One-step Conversion of Fructose to Furfuryl Alcohol in a Continuous Fixed-bed Reactor: The Important Role of Supports

Article abstract of DOI:10.1002/cctc.201900157

One-step catalytic conversion of fructose to furfuryl alcohol (FOL) is not only a challenge task but also to open a new pathway for conversion of hexose into useful chemicals. The favorable yield (41.6 %) of FOL can be directly obtained from fructose over the tandem of HY zeolite and Pd/Al₂O₃. However, the Pd/montmorillonite (Pd/MMT) was deactivated for production of FOL. The proportion of available Pd active sites (i. e. facets, corners and edges) vary with different supports, and the strong metal support interaction (SMSI) plays the pivotal role on the catalytic performance. For Al₂O₃ support, exposure of Pd (111) plane allows the hydroxymethylfurfural (HMF) to bind on the Pd (111) surface via planar arrangement. In this stance, the aldehyde group of HMF is readily eliminated by decarbonylation. Conversely, for Pd/MMT, the decarbonylation is completely inhibited due to the HMF absorbs on corner sites in perpendicular orientation. The XPS studies indicate that the density electron of Pd nanoparticles (NPs) could be manipulated by various supports, and the electron-donating effect of Al₂O₃ was observed. The collaborative effects between available sites and density electron of Pd are observed with respect to the different binding modes of HMF on the surface of Pd. The electronegative Pd NPs facilitate the decarbonylation of HMF as a consequence of the different adsorption strength between furan ring and aldehyde group.

Full text of DOI:10.1002/cctc.201900157

Accepted Article
Title: One-step Conversion of Fructose to Furfuryl alcohol in a
Continuous Fixed-bed Reactor: The Important Role of Supports
Authors: Yueqing Wang, Xiaohai Yang, Guoqiang Ding, Hongyan
Zheng, Xianqing Li, Yongwang Li, and Yulei Zhu
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: ChemCatChem 10.1002/cctc.201900157
Link to VoR: http://dx.doi.org/10.1002/cctc.201900157
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
One-step Conversion of Fructose to Furfuryl alcohol in a
Continuous Fixed-bed Reactor: The Important Role of Supports
[
a, b]
[a, b]
[c]
[d]
[c]
[a,
Yueqing Wang,
Xiaohai Yang,
Guoqiang Ding, Hongyan Zheng, Xianqing Li, Yongwang Li,
c]
[a, c]
and Yulei Zhu,*
Abstract: One-step catalytic conversion of fructose to furfuryl
alcohol (FOL) is not only a challenge task but also to open a new
pathway for conversion of hexose into useful chemicals. The
favorable yield (41.6%) of FOL can be directly obtained from
chemicals, furfuryl alcohol (FOL), which can supplement the
fossil products in industrial fields, have been identified as the
platform chemicals of interest. In particular, as starting feedstock
[
4]
for furan resins, adhesives and synthetic fibers , the FOL is
produced from the selective hydrogenation of furfural, and about
70 % of global furfural was consumed for production of FOL.
The global demand of FOL is still increasing with the increasing
concern of environmental problems. Currently, the FOL is only
2 3
fructose over the tandem of HY zeolite and Pd/Al O . However, the
Pd/montmorillonite (Pd/MMT) was deactivated for production of FOL.
The propotion of available Pd active sites (i.e. facets, corners and
edges) vary with different supports, and the strong metal support
interaction (SMSI) plays the pivotal role on the catalytic performance. produced from xylose or xylan by a multi-step catalytic process
[
5]
For Al
2
O
3
support, exposure of Pd (111) plane allows the
in plant scale .
Reducing of the oxygen content is the fundamental
challenge to upgrading biomass-derived chemical to useful
hydroxymethylfurfural (HMF) to bind on the Pd (111) surface via
planar arrangement. In this stance, the aldehyde group of HMF is
readily eliminated by decarbonylation. Conversely, for Pd/MMT, the
decarbonylation is completely inhibited due to the HMF absorbs on
corner sites in perpendicular orientation. The XPS studies indicate
that the density electron of Pd nanoparticles (NPs) could be
manipulated by various supports, and the electron-donating effect
[
6]
products . Therefore, based on the furfural and HMF, numerous
attempts were devoted to explore catalysts and catalytic system
. So far the reported routes for the production of FOL have
been mostly based on the selectivity hydrogenation of furfural,
moreover, various copper-based catalysts have been confirmed
an effective materials as a result of the preference of Cu to bind
[
7]
2 3
of Al O was observed. The collaborative effects between available
[
8]
sites and density electron of Pd are observed with respect to the
different binding modes of HMF on the surface of Pd. The
electronegative Pd NPs facilitate the decarbonylation of HMF as a
consequence of the different adsorption strength between furan ring
and aldehyde group.
of C=O over C=C . Compared with hydrogenation,
decarbonylation of hydroxymethylfurfural (HMF) by noble metal
catalysts is the most competitive route due to transform
abundant C6 sugar into FOL. Starting with HMF, Pd-based
catalysts are known to be capable of selectivity decarbonylation
[7a, 7b, 9]
of aldehyde moiety by scission of C-C bond
. Although
deoxygenation efficiency of HMF and furfural is high, thermal
unstable nature of these platform chemicals make it difficult to
separate it from solvent. Especially, the high boil point (291.5 °C
at 760 mmHg) of HMF significantly limits its application.
With increasing of concern about environmental problems
and cost-efficiency, the concept of “one-pot” reaction, using
multifunctional catalyst or catalyst combination, have been
Introduction
Catalytic conversion lignocellulosic biomass towards
valuable chemicals has drawn the intense interest, as it could
alleviate the dependence of fossil resource as well as improve
the environmental problems . As
lignocellulosic biomass is composed with cellulose (40-50 %),
hemicellulose (25-35 %) and lignin (15-20 %) . Some valuable
chemicals, including furfural, hydroxymethylfurfural, furfuryl
alcohol, tetrahydrofurfuryl alcohol, 2-methylfuran and 2,5-
dimethylfuran, can be obtained from lignocellulosic biomass and
[
1]
a renewable resource,
developed to directly convert monosaccharide (fructose, glucose
[10]
and xylose) to furan derivatives in batch or fixed-bed reactor
In particular, J. H. Dai
.
[2]
[10e]
and co-worker developed a batch
process combined dehydration and decarbonylation, in which
the fructose could be directly converted to FOL (yield 40.6 %).
[11]
Additionally, it was shown by Cui. et al that the conversion of
[
3]
its units by catalytic approach . Among all biomass-derived
xylose into FOL using Hβ and Cu/Zn/Al as catalysts, and 87.2 %
yield of FOL was directly obtained in fixed-bed reactor. In batch
tandem reactions, it is difficult to convert sugars on acidic
catalyst without degradation over the other coexisting catalyst. In
addition, the recovery of mixture catalysts is cumbersome.
However, one-step convert fructose to FOL within
continuously fixed-bed reactor has received little attention.
[
a]
Dr. Y. Wang, Dr. X. Yang, Prof. Y. Zhu, Prof. Y. Li
State Key Laboratory of Coal Conversion
Institute of Coal Chemistry
Chinese Academy of Sciences
Taiyuan 030001 (P.R. China)
E-mail: zhuyulei@sxicc.ac.cn
Dr. Y. Wang, Dr. X. Yang,
University of Chinese Academy of Sciences
Beijing, 100049 (P.R. China)
Prof. Y. Zhu, Prof. Y. Li, Dr. G. Ding, X. Li
Synfuels China Co. Ltd.
Beijing (P.R. China)
Developing
a continuously catalytic process to transform
[
[
[
b]
c]
d]
fructose to FOL is not only the challenging task, but also to open
a new way for conversion of hexose (i.e. fructose and glucose)
to valuable chemicals. Compared with batch process, integration
of tandem reaction of dehydration and decarbonylation in fixed-
bed reactor could substantially avoid side reaction.
Dr. H. Zheng
In this work, fructose is selected as model compound, as it
can be produced from lignocellulosic biomass. A green, eco-
Department of chemistry and chemical engineering
Taiyuan institute of technology
Taiyuan 030008 (P.R. China)
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
friendly and continuous process is explored to directly produce
FOL. Combined with the zeolites and Pd-based catalysts,
conversion of fructose into FOL is achieved in continuously
fixed-bed reactor, eliminating the need for separation step (see
Scheme 1 and Supporting Information S1). Unsatisfactorily, in
most reported examples dealing with the production of FOL from
HMF or carbohydrates, the toxic dioxane, which is hazardous to
reactor using mixture of GBL and water as solvent. The
presence of water is inevitable as it is beneficial solvent for
dissolving of sugars. The relative results have been listed in
Table 1, performance of all catalysts was evaluated at the
completely conversion of substrates. The HY and USY
contained cavities were preferential to choose because it
[
16]
exhibits high selectivity of HMF than Hβ and ZSM-5
.
[
7b, 7c, 10e, 12]
environment and operators, was used as solvent
.
Compared with USY, the desirable yield (52.2 %) of HMF was
achieved over HY with more little by-products (Entry 1-2).
Despite the highest yield can be achieved using A-35 under mild
condition, the amount of by-product was also increased (Entry 3).
Additionally, furfural, as the major by-product, could be easily
eliminated in next decarbonylation step by converting to volatile
furan (boiling point 31.3 °C).
[a]
Table 1. The dehydration of fructose over various catalysts.
Yield (%)
Scheme 1. The route for production of FOL from fructose, FOL= furfuryl
Temp
°C )
Conv.
(%)
Entry
Catal.
HMF
FAL
FA
LA
alcohol, HMF = hydroxymethylfurfural
(
1
2
3
4
5
6
7
HY
USY
A-35
HY
140
140
70
99.9
99.9
99.9
97.6
99.9
99.9
99.9
52.2
40.3
61.6
42.2
42.1
44.9
23.2
12.3
2.3
1.6
2.6
/
/
Therefore, in view of green and eco-friendly, the γ-butyrolactone
13.1
5.4
7.1
/
(GBL) is used as green solvent to replace the toxic dioxane
because GBL was cheaper, lower volatility and less toxicity. The
most important is that the GBL could be produced from
/
8.6
16.2
/
[
13]
130
150
80
biomass . Zeolite (HY) behaves as the effective active sites to
catalyze the dehydration of fructose to HMF, subsequently, the
FOL is selectively produced over Pd-catalysts under mild
reaction conditions. In addition, since the first example
associated with strong metal support interaction (SMSI) was
HY
2.3
3.9
7.9
0.5
17.3
31.5
A-35
A-35
90
/
[
14]
reported by S. J. Tauster
at 1978, the SMSI is frequently
viewed as a crucial issue for supported-metal catalysts, even the
SMSI is identified as a decisive factor for the distribution of
-1
[
a] WHSV = 0.018 h , carrier gas flow = 25 mL/min, atmospheric pressure,
sugar concentration: 2 wt %, GBL contained 6 wt % water, 4.0 g catalyst
was loaded for each runs. A-35 = Amberlyst 35,
[
15]
products . Interestingly, in the presented work, the SMSI was
also observed and it could govern the direction of selectivity.
The present results provide solid evidence that the supports not
only influenced the density electron of Pd nanoparticles (NPs)
but also changed the exposure of available active sites (i.e.
corners, edges and facets) other than only providing a substrate
upon for dispersion of metal. Therefore, the favorable yield
HMF = hydroxymethylfurfural, FAL = furfural, FA = formic acid, LA = levulinic
acid.
The effects of temperature also were investigated, the
decreased tendency was observed in all cases, in which the
degradation of products became significant at elevated
temperature. In particular, the HMF was greatly degraded into
levulinic acid over A-35 (Entry 6-7). Compared with A-35, more
fructose was converted into furfural over HY at high temperature
(41.6 %) of FOL was continuous obtained from fructose over the
tandem of HY zeolite and Pd/Al . The modification of available
2 3
O
active sites of Pd NPs could influence the absorption behavior of
reactants. These Pd catalysts, which are loaded on the suitable
supports, exhibit high efficiency selectivity in decarbonylation.
(
Entry 4-5). Based on the above investigation, the HY and A-35
will be selected as desirable candidate for production of HMF in
the continuous progress.
Results and Discussion
Combination
continuous system
dehydration
and
decarbonylation
in
Dehydration of fructose
Initially, various general solid acid catalysts, including
zeolites and acidic resin, were employed to convert fructose.
The dehydration of fructose was investigated in the fixed-bed
As fructose can be efficiently converted to HMF, a series of
Pd catalysts were evaluated for production of FOL in the tandem
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
reaction, with the data illustrated graphically in Figure 1. On the
comparison of the data over various Pd-based catalysts, the
interesting results can be observed. Using HY as dehydrated
catalyst, the highest yield of FOL (41.6 %) was directly achieved
for decarbonylation of HMF with high yield and stability.
Meanwhile, the production of 2-methylfuran, 2,5-dimethylfuran
and 5-methylfuran in the decarbonylation progress may arise
2
from the hydrogenolysis, in which the H source mainly derived
[
18]
from fructose over Pd/Al
O
2 3
in 3 % H
2
/N
2
(V/V) normal pressure,
from dehydrogenation of HMF (as shown in ESI†, Table S1).
These results clearly reveal that the supports act as the
important role in improving the selectivity of FOL. In order to
elucidate the important role of supports, the physicochemical
propeties of range catalysts have been summarized in Table 2.
The Pd dispersions of various catalysts were calculated based
on the CO adsorption measurement. The various dispersions of
Pd NPs can be achieved on the different supports. The
dispersions of Pd NPs were proportional to the particle size, but
it was not as a function of surface area of supports. The highest
while the FOL was no detected over Pd/MMT after running
about 2 h. In addition, the yield of FOL obtained over other
catalysts, such as Pd/TiO
2 2 2
, Pd/ZrO and Pd/SiO , was lower
than the Pd/Al . It is worthy note that some by-products,
2 3
O
including detectable and undetectable, are inevitably produced
in the dehydration section. The deactivation of Pd/MMT maybe
ascribe to these by-products. Therefore, the HMF was used as
initial substrate to exclude the effects of by-products. Even in the
case of using HMF as initial substrate, the poor yield of FOL was
also obtained over Pd/MMT (as shown in ESI†, Figure S1),
indicating that the poor performance of Pd/MMT is ascribe to its
intrinsic properties other than the by-products.
2 3
dispersion of 31.0 % can be achieved for the Pd/Al O with the
smallest particle size and the highest metallic surface. It is
expect that the high surface area of supports facilitate Pd NPs to
2
well disperse, but the high surface of SiO cannot prepare the
well dispersion of Pd NPs. There are clearly other parameters, in
addition to dispersion, that influence the activity and selectivity,
as can be observed by the catalytic performance of the Pd/MMT.
Table 2. Main physicochemical propeties of Pd-based catalysts.
Pd loading
wt %)
Metallic
Catal.
Pd
size
(nm)
(
S
m g )
(BET)
Dispersion
(%)
surface
area
2 -1 [c]
2
-1 [b]
[c]
(
[c]
[
a]
Normal
Actual
1.2
(m g )
Pd/Al
Pd/MMT
Pd/ZrO
Pd/TiO
Pd/SiO
2
O
3
1
1
1
1
1
126.6
90.9
3.6
6.2
5.9
5.4
5.7
31.0
17.8
18.7
20.3
19.7
138.4
79.6
83.5
90.7
87.9
1.3
2
0.8
11.9
Figure 1. Catalytic performance plots for production of FOL from fructose over
HY + Pd-catalysts. For each runs, decarbonylation temperature was 160°C,
2
1.1
24.0
-
1
2 2
WHSV: 0.018h (for fructose), 3 % H /N (V/V) gas flow: 25 mL/min, the ratio
of zeolites and Pd-based catalysts was 4:3 by weigh, and the sugar
concentration was 2 wt% contained 6 wt% water. FOL = furfuryl alcohol, MMT
2
1.1
380.9
=
montmorillonite.
[a] Determined by ICP-AES. [b] Measured by BET method. [c] Determined
by pulse CO-absorption method.Pd/MMT = Pd/montmorillonite
2 3
These results clearly imply that the performance of Pd/Al O and
Pd/MMT depend on its intrinsic properties other than the
experimental parameters. Despite the highest yield of HMF was
achieved on A-35, the undesirable yield (33.5 %) of FOL was
The FOL was not obtained over Pd/MMT with a dispersion of
7.8 % (Figure 1). These results indicate that the strong
interaction between Pd and supports plays the important role to
govern the selectivity of catalysts. In particular, the highest yield
1
obtained by the combination of A-35 and Pd/Al
2 3
O . The possible
reason is that the Pd-based catalyst was partly deactivated due
of FOL can be achieved over Pd/Al
deactivation in decarbonylation. Thus, the Pd/Al
2
O
3
, while the Pd/MMT is
and Pd/MMT
[
17]
to the leaching of sulfurated component from A-35 . On the
other hand, the low yield of FOL may arise from slower reaction
rate in decarbonylation section. On the constant WHSV for A-35,
this possible reason have been ruled out stem from the fact that
the yield of FOL was not increased with increase of retention
time in decarbonylation section. Conversely, the FOL was
further degraded into 2-methylfuran. Despite the HMF was not
detected in final products, the total yield of FOL and 2-
methylfuran was still low (as shown in ESI†, Table S2). The
most interesting is that the Pd/MMT cannot stably work for
2
O
3
were chosen to understand the SMSI in the following section.
Elucidation of Pd NPs available active sites
To elucidate the available active sites generated using the
different support, both Pd/Al O and Pd/MMT were evaluated
2 3
using CO as a probe molecule. The special adsorption sites in
the surface structure of Pd NPs were identified from the IR
frequency of the CO bond. As shown in the Figure 2, the spectra
2 3
decarbonylation section, but the Pd/Al O is desirable material
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
obtained for two catalysts exhibits noticeable differences in the
distribution of available active sites. Four types of adsorption
sites can be identified by special frequency of CO bond. The
adsorption band present at 2082, 2087 for Pd/Al O and
2 3
Pd/MMT should be generally assigned to CO linearly absorbed
on the corner sites of Pd NPs, and it is very prominent for
2 3
Pd/MMT but is only a weak peak for Pd/Al O . The broad peaks
at 1980 and 1969 include the contribution from bridging CO on
low-index Pd facets, and its shoulder peak at ~1930 should be
assigned to triply bridging CO species (3-fold hollow).
Figure 3. The different adsorptions of HMF on Pd surface
Conversely, the available sites that correspond to the
bridge-bonded and 3-fold CO absorption are able to bind the
1
HMF through both furan ring and the η (C) mode of aldehyde
group. This flat adsorption of HMF is beneficial to
[
19b, 20, 22]
decarbonylation of aldehyde group
. Therefore, the
superior activity of Pd/Al catalyst for production of FOL is
2 3
O
inferred to the decarbonylation of aldehyde group on the
accessible facets sites of Pd NPs. In contrast, the predominant
exposure of corner sites allows HMF to bind the Pd in a
perpendicular orientation, in which the decarbonylation of
aldehyde group is not readily carried out. It is worhty note that
the decline of activity may derived from adsorption of CO
molecule in reaction conditions. In order to investigate the
adsorption-desorption of CO, the temperature programmed
desorption (TPD) was carried out using CO as probe molecule
2 3
Figure 2. FTIR spectra from CO adsorption studies on Pd/Al O (red dashed
line) and Pd/MMT (solid black line)
(as shown in the ESI†, Figure S1). The desorption temperature
is 39.6 °C for Pd/MMT, 22.1 °C for Pd/Al . These results
2 3
O
The broad peak at 1811, only observed on Pd/Al
2 3
O , could be
indicate that the CO molecule could be completely desorbed
from the surface of catalysts at reaction temperature (160 °C ).
The deactivation of Pd/MMT is not to arise from the adsorption
of CO.
attributed to different types of CO species on the 3-fold hollow
[
12, 19]
sites
bonds, in comparison to those from linear-bonded CO, are much
higher for the catalyst prepared by Al support. In addition, the
. The relative intensities of the bridging adsorbed CO
2 3
O
Pd NPs loaded on the MMT have high proportion of the linear
adsorbed CO. It is clear that the various supports afford Pd NPs
with different surface characteristics. In other word, the Al
O
2 3
The XRD and TEM studies of Pd NPs
greatly facilitate Pd NPs to expose the facets Pd atoms with
small proportion of corners and different 3-fold sites, while the
Pd NPs loaded on montmorillonite (MMT) favors to expose more
corners Pd atoms with smaller facets and isolated-bonded sites
In order to understand the crystalline structure of metal NPs,
x-ray diffraction patterns of reduced Pd/Al O and Pd/MMT are
2 3
compared in Figure 4a. In addition to characteristic peaks of
MMT, the reduced Pd/MMT exhibits a very weak XRD peak at
2θ = 40.1° corresponding to the (111) plane of metallic Pd fcc
structure (JCPDS No. 05-0681). Whereas no clear peaks of
2 3
than Pd/Al O . It could be proposed that the different surface
sites are a consequence of the strong interaction between the
Pd NPs and supports. The HMF, which contains the furan ring
and pendant aldehyde group, could be adsorbed on these
available active sites via different orientations. The different
binding models of HMF have been shown in Figure 3, as
metallic Pd are observed in the case of Pd/Al
the weak and undistinguishable peaks of crystalline Pd indicate
that the small Pd NPs are homogeneously dispersed on Al O
3
2 3
O sample. Indeed,
2
[
20]
[23]
previous studies
proposed that the corner sites of Pd NPs
and MMT. Previous papers
mentioned that the MMT is
tend to bind furan ring in a perpendicular orientation, rather than
the flat across the surface. In this stance, the aldehyde group
composed by layers of negatively charged two dimensional
silicate sheets, and the silicate sheets are held together by
cationic species in the interlamellar space. Therefore, the small-
angle X-ray diffraction has been conducted to estimate the basal
spacings of MMT and Pd/MMT. The P (001), which indicates the
stacking space of silicate sheets, is chosen to determine the
intercalation of Pd NPs, and the basal spacings are calculated
by Bragg equation. As shown in the Figure 4b, the peak
reflecting the basal spacing of Pd-free MMT is located at 2θ =
2
adsorbs on the surface of Pd by the η (C, O) configuration,
[
19b, 21]
which is unfavorable for decarbonylation
.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
6.0° (basal spacing = 1.45 nm), which could be assigned to the
(001) planes.
Figure 5. The micrograph of the HAADF-STEM and HRTEM imagines. (a, b)
reduced Pd/Al O , (c, d) reduced Pd/MMT.
2 3
The high resolution TEM data was used to obtain the fast
Fourier transform (FFT) patterns. The Pd (111) plane with
interplanar spacing of 0.22 nm is confirmed by FFT patterns in
2 3
the Pd/Al O , despite the metallic Pd peaks cannot be identified
in XRD patterns (shown in Figure 5b inset). For the Pd/MMT
sample, the Pd (111) and Pd (220) planes are simultaneously
observed with interplanar spacing of 0.22 and 0.12, respectively
(shown in the Figure 5d inset). These results clearly suggest that
2 3
Figure 4. The XRD patterns of reduced samples a) Pd/Al O and Pd/MMT, b)
small-angle x-ray diffraction.
the available crystalline plane of Pd NPs can be manipulated by
SMSI. Correspondingly, the favorable yield of FOL can be
2 3
achieved over Pd/Al O with high propotion of facets sites, these
[
22]
For the Pd/MMMT samples, this peak position shifted to slightly
high angles (2θ = 6.9°) with a significant decline of the intensity
as compared to that the unloaded MMT. These observations
indicate that the Pd NPs are predominantly deposited on the
external surface of MMT, as also demonstrated by the peak at
Pd (111). The further evidence of crystalline structure could be
provided from the transmission electron microscopy (TEM). As
shown in the Figure 5, the Pd NPs are homogeneously
dispersed on the supports in spherical shape. As detected by
the HAADF-STEM imaging, the average size of Pd NPs is 2.9
results are consist with the previous studies
.
The chemical states of Pd NPs
As the surface state of catalyst is very crucial for the
catalytic behavior, the chemical status of the supported Pd
catalyst was further evaluated by in situ XPS. The palladium 3d
core level was found to be the best indicator of the near-surface
chemical state. Based on the XPS analysis, detailed information
about the chemical properties of Pd NPs with respect to the
different support can be convincingly declared.
2 3
nm for Pd/Al O , 2.4 nm for Pd/MMT (as shown in Figure 5a and
c). Significantly, there is no obvious aggregation of Pd NPs after
the reduction. The particle size obtained from TEM is smaller
than those measured from CO-absorption as a result from the
2 3
exposure of different available active sites on Al O and MMT.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
catalysts were carried out to evaluate its stability in the
conversion of fructose into FOL.
2 3
Figure 6. The XPS spectrum of reduced catalysts a) Pd/MMT and b) Pd/Al O
Figure 6 summarizes the state of Pd on different supports.
Specifically, for Pd/MMT, the binding energy at 335.4 eV and
3
40.4 eV are assigned to the 3d5/2 and 3d3/2 of zerovalent Pd,
2 3
Figure 7. Stability of the catalytic performance of HY + Pd/Al O
[
24]
respectively . In addition, the broad peak at around 337.5 eV is
[
25]
assigned to the PdO for Pd/MMT , indicating that the Pd/MMT
is difficult to be completely reduced under measurement
conditions. The strong metal support interaction (SMSI) could be
deduced from the shift of 3d peaks towards low binding
As shown in the Figure 7, no obvious decline in selectivity is
observed within the operating for the 40 h. In addition, the
activity of catalysts is also no decrease during the entire process.
These results clearly manifest that this tandem catalytic system
can steadily convert fructose into FOL with high selectivity.
[
26]
2 3
energy . With the Pd loading on the Al O , a 0.7 eV shift (334.7
eV) of 3d5/2 peak and a 0.4eV shift (334.0 eV) of 3d3/2 peak
are observed as compared with Pd/MMT, indicating that the Pd
[
27]
species are partially negative charged with high degree SMSI
.
The electronegative Pd reduce the strength of furan ring
adsorption on the Pd (111) surface, meanwhile, the
electropositive carbon of aldehyde group is strongly adsorbed on
Conclusions
Considering the eco-friendly and green aspects, a tandem
catalytic process has been explored to directly convert fructose
into FOL with respect to combination of zeolites and Pd-based
catalysts. The Pd particles were confirmed to be homogeneously
1
the Pd surface with the η (C) configuration. In this instance, the
decarbonylation is readily achieved, and these observations are
in agreement with our experimental results. Correspondingly, the
zerovalent Pd facilitate the hydrogenation of HMF rather than
loaded on the Al
scale. The average size of Pd was found to be 2.9 nm for
Pd/Al , 2.4 nm for Pd/MMT, which were prepared by simply
incipient wet impregnation method. The decarbonylation could
be effectively achieved on the Pd/Al combined with the
dehydration of fructose. Correspondingly, the favorable yield
41.6 %) of FOL can be continuously obtained starting from
2 3
O and montmorillonite (MMT) at the nano-
2
the decarbonylation, owing to the adsorption of η (C, O)-
hydroxymethylfurfural configuration. The poor yield of FOL
obtained over Pd/MMT may interpreted as evidence of the
difference density electron of Pd (as shown in ESI†, Table S1).
2 3
O
2 3
O
(
The stability test of catalytic system
fructose. The strong metal-support interaction (SMSI) is further
confirmed to play pivotal role in selectivity of catalysts. The
available sites and density electron of Pd NPs can be
manipulated with different supports. To be specific, the majority
Pd (111) facets, which are able to direct selectivity of the HMF
decarbonylation products, is exposed with small propotion of
The reaction temperature for the conversion fructose into FOL
have been studied (as shown in ESI†, Figure S2). For
production of FOL in decarbonylation section, the selectivity of
products exhibits the volcano trend with the increasing of
reaction temperature. The highest selectivity of FOL was
obtained at 160 °C. With the increasing of temperature, the
selectivity of FOL is decline on account of its degradation at high
temperature. The stability and durability of catalysts are the very
important aspects for liquid conversion various sugars into
2 3
corner sites on the Al O support. The inverse tendencies were
observed on the MMT, resulting in very poor selectivity of
decarbonylation. Additionally, the XPS studies point out that the
metallic Pd is negative charged owing to the electron donating of
Al O . Nevertheless, the Pd atoms exhibit zerovalent status on
2 3
MMT without negative charge. It is apparent that the
collaborative effect between available sites and density electron
2 3
useful chemicals. Long-term application of the HY and Pd/Al O
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
of Pd NPs is able to direct the catalytic performance. The
difference of available sites leads the HMF to adsorb on the Pd
surface by different orientation, moreover, the negative charged
Pd atoms improves the decarbonylation by repulsion of furan
ring over the complete aldehyde group. Conversely, the
zerovalent Pd atoms located on the MMT facilitate the aldehyde
group to adsorb on the corner sites of Pd with η (C, O)
configuration, in which the decarbonylation selectivity was
sharply decrease until to deactivation. Based on the above
2018, 10, 2326-2349; (d) M. J. Gilkey, B. Xu, ACS Catal. 2016, 6, 1420-
1436.
[
[
4]
5]
(a) A. Corma, S. Iborra, A. Velty, Chem.Rev. 2007, 107, 2411-2502; (b)
H. Li, S. Zhang, H. Luo, Mater. Lett. 2004, 58, 2741-2746; (c) P. D.
Vaidya, V. V. Mahajani, Ind. Eng. Chem. Res. 2003, 42, 3881-3885.
K. Yan, G. Wu, T. Lafleur, C. Jarvis, Renewable Sustainable Energy
Rev. 2014, 38, 663-676.
2
[6]
[7]
P. N. Vennestrom, C. M. Osmundsen, C. H. Christensen, E. Taarning,
Angew. Chem. Int. Ed. 2011, 50, 10502-10509.
(a) Q. Meng, D. Cao, G. Zhao, C. Qiu, X. Liu, X. Wen, Y. Zhu, Y. Li,
Appl. Catal. B Environ 2017, 212, 15-22; (b) F. M. Geilen, T. vom Stein,
B. Engendahl, S. Winterle, M. A. Liauw, J. Klankermayer, W. Leitner,
Angew. Chem. Int. Ed. 2011, 50, 6831-6834; (c) J. Mitra, X. Zhou, T.
Rauchfuss, Green Chem. 2015, 17, 307-313; (d) C. Wang, Z. Liu, L.
Wang, X. Dong, J. Zhang, G. Wang, S. Han, X. Meng, A. Zheng, F. S.
Xiao, ACS Catal. 2017, 8, 474-481; (e) S. Sitthisa, T. Pham, T.
Prasomsri, T. Sooknoi, R. G. Mallinson, D. E. Resasco, J. Catal. 2011,
experimental results and characterization data,
a strong
interaction (SMSI) between supports and Pd NPs has a crucial
influence on the catalytic performance, the nature of this
interaction is confirmed that the manipulation of available sites
and the change of the density electron of Pd NPs.
280, 17-27; (f) M. Lesiak, M. Binczarski, S. Karski, W. Maniukiewicz, J.
Rogowski, E. Szubiakiewicz, J. Berlowska, P. Dziugan, I. Witońska, J.
Mol. Catal. A: Chem. 2014, 395, 337-348.
Experimental Section
[
[
8]
9]
S. Sitthisa, T. Sooknoi, Y. Ma, P. B. Balbuena, D. E. Resasco, J. Catal.
The conversion of fructose to furfuryl alcohol was carried out in
continuous fixed-bed reactor, in which the reactor (600mm long, 12mm
i.d.) was packed with the catalysts (20-40 mesh) locating in the
thermostat segment of reactor. Before the run, the catalysts were pre-
2
011, 277, 1-13.
(a) W. Ouyang, A. Yepez, A. A. Romero, R. Luque, Catal. Today 2018,
08, 32-37; (b) G. J. S. Dawes, E. L. Scott, J. Le Nôtre, J. P. M.
Sanders, J. H. Bitter, Green Chem. 2015, 17, 3231-3250.
3
reduced in situ using a stream of pure H
h. The reaction solution was pumped into reactor using HPLC pump,
moreover, a stream of pure H or hybrid gas (H /N ) was introduced from
the top of reactor using the mass-flow controller.
2
(100mL/min) under 250 °C for 2
[10] (a) M. Braun, M. Antonietti, Green Chem. 2017, 19, 3813-3819; (b) W.
Ghezali, K. De Oliveira Vigier, R. Kessas, F. Jérôme, Green Chem.
2015, 17, 4459-4464; (c) Y. Liu, L. Zhu, J. Tang, M. Liu, R. Cheng, C.
2
2
2
Hu, ChemSusChem 2014, 7, 3541-3547; (d) Z. Z. Yang, J. Deng, T.
Pan, Q. X. Guo, Y. Fu, Green Chem. 2012, 14, 2986; (e) J. Dai, X. Fu,
L. Zhu, J. Tang, X. Guo, C. Hu, ChemCatChem 2016, 8, 1379-1385; (f)
J. Cui, J. Tan, X. Cui, Y. Zhu, T. Deng, G. Ding, Y. Li, ChemSusChem
The final products were collected in the gas-liquid separator. The
conversion of substrate was determined using HPLC (Agilent 1260)
equipped with a Shodex SH-1821 capillary column (300mm × 8 mm × 6
2016, 9, 1259-1262.
μm) and a refractive index detector. The mobile phase was 0.01M H
flowing at a rate of 0.6mL/min. The products were analyzed using GC
Agilent 7890A) equipped with HP-5 capillary column (30 m × 0.32 mm ×
.25 μm) and a flame ionization detector (FID). The standard solutions
2
SO
4
[11] V. V. Ordomsky, J. C. Schouten, J. van der Schaaf, T. A. Nijhuis, Appl.
Catal. A Gen 2013, 451, 6-13.
[
12] Q. Meng, C. Qiu, G. Ding, J. Cui, Y. Zhu, Y. Li, Catal. Sci. Technol.
016, 6, 4377-4388.
(
2
0
[
13] (a) D. W. Hwang, P. Kashinathan, J. M. Lee, J. H. Lee, U. h. Lee, J. S.
Hwang, Y. K. Hwang, J. S. Chang, Green Chem. 2011, 13, 1672; (b) G.
Budroni, A. Corma, J. Catal. 2008, 257, 403-408.
were used to obtain the calibration curves to calculate concentrations of
the compounds by external standard method.
[
[
14] S. J. Tauster, S. C. Fung, R. L. Garten, J. Am. Chem. Soc. 1978, 100,
170-175.
15] S. Li, Y. Xu, Y. Chen, W. Li, L. Lin, M. Li, Y. Deng, X. Wang, B. Ge, C.
Acknowledgements
Yang, S. Yao, J. Xie, Y. Li, X. Liu, D. Ma, Angew. Chem. Int. Ed. 2017,
5
6, 10761-10765.
The authors gratefully thank the financial supports of the
National Natural Science Foundation of China (No. 21875276),
Shanxi Science and Technology Innovation Team of Biofuels &
Chemicals, Institute of Coal Chemistry, Chinese Academy of
Sciences, Taiyuan, P R China.
[
16] (a) Y. Wang, G. Ding, X. Yang, H. Zheng, Y. Zhu, Y. Li, Appl. Catal. B
Environ 2018, 235, 150-157; (b) Y. Wang, X. Yang, H. Zheng, X. Li, Y.
Zhu, Y. Li, Molecular Catalysis 2019, 463, 130-139.
[17] X.L. Seoane, P. C. L'Argentiere, N.S. Figoli, A. Arcoya, Catal. Lett.
1992, 16, 137-148.
[
18] Q. Meng, H. Zheng, Y. Zhu, Y. Li, J. Mol. Catal. A: Chem. 2016, 421,
6-82.
7
Keywords: one-step • fructose • furfuryl alcohol • strong metal-
support interaction • decarbonylation
[
19] (a) T. Lear, R. Marshall, J. A. Lopez-Sanchez, S. D. Jackson, T. M.
Klapotke, M. Baumer, G. Rupprechter, H. J. Freund, D. Lennon, J.
Chem. Phys. 2005, 123, 174706; (b) S. M. Rogers, C. R. A. Catlow, C.
E. Chan Thaw, A. Chutia, N. Jian, R. E. Palmer, M. Perdjon, A.
Thetford, N. Dimitratos, A. Villa, P. P. Wells, ACS Catal. 2017, 7, 2266-
2274.
[
[
1]
2]
(a) L. C. Nhien, N. V. D. Long, S. Kim, M. Lee, Biotechnol Biofuels
017, 10, 81; (b) J. A. Melero, J. Iglesias, A. Garcia, Energy Environ.
2
Sci. 2012, 5, 7393-7420.
(a) J. B. Sluiter, R. O. Ruiz, C. J. Scarlata, A. D. Sluiter, D. W.
Templeton, J. Agric. Food. Chem. 2010, 58, 9043-9053; (b) R. Rinaldi,
F. Schüth, Energy Environ. Sci. 2009, 2, 610-626; (c) X. Li, P. Jia, T.
Wang, ACS Catal. 2016, 6, 7621-7640.
[20] S. H. Pang, J. W. Medlin, ACS Catal. 2011, 1, 1272-1283.
[21] S. T. Thompson, H. H. Lamb, ACS Catal. 2016, 6, 7438-7447.
[22] S. Wang, V. Vorotnikov, D. G. Vlachos, ACS Catal. 2014, 5, 104-112.
[23] (a) Q. Runliang Zhu , Qing Zhou,YunfeiXi,Jianxi Zhu,, H. He, Applied
Clay Science 2016, 123, 239-258; (b) B. Tyagi, C. D. Chudasama, R. V.
Jasra, Spectrochim Acta A Mol Biomol Spectrosc 2006, 64, 273-278.
[24] H. J. Chen, H. W. Liu, W. Liao, H. B. Pan, C. M. Wai, K. H. Chiu, J. F.
[
3]
(a) J. J. Bozell, G. R. Petersen, Green Chem. 2010, 12, 539-554; (b) L.
Hu, L. Lin, Z. Wu, S. Zhou, S. Liu, Renewable Sustainable Energy Rev.
2017, 74, 230-257; (c) K. Gupta, R. K. Rai, S. K. Singh, ChemCatChem
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201900157
FULL PAPER
Jen, Appl. Catal. B Environ 2012, 111-112, 402-408.
25] J. Han, H. Sun, J. Duan, Y. Ding, H. Lou, X. Zheng, Adv. Synth. Catal.
010, 352, 1805-1809.
26] T. Hengsawad, T. Jindarat, D. E. Resasco, S. Jongpatiwut, Appl. Catal.
A Gen 2018, 566, 74-86.
[27] M. G. v. Pitchon, H. Praliaud, Applied Catalysis 1990, 63, 333-343.
[
[
2
This article is protected by copyright. All rights reserved.