Selective hydrodeoxygenation of 5-hydroxymethylfurfural to 2,5-dimethylfuran on Ru-MoOx/C catalysts

Article abstract of DOI:10.1039/c7ra00605e

Selective hydrogenation of 5-hydroxymethylfurfural (HMF) has potential application in high quality biofuels. Herein, the catalytic hydrodeoxygenation (HDO) of HMF to 2,5-dimethylfuran (DMF) was investigated using bi-functional Ru-MoO_x/C catalyst prepared by initial wetness impregnation. The high dispersion and electronic transfer between Ru and MoO_x were demonstrated by a series of characterization techniques. During this HDO process, the synergy effect between metallic Ru and acidic MoO_x species in the Ru-MoO_x/C catalyst plays an essential role in obtaining maximized target product DMF (79.4%) via effective aldehyde group hydrogenation by Ru followed by dehydration over MoO_x. This work also elucidated that DMF production proceeded through two distinct pathways: the 2,5-hydroxymethyl furan intermediate was preferable by the aldehyde group hydrogenation of HMF over the Ru-MoO_x/C catalyst. Over MoO_x/C catalyst, comparatively, 5-methyl furfural was the key intermediate by direct hydrogenolysis of the hydroxyl group in HMF. This kind of catalyst is stable for the first two runs by maintaining the target product yield. After the third run, the catalyst showed deactivation gradually but could be almost completely recovered after regeneration by H₂ reduction.

Full text of DOI:10.1039/c7ra00605e

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Selective hydrodeoxygenation of
5-hydroxymethylfurfural to 2,5-dimethylfuran on
Cite this: RSC Adv., 2017, 7, 16311
Ru–MoOx/C catalysts†
a
b
Yue Yang, Qiying Liu, Dan Li,b Jin Tan,b Qi Zhang,b Chenguang Wangb
*
b
*
and Longlong Ma
Selective hydrogenation of 5-hydroxymethylfurfural (HMF) has potential application in high quality biofuels.
Herein, the catalytic hydrodeoxygenation (HDO) of HMF to 2,5-dimethylfuran (DMF) was investigated using
bi-functional Ru–MoOx/C catalyst prepared by initial wetness impregnation. The high dispersion and
electronic transfer between Ru and MoOx were demonstrated by a series of characterization techniques.
During this HDO process, the synergy effect between metallic Ru and acidic MoOx species in the Ru–
MoOx/C catalyst plays an essential role in obtaining maximized target product DMF (79.4%) via effective
aldehyde group hydrogenation by Ru followed by dehydration over MoOx. This work also elucidated that
DMF production proceeded through two distinct pathways: the 2,5-hydroxymethyl furan intermediate
was preferable by the aldehyde group hydrogenation of HMF over the Ru–MoOx/C catalyst. Over MoOx/
C catalyst, comparatively, 5-methyl furfural was the key intermediate by direct hydrogenolysis of the
hydroxyl group in HMF. This kind of catalyst is stable for the first two runs by maintaining the target
product yield. After the third run, the catalyst showed deactivation gradually but could be almost
completely recovered after regeneration by H2 reduction.
Received 15th January 2017
Accepted 2nd March 2017
DOI: 10.1039/c7ra00605e
rsc.li/rsc-advances
much attention, due to the oxygen-containing DMF being an
ideal renewable oxygen liquid fuel according to its higher energy
1. Introduction
density, higher octane number and higher boiling point
compared to the traditional cellulose based ethanol.22,23
With the decrease in fossil resources, the demand for exploring
new sustainable energy to produce chemicals and fuel is urgent
for the future.1–3 Renewable and sustainable lignocellulose is
composed of cellulose (40–50%), hemicellulose (25–35%) and
lignin (15–20%), and can be used as feedstock for value-added
chemicals/fuels production, which has attracted great atten-
tion during the last decade.4,5 However, the typical biomass-
derived compounds usually have high oxygen-content and
could not be used for fuel directly.6,7 Therefore, an additional
hydrogenation step is necessary to complete or selectively
remove the oxygen atoms in the compounds.8,9
5-Hydroxymethylfurfural (HMF), one of the important plat-
form compounds from cellulose,10–12 can be further converted
into a series of valuable high-quality chemicals/fuels, such as g-
valerolactone, levulinic acid, oxidative 2,5-furandicarboxylic, 2,5-
dimethylfuran and so on.13–17 Among those chemicals/fuels, 2,5-
dimethylfuran (DMF)19–21 by HMF hydrogenolysis has attracted
Catalysts played the key role in producing DMF with high
yields from HMF, lignocellulosic biomass and the derivatives.
In 2007, Dumesic et al.1 (Table 1, entry 1) rstly reported
a catalytic strategy using a biphasic system to extract HMF, and
a separated process using 6.8 bar of H2 over CuRu/C catalyst at
220 ꢀC to produce 71% yield of DMF. Under the same reaction
conditions, Binder and Raines25 (Table 1, entry 8) achieved 49%
DMF yield by hydrogenolysis of crude HMF which obtained
from corn stover. Chidambaram and Bell2 (Table 1, entry 7)
added acetonitrile as auxiliaries and obtained 32% DMF yield
over Pd/C catalyst in ionic liquids. For improving the yield of
DMF, Rauchfuss et al.27 (Table 1, entry 6) utilized HCOOH as the
hydrogen donor and obtained 95% DMF yield. However,
HCOOH is corrosive and environment unfriendly, which leads
to apply restrictedly.7 Recently, researchers reported to employ
adding transitional metal additives which can form Lewis acid
to obtain high DMF yield by adsorption and activation the
hydroxyl and aldehyde group in HMF. And the metal sites were
active for hydrogenation while the acidic sites for promoting
dehydration of CH2–OH group.28,29 Zu et al.24 (Table 1, entry 4)
used Ru/Co3O4 to produce 93.4% DMF yield at 130 ꢀC aer 24 h
and Huang et al.26 (Table 1, entry 5) reached up to 96% DMF
aDepartment of Chemistry, University of Science and Technology of China, Hefei
230026, P. R. China
bKey Laboratory of Renewable Energy, Guangzhou Institute of Energy Conversion,
Chinese Academy of Sciences, Guangzhou 510640, P. R. China. E-mail: mall@ms.
giec.ac.cn; liuqy@ms.giec.ac.cn; Tel: +86-20-3702-9721, +86-20-87057673
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c7ra00605e
ꢀ
yield over Ni-W2C/AC at 180 C aer 3 h in THF. Judging from
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 16311–16318 | 16311

View Article Online
RSC Advances
Paper
Table 1 Results of direct conversion of biomass into DMF
Order
Substrate
Medium
Catalyst
Condition
Yield%
References
1
2
3
4
5
6
7
8
HMF
HMF
HMF
HMF
HMF
HMF
Glucose
Cellulose
1-Butanol
n-Butyl alcohol
THF
THF
THF
THF + HCOOH
EMIMCl
DMA-LiCl/ EMIMCl
Cu/Ru/C
Ru/C
Zn/Pd/C
Ru/Co3O4
Ni–W2C/AC
H2SO4 + Pd/C
12-MPA + Pd/C
CrCl2 + HCl + Cu/Ru/C
220 ꢀC, 10 h
260 ꢀC, 1.5 h
150 ꢀC, 8 h
130 ꢀC, 24 h
180 ꢀC, 3 h
70 ꢀC, 15 h
120 ꢀC, 3 h
220 ꢀC, 10 h
71
60.3
85
93.4
96
95
1
18
22
24
26
27
2
32
49
25
the results, the bi-functional catalyst was an ideal pattern for at 100 ꢀC to remove excess water. The solid was dried at 100 ꢀC
HMF hydrogenolysis, but these process involved either long overnight and calcined under a N2 atmosphere at 400 ꢀC for 4 h.
reaction time or high H2 pressure and the interaction between Aer the preparation of Ru/C catalyst, a second impregnation
metal and acid sites was not clearly. Therefore, there is a need to with (NH4)6Mo7O24 solution by the similar procedure
develop a new-type metal–acid catalyst to conversion of HMF to mentioned above. Prior to the experiment, the dried sample was
DMF in mild condition and explore the synergy effect between reduced in owing H2 (40 ml minꢁ1) at 350 C for 4 h.
ꢀ
metal and acid sites.
On the basis of the above-mentioned research, we designed
the metal–acid Ru–MoOx/C catalyst by step wetness impregna-
tion for HMF hydrogenolysis to DMF in n-butyl alcohol. Due to
2.3. Catalyst performance
Hydrogenolysis of HMF was carried out in a stainless steel reactor
with a Teon inner. In a typical experiment, a mixture of 0.5 g
the different functional groups in the HMF, such as C]O, CH2–
HMF dissolved in 40 ml n-butyl alcohol was loaded to the auto-
OH, furan ring, hydrogenation of HMF to DMF needs C]O
clave, followed by adding 0.1 g of freshly reduced 5% Ru–10%
MoOx/C catalyst (reduced at 350 ꢀC for 4 h). The reaction system
was sealed, ushed by N2 to remove air residue, pressurized the
H2 pressure to 1.5 MPa, and then heated to a certain temperature
hydrogenation to the corresponding hydroxyl groups and
subsequent C–OH groups hydrogenolysis. Under the optimum
condition (180 ꢀC, 1.5 MPa, 1 h), the selectivity of DMF was
79.8%. In addition, the characterization of catalyst was con-
for a certain time. Aer the reaction, the reactor was cooled to
ducted to explain the interaction and electronic transfer
room temperature immediately in a water bath. The liquid
between Ru and MoOx, and the cause of catalyst deactivation. In
products was collected and qualitatively analyzed by a GC-MS
addition, various reaction parameters was modulated to inves-
instrument with an Agilent 7890A-5975C and quantitatively
tigate the pathway for HMF hydrogenolysis and the synergistic
determined by GC with SHIMADZU GC-2014C gas chromato-
graph equipped with a FID and a HP-INNOWAX column.
The product yields were calculated based on internal stan-
effect of the metal and acid sites.
2. Experimental
2.1. Materials
dard with using cyclohexanone as the internal standard. The
conversion (conv.) of HMF and the selectivity of product i (Si)
were dened as:
HMF (98%) and 5-MF (99%) was supplied by J&K chemical Co.
Ltd. DMF (99%) and RuCl3$3H2O was purchased from Shanghai
Aladdin Reagent Co. Ltd. 2,5-Hexanediol (98%) and 2,5-hex-
anedione (98%) was sourced by Macklin Reagent Co. Ltd. 2,5-
C0 ꢁ Ct
Conv:ðHMFÞ ¼
ꢂ 100%
C0
dihydroxymethylfuran
(95%),
(5-methyl-2-furyl)methanol
Ci ꢂ 6
(97%), (cis-tetrahydrofuran-2,5-diyl)di-methanol (95%) and
(NH4)6Mo7O24$4H2O, 2,5-dimethyltetrahydrofuran (98%) was
obtained from Maya Reagent, Energy chemical, Shu ya Reagent,
Tianjin Kaida chemical and TCI Co. Ltd respectively. All the
reagents was used without further purication.
Si ¼
ꢂ 100%
ðC0 ꢁ CtÞ ꢂ ai
here, C0 refers to the initial concentration of HMF and Ct means
the mole concentration of nal HMF aer reaction; Ci indicates
the concentration of product i, and ai represents the carbon
numbers of product i.
2.2. Catalyst preparation
2.4. Catalyst characterization
The catalyst was prepared by step-wise impregnation method.
In a typical synthesis, active carbon (40–60 mesh) was chose as X-ray photoelectron spectroscopy (XPS) was conducted on
support materials. Firstly it was pretreated by nitric acid a Thermo Fisher Scientic Escalab 250 Xi photoelectron spec-
(0.6 mol lꢁ1) at 110 ꢀC for 5 h, then ltered until cooled to room trometer, the values were revised by C 1s (284.8 eV). The specic
temperature, and dried at 100 ꢀC overnight. RuCl3$3H2O was surface area was measured by N2 isothermal (77 K) adsorption
dissolved in deionized water and a certain amount of active on Mesoporous ASAP-2010 automated system. The tested
ꢀ
carbon was added to a desired Ru loading by weight, and stirred samples was outgassed at 200 C for 8 h prior to analysis. The
16312 | RSC Adv., 2017, 7, 16311–16318
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
surface area was calculated by BET method. The morphology Ru–MoOx/C catalysts, and the results were shown in Fig. 1. For
and energy spectrum analysis of catalysts were investigated by 5% Ru/C catalyst, two reduction peaks at 100 ꢀC and 230 ꢀC were
transmission electron microscopy (TEM, JEM-2100F) and observed, these peaks were the reduction of RuOx species due to
scanning electron microscopy (SEM, SU-70). Temperature pro- the different interactions with support. The rst peak was
grammed reduction (TPR) equipped with a TCD detector was assigned to the weak interaction of Ru with support, which
used to measure the reduction behavior under H2. For H2-TPR could be easily reduce. While, the second peak was assigned to
experiments, 100 mg sample was laid into a quartz tube and the reduction of RuOx species which was strongly interaction
pretreated by Ar at 300 ꢀC for 1 h, then cooling to 50 ꢀC, tꢁh1e with support.30–32 the Ru catalytic activity can promote the break
catalysts were heated to 600 ꢀC at the rate of 5 ꢀC min
.
of C–C bond on the surface of active carbon, which leaded to the
Temperature programmed desorption of NH3 (NH3-TPD) uphill of TCD signal at 500 C.32,33 The prole of 10% MoOx/C
ꢀ
experiments were conducted on AutoChem II 2920 to analyze catalyst presented a main peak and a shoulder peak centered at
ꢀ
ꢀ
the acid content. The catalysts were purged by He at 673 K prior 450 C and 550 C, respectively. It was attributed to the reduc-
to adsorption, then cooling to 373 K to saturate with NH3 and tion of Mo6+ with octahedral coordination to Mo4+ at 450 ꢀC,
sweep by He to remove the physical absorbed NH3 for 30 min, while the H2 consumption peaks located at 550 ꢀC belonged to
whereaer, the samples were heated to 673 K at the speed of 10 the reduction of tetrahedrally coordinated Mo species to lower
K minꢁ1. All of the catalysts which needed to reduce were pre- oxidation state species.34–37 Over the 5% Ru–10% MoOx/C cata-
ꢀ
treated under H2 stream for 4 h at 350 C and cooled to room lyst, the small reduction peak at 180 ꢀC was attributed to RuOx
temperature, then passivated by 0.6% O2/N2 gas for 4 h.
to Ru0.38 However, the main Mo reduction peak located at
350 ꢀC and it was lower than the 10% MoOx/C catalyst (450 ꢀC).
These results suggested that Ru facilitated to the reduction of
Mo species,39 in which H2 was adsorbed on Ru and dissociated
to highly reactive hydrogen atoms.40 When Ru and MoOx species
was close on the support, the H2 spillover facilitated the
reduction of MoOx.39,41–43 This behavior increased a possible
electronic transfer between Ru and MoOx.39
3. Results and discussion
3.1. Catalyst characterization
Hydrogen temperature-programmed reduction (H2-TPR) was
conducted to investigate the reduction properties of the current
To verify the interaction between Ru and MoOx, the surface
chemical state of 5% Ru/C, 10% MoOx/C and 5% Ru–10%
MoOx/C were conducted by X-ray photoelectron spectroscopy
(XPS) characterization. Because of 5% Ru/C and 5% Ru–10%
MoOx/C catalyst were liable to oxidization, they were reduced by
in situ technology to detect the chemical state of elements. Due
to the photoemission line of Ru 3d was overlapped with C 1s,44
Ru 3p3/2 line was used to analyze the surface species. As shown
in Fig. 2a, for the fresh 5% Ru–10% MoOx/C catalyst, the
binding energy of 463.2 eV was ascribed to RuO2,45 and shied
to relatively low binding energy aer reduction by H2 (reduced
5% Ru–10% MoOx/C catalyst), assigning to a reduction of RuO2
to metal Ru. In addition, over the reduced 5% Ru/C catalyst, the
binding energy of metal Ru shied to higher binding energy
compared to the reduced 5% Ru–10% MoOx/C catalyst, which
Fig. 1 H2-TPR profiles of (a) 5% Ru/C, (b) 5% Ru–10% MoOx/C and (c)
10% MoOx/C catalysts.
Fig. 2 XPS spectra of Ru–MoOx/C catalysts: (a) Ru 3p and (b) Mo 3d. 5% Ru–10% MoOx/C fresh: the catalyst had not been reduced by H2; 5% Ru–
10% MoOx/C, 10% MoOx/C and 5% Ru/C reduced: the catalysts were reduced by H2 under 350 ꢀC; 5% Ru–10% MoOx/C used: the catalyst was
reduced by H2 under 350 ꢀC and then used at the reaction condition.
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 16311–16318 | 16313

View Article Online
RSC Advances
Paper
indicated that there have some electrons was transferred from catalyst. The MoOx species was close contact with the Ru moiety
MoOx to Ru.39 On the other hand, the chemical status of Mo 3d in this two metallic catalyst, which induces the electron trans-
was measured and the results were shown in Fig. 2b. The fresh ferring from MoOx to Ru. This result is well consistent with the
5% Ru–10% MoOx/C sample presented only one doublet located XPS measurement.39
at 232.6 eV which was assigned to Mo6+ species (MoO3).35 The
Surface acidity and acidic strength distribution of Ru–MoOx/
reduced 5% Ru–10% MoOx/C catalyst revealed three doublets, C with different MoOx loading was evaluated by NH3-TPD
and the values at 232.0 eV and 229.6 eV were contributed to the (Fig. S1†). The NH3-TPD proles of 5% Ru/C, 5% Ru–5% MoOx/
Mo5+ and Mo4+ formation during H2 reduction.46,47 The binding C, 5% Ru–10% MoOx/C, and 5% Ru–20% MoOx/C consisted
energy of Mo4+ over 5% Ru–10% MoOx/C reduced catalyst was a wider shoulder peak in the temperature range of 390–480 K,
shied to le compared to 10% MoOx/C catalyst, which was and these peaks could be attributed to NH3 adsorb on the weak
attributed to the fact that some electrons were transferred from acid site of the catalyst surface.51 For the Ru/C catalyst, almost
MoOx to Ru. The low chemical valence of MoOx (x < 3) could no peak was observed (Table S1,† entry 4), with the increase of
form surface detects, which promotes hydrogenolysis of MoOx loading, the peak area augmentation and the strength
hydroxyl groups as Lewis acid.46 In addition, from Fig. 2a, the intensied which indicated that the total acid was enhanced
value of the used 5% Ru–10% MoOx/C catalyst reected this (Table S1†).
pattern existed two chemical state, namely Ru and RuO2.
Fig. 3 showed the TEM images of specied catalysts and the
diameter of particles was calculated by the soware Nano
3.2. HMF hydrogenolysis to DMF
Measure. All of the particles was included to calculate the The catalyst was tested in HMF hydrogenolysis to produce DMF.
average particle size in the visible range. For the 10% MoOx/C During the reaction process, 2,5-dihydroxymethylfuran (DHMF)
catalyst, the MoOx particles (Fig. 3e) presented the average and 5-methylfurfural (MF) were the main intermediates along
diameter of about 7.3 nm. The supported 5% Ru/C (Fig. 3d) with different reaction pathways. Meanwhile, the byproducts
catalyst showed a better dispersion by observing the smaller mainly contained the O-containing intermediates (e.g. 5-methyl-
average diameter of about 2.6 nm. Compared to the previous 2-furanmethanol (MFA), 2,5-dihydroxymethyl tetrahydrofuran
two monometallic catalysts, however, the average particle size of (DHMTHF)) and the over-hydrogenated products such as 2,5-
5% Ru–10% MoOx/C catalyst (Fig. 3a) was about 1.8 nm which dimethyl tetrahydrofuran (DMTHF), 2,5-hexanedione and 2,5-
was highly dispersed on the support of activation carbon, owing hexanediol by furan ring opening. The carbon balance of all
to the interaction between Ru and MoOx nanoparticles and detected products was about 65–86% maybe due to the forma-
facilitates the dispersion.48,49 The high dispersion was benet to tion of undetected molecules7,24,52 and polymers and/or cokes.53
hydrogenolysis.48,50 While, the few larger particles in Fig. 3a In the studied condition, the n-butyl alcohol had no reactivity
were MoOx species (distinguished by the HRTEM on the over catalysts.
corner), which had not been combined with Ru, and their
3.2.1 Ru–MoOx/C catalysts. For all the tested Ru–MoOx/C
particle size was about 17 nm, it must be pointed out that this catalysts, the HMF conversion was 100% while the DMF selec-
size can not reect the actual size of MoOx because of few tivity signicantly varied based on Ru/Mo ratios. When the 2%
statistical particles. The average size of 5% Ru–5% MoOx/C Ru–10% MoOx/C was used, the target DMF selectivity was
(Fig. 3b) and 5% Ru–20% MoOx/C (Fig. 3c) were 1.1 nm and 60.8%, simultaneously accompanied with the over-
2.5 nm respectively, and the 5% Ru–20% MoOx/C was showed hydrogenated products DMTHF (3.6%), DHMTHF (0.2%) and
poor dispersion, which revealed the MoOx content could impact the furan ring opening byproducts 2,5-hexanedione (1.4%) and
the size of particle and dispersion. Meanwhile, the high- 2,5-hexanediol (0.7%) by trace amounts (Table 2, entry 1).
resolution electron microscopy (HRTEM) image showed the Increasing the Ru loading to 5 wt%, the 5% Ru–10% MoOx/C
morphology of the two species (Fig. 3f) of 5% Ru–10% MoOx/C catalyst showed a better performance with 79.8% selectivity of
DMF owing to the activity of hydrogenation increased simulta-
neously (Table 2, entry 2). Keep the Ru loading as a constant,
when MoOx was introduced by different loadings, the DMF
selectivity gradually increased with increasing the MoOx
content to 10 wt%. When the Mo content exceeded 10 wt%,
however, the DMF selectivity decreased and the furan ring
opening byproducts (2,5-hexanedione and 2,5-hexanediol)
increased gradually, indicating that excessive acidic site, which
had been veried by NH3-TPD (Table S1†), was benet to open
the furan ring (Table 2, entry 2–4).
In order to understand the role of Ru and MoOx in this
hydrogenolysis reaction, 5% Ru/C and 10% MoOx/C were
prepared and tested under the same reaction condition. Over
5% Ru/C catalyst (Table 2, entry 5), the target DMF selectivity
Fig. 3 TEM images of Ru–MoOx/C catalysts: (a) reduced 5% Ru–10%
MoOx/C; (b) reduced 5% Ru–5% MoOx/C; (c) reduced 5% Ru–20%
was 54.1%, in addition, over-hydrogenated DMTHF and
DHMTHF selectivity was 6.7% and 10.1% respectively, which
MoOx/C; (d) reduced 5% Ru/C; (e) reduced 10% MoOx/C and (f) HRTEM
image of reduced 5% Ru–10% MoOx/C.
16314 | RSC Adv., 2017, 7, 16311–16318
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
Table 2 The effect of Ru/MoOx ratio in conversion of HMF to DMFa
HMF
conv.%
DMF
sel.%
DMTHF
sel.%
5-MF
sel.%
MFA
sel.%
DHMF
sel.%
DHMTHF
sel.%
2,5-Hexanedione
sel.%
2,5-Hexanediol
sel.%
Entry
Catalyst
1
2
3
4
5
6
2% Ru–10% MoOx/C
5% Ru–10% MoOx/C
5% Ru–5% MoOx/C
5% Ru–20% MoOx/C
5% Ru/C
100
100
100
100
100
39.9
60.8
79.8
71.8
72.6
54.1
3.9
3.6
3.4
2.6
3.1
6.7
8.5
—
—
—
—
—
37.5
—
—
—
—
—
0.4
—
—
—
—
—
2.6
0.2
0.4
1.1
0.2
10.1
—
1.4
0.7
0.4
1.9
2.2
6.9
0.7
1.0
0.9
0.9
0.5
10% MoOx/C
11.6
a
Reaction condition: 180 ꢀC, 1.5 MPa of H2, 1 h, HMF 0.5 g, n-butyl alcohol 40 ml and 0.1 g catalyst; the Ru and Mo loading were dened by the mass
fraction on support.
indicated the good hydrogenation and medium DMF selectivity 6.7% and 10.1% over the 5% Ru/C catalyst, while the selectiv-
for the sole Ru particle. However, if 10% MoOx/C (Table 2, entry ities were decreased to 3.4% and 0.4% over 5% Ru–10% MoOx/C
6) was introduced to the reaction, the HMF conversion was only catalyst. These results reected that the hydrogenolysis ability
39.8% and the DMF selectivity was 3.9%. Besides the target on Ru particle was restrained when added MoOx, which leaded
DMF, 37.5% of the main intermediate 5-MF was detected. to obtain more target DMF over Ru–MoOx/C catalyst. The rela-
Apparently, compared to Ru particle, MoOx/C exhibited a good tively week hydrogenolysis for furan ring was attributed to the
deoxygenation activity to crack the CH2–OH bond in HMF, while electron enrichment of Ru, which was similar to the result of
its hydrogenation activity to activate aldehyde group (–CHO) hydrogenolysis of n-butane by Scott et al.54 In the characteriza-
was rather limited. Based on the above results, the synergy of Ru tion section, we had interpreted the increase in the Ru elec-
and MoOx with proper ratios played a key role in obtaining high tronic density due to the electronic transfer from partially
DMF yield in hydrogenolysis of HMF.
reduced MoOx to Ru. A higher electronic density over Ru
This catalytic performance could also be explained by the increased the hydrogenolysis of the minimum electron density
electronic modication between Ru and MoOx. Compared the C–OH bond since Ru electron-donating properties.39 On the
results of 5% Ru/C (Table 2, entry 5) with 5% Ru–10% MoOx/C other hand, the mutual electronic impact maybe was attributed
(Table 2, entry 2), the selectivity of DMTHF and DHMTHF were to the synergistic activity. Considering the highest DMF yield
Table 3 The effect of different conditions in conversion of HMF to DMFa
PH
(MPa)
T
t
(h)
HMF
conv.%
DMF
sel.%
DMTHF
sel.%
5-MF
sel.%
MFA
sel.%
DHMF
sel.%
DHMTHF
sel.%
2,5-Hexanedione
sel.%
2,5-Hexanediol
sel.%
2
Entry
(ꢀC)
1
2
3
4
5
6
7
8
9
10
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
1.5
120
140
160
180
200
120
120
120
120
120
1
1
1
1
97.1
99.8
100
100
100
93.2
95.6
99.6
99.7
99.8
21.7
47.9
74.4
79.8
69.0
17.8
17.4
46.4
56.1
50.9
0.2
0.8
2.1
3.4
9.5
0.8
0.5
0.6
0.6
0.9
3.6
1.0
0.1
—
11.2
3.5
0.8
—
36.3
5.0
0.1
—
—
29.4
46.6
13.2
6.8
0.2
0.4
0.6
0.4
0.4
0.1
0.3
0.4
0.5
1.0
0.5
1.5
0.8
0.7
0.3
0.3
0.4
1.3
0.6
0.3
0.4
0.6
0.8
1.0
2.3
0.3
0.4
0.6
0.7
0.8
1
—
—
0.25
0.5
2
4
6
2.2
3.4
1.5
0.6
0.9
5.4
11.3
5.3
4.7
5.4
6.9
a
Reaction condition: HMF 0.5 g, n-butyl alcohol 40 ml and 0.1 g 5% Ru–0% MoOx/C catalyst; the Ru and Mo loading were dened by the mass
fraction on support.
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 16311–16318 | 16315

View Article Online
RSC Advances
Paper
obtained, the 5–10% RuMo/C catalyst was selected for the decreased and some byproducts including DHMTHF and
following controlled experiments. DMTHF were produced, which is generated by furan ring
3.2.2 The pathway for DMF production from HMF. To gain saturation via over-hydrogenation. The Fig. S2† showed the
insights into the reaction network, a special reaction was con- concentration of the products as a function of time.
ducted by investigating the effect of various reaction tempera-
Based on the above results, the aldehyde group in HMF was
ture and the effect of various reaction time on products rstly attacked to form DHMF over the bimetallic Ru–MoOx/C
ꢀ
distribution at 120 C. The intermediates were analyzed by GC catalyst. Since DHMF could be converted to MFA and further
and determined by GC-MS.
be hydrogenated to DMF through cleavage of CH2–OH bond,
As shown in Table 3 (entry 1–5), when the temperature was MFA was a key precursor to be converted into DMF. Interest-
120 ꢀC, only 21.7% of DMF selectivity was observed even though ingly, over 10% MoOx/C catalyst (Table 2, entry 6), the result
at the 97.1% of HMF conversion. The intermediates (5-MF, MFA revealed that the main intermediate was 5-MF and the selec-
and DHMF) and byproducts (DMTHF, DHMTHF, 2,5-hex- tivity was 37.5%, which indicates that a different reaction
anedione and 2,5-hexanediol) were detected with DHMF pathway was observed via the rst step of dehydroxylation
(36.3%) and MFA (11.2%) as the largest two. With increasing the in the presence of aldehyde group, comparing to the bimetallic
ꢀ
ꢀ
reaction temperature from 120 C to 180 C, the HMF conver- Ru–MoOx/C catalyst. The proposed reaction pathway was
sion and DMF selectivity were correspondingly increased from shown in Scheme 1.
97.1% to 100% and from 21.7% to 79.8%, respectively, which
According to the rules of concentration intermediates
was attributed to that the intermediate products was conver- change, the catalytic step can be described as follows. Initially,
ted to the target DMF gradually. Additionally, the over- Ru metal site could activated H2 and further dissociative
hydrogenated byproducts were also risen but with the total absorbed on the metal site. Meanwhile, the O atom of aldehyde
amounts below 6%. However, when the temperature was group in HMF was activated by the MoOx vacant orbital through
further increased to 200 ꢀC, the selectivity of DMF was Mo–O bond.55 Firstly, activated H attacked the aldehyde group
decreased to 69.0%. Meanwhile, the byproducts (DMTHF, 2,5- to produce DHMF, which was a key intermediate in HMF
hexanedione and 2,5-hexanediol) were signicantly increased to hydrolysis. Then, the MoOx acid site accelerated dehydrox-
12.5%.
The inuence of reaction time was listed in Table 3, entry 6– DMF was desorbed and catalyst was used into the next round.19
10. At the initial time of 0.25 h, the conversion of HMF was 3.2.3 Catalyst stability. In order to reect the true activity of
ylation to generate DMF via MFA intermediate. Aer that, the
reached to 93.2%, nearly completely converted. However, the catalyst, 5% Ru–10% MoOx/C was chosen to further investigate
selectivity of the target DMF was only 17.8% and DHMF (29.4%) the reusability with the exception of DMF from HMF. The
and MFA (5.4%) were presented as the main products. Pro- optimum condition (180 ꢀC, 1.5 MPa of H2 and 1 h) was used for
longing the reaction time to 0.5 h, the conversion of HMF was the HMF hydrogenolysis to DMF. Aer the rst run, catalyst was
increased to 95.6% and the selectivity of DMF was almost separated by centrifugation, washed ve times with n-butyl
constant. While the selectivity of DHMF (46.6%) and MFA alcohol, and then directly used for the next run under the same
(11.3%) reached to the maximum. As the reaction time was reaction condition. From Fig. 4, more than 70% of DMF selec-
further extended to 4 h, 56.1% of DMF was obtained while the tivity and 100% of HMF conversion could be achieved aer
slightly reduced MFA and DHMF intermediates were obtained. three runs, indicating that Ru–MoOx/C catalyst was effective
When the time was exceeded 4 h, the yield of DMF was slightly and stable for the HMF hydrogenolysis. Aer the third run, the
Scheme 1 Proposed reaction network for hydrogenolysis of HMF to DMF.
16316 | RSC Adv., 2017, 7, 16311–16318
This journal is © The Royal Society of Chemistry 2017

View Article Online
Paper
RSC Advances
Table 4 Physicochemical property of the catalysts
BET Total
Mean particle
surface areaa pore volumea size diameterb
Catalyst
(m2 gꢁ1
447.9
455.6
194.9
384.0
)
(cm3 gꢁ1
)
(nm)
Fresh 5% Ru–10%
MoOx/C
0.26
1.5
Used 5% Ru–10%
MoOx/C aer rst run
Used 5% Ru–10%
MoOx/C aer ve run
Regenerated aer
ve run
0.29
1.7
0.15
2.6
0.23
2.2
a
The BET surface area, pore volume were determined by N2 physical
adsorption. b Particle size was observed by TEM.
Fig. 4 Reusability of the Ru–MoOx/C catalyst. Reaction condition:
180 ꢀC, 1.5 MPa of H2, 1 h, HMF 0.5 g, n-butyl alcohol 40 ml and 5%
Ru–10% MoOx/C catalyst 0.1 g.
4 also revealed that the surface area and total pore volume was
increased aer regenerated, which was well consistent with the
reaction results.
selectivity of DMF decreased gradually, and for the h run, the
selectivity of DMF decreased to 42.2%, which indicated that the
catalyst deactivated during the tests. The loss of reactivity of
catalyst was probably attributed to deposit of carbonaceous
species like humins, which formatted during the reaction and
immersed in the active sites, as notied in previous
reports.26,56,57 To get insights into the deactivation of catalyst,
the fresh and used catalysts were characterized by SEM, TEM
and BET as follows. The result of SEM was presented at Fig. 5a–
c. Compared to the fresh catalyst, aer the rst run, the size of
particle had become slight larger. Aer ve times, the diameter
increased largely which was in line with the result of TEM
(Table 4) and along with the agglomeration of carbon particle.
From Table 4, the used sample surface area and pore volume
were absorbed compounds with large molecular.
The above characterization results suggested that the deac-
tivation of our sample maybe due to the agglomeration of
carbon support and the decrease of surface area. Then the used
ve times catalyst was regenerated at 350 ꢀC and the owing H2
(40 ml minꢁ1) for 4 h. As depicted by Fig. 4, the HMF conversion
and DMF yield was regained 99.7% and 79%, indicating this
catalyst could be completely recovered aer regeneration. Table
4. Conclusion
The bi-functional Ru–MoOx/C catalysts were prepared by initial
wetness impregnation and their catalytic performance was
tested in DMF production from hydrogenolysis of biomass-
derived HMF by nely modulating the metal–acid properties
of catalysts and reaction parameters. The synergy effect between
metal and acid sites plays a key role in obtaining high selectivity
of DMF (79.8%). The characterizations of catalyst were inter-
preted the interaction and electron transfer between Ru and
MoOx and the cause of deactivation, which was attributed to the
agglomeration of carbon support and the shrink of surface area
of catalyst. Hydrogenation of HMF were proceeded through two
pathways: for the Ru–MoOx/C catalyst, hydrogenation of C]O
group primarily occurred, followed by hydrogenolysis of two
CH2–OH bonds in DHMF to obtained the target DMF;
comparatively, hydrogenolysis of CH2–OH bonds in HMF was
rst taken place to 5-MF over MoOx/C catalyst.
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (no. 51576199, 50536009 and 51376185).
References
1 Y. Roman-Leshkov, C. J. Barrett, Z. Y. Liu and J. A. Dumesic,
Nature, 2007, 447, 982–985.
2 M. Chidambaram and A. T. Bell, Green Chem., 2010, 12, 1253.
3 Y. B. Yi, J. W. Lee and C. H. Chung, Environ. Chem. Lett., 2015,
13, 173–190.
4 X. L. Tong, Y. Ma and Y. D. Li, Appl. Catal., A, 2010, 385, 1–13.
5 G. Bottari, A. J. Kumalaputri, K. K. Krawczyk, B. L. Feringa,
H. J. Heeres and K. Barta, ChemSusChem, 2015, 8, 1323–1327.
6 M. Y. Chen, C. B. Chen, B. Zada and Y. Fu, Green Chem., 2016,
18, 3858–3866.
Fig. 5 SEM images of (a) the fresh 5% Ru–10% MoOx/C catalyst, (b) the
used of 5% Ru–10% MoOx/C catalyst after first run, (c) the used of 5%
Ru–10% MoOx/C catalyst after five runs. And TEM image of (d) the used
of 5% Ru–10% MoOx/C catalyst after first run, (e) the used of 5% Ru–
10% MoOx/C catalyst after five runs and (f) the regenerated of 5% Ru–
10% MoOx/C catalyst after five runs.
This journal is © The Royal Society of Chemistry 2017
RSC Adv., 2017, 7, 16311–16318 | 16317

View Article Online
RSC Advances
Paper
7 A. S. Nagpure, N. Lucas and S. V. Chilukuri, ACS Sustainable 33 L. A. Chen, Y. L. Zhu, H. Y. Zheng, C. H. Zhang, B. Zhang and
Chem. Eng., 2015, 3, 2909–2916. Y. W. Li, J. Mol. Catal. A: Chem., 2011, 351, 217–227.
8 M. Besson, P. Gallezot and C. Pinel, Chem. Rev., 2014, 114, 34 T. Bhaskar, K. R. Reddy, C. P. Kumar, M. R. V. S. Murthy and
1827–1870. K. V. R. Chary, Appl. Catal., A, 2001, 211, 189–201.
9 J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius and 35 J. Chang, T. Danuthai, S. Dewiyanti, C. Wang and A. Borgna,
B. M. Weckhuysen, Chem. Rev., 2010, 110, 3552–3599. ChemCatChem, 2013, 5, 3041–3049.
10 N. Shi, Q. Y. Liu, T. J. Wang, Q. Zhang, J. L. Tu and L. L. Ma, 36 L. J. Feng, X. G. Li, D. B. Dadyburjor and E. L. Kugler, J.
Chinese J. Phys. Chem., 2014, 27, 711–717. Catal., 2000, 190, 1–13.
11 A. Takagaki, M. Ohara, S. Nishimura and K. Ebitani, Chem. 37 V. R. Surisetty, I. Eswaramoorthi and A. K. Dalai, Fuel, 2012,
Commun., 2009, 41, 6276–6278. 96, 77–84.
12 J. B. Binder, A. V. Cefali, J. J. Blank and R. T. Raines, Energy 38 P. G. J. Koopman, A. P. G. Kieboom and H. Vanbekkum, J.
Environ. Sci., 2010, 3, 765–771. Catal., 1981, 69, 172–179.
13 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539– 39 C. E. Scott, P. Betancourt, M. J. P. Zurita, C. Bolivar and
554. J. Goldwasser, Appl. Catal., A, 2000, 197, 23–29.
14 I. T. Horvath, H. Mehdi, V. Fabos, L. Boda and L. T. Mika, 40 H. Cheng, M. Wen, X. Ma, Y. Kuwahara, K. Mori, Y. Dai,
Green Chem., 2008, 10, 238–242.
15 L. Hu, X. Tang, J. Xu, Z. Wu, L. Lin and S. Liu, Ind. Eng. Chem.
Res., 2014, 53, 3056–3064.
B. Huang and H. Yamashita, J. Am. Chem. Soc., 2016, 138,
9316–9324.
41 P. A. Sermon and G. C. Bond, Catal. Rev., 1973, 8, 211–239.
16 A. Villa, M. Schiavoni, S. Campisi, G. M. Veith and L. Prati, 42 K. Fabicovicova, M. Lucas and P. Claus, Green Chem., 2015,
ChemSusChem, 2013, 6, 609–612. 17, 3075–3083.
17 Y. L. Zhang, Z. M. Xue, J. F. Wang, X. H. Zhao, Y. H. Deng, 43 N. Li, Y. Zheng, L. Wei, H. Teng and J. Zhou, Green Chem.,
W. C. Zhao and T. C. Mu, RSC Adv., 2016, 6, 51229–51237. 2016, 18, 1821–2242.
18 J. H. Zhang, L. Lin and S. J. Liu, Energy Fuels, 2012, 26, 4560– 44 J. L. Fu, K. X. Yang, C. J. Ma, N. W. Zhang, H. J. Gai,
4567.
J. B. Zheng and B. H. Chen, Appl. Catal., B, 2016, 184, 216–
19 A. B. Gawade, M. S. Tiwari and G. D. Yadav, ACS Sustainable
Chem. Eng., 2016, 4, 4113–4123.
20 T. S. Hansen, K. Barta, P. T. Anastas, P. C. Ford and
A. Riisager, Green Chem., 2012, 14, 2457.
222.
45 Y. J. Weng, S. B. Qiu, C. G. Wang, L. G. Chen, Z. Q. Yuan,
M. Y. Ding, Q. Zhang, L. L. Ma and T. J. Wang, Fuel, 2016,
170, 77–83.
21 J. Jae, W. Zheng, A. M. Karim, W. Guo, R. F. Lobo and 46 V. M. L. Whiffen and K. J. Smith, Energy Fuels, 2010, 24,
D. G. Vlachos, ChemCatChem, 2014, 6, 848–856. 4728–4737.
22 B. Saha, C. M. Bohn and M. M. Abu-Omar, ChemSusChem, 47 M. Shetty, K. Murugappan, T. Prasomsri, W. H. Green and
´
2014, 7, 3095–3101. Y. Roman-Leshkov, J. Catal., 2015, 331, 86–97.
23 S. H. Zhong, R. Daniel, H. Xu, J. Zhang, D. Turner, 48 P. P. Yang, Q. N. Xia, X. H. Liu and Y. Q. Wang, J. Energy
M. L. Wyszynski and P. Richards, Energy Fuels, 2010, 24,
2891–2899.
24 Y. H. Zu, P. P. Yang, J. J. Wang, X. H. Liu, J. W. Ren, G. Z. Lu
and Y. Q. Wang, Appl. Catal., B, 2014, 146, 244–248.
25 J. B. Binder and R. T. Raines, J. Am. Chem. Soc., 2009, 131,
1979–1985.
Chem., 2016, 25, 1015–1020.
49 W. T. Luo, Y. Lyu, L. F. Gong, H. Du, M. Jiang and Y. J. Ding,
Chin. J. Catal., 2016, 37, 2009–2017.
50 X. B. Fu, H. Yu, F. Peng, H. J. Wang and Y. Qian, Appl. Catal.,
A, 2007, 321, 190–197.
51 Y. H. Chu, T. T. Zhang, J. X. Guo, C. Liu, H. Q. Yin, X. F. Zhu
and Y. J. Liu, J. Rare Earths, 2015, 33, 371–381.
26 Y. B. Huang, M. Y. Chen, L. Yan, Q. X. Guo and Y. Fu,
ChemSusChem, 2014, 7, 1068–1072.
¨
52 A. Iriondo, A. Mendiguren, M. B. Guemez, J. Requies and
27 T. Thananatthanachon and T. B. Rauchfuss, Angew. Chem.,
2010, 49, 6616–6618.
28 Y. Kusunoki, T. Miyazawa, K. Kunimori and K. Tomishige,
Catal. Commun., 2005, 6, 645–649.
29 X. Kong, R. X. Zheng, Y. F. Zhu, G. Q. Ding, Y. L. Zhu and
Y. W. Li, Green Chem., 2015, 17, 2504–2514.
J. F. Cambra, Catal. Today, 2017, 279, 286–295.
53 J. J. Shi, Y. Y. Wang, X. N. Yu, W. C. Du and Z. Y. Hou, Fuel,
2016, 163, 74–79.
54 C. E. Scott, T. Romero, E. Lepore, M. Arruebarrena,
P. Betancourt, C. Bolivar, M. J. Perezzurita, P. Marcano and
J. Goldwasser, Appl. Catal., A, 1995, 125, 71–79.
30 A. S. Nagpure, A. K. Venugopal, N. Lucas, M. Manikandan, 55 C. M. Andrade, N. M. Lima, S. J. Melo and C. A. M. de Abreu,
R. Thirumalaiswamy and S. Chilukuri, Catal. Sci. Technol.,
2015, 5, 1463–1472.
31 S. Z. Zheng, X. Y. Cao, Q. Zhou, S. H. Wang, G. S. Hu, J. Q. Lu,
Fuel, 2015, 147, 125–132.
56 S. Nishimura, N. Ikeda and K. Ebitani, Catal. Today, 2014,
232, 89–98.
M. F. Luo and Y. J. Wang, J. Fluorine Chem., 2013, 145, 132–135. 57 J. Z. Chen, F. Lu, J. J. Zhang, W. Q. Yu, F. Wang, J. Gao and
32 I. Rossetti, N. Pernicone and L. Forni, Appl. Catal., A, 2003,
248, 97–103.
J. Xu, Chemcatchem, 2013, 5, 2822–2826.
16318 | RSC Adv., 2017, 7, 16311–16318
This journal is © The Royal Society of Chemistry 2017