Microwave-assisted catalytic conversion of glucose to 5-hydroxymethylfurfural using "three dimensional" graphene oxide hybrid catalysts

Article abstract of DOI:10.1039/d0ra01009j

Hybrids of reduced graphene oxide (rGO) and metal/metal oxide (Pt, NiO/Ni(OH)₂, CoO, Fe₃O₄) nano particle were prepared by reduction of graphene oxide (GO) and metal ion (Pt²⁺, Ni²⁺, Co²⁺, Fe²⁺) hybrids. The M-rGO hybrids (M = Pt, Ni-, Co and Fe) were justified for the transformation of glucose to 5-hydroxymethylfurfural (5-HMF). High glucose → 5-HMF conversion was yielded depending on the nature of the M-rGO catalyst. The Ni-rGO showed the highest 5-HMF yield. The conversion reaction tuned to the optimized state under a microwave-assisted reaction accomplished by using Ni-rGO. In such case, the conversion rate was 99% with a 5-HMF yield of 75%. In order to improve both the conversion and yield, NiGO-FD was prepared by a freeze-dry method. The NiGO-FD remarkably showed the highest conversion of 99% and 5-HMF yield of 95%. Beside the biomass transformation process, the physico-chemical strategy employed herein for multiplying the catalytic efficiency might be justified for catalyzing similar reactions.

Full text of DOI:10.1039/d0ra01009j

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Microwave-assisted catalytic conversion of glucose
to 5-hydroxymethylfurfural using “three
Cite this: RSC Adv., 2020, 10, 11727
dimensional” graphene oxide hybrid catalysts†
a
ab
a
a
Yui Hirano,* Jorge N. Beltramini, Atsushi Mori, Manami Nakamura,
cd
a
a
Mohammad Razaul Karim,
and Shinya Hayami
Yang Kim, Masaaki Nakamura
ae
*
Hybrids of reduced graphene oxide (rGO) and metal/metal oxide (Pt, NiO/Ni(OH)
2
, CoO, Fe
3
O
2+
4
) nano
2+
2
+
2+
particle were prepared by reduction of graphene oxide (GO) and metal ion (Pt , Ni , Co , Fe
)
hybrids. The M-rGO hybrids (M ¼ Pt, Ni–, Co and Fe) were justified for the transformation of glucose to
-hydroxymethylfurfural (5-HMF). High glucose / 5-HMF conversion was yielded depending on the
nature of the M-rGO catalyst. The Ni-rGO showed the highest 5-HMF yield. The conversion reaction
tuned to the optimized state under a microwave-assisted reaction accomplished by using Ni-rGO. In
such case, the conversion rate was 99% with a 5-HMF yield of 75%. In order to improve both the
conversion and yield, NiGO-FD was prepared by a freeze-dry method. The NiGO-FD remarkably showed
the highest conversion of 99% and 5-HMF yield of 95%. Beside the biomass transformation process, the
physico-chemical strategy employed herein for multiplying the catalytic efficiency might be justified for
catalyzing similar reactions.
5
Received 2nd February 2020
Accepted 14th March 2020
DOI: 10.1039/d0ra01009j
rsc.li/rsc-advances
1
selectivity. These methods rst fractionate the lignocellulosics
into its main components, cellulose, hemicellulose and lignin,
1
. Introduction
The current oil price volatility combined with high-level which allows for processing each fraction at different condi-
2
greenhouse gas emissions increases the possibility of an tions to achieve high yields of target products.
energy crisis amidst soaring global energy consumption that
Cellulose accounts for 30–50 wt% of lignocellulosic biomass
has driven the need to seek alternative energy sources. As such, and is a linear homopolymer made up of anhydro-D-glucopyr-
3
lignocellulosic biomass is a promising renewable feedstock for anose units linked via b-glycosidic bonds. Cellulose has
both carbon-based energy sources and sustainable chemical a degree of polymerization (DP) up to 15 000 glucopyranose
4
production. One of the main challenges in converting ligno- units with a network of intra- and inter-molecular hydrogen
cellulosic biomass is producing chemicals or fuels at high bonding. The formation of intramolecular hydrogen bonding
selectivities and yields at economical costs. This happens within cellulose structure is responsible for its crystallinity,
5
because of the recalcitrance of lignocellulosic biomass, thus rigidity and chemical inertness. Cellulose can be converted
6,7
requiring both physical and chemical pretreatments, which are into glucose by chemical or enzymatic hydrolysis and can be
still one of the most expensive stages of biomass conversion used to produce ethanol, platform chemicals, such as levulinic
strategy. Typically, two-step processing methods are employed acid (LA) and 5-hydroxymethylfurfural (5-HMF), and liquid
8,9
to control the reactivity of lignocellulosics and improve product fuels.
Amongst the numerous routes of lignocellulosic
biomass valorization, catalytic dehydration of cellulose to
produce 5-hydroxymethylfurfural (5-HMF) represents an
a
Department of Chemistry, Graduate School of Science and Technology, Kumamoto
10,11
attractive strategy.
This is because of the versatility of using
University, 2-39-1 Kurokami, Chuo-ku, Kumamoto 860-8555, Japan. E-mail:
hayami@kumamoto-u.ac.jp
5-HMF as an important platform molecule to synthesize an
b
array of intermediate chemicals, from which a diverse range of
biofuels as well as commodity and ne chemicals can be
generated as shown in Scheme 1. For example, 5-HMF can be
selectively oxidized to furan-2,5-dicarboxylic acid (FDCA) which
may be used as a replacement for terephthalic acid in the
production of polyethyleneterephthalate. Alternatively, 5-HMF
can be reduced to 2,5-bis(hydroxymethyl)furan (BHMF) and 2,5-
bis(hydroxymethyl)tetrahydrofuran (BHMTF) wherein both can
Centre for Tropical Crops and Bio-Commodities, Queensland University of
Technology, Brisbane, 4000, Australia
c
Chemistry Department, King Abdulaziz University, Jeddah 21589, Saudi Arabia
d
Department of Chemistry, School of Physical Sciences, Shahjalal University of Science
and Technology, Sylhet-3114, Bangladesh
e
Institute of Pulsed Power Science (IPPS), Kumamoto University, 2-39-1 Kurokami,
12
Chuo-ku, Kumamoto 860-8555, Japan
†
Electronic supplementary information (ESI) available. See DOI:
0.1039/d0ra01009j
1
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 11727–11736 | 11727

View Article Online
RSC Advances
Paper
species. In addition, the oxygen functional group on the gra-
phene oxide (GO) surface can be deoxygenated by heat, light,
hydrazine, and it becomes reduced type graphene oxide (rGO),
where its many defect structure has a high electron conductivity
property. Furthermore, since rGO has a two-dimensional
structure, it can not only provide surface area by doping
a metal but also is useful as a reaction eld, and it is expected
that it can be used as a catalyst with higher activity than the
32
conventional commercial catalyst. Thus, metal-supported
graphene oxide (M-GO) composite materials have been widely
used in different applications because of their high surface
area, stability under ambient conditions, and faster electron-
transport mechanism. Enhanced catalytic activities of several
metal such as Pd, Pt, Ni, Au, Ru supported on rGO have been
Scheme 1 5-HMF production and its utilization routes for biochemi-
cals and fuels.
serve as alcohol components in the production of polyesters,
providing completely biomass-derived polymers when
33,34
reported recently.
13
combined with FDCA. 5-HMF may also be rehydrated to lev-
ulinic acid from which gamma valerolactone can be produced,
Hence, in this work we focused on the effect of the synthesis
techniques of M-rGO (Pt, Ni, Co and Fe) for the microwave-
assisted catalytic decomposition of carbohydrates to 5-HMF,
where the effect of reaction conditions, amount of catalyst used,
effect of solvent, were also evaluated.
14
which can be utilized as a solvent and fuel additives. Another
important derivative of 5-HMF is 2,5-dimethylfuran (DMF) and
2
-methylfuran (2-MF), and both are potential liquid trans-
15
portation fuels. Moreover, 5-HMF can also serve as a precursor
16
in the synthesis of liquid alkanes for use as diesel fuel.
-HMF is preferentially obtained by acid-catalyzed dehydra-
tion of C carbohydrates, such as fructose and glucose using
2
. Experimental section
5
6
2.1. Materials
17,18
a variety of homogeneous and heterogenous catalysts.
The synthesis of M-rGO and on graphene oxide freeze-dried
However, because the synthesis efficiency of 5-HMF is poor with
low selectivity and yield, it is crucial to develop a low-cost and
highly efficient synthesis method. Over the years different
heterogeneous catalysts such as metal-supported mesoporous
zeolites, oxides, zeolites, ion exchange resins were extensively
(MGO-FD) are illustrated in Fig. S1.† In short, a metal chlo-
ride solution containing the prescribed amount of metal was
added into the GO dispersion as both linkages among graphene
sheets and metal precursors. As depicted in Scheme 2, during
ꢀ
the synthesis of M-rGO aer reduction at 120 C with hydrazine
19–22
investigated to improve 5-HMF conversion and selectivity.
From the results, it was clearly concluded that if a carbohydrate
cellulose, glucose, starch) was used as feed, a balance is needed
between the Lewis and Brønsted acid sites to isomerize glucose
solution under mild reaction conditions, a gel-like structure
could be formed between metal and rGO as a result of con-
necting rGO sheets with metal ions.
(
23,24
to fructose and dehydrate fructose to 5-HMF respectively.
It
2.2. Synthesis of graphene oxide
is also well known that the support plays a critical role in the
25
catalytic performance of a heterogeneous catalyst.
Graphene oxide was prepared by the modied Hummers'
35
As a new member of the carbon family, the outstanding method (Scheme 1). 1 g of graphite powder (95%) and 1 g of
properties of graphene make it an ideal candidate to support NaNO were mixed in a 500 mL round-bottom ask placed into
3
26,27
catalyst and metal precursors.
Graphene is a two- an ice bath. 50 mL of conc. H SO was added to it and the mixture
dimensional crystal that can be considered as a building was stirred for 30 min at 0 C. Aerward, 7 g KMnO was added
block for carbon materials of different structures. Graphene has and the mixture was then stirred for 40 min at 35 C. As such,
2 4
ꢀ
4
ꢀ
been known for long time since graphite is formed by 50 mL of distilled water was slowly added, and the mixture was
ꢀ
a combination of graphenes with van der Waals forces. stirred for another 30 min at 90 C. The reaction was stopped by
However, details of the properties were unclear until late years, adding 100 mL of distilled water followed by 3 mL of 30% H O2
2
because an isolation procedure of graphene from graphite was solution. The mixture was cooled at room temperature. Aer
28
not well developed for long time. As such, recently, graphene cooling, the product was centrifuged at 3000 rpm to remove the
has attracted comprehensive research interest because of its supernatant liquid. The remaining solid mixture was washed
extraordinary thermal, mechanical, electronic, and optical with distilled water and centrifuged at 3000 rpm to remove the
29
properties. Metal supported graphene nanocomposites are
30
a newly group of catalysts subject to current research. The
inherent properties of graphene can be enhanced by using it as
a support for metal nanoparticles. The large contact area
provided by the planar structure of graphene sheets can act as
31
a wonderful support for trapping guest metal particles.
Moreover, graphene sheets have the unique ability to promote
fast electron-transfer kinetics for a wide range of electroactive
Scheme 2 Synthetic strategy for M-rGO.
11728 | RSC Adv., 2020, 10, 11727–11736
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
supernatant liquid 3 times. The obtained solution was added into the analytical column and both RID (refractive index) and RID
a 500 mL Erlenmeyer ask and sonicated in an ultrasound bath (UV-Vis) detectors. The HPLC was operated under the following
ꢀ
for 2 h. This dispersion was centrifuged at 8000 rpm for 1 h to conditions: oven temperature, 50 C, mobile phase, 5 mM
ꢁ1
remove the impurities. This supernatant liquid was called gra- H SO ; ow rate, 0.6 mL min ; injection volume, 5 mL. The
2
4
phene oxide (GO) dispersion.
concentrations of glucose and 5-HMF were quantied through
the external standard method and calibration curves of
commercially available standard substrates.
2
.3. Synthesis of M-rGO hybrids
A 100 mL GO solution was mixed with 100 mg of metal chloride
solutions (NiCl , CoCl , FeCl ) to prepare M-GO solution. To
ꢀ
ꢁ
detected glucose ðmolÞ
Conversion ðmol%Þ ¼ 1 ꢁ
2
2
2
initial glucose input ðmolÞ
this M-GO solution, 10 mL hydrazine was added, and the
mixture was reuxed for 3 days at 120 C. Aer ltering the
powder thus obtained was the M-rGO solution.
ꢀ
ꢂ100
%
Y ¼
2
.4. Synthesis of NiGO-FD
ꢂ
ꢃ
number of moles of the product
ꢁ1
100 mL GO solution (1 g L ) was mixed with 100 mg nickel
I moles of glucose in feed ꢁ F moles of glucose in products
ꢂ 100
chloride (NiCl
2
). Obtained Ni-GO solution was freeze-dried by
placing the solution in a freeze-drying ask rotated in a bath,
with cooling facilities using dry ice. Aerward, the temperature
was raised to remove any physico-chemical interactions that
were formed between the water molecules and the frozen
materials, and the resulted solid material was grounded to
produce a NiGO-FD catalyst.
3
. Results and discussion
3.1. Characterization of M-rGO hybrid catalysts
The synthesis of M-rGO and on graphene oxide freeze-dried
MGO-FD) are illustrated in Fig. S1.† In short, a metal chlo-
2
.5. Catalytic activity with procedure and product analysis
(
Instead of conventional heating methods, we employed ride solution containing the prescribed amount of metal was
microwave-assisted reactions. The heat transferring process in added into the GO dispersion as both linkages among graphene
conventional heating is directly related to the diffusion of heat sheets and metal precursors. As depicted in Scheme 2, during
ꢀ
from the surface to bulk is comparatively slow and less effec- the synthesis of M-rGO aer reduction at 120 C with hydrazine
tive. In contrast, microwave heating conrms some advan- solution under mild reaction conditions, a gel-like structure
tages including short reaction time, high selectivity and high could be formed between metal and rGO as a result of con-
yields. The microwaves lead to the vibration and rotational necting rGO sheets with metal ions.
motion of polar molecules and can generate uniform heat in
On the other hand, another batch of GO was impregnated
a very short time. The microwave heating tools that we with nickel chloride solution as chosen metal and was subjected
employed were well equipped to control the W, Hz, time and to freeze-dried conditions. With the freeze-dried technique, the
temperature. When the power of the microwave was modu- 3D GO with uniform pore structure and larger surface area can
lated in the range of 200 to 800 W, the yield for glucose be obtained.
generation was 2.5 to 4.5 wt%. The high glucose yield observed
FTIR spectra of the parent GO, NiGO-FD and Ni-rGO are
at 800 W could be likely due to the sudden uneven energy shown in Fig. S2 and S3.† On NiGO-FD a strong adsorption
ꢁ1
distribution and high-temperature gradient known as “hot- band at around 3420 cm was detected that corresponds to
spots” or thermal runaway. Thus, low MW power was utilized a hydrogen-bonded O–H stretching vibration, with a very weak
in the succeeding experiments to minimize this occurrence. same signal found for the case of Ni-rGO parent material. The
ꢀ
ꢁ1
An optimized reaction temperature of 200 C was preferred band at 1048 cm corresponds to C–O stretching vibration
3
6
based on the previous report. The microwave-assisted cata- from C–O–C on the NiGO-FD and Ni-rGO respectively. Besides,
ꢁ
1
lytic conversion of cellulose, glucose and fructose was carried a band at 1635 cm of the C]C stretching vibration from the
out using microwave reaction vials. In a typical experimental benzene ring skeleton of graphene, totally disappears in the
run, 100 mg glucose, 4 mL solvent and 5 mg catalysts were case of Ni-rGO catalyst. As such, it is proved that GO was
charged into 10 mL reactor. Aer being sealed with a cap, the successfully modied by the use of the freeze-dried technique as
reactor containing the mixture was mounted in a microwave conrmed also by Raman and SEM. analysis. The Raman
reactor apparatus (Biotage Initiator+) and heated at a specied spectra for GO, NiGO-FD and Ni-rGO catalysts illustrated in
reaction time under magnetic stirring. Time zero of the reac- Fig. S4 and S5† also conrmed the ndings from FTIR spectra.
tion was dened as the time when the reactor reached its set It is known that two of the most intense D band and G band
point temperature. Reactions were performed in triplicate to Raman features can be ascribed to the characteristic peaks of
3
5
asses the reproducibility of results.
graphene materials. The D/G intensity ratio of the NiGO-FD is
The liquid samples were collected aer reaction and the 0.9, while on the Ni-rGO the corresponding ratio increases to
2
concentration of the product species were quantied using 1.43. The freeze-dried of GO facilitates the restoration of sp
Agilent Technologies HPLC with a ZORBAX Eclipse Plus C18 as carbon sites and the intensity of graphitic domains are smaller
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 11727–11736 | 11729

View Article Online
RSC Advances
Paper
in size that the ones in Ni-rGO sample. The increase in D/G on d show TEM images of Ni-rGO and NiGO-FD individually.
37
Ni-rGO is consistent with the results reported by Chen et al.
when studied GO reduction.
Fig. 1c indicates that the Ni-rGO is composed of folded single-
layer nanosheets, whereas Fig. 1d reveals the folding and
Fig. 1a and b shows the SEM images of the Ni-rGO and overlapping of the graphene sheets due to the freeze-drying
NiGO-FD samples respectively. While the as-synthesized Ni- process. In Fig. 1e and f, the selected EDX mapping from
rGO exhibits the presence of GO nanosheets stack tightly SEM conrmed the atomic percentage of Ni on the rGO surface
due to strong p–p interaction between the Ni metal nano- in a well distributed area.
particles and GO nanosheets. The spherical metal precursors
The variation of elemental composition and functional
grow on GO with an average diameter of 100–250 nm. Fig. 1a groups were analyzed via XPS. As shown in Fig. S6 and S7,† on
reveals the presence of larger (z250 nm) and smaller (z100 the C 1s spectra of the GO, NirGO and NiGO-FD seven
nm) particles by ꢃ17% and 36% respectively. Particles having different carbon entities: C]C, C–H, C–C, C–OH, C–O–C,
the intermediate sized were counted to be present by 47%. The C]O and O]C–O can be clearly identied. In the case of Ni-
NiGO-FD sample, conrmed the existence of a 3D structure rGO, the peaks of functional groups containing oxygen
revealing the presence of pores which are supposed to be completely disappears or decreases and the peak intensity for
created by the in situ self-assembly of graphene during the C]C dramatically increases, in agreement with Raman
3
8,39
freeze-dried process. As observed by other researchers,
it is results that shows an increase in D/G, demonstrating that the
assumed that the Ni metal nanoparticles could combine with GO reduction took place aer treatment with hydrazine. On
the graphene oxide sheets to allows the formation of a 3D GO the other hand, on GONi-FD, the presence of carboxylic acid
structure. These ndings were conrmed by the BET surface groups on GO is conrmed by the high intensity of C]O
area data where the total surface area of the NiGO-FD sample functional group in agreement with FTIR results shown in
is greatly larger than that of the Ni-rGO, as a result of the Fig. S2.†
formation of a 3D structure. On the other hand, Fig. 1c and
Then the presence and composition of the Ni on the GO
samples was determined by XRD and XPS analysis and
conrmed by TEM and TEM-EDX evaluation. It was found that
XPS spectrum of the Ni 2p for NirGo and NiGO-FD samples
(
8
Fig. S8†) exhibit the characteristic peaks at 871.7 eV and
54.2 eV respectively that are assigned to Ni 2p1/2 and Ni 2p3/2
These results clearly showed the presence of elemental
.
state nickel in the structure of the composite GO catalysts aer
reduction and freeze-dried respectively. XRD patterns of the
NirGO and NiGO-FD exhibited in Fig. S9 and S10† show a peak
ꢀ
of GO at reection (002) at 2q ¼ 25.4 , similar to previous
4
0,41
reports.
The reection GO peak in the case of NiGO-FD is
ꢀ
broader with the peak at 2q ¼ 23.3 which can suggest that
there is a disordered stacking among the graphene sheets due
to the formation of a 3D structure where the presence of Ni is
also conrmed. The presence of Ni species was also conrmed
by XRD analysis due to the formation of Ni-rGO as a conse-
quence of reduction of NiO nanoparticles in the GO. The peaks
ꢀ
ꢀ
at 44.49 (111) and 51.8 (200) 2q (JCPDS no. 01-071-3740)
ꢀ
along with the peak previously described at 25.4 for RGO
layers conrmed the formation of reduced Ni-rGO. The XRD
data is well supported by TEM and SEM observation (Fig. 1),
where the presence of Ni nanoparticles decorated rGO sheets
can be observed.
3.2. Catalytic properties of 2D M-GO materials
Metal supported on rGO and GO freeze-dried catalysts are
potential candidates for the conversion of glucose into 5-HMF
as they proved to be chemically stable and with a high func-
42
tionalized surface area. We can hypothesize that due to the
metal loaded and the formation of different oxygen functional
groups on the graphene oxide surface as shown by XPS and XRD
analysis, the hydrolysis of glucose to fructose and then the
dehydration to 5-HMF can be facilitated by the combination of
Lewis and Brønsted acid sites on the graphene oxide treated
Fig. 1 Surface morphology of Ni-rGO and NiGO-FD. SEM image of
fresh metal Ni-GO (a) and NiGO-FD (b). HRTEM image of Ni-rGO (c)
and NiGO-FD (d). EDX mapping of Ni-rGO (e) and NiGO-FD (f).
11730 | RSC Adv., 2020, 10, 11727–11736
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
Hence, the concentration of fructose observed on Ni-rGO was
lower when compared to Pt and Co on rGO respectively (Table
1). More so, with the presence of Co and Fe on rGO, 5-HMF
rehydrates to produce levulinic acid that may also react with
glucose or other reactive intermediates to form unwanted
humin compounds. The high hydrogenolytic activity of Pt on
rGO was conrmed by the presence of an increased yield of side
furanic compounds, where the lowest 5-HMF yield of all metal
rGO supported catalysts were found. Therefore, there is an
upward trend of both the yield of levulinic acid and glucose
conversion as shown in Table 1.
Regardless of the type of metal added to rGO, it is important
to note that fructose conversion to 5-HMF occurred relatively
faster in comparison to glucose isomerization. This indicates
that glucose proceeds via fructose to produce 5-HMF. The
mutual interaction of the pure metal and rGO provided
a moderate acid concentration suitable for the glucose-to-5-
HMF reaction. This cooperative interaction has been reported
in literature, wherein glucose is isomerized to fructose with
Fig. 2 Extent of Glucose conversion and yield of 5-HMF and other
43,44
products due to the catalytic action of M-rGO, blank, GO and rGO subsequent dehydration of fructose to 5-HMF.
Fig. 2 and
catalysts.
Table S1† also showed the individual metal rGO activity toward
glucose conversion to 5-HMF. This preliminary experimental
result encouraged us to investigate further how to optimize the
surface by reduction or by the freeze-dried technique. Initially, 5-HMF yield and glucose conversion.
metal-supported rGO reduced by hydrazine treatment were
Next, we examined the role of the reaction medium on the
studied for the conversion of glucose to 5-HMF using water as formation rate of 5-HMF. Ni-rGO catalyst was chosen as
the reaction medium. The microwave reaction was performed at showed the larger 5-HMF yield in water of all catalysts tested
ꢀ
2
00 C temperature during 30 minutes with a catalyst loading of (Fig. 2). Initially, single water-lean aqueous medium was
mg for 100 mg of glucose in 4 mL of water as the solvent. Four considered but the yield of 5-HMF found to be still relatively
5
hybrids were chosen. The performance of Pt-rGO, Ni-rGO, Co- low. Therefore, tetrahydrofuran (THF) was chosen as a co-
rGO and Fe-rGO were compared with a blank control, a GO solvent and added to water in different ratios. It is known
and rGO samples respectively. As can be seen in Fig. 2 under that THF lacks hydroxyl groups and has the potential to
4
5
identical reaction conditions the addition of a metal to the GO suppress side reactions. The addition of THF promoted a two
inuenced the catalytic activity, where all catalysts have solvent phases system; a reactive aqueous phase and an
increased glucose conversion and were more active toward extractive organic phase. This allowed in situ extraction of 5-
producing 5-HMF than GO and rGO alone except for Pt. Table 1 HMF from the aqueous phase into the organic phase, thereby
clearly shows that the yield increases in the following order: Ni- preventing undesired side reactions from 5-HMF decomposi-
rGO > Co-rGO > Fe-rGO > blank > Pt-rGO > rGO > GO.
tion. Moreover, the continuous extraction shis the reaction
Somehow surprisingly, the 5-HMF yield obtained during the equilibrium to produce more 5-HMF. Fig. 3 clearly demon-
blank experiment was higher than those obtained for GO, rGO strates the benecial effect of adding THF, which results in 5-
and Pt-rGO. This occurrence may be interpreted in relation to HMF yield increase up to about 70% at a near-complete
the metal and support properties of catalysts. rGO possesses glucose conversion. Not only does THF enable good parti-
a moderate concentration of acid sites that can dehydrate tioning of 5-HMF into the organic phase, but we also consider
fructose, an initial product of glucose isomerization, into 5- that the water reactive phase was modied by THF thereby
HMF.
Table 1 Extent of glucose conversion and yield of 5-HMF and other products due to the catalytic action of M-rGO, blank, GO and rGO catalysts
Glucose conversion
(%)
HMF yield
(%)
Fructose yield
(%)
FA yield
(%)
LA yield
(%)
Humin yield
(%)
Catalyst
Blank
GO
rGO
Pt-rGO
Ni-rGO
Co-rGO
45.1
31.5
35.2
58.6
61.4
69.3
12.5
4.3
7.4
9.8
28.1
17.8
10.3
9.6
9.1
4.8
7.3
5.6
0.7
1.2
0.6
7.3
3.1
2.3
0.1
0.8
0.2
3.5
1.5
1.4
21.5
15.6
17.9
33.2
21.4
42.2
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 11727–11736 | 11731

View Article Online
RSC Advances
Paper
Fig. 4 Extent of glucose conversion and yield of 5-HMF and other
products due to the catalytic action of blank, rGO Ni-rGO, GO-FD and
NiGO-FD catalysts.
Fig. 3 Extent of glucose conversion and yield of 5-HMF and other
products due to the variation in solvent composition in terms of water
to THF ration.
comparison GO-FD support have same catalytic behavior than
rGO support.
minimizing degradation of 5-HMF into humins. The outcome
of 5-HMF production using water–THF biphasic reaction
system is promising and provides a step further towards
achieving an efficient reaction medium for scaling-up.
As shown in Fig. 3, the 5-HMF yield increases from 22% to
As a result, the sharp catalytic conversion of glucose and 5-
HMF yield differences obtained when applying a freeze dried
technique can be explained based on the properties and char-
acterization of the NiGO-FD catalyst. As explained before, the
large surface area and porosity found on the freeze-dried GO
combined with the oxygen containing surface groups had an
impact on its catalytic activity. These groups linked to Ni metal
and Ni oxide might participate in the dehydration of fructose to
6
7% when the water/THF ratio rises from 1 : 1 to 1 : 4 implying
that the biphasic system worked successfully and prevented
the 5-HMF decomposition in the organic phase. On the other
hand, the ratio increases to 1 : 5, then 5-HMF yield and
glucose conversion started to decrease. It is important to note
that simultaneously levulinic acid and other compounds
increases (Table S2†). In addition, it was observed that the
water phase solution product colour changed from light yellow
to brown, an indication that also an amount of byproducts
principally humin compounds were formed. The decrease of 5-
HMF yield with increasing the THF concentration in two phase
solution may be due to a decreased transfer rate from the
reacting water phase to the extracting THF phase due to
formation of thicker layer partition, that favors also the
simultaneous formation of intermediate compounds and
other reactive intermediates via condensation, polymerization
5-HMF, while the large surface area of GO freeze-dried may
promote the conversion of glucose by increasing contact areas
42,43
between reactants.
As can be seen when comparing with Ni-
rGO catalyst, mainly the reaction selectivity and yield were
considerably lower than those under freeze dried treatment, as
the GO reduced catalyst contains a lesser amount of oxygen
groups on it surface as shown in Fig. 4 Furthermore this study,
there are also several reports in the literature of how the surface
of functionalized carbon materials play an important role in
44
catalytic reactions, such as alcohol dehydration, isomeriza-
45
tion. Then, due to its high 5-HMF catalytic yield (95%), NiGO-
FD catalyst was chosen for the optimization reaction studies.
For the study the biphasic water : THF system were used, but
the reaction time, reaction temperature, catalyst loading and
glucose concentration were changed in order to nd the
optimal reaction conditions for the conversion of glucose into 5-
HMF. Fig. S11 and Table S1† show that the zeta potential for
4
6
of glucose in the aqueous phase.
3.3. Inuence of 2D to 3D structural shis on catalytic
activities
Then, NiGO-FD catalyst synthesized by a freeze-dried technique NiGO-FD measured at pH 3, 5, 7 and 9 were found to be ꢁ21.75,
was tested for the glucose conversion to 5-HMF using the ꢁ22.79, ꢁ23.34 and ꢁ19.38 mV, respectively. The charge is
biphasic water : THF ¼ 1 : 4 and same reaction conditions as optimized at a pH of 7, whereas it tends to the zero point at both
per the case of Ni-rGO catalyst where a glucose conversion of the higher and lower pH value. No clear intersection at zero
1
00% and 5-HMF yield of 75% were achieved. As can be seen point is observed. The surface acidity of the materials-GO
from results in Fig. 4 NiGO-FD proved to be substantially more hybrid (20 wt%) in mmol unit (Table S2†) followed the trend
active than Ni-rGO, with practically 99% glucose conversion and as NiGO-FD (9.62) > Ni-rGO (2.68) > Co-rGO (2.47) > Fe-rGO (1.5)
5
-HMF yield close to 95%. However, it is interesting to note in > Pt-rGO (1.39) > rGO (1.28).
11732 | RSC Adv., 2020, 10, 11727–11736
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
aer 60 minutes could be due to the propensity of humin
formation via condensation, polymerization of glucose and
46
other reactive intermediates in the aqueous phase.
.3.2. Effect of reaction temperature on catalytic activities.
3
Temperature had a signicant impact on both the glucose
conversion and 5-HMF yield as well so high temperatures are
ꢀ
ꢀ
essential. Thus, temperature was varied from 160 C to 200 C. As
ꢀ
can be seen in Fig. 6 glucose conversion at 200 C reached 100%
with 95% 5-HMF yield aer 60 minutes reaction, whereas at 160 C
ꢀ
glucose conversion was only 30% with a mere 10% 5-HMF yield.
ꢀ
Then aer increasing temperature to 180 C there exists
a sharp increase in both glucose conversion and 5-HMF yield
ꢀ
that reach the maximum at 200 C. This is an indication that
longer time will be needed at a lower temperature, if would like
to reach similar glucose conversion at higher temperature.
47,48
However, as found previously
because 5-HMF is more reac-
tive at high temperature, a prolonged increase in the reaction
temperature beyond 200 C, can result in the decline of 5-HMF
Fig. 5 Reaction time dependent glucose conversion and yield of 5-
HMF and other products due to the catalytic action of NiGO-FD
catalysts.
ꢀ
yield. Based on these results, an optimized microwave condition
ꢀ
for this reaction was 60 minutes of reaction time at 200 C. At
these optimized conditions, we investigated also the effect of
initial glucose concentration.
3.3.1. Effect of reaction time on catalytic activities. As
3.3.3. Effect of initial glucose concentration on catalytic
shown in Fig. 5 5-HMF yield increases from 22% to 95% when
the reaction time increased from 10 to 60 minutes, respectively.
It was noted that, increasing reaction time, then 5-HMF yield
started to decrease, although the total conversion of glucose
continuously increases to reach 100%. It is important to note
that the amount of levulinic acid seems not to increase aer 60
minutes, which implies that the biphasic system worked
successfully and prevented the 5-HMF decomposition in the
organic phase. In addition, it was observed that as reaction time
increases the water phase solution product colour changed
from light to dark brown, an indication that an amount of
byproducts principally humin compounds were formed. As
such, a decrease of 5-HMF yield with increasing reaction time
activities. Fig. 7 represents the inuence of initial glucose
concentration on 5-HMF formation rate. Both glucose conver-
sion and 5-HMF yield decreased gradually with increased initial
glucose concentration.
The sharp loss of 5-HMF yield as a higher initial amount of
glucose in the feed may be due to the propensity self-
polymerization of 5-HMF or cross-polymerization between
glucose, 5-HMF and other intermediates that will occur easily at
a higher initial glucose concentrations, giving rise to the
formation of brown-black soluble polymers and insoluble
49,50
humin compounds in the aqueous phase.
Fig. 6 Reaction temperature dependent glucose conversion and yield Fig. 7 Initial concentration dependent glucose conversion and yield
of 5-HMF and other products due to the catalytic action of NiGO-FD of 5-HMF and other products due to the catalytic action of NiGO-FD
catalysts.
catalysts.
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 11727–11736 | 11733

View Article Online
RSC Advances
Paper
Fig. 9 Extent of glucose conversion and yield of 5-HMF and other
products with respect to cycle number during to the catalytic action of
NiGO-FD catalysts.
Fig. 8 Catalyst mass dependent glucose conversion and yield of 5-
HMF and other products due to the catalytic action of NiGO-FD
catalysts.
remaining cycles. We visually observed a change in colour of the
reaction solution from white to dark brown aer rst cycle. The
likely cause of this is the formation and then deposition of
humin compounds on the catalyst surface causing deactivation
of the active sites aer rst cycle that caused the drop of 5-HMF
yield as conrmed by. We hypothesized that the humins
deposited on NiGO-FD catalyst surface is most likely affecting
the dehydration step catalyzed by Brønsted acid from the
functionalized organic oxygen groups. 5-HMF is reported to
3
.3.4. Effect of catalyst loading on catalytic activities. The
effect of initial amount of catalyst loaded from 0 to 30 mg on the
conversion of glucose and 5-HMF yield is showed on Fig. 8. The
yield of 5-HMF in a control blank experiment was 40% at 70%
glucose conversion, but both increasing steadily as catalyst
loading increases, reaching a maximum of 100% glucose
conversion with 95% 5-HMF yield when 5 mg of catalyst was
used. However, increasing the catalyst loading up to 30 mg,
while the glucose conversion is maintained steady at 100%, the
52
have stronger adsorption affinity onto Brønsted acid site.
5
-HMF yield dropped sharply to 42%. Thus, large the amount of
Aerwards, adsorbed 5-HMF on the surface of the recycled
catalyst undergoes condensation reaction to humins, consistent
with the catalyst colour change. The strongly adsorbed 5-HMF
on the GO surface can also cause blockage of active sites as well
as block catalyst pores that can undergo further rehydration
reaction to organic acids. This is evidenced by the slightly
increased yields of fructose and levulinic acid as experimentally
detected. To test this hypothesis, we will be regenerating the
catalyst in future experiments by removing any trace of humin
compounds deposited on the catalyst surface before subse-
quent reaction runs, as catalyst deactivation by humin deposi-
tion is reversible by regeneration.
NiGO-FD catalyst added to the reaction system, consequently
a greater increase in Ni metal surface area and GO surface
oxygen functionalized area should speed up the glucose
conversion into 5-HMF increasing its yield. Unfortunately, it
was not the case here as the 5-HMF yield decreases at higher
catalyst loading due to side reactions (including humins)
formation in the reactive water phase. As reported in some
previous literature, THF, as the biphasic solvent for the reaction
system was preferred for its displayed optimized performance
51
to enhance the yield.
.3.5. Justication of catalyst recyclability. One of the most
3
important factors contributing to catalytic performance is the
catalyst stability and recyclability during the reaction. Recycla-
bility test with NiGO-FD catalyst was investigated by performing
the glucose dehydration reaction up to ve recycles. Aer
3.4. Comparison reported with catalytic systems
completion of each reaction cycle, the catalyst was recovered by The competitive reported values for the conversion of glucose
ltration, washed with acetone, dried and then used for the next into 5-HMF is presented in Table 2. A glucose conversion of

run and also submitted to characterization to assess if any 100% and 5-HMF yield of 75% that we achieved by using NiGO-
structural change occurred during the reaction. As a result, data FD is competitively higher than some reported values. The
from Fig. 9 conrmed that there was a slight decrease in the benchmark values for both the conversion of glucose and 5-
conversion of glucose aer all the cycle runs.
HMF yield as 100% were achieved by using P-VI-O/P-SO H-154
3
However, 5-HMF yield sharply declined over the rst cycle catalyst in THF/DMSO system. However present work repre-
runs from 95% to 55% and keeping constant that value for all sents the possibility of obtaining non-expensive metal nano-
particle scaffolding GO-based catalysts.
11734 | RSC Adv., 2020, 10, 11727–11736
This journal is © The Royal Society of Chemistry 2020

View Article Online
Paper
RSC Advances
Table 2 Reported values for glucose conversion and 5-HMF yields
Conversion
(%)
5-HMF yield
(%)
ꢀ
Catalyst
Solvent
T ( C)
T (h)
Ref.
SnPO
Sn-beta/HCl
EMIMBr
Water
Water/MlBK/CaCl
THF/water
TEAC
120
140
175
175
130
100
3
2
0.25
3
0.16
10
94.1
72.0
96.0
99.9
95.0
100
58.3
11.0
52.0
85.9
71.5
95.4
53
54
55
56
57
58
g-Al
2
O
3
2
TiO
2
–ZrO
2
/Amberlyst70
CrCl
3
$6H
2
O
P-VI-O/P-SO
3
H-154
THF/DMSO
2
S. G. Wettstein, D. M. Alonso, E. I. G u¨ rb u¨ z and J. A. Dumesic,
Curr. Opin. Chem. Eng., 2012, 1, 218–224.
4
. Conclusions
We demonstrated the conversion of glucose to 5-HMF in the
water-THF biphasic system catalyzed by Ni nanoparticles sup-
ported on GO NP synthesized by freeze-dried method. Modi-
cation of GO was achieved creating a 3D structure with large
surface area interconnected pores. Nickel metal was success-
fully incorporated into the GO surface as conrmed by XPS,
XRD and microscopy analysis. Furthermore, active catalytic
sites with Lewis and Brønsted type functionalities were created
on the GO surface as a result of the presence of oxygen func-
tional groups as conrmed by XPS that was highly effective for
the catalytic conversion of glucose to 5-HMF in comparison with
classical Ni-rGO structures. We also studied the effects of
temperature, time, amount of catalyst weight and glucose
substrate concentration. In general, all these parameters inu-
3 C. E. Wyman, B. E. Dale, R. T. Elander, M. Holtzapple,
M. R. Ladisch and Y. Lee, Bioresour. Technol., 2005, 96,
1959–1966.
4 R. Bodirlau, I. Spiridon and C. A. Teaca, BioResources, 2007,
2, 41–57.
5 T. Kondo, J. Polym. Sci., Part B: Polym. Phys., 1997, 35, 717–
723.
6 R. Rinaldi and F. Sch u¨ th, ChemSusChem, 2009, 2, 1096–1107.
7 M. Galbe and G. Zacchi, in Biofuels, Springer, 2007, pp. 41–
65.
8 J. C. Serrano-Ruiz, R. Luque and A. Sepulveda-Escribano,
Chem. Soc. Rev., 2011, 40, 5266–5281.
9 D. M. Alonso, J. Q. Bond and J. A. Dumesic, Green Chem.,
2010, 12, 1493–1513.
ence catalytic performance and yield of 5-HMF. Under opti- 10 T. Wang, M. W. Nolte and B. H. Shanks, Green Chem., 2014,
mized reaction conditions, remarkable high 5-HMF yield of ca.
16, 548–572.
5% and 100% glucose conversion was achieved in a biphasic 11 B. Saha and M. M. Abu-Omar, Green Chem., 2014, 16, 24–38.
water–THF system. Under comparable microwave reaction 12 A. Gandini, A. J. Silvestre, C. P. Neto, A. F. Sousa and
9
conditions, 5-HMF yield of 55% and 45% was attained at
glucose concentrations of 200 mg and 500 mg, respectively.
M. Gomes, J. Polym. Sci., Part A: Polym. Chem., 2009, 47,
295–298.
Lastly, catalyst stability under the given reaction conditions 13 B. Kamm, Angew. Chem., Int. Ed., 2007, 46, 5056–5058.
´ ´
´
´
afforded recycling for multiple reactions runs but at reduced 5- 14 A. Bereczky, K. Lukacs, M. Farkas and S. Dobe, Nat. Resour.,
HMF yield due to the formation of a high concentration of
humin products.
2014, 2014(5), 177–191.
15 R. Daniel, G. Tian, H. Xu, M. L. Wyszynski, X. Wu and
Z. Huang, Fuel, 2011, 90, 449–458.
1
1
6 J. O. Metzger, Angew. Chem., Int. Ed., 2006, 45, 696–698.
7 Y. Rom ´a n-Leshkov, J. N. Chheda and J. A. Dumesic, Science,
Conflicts of interest
2
006, 312, 1933–1937.
There are no conicts to declare.
1
1
8 M. Mascal and E. B. Nikitin, Green Chem., 2010, 12, 370–373.
9 J. N. Chheda, Y. Rom ´a n-Leshkov and J. A. Dumesic, Green
Chem., 2007, 9, 342–350.
Acknowledgements
2
2
2
2
2
0 M. Mascal and E. B. Nikitin, ChemSusChem, 2009, 2, 859–
This work was supported by KAKENHI Grant-in-Aid for Scien-
tic Research (A) JP17H01200. The authors acknowledge the
support of the Kumamoto University and Queensland Univer-
sity of Technology (QUT) for the preparation of this work. J. N.
Beltramini also acknowledge the nancial support of JSSP
Fellowship (No. L18519).
886.
1 G. Yang, C. Wang, G. Lyu, L. A. Lucia and J. Chen,
BioResources, 2015, 10, 5863–5875.
2 L. Hu, Z. Wu, J. Xu, Y. Sun, L. Lin and S. Liu, Chem. Eng. J.,
2014, 244, 137–144.
3 L. Atanda, S. Mukundan, A. Shrotri and Q. Ma andJ
Beltramini, ChemCatChem, 2015, 7, 781–790.
4 L. Atanda, A. Shrtri, S. Mukundan, Q. Ma, M. Kunarova and
J. Beltramini, ChemSusChem, 2015, 8, 2907–2916.
Notes and references
1
D. M. Alonso, S. G. Wettstein and J. A. Dumesic, Green Chem.,
013, 15, 584–595.
2
This journal is © The Royal Society of Chemistry 2020
RSC Adv., 2020, 10, 11727–11736 | 11735

View Article Online
RSC Advances
Paper
2
5 G. R. Jenness and J. R. Schmidt, ACS Catal., 2013, 12, 2881– 41 B. Li, H. Cao, C. Y. Park, J. G. Choi and W. C. Oh, J. Korean
890. Ceram. Soc., 2011, 48, 147–151.
6 M. R. Karim, H. Shinoda, M. Nakai, K. Hatakeyama, 42 H. Ham, T. V. Khai, N. H. Park, D. S. So, J. Lee, H. G. Na,
2
2
H. Kamihata, T. Matsui, T. Taniguchi, M. Koinuma,
K. Kuroiwa, M. Kurmoo, Y. Matsumoto and S. Hayami,
Adv. Funct. Mater., 2013, 23, 323–332.
Y. J. Kwon, H. Y. Cho and H. W. Kim, Nanotechnology,
2014, 25, 235601–235608.
43 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–
554.
2
7 H. Ohmagari, M. R. Karim, Y. Shudo, S. Ida, R. Ohtani and
S. Hayami, MRS Adv., 2018, 3, 2847–2854.
8 P. Johari and V. B. Shenoy, ACS Nano, 2011, 5, 7640.
44 J. N. Chheda, Y. Roman-Leshkov and J. A. Dumesic, Green
Chem., 2007, 9, 342–350.
2
2
9 A. L. Han, S. Jin, H. L. Cen, H. X. Ji, Z. J. Sun and P. W. Du, 45 L. Atanda, A. Shrotri, S. Mukundan, Q. Ma, M. Konarova and
Mater. Chem., 2015, 3, 1941–1946. J. Beltramini, ChemSusChem, 2015, 8, 2907–2916.
0 D. R. Dreyer, H. P. Jia and C. W. Bielawski, Angew. Chem., Int. 46 Y. Li and W. Shen, Chem. Soc. Rev., 2011, 10, 1543–1574.
3
Ed., 2010, 49, 6813–6820.
47 J. Bedia, J. M. Rosas, J. Martinez, J. Rodriguez-Mirasol and
31 Y. Murashima, M. R. Karim, R. Furue, T. Matsui,
T. Cordero, Carbon, 2009, 47, 286–294.
H. Takehira, K. Wakata, K. Toda, R. Ohtani, M. Nakamura 48 J. Han, J. Duan, P. Chen, H. Lou, X. Zheng and H. Hong,
and S. Hayami, Inorg. Chem. Front., 2016, 3, 842–848. Green Chem., 2011, 13, 2561–2568.
2 H. M. A. Hassan, V. Abdelsayed, A. E. R. S. Khder, 49 J. Gong, M. J. Katz and F. M. Kerton, RSC Adv., 2018, 8,
3
K. M. AbouZeid, J. Terner, M. S. El-Shall, S. I. Al-Reayes
and A. A. El-Azhari, J. Mater. Chem., 2009, 19, 3832–3837.
3 A. R. Siamaki, A. E. R. S. Khder, V. Abdelsayed, M. S. El-Shall
and B. F. Gupton, J. Catal., 2011, 279, 1–11.
4 J. Wang, X. B. Zhang, Z. L. Wang, L. M. Wang and Y. Zhang,
Energy Environ. Sci., 2012, 5, 6885–6888.
31618–31627.
50 X. Qi, M. Watanabe, T. M. Aida and J. R. L. Smith, Green
Chem., 2009, 11, 1327–1331.
51 L. Atanda, A. Silahua, S. Mukundan, A. Shrotri, G. T. Torres
and J. Beltramini, RSC Adv., 2015, 98, 80346–80352.
52 Y. Qu, Q. Wei, H. Li, P. Oleskowicz-Popiel and C. Huang,
Bioresour. Technol., 2014, 162, 358–364.
3
3
3
5 M. R. Karim, K. Hatakeyama, T. Matsui, H. Takehira,
T. Taniguchi, M. Koinuma, Y. Matsumoto, T. Akutagawa, 53 Q. Hou, M. Zhen, L. Liu, Y. Chen, F. Huang, S. Zhang, W. Li
T. Nakamura, S. Noro, T. Yamada, H. Kitagawa and
S. Hayami, J. Am. Chem. Soc., 2013, 135, 8097–8100.
6 E. G. Mission, A. T. Quitain, M. Sasaki and T. Kida, Green
Chem., 2017, 16, 19.
7 L. Chen and X. Zhang-Run, RSC Adv., 2016, 6, 56278–56286.
8 K. Bhowmik, A. Mukherjee, M. Kr Mishra and G. De,
Langmuir, 2014, 30, 3209–3216.
9 Y. Choi, H. S. Bae, E. Seo, S. Jang, K. H. Park and B. S. Kim, J.
Mater. Chem., 2011, 21, 15431–15436.
0 J. H. Li, Y. Li, H. Meng, S. Y. Xie, B. W. Zhang, L. F. Li,
and M. Ju, Appl. Catal., B, 2018, 224, 183–193.
54 E. Nikolla, Y. Rom ´a n-Leshkov, M. Moliner and M. E. Davis,
ACS Catal., 2011, 1, 408–410.
55 C. Garc ´ı a-Sancho, I. F u´ nez-N u´ n˜ ez, R. Moreno-Tost,
J. Santamar ´ı a-Gonz ´a lez, E. P ´e rez-Inestrosa, J. L. G. Fierro
and P. Maireles-Torres, Appl. Catal., B, 2017, 206, 617–625.
56 L. Atanda, A. Silahua, S. Mukundan, A. Shrotri, G. T. Torres
and J. Beltramini, RSC Adv., 2015, 98, 80346–80352.
57 L. Hu, Y. Sun and L. Lin, Ind. Eng. Chem. Res., 2012, 51, 1099–
1104.
3
3
3
3
4
H. J. Ma, J. Y. Zhang and M. Yu, J. Mater. Chem. A, 2014, 2, 58 L. Wang, H. Wang, F. Liu, A. Zheng, Q. Sun, J. Lewis, L. Zhu,
2934–2940.
X. Meng and F. S. Xiao, ChemSusChem, 2014, 7, 402–406.
11736 | RSC Adv., 2020, 10, 11727–11736
This journal is © The Royal Society of Chemistry 2020