Selective Hydrodeoxygenation of 5-Hydroxymethylfurfural to 2,5-Dimethylfuran over Heterogeneous Iron Catalysts

Article abstract of DOI:10.1002/cssc.201700105

This work provided the first example of selective hydrodeoxygenation of 5-hydroxymethylfurfural (HMF) to 2,5-dimethylfuran (DMF) over heterogeneous Fe catalysts. A catalyst prepared by the pyrolysis of an Fe-phenanthroline complex on activated carbon at 800 °C was demonstrated to be the most active heterogeneous Fe catalyst. Under the optimal reaction conditions, complete conversion of HMF was achieved with 86.2 % selectivity to DMF. The reaction pathway was investigated thoroughly, and the hydrogenation of the C=O bond in HMF was demonstrated to be the rate-determining step during the hydrodeoxygenation, which could be accelerated greatly by using alcohol solvents as additional H-donors. The excellent stability of the Fe catalyst, which was probably a result of the well-preserved active species and the pore structure of the Fe catalyst in the presence of H₂, was demonstrated in batch and continuous flow fixed-bed reactors.

Full text of DOI:10.1002/cssc.201700105

Accepted Article
Title: Selective Hydrodeoxygenation of 5-Hydroxymethylfurfural to 2,5-
Dimethylfuran over Heterogeneous Iron Catalysts
Authors: Jiang Li, Jun ling Liu, He Yang Liu, Guang yue Xu, Jun jie
Zhang, Jia xing Liu, Guang lin Zhou, Qin Li, Zhi hao Xu, and
Yao Fu
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemSusChem 10.1002/cssc.201700105
Link to VoR: http://dx.doi.org/10.1002/cssc.201700105
A Journal of
www.chemsuschem.org

10.1002/cssc.201700105
ChemSusChem
Selective Hydrodeoxygenation of 5-Hydroxymethylfurfural to 2,5-
Dimethylfuran over Heterogeneous Iron Catalysts
Jiang Li,*[a] Jun-ling Liu,[a] He-yang Liu,[a] Guang-yue Xu,[b] Jun-jie Zhang,[a] Jia-xing Liu,[a] Guang-lin
Zhou,[a] Qin Li,[a] Zhi-hao Xu,[a] and Yao Fu*[b]
Abstract: This work provided the first example of selective
hydrodeoxygenation of 5-hydroxymethylfurfural to 2,5-dimethylfuran
over heterogeneous iron catalysts. Catalyst prepared by the
pyrolysis of iron-phenanthroline complex on activated carbon at 800
oC was demonstrated to be the most active heterogeneous iron
catalyst. Under the optimal reaction condition, complete conversion
of HMF was achieved with an 86.2% selectivity of DMF. The reaction
pathway was investigated thoroughly, and the hydrogenation of C=O
bond in HMF was demonstrated to be the rate-determining step
during the hydrodeoxygenation, which will be accelerated greatly by
using alcohol solvents as additional H-donors. The excellent stability
of the iron catalyst was demonstrated in batch and continuous flow
fixed-bed reactors, which was probably due to the well reserved
active species and the pore structure of the iron catalyst in the
presence of H2.
The first example of the generation of DMF from biomass-
derived carbohydrates was reported by Dumesic et al.[6] The
whole process was achieved by selective dehydration of
fructose to HMF in
a biphasic system followed by the
hydrodeoxygenation (HDO) of HMF to DMF over a chloride-
resistant carbon-supported copper-ruthenium (CuRu/C) catalyst.
The yield of DMF from HMF was 76~79%. Inspired by their work,
various researches aimed at developing more efficient catalyst
systems for the HDO of HMF to DMF were carried out, and most
of them are based on noble metals such as Pd, Ru and Pt.
Rauchfuss et al. reported that the combination of Pd/C catalyst
and formic acid solvent enables an efficient one-pot synthesis of
DMF from fructose with a total yield of 51%.[13] Meanwhile, Bell
et al. reported a two-step process involving heteropoly acids
catalyzed dehydration of glucose to HMF and subsequent HDO
of HMF with Pd/C catalyst in ionic liquids with the addition of
acetonitrile.[14] Saha et al. reported a one-pot conversion of
lignocellulosic and algal biomass into DMF by using
a
multicomponent catalytic system consisted of [DMA]+[CH3SO3]-,
Ru/C, and formic acid.[15] Vlachos et al. reported the catalytic
transfer hydrogenation (CTH) of HMF to DMF using secondary
alcohols as the hydrogen donors over a Ru/C catalyst with a
DMF yield of up to 80%,[16] and then the role of Ru and RuO2
phases was studied in detail.[17] Hermans et al. reported the CTH
of HMF to DMF over Fe2O3-supported Pd catalyst with 2-
propanol as hydrogen donor, and the yield of DMF was 72%.[18]
Wang et al. reported the HDO of HMF over a Ru/Co3O4 catalyst
Introduction
The utilization of renewable sources especially lignocellulosic
biomass to replace fossil reserves to produce fuel and chemicals
has
recently
attracted
significant
attention.[1,2]
5-
hydroxymethylfurfural (HMF), which can be produced from
biomass-derived hexoses through acidic hydrolysis, has been
considered as one of the most important platform chemicals
during biomass conversion.[3] The high functionality of HMF
allows it to be converted to various biofuel molecules such as
ethyl levulinate,[4] 5-ethoxymethylfurfural,[5] 2,5-dimethylfuran,[6]
and value-added chemicals such as levulinic acid,[7] furfuryl
alcohol,[8] 2,5-diformylfuran,[9] 2,5-furandicarboxylic acid,[10]
terephthalic acid,[11] and caprolactone.[12] Among these HMF
derivatives, 2,5-dimethylfuran (DMF) is particularly attractive. It
was proposed as a more promising transportation fuel due to its
higher energy density, higher boiling point and lower solubility in
water as compared to ethanol.
o
at 130 C and 0.7 MPa H2 in Tetrahydrofuran (THF), and the
yield of DMF was 93.4%.[19] Hu et al. studied the HDO of HMF
o
over a Ru/C catalyst in THF at 200 C for 2 h, and 94.7% DMF
yield with 100% HMF conversion was achieved.[20] Kawanami et
al. reported the use of supercritical carbon dioxide–water on the
HDO of HMF over a Pd/C catalyst, and a very high yield (100%)
of DMF was observed at 80 °C for 2 hours.[21] More recently, the
HDO of HMF to DMF over Ru/NaY, Ru/HT, Pt/C, Pt/rGO, and
Pd-Cs2.5H0.5PW12O40/K-10 clay catalysts have also been
reported.[22-26]
Besides above-mentioned monometallic catalyst systems, the
HDO of HMF to DMF over noble metal-based bimetallic
catalysts has also been reported. Notably, Schuth et al. reported
that PtCo bimetallic nanoparticles could be used as highly
efficient catalyst for the HDO of HMF to DMF in butanol.[27] The
conversion of HMF achieved 100% within 10 min, and the yield
of DMF reached 98% after 2 h. Subsequently, Gorte et al.
further explored the mechanism for this high selectivity.[28] The
formation of a monolayer oxide on the surface of the metallic
core was suggested as the fundamental principle, which
interacts weakly with the furan ring to prevent
overhydrogenation and ring opening of DMF while providing
active sites for the HDO reaction. In addition, Pt-Ni, Pt-Zn and
Pt-Cu alloyed nanocrystals have also been demonstrated to be
effective catalysts for HMF HDO with 98% DMF yield.[29]
Dumesic et al. reported that DMF can be produced with 46%
[a]
Dr. J. Li, J.-L. Liu, H.-Y. Liu, J.-J. Zhang, J.-X. Liu, Dr. G.-L. Zhou,
Q. Li, Z.-H. Xu
State Key Laboratory of Heavy Oil Processing
Institute of New Energy
China University of Petroleum (Beijing)
Beijing 102249, China
Fax: (+86) 010-89731300
Email: lijiang@cup.edu.cn
G.-Y. Xu, Prof. Dr. Y. Fu
Anhui Province Key Laboratory of Biomass Clean Energy
Department of Chemistry
University of Science and Technology of China
Hefei 230026, China
[b]
Fax: (+86) 551-3606689
E-mail: fuyao@ustc.edu.cn
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
yield by hydrogenolysis of HMF in the presence of lactones
using a RuSn/C catalyst at 200 °C.[30] Abu-Omar et al. reported
that a bimetallic catalyst which contained a Lewis-acidic ZnII and
Pd/C components was effective for the HDO of HMF to DMF
with high conversion (99%) and selectivity (85%).[31] Ebitani et al.
reported a bimetallic Pd50Au50/C catalyst catalyzed HDO of HMF
to DMF in the presence of hydrochloric acid (HCl) under an
atmospheric hydrogen pressure.[32]
selectivity of DMF. The detail study of the reaction pathway
indicated that the hydrogenation of C=O bond was the rate-
determining step, which will be boosted greatly by using alcohols
as both solvents and additional H-donors. The excellent stability
of iron catalysts was proved in batch and continuous flow fixed-
bed reactors. Thus, nitrogen-doped activated carbon-supported
iron catalysts were demonstrated to be an effective, stable
catalyst system for the HDO of HMF to DMF.
Considering the high cost and scarcity of noble metals, the
HDO of HMF to DMF over non-noble metal catalysts was
investigated recently. The activities and stabilities of six carbon-
supported noble or non-noble metal catalysts had been
compared in a continuous flow reactor by Gorte et al..[33] Barta et
al. reported the catalytic conversion of HMF over a Cu-doped
porous metal oxide in supercritical methanol, and a combined
yield (DMF + DMTHF) of 58% was achieved.[34] Subsequently,
they studied the conversion of HMF over Cu0.61Mg2.33Al0.98Ru0.02
and copper–zinc nanoalloy catalysts, and up to 97% combined
Results and Discussion
Catalyst screening
The fabrication process for the iron catalysts was the same as
our previous study, which was generally involved the
simultaneous pyrolysis of metal precursors and nitrogen
precursors on several supports at a pyrolysis temperature
o
products yields were obtained at 200–220 C using 20–30 bar
H2.[35,36] Fu et al. used nickel–tungsten carbide and perovskite
type oxide supported Ni catalysts for the conversion of HMF into
DMF.[37,38] Zhu et al. reported a series of researches concerning
about the catalytic conversion of HMF over non-noble metal
catalysts such as Raney Ni, NiSi-PS, Ni/Al2O3 and mineral-
derived Cu catalysts.[39-42] They also studied the one-step
continuous conversion of fructose to DMF over combined HY
zeolite and hydrotalcite (HT)-Cu/ZnO/Al2O3 in a fixed-bed reactor
with a DMF yield of 40.6%.[43] Wang et al. studied the catalytic
transfer hydrogenation/hydrogenolysis of HMF to DMF over a
carbon supported nickel-cobalt catalyst with formic acid as
hydrogen donor with the highest DMF yield of 90.0%.[44] Yuan et
al. recently reported that Cu-Co bimetallic nanoparticles coated
with carbon layers were very efficient catalyst for the HDO of
HMF to DMF.[45] The key issue for these Cu and Ni-based
promising catalyst system is how to suppress the
overhydrogenation of the aromatic furan ring and the ring-
opening of DMF to improve the selectivity of DMF.
Iron is an abundant, eco-friendly, relatively nontoxic, and
inexpensive element, and thus, a very promising alternative to
precious metals in catalysis. However, to the best of our
knowledge, the possibility of iron-catalyzed HMF HDO to DMF
was still questionable. For example, a DMF yield of 91.3% could
be achieved over carbon nanotube-supported bimetallic Ni-Fe
(Ni-Fe/CNT) catalysts, but trace yield of DMF was obtained over
Fe/CNT catalyst with only 3.2% conversion of HMF.[46]
o
o
ranged from 400 C to 1000 C (see figure S1 in the supporting
information).[52] Five compounds are used as nitrogen precursors,
including 1,10-phenanthroline (L1), 2,2'-bipyridine (L2), 2,2';6',2"-
terpyridine (L3), 8-hydroxyquinoline (L4) and phenylglycine (L5).
A native iron complex, hemin (6) was directly used as the
precursor of metal and nitrogen. Activated carbon, SiO2, and
Al2O3 were used as the supports. The as-prepared catalysts
were denoted as Fe-Ln/S-T, where Ln represents the type of
nitrogen precursor, S represents the type of support, and T
represents the pyrolysis temperature. The catalyst prepared
from hemin was denoted as 6/C-800.
The HDO of HMF to DMF over heterogeneous iron catalysts
was first investigated at 240 oC 1MPa H2 in THF, which has
been demonstrated to be a good solvent for heterogeneous iron
catalysis and HMF conversion.[37,49] A glass vial was used to
exclude the influence of reactor body. Besides the target product
DMF, 5-methylfurfural (MF), 5-methyl-2-furanmethanol (MFM),
2,5-dihydroxymethylfuran (DHMF), and 2,5-diformylfuran (DFF)
were also observed as by-products. In addition, the absence of
ring-hydrogenated
products
such
as
2,5-dimethyl
tertrahydrofuran (DMTHF), 2-hexanone, and tetrahydrofurfural
alcohol demonstrated that the iron catalysts did not hydrogenate
the aromatic rings of furanic compounds. The carbon mass
balance(CMB) was calculated based on the detected known
products.
In the blank test, only 1.1% yield of DMF was observed (Table
1, entry 1), indicating that the HDO of HMF to DMF was unlikely
to occur without any catalysts. In addition, 11.3% yield of MF
indicated the occurrence of C-O bond HDO. The yield of DMF
increased significantly in the presence of heterogeneous iron
catalysts. Catalysts prepared from different nitrogen precursors
were investigated at first (entry 2-7). Fe-L1/C-800 catalyst gave
a highest DMF yield of 32.5%, and MF and DFF were observed
as major by-products with yields of 30.3% and 7.5%,
respectively. The generation of DFF during HMF HDO, which
was further demonstrated by the GCMS spectrum (figure S2A),
has rarely been reported before. While the iron-catalyzed CTH
was confirmed in a later section, we proposed that DFF was
Recently, nitrogen-doped carbon materials supported iron
catalysts has attracted much attention in some important
chemical reactions such as oxygen reduction reaction (ORR)
and nitroarenes hydrogenation.[47-49] Beller et al. reported some
pioneer works using heterogeneous iron catalysts for organic
synthesis, for example selective hydrogenation of nitroarenes,
green synthesis of nitriles and green reductive aminations.[49-51]
Inspired by their works, we recently reported the efficient CTH of
furfural to furfuryl alcohol over heterogeneous iron catalysts.[52]
In this work, the first example of iron-catalyzed HDO of HMF to
DMF was reported. Under the optimal reaction condition,
complete conversion of HMF was achieved with an 86.2%
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
Table 1. Catalyst screening for HDO of HMF to DMF.[a]
nickel catalysts have been previously
reported to be more effective than iron
catalysts in the HDO of furanic
compounds,[53] the pyrolysis of metal-
phenanthroline complexes on activated
carbon seemed to generate more active
iron catalysts than nickel catalysts.[54]
In our previous research, we have
demonstrated that the catalytic activities
of the iron catalysts were determined by
the nitrogen content and the percentage
of N-Fe specie by using various
Entry
Catalysts
Conversion
[%]
Yield [%]
CMB[b]
[%]
DMF
1.1
32.5
6.6
15.8
3.9
14.9
30.7
11.4
14.1
9.8
5.0
12.5
5.2
MF
MFM
0.4
0.1
<0.1
0.2
<0.1
<0.1
0.3
0.1
0.4
4.1
0.5
0.6
<0.1
0.3
<0.1
0.9
0.2
1.3
-
DHMF
0.2
0.5
0.1
0.6
0.1
0.1
0.3
0.5
0.2
1.4
0.2
0.1
0.9
0.2
0.2
0.9
0.2
0.5
-
DFF
-
1
2
3
4
5
6
7
8
-
19.6
73.0
34.4
55.0
35.3
41.9
70.7
67.8
35.7
44.3
70.4
74.9
34.5
29.2
45.6
17.0
28.9
30.1
23.4
96.8
11.3
30.3
23.7
26.6
15.2
12.9
30.8
45.1
11.7
19.1
20.3
30.5
13.3
5.8
13.0
70.9
31.2
50.5
29.2
38.2
63.2
63.4
26.8
39.7
42.4
48.0
24.9
20.1
35.4
14.6
17.8
24.2
15.0
72.7
Fe-L1/C-800
Fe-L2/C-800
Fe-L3/C-800
Fe-L4/C-800
Fe-L5/C-800
6/C-800
Fe-L1/SiO2-800
Fe-L1/Al2O3-800
Fe-L1/TiO2-800
Fe-L1/C-400
Fe-L1/C-600
Fe-L1/C-1000
Fe/C-800
L1/C-800
Fe-L1
7.5
0.8
7.3
10.0
10.3
1.1
6.3
0.4
5.3
16.4
4.3
5.5
3.4
5.8
3.9
4.5
0.8
6.9
6.4
technologies
especially
XPS
9
characterization.[52] Moreover, a recent
study by Beller et al. demonstrated that
the active site in this kind of pyrolyzed
iron catalyst was consisted of a unique
structure, where iron oxides were
surrounded by nitrogen-doped-graphene
shells then immobilized on carbon
10
11
12
13
14
15
16
17
18
19
20
10.4
8.9
0.5
4.2
1.8
20.5
8.4
Fe
L1
C
8.7
19.8
6.9
support.[55]
We
explored
the
1.2
41.5
morphologies of Fe-L1/C-800, Fe-L1/C-
400, and Fe-L4/C-800 catalysts by TEM
and Energy dispersive X-ray (EDX)
analysis (figure S3). No obvious iron
particles could be observed in the TEM
image of Fe-L1/C-800 catalyst at different
Co-L1/C-800
23.6
0.3
0.9
21
Ni-L1/C-800
52.2
24.6
24.3
0.2
0.1
0.3
49.5
[a] Reaction conditions: 0.5 mmol HMF, 20 mL THF, 0.1 g iron catalyst, 1 MPa H2, t=12h. The molar
ratio of iron/HMF was kept at 11.2 mol%. [b] CMB is carbon mass balance, which is calculated based
on the detected known products.
generated by the dehydrogenation of HMF when HMF itself
areas, and EDX analysis showed extremely weak signals of iron
species. This suggested a homogeneous distribution of iron
species on the catalyst surface. In contrast, cloudy
agglomerates can be observed in the TEM images of Fe-L1/C-
400 and Fe-L4/C-800 catalysts, and stronger signals of iron
species can be detected by EDX analysis.
acted as H-donor during the reaction. Catalyst prepared from
native iron complex (6/C-800) gave a slightly lower DMF yield of
30.7%. In contrast, catalysts prepared from L2-L5 gave much
lower DMF yields. Thus, 1,10-phenanthroline (L1) was the most
effective nitrogen precursor.
Besides using activated carbon as the support, some metal
oxides such as SiO2, Al2O3, and TiO2 were also examined (entry
8-10). Much lower yields of DMF were obtained, confirming that
activated carbon was the most effective support. Subsequently,
the effect of pyrolysis temperature was also investigated (entry
To further explore the Fe configuration, H2-TPR analysis was
performed. It is well known that the Fe2O3 species in iron
catalysts have two characteristic peaks in the H2-TPR
analysis.[53] The first broad peak ranging from 410 to 570 oC
could be ascribed to the reduction of -Fe2O3 into -Fe3O4. The
second broad peak ranging from 590 to 730 oC is typically
attributed to the reduction of iron oxides to -Fe. As shown in
figure 1, the characteristic peaks of the Fe2O3 species could be
clearly observed in the Fe/C-800 catalyst, demonstrating that the
Fe2O3 species could be formed by the direct pyrolysis of iron
acetate onto activated carbon. In addition, we did not observe
any diffraction peaks of the Fe2O3 phase in the XRD pattern of
the Fe/C-800 catalyst,[52] suggesting that amorphous Fe2O3
phase could also be detected by the H2-TPR analysis. When
nitrogen precursors were added into the pyrolysis process, the
characteristic peaks of the Fe2O3 species could only be found in
the Fe-L3/C-800 and Fe-L4/C-800 catalysts. In contrast, only
one broad peak ranging from 200 to 800 oC was observed in the
H2-TPR results of the most active Fe-L1/C-800 and 6/C-800
catalysts, and it started from 500 oC and 300 oC for the Fe-L2/C-
800 and Fe-L5/C-800 catalysts, respectively. The curve fitting
results for the H2-TPR analysis of Fe-L1/C-800 catalyst
suggested that there should be an intense peak at 620 oC
besides peaks of Fe2O3 species. We proposed that this peak
o
11-13). The decrease of pyrolysis temperature from 800 C to
600 or 400 oC leads to a sharp decrease in DMF yield. In
addition, when the pyrolysis temperature was further raised to
1000 oC, significant decrease in both HMF conversion and DMF
yield were also observed. These results confirmed that the
optimal pyrolysis temperature was 800 oC.
Catalysts prepared by the pyrolysis of iron precursor or
nitrogen precursor on carbon support exhibited an inferior
catalytic activity towards the HDO of HMF (entry 14, 15). In
addition, the direct use of iron complex (Fe-L1), iron acetate (Fe),
1,10-phenanthroline (L1), or carbon support (C) also exhibited
poor catalytic performances (entry 16-19). Thus, the
simultaneous pyrolysis of iron precursor and nitrogen precursor
on carbon support was critical to its catalytic activity.
The replacement of the metal center iron with cobalt led to
enhanced HMF conversion and DMF yield. However, the yield of
unidentified products also increased to 24.1%. This cobalt-
catalyzed HMF HDO will be studied in detail in our further works.
In addition, when the metal center was replaced by nickel, a
decrease in the catalytic performance was observed. Although
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
the maximum at a higher temperature around 800 oC. The
desorption peak disappeared when the pyrolysis temperature
o
exceeded 600 C, and only a broad peak could be observed in
the H2-TPR analysis results of the Fe-L1/C-600 and
6/C-800
Fe-L5/C-800
Fe-L4/C-800
Fe-L3/C-800
Fe-L1/C-800 catalysts. When the pyrolysis temperature was
further increased to 1000 oC, the characteristic peaks of the
o
o
Fe2O3 species at 490 C and 650 C could be observed in the
H2-TPR result (see figure S4 for clearer observations). The
changes in the H2-TPR results were consistent with some
previous findings, for example the structure of iron oxides
surrounded by nitrogen-doped-graphene shells was formed at
pyrolysis temperature over 400 oC,[49] and the active sites of
FeNx species would be decomposed to Fe2O3 at a pyrolysis
Fe-L2/C-800
Fe-L1/C-800
Fe/C-800
200
400
600
800
temperature of 1000 C.[52] Thus, the peak at 620 C in the H2-
TPR results is likely the characteristic peak of FeNx species,
which were considered to govern the unique catalytic activity of
the iron catalysts in some previous studies.[49]
o
o
T/ o
C
Fe-L1/C-1000
Fe-L1/C-800
Fe-L1/C-600
Solvent effect
The unsatisfactory catalytic performance of Fe-L1/C-800 catalyst
in THF (aprotic polar solvent) required the further optimization of
the reaction parameters. Some previous studies indicated that
the chemical nature of the solvent have a significant impact on
the HDO of HMF to DMF.[23,35] We further examined the solvent
effect in the iron-catalyzed HDO reaction (table 2). When
hexane (non-polar solvent) was used as the solvent, HMF
conversion and DMF yield was improved slightly, but still far
from satisfying. Considering that we have recently reported the
efficient catalytic transfer hydrogenation of C=O bond in furfural
over iron catalysts by using alcohols as H-donors, it was
speculated that the hydrogenation of C=O bond in HMF maybe
O
Cl
N
N
N
N
OH
Fe
O
Fe-L1/C-400
OH
6
200
400
600
800
T/ o
C
Figure 1. H2-TPR of various iron catalysts.
could be attributed to the reduction of
iron cations coordinated by pyridinic
nitrogen functionalities (FeNx species) to
metallic iron. This is hard to be confirmed
because of the lack of previous studies
about the H2-TPR analysis of Fe-N-C
catalysts to our knowledge.
Table 2. Catalytic HDO of HMF to DMF over Fe-L1/C-800 catalyst in different solvents.[a]
Entry
Solvent
Atmosphere
Conversion
[%]
Yield[%]
CMB[b]
[%]
DMF
32.5
MF
30.3
MFM
0.1
DHMF
0.5
DFF
7.5
1 MPa H2
1 MPa H2
1 MPa H2
Without H2
1 MPa H2
Without H2
1 MPa H2
Without H2
1 MPa H2
Without H2
1 MPa H2
Without H2
1 MPa H2
Without H2
73
70.9
53.1
50.6
29.8
65.5
54.3
70.6
50.3
60.8
52.4
76.4
67.4
61.4
55.3
1
2
THF
Hexane
100
100
100
100
100
100
100
100
100
100
100
99.7
100
45.5
42.1
15.0
64.1
54.1
69.1
50.3
60.8
52.1
75.3
67.4
60.9
55.3
5.1
7.2
5.4
1.4
0.2
0.5
-
-
-
2.5
3
4
5
6
7
8
MeOH
0.2
<0.1
1.1
To test our hypothesis, we tried to
explore the H2-TPR of hemin (previously
denoted as 6), which was a native iron
complex contained FeN4 species, and
Fe-L1/C catalysts prepared at different
pyrolysis temperatures. For hemin, the
reduction peak started from 250 oC,
which was generally consistent with Fe-
L1/C-800 catalyst. Unfortunately, this
compound was unstable during H2-TPR.
Some gases were released, leading to a
large desorption peak after 500 oC. Thus,
the H2-TPR result of hemin could not
directly help us justify previous peak at
9.3
-
0.1
EtOH
<0.1
-
-
-
-
-
1-PrOH
2-PrOH
1-BuOH
2-BuOH
0.4
0.1
0.5
-
-
-
-
-
-
-
-
-
-
-
-
0.3
0.6
-
-
-
-
0.5
-
-
0.1
-
0.2
-
0.2
-
o
620 C. The desorption peak could also
be observed in the H2-TPR analysis of
the Fe-L1/C-400 catalyst, but it achieved
[a] Reaction conditions: 0.5 mmol HMF, 20 mL solvent, 0.1 g Fe-L1/C-800 catalyst, t=12h. [b] CMB is
carbon mass balance, which is based on the detected known products.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
could also be boosted by using alcohols as the solvent. Thus,
various primary and secondary alcohols were used to
investigate the solvent effect (Table 2, entry 3-8). It can be
clearly seen that the yield of DMF was increased dramatically
with a significant drop in the yield of MF, confirming that there
was a substantial enhancement in the hydrogenation of C=O
bond. Moreover, we still do not observe any ring-hydrogenated
products, indicating that using alcohols as the solvent will not
promote the hydrogenation of the furan ring. However, an
unsatisfactory carbon mass balance was observed during the
reaction. The polymerization of HMF to humins was probably the
main reason for the low carbon balance, which will be discussed
later in the reaction pathways section.
Conversion
DMF
MFM
MF
DFF
100
80
60
40
20
0
DHMF
120
160
200
240
280
T/ o
C
To further check the role of the alcohols, reactions were also
performed in an inert atmosphere (0.5 MPa N2). While using
methanol as the solvent, a significant drop in DMF yield was
observed in the absence of H2. By comparison, only slight
decreases in DMF yields were observed while using alcohols
with a carbon number of >2. Thus, it is concluded that the
alcohols with medium chain lengths will act as H-donors in the
HDO of HMF and provide alternative hydrogen source besides
gaseous hydrogen. The GCMS spectrum strongly confirmed that
the CTH reaction occurred even at a H2 pressure of 4 MPa.
Large amount of butyraldehyde dibutyl acetal (see figure S2B,
RT=18.088 min) could be observed when n-butanol was used as
the solvent at 4 MPa H2. It is the aldol condensation product
formed from n-butanol and butyraldehyde (the dehydrogenation
product of n-butanol produced in the CTH process). The molar ratio
of butyraldehyde dibutyl acetal to DMF was 0.23 at 0.5 h, and
then increased to 0.6-0.7 after 3 h. In comparison, the molar
ratio of butyraldehyde dibutyl acetal to DMF was 2.2 when the
reaction was performed under inert atmosphere. Thus, it could
be concluded that both the H-donor and gaseous H2 supplied
the hydrogen consumed during the HDO of HMF to DMF.
Figure 2. Effect of reaction temperature on the HDO of HMF to DMF. Reaction
conditions: 0.5 mmol HMF, 20 mL n-butanol, 0.1 g Fe-L1/C-800 catalyst, 1
MPa H2, t=12h.
1h
12h
100
80
60
40
20
0
1
2
3
4
5
PH2 / MPa
The yield of DMF increased along with the increase in alcohol
chain length from methanol to n-butanol, which is probably due
to the dehydrogenation activity of the alcohol (Table S1, see the
supporting information). In addition, reactions performed in
primary alcohols gave slightly higher DMF yields than in
secondary alcohols, although secondary alcohols represented
higher dehydrogenation activity than primary alcohols. This
solvent effect was probably attributed to solubility of hydrogen,
thermal interaction between HMF and the solvent molecules,
and competitive adsorption onto the catalytic sites, [56] which was
still needed to confirm in future works.
Figure 3. Effect of hydrogen pressure on the HDO of HMF to DMF. Reaction
conditions: 0.5 mmol HMF, 20 mL n-butanol, 0.1 g Fe-L1/C-800 catalyst, T=
240 oC, t= 1h or 12h.
DFF yield of 4.6%, indicating that the HDO of HMF was still
inefficient. In addition, the HMF conversion was found to be
much higher than the total yield of furanic compounds at these
two reaction temperatures. This could be attributed to the
generation of acetals or ethers byproducts from HMF confirmed
by GCMS analysis. A significant enhancement in DMF yield was
observed when the reaction temperature was further increased
to 200 oC. However, some intermediates such as MF, MFM, and
DHMF were still present in considerable quantities. The highest
Effect of reaction temperature and H2 pressure
o
DMF yield of 75.3% was observed at 240 C, and its yield was
The effect of reaction temperature was investigated from 120 to
o
decreased slightly while the reaction temperature was further
raised to 280 oC. In addition, acetals or ethers byproducts
280 C at 1 MPa with the Fe-L1/C-800 catalyst, and the results
were shown in figure 2. Accordingly, it was found that the
reaction temperature played an important role in the HDO of
HMF. The absence of DMF with low yields of intermediates such
as MFM, MF and DHMF indicated that the HDO of HMF to DMF
o
disappeared when the reaction temperature exceeded 240 C.
o
Thus, 240 C was chosed as the optimal reaction temperature
for further investigations.
o
was unlikely to occur at 120 C. When the reaction temperature
was increased to 160 oC, the DMF yield increased slightly with a
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
Subsequently, as shown in figure 3, the effect of hydrogen
pressure on the HDO of HMF to DMF was investigated at 240 oC
with the Fe-L1/C-800 catalyst. At a reaction time of 1 h, the
highest DMF yield of 39.1% was obtained at H2 pressure of 4
MPa. In comparison, the DMF yield was only 6.2% at 3MPa H2.
When the reaction was prolonged to 12 h, trace yields of
intermediates such as MF and DHMF were observed at a
Reaction Pathways
100
Conversion
DMF
MFM
MF
DFF
DHMF
80
60
40
20
0
hydrogen pressure of
1 MPa, and these products were
undetectable at higher hydrogen pressures. In addition, no ring-
hydrogenated products were observed even at a hydrogen
pressure of 5 MPa. Full conversion of HMF was achieved in
each case, and the highest DMF yield of 86.2% was observed at
4 MPa. Thus, a hydrogen pressure of 4 MPa was used in
following experiments.
0
2
4
6
8
10
12
t / h
Effect of iron loading
100
80
60
40
20
0
Conversion
DMF
100
80
60
40
20
0
Conversion
DMF
MFM
MF
DHMF
0
2
4
6
8
10
12
2
3
5
10
t / h
Theoretical iron content / wt%
Figure 5. Product distributions of the HDO of HMF as a function of reaction
time in THF (top) and n-butanol (bottom). Reaction conditions: 0.5 mmol HMF,
20 mL solvent, 0.1 g Fe-L1/C-800 catalyst, 4 MPa H2.
Figure 4. Effect of iron loading on the HDO of HMF to DMF. Reaction
conditions: 0.5 mmol HMF, 20 mL n-butanol, 0.1 g Fe-L1/C-800 catalyst, 4
MPa H2, t=12h.
On the basis of the above catalytic results, two preliminary key
Iron catalysts with four different iron contents were prepared to
explore the effect of iron loading. The theoretical iron contents
were determined as 2, 3, 5, 10 wt%, which were calculated by
dividing the weight of iron in the iron precursors by the weight of
carbon supports. However, the actual iron contents were 1.52,
2.19, 3.14, and 4.87 wt%, respectively. The differences could be
attributed to the weight of the nitrogen precursors. In addition,
the Fe/N molar ratios were 0.22, 0.26, 0.34, and 0.62,
respectively. The catalyst dosage was kept constant during the
examination. As shown in figure 4, full HMF conversion was
observed in each catalytic experiment, and the highest DMF
yield was achieved at a theoretical iron content of 5 wt%.
Besides the target product DMF, only DFF was detected as the
by-product at the lowest theoretical iron content of 2 wt% with a
yield of 7.7%. Thus, it seems that a low iron loading would lead
to an inefficient hydrogenation activity. The inferior catalytic
performance of iron catalyst with 10 wt% loading suggested that
higher Fe/N molar ratio would lead to lower DMF selectivity.
Thus, the optimal iron loading was 5%.
points could be summarized: (a) MF is
a possible key
intermediate in the catalytic reaction pathway; (b) DMF yield
could be enhanced by using alcohols as both solvents and H-
donors. In order to better understand the reaction pathways, we
explored the product distribution of HMF HDO at different
reaction time in THF and n-butanol (figure 5). While using THF
as the solvent, the conversion of HMF increased gradually
throughout the reaction time. In contrast, total conversion of
HMF was achieved within 3 h with n-butanol as the solvent.
Thus, the conversion rate of HMF was enhanced dramatically by
promoting the hydrogen-donating ability of the solvent.
In addition, a significant difference in the intermediate product
distribution in THF and n-butanol was also observed. When THF
was used as the solvent, only 34.7% yield of the target product
DMF was obtained at 12 h. MF was observed as the major by-
product, which achieved a maximum yield of 60.6% at 6 h, and
then
decreased
gradually.
This
product
distribution
demonstrated that the reaction rate of HMF HDO to MF was
much higher than that of MF HDO to DMF, leading to the
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
100
80
60
40
20
0
100
80
60
40
20
C onversion
D M F yield
Conversion
DMF
0
0
2
4
6
8
10
12
0
2
4
6
8
10
12
t / h
t / h
100
80
60
40
20
100
80
60
40
20
0
Conversion
DMF yield
C onversion
D M F yield
0
0
1
2
3
4
5
6
t/ h
0
1
2
3
4
5
6
t/ h
Figure 6. Product distributions of the HDO of MFM as a function of reaction
time in THF (top) and n-butanol (bottom). Reaction conditions: 0.5 mmol MFM,
20 mL solvent, 0.1 g Fe-L1/C-800 catalyst, 4 MPa H2.
Figure 7. Product distributions of the HDO of MF as a function of reaction time
in THF (top) and n-butanol (bottom). Reaction conditions: 0.5 mmol MF, 20 mL
solvent, 0.1 g Fe-L1/C-800 catalyst, 4 MPa H2.
accumulation of MF in the solution. Trace yields of other alcohol
intermediates such as MFM and DHMF could be attributed to
the high reaction rate of C-O bond HDO. When n-butanol was
used as the solvent, the production rate of DMF was enhanced
significantly, and a maximum yield of 86.2% was obtained within
5 h. MF achieved a maximum yield of 9.9% at 1 h, and then be
converted completely after 6 h. In addition, DFF, which was the
dehydrogenation product of HMF, achieved a maximum yield of
4.4% in THF, but could not be detected with n-butanol as the
solvent. Thus, we proposed that promoting the hydrogen-
donating ability of the solvent would significantly improve the
reaction rate of the HDO of MF to DMF, thus leading to a higher
yield and production rate of DMF.
We further carried out a series of experiments to demonstrate
our hypothesis. In general, the HDO of MF could be divided into
two reactions: (a) the hydrogenation of MF to MFM; (b) the HDO
of MFM to DMF. We first explored the solvent effect on the HDO
of MFM to DMF (figure 6). The complete conversion of MFM in
THF was achieved within 2 h with a 65.6% yield of DMF. In
comparison, slightly higher conversion rate of MFM and DMF
yield were observed while using n-butanol as the solvent. Thus,
it can be concluded that the hydrogen-donating ability of the
solvent has considerable influence on the reaction rate of the
HDO of C-O bond in MFM and the selectivity of target product
DMF.
Subsequently, we explored the product distribution of MF
HDO in THF and n-butanol (figure 7). It can be clearly seen that
the HDO of MF proceeded fast in n-butanol, and a maximum
DMF yield of 89.8% was observed at a reaction time of 2 h. In
contrast, only 50.6% yield of DMF with a 66.8% conversion of
MF was obtained in THF even though the reaction time was
prolonged to 12 h. Thus, the HDO of MF was significantly
accelerated by using solvent with stronger hydrogen donating-
ability. This improvement was much more significant than that
observed in the HDO of MFM to DMF. This enhancement could
also be demonstrated by our latest work, in which the selective
catalytic transfer hydrogenation of C=O bond in furfural was
successfully achieved over this kind of heterogeneous iron
catalysts with alcohols as the H-donors.[42]
The HDO of MF and MFM over Fe/C-800 catalyst in n-butanol
was explored to check the role of FeNx species in the HDO of
C=O and C-O bond (see figure S5 in the supporting information).
The DMF yields were 44% and 53% for the HDO of MF and
MFM, respectively. The inferior catalytic performance
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
spectra were shown in figure S2 (see the
Table 3. Catalytic HDO of various possible intermediates over Fe-L1/C-800 catalyst.[a]
supporting
information).
The
Entry
Solvent
Substrate
Conversion
[%]
Yield [%]
CMB[b]
[%]
overhydrogenation and ring-opening products
of the furan ring in HMF were reported to be
the major by-products in previous researches
on the catalytic conversion of HMF to DMF
over other transition metals with stronger
hydrogenation ability. However, in this work,
after checking all the experiments, including
the time course during the reaction, we did not
observe any overhydrogenation and ring-
opening products in the GC-MS spectra. Thus,
it was demonstrated that the iron catalysts did
not destroy the aromatic structure in HMF
during the HDO. Moreover, the only possible
DMF
63.1
9.1
-
MF
8.7
0.5
-
MFM
DHMF
DFF
1
2
3
THF
THF
THF
DHMF
DFF
100
-
-
15.3
-
-
71.8
80.2
-
94.6
<5%
54.3
-
1.0
DMF
-
-
-
-
85.2
81.4
-
4
5
n-butanol
n-butanol
DFF
96.3
100
85.2
81.4
-
-
-
-
-
DHMF
-
6
n-butanol
DMF
<5%
-
-
-
-
[a] Reaction conditions: 0.5 mmol substrate, 20 mL solvent, 0.1 g Fe-L1/C-800 catalyst, 4 MPa
H2, t=12h. [b] CMB is carbon mass balance, which is based on the detected known products.
demonstrated that the presence of FeNx species was critical to
HMF-derived by-product detected in GC-MS spectra was a
dimer with a molecular weight of 190 (the recommended
structure by NIST spectrum library is α-Furil) with a yield of ca.
4%, indicating that the polymerization of furanic compounds
could be catalyzed by the lewis acidity of the iron catalysts. Thus,
we proposed that the unsatisfactory carbon balance was
probably due to the lewis acid-catalyzed polymerization, leading
to the formation of humins, which almost could not be detected
in GC-MS. Hence, further effort will be devoted in the careful
control of the hydrogenation ability and lewis acidity of the iron
catalysts to achieve higher DMF selectivity.
At last, on the basis of these experiments and previous
researches, we proposed the following reaction pathway for the
HDO of HMF to DMF over heterogeneous iron catalysts
(scheme 1). When the reaction was performed in solvents with
weak hydrogen-donating ability such as THF, the HDO of HMF
to MF proceeded much faster than the HDO of MF to DMF,
resulting in a large amount of MF in the product mixture with a
low DMF yield. The hydrogenation of C=O bond was the rate-
determine step during the reaction. In addition, the
dehydrogenation of HMF to DFF also occurred. When solvents
with higher hydrogen-donating ability were used, the
hydrogenation of C=O bond was improved significantly by the
catalytic transfer hydrogenation. Thus, both the hydrogenation of
C=O bond and the HDO of C-O bond proceeded rapidly in
alcohol solvents, and MF and DHMF were the possible
intermediates during the reaction.
both HDO reactions.
We also conducted experiments with several other possible
reaction intermediates such as DHMF, DFF, and product DMF
(Table 3). The complete conversion of DHMF in THF was
achieved with 63.1% yield of DMF and 8.7% yield of MF. The
presence of MF demonstrated that the dehydrogenation of C-O
bond occurred in THF. When n-butanol was used as the solvent,
the DMF yield was increased to 81.4%, and no furanic by-
products were detected. While we have demonstrated that the
hydrogenation of C=O bond in MF was enhanced significantly by
using n-butanol as the solvent, the hydrogenation of C=O bond
in HMF was also possibly improved to form DHMF as the
intermediate to produce DMF.
The catalytic results of DFF conversion further confirmed the
difference between the HDO of C=O bond in THF and n-butanol.
While using THF as the solvent, the major product was MF with
a yield of 54.3%, and the yield of the target product DMF was
only 9.1%. In addition, 15.3% yield of HMF was detected. In
contrast, a 96.3% conversion of DFF with 85.2% DMF yield was
observed by using n-butanol as the solvent. Thus, for the
hydrogenation of C=O bond over heterogeneous iron catalysts,
the catalytic transfer hydrogenation by H-donors was more
preferable than direct hydrogenation by gaseous hydrogen.
The reactivity of product DMF in both solvents was also
investigated. The conversion of DMF was lower than 5%,
confirming the stability of DMF in the reaction condition.
Due to the unsatisfactory carbon balance (usually <90%)
during the HMF conversion especially when using alcohols as
the solvent, GC-MS analysis was performed to further identify
Recyclability
the by-products during the reaction, and two typical GC-MS
in THF
HO
OH
O
The recyclability of the heterogeneous iron catalysts was
examined in five consecutive runs under identical reaction
conditions, and the results were shown in figure 8. The used
catalyst was recovered by centrifugation, washed with n-butanol
for several times, and then used directly in next run. Complete
conversions of HMF were achieved in each recycle run with
similar DMF yields, and no other byproducts were observed.
Thus, the catalytic activity of the iron catalyst maintained very
well during the recycle test.
in n-butanol
O
OH
HO
O
O
O
O
O
O
O
O
The excellent stability of the iron catalysts in the HDO of HMF
was unexpected because this catalyst had represented an
Scheme 1. A plausible reaction pathway for the HDO of HMF to DMF over
heterogeneous iron catalysts in THF and n-butanol.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
Table 4. Textural properties of the fresh and reused Fe-L1/C-800 catalysts..
BET surface area
Pore Volume
[cm³/g]
Pore Size
[nm]
Sample
Reused under H2
Reused under N2
[m²/g]
100
80
60
40
20
0
Fresh
830
635
495
0.89
0.79
0.53
4.27
5.56
5.03
Reused under H2
Reused under N2
The XPS spectrum of the Fe-L1/C-800 catalyst reused under H2
clearly showed that the N•••Fe specie was retained but with
lower percentage content. In comparison, the peak of N•••Fe
specie slightly shifted to higher binding energies (from 399.7 eV
to 399.9 eV) for the iron catalyst reused under N2, and the
1
2
3
4
5
percentage content of graphitic N increased dramatically,
Number of cycles
leading to the formtion of a new peak at 401.3 eV in the XPS
spectrum.
The XRD and BET characterization results also implicated
that the structure of the iron catalyst was more easily destroyed
under N2 (figure 9). The peak of metallic iron at 44.6o in the XRD
pattern of fresh Fe-L1/C-800 catalyst disappeared after recycling
under H2 for five runs. In comparison, significant changes in the
XRD pattern were observed when the catalyst was reused under
N2, in which three broad humps at around 18o, 30o, and 41.5o
were present. Some possible species such as amorphous
carbon generally shows typical diffraction peaks of 26o and 43o,
and Fe2O3 phase has a distinct peak at 35o. Thus, the peaks in
the XRD pattern of catalyst reused under N2 cannot be attributed
to those species because of the considerable deviations.
Nonetheless, the XRD patterns still suggested that the structure
of the catalyst was destroyed more significantly when reused
under N2. The BET surface area and pore volume of Fe-L1/C-
800 catalyst also decreased more dramatically while recycled
under N2 (Table 4), indicating that the porous structure of the
catalyst was destroyed more significantly. In summary, the
unexpected stability of the catalyst for the HDO of HMF could be
attributed to the better reservation of the active species and the
pore structure in the presence of H2.
Figure 8. Recycling results of Fe-L1/C-800 for the HDO of HMF to DMF under
two different reaction atmospheres. Reaction condition: 0.5 mmol substrate,
20 mL solvent, 0.1 g Fe-L1/C-800 catalyst, t=12h, 4 MPa H2 or 1MPa N2.
unsatisfactory stability in the catalytic transfer hydrogenation of
furfural to furfuryl alcohol in our latest report. Accordingly, the
destruction of N•••Fe species, the presence of crystallized Fe2O3
phase, and the pore structure change were the main reasons for
the catalyst deactivation.[52] The only difference between these
two reactions is the reaction atmosphere. Thus, we proposed
that the presence of hydrogen is essential to the maintenance of
the catalytic activity. To test our hypothesis, the recyclability
tests were then performed under an inert atmosphere. The yield
of DMF increased slightly in the second run, and then dropped
significantly in the third run. In addition, MF was detected as a
byproduct with a yield of ca. 5% after the third cycle, suggesting
a decrease in the hydrogenation ability of the iron catalyst after
recycling under N2. Although a considerable recovery of catalytic
activity was observed in the fourth run, this catalyst was
definitely unstable under the inert atmosphere.
To explore the reasons for the different stabilities of the iron
catalysts under two different atmospheres, the reused catalysts
were further characterized by XPS, XRD, and BET technologies.
Continuous test
To further demonstrate the stability of the iron catalyst, the HDO
of HMF was also performed in a continuous flow fixed-bed
reactor in down-flow mode. The image and the diagram of the
reactor could be seen in figure S7. The reactor system was
consisted of a pump that introduced the liquid feed (25mM HMF
n-butanol solution with toluene as the internal standard), into a
stainless steel tubular reactor with an internal diameter of 17 mm
and a length of 670 mm. The pressure over the system was
controlled through a back-pressure regulator installed at the
outlet of the reactor. The catalyst weight was 5.4 g (ca. 15 mL)
and diluted with quartz sand (10 mL). The reaction was carried
out at 240 oC with a LHSV of 0.4 h-1. Owing to some limitations,
the pressure was kept at only 0.7 MPa at a H2 flow of 22 mL
min−1. The continuous test was performed for 72 h, and the
results were shown in figure 10. To our delight, the catalysts did
not deactivate after 72 h, suggesting its high stability during the
Reused under N2
Reused under H2
Fresh catalyst
20
40
2
60
80
o
/
Figure 9. XRD patterns of fresh and reused Fe-L1/C-800 catalysts.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
Thus, future work will focus on developing enhanced iron
catalyst system with carful control of the hydrogenation ability
and lewis acidity to achieve higher DMF selectivity.
100
80
60
40
20
Experimental Section
List of chemicals
HMF, DHMF, and DFF were generous gifts from Hefei Leaf Energy
Biotechnology Co., Ltd. DMF, MF, iron(II) acetate, 1,10-phenanthroline,
2,2'-bipyridine, 2,2';6',2"-terpyridine, hemin, and activated carbon were
purchased from TCI. MFM was purchased from J&K Chemical. 8-
hydroxyquinoline, phenylglycine, cobalt (II) acetate tetrahydrate, nickel
acetate tetrahydrate, and alcohol solvents were purchased from
Sinopharm Chemical Reagent Co. Ltd. Metal oxide supports such as
titanium dioxide, silicon dioxide, and aluminium oxide were purchased
from Aladdin Reagent Co. Ltd.
0
0
20
40
60
80
t/ h
Figure 10. Dependence of DMF yield on time in the fixed-bed reactor.
Reaction condition: 240 oC, 25mM HMF n-butanol solution mixed with with
toluene as the internal standard, LHSV=0.4 h-1, p=0.7 MPa, H2 flux=22 mL
min-1.
Catalyst preparation
The preparation of catalyst used a simultaneous pyrolysis of metal
complex and carbon material similar to beller et al.[57] A typical procedure
for catalyst preparation is described as follows: 0.5 mmol iron precursor
and corresponding ligands were added to 50 mL ethanol under
vigorously stirring at room temperature. Then 1 g activated carbon was
added into the solution and the mixture was stirred at 60 oC for 15 h
followed by rotary evaporation at 30 C. Obtained solid was grinded into
fine powder, and then pyrolyzed under an argon atmosphere with a gas
flow rate of 100 mL/min in a tubular furnace.
continuous test. The DMF yield was ca. 80%, which was slightly
lower than the best results obtained in the batch reactor.
The catalyst after reaction was further characterized by XPS,
XRD, and BET technologies. The XPS spectrum and the XRD
pattern was similar to those of the fresh catalyst (figure S8). The
new peak at 26.5o in the XRD pattern could be attributed to the
quartz sand, which was hard to be isolated completely from the
catalyst. The BET suface area was decreased to 630 m²/g,
which was nearly the same as the catalyst resued under H2 in
the batch reactor (Table S2). To sum up, the iron catalyst was
demonstrated to be stable during the continuous test in the
presence of H2.
o
Catalyst characterization
Transmission electron microscopy (TEM) was performed on a JEOL-
2010 electron microscope. The samples were deposited on a Cu grids
after ultrasonic dispersion of the samples in ethanol.
Nitrogen adsorption measurements were performed using an
ASAP2020M adsorption analyzer which reports adsorption isotherm,
specific surface area and pore volume automatically. The Brunauer-
Emmett-Teller (BET) equation was used to calculate the surface area in
the range of relative pressures between 0.05 and 0.20. The pore size
was calculated from the adsorption branch of the isotherms using the
thermodynamic 60 based Barrett-Joyner-Halenda (BJH) method.
XRD analysis was conducted on an X-ray diffractometer 65 (TTR-III,
Rigaku Corp., Japan) using Cu Kα radiation (λ=1.54056 Å). The data
were recorded over 2θ ranges of 10-70°.
Conclusions
The Iron-catalyzed HDO of HMF to DMF was explored in this
work. The examination of the effect of the nitrogen precursor,
support, pyrolysis temperature, and the metal center
demonstrated that catalyst prepared by the pyrolysis of iron
actate and 1,10-phenanthroline on activated carbon at 800 C
was the most active heterogeneous iron catalyst. Complete
conversion of HMF with an 86.2% selectivity of DMF was
o
XPS was obtained with an X-ray photoelectron spectroscopy
(ESCALAB250, Thermo-VG Scientific, USA) using monochromatized Al
Kα radiation (1486.92 eV).
o
achieved in n-butanol at 240 C for 5 h. Alcohol solvents could
Temperature-programmed reduction (TPR) was conducted on a Quanta
Chembet chemisorption instrument with a thermal conductivity detector
(TCD). About 100 mg of samples was loaded in a quartz reactor, and
then heated to 800 °C with a heating ramp rate of 10 °C min–1 in a
act as H-donors during the HDO, enhancing the reaction rate of
the HDO reactions especially the hydrogenation of C=O bond.
The excellent stability of the iron catalysts was demonstrated in
batch and continuous flow fixed-bed reactors. The well reserved
active species and the pore structure of the iron catalyst in the
presence of H2 was the key point. One of the main advantages
of the iron catalyst system is that it did not hydrogenate the
aromatic ring in HMF, which was beneficial to avoid the
formation of overhydrogenation products and ring-opening
products. However, the lewis acidity of the iron catalyst will
catalyze the polymerization of the furanic compounds to humins,
causing a still unsatisfactory carbon balance during the reaction.
stream of 5% H2/Ar with a total flow rate of 50 mL min–1
.
Experimental procedure
The catalytic HDO of HMF was carried out using a 50 mL Zr alloy
autoclave provided by Anhui Kemi Machinery Technology Co., Ltd. For a
typical procedure, 0.5 mmol HMF, 100 mg heterogeneous iron catalyst,
and 20 mL solvent were added into the autoclave with a quartz lining.
After purging the reactor with H2, the reaction was conducted with 4 MPa
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
H2 (room temperature) at 240 oC for 12 h with a stirring speed of 800 rpm.
After reaction, the gaseous products were analyzed by using a Fuli 9790
II Gas Chromatograph with a thermal conductivity detector (TCD). The
liquid products were analyzed by using both gas chromatography and
Gas chromatograph-Mass spectrometry.
GC-MS analyses were performed on an Agilent 7890 Gas
Chromatograph equipped with a DB-WAXETR 30 m × 0.25 mm × 0.25
μm capillary column (Agilent) or a HP-5MS 30 m × 0.25 mm × 0.25 μm
capillary column (Agilent). The GC was directly interfaced to an Agilent
5977 mass selective detector (EI, 70 eV). The following GC oven
temperature programs were used: 40 °C hold for 1 min, ramp 5 °C/min to
a temperature of 120 °C, and then ramp 10 °C/min to 240 °C and hold for
5 min.
[21] M. Chatterjee, T. Ishizaka, H. Kawanami, Green Chem. 2014, 16,
1543–1551.
[22] A. S. Nagpure, N. Lucas, S. V. Chilukuri, ACS Sustainable Chem. Eng.
2015, 3, 2909-2916.
[23] A. S. Nagpure, A. K. Venugopal, N. Lucas, M. Manikandan, R.
Thirumalaiswamy, S. Chilukuri, Catal. Sci. Technol. 2015, 5, 1463-1472.
[24] J. Luo, L. Arroyo‐Ramírez, R. J. Gorte, AlChE J. 2015, 61, 590-597.
[25] J. J. Shi, Y. Y. Wang, X. N. Yu, W. C. Du, Z. Y. Hou, Fuel 2016, 163,
74-79.
[26] A. B. Gawade, M. S. Tiwari, G. D. Yadav, ACS Sustainable Chem. Eng.
2016, 4, 4113–4123.
[27] G. H. Wang, J. Hilgert, F. H. Richter, F. Wang, H. Bongard, B.
Spliethoff, C. Weidenthaler, F. Schüth, Nature Mater. 2014, 13, 293-
300.
[28] J. Luo, H. Yun, A. V. Mironenko, K. Goulas, J. D. Lee, M. Monai, C.
Wang, V. Vorotnikov, C. B. Murray, D. G. Vlachos, P. Fornasiero, R. J.
Gorte, ACS Catal. 2016, 6, 4095–4104.
Acknowledgements
[29] J. Luo, J. D. Lee, H. Yun, C. Wang, M. Monai, C. B. Murray, P.
Fornasiero, R. J. Gorte, Appl. Catal. B 2016, 199, 439-446.
[30] J. M. R. Gallo, D. M. Alonso, M. A. Mellmer, J. A. Dumesic, Green
Chem. 2013, 15, 85-90.
This work was supported by Science Foundation of China
University of Petroleum, Beijing (No. 2462014YJRC037,
C201604).
[31] B. Saha, C. M. Bohn, M. M. Abu-Omar, ChemSusChem, 2014, 7, 3095-
3101.
Keywords: biomass • heterogeneous catalysis • 5-
[32] S. Nishimura, N. Ikeda, K. Ebitani, Catal. Today, 2014, 232, 89-98.
[33] J. Luo, L. Arroyo-Ramirez, J. F. Wei, H. Yun, C. B. Murray, R. J. Gorte,
Appl. Catal. A 2015, 508, 86-93.
hydroxymethylfurfural • hydrodeoxygenation • iron
[34] T. S. Hansen, K. Barta, P. T. Anastas, P. C. Ford, A. Riisager, Green
Chem. 2012, 14, 2457-2461.
[1]
[2]
[3]
G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044– 4098.
A. Corma, S. Iborra, A. Velty, Chem. Rev. 2007, 107, 2411 –2502.
R. J. van Putten, J. C. van der Waal, E. D. De Jong, C. B. Rasrendra, H.
J. Heeres, J. G. de Vries, Chem. Rev. 2013, 113, 1499-1597.
R. L. Liu, J. Z. Chen, X. Huang, L. M. Chen, L. L. Ma, X. J. Li, Green
Chem. 2013, 15, 2895-2903.
[35] A. J. Kumalaputri, G. Bottari, P. M. Erne, H. J. Heeres, K. Barta,
ChemSusChem 2014, 7, 2266 – 2275.
[36] G. Bottari, A. J. Kumalaputri, K. K. Krawczyk, B. L. Feringa, H. J.
Heeres, K. Barta, ChemSusChem 2015, 8, 1323-1327.
[37] Y. B. Huang, M. Y. Chen, L. Yan, Q. X. Guo, Y. Fu, ChemSusChem
2014, 7, 1068 – 1070.
[4]
[5]
[6]
[7]
[8]
H. L. Wang, T. S. Deng, Y. X. Wang, Y. Q. Qi, X. L. Hou, Y. L. Zhu,
Bioresour. Technol. 2013, 136, 394-400.
[38] M. Y. Chen, C. B. Chen, B. Zada, Y. Fu, Green Chem. 2016, 18, 3858-
3866.
Y. Román-Leshkov, C. J. Barrett, Z. Y. Liu, J. A. Dumesic, Nature, 2007,
447, 982-985.
[39] X. Kong, Y. F. Zhu, H. Y. Zheng, F. Dong, Y. L. Zhu, Y. W. Li, RSC Adv.
2014, 4, 60467–60472.
B. Girisuta, L. P. B. M. Janssen, H. J. Heeres, Green Chem. 2006, 8,
701-709.
[40] X. Kong, Y. F. Zhu, H. Y. Zheng, X. Q. Li, Y. L. Zhu, Y. W. Li, ACS
Catal. 2015, 5, 5914−5920.
F. Geilen, T. vom Stein, B. Engendahl, S. Winterle, M. A. Liauw, J.
Klankermayer, W. Leitner, Angew. Chem. 2011, 123, 6963-6966;
Angew.Chem. Int. Ed. 2011, 50, 6831-6834.
[41] X. Kong, R. X. Zheng, Y. F. Zhu, G. Q. Ding, Y. L. Zhu, Y. W. Li, Green
Chem. 2015, 17, 2504-2514.
[9]
J. P. Ma, Z. T. Du, J. Xu, Q. H. Chu, Y. Pang, ChemSusChem 2011, 4,
51-54.
[42] Y. F. Zhu, X. Kong, H. Y. Zheng, G. Q. Ding, Y. L. Zhu, Y. W. Li, Catal.
Sci. Technol. 2015, 5, 4208-4217.
[10] N. K. Gupta, S. Nishimura, A. Takagaki, K. Ebitani, Green Chem. 2011,
13, 824-827.
[43] X. M. Xiang, J. L. Cui, G. Q. Ding, H. Y. Zheng, Y. L. Zhu, Y. W. Li,
ACS Sustainable Chem. Eng. 2016, 4, 4506−4510.
[44] P. P. Yang, Q. N. Xia, X. H. Liu, Y. Q. Wang, Fuel 2017, 187, 159-166.
[45] B. F. Chen, F. B. Li, Z. J. Huang, G. Q. Yuan, Appl. Catal. B 2017, 200,
192-199.
[11] J. J. Pacheco, M. E. Davis, Proc. Natl. Acad. Sci. USA 2014, 111,
8363-8367.
[12] T. Buntara, S. Noel, P. H. Phua, I. Melián-Cabrera, J. G. de Vries, H. J.
Heeres, Angew. Chem. 2011, 123, 7221-7225; Angew.Chem. Int. Ed.
2011, 50, 7083-7087.
[46] L. L. Yu, L. He, J. Chen, J. W. Zheng, L. M. Ye, H. Q. Lin, Y. Z. Yuan,
ChemCatChem 2015, 7, 1701-1707.
[13] T. Thananatthanachon, T. B. Rauchfuss, Angew. Chem. 2010, 122,
6766-6768; Angew.Chem. Int. Ed. 2010, 49, 6616-6618.
[14] M. Chidambaram, A. T. Bell, Green Chem. 2010, 12, 1253–1262.
[15] S. De, S. Dutta, B. Saha, ChemSusChem 2012, 5, 1826-1833.
[16] J. Jae, W. Q. Zheng, R. F. Lobo, D. G. Vlachos, ChemSusChem 2013,
6, 1158 – 1162.
[47] M. Lefèvre, E. Proietti, F. Jaouen, J. P. Dodelet, Science, 2009, 324,
71-74.
[48] G. Wu, K. L. More, C. M. Johnston, P. Zelenay, Science, 2011, 332,
443-447.
[49] R. V. Jagadeesh, A. E. Surkus, H. Junge, M. M. Pohl, J. Radnik, J.
Rabeah, M. Beller, Science, 2013, 342, 1073-1076.
[50] R. V. Jagadeesh, H. Junge, M. Beller, ChemSusChem 2015, 8, 92-96.
[51] T. Stemmler, A. Surkus, M. Pohl, K. Junge, M. Beller, ChemSusChem
2014, 7, 3012-3016.
[17] J. Jae, W. Q. Zheng, A. M. Karim, W. Guo, R. F. Lobo, D. G. Vlachos,
ChemCatChem 2014, 6, 848 – 856.
[18] D. Scholz, C. Aellig, I. Hermans, ChemSusChem 2014, 7, 268 – 275.
[19] Y. H. Zu, P. P. Yang, J. J. Wang, X. H. Liu, J. W. Ren, G. Z. Lu, Y. Q.
Wang, Appl. Catal. B 2014, 146, 244-248.
[52] J. Li, J. L. Liu, H. J. Zhou, Y. Fu, ChemSusChem 2016, 9, 1339-1347.
[53] S. Sitthisa, W. An, D. E. Resasco, J. Catal. 2011, 284, 90-101.
[53] R. A. Rajadhyaksha, S. L. Karwa, Chem. Eng. Sci. 1986, 41, 1765-
1770.
[20] L. Hu, X. Tang, J. X. Xu, Z. Wu, L. Lin, S. J. Liu, Ind. Eng. Chem. Res.
2014, 53, 3056−3064.
[54] R. V. Jagadeesh, H. Junge, M. Beller, Nat. Commun. 2014, 5, 4123.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
[55] X. J. Cui, Y. H. Li, S. Bachmann, M. Scalone, A. Surkus, K. Junge, C.
Topf, M. Beller, J. Am. Chem. Soc. 2015, 137, 10652−10658.
[56] R. A. Rajadhyaksha, S. L. Karwa, Chem. Eng. Sci. 1986, 41, 1765-
1770.
[57] R. V. Jagadeesh, T. Stemmler, A. E. Surkus, H. Junge, K. Junge, M.
Beller, Nat. Protoc. 2015, 10, 548 –557.
This article is protected by copyright. All rights reserved.

10.1002/cssc.201700105
ChemSusChem
Entry for the Table of Contents
FULL PAPER
Selective catalytic
Jiang Li,* Jun-ling Liu, He-yang Liu,
hydrodeoxygenation of 5-
hydroxymethylfurfural to 2,5-
Guang-yue Xu, Jun-jie Zhang, Jia-xing
Liu, Guang-lin Zhou, Qin Li, Zhi-hao Xu,
and Yao Fu*
dimethylfuran over heterogeneous
iron catalysts is demonstrated in batch
and continuous flow fixed-bed
reactors. The iron catalyst exhibited
excellent stability, which is probably
due to the well reserved active
species and the pore structure of the
iron catalyst in the presence of H2.
Page No. – Page No.
Selective Hydrodeoxygenation of 5-
Hydroxymethylfurfural to 2,5-
Dimethylfuran over Heterogeneous
Iron Catalysts
This article is protected by copyright. All rights reserved.