Highly Active Mesoporous Cu?Al2O3 Catalyst for the Hydrodeoxygenation of Furfural to 2-methylfuran

Article abstract of DOI:10.1002/cctc.201901312

While furfural has great potential as a platform chemical through diverse conversion pathways, 2-methylfuran is one of the key components as a fuel additive and a precursor of derived biofuels. However, catalyst design to control reactivity of active species to furfural has been challenging. Herein, we demonstrate that a mesoporous Cu?Al₂O₃ catalyst prepared by solvent-deficient precipitation shows the enhancement of furfural hydrodeoxygenation performance. Not only is the unique pore structure provided, but also the synergistic effects of improved Cu accessibility, co-existence of Cu⁰/Cu⁺ states, and strong metal-support interaction contribute to the superior activity and stability. Additionally, essential parameters for the synthesis and the test reaction are studied to improve the selectivity to 2-methylfuran. The employed catalyst preparation method opens a new channel for tailoring the bifunctionality of supported Cu catalysts.

Full text of DOI:10.1002/cctc.201901312

Accepted Article
Title: Highly active mesoporous Cu-Al₂O₃ catalyst for the
hydrodeoxygenation of furfural to 2-methylfuran
Authors: Seyeon Park, Hari Prasad Reddy Kannapu, Cheonwoo
Jeong, Jinsung Kim, and Young-Woong Suh
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.201901312
Link to VoR: http://dx.doi.org/10.1002/cctc.201901312
A Journal of
www.chemcatchem.org

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
Highly active mesoporous Cu-Al2O3 catalyst for the
hydrodeoxygenation of furfural to 2-methylfuran
Seyeon Park,[a] Hari Prasad Reddy Kannapu,[a, b] Cheonwoo Jeong,[c] Jinsung Kim,[a] and Young-
Woong Suh*[a, b]
Abstract: While furfural has great potential as a platform chemical
through diverse conversion pathways, 2-methylfuran is one of the
key components as a fuel additive and a precursor of derived
biofuels. However, catalyst design to control reactivity of active
species to furfural has been challenging. Herein, we demonstrate
that a mesoporous Cu-Al2O3 catalyst prepared by solvent-deficient
precipitation shows the enhancement of furfural hydrodeoxygenation
performance. Not only is the unique pore structure provided, but also
the synergistic effects of improved Cu accessibility, co-existence of
Cu0/Cu+ states, and strong metal-support interaction contribute to
the superior activity and stability. Additionally, essential parameters
for the synthesis and the test reaction are studied to improve the
selectivity to 2-methylfuran. The employed catalyst preparation
method opens a new channel for tailoring the bifunctionality of
supported Cu catalysts.
with the orientation of the adsorption, the less oxophilic nature of
Cu results in the selective hydrogenation of FUR to furfuryl
alcohol (FOL).[9,10] When considering two step reactions for the
transformation of FUR to FOL, and then to 2MF, a plausible
strategy to propagate the second step is to introduce acidic
support materials.[11,12] Through the cooperation of metal and
acidic sites, cleavage of the C–O bond and the hydrogenation of
the remaining carbon species can be accomplished.
Binary and ternary Cu-based systems for the HDO of FUR
have been thoroughly researched, employing acidic materials
such as H-ZSM-5, TiO2, SiO2, and Al2O3.[11,13–17] Their role is to
support mono- or bi-metallic catalysts[18–21] or to enhance the
acidity of the catalyst at minor elemental composition.[9,22] From
various combinations of metal, support, and sometimes
promoter, we seek a simplistic binary catalyst system to allow for
a straightforward investigation of the catalytic activity and the
role of the support. In particular, we have studied Al2O3 as a
support because it is one of the most commonly used supports
in the corporate world. Therefore, our interest is focused on the
introduction of Cu for its selectivity toward the aldehyde group of
FUR and the addition of alumina for its large surface area[23] and
moderate acidity.[24]
In addition to the simplicity of the binary system, a facile and
scalable synthetic route could increase potential application of
the system in industrial level processes. Although a great
number of elegant synthetic methods have been devised, we
find the solvent-deficient precipitation (SDP) method reported by
Smith et al.[23] greatly promising due to its rapid time-scale, easy
scale-up, and potential applications. Furthermore, SDP is
environmentally benign as it minimizes post-synthetic catalyst
treatment including filtering, washing, and drying. When
considering conventional co-precipitation (CP) and wet
impregnation (WI), not only is their use of large amounts of
water environmentally deleterious, but the post-treatment of
waste water, which probably contains acidic and/or basic ions,
also increases operation costs.
Herein, we report a highly active mesoporous Cu-Al2O3
catalyst for the gas-phase HDO of FUR into 2-MF, where CuO
loading is 20 wt%. The utilization of the solvent-deficient method
satisfies our requirements as the simple binary catalyst is
afforded through a fast and water-saving synthesis. The high-
quality catalyst named m20CA exhibits superior HDO
performance compared to the other supported Cu catalysts with
the same CuO loading that are synthesized by CP and WI
methods (labelled as p20CA and i20CA, respectively). This work
will discuss the catalytic advantages of the newly prepared Cu-
Al2O3 catalyst through various characterization and activity tests.
The catalyst m20CA exhibits unique structural characteristics
which are readily distinguishable with powder X-ray diffraction
(XRD) and N2 physisorption. Despite similar CuO loading for all
samples as shown in Table 1, the XRD pattern of m20CA show
the smallest CuO reflections (Fig. S1), possibly due to the
scarcity of water during synthesis. Specifically, water molecules
located between the metal and support atoms are prone to
Due to the depletion of fossil fuels and the increase in global
warming, biomass conversion has great potential as one of the
most attractive candidates for environmentally friendly energy
sources not only for its abundance as a non-food carbon source,
but also for its ability to produce platform chemicals such as
furfural. Furfural (FUR) has garnered significant attention due to
its annual production of approximately 300 kt and the enhanced
chemical reactivity of its two functionalities, the aldehyde group
and aromatic ring.[1] These functionalities allow FUR to undergo
various reaction pathways including hydrogenation,[2] reductive
amination,[3] decarbonylation,[4] and aldol condensation.[5]
Specifically, 2-methylfuran (2MF) produced through the
hydrodeoxygenation (HDO) of FUR is of enormous interest as a
fuel additive[6] and a precursor of derived biofuels.[7]
It is challenging to design and control the reactivity of metal
species with FUR due to the different modes of FUR adsorption
onto Cu and group VIII catalysts. Group VIII elements, including
Pd, Pt, and Ni, tend to facilitate decarbonylation and ring-
opening as they favor planar binding to the FUR ring.[1,8] In
contrast, Cu causes repulsion due to the overlap of the 3d band
of the metal surface and the furan ring of FUR, enabling
perpendicular adsorption of FUR to the Cu surface. Combined
[a]
S. Park, Dr. H. P. R. Kannapu, J. Kim, Prof. Dr. Y.-W. Suh
Department of Chemical Engineering, Hanyang University
222 Wangsimni-ro, Seongdong-gu, Seoul 04763 (Republic of Korea)
E-mail: ywsuh@hanyang.ac.kr
[b]
[c]
Dr. H. P. R. Kannapu, Prof. Dr. Y.-W. Suh
Research Institute of Industrial Science, Hanyang University
222 Wangsimni-ro, Seongdong-gu, Seoul 04763 (Republic of Korea)
Dr. C. Jeong,
Gwangyang Research Group, Research Institute of Industrial
Science and Technology
187-12 Geumho-ro, Gwangyang, Jeollanam-do 57801 (Republic of
Korea)
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
Table 1. Structural characteristics of catalysts synthesized under different conditions.
CuO a
(wt%)
SBET
Vpore
DCu-XRD
SCu
DCu
Nacid
b
b
c
d
d
e
Catalyst
(m2 g−1)
(cm3 g−1)
(nm)
(m2 g−1)
(%)
(μmol g−1)
m20CA
p20CA
i20CA
21.3
20.1
23.1
407
303
262
0.53
0.48
0.42
9.9 (10.6)
13.6 (19.0)
38.9
16.0
9.4
14.6
9.1
26
31
19
6.7
5.6
[a] Actual CuO loading of the calcined samples was measured by ICP-OES. [b] BET surface area (SBET) and pore volume (Vpore) of metal oxides were
obtained from N2 physisorption at −196 °C. [c] Crystalline size of active Cu (111) was estimated by Scherrer’s equation, where the value in parenthesis
represents Cu size of the spent catalyst obtained after the long-term test at 250 °C. [d] Cu surface area (SCu) and dispersion (DCu) of reduced samples
were calculated through N2O reactive frontal chromatography (N2O-RFC). [e] Acidity (Nacid) was computed from temperature-programmed desorption of
NH3 (NH3-TPD) on reduced catalyst.
come together, causing growth of the nanoparticles. That is, the
hydration effect in p20CA and i20CA prompts Cu nanoparticle
growth. On the other hand, an unambiguous difference in the
BET surface area (SBET) and pore volume (Vpore) was observed.
The largest SBET (407 m2 g−1) and Vpore (0.53 cm3 g−1) values
were observed for m20CA, due to the formation of NH4NO3
during synthesis that can subsequently serve as a spacer.[23,25]
Moreover, it contributes to the narrowest BJH pore size
distribution as well as the exceptional N2 adsorption (Fig. S2).
The unique pore structure of m20CA prompted us to conduct
activity tests. To prepare the samples for testing, the calcined
samples were reduced at 300 °C for 1 h under H2 at a flow rate
of 80 mL min−1 in a stainless-steel reactor. We first tested the
three prepared catalysts under kinetically controlled conditions
showing FUR conversion lower than 15%: 190 °C, weight hourly
space velocity (WHSV) = 23 h−1, and H2:FUR molar ratio = 20
(H2 flow = 35 mL min–1). FUR activity was measured to be 31.9,
Then, the catalytic performance was examined at 230 °C,
WHSV = 1.5 h−1, and H2:FUR molar ratio = 40.8 (H2 flow = 35
mL min–1) for 8 h (indicated by colored bars in Fig. 1a), similar to
conditions of previous studies.[18,26] While i20CA exhibited the
poorest activity under these conditions, similar activities were
found for m20CA and p20CA, leading to further stability testing
at 250 °C while the other parameters were the same. As a result,
significant differences in selectivity and stability were observed
over 24 h of reaction time (Fig. 1b). While m20CA deactivated
more slowly and yielded 2MF as the dominant product, p20CA
produced less 2MF with an increase in FOL selectivity after 16 h.
To understand the superior activity and stability of m20CA, we
utilized various characterization techniques.
Active Cu-related characteristics were analyzed with XRD
and N2O reactive frontal chromatography (N2O-RFC). The
crystalline size calculated by the Cu (111) reflection at 43.3°
(DCu-XRD) was 0.25- to 0.72-fold smaller for m20CA (9.9 nm)
compared to i20CA (38.9 nm) and p20CA (13.6 nm).
Accordingly, the most accessible Cu sites are attained at Cu
surface area (SCu) = 16.0 m2 g–1 and Cu dispersion (DCu) =
14.6% for m20CA (Table 1). Note that more dispersed Cu
particles are seen in the transmission electron microscopy (TEM)
image of m20CA than in those of p20CA and i20CA (Fig. S3).
The poor performance of i20CA could be explained by the fact
that Cu aggregation limits its exposure to FUR. When the spent
m20CA and p20CA (obtained after the long-term test at 250 °C)
were characterized by XRD, the calculated DCu-XRD values were
10.6 and 19.0 nm, respectively. The considerable change of Cu
crystalline size in p20CA indirectly explains the diminshed
selectivity to 2MF from 72.6% at 230 °C to 42–51% at 250 °C,
which is caused by strong Cu sintering at the latter temperature.
In contrast, the growth of Cu particles in m20CA was not
significant, meaning that Cu sintering is not a major contributor
to the changes in product selectivities after the time-on-stream
of 7 h (discussed later).
17.4, and 7.6 mmol gcat h−1 for m20CA, p20CA, and i20CA,
−1
respectively (empty bars in Fig. 1a). However, the selectivity to
2MF was much lower compared to that to FOL (ca. 80±4%) in
these activity runs.
Since it is anticipated that acidity plays a role as a synergistic
component in the activity of m20CA and p20CA at 230 and
250 °C, NH3 temperature-programmed desorption (NH3-TPD)
experiments were carried out, where we only took into account
the m/z = 16 profiles for catalyst acidity (Nacid) due to
interference of H2O at m/z = 17 (Fig. S4a). The acidity of m20CA
and p20CA was higher than that of i20CA (Fig. S5a). In addition,
acidic sites of the spent catalyst were only considered up to
260 °C as a certain amount of water inherently present within
the catalyst evaporates (Fig. S4b). Although the pre-treatment of
the samples at 200 °C for 2 h was implemented to assure
complete removal of water, the presence of the innate water
possibly originates from 1) surface water binding[27] or 2)
transition from residual AlOOH to γ-Al2O3.[28] While the acidity of
the reduced m20CA and p20CA were comparable to each other,
Figure 1. a) FUR activity under the condition of 190 °C, WHSV = 23 h–1, and
H2:FUR = 20 (empty bars), and FUR activity and product selectivity obtained
after 8 h reaction under 230 °C, WHSV = 1.5 h–1, and H2:FUR = 40.8 (colored
bars for FUR activity (gray), and the selectivity to 2MF (blue), FOL (red), and
ring-opening products such as 2-pentanone and 1-pentanol (ROP, green). b)
Stability test results of m20CA (black circles) and p20CA (blue squares) at
250 °C, WHSV = 1.5 h–1, and H2:FUR = 40.8 (FUR conversion: full-filled, 2MF
selectivity: top-filled, FOL selectivity: bottom-filled, and ROP selectivity: empty).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
the acidity change in p20CA was more substantial as it
decreased from 31 to 14 μmol g−1 (by 54.8%) after the long-term
test at 250 °C and 1.5 h−1 WHSV (Fig. 2a). In contrast, the
acidity of m20CA decreased from 26 to 19 μmol g−1 (by 26.8%).
This is one possible reason for the changes in FUR activity and
product selectivity observed throughout the reaction over
m20CA.
To determine why this acidity transformation occurs, XPS and
H2 temperature-programmed reduction (H2-TPR) were utilized.
X-ray photoelectron spectroscopy (XPS) of Cu 2p confirmed the
oxidation states of the reduced samples. While Cu2+ is
distinguishable with its strong satellite features near 943 eV, Cu+
state has a weak satellite at around 945 eV. In this sense, Fig.
2b illustrates the co-existence of Cu0 and Cu+ in m20CA and
p20CA. From the curve fitting results, the abundance of Cu0/Cu+
mixture was estimated to be 98.6% and 97.6% for the fresh
m20CA and p20CA samples, respectively (Fig. S6). More
specifically, Cu LMM Auger spectra reveal that the fresh m20CA
and p20CA contain 12.1% and 11.2% Cu+, respectively, along
with 86.2% Cu0 (Fig. S7). In case of the spent (obtained after the
long-term test at 250 °C and 1.5 h−1 WHSV) m20CA and p20CA
samples, the abundance of Cu0/Cu+ mixture was negligibly
changed (by ca. 1%) while the small increase in Cu+ (14.2% for
m20CA and 13.3% for p20CA) was observed with the decrease
in Cu0 to 83.3%. Therefore, it is believed that the mixture of
Cu0/Cu+ is preserved due to supply of abundant H2 for the titled
reaction. According to previous studies,[29,30] the harmony of Cu0
and Cu+ permits optimal catalytic performance for the activation
of molecular H2 on the metallic Cu surface; wherein Cu+ is also
involved in the catalytic cycle. Furthermore, Cu+ contributes to
enhancing the oxophilic nature and, by extension, to increasing
the acidity of the two catalysts compared to Cu0-governed i20CA
(Fig. S8).[11] The acidity increase by Cu loading was examined
by NH3-TPD profile of Cu-free mesoporous Al2O3 (mA) prepared
by SDP method followed by calcination at 600 °C. As presented
in Fig. S5b, the ammonia desorption from mA was measured in
the temperature range of 100 to 600 °C centered at around
300 °C. In contrast, the ammonia desorption from Cu/mA
prepared by wet impregnation occurred from 50 to 400 °C
although the acid site densities of mA and Cu/mA were not
greatly different (15.5 and 16.7 μmol g–1, respectively). The
catalyst m20CA showed a similar NH3-TPD profile with the
higher acid site density of 26 μmol g–1, which supports the
contribution of Cu+ to the total acidity of m20CA. Therefore, a
mixture of Cu0/Cu+ oxidation states are responsible for the
superior activity of m20CA and p20CA.
Complementary to the XPS analysis, H2-TPR provided insight
on the structural stability of the catalyst through strong metal-
support interaction (SMSI). The SDP synthesis induced SMSI as
m20CA not only begins its reduction at a higher temperature, but
also accomplishes full reduction at a later stage (Fig. 2c). Thus,
Cu particles are well bound to surrounding Al particles such that
the support firmly maintains its mesoporous Cu-Al2O3 structure.
Hence, it can be concluded that the smaller active Cu size,
larger Cu surface area, moderate acidity, cooperation of Cu0 and
Cu+, and configurational stability of the catalyst contribute in
multiple ways to its highly active and stable performance.
Toward our aim of stable 2MF production, we further tuned
reaction parameters such as reaction temperature, H2:FUR
molar ratio, and WHSV. Based on the results in Fig. S9, 250 °C,
H2:FUR = 40.8 (feed flow = 0.064 mL min–1 and H2 flow = 35 mL
min–1), and WHSV of 0.4 h-1 (catalyst loading = 500 mg) were
Figure 2. a) Mass profiles at m/z = 16 from NH3-TPD and b) XPS Cu 2p
spectra of the fresh and spent (obtained after the long-term test at 250 °C and
1.5 h−1 WHSV) catalyst samples, and c) cumulative reduction profiles from H2-
TPR results for the fresh m20CA and p20CA catalyst samples.
chosen as the conditions for long-term stability testing. Next, the
calcination temperature of the mesoporous Cu catalyst was
optimized to maximize 2MF production since the calcination
temperature is one of the most influential factors on the
morphology and activity of Cu-Al2O3 catalysts.[31] To explore this
relationship, we varied the calcination temperature from 500 to
800 °C with an interval of 100 °C. The prepared catalysts are
named m20CA-Cx in which C stands for calcination and x is the
calcination temperature (unit: °C). The variation was determined
to have a direct effect on FUR HDO (Fig. 3a), which offers a
volcano-shaped relationship between calcination temperature
and 2MF selectivity. The apex is lying at 600 °C and 2MF
selectivity decreases in response to elevated calcination
temperature. To further distinguish the cause of this correlation,
we exploited diverse characterization techniques as described
below.
The samples calcined at 700 and 800 °C showed inferior
2MF selectivity due to the spinel structure of CuAl2O4. The XRD
patterns in Fig. 4a demonstrate explicit phase differences along
with spinel formation. In particular, the sample calcined at
700 °C is shown to be in a transitional period, exhibiting both
CuO and CuAl2O4 reflections. As a result, it not only shows
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
Table 2. Physicochemical properties of the mesoporous Cu-Al2O3 calcined at different temperature.
CuO a
(wt%)
SBET
Vpore
SCu
DCu
Nacid
b
b
c
c
d
Catalyst
(m2 g−1)
(cm3 g−1)
(m2 g−1)
(%)
(μmol g−1)
m20CA-C500
m20CA-C600
m20CA-C700
m20CA-C800
21.3
21.3
21.4
21.1
419.1
406.6
210.1
180.8
0.526
0.531
0.393
0.309
14.4
16.0
12.9
11.5
13.2
14.6
11.7
10.6
26
26
20
15
[a] Actual CuO loading of the calcined samples was measured by ICP-OES. [b] BET surface area (SBET) and pore volume (Vpore) of calcined
samples were obtained from N2 physisorption. [c] Cu surface area (SCu) and dispersion (DCu) were computed from the results of N2O-RFC. [d]
Acidity (Nacid) was calculated by NH3-TPD results.
metallic Cu reflections after reduction, but also exhibits non-
reduced copper aluminate reflections. Non-reducible CuAl2O4
reflections were observed from the calcined and reduced
samples of m20CA-C800. Moreover, in the transitional period of
m20CA-C700, 41.8% of Cu2+ and 58.2% of Cu0/Cu+ coexist due
to fragmental CuAl2O4 formation (Fig. S10). The reduced
m20CA-C800 exhibited two distinct peaks, corresponding to the
results of Kwak et al.[30] While the XRD patterns exhibit CuAl2O4
spinel-governed reflections, the XPS analysis in Fig. 4b clarifies
fine CuO and, in turn, Cu0 configurations whose reflections are
undetected by XRD. The non-reduced Cu2+ state from the spinel
and the partially reduced Cu0 states are taken into account by
the higher Cu2+ percentage of 74.2% and by the metallic Cu
presence of 25.8%. Over a range of calcination temperatures,
the physicochemical properties such as SBET, Vpore, SCu, DCu, and
Nacid are altered. Lower numbers of active Cu particles are
exposed, and porosity and acidity decrease due to the
incorporation of Cu and Al particles inside the spinel structure
(Table 2). Combining the analytical methods, we have
determined that increasing the calcination temperature leads to
the evolution of spinel structures, exhibiting a substantial impact
on the structure-activity relationship.
exhibited poorer 2MF selectivity for two reasons: 1) possible
existence of pseudo-boehmite phases in the support[28] and 2)
smaller SCu and DCu stemming from the hydration effect. More
specifically, the CuO reflections of m20CA-C500 are larger than
those of m20CA-C600 due to water molecules intrinsic to the
hydrous salt. Along the way, less approachable Cu sites are
attained, resulting in smaller SCu and DCu of m20CA-C500. On
the other hand, the similar physicochemical properties of the two
catalysts are revealed (Figs. S11 and S12). Hence, calcination
at 500 °C results in a pseudo-boehmite phase and less
accessible Cu sites which directly decreases 2MF selectivity.
Such modifications in the morphology of mesoporous Cu-
based catalysts are in charge of the volcano-shaped structure-
Another limiting aspect was noted for the samples calcined at
500 and 600 °C; the Cu-Al2O3 catalyst calcined at 500 °C
Figure 3. a) Volcano-shaped relation between calcination temperature and
selectivity (reaction temperature: 250 °C, WHSV = 0.4 h–1 and H2:FUR = 40.8)
and b) stability tests under the same reaction condition with different WHSV
values of 0.4 h-1 in black and 1.5 h-1 in pink (FUR conversion: full-filled, 2MF
selectivity: top-filled, FOL selectivity: bottom-filled, and ROP selectivity: empty).
Figure 4. a) XRD patterns of the calcined and reduced m20CA catalyst series
and b) Cu 2p XP spectra of the reduced m20CA catalyst series.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
activity relationship. Compared to the other investigated
temperatures, calcination at 600 °C resulted in smaller metallic
Cu and CuO reflections, a greater number of approachable Cu
sites (SCu and DCu), and complete removal of water. Similar to
Our work elaborates the potential of the solvent-deficient
precipitation method for the morphological control of bifunctional
Cu-based catalysts in furfural hydrodeoxygenation to 2-
methylfuran. This capability could be extended to various Cu-
promoted hydrogenation and dehydrogenation reactions with
optimized acidity.
the temperature-dependent structure-activity correlation,
a
unique dependence of the physicochemical properties on the
calcination temperature was also observed (Fig. S13). In brief,
the relationship between Cu surface and dispersion, and
calcination temperature peaked at 600 °C. Thus, the greater
porosity, highly accessible active Cu, and higher acidity of
m20CA-C600 account for its superior 2MF selectivity compared
to the samples calcined at the other temperatures.
Experimental Section
Catalyst preparation: We modified the solvent-deficient precipitation
method reported by Smith et al.[23] to synthesize mesoporous Cu-Al2O3. A
certain amount of Cu(NO3)2·3H2O, Al(NO3)3·9H2O, and NH4HCO3 were
added to a mortar. The weight ratio of CuO:Al2O3 was calculated as
20:80 and the moles of bicarbonate were equivalent to the moles of
anion from metal salts. The salts and bicarbonate were manually ground
together with a pestle for 21 min, the paste was transferred to a crucible,
and the sample was directly calcined at different temperatures (500, 600,
700 or 800 °C) for 5 h at a ramping rate of 2 °C min−1. This synthesis
recipe was examined to be quite reproducible by identifying the similar
activity and characteristics of the catalyst samples prepared in a number
of different batches. Each sample was designated as m20CA-Cx where
m, 20, CA, and Cx indicate mesoporous, 20 wt% CuO, Cu-Al2O3, and
calcination temperature, respectively.
For comparison, we utilized co-precipitation and wet impregnation
methods. For synthesis via co-precipitation,[32,33] 87.5 mL of a 1.2 M
aqueous solution of Cu(NO3)2·3H2O and Al(NO3)3·9H2O was added drop
wise (14 mL min−1) to 2100 mL of a 0.14 M aqueous solution of
NH4HCO3 in a large scale 5 L precipitation reactor at 70 °C. After aging
of the Cu and Al precipitate under vigorous stirring at the same
temperature for 90 min, it was repeatedly filtered and washed with
deionized water to assure complete removal of sodium and nitrate ions.
Then, the filter cake was dried overnight in a convection oven at 105 °C.
To impregnate commercial γ-alumina (from Sasol) with 20 wt% CuO, we
vigorously stirred 5 g of γ-Al2O3 and a certain amount of Cu(NO3)2·3H2O
in 50 mL of deionized water for 90 min. Following the evaporation of
water in a rotary evaporator (EYELA) under reduced pressure at 80 °C,
the sample was dried at 105 °C overnight and calcined at 600 °C for 5 h
at a heating rate of 2 °C min−1. The catalysts prepared by co-precipitation
and wet impregnation are denoted as p20CA and i20CA, respectively.
We next carried out long-term activity test under the optimal
condition investigated above with m20CA-C600. Fig. 3b
indicates that stabilized 2MF production was achieved at WHSV
= 0.4 h-1, FUR activity was maintained at around 3.8±0.1 mmol
gcat−1 h−1 (FUR conversion of nearly 100%), and 2MF productivity
was averaged at 3.1 mmol gcat h−1 (2MF selectivity of ca.
−1
75.9%) with significantly low FOL productivity. The enhancement
in stability was due to the presentation of more sufficient active
sites compared to the molar flow rate of FUR. Note that the
higher WHSV of 1.5 h-1 produced a considerable amount of coke
at the surface of m20CA, which was identified by the increase in
the carbon content to 12 wt% that was measured by XPS C 1s
peak of the spent m20CA catalyst (obtained after the long-term
test under the condition specified for Fig. 1b), as presented in
Fig. S14. This justifies the changes in product selectivities
observed after the time-on-stream of 7 h.
When compared to activity results reported in literature, the
results reported herein are quite promising. First of all, binary
Cu-Al2O3 catalyst systems for FUR conversion to 2MF have
mostly been tested in a batch reactor,[16,20] indicating that our
activity results obtained with a continuous reaction system will
be useful in future reports. On the other hand, Jiménez-Gómez
et al.[18] reported selective production of 2MF over supported Cu
on mesoporous silica (denoted as 10Cu-MS). When this catalyst
was tested in a fixed-bed reactor (conditions: 210 °C, WHSV =
0.75 h–1, and H2 flow = 10 mL min–1), FUR conversion of close to
90% was attained but 2MF selectivity dropped to 50% after 24 h.
This result is quite similar to that displayed in Fig. 1b. Therefore,
our attempt to reduce the WHSV to 0.4 h-1 is so reasonable that
the catalytic performance of m20CA would be stable throughout
the reaction.
Activity test: In a stainless-steel reactor (i.d. = 10 mm, length = 47 mm),
110–500 mg of catalyst was loaded between three beds of glass beads.
The catalysts were reduced at 300 °C for 1 h under 80 mL min−1 H2
preceding the catalytic activity test. Then, at 210–250 °C, 5 vol% of
furfural in cyclopentyl methyl ether (CPME) was fed into the reactor by an
HPLC pump at 0.064 mL min−1 while certain amount of H2 was provided
during reaction. CPME is reported to be a green solvent for its stability in
acidic and basic media, lower formation of peroxides, and high boiling
point.[34] It also could be used as a reaction medium in dehydration of
lignocellulosic pentose to FUR and in HDO of FUR into 2MF.[35] Moreover,
diluted furfural was preferred in our system to avoid the polymerization of
furfural that possibly causes the line blockage in the continuous
reactor.[26] Owing to the use of CPME, the carbon balance was in the
range of 95–98% in all reaction runs. The product was condensed and
collected for analysis using a GC equipped with an FID. The conversion,
To conclude, this work presents the synthesis of
a
bifunctional Cu-Al2O3 catalyst by solvent-deficient precipitation.
The optimized catalyst exhibited improved reactivity, selectivity,
and stability in furfural hydrodeoxygenation. Equipped with
unique structural properties including small-sized particles and
greater porosity, the highly active mesoporous Cu-Al2O3 showed
superior catalytic activity. The active Cu particles exhibited
enhanced accessibility due to their dispersion and small size.
Also, cooperation between Cu0 and Cu+ oxidation states not only
increased the acidity of the catalyst, but also yielded synergistic
effects on furfural hydrodeoxygenation. Furthermore, the strong
metal-support interaction induced by the water scarcity during
synthesis increased the stability of the catalyst. Facile tuning by
altering the calcination temperature enabled us to explore the
structure-activity relationship of the catalyst. Calcination at an
intermediate temperature of 600 °C avoided the evolution of
spinel structures thereby avoiding decreases in Cu active sites,
porosity, and acidity. This calcination temperature also achieved
complete removal of inherent water from the hydrous precursors.
yield, and selectivity were calculated by the following equations:
푚푚푚퐹퐹퐹,푖푖 − 푚푚푚퐹퐹퐹,표표표
Conversion (%) =
× 100
푚푚푚퐹퐹퐹,푖푖
푚푚푚푃푃표푃표푃표,표표표
Yield (%) =
× 100
푚푚푚퐹퐹퐹,푖푖
푚푚푚푃푃표푃표푃표,표표표
푚푚푚퐹퐹퐹,푖푖 − 푚푚푚퐹퐹퐹,표표표
푌푌푌푚푌
( )
% =
Selectivity
× 100 =
× 100
퐶푚퐶퐶푌퐶퐶푌푚퐶
Analytic characterization: Powder X-ray diffraction (XRD) analysis was
carried out with a Rigaku MiniFlex600 with a radiation source of Cu Kα
operating at 40 kV and 15 mA. Diffraction patterns were obtained at a
scan speed of 10° min−1 at a step of 0.02° with a 2θ range of 10 to 80°.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
The BET (Brunauer-Emmett-Teller) surface area (SBET), pore volume
(Vpore), and BJH pore size distribution were measured in a Micromeritics
ASAP 2020 instrument, following degassing of a 1000 mg sample at
105 °C for 1 h under vacuum. Inductively coupled plasma optical
emission spectrometry (ICP-OES) was performed with a Perkin Elmer
OPTIMA 8300 to evaluate the metal loading of the oxide samples.
[10] C. P. Jiménez-Gómez, J. A. Cecilia, D. Durán-Martín, R. Moreno-Tost,
J. Santamaría-González, J. Mérida-Robles, R. Mariscal, P. Maireles-
Torres, J. Catal. 2016, 336, 107–115.
[11] F. Dong, Y. Zhu, H. Zheng, Y. Zhu, X. Li, Y. Li, J. Mol. Catal. A 2015,
398, 140–148.
[12] M. J. Gilkey, P. Panagiotopoulou, A. V. Mironenko, G. R. Jenness, D. G.
Vlachos, B. Xu, ACS Catal. 2015, 5, 3988–3994.
Utilizing N2O reactive frontal chromatography (N2O-RFC), we
computed the Cu surface area of the samples. A calcined sample (105
mg) was reduced with a BELCAT-B instrument (BEL Japan, Inc.) at
300 °C with a ramping rate of 5 °C min−1 for 1 h under 10% H2/Ar at 30
mL min−1. After cooling to 40 °C under H2, the introduction of 1% N2O/He
at 5 mL min−1 (m/z = 44) and the production of N2 (m/z = 28) were
recorded by a BEL-Mass (BEL Japan, Inc.). The calculation of Cu
surface area was based on two assumptions: the reaction stoichiometry
[13] Z. Fu, Z. Wang, W. Lin, W. Song, S. Li, Appl. Catal. A 2017, 547, 248–
255.
[14] S. Srivastava, G. C. Jadeja, J. Parikh, RSC Adv. 2016, 6, 1649–1658.
[15] H. Zhang, C. Canlas, A. Jeremy Kropf, J. W. Elam, J. A. Dumesic, C. L.
Marshall, J. Catal. 2015, 326, 172–181.
[16] B. Seemala, C. M. Cai, R. Kumar, C. E. Wyman, P. Christopher, ACS
Sustain. Chem. Eng. 2017, 6, 2152–2161.
of Cu/O is 2:1 and Cu surface density is 1.46 × 1019 Cu atoms m−2
The acidity of the samples was studied through ammonia
temperature-programmed desorption (NH3-TPD) experiments with
.
[17] Z. Zhang, Z. Pei, H. Chen, K. Chen, Z. Hou, X. Lu, P. Ouyang, J. Fu,
Ind. Eng. Chem. Res. 2018, 57, 4225–4230.
a
[18] C. P. Jiménez-Gómez, J. A. Cecilia, R. Moreno-Tost, P. Maireles-
Torres, ChemSusChem 2017, 10, 1448–1459.
BELCAT-B instrument. Prior to analysis, the calcined sample was
reduced following the same procedure described above for the N2O-RFC
analysis. On the other hand, the spent catalyst was stabilized at 200 °C
under He flow (30 mL min−1) for 2 h to assure complete removal of
residual species on the catalyst surface. After cooling to 50 °C, the
samples were exposed to 5% NH3/He at a flow of 30 mL min−1 for 1 h.
Then, after He purging, we monitored the desorption of ammonia with
heating from 50 to 800 °C at a ramping rate of 15 °C min−1. For
confirmation of water removal during the TPD, the NH3/He flow was
[19] J. Zhang, J. Chen, ACS Sustain. Chem. Eng. 2017, 5, 5982–5993.
[20] B. Seemala, C. M. Cai, C. E. Wyman, P. Christopher, ACS Catal. 2017,
7, 4070–4082.
[21] H. Sheng, R. F. Lobo, ChemCatChem 2016, 8, 3402–3408.
[22] H. Niu, J. Luo, C. Li, B. Wang, C. Liang, Ind. Eng. Chem. Res. 2019, 58,
6298–6308.
[23] S. J. Smith, B. Huang, S. Liu, Q. Liu, R. E. Olsen, J. Boerio-Goates, B.
F. Woodfield, Nanoscale 2015, 7, 144–156.
replaced with flowing He at 30 mL min−1
Temperature-programmed reduction (TPR) experiments were
conducted in Micromeritics AutoChem 2910 to observe the H2
consumption profiles of the samples. The calcined sample (65 mg) was
heated to 800 °C at a rate of 2 °C min−1 under 10% H2/Ar at 50 mL min−1
X-ray photoelectron spectroscopy (XPS) spectra were obtained with a K-
Alpha Plus spectrometer (Thermo Scientific Co.) with a monochromatic
Al Kα X-ray source (1486.6 eV). The recorded spectra were calibrated by
a standard C 1s peak at 284.6 eV.
.
[24] B. Huang, C. H. Bartholomew, B. F. Woodfield, Micropor. Mesopor.
Mater. 2014, 183, 37–47.
a
[25] J. Oh, H. B. Bathula, J. H. Park, Y.-W. Suh, Commun. Chem. 2019, 2,
68.
.
[26] C. P. Jiménez-Gómez, J. A. Cecilia, R. Moreno-Tost, P. Maireles-
Torres, ChemCatChem 2017, 9, 2881–2889.
[27] J. M. McHale, A. Navrotsky, A. J. Perrotta, J. Phys. Chem. B 1997, 101,
603–613
[28] S. J. Smith, S. Amin, B. F. Woodfield, J. Boerio-Goates, B. J. Campbell,
Inorg. Chem. 2013, 52, 4411–4423.
[29] R. S. Rao, R. T. K. Baker, M. A. Vannice, Catal. Lett. 1999, 60, 51–57.
[30] B. M. Nagaraja, A. H. Padmasri, B. David Raju, K. S. Rama Rao, J. Mol.
Catal. A: Chem. 2007, 265, 90–97.
Acknowledgements
[31] B. K. Kwak, D. S. Park, Y. S. Yun, J. Yi, Catal. Commun. 2012, 24, 90–
95.
This work was financially supported by the Korea Institute of
Energy Technology Evaluation and Planning under the Ministry
of Trade, Industry, and Energy, Republic of Korea (KETEP-
20163010092210), as well as by the Next Generation Carbon
Upcylcing Program through the National Research Foundation
of Korea under the Ministry of Science and ICT, Republic of
Korea (NRF-2017M1A2A2043148).
[32] C. Jeong, H. Ham, J. W. Bae, D.-C. Kang, C.-H. Shin, J. H. Baik, Y.-W.
Suh, ChemCatChem 2017, 9, 4484-4489.
[33] C. Jeong, J. Park, J. Kim, J. H. Baik, Y.-W. Suh, Korean J. Chem. Eng.
2019, 36, 191-196.
[34] K. Watanabe, N. Yamagiwa, Y. Torisawa, Org. Process Res. Dev. 2007,
11, 251–255.
[35] M. J. Campos-Molina, R. Mariscal, M. Ojeda, M. López Granados,
Bioresour. Technol. 2012, 126, 321–327.
Keywords: heterogeneous catalysis • bifunctional catalyst •
furfural • green chemistry • mesoporous Cu-Al2O3
[1]
[2]
[3]
[4]
[5]
[6]
[7]
R. Mariscal, P. Maireles-Torres, M. Ojeda, I. Sádaba, M. L. Granados,
Energy Environ. Sci. 2016, 9, 1144–1189.
L. Burnett, I. Johns, R. Holdren, R. Hixon, Ind. Eng. Chem. 1948, 40,
502–505.
M. Chatterjee, T. Ishizaka, H. Kawanami, Green Chem. 2016, 18, 487–
496.
V. V. Pushkarev, N. Musselwhite, K. An, S. Alayoglu, G. A. Somorjai,
Nano Lett. 2012, 12, 5196–5201.
H. J. Cho, D. Kim, J. Li, D. Su, B. Xu, J. Am. Chem. Soc. 2018, 140,
13514–13520.
C. Wang, H. Xu, R. Daniel, A. Ghafourian, J. M. Herreros, S. Shuai, X.
Ma, Fuel 2013, 103, 200–211.
M. N. Gebresillase, R. Shavi, J. G. Seo, Green Chem. 2018, 20, 5133–
5146.
[8]
[9]
S. Sitthisa, D. E. Resasco, Catal. Lett. 2011, 141, 784–791.
G. S. Hutchings, W. Luc, Q. Lu, Y. Zhou, D. G. Vlachos, F. Jiao, Ind.
Eng. Chem. Res. 2017, 56, 3866–3872.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201901312
ChemCatChem
COMMUNICATION
Entry for the Table of Contents
COMMUNICATION
S. Park, H. P. R. Kannapu, C. Jeong, J.
Kim, and Y.-W. Suh*
Page No. – Page No.
Highly active mesoporous Cu-Al2O3
catalyst for the hydrodeoxygenation
of furfural to 2-methylfuran
A mesoporous Cu-Al2O3 catalyst shows the enhanced performance in the
hydrodeoxygenation of furfural to 2-methylfuran. The solvent-deficient precipitation
method provides unique morphology of smaller Cu crystalline size and larger
porosity, and induces strong metal-support interaction. Equipped with these
remarkable characteristics, the developed catalyst contributes in various aspects to
its highly active and stable performance.
This article is protected by copyright. All rights reserved.