Green synthesis of heterogeneous copper-alumina catalyst for selective hydrogenation of pure and biomass-derived 5-hydroxymethylfurfural to 2,5-bis(hydroxymethyl)furan

Article abstract of DOI:10.1016/j.apcata.2020.117892

In this work novel copper-alumina catalysts were prepared through a solvent-free solid-state grinding method ― a low cost and green catalyst preparation method for selective hydrogenation of 5-hydroxymethylfurfural (5-HMF) into 2,5-bis(hydroxymethyl)furan (BHMF). Under the optimized reaction conditions (3 MPa H₂, 130 °C, 1 h), >99 % 5-HMF conversion and 93 % BHMF yield were obtained by using a 20CA (20 mol%Cu-Al₂O₃) catalyst. The catalyst characterization results could reveal that the high catalytic activity and selectivity could be attributed to the presence of both metallic and electrophilic copper (Cu°/Cu²⁺) species and the uniformly distributed copper nanoparticles. Furthermore, an integrated catalytic process was demonstrated for the first time for direct conversion of mono, di, and polysaccharides into the corresponding BHMF, obtained overall BHMF yield in the range of 25 %–48 %.

Full text of DOI:10.1016/j.apcata.2020.117892

Journal Pre-proof
Green synthesis of heterogeneous copper-alumina catalyst for selective
hydrogenation of pure and biomass-derived 5-hydroxymethylfurfural to
2,5-bis(hydroxymethyl)furan
Kasanneni Tirumala Venkateswara Rao (Methodology)
(Investigation) (Formal analysis) (Data curation) (Visualization)
(Writing - original draft) (Writing - review and editing), Yulin Hu
(Investigation), Zhongsun Yuan (Methodology), Yongsheng Zhang
(Investigation) (Validation), Chunbao Charles Xu
(Conceptualization) (Funding acquisition) (Methodology) (Project
administration) (Supervision) (Validation) (Writing - review and
editing)
PII:
S0926-860X(20)30485-3
DOI:
https://doi.org/10.1016/j.apcata.2020.117892
APCATA 117892
Reference:
To appear in:
Applied Catalysis A, General
Received Date:
Revised Date:
Accepted Date:
22 July 2020
18 October 2020
19 October 2020
Please cite this article as: Rao KTV, Hu Y, Yuan Z, Zhang Y, Xu CC, Green synthesis of
heterogeneous copper-alumina catalyst for selective hydrogenation of pure and
biomass-derived 5-hydroxymethylfurfural to 2,5-bis(hydroxymethyl)furan, Applied Catalysis A,
General (2020), doi: https://doi.org/10.1016/j.apcata.2020.117892

This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. Please note that, during the production process, errors may be
discovered which could affect the content, and all legal disclaimers that apply to the journal
pertain.
© 2020 Published by Elsevier.

Green synthesis of heterogeneous copper-alumina catalyst for selective
hydrogenation of pure and biomass-derived 5-hydroxymethylfurfural to
2,5-bis(hydroxymethyl)furan
Kasanneni Tirumala Venkateswara Raoa, Yulin Hua, Zhongsun Yuana, Yongsheng Zhangb,
Chunbao Charles Xua*
a Department of Chemical and Biochemical Engineering, Western University, London,
Ontario, Canada N6A 5B9
b School of Chemical Engineering, Zhengzhou University, Zhengzhou, PR China
*Corresponding Author:cxu6@uwo.ca
Graphical abstract
5-HMF
BHMF
Cellulosic biomass
H2
Cu(NO3)2 6H2O
Calcination & Reduction
Al(NO3)2 9H2O
Oxalic acid
1

Highlights
 Catalytic conversion of biomass-derived carbohydrates into bio-based chemicals
 Conversion of 5-hydroxymethylfurfural (5-HMF) to 2,5-bis(hydroxymethyl)furan
(BHMF)
 Copper-alumina catalysts were prepared through solvent-free solid-state grinding
 Optimized reaction conditions: 3 MPa H2, 130 oc, 1h
 >99% 5-HMF conversion and 93% BHMF yield obtained
Abstract
In this work novel copper-alumina catalysts were prepared through a solvent-free solid-state
grinding method  a low cost and green catalyst preparation method for selective
hydrogenation of 5-hydroxymethylfurfural (5-HMF) into 2,5-bis(hydroxymethyl)furan
(BHMF). Under the optimized reaction conditions (3 MPa H2, 130 oC, 1h), >99% 5-HMF
conversion and 93% BHMF yield were obtained by using a 20CA (20 mol%Cu-Al2O3)
catalyst. The catalyst characterization results could reveal that the high catalytic activity and
selectivity could be attributed to the presence of both metallic and electrophilic copper
(Cu0/Cu2+) species and the uniformly distributed copper nanoparticles. Furthermore, an
integrated catalytic process was demonstrated for the first time for direct conversion of mono,
di, and polysaccharides into the corresponding BHMF, obtained overall BHMF yield in the
range of 25% to 48%.
Keywords: Selective hydrogenation, 5-Hydroxymethylfurfural, 2,5-
bis(hydroxymethyl)furan, Copper-alumina catalysts, Solvent-free solid-state grinding
2

1. Introduction
Due to the fast depletion of petroleum-based resources and the awareness of environmental
problems triggered by the continuous consumption of fossil deposits, searching for new and
alternative feedstocks to synthesize commodity chemicals, fuels, and polymers has attracted
significant attention in the past few years [1,2]. Low cost, highly abundant, and carbon-
neutral lignocellulosic biomass are considered an ideal feedstock for the synthesis of a variety
of platform chemicals and useful industrial products [3]. 5-Hydroxymethylfurfural (5-HMF)
is a versatile important platform chemical that can be synthesized directly from
lignocellulosic biomass or biomass-derived C6-carbohydrates by the acid-catalyzed
dehydration reaction [4,5]. The obtained 5-HMF can be used for the synthesis of various
useful industrial chemicals e.g., 2,5-furandicarboxylic acid (FDCA) [6][7], hexanediol, 2,5-
dimethylfuran, and 2,5-bis(hydroxymethyl)furan (BHMF) [4] by performing different types
of chemical reactions such as oxidation and hydrogenation. For instance, 2,5-
furandicarboxylic acid can replace the petroleum-derived terephthalic acid for the synthesis
of bio-based polyethylenefuranoate (PET), which can be used for water bottles and packaging
applications [8]. Also, 5-HMF can be used in the synthesis of phenol-hydroxymethyl furfural
(PHMF), a green alternative to PF resins [9].
Since 5-HMF contains two types of functional groups such as C=O and C=C; however,
hydrogenation of this molecule can produce different reaction products [10]. For example, the
carbonyl group selective hydrogenation of 5-HMF can produce a diol - 2,5-
bis(hydroxymethyl)furan (BHMF). BHMF can be used as an essential intermediate chemical
for various potential industrial applications such as polymers, polyurethane foams, drugs, and
other products of the fine chemical industry [10]. Traditionally, BHMF can be synthesized by
the stoichiometric hydrogenation of 5-HMF in the presence of NaBH4 or LiAlH4 [11,12].
However, these reagents are moisture sensitive and generate much environmental waste.
3

Alternatively, the catalytic hydrogenation of 5-HMF with hydrogen has been widely studied
in recent years [10]. A homogeneous ruthenium-based dinuclear complex was found to be
very active for the selective hydrogen of 5-HMF to BHMF, achieved 99% BHMF yield under
mild reaction conditions [13]. However, this catalyst is very expensive, difficult to recover
and reuse, and hence has less industrial potential. Alternatively, heterogeneous catalysts that
are simple to recover and can reuse for many catalytic cycles are more promising. In this
regard, various supported heterogeneous noble metal (Ru, Pt, Pd, Au, Ir, and Re) based
catalysts have been reported for selective hydrogenation of 5-HMF to BHMF[14–25][26].
Although all these catalysts would show remarkable catalytic activity and selectivity,
preparation of these metal catalysts' is very costly. Thus, the development of inexpensive and
heterogeneous recyclable supported metal-based catalysts for selective hydrogenation of 5-
HMF is highly desirable.
Copper-based heterogeneous catalysts are economical and selective catalysts for side-chain
carbonyl group hydrogenation without affecting double bonds in 5-HMF molecule [10]. Cao
et al. synthesized copper supported on silica (Cu/SiO2) catalyst for the selective
hydrogenation of 5-HMF, obtained high BHMF yield after 8h under 2.5 MPa H2 at 100 oC
[27]. In another work, Lima et al. used a commercial Raney®-Cu catalyst for continuous
flow hydrogenation of aqueous 5-HMF at 90 oC under 90 bar H2 pressure, attaining 94% 5-
HMF conversion and 84% BHMF yield, respectively. In contrast, Raney ®-Ni catalyst was
less selective under the same reaction conditions [28]. Hassan et al. synthesized bimetallic
copper-iron oxide nanoparticles for selective hydrogenation of 5-HMF in the absence of
external hydrogen at 150 oC for 5h in ethanol solvent, and obtained 97% 5-HMF conversion
and 92% BHMF yield, respectively [29]. Besides, various other heterogeneous copper-based
catalysts have also been reported under different reaction conditions, achieving BHMF yields
are in the range of 62 to 99% [30–33]. Even though all these catalysts were active for the
4

synthesis of BHMF, most of the reported studies were focussed on hydrogenation of
commercially available 5-HMF. Recently, an integrated catalytic process for direct
conversion of fructose to BHMF was reported [34]. High BHMF yields were reported,
whereas fructose is an expensive feedstock for large scale production of BHMF. Therefore,
catalytic conversion of inexpensive biomass-derived carbohydrates into BHMF is highly
attractive. It should be noted that most of the reported copper-based catalysts for
hydrogenation of 5-HMF to BHMF were prepared through conventional precipitation or co-
precipitation and impregnation methods, which not only are time consuming, but generate
much environmental waste and involve multiple preparation steps. Accordingly, the
development of heterogeneous supported copper catalyst prepared through a simple and
greener solvent-free method is highly desirable for the selective hydrogenation of 5-HMF to
BHMF.
Simple solid-state grinding synthesis of supported metal or metal oxide catalysts have
attracted much attention in the past years [35–42]. In this method, various heterogeneous
catalysts were prepared through simple physical mixing of metal precursors followed by heat
treatment. Interestingly, these catalysts demonstrated better catalytic performance than those
prepared via the conventional wet chemistry route. We recently reported the solid-state
synthesis of tin phosphate and Co-Mn mixed oxide catalysts [43,44], which also showed
much better catalytic performance than the same catalysts prepared in a conventional
precipitation method. Moreover, the main advantages of the solid-state grinding method are:
it is quick, inexpensive, easy to scale up, and no waste is produced during the catalyst
preparation. To the best of our knowledge, no report on the solid-state grinding synthesis of
copper-alumina catalysts for selective hydrogenation of biomass-derived 5-HMF into BHMF.
Herein, we report synthesis of various copper-alumina (Cu-Al2O3) catalysts by the one-step
solid-state grinding of copper and alumina precursors with oxalic acid followed by
5

calcination and reduction. The as-synthesized catalysts were directly used for the selective
hydrogenation of biomass-derived 5-HMF. The effects of various other reaction parameters
such as catalyst support, copper loading, hydrogen pressure, solvent, reaction temperature,
and time, and the influence of catalyst preparation methods were examined in this work to
produce a high BHMF yield. Different catalyst characterization techniques were used to
understand the structural properties of the catalysts and their catalytic activities. Additionally,
a two-step integration method was demonstrated for direct conversion of simple and complex
sugars into BHMF without isolation of the 5-HMF intermediate.
2. Experimental
2.1 Materials
Reagent grade Al(NO3)3 9H2O, Cu(NO3)2 3H2O, oxalic acid, 5-hydroxymethylfurfural
(99%), glucose, microcrystalline cellulose, cellobiose, sucrose, starch, and 5-methylfurfural
were all purchased from Sigma-Aldrich. Niobium phosphate hydrate (NbOPO4.nH2O) was
supplied by CBMM (Companhia Brasileira de Metalurgia e Mineraçã). Ethanol, methanol,
THF, Toluene, methyl isobutyl ketone (MIBK), Hexane, Water (HPLC grade), acetonitrile
(HPLC grade) were obtained from Caledon Laboratories. All reagents were used as received
without further purification.
2.2 Catalyst preparation
Copper-alumina catalysts were prepared via the solid-state grinding method. Taking the
20mol%Cu-Al2O3 catalyst as an example, it was prepared with the required amounts of
copper nitrate, aluminum nitrate, and oxalic acid (OA) at a specific mole ratio
(OA/Cu+Al=1.4) by grinding manually in a mortar and pestle for about 10-15 minutes. We
prepared each catalyst three times and tested their catalytic activity to ensure reproducibility
of the sample preparation method and reported the average values of these catalytic results in
the manuscript. The obtained blue coloured solid paste was dried in an oven at 90 oC
6

overnight, followed by calcination in air at 600 oC for 2h at 2 oC min-1. This catalyst was
denoted as 20CA.
Similarly, by varying the copper nitrate mole ratio, different catalysts were prepared, denoted
as 10Cu-Al2O3 (10CA), 15Cu-Al2O3 (15CA) and 25Cu-Al2O3 (25CA). For comparison,
the 20Cu-Al2O3 catalyst was also prepared by the conventional co-precipitation method, as
reported elsewhere[45], and the resulting catalyst was denoted as 20CA-P.
2.3 Catalyst characterization
The acidic and basic properties of the reduced copper-alumina catalysts (in H2 flow in a
tubular furnace at 350 oC for 2h) were measured by NH3-TPD and CO2-TPD. The
temperature-programmed desorption (TPD) experiments were carried out on a ChemBET
Pulsar TPR/TPD automated chemisorption analyzer equipped with a thermal conductivity
detector. In a typical run, 0.1 g of catalyst sample was loaded into a U-shaped quartz tube,
and the catalyst was pre-treated at 300 oC for 1h under a flow of helium gas (99.9%, 120 mL
min-1). After pre-treatment, the catalyst sample was exposed to pure NH3 or CO2 gas at 50
oC for 10 min. The physisorbed NH3 or CO2 was removed by flushed with He gas at 100 oC
for 30 min. The catalyst's total acidity or basicity was calculated from the amount of NH3 or
CO2 desorbed from ambient temperature to 800 oC at a heating rate of 10 oC min-1. In
addition, temperature-programmed reduction (TPR) analysis of the copper-alumina catalysts
was measured on a ChemBET Pulsar TPR/TPD automated chemisorption analyzer in the
following procedure: approximately 0.1 g of the unreduced catalyst was loaded into the U-
shaped quartz tube, and the sample was pre-treated at 300 oC for 1h with a heating rate of 10
oC min-1 under a flow of pure He gas (99.9%, 120 mL min-1). Then, the furnace temperature
was cooled down to 50 oC in the same He atmosphere. The treated sample was then heated
from 50 oC to 800 oC at linear heating of 10 oC min-1 in 5% H2/N2 with a flow rate of 120 mL
min-1.
7

Powder X-ray diffraction (XRD) patterns of all fresh catalysts were obtained using an X-ray
diffractometer (Rigaku RINT 2500, Tokyo, Japan) with Ni-filtered CuKa radiation (λ=1.5406
Å) with a scan speed of 2o min-1 and a scan range of 2–80o at 30 kV and 15 mA. N2
physisorption was performed using Quadrasorb SI automated surface area and a pore size
analyzer. Before the N2 adsorption measurement, the sample was degassed at 200 oC until a
stable static vacuum of less than 5×10-3 torr was achieved. Subsequently, N2 adsorption was
performed at -196 oC. The surface area of the catalysts was measured using multipoint BET
in the relative pressure range of 0.05–0.3. The total pore volume was estimated at a relative
pressure of 0.99. XPS measurements were conducted on an Axis Nova spectrometer using a
monochromatic AlKα source (15 mA, 14 kV), calibrated using Au as the standard. For
energy calibration, the C1s photoelectron lines were used. The C1s binding energy was taken
as 285 eV. TEM images were acquired using an FEI Tecnai Osiris S/TEM equipped with a
super X-field emission gun (FEG), X-ray detection and Kratos charge neutralizer system.
STEM analysis was executed at an accelerating voltage of 200 kV. Thermogravimetric
analysis (TGA) of fresh and spent catalyst samples was performed using a Mettler Toledo
TGA/DTA 851e model (Switzerland) with a Stare software (version 6.1) by heating the
samples from 25 to 800 °C under an oxygen flow of 100 mL min−1 with a heating rate of 10
°C min−1.
2.3 Catalytic studies
2.3.1 Catalytic hydrogenation of pure 5-HMF to BHMF
The selective hydrogenation of 5-HMF to BHMF was carried out in a batch type custom-
made 25 mL mini reactor constructed with standard SS 316L Swagelok tubing and fitting,
equipped with a pressure gauge and magnetic stir bar. Before the reaction, the catalyst was
reduced in the flow of H2 in a tubular furnace at 350 oC for 2h. In a typical run, 0.25 mmol
HMF, 5 mL methanol, and 16 mg 20CA reduced catalyst were charged into the reactor. The
8

air inside the reactor was flushed three times with pure N2 and H2, and finally, the reactor
was pressurized with 3MPa H2. The reactor was placed in an oil bath on a hot plate and
stirred magnetically at 130 oC for 1 h. After the reaction, the reaction mixture was cooled to
room temperature and depressurized carefully. The catalyst was separated by filtration, and
the liquid products were analyzed by gas chromatograph-mass spectrometer (GC-MS)
(Agilent Technologies, 5977 A MSD with an SHRXI 5MS column (30 m × 250 mm × 0.25
mm)). The products were quantified by an external calibration method using known
standards. The conversion of 5-HMF and the yield of products were calculated as follows:
HMF conversion (%) = 푚표푙 표푓 퐻푀퐹 푟푒푎푐푡푒푑 × 100%
(1)
(2)
퐼푛푖푡푖푎푙 푚표푙 표푓 퐻푀퐹
BHMF yield (%) = 푚표푙 표푓 퐵퐻푀퐹 푝푟표푑푢푐푒푑 × 100%
푖푛푖푡푖푎푙 푚표푙푒 표푓 퐻푀퐹
2.3.2 Synthesis of 5-HMF from different mono, di and polysaccharides
The dehydration glucose, fructose, sucrose, cellobiose, starch and microcrystalline cellulose
was carried out in a custom made 30 mL batch type reactor equipped with a pressure gauge
and magnetic stir bar. In a typical run, 0.3 g feedstock, 3 mL water saturated with NaCl, 15
mL MIBK and 75 mg niobium phosphate (NbP) catalyst was loaded into the reactor. Before
the experiment, the reactor was flushed and pressurized with 0.5 MPa nitrogen gas to remove
the air inside the reactor and avoid boiling the reaction medium under the experimental
conditions. The reactor was placed in an oil bath on a hot plate and stirred magnetically at
160 oC for 1-6 h depends on the feedstock solution. After completing the reaction, the reactor
was removed from the hot plate and placed in an ice-water bath to quench the reaction. The
catalyst was separated by simple filtration, and the reaction mixture containing 5-HMF in
both aqueous and organic phases was quantified using high-performance liquid
chromatography (HPLC) using HPLC (Waters 2690 HPLC) equipped with a refractive index
detector (Waters 410) and photodiode array detector (Waters 2996) using an XBridge Amide
analytical column (3.5 μm, 4.6 × 250 mm), 80/20 acetonitrile/water with 0.2% triethylamine
9

as mobile phase with a flow rate of 0.6 mL/min. The substrate conversion and 5-HMF yields
were quantified by external calibration method with calibration curves constructed with pure
glucose, fructose and 5-HMF substrates.
3. Results
Surface and structural properties of copper-alumina catalysts were characterized by BET,
XRD, CO2 and NH3-TPD, H2-TPR, XPS, and HAADF-STEM techniques to gain insight
into the structural characteristics of the catalyst materials.
3.1 Structural characteristics of the catalyst materials
It should be noted that all synthesized catalysts were reduced under H2 at 350 0C for 2h prior
to characterization with different instrumental techniques.
3.1.1 BET
The textural properties of the catalyst can play a vital role in heterogeneous catalysis. Table 1
summarizes the BET surface area and total pore volume of the pure alumina and copper-
alumina based catalysts prepared by solid-state grinding, co-precipitation and impregnation
methods. As shown in Table1, the BET surface area and total pore volume of the pure Al2O3
prepared by the solid-state grinding method are 18.1 m2/g and 0.03 cc/g, respectively. Both
BET surface area and total pore volume of the copper-alumina catalysts are significantly
higher relative to the pure alumina, whereas depending on the copper loading. For example,
the specific BET surface area and total pore volume of the 10CA catalyst are 18.3 m2/g and
0.07 cc/g, respectively, while these values increase to 77.7 m2/g and 0.22 cc/g, respectively,
when increasing the copper loading to 20 mol% for 20CA catalyst. Whereas, further
increasing the copper loading to 25mol%, both specific surface area and pore volume of the
catalyst are significantly reduced to 41.2 m2/g and 0.16 cc/g, respectively. The average pore
size of all these catalysts was also measured, as presented in Table 1. The BJH pore size
10

distribution curves (Fig. S1) display that these catalysts contain both micropores and
mesopores and all the catalyst samples have a pore size distribution in the range of 1-20 nm.
On the other hand, the 20CA-P catalyst prepared by the conventional co-precipitation method
for comparison has a specific surface area of (71.7 m2/g) that is slightly lower than that of the
20CA catalyst (77.7 m2/g), but there is a noticeable difference in the total pore volume of
these two catalysts. For comparison, the 20CA-I catalyst was also prepared by impregnation
of copper on commercial Al2O3 support that has a high surface area (200 m2/g). After
impregnation of copper, the BET surface area is significantly reduced compared to the bare
Al2O3 support. The pore size distribution of these three catalysts prepared through different
methods is shown in Fig. S2. Both 20CA-P and 20CA-I catalysts contain micro- and
mesopores and the particle size distribution is in the range of 1-100 nm. Overall, compared
with the pure Al2O3, the one-step preparation of copper-alumina catalysts via the solvent-free
solid-state grinding method could improve the textural properties by forming a Cu-O-Al bond
that enhances the catalyst porosity [46,47].
3.1.2 XRD
The powder X-ray diffraction measurements are used to determine the crystalline structure of
the catalyst samples. Fig. 1 shows the XRD patterns of supported copper catalysts. As shown
in Fig.1, the bare Al2O3 does not exhibit any XRD peaks, indicating the amorphous nature of
the support. On the other hand, the XRD patterns of copper-alumina catalysts reduced in the
flow of H2 exhibit the main crystalline peaks at 2θ values of 36.6o, 43.30, 50.3o, and 74.1o.
The first peak at 2θ = 36.6o corresponds to the spinel structure of copper aluminate (CuAl2O4)
phase JCPDS 33-0448 [48–50]. It can be observed that the crystallinity of the copper
aluminate phase increases with the increase of copper loading. The other three XRD peaks at
around 43.3o, 50.3o, and 74.2o correspond to the (111), (200), and (220) planes of metallic
Cu, respectively, which can be indexed to the fcc structure of Cu (JCPDS No. 85–1326)
11

[51,52]. The metallic copper peaks' intensity increases with the increase of copper loading
from 10 to 20 mol% but slightly decreases with a further rise of copper loading to 25 mol%.
In contrast, the 20CA-P catalyst displays only one broad XRD peak at 2θ of 36.8o, which
corresponds to the copper aluminate phase, and no other peaks were detected, suggesting the
amorphous nature of the catalyst or highly dispersed status of the copper species in the Al2O3
matrix. In contrast, the XRD pattern (Fig.S3) of the 20CA-I catalyst prepared by
impregnation of copper on commercial γ-Al2O3 shows both metallic copper and γ-Al2O3
peaks, implying poorer dispersion of the copper particles in the 20CA-I than that of the
20CA-P.
Furthermore, the crystallite sizes of copper-aluminate spinel and metallic Cu species of
different copper-alumina catalysts were calculated using the Scherrer equation based on the
highest intensity XRD peak at 2θ = 36.6o and 43.3o, respectively, and the crystallite size
values are presented in Table 1. These results demonstrate that the crystallite sizes of both
species increase with the increase of copper loading.
Al2O3
10CA
15CA
20CA
25CA
20CA-P
@Cu0
# CuAl2O4
@
@
@
#
5
15
25
35
45
55
65
75
85
2-Theta (degree)
12

Figure 1. XRD patterns of Al2O3 and different copper-alumina catalysts
3.1.3 NH3-TPD
The acidic properties of the copper-alumina catalysts were measured by NH3-TPD, and the
results are presented in Fig. 2. The total acidity values are summarized in Table 1. As shown
in Fig. 2, the NH3-TPD pattern of the bare Al2O3 exhibits a one broad NH3 desorption peak in
the temperature range of 300 to 800oC, indicates the presence of medium and strong acid sites
on the catalyst surface. However, the copper-alumina mixed oxide catalysts prepared by
solid-state grinding followed by calcination and reduction, the acidity of these catalysts was
slightly altered, depending on the copper loading. Except for the 10CA catalyst, all other
copper-alumina catalysts exhibit low temperature and high-temperature NH3 desorption
peaks. It indicates weak, medium and strong acid sites present on the catalyst surface. The
low-temperature NH3 desorption peak could be attributed to Lewis acid sites, essential for
hydrogenation reactions because they act as electron-pair acceptors [53]. Souza et al. [54]
reported that with the incorporation of copper into alumina, the total acidity of the catalyst is
decreased relative to pure Al2O3 caused by the covering of acid sites with metal clusters.
Nevertheless, in the present study, increasing the copper content from 10 mol% to 25 mol%,
the different copper-alumina catalysts' total acidity did not follow a particular trend.
Compared with pure Al2O3, both 15CA and 20CA catalysts exhibit increased acidity, whereas
10CA and 25CA catalysts display reduced acidity values.
Furthermore, the 20CA-P catalyst also exhibits weak, medium, and strong acid sites;
however, while compared with the 20CA catalyst, some difference in the peak intensity was
observed, especially at low and high- temperature regions. As shown in Fig. 2, the 20CA-P
catalyst contains moderately strong acid sites, as indicated by a high-intensity peak at the
high-temperature region. However, the total acidity of the 20CA-P catalyst is moderately
13

low, if compared with the 20CA catalyst (Table 1). This difference might be attributed to the
different methods adopted to prepare these catalysts.
Al2O3
10CA
15CA
20CA
25CA
100
200
300
400
500
600
700
800
Temperature (oC)
Figure 2. NH3-TPD profiles of the copper-alumina catalysts
3.1.4 CO2-TPD
CO2-TPD measurements were used to evaluate the copper-alumina catalysts' surface basicity,
and the corresponding CO2-TPD profiles are illustrated in Fig. 3. The CO2-TPD profiles of all
catalyst materials, including the pure Al2O3, display a strong CO2 desorption peak in the
temperature range of 300 to 800 oC, which could characterize the presence of medium to
strong basic sites [55]. The medium strength of basic sites could be originated from the
interaction of CO2 with the Lewis acid-base pairs (M+- O-), forming the monodentate or the
bidentate carbonate species [56]. The strong basic sites could be ascribed to the interaction of
CO2 with the coordinately unsaturated O2- species [55].
The total basicity of all these catalysts was quantified, and the values are presented in Table
1. For comparison, the total basicity value of pure Al2O3 prepared by solid-state grinding is
14

also listed in Table 1, being the maximum value of 0.49 mmol/g. The total basicity of the
copper-alumina catalysts was altered significantly relative to that of the pure Al2O3. From
Table 1, the total basicity of these catalysts decreases in the following order: Al2O3 > 10 CA
> 20 CA-P > 20 CA >25 CA > 15 CA.
Al2O3
10CA
15CA
20CA
25CA
20CA-P
100
200
300
400
500
600
700
Temperature (oC)
Figure 3. CO2-TPD profiles of the pure Al2O3 and different copper-alumina catalysts
3.1.5 H2-TPR
H2-TPR technique provides information about the reducibility and metal-support interaction
of a catalyst [57]. The H2-TPR profiles of the unreduced copper-alumina catalysts are shown
in Fig. 4. The copper-alumina catalysts prepared by solid-state grinding method exhibits a
single reduction peak in the temperature range of 150 oC to 500 oC, which could characterize
the reduction of CuO to metallic copper [58]. As expected, the intensity of the hydrogen
consumption peak increases with the increase of copper loading, which is consistent with the
observations in the study of Dieuzeide et al. [51]. According to the literature survey, the
maximum reduction temperature of coper-alumina catalysts is below 300 oC [57,59],
15

whereas, in the present study, it is reduced at a maximum temperature of 345 oC, which is in
good agreement with Dong et al. work [60].
Nevertheless, compared with the literature results [57,59,61], the reduction temperature of
copper catalyst shifts to a high-temperature region, which might be due to the difference in
the catalyst preparation methods. Indeed the copper-aluminate compounds formed by solid-
state reactions during the calcination process would result in strong interaction and a higher
reduction temperature [59,60]. On the other hand, the reduction peak slightly shifts to a lower
temperature (325 oC) in the 20CA-P catalyst, indicating weak interaction between copper and
alumina.
20CA
10CA
15CA
25CA
20CA-P
100
200
300
400
500
600
Temperature (oC)
Figure 4. H2-TPR profiles of the unreduced copper-alumina catalysts
3.1.6 XPS
XPS was used in this work to quantify the surface atomic composition and identify each
element's oxidation state on the catalyst surface. The survey scan spectra of the 20CA catalyst
display copper, alumina, and oxygen elements (Fig. 5a). The deconvoluted XPS spectra of Cu
2p (Fig. 5b) show four different binding energy peaks. The binding energy peak at 930.8 eV
16

and another peak at 950.01 eV could be ascribed to Cu 2p3/2 and Cu 2p1/2, signifying the
presence of Cu0 oxidation state [58]. Whereas the XPS peaks at 932.4 and 952 eV
corresponds to the Cu 2p3/2 and Cu 2p1/2 of Cu2+ oxidation state[62]. It should be noted
that the presence of Cu1+ cannot be ruled out due to practically the same binding energies of
both Cu0 and Cu1+ [62,63]. However, compared with literature results[64–66], the binding
energy values of both Cu0 and Cu2+ slightly shift to lower values, implying the strong
interaction between Cu and Al2O3 due to the electrons transfer from alumina to the active
metal, which is well supported by the results of H2-TPR measurements. In addition to the
(a)
(b)
2p3/2
Cu2+
2p1/2
Cu 0
17
Binding energy (eV)

Figure 5. XPS survey scan spectra (a) and Cu 2p XPS spectra (b) of the 20CA catalyst
main peaks, small satellite peaks of Cu 2p3/2 and Cu 2p1/2 were detected at 938.1 and 962.5
eV, respectively, which are formed due to the ejection of photoelectron with another valence
band electron and/or to the metal-to-ligand charge transfer, indicating the presence of Cu2+
ions in the 20CA catalyst[67]. Thus, the XPS results reveal that the catalyst surface is not
completely reduced and contains oxidized Cu2+ ions. Furthermore, these results are well
supported with the XRD results, where the presence of the CuAl2O4 phase was detected in the
XRD pattern of the 20CA catalyst, which indicates that the presence of oxidized copper in the
catalyst. Moreover, the XPS results also align well with the H2-TPR measurement, where the
maximum reduction temperature of 20CA catalyst is 345 oC, and the copper ions are not
entirely reduced before 350 oC, i.e., the reduction temperature used in this study (Fig. 4).
Therefore, XPS, XRD, and H2-TPR results signify the presence of both metallic and oxidized
copper species in the 20CA catalyst.
3.1.7 STEM
Fig. 6a shows the high-angle-annular darkfield scanning electron microscopy (HAADF-
STEM) images of the 20CA catalyst, and the corresponding particle size distribution
histogram is shown in Fig. 6b. As can be observed from Fig. 6a, many bright spots are
dispersed on the alumina surface, indicating the atomic dispersion of metallic copper
particles. The particle size distribution histogram displays that the copper particle size is in
the range of 1-9 nm, and the average particle diameter is 2.6 nm. The elemental mappings of
the 20CA catalyst are shown in Figs. 6c-6e, which demonstrates uniform dispersion of copper
on alumina across the surface (Fig. 6f).
18

40
30
20
10
0
Mean = 2.565 SD = 1.556
(a)
(b)
0
2
4
6
8
10
12
14
Particle size (nm)
(c)
(d)
(f)
(e)
Figure 6. Representative HAADF-STEM image of the 20CA catalyst (a), copper particle size
distribution (b), and elemental mapping Cu (c), Al (d), O (e), and overlaid Cu/O/Al (f).
3.2 Catalytic performance
19

3.2.1. Effects of reaction temperature and time
The copper-alumina catalysts prepared through the solvent-free solid-state grinding method
were evaluated for selective hydrogenation of biomass-derived 5-hydroxymethylfurfural to
2,5-bis(hydroxymethyl)furan (BHMF). Generally, the hydrogenation of 5-HMF can produce
different reaction products, such as BHMF, 5-methylfurfural (5-MF), 5-methyl-2-
furanmethanol (MFM), 2,5-dimethylfuran (DMF), and the ring-opening rearrangement
products [68–70]. However, selective hydrogenation of 5-HMF to BHMF can be achieved by
varying the reaction parameters, particularly the reaction temperature and time. In this study,
initially, the effects of reaction temperature and time on selective hydrogenation of 5-HMF
over 20CA catalyst were examined under the following reaction conditions: 5-HMF/cat. 2:1
w/w, methanol (5 mL), and H2 pressure (3 MPa). The hydrogenation of 5-HMF was carried
out between the temperature range of 110 to 140 oC. As shown in Fig. 7a, the conversion of
5-HMF was very low, especially at the lower reaction temperature (110 oC and 120 oC).
However, it gradually increased with the increase of reaction temperature and time. Indeed an
almost complete conversion of 5-HMF was achieved at 130 oC after a one-hour course of
reaction time. If further increasing the reaction temperature and time, a constant conversion
of 5-HMF could be achieved. In contrast, an extended reaction time (2h) was required for the
complete conversion of 5-HMF at lower reaction temperatures (110 oC and 120 oC). In this
conversion, BHMF was the main reaction product, and 5-MF a byproduct, whose yields
varied with reaction temperature and time.
Figs. 7b and 7c show the variation of BHMF and 5-MF yields with the reaction temperature
and time. Both BHMF and 5-MF yields increased while increasing the reaction temperature
from 110oC to 140oC. A maximum of 94% and 91% BHMF yield were obtained at 130 oC
and 140 oC after one hour course of time. When extending the reaction time to 2 h, a
considerable decrease in BHMF yield and a simultaneous rise in 5-MF yield were noticed.
20

The formation of 5-MF results from the hydrogenolysis reaction of 5-HMF at higher reaction
temperatures [71,72]. Accordingly, 130 oC for 1h appears to be the best reaction conditions
for selective hydrogenation of 5-HMF to BHMF over the 20CA catalyst.
30 min
60 min
90 min
120 min
(a)
100
80
60
40
20
0
110
120
130
140
Reaction temperature (oC)
100
(b)
30 min
60 min
90 min
120 min
80
60
40
20
0
110
120
130
140
Reaction temperature (oC)
21

30
25
20
15
10
5
30 min
60 min
90 min
120 min
(c)
0
110
120
130
140
Reaction temperature (oC)
Figure 7. Effects of reaction temperature and time on selective hydrogenation of 5-HMF to
BHMF and 5-MF over 20CA catalyst: 5-HMF conversion (a), BHMF yield (b) and 5-MF
yield (c). Other reaction conditions: 5-HMF/catalyst 2:1 w/w, methanol 5 mL, 3 MPa H2.
3.2.2 Effects of copper loading
Next, we studied the effects of copper loading for the selective hydrogenation of 5-HMF
under the above-optimized reaction conditions of temperature and time (130oC, 1h), and these
results are depicted in Fig. 8. Again, it should be mentioned that 5-MF formed as a reaction
byproduct. Both 5-HMF conversion and BHMF yield increased while increasing the copper
content from 10 mol% to 20 mol%. More than 99 % of 5-HMF conversion and 93% of
BHMF yield were achieved over the 20CA catalyst. However, with a further increase of
copper loading to 25 mol%, an almost constant conversion of 5-HMF and BHMF yields were
achieved. In fact, the BHMF yield was slightly reduced. Thus, these results demonstrate that
the 20 mol% Cu loading could be optimum for this reaction.
The variation in the catalytic activity and the BHMF yield among different copper-alumina
catalysts could be related to the differences in the structural properties of the catalysts. It was
believed that metallic and electrophilic copper species are played a crucial role in the
selective hydrogenation of 5-HMF to BHMF [31]. As illustrated previously in Fig. 1, the
22

relative XRD peak intensity of metallic copper, particularly at 2θ = 43.3o, was found
maximum in the 20CA and 25CA catalysts and copper-aluminate species were observed in
all catalyst samples. Whereas, the presence of uniformly distributed metallic copper
nanoparticles (Fig. 6) was believed to play a pivotal role in the greater production of BHMF
[31].
100
80
60
HMF conversion (%)
BHMF yield (%)
40
MF yield (%)
20
0
10CA
15CA
20CA
25CA
Catalyst
Figure 8. Effects of copper loading in the copper-alumina catalysts. Other reaction
conditions: 5-HMF/catalyst 2:1 w/w, methanol 5 mL, 3 MPa H2, 130 oC, 1 h.
3.2.3 Effects of reaction media
The reaction medium could significantly influence the selective hydrogenation of 5-HMF to
BHMF [18]. In this study, the hydrogenation of 5-HMF was carried out in different solvents
over the 20CA catalyst under the so far optimized reaction conditions, and these results are
summarized in Table 2. Initially, the hydrogenation of 5-HMF was conducted in polar protic
solvents like water, methanol, and ethanol. A complete conversion of 5-HMF was achieved in
pure water solvent, but the BHMF yield was very low (< 35%). Excellent yields of BHMF
(93% and 90%) were achieved when the hydrogenation of 5-HMF was carried out in
methanol or ethanol reaction medium, as similarly observed in the literature [18].
23

In contrast to the previous literature reports [16, 18], the polar aprotic solvents such as methyl
isobutyl ketone (MIBK), THF, and acetonitrile solvents were also efficient for hydrogenation
of 5-HMF, producing 90%, 74%, and 89% of BHMF yields, respectively. Since MIBK and
THF were used as biphasic reaction media to synthesize 5-HMF from sugars[73], it would be
advantageous to directly convert biomass-derived sugars into BHMF without isolation of 5-
HMF intermediate. On the other hand, low yields of BHMF were reported with nonpolar
organic solvents (toluene and hexane) [18]. Therefore, the polar protic (methanol and
ethanol) and polar aprotic (MIBK and acetonitrile) solvents appear to be best for the selective
hydrogenation of 5-HMF to BHMF over a 20CA catalyst. Since a little higher BHMF yield
was obtained with the polar protic methanol solvent, methanol was used as the reaction
medium to study the effects of other reaction conditions.
3.2.4 Influence of hydrogen pressure
Hydrogen pressure also could influence the selective hydrogenation of 5-HMF to BHMF. In
this study, the catalytic performance of the 20CA catalyst was investigated by varying the
hydrogen pressure from 1 MPa to 4 MPa, while maintaining the other reaction parameters
constant. The catalytic activity results are depicted in Fig. 9. It clearly shows that the 5-HMF
conversion increased with the increase of hydrogen pressure and reached a maximum with 3
MPa of H2 used. A further rise in hydrogen pressure to 4 MPa led to nearly constant
conversion of 5-HMF. BHMF yield increased from 70 to 93%, along with increasing H2
pressure from 1MPa to 3 MPa, but an almost constant yield of BHMF (92%) was obtained
when further increasing hydrogen pressure to 4 MPa. Besides BHMF, 5-MF was also
produced under all these reaction conditions, but at a yield below 10%. In the absence of
hydrogen, 5-HMF could convert to other products such as humins, although they were not
detected by GC-MS likely due to their poor volatility. Consequently, 3 MPa of hydrogen
pressure was chosen for studying other reaction parameters.
24

HMF conversion (%) BHMF yield (%) MF yield (%)
100
80
60
40
20
0
1
2
3
4
Hydrogen pressure (MPa)
Figure 9. Effects of hydrogen pressure on selective hydrogenation of 5-HMF over 20CA
catalyst. Other reaction conditions: HMF/catalyst 2:1 w/w, methanol 5 mL, 130 oC, 1 h
3.2.5 Effects of catalyst preparation method
In this study, the 20CA-P catalyst was prepared through the conventional co-precipitation
method and its catalytic activity was compared with the 20CA catalyst prepared by the
solvent-free solid-state grinding method. As shown in Fig. 10, both catalysts attained
complete conversion of 5-HMF but exhibited a slight variation in the product distribution.
The 20CA catalyst prepared by the solvent-free grinding method is slightly superior to the
20CA-P catalyst, although ICP-AES analysis found that these two catalysts contain different
amounts of copper (18.4 mol% Cu in 20CA vs 15.1 mol% Cu in 20CA-P). Thus, the little
variation in the BHMF yield might be attributed to the difference in copper content between
these two catalysts.
25

100
80
60
40
20
0
HMF conversion (%)
BHMF yield (%)
MF yield (%)
Grinding
Precipitation
Figure 10. Effects of catalyst preparation method on selective hydrogenation of 5-HMF to
BHMF. Other reaction conditions: HMF/catalyst 2:1 w/w, methanol 5 mL, 3 MPa H2, 130
oC, 1 h.
3.2.6 Integrated catalytic process for conversion of mono, di, and polysaccharides to BHMF
So far, most of the literature studies were focused on the hydrogenation of commercially
available expensive 5-HMF. However, direct synthesis of BHMF from biomass-derived
carbohydrates via catalytic dehydration through the formation of 5-HMF intermediate
followed by hydrogenation is more economically attractive and sustainable. Niobium
phosphate (NbP) catalyst proved active for converting glucose and industrial-grade sugars
into 5-HMF in a continuous flow process [74]. In this study, the dehydration of biomass-
derived mono, di and polysaccharides (10 wt%) was first carried out in a biphasic reaction
medium water and MIBK using a NbP catalyst in a batch process at 160 oC for 1 to 6h. After
the dehydration reaction, the NbP catalyst was separated by simple filtration, and the filtrate
could be used directly for the second step hydrogenation reaction. However, since water is
not a suitable reaction medium for the conversion of 5-HMF to BHMF over the 20CA
catalyst (Table 2, entry 1), it was removed by evaporation after the dehydration reaction,
26

followed by the extraction of 5-HMF with MIBK. The organic phase containing 5-HMF was
then directly used without any further treatment for the second step hydrogenation reaction.
As shown in Table 3, monosaccharide glucose was effectively dehydrated over the NbP
catalyst to form 5-HMF, resulting in 51.5% yield, followed by hydrogenation over 20CA
catalysts, producing BHMF at 15% overall yield under the reaction conditions of (3MPa H2
pressure, 1300C, 1h). On one hand, it needs to note that the ketone group in a MIBK solvent
is unreactive under these conditions. On the other hand, the low catalytic activity of the 20CA
catalyst under these conditions is more likely due to the competition between the adsorption
of the 5-HMF and some MIBK-soluble impurities generated during the first-step dehydration
reaction. However, increasing the reaction temperature from 130 to 150oC and extending the
reaction time to 4.5 h, achieved complete conversion of 5-HMF and obtained a maximum
overall yield (42.2%) of BHMF. Henceforth, these reaction conditions were used for the
conversion of other biomass-derived 5-HMF to BHMF. It was found that the monosaccharide
fructose could also easily undergo dehydration over the NbP catalyst, producing 58% 5-HMF
yield and attained 48% overall yield of BHMF after hydrogenation of 5-HMF. In addition,
disaccharides such as cellobiose and sucrose could be easily converted into BHMF under the
above established reaction conditions, achieving 40% and 39.7% overall yield, respectively.
Furthermore, the integrated catalytic process was also demonstrated applicable for converting
polysaccharides such as starch and microcrystalline cellulose, obtained 34.4% and 25.2%
BHMF overall yield, respectively. From these above results, it can conclude that it is viable
to produce BHMF from carbohydrates via an integrated catalytic process. Our group has
recently developed a bifunctional carbon-based solid acid catalyst for production of 5-HMF
(93%) from glucose [75]. Such carbon-based solid acid catalyst will be tested in the
integrated process along with the 20CA catalyst to produce BHMF from carbohydrates.
27

4. Discussion
4.1 Reaction mechanism
5-HMF contains reactive C=C and C=O functional groups; however, the bond energy of C=O
(715 kJ mol-1) is much higher than that of C=C (615 kJ mol-1), allowing the hydrogenation
of furan ring take place easier especially in the presence of a metal catalyst [76].
Nevertheless, owing to the conjugation of the carbonyl group with furan ring in a 5-HMF
molecule, the selective hydrogenation of C=O for the synthesis of BHMF could occur at the
lower reaction temperature. Based on XPS results, the 20CA catalyst contains both Cu2+ and
Cu0 species on the catalyst surface. These could be active sites for hydrogenation of carbonyl
group in a 5-HMF molecule in the presence of external hydrogen. Controlled experiments
were carried out to understand the role of both Cu2+ and Cu0 species' in the selective
hydrogenation of 5-HMF. With unreduced 20CA catalyst in the absence of Cu0 species, only
< 5% yield of BHMF was obtained, although the 5-HMF was completely converted. Very
recently, Li et al. [77] and Hellgardt et al.[78] reported RANEY® Cu catalyst for
hydrogenation of 5-HMF to BHMF, obtained 59% and 84% BHMF yield, respectively.
Nevertheless, in the present study, while using the metallic copper nanopowder (Cu0) as a
catalyst, very low yield of BHMF (< 15%) were obtained under the established reaction
conditions. Surprisingly, with physically mixed copper nanopowder and unreduced 20CA
catalyst in an equal mass and a substrate-to-catalyst mass ratio of 2, the BHMF yield was
abruptly improved to 62%, much higher than that with the unreduced 20CA or copper
nanopowder alone. These mechanistic studies demonstrate that the co-existence of both
metallic and electrophilic copper species is likely a prerequisite for selective hydrogenation
of 5-HMF, as similarly proposed in the literature [79,80]. Based on the above experimental
results and the literature reports [79,81,82], a plausible reaction mechanism for selective
hydrogenation of 5-HMF to BHMF over copper-alumina catalyst is proposed, as shown in
28