Exploration on Hydroxymethylation of Furfural to 5-Hydroxymethylfurfural Over Mesoporous Niobium Phosphate Catalyst

Article abstract of DOI:10.1007/s10562-020-03261-5

Abstract: 5-Hydroxymethylfurfural (HMF) is mainly produced by dehydration of hexasaccharides which were limited by the source availability and low yield of the process. Accounting to this, the current study was thus designed to first identify the mesoporous niobium phosphate was a tailored catalyst and further to develop a suitable reaction system for hydroxymethylation of furfural to HMF. The mesoporous niobium phosphate was prepared by a two-step sol–gel hydrothermal method used to catalyze this hydroxymethylation reaction either in a pure water solvent or in an aqueous organic solvent system. Detailed investigation on reaction activity and HMF selectivity was carried out by varying reaction conditions. A highest HMF yield of 51.1% was obtained with a 79.2% furfural conversion under the aqueous dimethyl sulfoxide (DMSO) at 90?°C for 20?h over the current catalyst. Moreover, the mesoporous niobium phosphate could be recovered and reused at least for 6 runs without obvious loss in catalytic activity. Grpahic Abstract: Mesoporous niobium phosphate was synthesized by a two-step sol–gel hydrothermal method and firstly investigated for the catalytic hydroxymethylation of furfural to HMF. The results suggest that mesoporous niobium phosphate shows a superior activity and selectivity. The highest yield of 51.1% of HMF was obtained with a 79.2% furfural conversion. The excellent catalytic performance could be attributed to attributed to the strong acid sites and larger porous structure in the catalyst. [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-020-03261-5

Catalysis Letters
https://doi.org/10.1007/s10562-020-03261-5
Exploration on Hydroxymethylation of Furfural
to 5‑Hydroxymethylfurfural Over Mesoporous Niobium Phosphate
Catalyst
1
1
2
1
Huifa Zhang · Tong Geng · Huixian Zhang · Fahai Cao
Received: 7 January 2020 / Accepted: 14 May 2020
©
Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract
-Hydroxymethylfurfural (HMF) is mainly produced by dehydration of hexasaccharides which were limited by the source
5
availability and low yield of the process. Accounting to this, the current study was thus designed to ꢀrst identify the
mesoporous niobium phosphate was a tailored catalyst and further to develop a suitable reaction system for hydroxymeth-
ylation of furfural to HMF. The mesoporous niobium phosphate was prepared by a two-step sol–gel hydrothermal method
used to catalyze this hydroxymethylation reaction either in a pure water solvent or in an aqueous organic solvent system.
Detailed investigation on reaction activity and HMF selectivity was carried out by varying reaction conditions. A highest
HMF yield of 51.1% was obtained with a 79.2% furfural conversion under the aqueous dimethyl sulfoxide (DMSO) at 90 °C
for 20 h over the current catalyst. Moreover, the mesoporous niobium phosphate could be recovered and reused at least for
6
runs without obvious loss in catalytic activity.
Grpahic Abstract
Mesoporous niobium phosphate was synthesized by a two-step sol–gel hydrothermal method and ꢀrstly investigated for the
catalytic hydroxymethylation of furfural to HMF. The results suggest that mesoporous niobium phosphate shows a superior
activity and selectivity. The highest yield of 51.1% of HMF was obtained with a 79.2% furfural conversion. The excellent
catalytic performance could be attributed to attributed to the strong acid sites and larger porous structure in the catalyst.
Bronsted acid sites
OH OH OH
Nb
O O
P
P
O
O
O
O
Nb
P
P
O
OH
O
O
O
O
O
Lewis acid sites
O
+
H
H
Aqueous or aqueous organic solvent
HMF
Furfural
Formaldehyde
1
Introduction
Electronic supplementary material The online version of this
article (https://doi.org/10.1007/s10562-020-03261-5) contains
supplementary material, which is available to authorized users.
The molecular structure of HMF has a furan ring with a
hydroxyl group and an aldehyde-function which makes
HMF one of the most important intermediate products and
also being identiꢀed as one of the top 10 value-added bio-
based chemicals by the U.S. Department of Energy [1].
It is a versatile intermediate and can be transformed into
various major high value chemicals, such as furan-2,5-di-
methylcarboxylate (FDMC) [2], 2,5-diformyl furan (DFF)
*
Fahai Cao
fhcao@ecust.edu.cn
1
Engineering Research Centre of Large Scale Reactor
Engineering and Technology of Ministry of Education,
East China University of Science and Technology,
Shanghai 200237, China
2
SINOPEC North China E&P Company, Zhengzhou 450006,
China
[3], 5-hydroxymethyl-2-furan carboxylic acid (HMFCA) [4],
Vol.:(0
1
123456789)
3

H. Zhang et al.
2
,5-furandiformic acid (FDCA) [5, 6], 2,5-dimethylfuran
This work was thus designed and carried out to investi-
gate the feasibility of using niobium phosphate as a solid
acid catalyst for hydroxymethylation of furfural to HMF.
The choice of niobium phosphate to deepen the investigation
lies in our understanding on its tunable and water-tolerable
Brönsted and Lewis acid sites. A two-step sol–gel hydro-
thermal method was then applied to prepare niobium phos-
phate in a mesoporous form. The catalyzed reaction was
performed either in a pure water solvent or in an aqueous
organic solvent system. Reaction activity and HMF selec-
tivity was studied and compared extensively. It is believed
that this work would not only provide a promising catalyst
of high eꢂciency to convert furfural to HMF, but also with
a readily operation-route for industrial production.
(
DMF) [7] and 5-hydroxymethylfurfurylamine (HMFA) [8],
which can further be used as starting materials for various
products and are of great impacts on the green economy.
HMF is mainly synthesized by dehydration of C6 carbo-
hydrates, such as fructose, glucose and sucrose. Its current
production was limited by the source availability and low
yield of the process. The biomass derived C6 carbohydrates
are now considered as the only sustainable chemical car-
bon source for HMF [9, 10], which arouses the concern of
squeezing the productions of food crops [11], even though
the sustainability of biomass was well acknowledged to alle-
viate the energy crisis. The low yield was due to excessive
HMF degradation and cross-polymerization in the produc-
tion process. Direct conversion of lignocellulose to HMF in
a large scale for commercial use, whereas, was not success-
ful till now [12].
2 Materials and Methods
On the other hand, industrial production of furfural is
now very mature to make it widely available [13]. It would
thereby be extremely ingenious and attractive if a feasible
route from furfural to HMF could be explored. Early eꢁorts
on the synthesis of HMF from furfural could be traced back
to 1999 by Lecomte and later by Nishimura et al. [14–16].
The reaction was taken place in a formaldehyde solution
with the aid of acids as a catalyst. By using microporous
solid acids as a catalyst in a bath reactor, Jerome Lecomte
achieved this with a selectivity of nearly 30%. Nishimura
et al. further developed it using homogeneous and solid
acids including inorganic acid, zeolite, cation resin and
metal oxide. In their study on homogeneous acids based
catalysts, a positive correlation between furfural conversion
and acid strength was established, but with a negative corre-
lation for HMF selectivity. Furthermore, the solid acid based
catalysts would yield a superior selectivity. Of note, when
Amberlyst-15 was used as the catalyst, a 43.1% yield and
2.1 Materials
Furfural (99%) and HMF (98%) were of analytical grade
and purchased from Aladdin Chemicals Co., Ltd. (Shang-
hai, China). Cetyltrimethyl ammonium bromide (CTAB)
and niobium oxide were purchased from TCI (Shanghai)
Chemical Industry Co., Ltd. Formalin was purchased from
Tianjin Guangfu Chemical Industry (Tianjin, China). H-Beta
(100) was provided by Novel Environmental Protection New
Materials (Shanghai, China) Co., Ltd. H-ZSM-5 (50) was
purchased from Nankai University Catalyst Co., Ltd (Tian-
jin, China). All other chemicals and solvents were purchased
from Shanghai Titan Scientiꢀc Co., Ltd. (Shanghai, China)
and they were analytically pure which were used without
further puriꢀcation. Furfural was distilled before used.
2.2 Methods Section
5
7.5% selectivity for HMF production were achieved aston-
ishingly. Despite the fact that Amberlyst-15 is an excellent
solid acid catalyst, it shows poor solvent tolerant because
of the swelling eꢁect [15], which may result in cracking of
resin beads. In addition, the acid concentration of Amber-
lyst-15 was shown to be unstable [17]. What’s more, resin
is a more expensive solid acid catalyst because of its more
complex fabricating process than other solid acid catalysts
2.2.1 Preparation of Mesoporous Niobium Phosphate
Catalyst
Mesoporous niobium phosphate catalyst was prepared via a
sol–gel hydrothermal method in two steps with minor modi-
ꢀcation on previously reported [18]. First, niobium tartrate
was prepared. Niobium oxide of 5.136 g was dissolved in
60 ml hydroꢃuoric acid solution in a stainless steel reac-
tor containing a Teꢃon liner at 100 °C for 6 h. Ammonium
oxalate aqueous solution was prepared by dissolving 7.316 g
ammonium oxalate in 40 ml deionized water in an oven at
80 °C for 20 min. The two solutions were added into a poly-
propylene plastic bottle and were stirred vigorously. The
resulting solution was titrated with ammonia until the solu-
tion became cloudy, indicating the completion of precipita-
tion. It was then ꢀltered and rinsed with deionized water.
Aqueous solution of tartaric acid which was prepared by
[
16]. Further exploration was carried out on beta zeolites
with diꢁerent ratios of Si–Al to catalyze the reaction. It was
interesting to identify that the ratio of Si–Al has a strong
eꢁect on the target reaction with the highest yield of 31.6%
achieved over a beta zeolite of a 52 (Si–Al) ratio (SiO /
2
Al O =104, selectivity=46.5%, reaction duration=12 h)
2
3
[
16]. Above all, a solvent tolerant and stable catalyst is the
key for eꢂcient hydroxymethylation of furfural to HMF and
remains to be surmounted.
1
3

Exploration on Hydroxymethylation of Furfural to 5-Hydroxymethylfurfural Over Mesoporous…
dissolving 9 g tartaric acid in 80 ml deionized water was
added to the ꢀltered solid with stirring vigorously at 80 °C
for 15 min. Niobium tartrate solution was obtained after the
addition of 5 g (30%) H O aqueous solution to the above
size distributions of H-beta and H-ZSM-5 were measured
by the DFT method. Temperature-programmed desorption
of ammonia (NH -TPD) of the samples were carried out on
3
Micromeritics ChemiSorb 2920 with a thermal conductiv-
ity detector (TCD). First, the samples (0.12 g) were pre-
treated at 300 °C under He atmosphere at the ꢃow rate of
2
2
solution with stirring for 6–8 h at 80 °C. In the second step,
2
0 ml diammonium hydrogen phosphate aqueous solution
−
1
was prepared by dissolving 1.32 g diammonium hydrogen
phosphate ((NH ) HPO ) in 20 ml of deionized water and
50 ml·min for 2 h. After the temperature was decreased to
50 °C under He atmosphere, NH adsorption was carried out
4
2
4
3
adjusted with phosphoric acid to pH 2. The 20 ml niobium
tartrate aqueous solution that was adjusted with phosphoric
acid to pH 2 and the 20 ml diammonium hydrogen phosphate
aqueous solution were mixed and stirred, obtaining a mixed
solution A. CTAB aqueous solution which was prepared by
dissolving 1.0 g of CTAB in 13 ml deionized water was
added into the solution A dropwise, and stirred vigorously
at 35 °C for 1 h. The obtained suspension was aged at 160 °C
for 24 h in a 150 ml stainless steel reactor with a Teꢃon liner.
After being cooled down with iced water, the precipitate
was ꢀltered and washed with deionized water. The resulting
solids were dried at 100 °C in a vacuum oven for 4 h. Finally,
mesoporous niobium phosphate was obtained by calcination
of the resulting solids in air atmosphere at 550 °C for 3 h
with a ramp rate of 2 °C·min− and denoted as NbPO.
at 50 °C for 60 min and NH desorption was conducted from
3
−1
50 to 700 °C under He atmosphere, at a rate of 10 °C·min .
The desorption gas was detected with a TCD detector. As
a comparison, the amounts and strengths of acidity of the
H-beta (100) zeolite and H-ZSM-5 (50) were also tested
by NH -TPD method. Pyridine was used as one probe mol-
3
ecule to investigate the nature of acid sites of the catalyst.
Pyridine infrared spectroscopy (Py-IR) was recorded on a
Nicolet NEXUS 670 FT-IR spectrometer between 1400 and
−
1
1650 cm at diꢁerent desorption temperatures. Self-sup-
porting disks (6.5 mm diameter, 10 mg) of the samples were
prepared applying 6-ton pressure. Before adsorbing pyridine,
the samples were previously dehydrated by heating at 400 °C
for 3 h under vacuum. After cooling, pyridine vapor was
admitted until equilibrium was reached. Subsequent pyri-
dine desorption was carried out at diꢁerent temperature for
20 min and the spectra were obtained in situ.
1
2.2.2 Catalysts Characterization
Scanning electron microscope (SEM) was used to character-
ized the surface morphology of the prepared samples on a
Nova NanoSEM 450 ꢀeld emission scanning electron micro-
scope. Transmission electron microscopy (TEM) was used
to investigate the microstructure of the prepared samples on
a FEI Tecnai F20 s-TWIN instrument and the acceleration
voltage was 200 kV. Inductively coupled plasma mass spec-
trometry (ICP-MS) measurements were performed using an
Agilent ICP-MS 4500 instrument to measure the actual Nb,
P and O wt% present. The samples were dissolved in nitro-
hydrochloric acid before measurement. Powder X-ray dif-
fraction (XRD) patterns of the prepared were recorded on a
Bruker D8 FOCUS diꢁractometer with Cu Kα (λ=1.541 Å)
operating at room temperature. The diꢁraction patterns were
collected in the 2 θ range of 10°–80° and 0.7°–10° with
2.2.3 Catalytic Reaction
In a typical reaction (Scheme 1), 2 mmol furfural, 10 ml
formalin, 0.4 g mesoporous NbPO catalyst, and 10 ml
organic solvent were successively added into a 50 ml Tef-
lon-lined stainless steel autoclave. After being purged with
N for 10 min, the reactor was heated in an oil bath with
2
a magnetic stirrer. Once the system reached the desired
temperature (e.g. 90 °C), the reaction was initiated by vig-
orous stirring with a magnetic stirrer at a stirring speed of
500 rpm. After the reaction, the reactor bath was cooled
to room temperature with iced water. The products were
collected and separated by centrifugation. The obtained
liquid was analyzed by High Performance Liquid Chro-
matography (HPLC) equipped with an XDB-C18 reversed
phase column and a UV detector.
−
1
a scanning rate of 0.2°·s . Samples surface area analyses
and the pore size distributions were measured by nitrogen
adsorption at−196 °C using an ASAP 2000 analyses. The
samples were degassed at 150 °C for 10 h prior to adsorp-
tion measurements. Previous work found that H-beta zeolite
and H-ZSM-5 have certain catalytic activity for this reaction
O
OH
O
O
O
O
+
[
15]. As a comparison, the surface area and the pore size
H
H
distributions of the H-beta (100) zeolite (SiO /Al O =100)
2
2
3
and H-ZSM-5(50) (SiO /Al O = 50) were also measured
HMF
2
2
3
Furfural
Formaldehyde
with the same method. Surface areas were determined by
BET method. Pore size distribution of NbPO was calcu-
lated by the Barrett–Joyner–Halenda (BJH) method. Pore
Scheme 1 Hydroxymethylation of furfural with formaldehyde to
HMF
1
3

H. Zhang et al.
3
Results and Discussion
3.1 Characterization of Catalysts
As shown in Fig. 1, the small-angle X-ray diꢁraction (XRD)
pattern of the prepared sample displayed a broad peak at
around 2.3 degrees (at 2 θ), indicating that it has a disor-
dered, wormhole-like structure. There were two broad peaks
in the wide-angle X-ray diꢁraction (XRD) curve of the pre-
pared sample at around 25 degrees and 53 degrees (at 2 θ),
respectively, which means the structure of the catalyst is
amorphous with low crystallinity. In addition, the XRD char-
acteristic peaks of the synthesized materials are consistent
with those of niobium phosphate materials in the literature,
indicating that niobium phosphate catalyst has been success-
fully synthesized [18].
Fig. 2 SEM (left) and TEM (right) images of the prepared niobium
phosphate material
broad peaks can be observed in all the tested samples at the
temperature region of 150 °C and 400 °C. The desorption
peak below 200 °C corresponds to the weak acid sites. The
peak at the temperature region of 200–400 °C is referred to
medium acid sites while the peak at the temperature region
of 400–700 °C is attributed to strong acid sites in the tested
catalysts, which are consistent to the literature [19]. It is
clear from Fig. 3 that the medium strength acid sites are
much higher for NbPO than the compared catalysts. Larger
amount of strong acid site ratio is observed on the H-Beta
The surface morphology and microstructure of the sample
were examined by SEM and TEM. The representative SEM
image (Fig. 2, left) exhibits a dominant ꢃuꢁy and porous
like marshmallow morphology for the prepared NbPO. In
order to further understand its microstructure, TEM was per-
formed. As shown in Fig. 2 (right), a random distribution of
the pores of the material can be observed, which means the
prepared NbPO has an amorphous mesostructure.
(
100) samples.
In order to quantify and investigate the nature of acid sites
Inductively coupled plasma mass spectrometry (ICP-MS)
was used to measure the actual elementary composition of
the prepared NbPO. According to the ICP results, the actual
Nb, P and O wt% of the catalyst after calcination is 9.4%,
of the NbPO, Py-IR spectra were recorded. Figure 4 shows
the IR spectra of adsorbed pyridine at diꢁerent temperature
on the NbPO, H-beta (100) zeolite and H-ZSM-5 (50). All
−
1
the samples exhibit three reꢃection bonds at 1540 cm ,
9
.4%, and 81.2%, respectively.
−
1
−1
−1
1
490 cm and 1450 cm . The peaks at 1540 cm have
In order to study the amount and strength of acidity of the
been generally ascribed to Brønsted acidic sites (B). The
prepared NbPO, NH -TPD was performed. The NH -TPD
3
3
−
1
peaks at 1490 cm could be attributed to Brønsted and
Lewis acidic sites (B + L) at the same time. The peaks at
proꢀles of the tested samples are shown in Fig. 3. Two
Fig. 1 Small-angle (before calcination) and wide-angle (inset) XRD
Fig. 3 NH -TPD proꢀles of a H-ZSM-5(50), b NbPO, and c H-Beta
3
patterns of the prepared niobium phosphate material
(100)
1
3

Exploration on Hydroxymethylation of Furfural to 5-Hydroxymethylfurfural Over Mesoporous…
Fig. 4 Py-IR spectra of a NbPO, b H-beta (100) zeolite and c H-ZSM-5 (50)
Table 1 Summarized acid
distribution on the catalysts
used
Acid sites (mmol Py/gcat)a
Catalyst
Weak
B
Medium
B
Strong
B
Total
B
Total
B+L
L
L
L
L
NbPO
H-Beta (100)
H-ZSM-5 (50)
0.16
0.08
0.19
0.03
0.13
0.03
0.19
0.02
0.28
0.08
0.03
0.00
0.03
0.11
0.08
0.03
0.14
0.00
0.38
0.21
0.55
0.13
0.30
0.03
0.52
0.51
0.58
aEstimated by Py-IR spectra of pyridine adsorption analysis
450 cm− could be ascribed to the Lewis acidic sites (L).
1
1
The acid sites of pyridine desorbed below 200 °C are con-
sidered to be weak acid sites; the sites pyridine desorbed at
2
00–300 °C is ascribed to be the medium strong acid sites,
and the remain adsorbing pyridine sites after evacuation at
4
00 °C are deꢀned as the strong sites. The amounts of acid
sites obtained by the integral peak areas according to the
literature [20] are depicted in Table 1. As can be seen, the
−
1
medium acidic sites of the NbPO are more (270 μmol g )
−1
than weak acidic sites (190 μmol g ) and strong acidic sites
−
1
(
60 μmol g ), that is coincident to the NH -TPD results. It
3
is interesting to note that the prepared NbPO and the H-beta
(
100) zeolite almost have a comparable acid content which
is coincident to the NH -TPD results.
3
Fig. 5 N adsorption–desorption isotherms of catalysts used
The nitrogen adsorption–desorption isotherms were
used for surface area and porosity measurements of the
catalysts used. As shown in Fig. 5, the H-beta (100) and
H-ZSM-5 (50) are typical type I isotherms with a saturated
adsorption platform on the adsorption isotherm, which
show that the pore diameters are evenly distributed. The
nitrogen adsorption–desorption isotherm of the prepared
NbPO is a type IV isotherm with a hysteresis loop at the
middle and high pressure region, generally ascribed to the
mesoporous structure according to IUPAC classiꢀcation.
The results show that NbPO is a kind of porous mate-
rial with a mesoporous and complex pore structure. The
detailed information of pore size distributions and speciꢀc
surface areas are summarized in Table 2. It can be seen
2
that the H-beta (100) and H-ZSM-5 (50) samples show a
relatively high intensive pore distribution in 0.62 nm and
0.55 nm respectively. The pores of the prepared NbPO are
in the main range of 3.8 nm and the surface area is 374.2
2
−1
m g , which is smaller than the surface area of H-beta
2
−1
(100) sample (641.2 m g ) and H-ZSM-5 (50) sample
2
−1
(394.3 m g ).
As a result, the above XRD, TEM, SEM and BET
results strongly demonstrate the successful synthesis of
mesoporous structured niobium phosphate.
1
3

H. Zhang et al.
Table 2 Summary of
physicochemical properties of
catalysts used
2
Entry
Catalyst
SBET (m /g)
Pore diameter
Pore
volume
Micropore diam- Mesopore diam-
3
(
cm /g)
eter (nm)
eter (nm)
1
2
3
NbPO
H-Beta (100)
H-ZSM-5 (50)
374.2
641.2
394.3
3.8
–
–
0.28
0.57
0.36
0.62
0.55
3
.2 Hydroxymethylation of Furfural
3.2.2 Catalytic Hydroxymethylation of Furfural
to 5‑Hydroxymethylfurfural
to 5‑Hydroxymethylfurfural Over Mesoporous
Niobium Phosphate Catalyst in Various Reactions
Medium
3
.2.1 Catalytic Hydroxymethylation of Furfural
to 5‑Hydroxymethylfurfural Over Various Catalysts
in Aqueous Solution
Previously, numerous studies have been done on the produc-
tion of HMF over various heterogeneous catalysts by dehy-
dration of hexose or lignocellulose. However, because of the
excessive HMF degradation and cross-polymerization, there
is usually a low yield of HMF in a pure aqueous solution. In
order to improve the yield of HMF, organic solvents includ-
In order to test the activity of the prepared mesoporous
NbPO, hydroxymethylation of furfural was ꢀrst conducted
in an aqueous solution. For comparison, parallel reactions
using H-Beta (100) and H-ZSM-5 (50) with same amount
of acid sites were also carried out and the results are sum-
marized in Table 3. It is found that the mesoporous NbPO
shows a superior activity and selectivity, with the conver-
sion of furfural and the selectivity of HMF being 63.8%
and 57.5% respectively (Table 3, Entry 1). The conver-
sion of furfural over the two comparison catalysts (H-Beta
ing acetonitrile (CH CN), dimethyl sulfoxide (DMSO),
3
dimethyl formamide (DMF), and N-Methyl pyrrolidone
(NMP) have been used to avoid HMF degradation during the
reaction. Therefore, the hydroxymethylation of furfural to
HMF was ꢀrstly carried out in four diꢁerent solvent systems
in the presence of the mesoporous NbPO catalyst in order to
ꢀnd out the best reaction system. Eꢁects of the solvents on
the hydroxymethylation of furfural to HMF are summarized
in Table 4. Generally, the addition of organic solvents shows
a positive eꢁect on the product yield. This may be due to
the high stability of both furfural and HMF in these solvent
systems. Among the solvents used, DMF and DMSO oꢁered
a desirable furfural conversion as well as HMF selectivity.
The highest yield of HMF (45.5%) was achieved in the reac-
tion by using aqueous DMSO as the solvent system, in which
the conversion of furfural and the selectivity of HMF was
62.7% and 72.6% respectively. This may be attributed to
the fact that the presence of DMSO or DMF could protect
HMF from further degradation and cross-polymerization by
forming a shielding ꢀlm around HMF. However, the yield of
HMF is only 36.9% under the same condition with aqueous
acetonitrile as the solvent, which is slightly higher than the
(
100) and H-ZSM-5 (50)) is respectively 58.5%, 37.3%,
and the selectivity of HMF is respectively 41.7%, 36.6%
(
Table 3, Entry 2–3). Considering that diꢁerent types of
acid sites may contribute diꢁerent reactivity, H-Beta (100)
with equivalent amounts of total of medium strength and
strong acid sites was used for comparison (Table 3, Entry
4
). It is discovered that the two catalysts present diꢁer-
ent reaction selectivity toward the hydroxymethylation of
furfural to HMF although they have a comparable fur-
fural conversion. This may be attributed to the fact that
mesoporous NbPO has relatively larger porous structure
which is beneꢀcial to the diꢁusion of products of HMF in
the material. Based on these data, it can be concluded that
the mesoporous NbPO is a more suitable catalyst for the
production of HMF under the speciꢀed hydroxymethyla-
tion reaction conditions.
Table 3 Hydroxymethylation
Entry
Catalyst
Catalyst loading Furfural Con. HMF yield (%)
HMF Sel. (%)
of furfural to HMF on various
catalysts in the aqueous solution
(
mg)
(%)
1
2
3
4
NbPO
400
407
355
440
63.8
58.5
37.3
63.1
36.7
24.4
13.6
25.8
57.5
41.7
36.6
40.9
H-Beta (100)
H-ZSM-5 (50)
H-Beta (100)
Reaction conditions: furfural (2 mmol), formalin (10 mL), time (12 h), temp. (90 °C), stirring (500 rpm).
All catalysts were calcined at 550 °C for 3 h before use
1
3

Exploration on Hydroxymethylation of Furfural to 5-Hydroxymethylfurfural Over Mesoporous…
Table 4 Eꢁects of the solvents on the hydroxymethylation of furfural
to HMF
Entry Co-solvent Furfural
Con. (%)
HMF yield (%) HMF Sel. (%)
1
2
3
4
5
6
H O
61.5
59.3
62.3
66.6
62.7
7.7
36.1
36.9
38.1
44.8
45.5
1.8
58.7
62.2
61.1
67.3
72.6
23.5
2
CH CN
3
NMP
DMF
DMSO
DMSO
a
Reaction conditions: furfural (2 mmol), formalin (10 mL), sol-
vent (10 ml), catalyst (400 mg), time (12 h), temp. (90 °C), stirring
(
500 rpm). All catalysts were calcined at 550 °C for 3 h before used
Without catalyst loading
a
Fig. 6 Eꢁects of catalyst dosage on furfural conversion, HMF selec-
tivity, and HMF yield. Reaction conditions: furfural (2 mmol), for-
malin (10 mL), DMSO (10 ml), time (12 h), temp. (90 °C), stirring
(500 rpm).
result in the reaction carried out in the aqueous solution.
When aqueous N-Methyl pyrrolidone solvent was used, the
conversion of furfural was 62.3%, and the HMF selectiv-
ity was 61.1%. When aqueous DMSO solvent alone (with-
out any catalysts) was used as a solvent, the conversion of
furfural was 7.7% and the selectivity of HMF was 23.5%,
which conꢀrm that DMSO doesn’t play a catalytic role in
the course of reaction. All these results indicate that DMSO
is a reliable solvent for the hydroxymethylation of furfural
to HMF.
and the yield of HMF, the dosage of catalyst was kept at
0.4 g in the following experiments if not otherwise indicated.
3.2.4 Catalytic Hydroxymethylation of Furfural
to 5‑Hydroxymethylfurfural Over Mesoporous
Niobium Phosphate Catalyst at Diꢀerent Reaction
Temperatures
3
.2.3 Catalytic Hydroxymethylation of Furfural
to 5‑Hydroxymethylfurfural Over Mesoporous
Niobium Phosphate Catalyst at Diꢀerent Catalyst
Dosages
Eꢁects of reaction temperature on furfural conversion, selec-
tivity to HMF, and HMF yield were further investigated
from 70 to 120 °C in the presence of mesoporous NbPO
To further optimize the reaction conditions, eꢁects of cata-
lyst dosage to the reaction were investigated in the presence
of diꢁerent amounts of mesoporous NbPO with aqueous
DMSO as solvent. Figure 6 shows the correlations between
catalyst dosage and furfural conversion, HMF selectivity
and yield. It can be seen that the increase in the dosage of
catalyst from 0.2 to 1 g can greatly increase the conversion
of furfural from 41.2 to 88.7% at the reaction temperature
of 90 °C for 12 h, while the selectivity of HMF decreases
gradually from 75.3 to 43.5%. This might be ascribed to the
strong acid of mesoporous NbPO, which can not only cata-
lyze the hydroxymethylation of furfural to HMF, but also
further hydrolyze HMF to produce levulinic acid and formic
acid during the reaction process. Actually, a small amount
of by-products including levulinic acid and formic acid
were detected during products analysis by HPLC. When the
dosage of the catalyst was 0.6 g, the highest yield of HMF
Fig. 7 Eꢁects of reaction temperature on furfural conversion,
(
47.6%) was achieved, while the conversion of furfural and
selectivity to HMF, and HMF yield. Reaction conditions: furfural
the selectivity of HMF was 73.1% and 65.2%, respectively.
In order to seek a balance between the selectivity of HMF
(
2 mmol), formalin (10 mL), DMSO (10 ml), catalyst (400 mg), time
. 12 h), stirring (500 rpm).
1
3

H. Zhang et al.
with aqueous DMSO as solvent. The results are summarized
in Fig. 7. It can be seen that both the conversion of furfural
and the selectivity of HMF are sensitive to the temperature
in the hydroxymethylation. Generally, the higher the reaction
temperature, the higher the conversion of furfural can be
obtained. When the reaction temperature increased from 80
to 120 °C, the conversion of furfural increased from 48.6 to
HMF decreased from 64.5 to 53.4%. This may be attributed
to the further conversion of HMF into byproducts.
3.2.6 Stability of Mesoporous Niobium Phosphate Catalyst
Stability of the mesoporous NbPO catalyst was inves-
tigated since it is a very important property for the solid
catalyst. The hydroxymethylation of furfural to HMF over
mesoporous NbPO catalyst was carried out at 90 °C for
20 h with aqueous DMSO as solvent, which was used as a
model reaction to evaluate the stability of current prepared
mesoporous NbPO catalyst. After every reaction run, the
catalyst was recovered by centrifugation and thoroughly
washed with 75% ethanol. It was then dried at 100 °C in a
vacuum oven for 4 h and calcined at 550 °C for 3 h under
an air atmosphere. After that, the catalyst was reused under
the same conditions. In the catalyst recycling test, in order
to avoid the impact of catalyst loss, each reuse was scaled
down. The results in Fig. 9 show that even after six cycles,
furfural conversion, and HMF selectivity and yield are
almost the same, demonstrating that the catalyst is stable
without an obvious decrease in reactivity after 6 runs.
In fact, lots of homogeneous acid catalysts have been used
to produce HMF. However, homogeneous acid catalysts
still suꢁer from many defects, such as equipment corrosion,
diꢂculty in separation and toxicity, which could be elimi-
nated by the use of solid acid catalysts. In recent years, more
attention has been drawn to the use of solid acid catalysts,
including H-zeolites, strong acidic ion exchange resins and
metal oxides. Being a strong solid acid catalyst, niobium
phosphate is extensively used in lots of reactions, such as
esteriꢀcation [21], hydrodeoxygenation [22], dehydrogena-
tion [23], dehydration [24] and etheriꢀcation [25]. Niobium
phosphate containing Brönsted and Lewis acid sites which
are tunable is considered as a very important water tolerant
9
6.2%, while the selectivity of HMF decreased dramatically
from 76.2 to 39.5%. In spite of the fact that high temperature
can accelerate the conversion of furfural, it can also aggra-
vate the degradation of HMF, leading to a low yield of HMF.
The highest HMF yield is 45.5%, which is obtained at a reac-
tion temperature of 90 °C. This temperature was selected as
the reaction temperature for the further investigation.
3
.2.5 Catalytic hydroxymethylation of furfural
to 5‑hydroxymethylfurfural over mesoporous
niobium phosphate catalyst at diꢀerent reaction
times
The eꢁects of reaction time on the conversion of furfural and
selectivity to HMF were also investigated in the presence
of mesoporous NbPO with aqueous DMSO as solvent at
9
0 °C, as shown in Fig. 8. It was found that the conversion
of furfural and the yield of HMF increased progressively
with the increase in reaction time, although the selectivity to
HMF decreased. The highest yield of HMF of about 51.1%
is obtained after 20 h of reaction while furfural conversion
and HMF selectivity is 79.2% and 64.5% respectively. Fur-
ther prolonging reaction time would enhance the conversion
of furfural from 79.2 to 84.5%. Meanwhile, the selectivity of
Fig. 8 Eꢁects of reaction time on furfural conversion, selectivity to
HMF, and HMF yield. Reaction conditions: furfural (2 mmol), forma-
lin (10 mL), DMSO (10 ml), catalyst (400 mg), temp. (90 °C), stir-
ring (500 rpm).
Fig. 9 Cycle usage of mesoporous niobium phosphate catalyst
1
3

Exploration on Hydroxymethylation of Furfural to 5-Hydroxymethylfurfural Over Mesoporous…
catalyst [18]. As mentioned by various researchers, organic
solvents such as dimethyl sulfoxide (DMSO) [26], dimeth-
ylformamide (DMF) [27] and g-valerolactone (GVL) [28]
could form a shielding ꢀlm around HMF to stabilize HMF
in the synthesis process, leading to a high yield of HMF
References
1. Teong SP, Yi GS, Zhang YG (2014) Green Chem 16:2015–2026
2
.
Cho A, Byun S, Cho JH, Kim BM (2019) Chemsuschem
1
2:2310–2317
3
.
Casanova O, Iborra S, Corma A (2009) J Catal 265:109–116
[
29]. Based on the advantages of niobium phosphate shown
4. Xia Q, Xia Y, Xi J, Liu X, Zhang Y, Guo Y, Wang Y (2017)
Chemsuschem 10:747–753
above and accounting for the fact that organic solvents can
improve the yield of HMF by stabilizing HMF in the syn-
thesis process, niobium phosphate was expected to give a
better catalytic performance when it is used as a catalyst
for hydroxymethylation of furfural to HMF under an aque-
ous organic solvent. The comprehensive evaluation results
revealed that the mesoporous niobium phosphate demon-
strated an excellent catalytic activity in the hydroxymeth-
ylation of furfural to HMF, which may provide a possibility
for commercial HMF production. The highest HMF yield of
5
.
Ardemani L, Cibin G, Dent AJ, Isaacs MA, Kyriakou G, Lee AF,
Parlett CMA, Parry SA, Wilson K (2015) Chem Sci 6:4940–4945
Kroger M, Prusse U, Vorlop KD (2000) Top Catal 13:237–242
6
.
7. Zhang F, Liu Y, Yuan F, Niu X, Zhu Y (2017) Energy Fuels
31:6364–6373
8
9
.
.
Petri A, Masia G, Piccolo O (2018) Catal Commun 114:15–18
Hou QD, Zhen MN, Li WZ, Liu L, Liu JP, Zhang SQ, Nie YF, Bai
CYL, Bai XY, Ju MT (2019) Appl Catal B-Environ 253:1–10
10. Swift TD, Nguyen H, Anderko A, Nikolakis V, Vlachos DG
(2015) Green Chem 17:4725–4735
1
1. Dick GR, Frankhouser AD, Banerjee A, Kanan MW (2017) Green
Chem 19:2966–2972
5
1.1%, more than the yields reported previous [15, 16], was
1
2. Menegazzo F, Ghedini E, Signoretto M (2018) Molecules
23:2201–2218
obtained with a 79.2% furfural conversion under an aque-
ous organic solvent system of DMSO and water at 90 °C for
13. Mittal A, Black SK, Vinzant TB, O’Brien M, Tucker MP, Johnson
DK (2017) ACS Sustain Chem Eng 5:5694–5701
2
0 h. Considering that DMSO is not sustainable, the cata-
1
4. Jerome Lecomte AF (1999) Claude Moreau. Ind Crops Prod
lytic performances of NbPO in water was also investigated
systematically (see the supporting information). The highest
HMF yield was 44.8%, which is signiꢀcantly lower than the
HMF yield in a water-DMSO mixed solvent. This can be
attributed to that the organic solvents could form a shielding
1
9:235–241
15. Nishimura S, Shibata A, Ebitani K (2018) ACS Omega
3:5988–5993
1
1
6. Nishimura S, Shibata A (2019) Catalysts 9:314
7. Wang Z, Liu H, Cui H, Zhang M, Zhang Z (2015) Ind Eng Chem
Res 54:7219–7225
ꢀ
lm around HMF to stabilize HMF in the synthesis process,
18. Zhang Y, Wang J, Ren J, Liu X, Li X, Xia Y, Lu G, Wang Y
(2012) Catal Sci Technol 2:2485–2491
leading to a high yield of HMF.
1
2
9. Sato H (1997) Catal Rev 39:395–424
0. Rocha AS, Forrester AMS, de la Cruz MHC, da Silva CT, Lachter
ER (2008) Catal Commun 9:1959–1965
4
Conclusion
21. Rade LL, Lemos COT (2019) Barrozo MAdS, Ribas RM, Mon-
teiro RdS, Hori CE. Renew Energy 131:348–355
2
2. Schade OR, Kalz KF, Neukum D, Kleist W, Grunwaldt J-D (2018)
Green Chem 20:3530–3541
In summary, one kind of acid adjustable mesoporous nio-
bium phosphate was successfully synthesized and used as a
catalyst in the hydroxymethylation of furfural to HMF. Com-
pared with H-ZSM-5 (100) and H-Beta (50), the prepared
mesoporous niobium phosphate demonstrated a superior
activity and selectivity. The highest yield of HMF of 51.1%
could be achieved under the optimal reaction condition over
the current catalyst. Moreover, the prepared mesoporous
niobium phosphate demonstrated a stable catalytic activity.
Such an acidity adjustable and mesoporous catalyst could
open a promising prospective for future research and large-
scale production of HMF.
2
3. Weng WH, Davies M, Whiting G, Solsona B, Kiely CJ, Carley
AF, Taylor SH (2011) Phys Chem Chem Phys 13:17395–17404
24. Tanabe K (2003) Catal Today 78:65–77
2
5. Kumari PK, Rao BS, Dhana Lakshmi D, Sai Paramesh NR,
Sumana C, Lingaiah N (2019) Catal Today 325:53–60
2
6. van Putten RJ, van der Waal JC, de Jong E, Rasrendra CB, Heeres
HJ, de Vries JG (2013) Chem Rev 113:1499–1597
27. Fan G, Wang Y, Hu Z, Yan J, Li J, Song G (2018) Carbohydr
Polym 200:529–535
2
8. Gallo JMR, Alonso DM, Mellmer MA, Dumesic JA (2013) Green
Chem 15:85–90
29. Mushrif SH, Caratzoulas S, Vlachos DG (2012) Phys Chem Chem
Phys 14:2637–2644
Publisher’s Note Springer Nature remains neutral with regard to
Acknowledgements This research did not receive any speciꢀc grant
from funding agencies in the public, commercial, or not-for-proꢀt
sectors.
jurisdictional claims in published maps and institutional aꢂliations.
Compliance with Ethical Standards
Conflict of interest There are no conꢃicts to declare.
1
3