosen to be the optimal pyrolysis temperature in the fol-
lowing study.
of pyrrole-type nitrogen and graphene-type nitrogen as well as
the acceleration of pyridine-type nitrogen and cobalt oxides.
From 8008C, the appearance of Co0 is detected due to the re-
ducibility of carbon. At a pyrolysis temperature of 10008C, as
shown in the spectra, interaction between Co and 1,10-phen
no longer exists and the whole catalyst begins to decompose.
Thus, the catalyst with the highest activity can be obtained
through a pyrolysis process at 8008C. Additionally, changes in
satellite peaks of Co illustrate that, in CoxOy–N@C catalysts, the
amounts of Co2+ are variable.
To elucidate the effect of pyrolysis temperature on catalytic
activity, we performed detailed investigation using thermo
gravimetric analysis (TGA) and X-ray photoelectron spectrosco-
py (XPS). Result of TGA tests for nonpyrolyzed CoxOy–N@C cat-
alysts are shown in Figure 3a). The Figure reveals the following
information:[17] 1) Below 1008C, the loss of weight is mainly
due to the loss of crystal water. 2) From 1008C to 2008C, ace-
tate begins to change. Therefore, catalysts pyrolyzed at 2008C
show little activity on HMF esterification. 3) Over 2008C, con-
tinued rise of temperature brings about a cleavage of Co(1,10-
phen) to free 1,10-phen and cobalt, leading to a formation of
cobalt oxides from the free cobalt and oxygen. In this way, the
catalysts adopt the active state. This is consistent with the ex-
perimental results. A significant difference of FDCAM yield was
obtained between pyrolysis temperature 2008C and 4008C.
4) In the range of 4008C to 8008C, further conversion of 1,10-
phen results in the formation of pyrrole-type nitrogen and gra-
phene-type layers. Thus, the most active catalysts can be ob-
tained. In the experiment data, catalysts pyrolyzed at 8008C
showed the best performance. 5) More than 8008C, CoxOy–
N@C catalysts begin to decompose.
Characterization by energy-dispersive X-ray spectroscopy el-
emental mapping (EDS) also confirms the existence of cobalt
oxides. As shown in Figure S7, Supporting Information, in the
combined image, the area in which cobalt is present is also
abundant in oxygen while the amount of carbon is low.
Similar results can be summed up from previous works.[18]
Generally, in inert atmosphere, Co–N@C catalysts pyrolyzed at
6008C to 8008C has the best catalytic activity. In this range,
the metal–nitrogen chelating structures as catalytic active sites
are uniformly dispersed on the carbon support. Above 8008C,
metal–nitrogen chelating bond would break, resulting in a de-
cline of catalytic activity. Two possible explanations for this
phenomenon have been raised as follows: 1) high tempera-
tures favor the generation of inert particles of metals, metal
oxides, or metal carbides;[19] 2) as the metal–nitrogen bond
breaks, the density of active sites distributing on the surface of
catalyst also decreases.[20] Another view point states that, of
the two active sites, pyridine-type nitrogen and graphene-type
nitrogen, the former one is more active. With the increase of
pyrolysis temperature, amounts of graphene-type nitrogen will
increase with a drop in the amount of pyridine-type nitro-
gen.[13] Thus, above the optimal pyrolysis temperature, the cat-
alytic activity will decrease.
Corresponding results analyzed by XPS are shown in Fig-
ure 3b) and c). In the N1s spectra, four distinct peaks with
electron binding energy of 403.6 eV, 400.8 eV, 399.7 eV, and
398.6 eV were detected and can be attributed to graphene-
type nitrogen, pyrrole-type nitrogen, pyridine-type nitrogen
(bonded with cobalt), and nitrogen of 1,10-phen (pyridine-type
nitrogen), respectively.[14] For the Co2p3/2 spectra, peaks with
electron binding energy of 785.6 eV to 786.6 eV, 781.0 eV,
779.3–780.0 eV, and 778.5 eV can be attributed to Co satellite,
Co–N, Co-O, and Co0.[15] Below 2008C, only the peak of
Co(1,10-phen) is observed. There is an emergence of pyridine-
type nitrogen and Co-O at pyrolysis temperatures of over
4008C. Along with the increase of temperature, from 6008C,
sustaining cleavage of Co(1,10-phen) results in the formation
Furthermore, reactions were also conducted to probe the ef-
fects of Co loading and catalyst amount on the catalytic effi-
ciency. Typically, we used 3 wt% Co with 1,10-phen ligand ad-
sorbed on active carbon and pyrolyzed at 8008C for 2 h. All re-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
4
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
actions were performed over 25 mg CoxOy–N@C catalysts and
25 mgK-OMS-2. Both decreasing the Co loading to 1% or in-
creasing it to 5% led to a drop of FDCAM yield. The yield de-
clined from 96% to 43% and 87% respectively. Similarly, re-
ducing the amount of catalysts to 10 mg afforded a decline of
FDCAM from 96% to 82%. Similar results of metal loading
effect have been discussed in previous literature.[21] Over-load-
ing metal may lead to the formation of particles with no activi-
ty. And the inactive metals, metal oxides, and metal carbides
hinder the reactants entering the pores of carbon support.
Hence, no effective interaction between reactants and active
sites leads to a loss of catalytic activity.[22] Brunauer–Emmett–
Teller (BET) data of 3% and 5% CoxOy–N@C is listed in ESI. Re-
sults show a significant loss of BET surface area and pore
volume with an increase of average pore diameter when Co
loading increases from 3% to 5%.
Figure 5. Product distribution under different reaction temperatures.
FDCMM, and less than 1% FDCA. In the range of 1008C to
1208C, yield of FDCAM fell to 93%. This might be due to a fur-
ther decomposition of FDCMM and FDCDM.
Effects of reaction conditions
In the oxidative esterification of furfural, a 12 h reaction time
afforded the best result. Because we increased the reaction
pressure and temperature, a series experiments were per-
formed to verify the effects on reaction rate acceleration. Reac-
tion time 1 h, 2 h, 4 h, 6 h, and 12 h were tested separately
under 1 MPa O2 at 1008C (Figure 6). Gratifyingly, a reaction
time of 6 h achieved the best performance of 96% yield of
FDCAM.
As shown in Figure 4, we examined the influence of oxygen
pressure on the reaction yield. All the data were collected at
608C for a 12 h reaction time over 25 mg CoxOy–N@C and K-
OMS-2. Initially, as in the oxidation of furfural, reaction pressure
Figure 4. Product distribution under different reaction pressures.
was set as 0.1 MPa, and a 75% yield of FDCAM was obtained.
For the purpose of getting better results, oxygen pressure was
expanded to 1 MPa, 2 MPa, and 4 MPa separately. As expected,
the yield of FDCAM increased gradually when the pressure was
increased. Due to an increase of only 4% yield of FDCAM
when pressure was increased from 1 MPa to 4 MPa, we chose
1 MPa as the most cost-effective oxygen pressure.
Figure 6. Product distribution under different reaction times.
Effects of alkaline intensity
Despite all the factors discussed above, alkaline intensity was
another essential element in the oxidative esterification of
HMF (Table 3). Starting with K2CO3, salts with different alkaline
were tested under the optimal reaction conditions. Along with
the reduction of alkaline, a drop of FDCAM yield was clearly
detected. The possible intermediates of oxidative esterification
of HMF, aldehyde and hemiacetal, indicated an elimination of
the alpha-H of the hydroxyl. On the other hand, the presence
of alkali was conducive to the occurrence of dehydrogenation.
Studies on reaction temperature were conducted at 608C,
808C, 1008C, and 1208C under 1 MPa O2 for 12 h (Figure 5).
Within the range of 608C to 1008C, a continuous growth of
FDCAM yield and drop of FDCDM yield were observed concur-
rently. One reason for the growth was that the conversion of
DFF (categorized in the series “others”) to FDCAM improved
with increasing the reaction temperature. The drop was due to
the accelerated hydrolysis of FDCDM. The highest yield of 95%
was obtained at 1008C, composed of 50% FDCDM, 45%
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
5
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Table 3. Effects of alkaline intensity on the esterification of HMF.[a]
Entry Base[b]
Conversion[c] [%]
Yield[c] [%]
HMFM FDCDM FDCMM FDCA FDCAM
1
2
3
4
5
K2CO3
Na2CO3 >99
KHCO3
NaOAc
Na2HPO4
99
2
0
1
4
3
53
58
68
61
42
42
39
20
3
1
1
0
0
0
96
98
88
64
44
>99
91
70
2
[a] Conditions: 1 MPa O2, 1008C, 6 h, 25 mg CoxOy–N@C-800 and 25 mgK-
OMS-2. [b] 0.2 Equiv. [c] Determined by HPLC.
Effects on product distribution
Figure 7. Reusability of CoxOy–N@C catalysts.
Our original intention was to explore a method for direct con-
version of HMF to FDCDM, and then obtain FDCA through pu-
rification and hydrolysis. However, the target product was ap-
pearing in three forms because the FDCDM hydrolyzed. There
are two reasons for this: Both the alkaline reaction base and
the increase of reaction temperature accelerated the hydrolysis
process. As shown in Figure 3 and Table 3, with the increase of
alkaline and temperature, the proportion of FDCMM and FDCA
increased continuously while an apparent decline in the pro-
portion of FDCDM was observed.
Conclusions
We verified the feasibility of CoxOy–N@C catalysts applied to
the oxidative esterification of HMF and furfural. In addition, full
conversion of reactants and high yield of esters were achieved
simultaneously. Briefly, we presented a novel, cost-efficient,
eco-friendly and recyclable method for the production of
FDCAM. The catalyst with the highest activity was prepared
using homogeneous cobalt with 1,10-phen as ligand pyrolyzed
at 8008C for 2 h after adsorption on active carbon. The reason
can be well-explained by the measurements and data of XPS
and TGA. Under the optimal conditions, 1 MPa O2 at 1008C for
6 h, we obtained FDCAM in the yield of 96%. Reusability stud-
ies were performed with a view to industrial scale production.
To our delight, no significant loss of activity was detected up
to the fifth run.
Furthermore, considering the byproduct was water, we also
examined the effect of water on the product distribution
(Table 4). Separately, H2O of 3 vol%, 5 vol% and 10 vol% was
added into the reaction system. Under the optimal reaction
conditions, the yield of FDCAM was significantly affected as
shown in Table 4. A drop from 66% to 46% was determined
by HPLC as the volume of water increased from 3% to 10%.
When water was as solvent, no effective activity was detected
even though there was a 45% conversion of HMF.
Experimental Section
Table 4. Effects of water on product distribution.[a]
Materials
Entry Water Conversion[c] [%] Yield[c] [%]
[Vol%]
All the chemicals involved in the preparation of catalysts were
commercial available. Co(OAc)2·4H2O and EtOH from SCR, 1,10-
phen and 2,2-diPy by Alfa and active carbon from TCI were used
as purchased in the preparation of the CoxOy–N@C catalysts.
Materials used in reactions including HMF, FDCA, and FDCDM were
generous gifts from Hefei Leaf Energy Biotechnology Co., Ltd.
FDCMM was self-made; the detailed preparation process was
shown in the Supporting Information. And the left chemicals in-
cluding HMFM, DFF, methanol, K2CO3, Na2CO3, KHCO3, NaOAc and
Na2HPO4 were commercial available, reagent water was purchased
from Wahaha.
HMFM FDCDM FDCMM FDCA FDCAM
1
2
3
4[b]
3
5
10
100
92
90
84
45
4
9
8
0
33
30
19
0
31
31
26
0
1
1
1
1
66
62
46
1
[a] Conditions: 1 MPa O2, 1008C, 6 h, 25 mg CoxOy–N@C-800, and
25 mgK-OMS-2. [b] Reaction performed in water. [c] Determined by HPLC.
Reusability
In order to apply the catalysts in commercial-scale production,
a test of reusability was performed in five consecutive runs
(Figure 7). Under identical reaction conditions, results showed
a slight efficiency loss in the oxidation of HMF in five continu-
ous runs. A yield of 80% FDCAM was detected at the fifth run.
XRD pattern of CoxOy–N@C and K-OMS-2 as prepared, mixture
of CoxOy–N@C and K-OMS-2 after the fifth run were exhibited
in the Figure S5. No significant difference in structure of cata-
lysts between the two states was observed.
Catalyst preparation
Catalyst CoxOy–N@C was prepared as demonstrated in the pub-
lished literature.[12] Co(OAc)2·4H2O (3 mmol) and 1,10-phen
(6 mmol) with a molar ratio 1:2 was stirred in 300 mL ethanol for
30 min at room temperature. Then, 4.14 g active carbon was
added in for absorption of the products. The reaction system was
refluxed at 1108C for 4 h followed by rotary evaporation at 258C.
Obtained solid was dried at 608C for 12 h in vacuum, and finally
pyrolyzed in N2 (20 mLminÀ1) at 8008C (a rate of 108CminÀ1 from
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
6
&
ÞÞ
These are not the final page numbers!

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
vestre, B. Reis, J. Polym. Sci. Part A 2011, 49, 3759–3768; e) A. J. J. E. Eer-
hart, A. P. C. Faaij, M. K. Patel, Energy Environ. Sci. 2012, 5, 6407–6422;
f) J. Ma, X. Yu, J. Xu, Y. Pang, Polymer 2012, 53, 4145–4151; g) J. Ma, Y.
Pang, M. Wang, J. Xu, H. Ma, X. Nie, J. Mater. Chem. 2012, 22, 3457–
3481.
RT to 8008C) for 2 h. An coupled plasma (ICP) test showed the Co
loading in the catalysts was 3.36%.
Catalytic oxidation of HMF
[4] S. E. Davis, L. R. Houk, E. C. Tamargo, A. K. Datye, R. J. Davis, Catal. Today
2011, 160, 55–60.
All the oxidative esterification reactions was performed in the auto-
clave provided by Anhui Kemi Machinery Technology Co., Ltd. (Fig-
ure S9, Supporting Information). Typically, a mixture of 4 mL meth-
anol and 0.5 mmol HMF reacted with the aid of 25 mg CoxOy–N@C,
25 mgK-OMS-2 and 0.1 mmolK2CO3. O2 was used as the oxidant
with a pressure of 1 MPa. Then the reactor was closed, heated to
1008C and stirred for 6 h. Products were separated by Hitachi
L2000 HPLC System, Alltech C18 column at 308C at a wavelength
of 265 nm. The mobile phase was 30% methanol and 0.1% phos-
[5] a) M. A. Lilga, R. T. Hallen, M. Gray, Top Catal. 2010, 53, 1264–1269;
b) S. E. Davis, B. N. Zope, R. J. Davis, Green Chem. 2012, 14, 143.
[6] a) B. Saha, S. Dutta, M. M. Abu-Omar, Catal. Sci. Technol. 2012, 2, 79–81;
b) O. Casanova, S. Iborra, A. Corma, ChemSusChem 2009, 2, 1138–1144;
c) Y. Y. Gorbanev, S. K. Klitgaard, J. M. Woodley, C. H. Christensen, A. Riis-
ager, ChemSusChem 2009, 2, 672–675; d) J. Cai, H. Ma, J. Zhang, Q.
Song, Z. Du, Y. Hu, J. Xu, Chem. Eur. J. 2013, 19, 14215–14223; e) N. K.
Gupta, S. Nishimura, A. Takagaki, K. Ebitani, Green Chem. 2011, 13, 824;
f) G. Yi, S. P. Teong, X. Li, Y. Zhang, ChemSusChem 2014, 7, 2131–2135;
g) S. P. Teong, G. Yi, X. Cao, Y. Zhang, ChemSusChem 2014, 7, 2120–
2124.
phoric acid aqueous with a rate of 1 mLminÀ1
.
[7] a) T. Stꢁhlberg, E. Eyjꢂlfsdꢂttir, Y. Y. Gorbanev, I. Sꢃdaba, Catal. Lett.
2012, 142, 1089–1097; b) N. Lucas, N. R. Kanna, A. S. Nagpure, G.
Kokate, S. Chilukuri, J. Chem. Sci. 2014, 126, 403–413; c) Y. Y. Gorbanev,
S. Kegnaes, A. Riisager, Catal. Lett. 2011, 141, 1752–1760.
[8] a) A. Villa, M. Schiavoni, S. Campisi, G. M. Veith, L. Prati, Chemsuschem
2013, 6, 609–612; b) T. Pasini, M. Piccinini, M. Blosi, R. Bonelli, S. Albo-
netti, N. Dimitratos, J. A. Lopez-Sanchez, M. Sankar, Q. He, C. J. Kiely,
G. J. Hutchings, F. Cavani, Green Chem. 2011, 13, 2091–2099; c) H. Ai-
t Rass, N. Essayem, M. Besson, Green Chem. 2013, 15, 2240–2251.
[9] M. L. Ribeiro, U. Schuchardt, Catal. Commun. 2003, 4, 83–86.
[10] E. Taarning, I. S. Nielsen, K. Egeblad, R. Madsen, C. H. Christensen, Chem-
suschem 2008, 1, 75–78.
[11] O. Casanova, S. Iborra, A. Corma, J. Catal. 2009, 265, 109–116.
[12] a) R. V. Jagadeesh, G. Wienhçfer, F. A. Westerhaus, A. Surkus, M. Pohl, H.
Junge, K. Junge, M. Beller, Chem. Commun. 2011, 47, 10972–10974;
b) R. V. Jagadeesh, H. Junge, M. Pohl, J. Radnik, A. Brꢄckner, M. Beller, J.
Am. Chem. Soc. 2013, 135, 10776–10782; c) F. A. Westerhaus, R. V. Jaga-
deesh, G. Wienhçfer, M. Pohl, J. Radnik, A. Surkus, J. Rabeah, K. Junge,
H. Junge, M. Nielsen, A. Brꢄckner, M. Beller, Nat. Chem. 2013, 5, 537–
543; d) R. V. Jagadeesh, A. Surkus, H. Junge, M. Pohl, J. Radnik, J.
Rabeah, H. Huan, V. Schꢄnemann, A. Brꢄckner, M. Beller, Science 2013,
342, 1073–1076; e) D. Banerjee, R. V. Jagadeesh, K. Junge, M. Pohl, J.
Radnik, A. Brꢄckner, M. Beller, Angew. Chem. Int. Ed. 2014, 53, 4359–
4363; Angew. Chem. 2014, 126, 4448–4452.
Catalyst characterization
Catalysts were characterized by XRD, TGA, XPS, TEM, ICP, BET sur-
face area analysis, and energy-dispersive X-ray spectroscopy ele-
mental mapping (EDS). Relevant measurements and data are pro-
vided in part 3 of the Supporting Information.
Acknowledgements
This work was supported by the 973 Program (2012CB215305,
2013CB228103), NSFC (21402181, 21325208, 21172209,
21272050), CAS (KJCX2-EW-J02), FRFCU (WK2060190025,
WK2060190033), SRFDP (20123402130008) and China Postdoctor-
al Science Foundation (2014M561835). The authors thank the
Hefei Leaf Energy Biotechnology Co., Ltd. and Anhui Kemi Machi-
nery Technology Co., Ltd. for free sample sponsoring in this study.
Keywords: 2,5-furandicarboxylicacid dimethyl ester
hydroxymethylfurfural cobalt oxides heterogeneous
catalysis · oxidative esterification
·
5-
·
·
[13] G. Liu, X. Li, P. Ganesan, B. N. Popov, Electrochim. Acta 2010, 55, 2853–
2858.
[14] J. Casanovas, J. M. Ricart, J. Rubio, F. Illas, J. M. Jimꢅnez-Mateos, J. Am.
Chem. Soc. 1996, 118, 8071–8076.
[15] A. Morozan, P. Jꢅgou, B. Jousselme, S. Palacin, Phys. Chem. Chem. Phys.
2011, 13, 21600–21607.
[16] Z. Yang, J. Deng, T. Pan, Q. Guo, Y. Fu, Green Chem. 2012, 14, 2986–
2989.
[1] a) G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098;
b) J. N. Chheda, G. W. Huber, J. A. Dumesic, Angew. Chem. Int. Ed. 2007,
46, 7164–7183; Angew. Chem. 2007, 119, 7298–7318; c) D. R. Dodds,
R. A. Gross, Science 2007, 318, 1250–1251; d) A. Corma, S. Iborra, A.
Velty, Chem. Rev. 2007, 107, 2411–2502; e) J. C. Serrano-Ruiz, J. A. Du-
mesic, Energy Environ. Sci. 2011, 4, 83–99; f) Y. Huang, Y. Fu, Green
Chem. 2013, 15, 1095–1111; g) M. Besson, P. Gallezot, C. Pinel, Chem.
Rev. 2014, 114, 1827–1870.
[2] a) G. Yong, Y. Zhang, J. Y. Ying, Angew. Chem. Int. Ed. 2008, 47, 9345–
9348; Angew. Chem. 2008, 120, 9485–9488; b) J. B. Binder, R. T. Raines,
J. Am. Chem. Soc. 2009, 131, 1979–1985; c) X. Tong, Y. Ma, Y. Li, Appl.
Catal. A 2010, 385, 1–13; d) A. Corma, O. d. L. Torre, M. Renz, N. Villandi-
er, Angew. Chem. Int. Ed. 2011, 50, 2375–2378; Angew. Chem. 2011, 123,
2423–2426; e) L. Hu, G. Zhao, W. Hao, X. Tang, Y. Sun, L. Lin, S. Liu, RSC
Adv. 2012, 2, 11184–11206; f) R. Van Putten, J. C. V. D. Waal, E. D. Jong,
C. B. Rasrendra, H. J. Heeres, J. G. D. Vries, Chem. Rev. 2013, 113, 1499–
1597; g) V. Choudhary, S. H. Mushrif, C. Ho, A. Anderko, V. Nikolakis, N. S.
Marinkovic, A. I. Frenkel, S. I. Sandler, D. G. Vlachos, J. Am. Chem. Soc.
2013, 135, 3997–4006.
´
[17] D. Poleti, D. R. Stojakovic, Thermochim. Acta 1992, 205, 225–233.
[18] a) Y. Ma, H. Zhang, H. Zhong, T. Xu, H. Jin, Y. Tang, Z. Xu, Electrochim.
Acta 2010, 55, 7945–7950; b) Y. Ji, Z. Li, S. Wang, G. Xu, X. Yu, Int. J. Hy-
drogen Energy 2010, 35, 8117–8121.
[19] a) C. W. B. Bezerra, L. Zhang, K. Lee, Electrochim. Acta 2008, 53, 7703–
7710; b) S. Li, L. Zhang, H. Liu, Electrochim. Acta 2010, 55, 4403–4411.
[20] a) X. Li, G. Liu, B. N. Popov, J. Power Sources 2010, 195, 6373–6378; b) U.
Koslowski, I. Herrmann, P. Bogdanoff, ECS Trans. 2008, 13, 125–141.
[21] M. Bron, S. Fiechter, M. Hilgendorff, J. Appl. Electrochem. 2002, 32, 211–
216.
[22] a) L. Zhang, K. Lee, C. W. B. Bezerra, Electrochim. Acta 2009, 54, 6631–
6636; b) M. Lefꢆvre, J. P. Dodelet, P. Bertrand, J. Phys. Chem. 2000, 104,
11238–11247.
[3] a) A. Gandini, Macromolecules 2008, 41, 9491–9504; b) A. Gandini, A.
Silvestre, C. P. Neto, A. F. Sousa, M. Gomes, J. Polym. Sci. Part A 2009, 47,
295–298; c) A. Gandini, D. Coelho, M. Gomes, B. Reis, A. Silvestre, J.
Mater. Chem. 2009, 19, 8656–8664; d) M. Gomes, A. Gandini, A. J. D. Sil-
Received: August 17, 2014
Revised: August 28, 2014
Published online on && &&, 0000
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
7
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
J. Deng, H.-J. Song, M.-S. Cui, Y.-P. Du,
Y. Fu*
Happy Ester: The oxidative esterifica-
tion of 5-hydroxymethylfurfural (HMF)
and furfural over CoxOy–N@C catalysts is
performed using O2 as benign oxidant,
obtaining the corresponding esters.
High yield and selectivity of 2,5-furandi-
carboxylic acid methyl and methyl 2-fu-
roate are achieved under optimized
conditions.
&& – &&
Aerobic Oxidation of
Hydroxymethylfurfural and Furfural
by Using Heterogeneous CoxOy–N@C
Catalysts
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
8
&
ÞÞ
These are not the final page numbers!

ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 0000, 00, 1 – 9
&
9
&
ÞÞ
These are not the final page numbers!