Bifunctional Mg?Cu-Loaded β-Zeolite: High Selectivity for the Conversion of Furfural into Monoaromatic Compounds

Article abstract of DOI:10.1002/cctc.201800450

Mg?Cu-loaded β-zeolite was prepared as an acid?base bifunctional catalyst for the conversion of furfural to monoaromatic compounds. It is found that this catalyst exhibited high selectivity towards benzene, toluene and xylenes (BTXs) production with anti-polycyclic-aromatic-hydrocarbon formation as well as anti-coking ability when compared with Cu-loaded β-zeolite and the parent β-zeolite. The product distribution indicated that addition of Cu species significantly promoted the deoxygenation and aromatization of an intermediate product of furan while Mg apparently suppressed the polyaromatization. Especially, an optimum loading amount of 0.5 wt%Mg-1 wt%Cu on β-zeolite was obtained, which showed lower catalytic deoxygenation temperature and interestingly, only benzene was detected in the liquid product at a reaction temperature over 700 °C. Also, 0.5 wt%Mg-1 wt%Cu/β-zeolite exhibited long-term stability for 10 cycles with 100 % of furfural conversion to monoaromatic compounds. The physicochemical properties of β-zeolite after Cu and Mg loadings were characterized using N₂ sorption, XRD, SEM-EDX, TEM, H₂-TPR, UV-Vis, XPS, NH₃-TPD and CO₂-TPD techniques. The significant changing of acidity and basicity of β-zeolite were found after Cu and Mg loadings, which should be the main factors for the improvement of activity, selectivity and stability of the developed catalyst.

Full text of DOI:10.1002/cctc.201800450

Accepted Article
Title: Bifunctional Mg-Cu-Loaded β-Zeolite: High Selectivity for the
Conversion of Furfural into Monoaromatic Compounds
Authors: Surachai Karnjanakom, Akihiro Yoshida, Asep Bayu, Irwan
Kurnia, Xiaogang Hao, Panya Maneechakr, Abuliti Abudula,
and Guoqing Guan
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.201800450
Link to VoR: http://dx.doi.org/10.1002/cctc.201800450
A Journal of
www.chemcatchem.org

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
Bifunctional Mg-Cu-Loaded β-Zeolite: High Selectivity for the
Conversion of Furfural into Monoaromatic Compounds
Surachai Karnjanakom, [a,b] Akihiro Yoshida,[a,b] Asep Bayu,[a,b] Irwan Kurnia,[a,b] Xiaogang Hao,[c]
Panya Maneechakr,[d] Abuliti Abudula,[b] and Guoqing Guan*[a,b]
Abstract: Mg-Cu-loaded β-zeolite was prepared as an acid–base
bifunctional catalyst for the conversion of furfural to monoaromatic
compounds. It is found that this catalyst exhibited high selectivity towards
benzene, toluene and xylenes (BTXs) production with anti-polycyclic-
aromatic-hydrocarbon formation as well as anti-coking ability when
compared with Cu-loaded β-zeolite and the parent β-zeolite. The product
distribution indicated that addition of Cu species significantly promoted
the deoxygenation and aromatization of an intermediate product of furan
while Mg apparently suppressed the polyaromatization. Especially, an
optimum loading amount of 0.5 wt%Mg-1 wt%Cu on β-zeolite was
obtained, which showed lower catalytic deoxygenation temperature and
interestingly, only benzene was detected in the liquid product at a
reaction temperature over 700 °C. Also, 0.5wt%Mg-1wt%Cu/β-zeolite
exhibited long-term stability for 10 cycles with 100% of furfural
conversion to monoaromatic compounds. The physicochemical
properties of β-zeolite after Cu and Mg loadings were characterized
using N2 sorption, XRD, SEM-EDX, TEM, H2-TPR, UV-Vis, XPS, NH3-
TPD and CO2-TPD techniques. The significant changing of acidity and
basicity of β-zeolite were found after Cu and Mg loadings, which should
be the main factors for the improvement of activity, selectivity and
stability of the developed catalyst.
the catalytic deoxygenation of bio-oil into aromatic
hydrocarbons.[3,4] Jae et al.[5] investigated the shape selectivity of
zeolite on the conversion of glucose into aromatics, and found
that glucose-derived vapors can easily diffuse into the pores of
ZSM-5 and ZSM-11, resulting in the formation of aromatics with
high yield. Park et al.[6] found that mass transfer rate was
improved when using mesoporous zeolites, which enhanced the
product selectivity to monocyclic aromatic hydrocarbons (MAHs)
such as BTXs and other alkylbenzenes. However, even though
these zeolites showed good catalytic activity for deoxygenation
with high selectivity toward aromatic hydrocarbons, a large
amount of undesirable polycyclic aromatic hydrocarbons (PAHs)
such as indenes and naphthalenes were also generated in the
upgraded bio-oil due to the further aromatization and alkylation
of MAHs.[7] Moreover, owing to the strong acidity of zeolite,
coking always takes place simultaneously via polymerization of
small molecule oxygenates, leading to the blocking of pores and
rapid deactivation.[8] Therefore, it is necessary to improve the
activity, selectivity and stability of zeolite for BTXs production by
avoiding PAHs and coke formations.
Introduction of transition metals into zeolite or other porous
materials to adjust the acidity and generate new acid sites were
found to be the effective way for upgrading of bio-oils and
increasing aromatic selectivity.[9,10] Mullen et al.[11] loaded Fe
species on HZSM-5 for upgrading of the bio-oil in catalytic fast
pyrolysis of biomass and found that the yield of aromatic
hydrocarbon was increased to a great extent. In our previous
study, Cu loaded β-zeolite was utilized for in-situ upgrading of
bio-oil from biomass pyrolysis, resulting in obvious increasing of
hydrocarbons amount in the final product.[12] Iliopoulou et al.[13]
reported that the addition of Co on ZSM-5 strongly enhanced
aromatization reactions, which resulted in an increase in the
selectivity towards aromatics in the upgraded bio-oil. Recently,
alkali and alkaline earth metals (AAEMs) species were found to
have coking resistance ability due to their strong basic
properties. Among them, MgO and CaO were considered as the
most attractive basic catalysts for upgrading of bio-oil because
they are abundant, stable and relatively inexpensive.[14]
Stefanidis et al.[15] compared the catalytic performances of MgO
with ZSM-5 for the upgrading of bio-oil, and found that MgO had
high ability for retarding the formations of PAHs and hard coke.
However, its surface area was much lower than that of zeolite so
that only a few active sites can be provided for the reactions. To
solve this problem, it is possible to disperse AAEMs species on
the supports such as zeolite with high surface area.
Introduction
Pyrolysis of lignocellulose biomass is one of the promising
alternatives to produce bio-oil.[1] However, the obtained bio-oil
always contains many oxygenated compounds, which make bio-
oil corrosive, unstable, immiscible and have low heating value.[2]
Hence, the original bio-oil should be upgraded in advance of
utilization as transportation fuels. Zeolite is generally used for
[a]
Dr. S. Karnjanakom, Prof. Dr. A. Yoshida, Dr. A. Bayu, I. Kurnia,
Prof. Dr. G. Guan
Energy Conversion Engineering Group, Institute of Regional
Innovation (IRI), Hirosaki University,2-1-3, Matsubara, Aomori 030-
0813, Japan. E-mail:guan@hirosaki-u.ac.jp
Dr. S. Karnjanakom, Prof. Dr. A. Yoshida, Dr. A. Bayu, I. Kurnia,
Prof. Dr. A. Abudula, Prof. Dr. G. Guan
Graduate School of Science and Technology, Hirosaki University, 1-
Bunkyocho, Hirosaki 036-8560, Japan
Prof. Dr. X. Hao
[b]
[c]
[d]
Department of Chemical Engineering
Taiyuan University of Technology
Taiyuan 030024, China
Prof. P. Maneechakr
In this study, a try was performed to co-load Cu and Mg on
β-zeolite as an acid–base bifunctional catalyst for upgrading of
bio-oil, especially for BTXs production from bio-oil. Here, in order
to understand the catalytic mechanism, furfural was selected as
the model compound of oxygenates since it is one of most
popular compounds in the bio-oils derived from the pyrolysis of
biomass.[16] Importantly, there are no report on the role and the
Department of Chemistry
Rangsit University
Pathumthani 12000, Thailand
Supporting information for this article is given via a link at the end of
the document.
1
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
nature of the active Cu and Mg species and their interaction for
selective conversion of furfural into BTXs. The Cu and Mg
loading amounts (wt%) were optimized and the surface area,
porosity, morphology, crystal structure, metal dispersion,
reduction behaviour, acidity and basicity of prepared catalysts
were characterized. The product distributions derived from
furfural conversion at different reaction temperatures and weight
hourly space velocities (WHSVs), the anti-coking and anti-PAHs
formation ability, and the long-term stability were systematically
investigated. The possible reaction routes for catalytic
conversion of furfural into aromatic hydrocarbons were proposed
resulted from the covering by excessive metal species or from the
sintering of metal particles on the surface of β-zeolite.[24,25]
Figure S2 shows XRD patterns of β-zeolite before and after
metal doping. The diffraction peaks corresponding to CuO was
observed in case that Cu loading amounts were higher than 1
wt% (Figure S2A). Interestingly, any diffraction peaks of MgO
did not appear with Mg loading amounts from 0.5 to 5 wt%
applied in this work (Figure S2B), indicating that Mg species
were well dispersed in the β-zeolite framework without any
accumulation and sintering during the calcination process.[27] As
Cu loading amount was increased to 3wt%, weak diffraction
based on the experimental results. To the best of our knowledge, peaks corresponding to CuO were observed at 35.5, 38.7, 48.7,
bifunctional Mg-Cu-loaded β-zeolite has not been reported in
literatures on catalytic deoxygenation and aromatization of
furfural. It is expected that this catalyst can be applied for the
upgrading of bio-oil in a practical process.
53.5 and 58.2°.[17] When the Cu loading amount was further
increased from 3 wt% to 10 wt%, the intensities of diffraction
peaks corresponding to CuO increased to some extent while the
CuO crystal size was also increased from 7.3 to 25.5 nm,
indicating that high Cu loading amount resulted in particle
accumulation.[27] In contrast, for Mg loading, the occurrence of
diffraction peaks at around 36.9 and 42.9°corresponding to MgO
were only found when the loading amount of Mg was as high as
10 wt% (Figure S2B). Moreover, it should be noted that the
loading of Cu and Mg with various loading amounts did not
change the crystalline structure of β-zeolite except slight
reduction of its diffraction peak intensity due to the covering of
metal spices on its surface.[28]
Figures S3 and S4 show the SEM-EDX mapping images of
Cu and Mg loaded β-zeolite after calcination in air. All metal-
loaded samples with Cu loading amount less than 10wt%
showed the similar structural morphology as that of the parent β-
zeolite. Moreover, no bulk oxide species of Cu and Mg were
formed on the surface of β-zeolite, suggesting that they were
well dispersed inside β-zeolite structure. EDX mappings
confirmed that Cu and Mg elements with loading amounts lower
than 5wt% were homogeneously dispersed on β-zeolite. For
10wt%Cu/β-zeolite, Cu-rich regions appeared, which should be
resulted from the existence of high Cu amount as well as the
large accumulation of copper oxide particles. Figure S5 shows
TEM images of Cu and Mg loaded β-zeolite catalysts after
calcination in air. Dark spots which correspond to copper oxides
were not observed on the catalyst with 1wt% Cu loading (Figure
S5A, E and F), while copper oxide nanoparticles observed inside
and outside β-zeolite structures in the cases of higher loading
amounts of copper than 3wt% (Figures S5B, C and D). However,
with the co-loading of Mg species, no MgO particles were
observed on the TEM images for all catalysts,indicating that the
Mg species were homogeneously dispersed on β-zeolite, and the
loaded Mg did not result in the significant formation of MgO
species with observable size on the β-zeolite. These results were
consistent with the XRD analysis results described above.
Results and Discussion
Catalyst Characterization
Table S1 shows the textural properties of parent β-zeolite and
metal-loaded β-zeolites. Interestingly, loading of Cu and Mg on β-
zeolite resulted in the obvious increase in specific surface area
and pore volume when compared with the parent β-zeolite.
Moreover, after co-loading with Mg on 1wt%Cu/β-zeolite, the
surface area was further increased, comparing with that loading
with 1wt%Cu only. This result is different from those reported ones,
in which the surface area and pore size were always decreased
by metal loading due to the metal sintering and/or the pore
blocking.[17,18] Figure S1 shows the N2 adsorption-desorption
isotherms of β-zeolite before and after metal loading. It can be
seen that the parent β-zeolite exhibited the type I sorption
isotherm based on the IUPAC classification for microporous
structure.[12] However, hysteresis characteristic appeared after
metal loading, suggesting that mesoporous structure was
generated.[19,20] Here, it is possible that these new extra pore
structures were formed due to ion-exchange, metal substitution
and calcination. Also, the small metal species might enter the
channel in β-zeolite, resulting in the formation of new surface area
within channel by the increase of inner surface roughness.[21,22]
Another possible reason may be due to the extraction of
aluminium species from the zeolite structure and reaccumulation
of aluminium hydroxide species on the zeolite framework which
possibly occurs at the slightly acidic condition (pH = ~4) of the
mixture solution of magnesium and copper nitrates.[23] In order to
confirm it, the surface area of the β-zeolite treated in the acidic
solution (pH = 4, adjusted by 0.1 M HNO3) was measured. As
shown in Table S1, although a little higher surface area was
observed for the acid-treated β-zeolite when compared with parent
β-zeolite, it was still lower than those of metal-loaded β-zeolites
such as 0.5wt%Cu/β-zeolite, 0.5wt%Mg-1wt%Cu/β-zeolite and
others. Thusly, it can be considered that the formation of extra
porous structure on the zeolite framework from the loading of
metal species should be the main reason for increasing catalyst
surface area. As such, loading of Cu and Mg on β-zeolite could
provide more active sites for the conversion of furfural molecules.
However, it should be noted that the surface area was decreased
to some extent in the case of overloading of metal species amount,
such as 10wt%Cu and 3wt%Mg loadings, which should be
Figure S6 shows TPR profiles of the catalysts after calcination
in air. One can see that a reduction peak appeared in the range of
about 120-260 °C for 3-10wt%Cu/β-zeolite samples (Figure S6A),
which correspond to the reduction of bulk CuO to metallic Cu.[22,29]
Moreover, with the increase in Cu loading amount, more H2 was
consumed (Table S1), and due to the change of interaction
between Cu species and β-zeolite, the reduction temperature was
shifted to some extent. It should be noted that no reduction peak
appeared in the case of 0.5wt%Cu loading, indicating that Cu
could be completely incorporated into the framework of zeolite
structure at a low loading amount. In the case of 1wt%Cu loading,
two reduction peaks were observed at temperatures of 180 °C and
2
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
amount from 1 to 3wt% resulted in the
increasing of acidity from 0.33 to 0.67
mmol/g (Table S2). As Cu loading amount
was over 3wt%, the density of weak acid
sites significantly increased. This might be
resulted from the increase in the amounts
of new Lewis acid sites derived from CuO
on zeolite. In addition, three NH3 desorption
peaks (weak, medium and strong acid
sites) appeared at 3wt% and 5wt%Cu
loading
amounts.
These
different
distributions of acid sites could be
attributed to the existence of different Cu
species such as bulk CuO, Cu+ and Cu2+
located at different places in zeolite
structure.[34] However, when the Cu loading
amount was increased to 10wt%, the
acidity decreased, probably due to the
existence of large amount of large CuO
particles. For Fig. 1B, the addition of only
Cu resulted in the creation of weak basic
sites on β-zeolite for CO2 desorption peak
at low temperature. Interestingly, co-
loading of Mg clearly resulted in the shifting
of CO2 desorption peaks from low to high
temperature range (Figure 1F), indicating
that Mg also had great effect on the
generation of new basic sites so that the
obtained Mg-Cu-loaded β-zeolite had acid–
base bifunctional properties and exhibited
new catalytic performance. Asthana et
al.[35] reported that the acidity of γ-Al2O3
was decreased due to the addition of MgO.
However, in this study, when Mg was co-
loaded on 1wt%Cu/β-zeolite, as shown in
Figure 1E, the densities of both weak and
strong acid sites were slightly increased. It
Figure 1. (A,C,E) NH3-TPD profiles and (B,D,F) CO2-TPD profiles
of Cu and Mg loaded on β-zeolite with various loading amounts.
is possible that the loaded Mg did not result in the significant
formation of MgO species on 1wt%Cu/β-zeolite. However, the
new Lewis acid site was obviously observed, especially at 500 °C,
suggesting that co-Mg loading could result in the change of
isolated Cu species or Cu particle size located on β-zeolite and
promotion of Cu species to bind with aluminate species in
zeolite.[30] Herein, the proton sites of zeolite could be easily ion-
exchanged by Mg2+ and Cu2+, which resulted in the regeneration
of new active sites.
400 °C, which should be attributed to the reduction of CuO to Cu0
or Cu2+ to Cu+ and from Cu+ to Cu0, respectively.[22] In the case of
Mg co-loading with 1wt%Cu, as shown in Figure S6B, the
reduction peak was shifted to the higher temperature, and with the
increase in Mg loading amount, the reduction temperature
became higher. It indicates that Mg species could promote Cu
species to bind with aluminate species in zeolite.[30]
Figure S7 shows UV-Vis diffuse reflectance spectra of the as-
prepared catalysts. One can see that all of samples displayed the
absorption band at 207 nm, which could be attributed to the charge-
transfer from O-to-metal, relating to the Cu+ and Cu2+ species
stabilized in the β-zeolite framework.[36] The broader absorption
band appeared with the increase in Cu loading amount from 3wt%
to 10wt%, indicating that more bulk CuO species were formed on
the β-zeolite structure.[37] Here, it is also possible that the increasing
of bulk CuO species may be one reason for promoting the PAHs
and coke formations. Moreover, as shown in Figure S7B, after co-
loading with Mg, the absorption band was shifted to higher
wavelength around 228 nm, indicating the co-existences of Cu2+
and Mg2+ species. With the increase in Mg loading amount, the
intensity of absorption band was obviously increased due to the
domination of more Mg2+ and Cu2+ species, resulting in the
generation of new active sites in β-zeolite.[38] XPS analysis was
The acid-base property of catalysts was also investigated by
TPD analyses. Figure 1 shows NH3-TPD and CO2-TPD profiles of
the as-prepared catalysts. As shown in Figure 1A, for the parent
β-zeolite, two NH3 desorption peaks at low temperature of about
≤280 °C and high temperature of about ≥280°C should be
attributed to the weak acid sites and strong acid sites
respectively.[31,32] The amount of total acid sites on the parent β-
zeolite was estimated to 0.29 mmol/g, which was quantified from
the peak area by calibration using the standard gas. The results
are also shown in Table S2. After 0.5wt% of Cu was loaded on β-
zeolite, the NH3 desorption peaks shifted to lower temperature
with the increase in their intensities. This should be the
contribution of copper species, which can be exchanged with the
protons of Brønsted acid sites, resulting in the obvious increase of
Lewis acid site amount.[33] Furthermore, the increase of Cu loading
3
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
also investigated on the 1wt% of Cu loaded and the Mg co-
loaded zeolite catalysts (Figure S8). As observed, the Cu 2p3/2
appeared at around 933.6 eV, which well ascribed to the
characteristics of the Cuδ+ (δ < 2) species.[36,39] Also, the
existence of Mg2+ on catalyst was confirmed with a binding
energy value at 1304 eV.[40] These results are also in agreement
with observations of TEM and H2-TPR.
Catalytic Conversion of Furfural
Figure 2 and Table S3 show the product distributions derived
from catalytic conversion of furfural at 600 °C over Cu/β-zeolite
calcined in air. Here, the overall product yield was quantified and
calculated based on carbon balance. When the parent β-zeolite
without Cu loading was employed as a catalyst, 96% of furfural
was converted, and furan and carbon monoxide were yielded as
the abundant products. This product distribution indicates that
the decarbonylation reaction of furfural mainly occurred on the
parent β-zeolite catalyst. In contrast, when 0.5wt%Cu/β-zeolite
was used, furfural was completely converted and 41% of
aromatic hydrocarbons were yielded (Figure 1B), indicating that
Cu loaded β-zeolite not only favoured the deoxygenation but
also promoted the aromatization, oligomerization and alkylation.
As indicated above (Figure 1 and Table S2), it should be
attributed to the increase in Lewis or weak acid sites on the Cu
loaded β-zeolite.[12] Then, further increase in Cu loading amount
resulted in the obvious reduction of CH4 and olefins yields due to
the promotion of aromatization and alkylation reactions. Here, it
should be noted that the main component in olefins was
ethylene.
Figure 3. (A) Product yields and (B) aromatic distributions in
liquid products obtained from catalytic deoxygenation of furfural
at a reaction temperature of 600 °C, a TOS of 1 h and a WHSV
of 1.2 h-1 using Mg-1wt%Cu/β-zeolite with various Mg loading
amounts.
From Figure 2, it is also found that the PAHs selectivity was
increased to some extent as Cu loading amount was increased
from 1wt% to 10wt%. Therefore, it is expected that the
generated olefins could be consumed for PAHs formation,
leading to the reduction of MAHs selectivity. Moreover, CO yield
remained unchanged but CO2 yield continuously increased with
the increase in Cu doping amount from 1wt%. Meanwhile, it
should be noted that more Cu loading enhanced the coke
formation on the surface of catalyst since the increase of Lewis
acid sites and acidity by the loading of Cu could promote the
polycondensation of intermediates derived from catalytic cracking.
Also, bulk copper species are assumed to participate in the coke
formation. Accordingly, low loading amount of Cu on β-zeolite was
required for the selective production of MAHs in this condition.
From the above results, it can be concluded that Cu loading with
an optimum amount of 1wt% improved the catalytic performance
of β-zeolite for the selective furfural conversion into aromatic
hydrocarbon. However, the generated coke and PAHs amounts
were still high, resulting in the decrease in the stability of catalyst.
To solve this problem, in the following study, the co-doping
effect of Mg was investigated on the 1wt%Cu/β-zeolite catalyst.
Figure 3 and Table S4 show the product distributions derived
from catalytic deoxygenation of furfural at 600 °C using
1wt%Cu/β-zeolite with different Mg loading amounts. One can
see that the co-loading of Mg enhanced the aromatic yield,
especially the MAHs yield. Interestingly, only BTXs were
selectively produced without PAHs formation by using 0.5wt% or
1wt% of Mg loaded 1wt%Cu/β-zeolite, clearly indicating that co-
Figue 2. (A) Product yields and (B) aromatic distributions in
liquid products obtained from catalytic deoxygenation of furfural
at a reaction temperature of 600 °C, a TOS of 1 h and a WHSV
of 1.2 h-1 by using Cu/β-zeolite with various Cu loading amounts.
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
loading of Mg with Cu is quite effective for inhibiting PAH
formation. Here, the alkylation and aromatization of MAHs to
PAHs could be retarded by the new acid-base sites generated
from co-loading of Cu and Mg. This result was confirmed by
NH3-TPD and CO2-TPD in Figures 1E and F, implying that acid-
base properties of catalysts and the metal species with their
interaction at the exchange sites had the great effect on BTXs
formation. As shown in Figure S9 and Table S5, the catalytic
activity was not improved by sole loading of Mg on β-zeolite,
indicating that Mg loading did not create active sites for the
furfural conversion but created basic sites which suppressed
excessive aromatization, alkylation and polymerization.
In addition, the H2 formation might occur via redox cycle
reaction in the presence of AAEMs with in-situ generated H2O
during the reaction (Eq. (1)). In this case, some reactions such
as hydrogenation, hydrocracking, and deoxygenation could be
promoted by the produced H2 via capping free radicals, resulting
in the increase in the productions of some hydrocarbon such as
alkane, which were the sources of MAHs formations.[41,42]
production of BTXs over the 0.5wt%Mg-1 wt%Cu/β-zeolite catalyst
prepared by the co-doping method. The 0.5wt%Mg-1wt%Cu/β-
zeolite catalyst prepared by the ion-exchange method also showed
slightly lower activity than the catalyst prepared by the co-
impregnation method, although the EDX analysis revealed that the
elemental compositions of Mg and Cu (Mg: 0.48 wt%, Cu: 0.99
wt%) were almost identical to those prepared by the co-doping
method. When the catalyst was prepared by the ion-exchange
method, the cationic species, namely Cu2+ and Mg2+, should be
separately fixed at the cation exchange sites on β-zeolite. Both the
mixed catalyst and ion-exchanged catalyst showed lower activity
and selectivity towards BTXs, indicating that the vicinity of Cu and
Mg species is essential for the appearance of co-loading effect of
Cu and Mg. This hypothesis is supported by the facts that the
catalyst prepared by the successive impregnation of Cu and Mg
showed almost similar catalytic performance as the catalyst
prepared by the co-impregnation method.
αC + H2O  H2 + (1 – α)CO2 +(2α – 1)CO (Overall redoxreaction) (1)
Furthermore, with the increase in Mg loading amount, the coke
yield was also obviously reduced (Figure 3), suggesting that the
polycondensation reaction was effectively hindered due to the
addition of Mg metal. The possible reaction mechanisms of these
phenomena were reported in details in previous literatures.[43,44]
Also, as shown in Figure 3, more olefins and CH4 were produced
by increasing of Mg loading amount on 1wt%Cu/β-zeolite,
indicating that polymerization and alkylation of BTXs to PAHs
were inhibited by co-loading effect of Mg with Cu. However,
when Mg loading amount was further increased to 3wt%, the
oxygenated compounds were again detected in the produced
liquid while the aromatic hydrocarbon yield decreased to some
extent. Therefore, the optimum doping amount should be existed.
In this study, 0.5wt%Mg-1 wt%Cu/β-zeolite exhibited the highest
activity, selectivity and stability for BTXs only production with a
total aromatic hydrocarbon yield of 44.5%.
In addition, Mg and Cu doped on SiO2 and Al2O3 were also
tested for comparison (Figures S10 and S11, Tables S3 and S4).
Especially, when 1wt%Cu/SiO2, whose acid amount is much
lower than that in 1wt%Cu/β-zeolite, was used as the catalyst,
only a few aromatic hydrocarbons were detected, indicating that
the acid sites of zeolite had more important role than Cu metal
particle for the production of aromatic hydrocarbons. It is also
found that the use of 1wt%Cu/Al2O3, which had relatively
stronger acid sites than 1wt%Cu/β-zeolite, resulted in facile
formation of PAHs and coke from furfural. This result indicates
that the stronger acid sites may promote the undesirable
polyaromatization. Even though those supports were less
effective for furfural conversion than β-zeolite, both the
conversion of furfural and the yields of MAHs were increased by
co-doping of Mg with Cu. These results indicate that co-loading
of Mg with Cu on each type of support material was necessary.
The effect of catalyst preparation way was also investigated in
this work (Figures. S12, S13, S14 and Table S6). The preparation
method of each catalyst was provided in SI. As shown in Figure S12,
when the physical mixture of 0.5wt%Mg/β-zeolite and 1wt%Cu/β-
zeolite (weight ratio of 1 to 1) was applied, the total yield of BTXs
was decreased accompanying with the formation of certain amount
of furan and polyaromatics as well as unreacted furfural, which is in
contrast with the complete conversion of furfural and the selective
Figure 4. Effect of WHSV on product yields and aromatic
distributions in liquid products obtained from the catalytic
deoxygenation of furfural at a reaction temperature of 600 °C
and a TOS of 1 h using (A,B) β-zeolite, (C,D) 1wt%Cu/β-zeolite
and (E,F) 0.5wt%Mg-1wt%Cu/β-zeolite.
5
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
As such, in the following study, 0.5wt%Mg-1wt%Cu/β-zeolite with
the best performance was selected to investigate the effects of WHSV
and reaction temperature on furfural conversion by comparing with β-
zeolite and 1wt%Cu/β-zeolite. Figure 4 and Table S7 show the effect of
WHSV on the catalytic conversion of furfural. In this study, the WHSV
was changed by differentiating mass of loaded catalysts. For the parent
β-zeolite, 1wt%Cu/β-zeolite and 0.5wt%Mg-1wt%Cu/β-zeolite, furan
and carbon monoxide were the abundant products at the highest
WHSV of 4.8 h-1. These results clearly indicate that the
decarbonylation of furfural should be the initial reaction on these
catalysts regardless of the loading of Cu and Mg. However,
similarity of the product distribution on the parent β-zeolite at a
WHSV of 2.4 h-1 and those on 1wt%Cu/β-zeolite and 0.5wt%Mg-
1wt%Cu/β-zeolite at a WHSV of 4.8 h-1 indicates that the reaction
rate of this initial reaction was increased almost twice by Cu
loading but not significantly promoted by the co-loading of Mg with
Cu. Moreover, in comparison of the parent β-zeolite and
1wt%Cu/β-zeolite, similar product distribution was observed at the
lowest WHSV of 0.6 h-1 but drastically different at the WHSVs of
1.2 h-1 and 2.4 h-1. These facts imply that the deoxygenation of
furan to aromatic compounds, which is relatively slow on the
parent β-zeolite, was effectively promoted by Cu loading.
According to the above discussion, it can be concluded that the
conversion of furfural was promoted by the acid sites on β-zeolite
while Cu was mainly served as a modifier to enhance the catalytic
activity of β-zeolite especially on the deoxygenation and
aromatization of furan. Meanwhile, on both of the parent β-zeolite
and 1wt%Cu/β-zeolite, the production of naphthalene was
observed even at the high WHSV, indicating that the
polyaromatization should easily occur during the conversion of
furan to aromatic compounds on these catalysts. However, on the
0.5wt%Mg-1wt%Cu/β-zeolite, naphthalene and indene were not
observed unless the WHSV was quite low. This comparison
obviously indicates that the addition Mg prevented the
polyaromatization, which could be resulted from the modification
of acid-base sites in the catalyst.
As shown in Figure
5 and Table S8, the reaction
Figure 5. Effect of reaction temperature on product yields and
aromatic distributions in liquid products obtained from catalytic
deoxygenation of furfural at a TOS of 1 h and a WHSV of 1.2 h-1
using (A,B) β-zeolite, (C,D) 1wt%Cu/β-zeolite and (E,F)
0.5wt%Mg-1wt%Cu/β-zeolite.
temperature should be over 600 °C for 100% furfural conversion
over the tested catalysts at the WHSV of 1.2 h-1. However, the
increase in reaction temperature resulted in the slight decrease
in aromatic yield with the increase in gas yield. In general, the
catalytic cracking can be enhanced at a high temperature, which
could increase the coke yield as well. However, it should be
noted that the coke yield was decreased again at 700 °C,
suggesting that coke could be reformed into the gaseous
In addition, the increase of temperature inhibited the alkylation
and favoured the dealkylation, leading to the increase in CH4
and olefins yields. Particularly, only pure benzene was detected
in the liquid product for 1wt%Cu/β-zeolite and 0.5wt%Mg-
1wt%Cu/β-zeolite with the reaction temperatures of 700 °C and
750 °C, respectively. This indicates that 1wt%Cu/β-zeolite and
0.5wt%Mg-1wt%Cu/β-zeolite had high selectivity for the
benzene production from furfural conversion at a reaction
temperature over 700 °C. Based on these results, it can be
concluded that the tuning of acid and basic sites of β-zeolite and
the interaction of metal species on β-zeolite by co-doping of Mg
with Cu, the changing of reaction temperature as well as WHSV
can effectively control the product distribution from the furfural
conversion.
products at
a higher temperature. More interestingly, the
oxygenated compounds were completely converted to aromatic
hydrocarbons at a reaction temperature over 500 °C when
0.5wt%Mg-1wt%Cu/β-zeolite was used. In contrast, the
temperature should be increased over 600 °C and 700 °C for
1wt%Cu/β-zeolite and the parent β-zeolite respectively for
achieving the complete conversion of oxygenated compounds.
This indicates that Mg-Cu/β-zeolite was beneficial for the
catalytic deoxygenation at a lower temperature when compared
with the parent β-zeolite as well as Cu/β-zeolite. Furthermore,
with the increasing of reaction temperature from 600 °C to
650 °C, for 1wt%Cu/β-zeolite and 0.5wt%Mg-1wt%Cu/β-zeolite,
the relative benzene amount was obviously increased while the
relative amounts of toluene and xylenes were decreased to
some extent. It indicates that the demethylation of toluene and
xylene to benzene can be promoted at higher temperatures.
According to the presented results, the possible mechanism
for the catalytic conversion of furfural over Mg-Cu-doped β-
zeolite can be proposed following based on previous
literature.[45] Firstly, furan is produced by the decarbonylation of
furfural. Assuming from the formation of certain amount of
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
olefins, allene is possibly formed by decarbonylation of furan
and the produced allene is trapped on the acid sites of zeolite
and converted into various olefins. In addition, benzofuran was
also detected especially at a high WHSV and consumed with
increasing of the WHSV, indicating that benzofuran, which is
formed by the dehydrative condensation of furan, is also a
possible intermediate for the formation of MAHs. These
intermediates are further transformed into aromatic, olefins and
coke through a series of reactions such as decarbonylation,
oligomerization, cracking, dehydrogenation, dehydration, Diels-
Alder condensation, alkylation and aromatization. In addition, the
metal loaded β-zeolite serves Lewis acid sites or electron pair
acceptors where hydride ions release to promote the rate of
olefin oligomerization, leading to the facile formation of aromatic
hydrocarbons.[46,47] However, if the Cu loading amount is too
high, higher acidity as well as bulk CuO formed on β-zeolite
could promote the secondary reactions of aromatization and
polymerization of BTXs to form PAHs and coke. Therefore, the
optimum acid-base properties of catalyst and the interaction of
metal species with zeolite support for preventing the formation of
bulk copper species should be the key to control the final
aromatic distribution. Moreover, simultaneous introduction of
basic sites (MgO) on catalyst by loading alkali earth metal
displays the significant reduction of coke yield together with
lower PAHs yield while BTXs yield is increased.[48] In other
words, basic MgO is able to deactivate strong Lewis acid sites
derived from extraframework Al species of β-zeolite and ionic
Cu(Cu2+)/CuO/Cu2O
which
can
preferably
catalyze
polyaromatization and coke formation, and thusly decrease the
amounts of coke as well as PAHs generations. As stated above,
in this study, 0.5wt%Mg-1wt%Cu/β-zeolite exhibited the
optimum physicochemical properties for the selective production
of BTXs from furfural.
Catalyst stability and reusability
Figure 6 shows the stability and reusability of 1wt%Cu/β-zeolite
and 0.5wt%Mg-1wt%Cu/β-zeolite, which were tested for 10 cycles.
In each cycle, the reaction was performed at 700 °C for 1 h with a
WHSV of 1.2 h-1. Here, a reaction temperature of 700 °C was
selected for this study since only pure benzene in liquid product was
detected at this temperature together with a reduction of coke
amount on the catalysts. As shown in Figures 6A and C, one can
see that 0.5wt%Mg-1wt%Cu/β-zeolite exhibited long-term stability
for 10 cycles with 100% of furfural conversion. In contrast,
1wt%Cu/β-zeolite was stable in only 4 cycles. This indicates that
0.5wt%Mg-1wt%Cu/β-zeolite was much more stable than
1wt%Cu/β-zeolite. The aromatic distribution during the catalyst
stability test and the reusability are also shown in Figure 6. It is
found that 1wt%Cu/β-zeolite showed high benzene selectivity for
only 3 cycles. In contrast, the stable benzene selectivity of
0.5wt%Mg-1wt%Cu/β-zeolite was maintained up to 6 cycles, and
thereafter, the toluene and xylene were also detected in the
product. It should be noted that the aromatic hydrocarbon
selectivity towards BTXs only was maintained for 10 cycles over
both catalysts. Here, the deactivation of catalyst should be resulted
from the coke deposition on the catalyst. It is found that the amount
of coke also increased with the increase of the reused cycle
number (Table S9).
Figure 6. Stability and reusability of (A) fresh 1wt%Cu/β-zeolite,
(B) regenerated 1wt%Cu/β-zeolite, (C) fresh 0.5wt%Mg-1wt%
Cu/β-zeolite and (D) regenerated 0.5wt%Mg-1wt%Cu/β-zeolite
for catalytic deoxygenation of furfural at a reaction temperature
of 700 °C, a TOS of 1 h and a WHSV of 1.2 h-1.
The coke formation could block the accessibility to the active
sites and thus diminish the efficiency of catalyst. This resulted in
the change of BTXs selectivity in each reaction cycle. The SEM
image of spent catalyst indicated the existence of carbon
deposited on the surface of catalyst (Figure S15). Figure S16
7
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
shows the TEM image of spent catalyst. Distinctly different
morphologies from β-zeolite support were observed, which
should be assigned to the coke deposits. The coke
morphologies exhibited the appearance of wormhole-like
structure with significant disorder of pore arrangement,
corresponding to the feature of amorphous coke. Figure S17
shows DTG profiles for the spent catalysts. It is reported that the
coke with a lower removal temperature range belonged to the
soft coke species while that with a higher removal temperature
range was hard coke species.[49,50] One can see that the obvious
thermal decomposition of coke occurred in the range of 450-
625 °C for the 0.5wt%Mg-1wt%Cu/β-zeolite but this
decomposition temperature was in the range of 550-750 °C for
1wt%Cu/β-zeolite, indicating that Mg loading retarded the hard
coke formation. Furthermore, the coke amount on 1wt%Cu/β-
zeolite after the stability test for 10 cycles was 29.1%, which was
also higher than that on 0.5wt%Mg-1wt%Cu/β-zeolite (22.6%).
Here, the changing of coke yield and formed coke type might be
also related with the existence of Mg and Cu species in the
vicinity of zeolite, which could have the synergy effect between
H+ sites with metal/metal oxide sites in the zeolite structure.
The spent catalysts were regenerated by calcination of them at
750 °C for 30 min under air atmosphere. As shown in Figures S15B
and D, the carbon deposited on the surface of catalyst was
successfully removed by this method. Also, it was confirmed by
investigating the textural properties as well as the physical structure
of catalysts before and after reaction (Table S10 and Figure S18).
Non-significant changing of surface area as well as pore size of
catalyst after the reaction was observed. Moreover, XRD patterns of
catalysts before and after the 10 reused cycles were almost similar,
suggesting that the β-zeolite structure was still preserved during the
stability test. Meanwhile, as expected, no serious reductions of
furfural conversion and relative aromatic hydrocarbon amount were
observed for the regenerated catalysts when compared with the
fresh ones. Furthermore, 0.5wt%Mg-1wt%Cu/β-zeolite exhibited
much better reusability than 1wt%Cu/β-zeolite. It is possible that
0.5wt%Mg-1wt%Cu/β-zeolite preserved the acidic and basic sites
during the regeneration process, but the basic sites of 1wt%Cu/β-
zeolite were obviously lost (Figure S19 and Table S10), which
should be resulted from the stable graphite-like coke formation on
the spent catalyst. Therefore, it can be concluded that the co-
loading of Cu and Mg can significantly improve the stability and the
selectivity towards BTXs.
polyaromatization was apparently suppressed on the Mg loaded
catalyst, while the decarbonylation reaction of furfural took place
as an initial reaction regardless of Cu and Mg loadings. Cu
loading may create Lewis acid sites which would be responsible
for the improvement of deoxygenation and aromatization. The
increase of reaction temperature over 700 °C resulted in the
pure benzene generation, suggesting that such
a high
temperature was beneficial for the dealkylation of toluene and
xylenes as well as the cracking of larger aromatics compounds.
Especially, 0.5wt%Mg-1wt%Cu/β-zeolite exhibited much higher
stability and better reusability than 1wt%Cu/β-zeolite. These
results indicate that Cu and Mg species distributed on the zeolite
structure displayed crucial roles in improving the selectivity
towards BTXs production as well as the long-term stability and
reusability by tuning the acid and basic sites of β-zeolite.
Experimental Section
Catalyst preparation
Commercial β-zeolite (HSZ-960, Si/Al = 100) in protonic form (H+
cation) was purchased from TOSOH Corporation, Japan and
applied as
a catalyst and/or a catalyst support for the
deoxygenation of furfural. In metal loading process, the
designed amounts of Cu and Mg were loaded on β-zeolite by
wet impregnation method. In a typical preparation, a certain
amount of Cu(NO3)2•3H2O or Mg(NO3)2•6H2O (Wako, Japan)
was dissolved in the deionized water and stirred until a
homogeneous solution was obtained. Then, β-zeolite powder
was added into the mixture and stirred at ambient temperature
for 2 h. Thereafter, the slurry was dried at 80 °C and calcined in
an air atmosphere at 750 °C for 2 h. The metal loading amount
was calculated based on the weight percentage. For comparison,
Cu and Mg loaded catalysts were prepared by using SiO2 (Saint-
Gobain, Norpro, France) and γ-Al2O3 (Alfa Aesar, USA) as
supporting materials by the same impregnation method.
Characterizations
N2 adsorption/desorption isotherms were measured at -196 °C
using a Quantachrome instrument (NOVA 4200e, USA). Before
the N2 sorption measurements, the catalysts were preheated
under vacuum condition at a temperature of 200 °C for 2 h. The
specific surface area and total pore volume of catalyst were
calculated by the Brumauer–Emmett–Teller (BET) method and
the Barrett–Joyner–Halanda (BJH) method, respectively.
XRD patterns of β-zeolite structure before and after loading
with Cu and/or Mg metals were determined using an X-ray
diffractometer (XRD, Rigaku Smartlab, Japan) in 2θ range of 10-
60° with a scanning step of 0.02° at each point using Cu Kα
radiation (λ = 0.1542 nm).
Morphology of catalyst as well as the dispersion of Cu and/or
Mg species on the framework of β-zeolite were investigated by
using a transmission electron microscope (TEM, JEM-2100F,
JEOL, Japan) and a scanning electron microscope (SEM,
SU8010, Hitachi, Japan) coupled with an energy dispersive X-
ray detector (EDX).
Conclusions
The catalytic activity, selectivity and stability of β-zeolite in the
furfural conversion to BTXs were successfully improved by co-
loading of Mg with Cu. It is found that the loading amounts of Cu
and Mg and the reaction temperature had great effects on the
final monoaromatic compounds distribution in the liquid product.
The enhancement of selective BTXs production and the
significant reduction of coke deposition amount on catalysts
were achieved when 0.25-1wt%Mg and 1wt%Cu were loaded on
β-zeolite. Avoidance of formation of bulk copper species, which
promoted coke and polyaromatic production, at the low Cu
loading amount (1wt%) was found to be important for achieving
the high BTXs selectivity. The product distributions in varying
WHSV indicate that deoxygenation and aromatization of furan
were significantly promoted on the Cu loaded catalyst and the
Diffuse reflectance (DR) UV–Vis spectra of Cu and Mg
loaded β-zeolite catalysts were recorded at wavelengths ranged
8
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
from 190-450 nm on a JASCO V-650 spectrophotometer using
an ISV-722 integrating sphere coated with BaSO4.
X-ray photo electron spectroscopy (XPS) was conducted in
order to recognize the oxidation states of Cu and Mg species on
β-zeolite structure using a VG Scientific ESCALab250i-XL unit
(UK).
Ultra ALLOY+ 5 capillary column. The operation conditions in
column were conducted at temperature from 50 to 300 °C with a
heating rate of 10 °C/min and held at 300 °C for 10 min. The
compounds were identified by comparison with the mass spectra
in the NIST MS library. The conversion of furfural and the yields
of liquid components such as BTXs, indene, naphthalene, furan,
benzofuran were quantified using external standards purchased
from Wako Chemical Company, Japan.
The acid and base properties of catalysts were determined
by NH3-Temperature-Programmed Desorption (NH3-TPD) and
CO2-Temperature-Programmed
Desorption
(CO2-TPD),
The gas products such as CO, CO2, CH4 and H2 were
analyzed by using a gas chromatograph (Agilent 7890A GC
system, USA) equipped with a thermal conductivity detector
(TCD) and three packed columns (two molecular sieve 5A
columns and one HayeSep Q column) for separation of CO, CH4
and CO2 using He as a carrier gas, while a molecular sieve 5A
for H2 detection using Ar as a carrier gas. For olefin gases such
as C2H4, C2H6, C3H6, C3H8, C4H8 and C4H10 were analyzed by a
gas chromatograph (GC-2014, Shimadzu, Japan) with a flame
ionization detector (FID, column: Gaskuropack). The gas yield
was quantified from the external standards. Here, the yield of
each product was expressed in unit of weight percentage of
carbon.
respectively using a BEL-CAT catalyst analyzer (BEL, Japan)
equipped with a thermal conductivity detector (TCD). Before the
acidity measurement, the catalyst was preheated with a He flow
at 750 °C for 1 h. The preheated catalyst was then saturated in a
gas mixture (5%NH3/95%He) with a flow rate of 50 cm3/min at
50 °C for 1 h. Thereafter, NH3 desorption was detected by TCD
from 50 to 800 °C with a heating rate of 10 °C/min under a He
flow (50 cm3/min). For basicity measurement, the same
measurement conditions were applied, in which only NH3 gas
was replaced by CO2.
Reduction behaviours of Cu and Mg loaded β-zeolite were
measured by H2-Temperature-programmed reduction (H2-TPR)
using the same instrument and pretreatment method as that for
TPD. H2 consumption of each catalyst was detected by TCD
from 50 to 650 °C with a heating rate of 5 °C/min under 5%H2/Ar
steam with a total flow rate of 50 cm3/min.
Acknowledgements
The total coke amount deposited on the spent catalyst was
quantified using a thermogravimatric analyzer (TGA, DTG-60H,
Shimadzu, Japan). All the samples were preheated at 120 °C for
30 min to remove the moisture and then heated with a ramp rate
of 10 °C/min to 800 °C under an air flow.
This study is supported by Aomori City Government, Japan, and
the International Joint Research Project of Shanxi Province
(No.2015081051 and 2015081052), P.R. China. S. Karnjanakom,
A. Bayu and I. Kurnia greatly acknowledge the Ministry of
Education, Culture, Sports, Science, and Technology (MEXT) of
Japan for the scholarship.
Catalytic activity testing
Keywords: Mg-Cu-loaded β-zeolite
• Furfural • BTXs •
Deoxygenation • Aromatization • Stability
The catalytic conversion of furfural was carried out in a
continuous-flow fixed bed reactor equipped with a syringe pump
[1] X. Zhang, H. Lei, S. Chen, J. Wu, Green Chem. 2016, 18,
4145-4169.
[2] J. Zhang, Y. S. Choi, B. H. Shanks, Catal. Sci. Technol.
2016, 6, 7468-7476.
[3] R. H. Venderbosch, ChemSusChem 2015, 8, 1306-1316.
[4] P. S. Rezaei, H. Shafaghat, W. M. A. W. Daud, Green Chem.
2016, 18, 1684-1693.
[5] J. Jae, G. A. Tompsett, A. J. Foster, K. D. Hammond, S. M.
Auerbach, R. F. Lobo, G. W. Huber, J. Catal. 2011, 279,
257-268.
[6] H. J. Park, H. S. Heo, J. K. Jeon, J. Kim, R. Ryoo, K. E.
Jeong, Y. K. Park, Appl. Catal. B 2010, 95, 365-373.
[7] C. Mukarakate, M. J. Watson, J. T. Dam, X. Baucherel, S.
Budhi, M. M. Yung, H. Ben, K. Iisa, R. M. Baldwin, M. R.
Nimlos, Green Chem. 2014, 16, 4891-4905.
[8] F. Fan, Y. Cai, X. Li, H. Yin, J. Xia, J. Ind. Eng. Chem.
2017, 46, 139-149.
[9] Y. Shi, Y. Cao, Y. Duan, H. Chen, Y. Chen, M. Yang, Y.
Wu, Green Chem. 2016, 18, 4633-4648.
[10]M. Conte, J. A. Lopez-Sanchez, Q. He, D. J. Morgan, Y.
Ryabenkova, J. K. Bartley, A. F. Carley, S. H. Taylor, C. J.
Kiely, K. Khalid, G. J. Hutchings, Catal. Sci. Technol.
2012, 2, 105-112.
(Fusion 100 Infusion Pump, CHEMYX) connected to
a
vaporization line at 240 °C. The reactor was made by quartz
tube with a height of 500 mm and an internal diameter of 8 mm.
The temperature of reactor was measured by
a K-type
thermocouple inserted into the catalyst bed. A schematic
diagram of the experimental setup was presented in Figure S20.
In a typical run, 1 g of catalyst was packed in the reactor with the
quartz wool. Furfural (95%, Wako, Japan) was evaporated at
240 °C in the pipeline before entering into the catalyst bed.
During the reaction, furfural (1.17 g/h) was carried by pure Ar
(50 cm3/min) and fed into the reactor. Before and after the
reaction, the reactor was purged with Ar gas flow for 30 min until
no other gases were detected by GC. The operating conditions
were as follows: reaction time of 1 h; reaction temperatures
ranged from 400-750°C; WHSVs ranged from 0.6-4.8 h-1. The
condensable liquid was trapped with acetone in an ice-cooling
bottle and the gas product was collected in a gas bag for further
analysis. The spent catalyst was analyzed by TGA to determine
the coke amount.
Analysis of liquid and gas products
[11] C. A. Mullen, A. A. Boateng, ACS Sustainable Chem.
Eng. 2015, 3, 1623-1631.
Chemical distribution in liquid product was determined by using
a gas chromatograph (GC-2010 Plus, Shimadzu, Japan)/mass
spectrometer (GCMS-QP2010 Ultra, Shimadzu, Japan) with an
[12]W. B. Widayatno, G. Guan, J. Rizkiana, J. Yang, X. Hao,
A. Tsutsumi, A. Abudula, Appl. Catal. B 2016, 186, 166-
172.
9
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
[13]E. F. Iliopoulou, S. Stefanidis, K. Kalogiannis, A. C.
Psarras, A. Delimitis, K. S. Triantafyllidis, A. A. Lappas,
Green Chem. 2014, 16, 662-674.
[31]D. Zhang, R. Wang, X. Yang, Catal. Commun. 2011, 12,
399-402.
[32]C. Li, K. Fujimoto, Catal. Sci. Technol. 2015, 5, 4501-
[14]A. Veses, M. Aznar, I. Martínez, J. D. Martínez, J. M.
López, M. V. Navarro, M. S. Callén, R. Murillo, T. García,
Bioresour. Technol. 2014, 162, 250-258.
4510.
[33]T. E. Tshabalala, M. S. Scurrell, Catal. Commun. 2015
72, 49–52.
,
[15]S. D. Stefanidis, S. A. Karakoulia, K. G. Kalogiannis, E. F.
Iliopoulou, A. Delimitis, H. Yiannoulakis, T. Zampetakis, A.
A. Lappas, K. S. Triantafyllidis, Appl. Catal. B 2016, 196,
155-173.
[34]M. Rutkowska, I. Pacia, S. Basąg, A. Kowalczyk, Z.
Piwowarska, M. Duda, K. A. Tarach, K. Góra-Marek, M.
Michalik, U. Díaz, L. Chmielarz, Microporous Mesoporous
Mater. 2017, 246, 193-206.
[16]W. L. Fanchiang, Y. C. Lin, Appl. Catal. A 2012, 419-420,
102-110.
[17]S. Karnjanakom, G. Guan, B. Asep, X. Hao, S.
Kongparakul, C. Samart, A. Abudula, J. Phys. Chem. C
2016, 120, 3396-3407.
[18]R. Nava, B. Pawelec, P. Castaño, M. C. Álvarez-Galván,
C.V. Loricera, J. L. G. Fierro, Appl. Catal. B 2009, 92,
154-167.
[35]S. Asthana, C. Samanta, A. Bhaumik, B. Banerjee, R.K.
Voolapalli, B. Saha, J. Catal. 2016, 334, 89-101.
[36]X. Zheng, W. Fu, J. Xiong, J. Xi, X. Ni, T. Tang, Catal.
Today 2016, 264, 152-157.
[37]M. Karthik, L. Y. Lin, H. Bai, Microporous Mesoporous
Mater. 2009, 117, 153-160.
[38]J. Zang, Y. Ding, L. Yan, T. Wang, Y. Lu, L. Gong, Catal.
Commun. 2014, 51, 24-28.
[19]Z. Hu, H. Zhang, L. Wang, H. Zhang, Y. Zhang, H. Xu, W.
Shen, Y. Tang, Catal. Sci. Technol. 2014, 4, 2891-2895.
[39]S. Chen, Z. Shao, Z. Fang, Q. Chen, T. Tang, W. Fu, L.
Zhang, T. Tang, J. Catal. 2016, 338, 38-46.
[20]N. Zhang, D. Mao, X. Zhai, Fuel Process. Technol. 2017
167, 50-60.
,
[40]C. Yuan, Y. Zhang, Y. Pan, X. Liu, G. Wang, D. Cao,
Electrochim. Acta 2014, 116, 404-412.
[21]M. Rutkowska, Z. Piwowarska, E. Micek, L. Chmielarz,
[41]F. Melligan, M. H. B. Hayes, W. Kwapinski, J. J. Leahy,
Energy Fuels 2012, 26, 6080-6090.
Microporous Mesoporous Mater. 2015, 209, 54-65.
[22]T. Intana,
K. Föttinger, G. Rupprechter,
P.
[42]H. Wan, P. Chitta, J. Anal. Appl. Pyrolysis 2016, 121,
369-375.
Kongkachuichay, Colloids Surf. A 2015, 467, 157–165.
[23]B. M. Abu-Zied, Microporous Mesoporous Mater. 2011
139, 59-66.
[24]C. M. A. S. Freitas, O. S. G. P. Soares, J. J. M. Órfão, A.
M. Fonseca, M. F. R. Pereira, I. C. Neves, Green Chem.
2015, 17, 4247-4254.
[25]S. Karnjanakom, A. Bayu, X. Hao, S. Kongparakul, C.
Samart, A. Abudula, G. Guan, J. Mol. Catal. A Chem.
2016, 421, 235-244.
[26]H. Wei, K. Liu, S. Xie, W. Xin, X. Li, S. Liu, L. Xu, J. Catal.
2013, 307, 103-110.
,
[43]D. W. Mckee, Fuel 1983, 62, 170-175.
[44]S. Karnjanakom, T. Suriya-umporn, A. Bayu, S.
Kongparakul, C. Samart, C. Fushimi, A. Abudula, G.
Guan, Energy Convers. Manage. 2017, 145, 272-285.
[45]Y. T. Cheng, J. Jae, J. Shi, W. Fan, G. W. Huber, Angew.
Chem. Int. Ed. 2011, 51, 1387-1390.
[46]S. Kelkar, C. M. Saffron, Z. Li, S. S. Kim, T. J. Pinnavaia,
D. J. Miller, R. Kriegel, Green Chem. 2014, 16, 803-812.
[47]X. Y. Ren, J. P. Cao, X. Y. Zhao, Z. Yang, T. L. Liu, X.
Fan, Y. P. Zhao, X. Y. Wei, Fuel 2018, 218, 33-40.
[48]A. Martínez, E Peris, Appl. Catal. A 2016, 515, 32-44.
[49]H. Zhang, S. Shao, R. Xiao, D. Shen, J. Zeng, Energy
Fuels 2014, 28, 52-57.
[27]S. Karnjanakom, A. Bayu, P. Xiaoketi, X. Hao, S.
Kongparakul, C. Samart, A. Abudula, G. Guan, RSC Adv.
2016, 6, 50618–50629.
[28]N. Blanch-Raga, A. E. Palomares, J. Martínez-Triguero, S. [50]Z. Han, S. Li, F. Jiang, T. Wang, X. Ma, J. Gong,
Valencia, Appl. Catal. B 2016, 187, 90-97.
[29]J. Qiu, K. Zhuang, M. Lu, B. Xu, Y. Fan, Catal. Commun.
2013, 31, 21-24.
Nanoscale 2014, 6, 10000-10008.
[30]M. V. L. Pereira, A. Nicolle, D. Berthout, Catal. Today
2015, 258, 424-431.
10
This article is protected by copyright. All rights reserved.

10.1002/cctc.201800450
ChemCatChem
FULL PAPER
Entry for the Table of Contents
Layout 1:
FULL PAPER
High selectivity for BTXs production
with anti-polycyclic-aromatic-
hydrocarbon formation as well as
anti-coking ability was achieved by
using bifunctional Mg-Cu-loaded
β-zeolite.
Surachai Karnjanakom, Akihiro Yoshida,
Asep Bayu, Irwan Kurnia, Xiaogang
Hao, Panya Maneechakr, Abuliti
Abudula and Guoqing Guan*
Page No. – Page No.
Bifunctional Mg-Cu-Loaded β-Zeolite:
High Selectivity for the Conversion of
Furfural into Monoaromatic
Compounds
11
This article is protected by copyright. All rights reserved.