Glucose dehydration to 5-hydroxymethylfurfural by a combination of a basic zirconosilicate and a solid acid

Article abstract of DOI:10.1007/s10562-014-1390-6

A recently reported layered zirconosilicate Na₂ZrSi₄O₁₁ displays good activity in the isomerization of glucose to fructose in water at mild conditions. Part of the activity derives from the homogeneous base-catalyzed reaction due to exchange of the sodium ions of the layered zirconosilicate in water. Following ion-exchange, the isomerization is mainly catalyzed by the basic sites of the re-used heterogeneous zirconosilicate catalyst. Combined with the solid acid Amberlyst-15, 5-hydroxymethylfurfural (5-HMF) can be produced from glucose in a one-pot reaction. In a THF/H₂O mixture solvent system, 5-HMF was obtained with 45 % selectivity at 87 % glucose conversion at a temperature of 180 °C in 1.5 h. Graphical Abstract: [Figure not available: see fulltext.]

Full text of DOI:10.1007/s10562-014-1390-6

Catal Lett
DOI 10.1007/s10562-014-1390-6
Glucose Dehydration to 5-Hydroxymethylfurfural
by a Combination of a Basic Zirconosilicate and a Solid Acid
•
Chaochao Yue Marcello S. Rigutto
•
Emiel J. M. Hensen
Received: 10 July 2014 / Accepted: 3 October 2014
Ó Springer Science+Business Media New York 2014
Abstract A recently reported layered zirconosilicate
Na2ZrSi4O11 displays good activity in the isomerization of
glucose to fructose in water at mild conditions. Part of the
activity derives from the homogeneous base-catalyzed
reaction due to exchange of the sodium ions of the layered
zirconosilicate in water. Following ion-exchange, the
isomerization is mainly catalyzed by the basic sites of the
re-used heterogeneous zirconosilicate catalyst. Combined
with the solid acid Amberlyst-15, 5-hydroxymethylfurfural
(5-HMF) can be produced from glucose in a one-pot
reaction. In a THF/H2O mixture solvent system, 5-HMF
was obtained with 45 % selectivity at 87 % glucose con-
version at a temperature of 180 °C in 1.5 h.
raw material for the production of 5-hydroxymethyl fur-
fural (5-HMF), which is a prospective platform molecule
for the synthesis of polymers and biofuels [3, 4].
The conversion of glucose to 5-HMF is thought to
proceed via isomerization of glucose to fructose followed
by dehydration of fructose to 5-HMF [5]. The isomeriza-
tion reaction can be promoted by a variety of catalysts,
including bases and Lewis acids [6, 7]. Industrially, glu-
cose isomerization is carried out by enzymes for the pro-
duction of fructose syrups [8]. Fructose dehydration to
5-HMF is a very facile reaction and can be catalyzed by a
wide range of mineral as well as solid Brønsted and Lewis
acids (Scheme 1) [9, 10]. Whereas fructose dehydration
can be made to react to 5-HMF with yields as high as 90 %
[11], conversion of glucose to 5-HMF remains problematic
with much lower yields. A potentially effective approach is
to combine two functionalities in one pot [12], that is, the
isomerization of glucose to fructose and dehydration of the
latter to 5-HMF. For instance, in a biphasic system a
5-HMF yield of 68 % was achieved by combining enzy-
matic and acid catalysts [13]. In another work, glucose
conversion and 5-HMF selectivity of 73 and 58 %,
respectively, were reported by combining basic hydrotal-
cites with acidic Amberlyst-15 in N, N-dimethylformamide
(DMF) [14]. Also, a bifunctional sulfated ZrO2 catalyst
with acid as well as base properties was applied in glucose
dehydration [15]. It was reported that glucose is first
isomerized by Lewis acid and base sites of ZrO2 and the
formed fructose was subsequently dehydrated to 5-HMF by
Brønsted acid sites involving mono- or bidentate sulfate
groups. Sn-containing Beta zeolite has most recently been
identified as an excellent catalyst for the isomerization of
glucose to fructose in water [7]. Combined with HCl,
glucose can be converted to 5-HMF with 72 % selectivity
at 79 % conversion of the starting sugar at 180 °C [16].
Keywords Zirconosilicate Á 5-hydroxymethylfurfural Á
Glucose isomerization
1 Introduction
The impending oil production peak and the ever increasing
global energy demand stimulate substantial efforts in
identifying renewable sources of carbon for the production
of fuels and chemicals [1, 2]. Glucose, the most abundant
monosaccharide and the cheapest hexose, is a promising
C. Yue Á E. J. M. Hensen (&)
Inorganic Materials Chemistry, Eindhoven University of
Technology, P.O. Box 513, 5600 MB Eindhoven,
The Netherlands
e-mail: e.j.m.hensen@tue.nl
M. S. Rigutto
Shell Global Solutions International B.V., P.O. Box 38000,
1030 BN Amsterdam, The Netherlands
e-mail: marcello.rigutto@shell.com
123

C. Yue et al.
Scheme 1 Schematic
representation of the mechanism
of glucose isomerization and
dehydration
2.1.3 Preparation of ZrO2 (Tetragonal ZrO2)
Earlier, Zhao et al. have reported the use of CrCl2 in
1-ethyl-3-methylimidazolium chloride resulting in
a
ZrOx(OH)4-2x was prepared by the hydrolysis of 0.5 M
ZrOCl2 (Aldrich Chemicals,[98 %) aqueous solutions fed
at 500 cm3 h-1 into a well-stirred vessel with a pH of 10
held constant by the controlled addition of 14 M NH4OH.
The precipitate was dried at 150 °C overnight after residual
Cl- ions were removed (\10 ppm) by thorough washing.
Dried ZrOx(OH)4-2x was calcined at 550 °C for 6 h. The
product was tetragonal ZrO2.
5-HMF selectivity of 72 % at a glucose conversion of 94 %
at 80 °C [17]. A heterogeneous version of this catalyst
consisting of ionic liquid PMIm?Cl- (PMIm? = 1-propyl-
3-methylimidazolium), coordinating the CrCl2 catalytic
centers, covalently grafted on SBA-15 gave a 5-HMF
selectivity of 69 % at a glucose conversion of 50 % [18].
Despite the promise of some of the approaches above, there
is substantial room for the development of novel cheap
catalysts, which can operate in green solvents. As such, it
may also be mentioned that the synthesis of Sn-Beta is
cumbersome and tedious, while the use of ionic liquids as
solvent has also drawbacks such as the need to extract the
desired dehydration products.
2.2 Characterization
Inductively coupled plasma optical emission spectros-
copy (ICP-OES) measurements were performed on a
Spectro CIROS spectrometer equipped with a free run-
ning 27.12 MHz generator at 1,400 W. Prior to mea-
surements, the samples were melted with lithium
tetraborate and dissolved in diluted nitric acid. Powder
X-ray diffraction (XRD) patterns were measured on a
Bruker D4 Endeavor powder diffraction system using Cu
Ka radiation. Temperature programmed desorption
(TPD) of pre-adsorbed CO2 was carried out to determine
the surface basicity. An amount of 150 mg catalyst was
placed in a tubular quartz reactor. The sample was
treated in a He flow of 100 ml/min at 200 °C followed
by cooling to 50 °C. The sample was then exposed to
CO2 pulses until saturation occurred. After the material
was swept with He (flow rate 100 ml/min) at 50 °C for
5 h, the temperature was increased at a rate of 10 °C/
min in 50 ml/min He to 650 °C. The signals of He and
CO2 were recorded by online mass spectrometry (quad-
rupole mass spectrometer, Balzers TPG-300).
In this work, we report about the catalytic properties of a
basic layered zirconosilicate for glucose isomerization in
water and its combination with Amberlyst-15 solid acid to
obtain 5-HMF from glucose in a one-pot reaction. We also
investigate the use of other solvents than water and opti-
mize the reaction conditions.
2 Experimental
2.1 Synthesis of ZrSil, ETS-10, AV-3 and Magadiite
2.1.1 Synthesis of Zirconosilicate (ZrSil), ETS-10
and AV-3
ZrSil was prepared by the method recently reported by us
[19]. Briefly, a gel with the composition 1.75 Na2O: 0.28
K2O: 1.0 SiO2: 0.15 ZrO2: 25 H2O was crystallized in a
Teflon-lined autoclave under static conditions at 230 °C for
10 days. AV-3 and ETS-10 were synthesized according to
known literature procedures [19–21].
2.3 Catalytic Activity Measurements
The isomerization of glucose to fructose was performed in
a closed, magnetically stirred glass reactor tube (Ace
Pressure Tube, Sigma-Aldrich) under autogenous pressure.
The glass reactor was heated by immersion in a thermo-
stated oil bath. The reaction mixture consisted of 75 mg D-
glucose, different amounts of base and/or acid catalysts in
1.5 mL solvent. The reaction was carried out at 80, 100,
120, 140 or 150 °C at a stirring rate of 700 rpm. The
dehydration of glucose to 5-HMF was carried out in a
2.1.2 Synthesis of Magadiite
Magadiite was prepared according to a literature procedure
[22]. A gel was made by mixing a silicate solution (30 wt%
colloidal silica suspension in water), NaOH and water. The
gel with the final composition 1.0 NaOH: 3.0 SiO2: 50.0
H2O was crystallized in a Teflon-lined autoclave under
static conditions at 150 °C for 3 days.
123

Combination of a Basic Zirconosilicate and a Solid Acid
stainless steel autoclave heated by immersion in a
thermostated oil bath. The reaction mixture consisted of
75 mg D-glucose, different amounts of base and/or acid
catalysts in 1.5 mL solvent at 160 or 180 °C. After reac-
tion, the glass or stainless steel autoclave was quenched in
water and its contents was analyzed by a Shimadzu HPLC
system with ELSD and UV detectors. Glucose, fructose
and mannose were detected by an ELSD detector using a
Prevail Carbohydrate ES (Grace) column. The mobile
phase (1.0 mL/min) was acetonitrile/water (70:30 v/v) and
the column temperature 50 °C. 5-HMF was detected by
using a UV detector (320 nm) with a Pathfinder PS C18
reversed phase column. The mobile phase (0.4 mL/min)
was methanol/water (20:80 by volume) and the column
temperature 30 °C. Lactic acid content was evaluated by
use of a Prevail Organic Acid column in combination with
a UV detector (220 nm). The mobile phase (0.5 mL/min)
was 25 mM potassium phosphate with pH 2.5 and the
column temperature 45 °C. The conversion was deter-
mined from the glucose concentration determined by
HPLC analysis in the starting mixture and in the mixture
after reaction. Selectivity was based on the concentration
of the products determined by HPLC analysis.
Fig. 1 XRD patterns of a ZrSil, b used ZrSil after three rcycles,
c Magadiite, d ETS-10 and e AV-3
[23]. It is formed by protonation of the enol intermediate
occurring during glucose isomerization [7]. The glucose
isomerization reaction rate and selectivity depend on
reaction temperature. The total saccharide yield decreases
with increasing temperature due to degradation of sugars in
the basic solution [24, 25]. The ZrSil catalyst outperforms
other basic materials such as the sodium silicate magadiite
and the metallosilicates ETS-10 and AV-3. As shown in
Table 1, the fructose yields are only 22, 20 and 17 % for
ETS-10, AV-3 and magadiite, respectively, after a reaction
time of 1 h. Zirconia (ZrO2) exhibits low activity in glu-
cose isomerization, which should be due to the lower
basicity of ZrO2 compared to ZrSil. Besides solid bases,
enzymes and solid Lewis acids have also been used for the
isomerization of glucose to fructose [7, 8]. Although strict
reaction conditions need to be followed when using
enzymes, fructose yields are generally as high as 42 % at
60 °C [8]. The fructose yield in water using Sn-Beta zeolite
as the catalyst is 32 % at a glucose conversion of 55 % at a
reaction temperature of 110 °C after 30 min [7]. Thus, the
glucose isomerization activity of our zirconosilicate is
comparable to that of Sn-Beta zeolite.
3 Results and Discussion
3.1 Isomerization of Glucose to Fructose Using Solid
Base
The powder XRD patterns of AV-3, magadiite, ETS-10
and the layered zirconosilicate are shown in Fig. 1. The
patterns of the first three materials agree with published
patterns [21, 22]. The layered zirconium silicate ZS with
the chemical formula Na2ZrSi4O11 (Fig. 1a) comprising
SiO4 tetrahedra and ZrO6 octahedra contains two types
of Na sites. One type of sodium ion is likely located in
the interlayer, the other one in the intralayer. More
detailed characterization of this ZrSil zirconosilicate has
been reported elsewhere [19]. Briefly, the surface area of
ZrSil is about 10 m2/g. The Na/Zr ratio is 2.1. The Na
cations balancing the negative charge in the framework
of zirconium silicate can be exchanged for protons and,
in this way, impart basic properties in contact with
water.
The reusability of ZrSil was tested by performing three
consecutive isomerization cycles at 100 °C for 1 h. After
each reaction, the catalyst was filtered and washed with
water before adding fresh glucose solution. From Table 1,
it follows that the reused catalysts exhibit lower activity
than the fresh one. After three reaction cycles, the fructose
yield is *18 %, which is almost the same as during the
first reuse. The Na-to-Zr ratio is about 1.4 in spent ZrSil,
lower than in the parent material. We speculate that the
sodium ions located in the interlayer space of ZrSil were
lost by exchange with water during the first reaction cycle,
whereas sodium ions located in the small cages of ZrSil
remain in the material. Consistent with the loss of
Results of application of these materials in the isomer-
ization of glucose to fructose are shown in Fig. 2. ZrSil is a
highly active catalyst for the isomerization of glucose to
fructose in water at 100 °C. The fructose yield is highest
(26 %) after 7 min. After 1 h, the glucose conversion
reached its maximum value and the reaction mixture con-
tained 43 % glucose, 26 % fructose and 17 % mannose.
Mannose is the product of the epimerization of glucose
123

C. Yue et al.
Table 1 Glucose isomerization in water by solid bases (75 mg glu-
cose, 30 mg ZrSil, 1.5 mL H2O, T = 100 °C, reaction time 1 h)
Catalyst
Xglucose (%) Sfructose (%) Yfructose (%)
ZrSila
56
37
30
24
48
52
63
76
27
19
19
18
ZrSil-reuse1b
ZrSil-reuse2b
ZrSil-reuse3b
ZrSil
80 °C
120 °C
160 °C
ZrSil—hot filtratec
ZrSil-reuse1—hot filtratec
ZrO2
46
57
65
25
8
52
33
32
40
38
25
23
56
33
24
19
16
10
3
8
2
Fig. 2 Glucose isomerization as a function of reaction time for the
zirconosilicate ZS: (black filled square) glucose conversion, (black
filled triangle) fructose yield and (black filled circle) mannose yield
(75 mg glucose, 30 mg ZrSil, 1.5 mL H2O, T = 100 °C)
AV-3
74
39
64
17
22
21
Magadiite
ETS-10
a
The yield of lactic acid was lower than 0.5 %
b
Spent ZrSil was separated by filtration, washed with water and
dried at 110 °C for 10 h, before being reused under similar conditions
exchangeable sodium, we observed that the pH of an
aqueous suspension of reused ZrSil (30 mg reused ZrSil in
1.5 mL water, 1 h) was almost neutral (pH & 7) in con-
trast to the pH of an aqueous suspension of fresh ZrSil
(same conditions, pH & 8). It was also found that the
fructose selectivity was higher when the reaction was
carried out in the lower pH solvent. The crystallinity of
ZrSil was found to be lower after three recycles (Fig. 1b),
which may further contribute to the activity loss.
c
Catalyst stirred in water for 30 min and removed by filtration,
glucose was then added to the filtrate and a reaction time of 1 h was
allowed for this solution
X conversion, S selectivity, Y yield
We speculate that the additional strong basic sites repre-
sented by the peak at 640 °C may be due to super basic
sites in a Na/ZrO2-SiO2-type phase, similar to the basic
sites obtained by loading ZrO2/SBA-15 with Na? or K?
[28]. In previous work, metal oxides were compared as
solid bases for glucose isomerization to fructose [27]. It
was found that glucose isomerization is catalyzed by weak
and medium basic sites, whereas very strong basic sites
were not active. Our data also suggest that very strong
basic sites formed in the reused catalyst are probably not
active for glucose isomerization. Figure 3 shows that the
strength of medium basic sites is weaker in reused ZrSil
than in fresh ZrSil, which may also contribute to the lower
isomerization activity of reusing ZrSil.
We investigated in more detail the heterogeneous
nature of the catalyst system. After stirring ZrSil or
reused ZrSil for 30 min in water and removing the solid
by filtration, we added glucose to the filtrate and ran a
reaction experiment at 100 °C for 1 h. For the fresh
catalyst, it was observed that the filtrate could convert
glucose due to its basicity, arising from partial exchange
of Na? in ZrSil with protons in the solution. This clearly
shows that there is a homogeneous contribution to the
reaction. After the first recycle, the contribution of the
homogeneous reaction was much lower, indicating that
the activity of the recycled ZrSil materials is mainly
heterogeneous in nature.
When carried out in DMSO, the ZrSil exhibited lower
isomerization activity (Table 2). At 100 °C, the glucose
conversion and fructose yield were only 22 and 6 %,
respectively. The yield was higher at a temperature of
120 °C and remained fairly constant at higher reaction
temperatures. It is seen that increasing temperature leads to
higher glucose conversion but at the expense of fructose
selectivity. The latter observation should be due to sugar
degradation. The glucose isomerization activity in aprotic
DMSO solvent is lower for ZrSil than earlier observed for
hydrotalcites [14].
The basicity of the ZrSil samples was evaluated by TPD
of adsorbed CO2 (Fig. 3). For fresh ZrSil, we observed
three desorption peaks at 140, 470 and 570 °C. The latter
two peaks correspond to strong basic oxygen sites [26],
while the one at 140 °C represents weak basic sites. After
the first use of ZrSil, the CO2 TPD patterns shows four
peaks at 115, 356, 560 and 640 °C. The desorption peak at
around 560 °C corresponding to the strong basic sites are
retained in ZrSil. Medium basic sites represented by the
peak at 356 °C are associated with collapsed ZrSil [27].
123

Combination of a Basic Zirconosilicate and a Solid Acid
experiments, consistent with the notion that it is one of
the reaction intermediates in the conversion of glucose
into 5-HMF. By-product formation in the form of higher
molecular weight products (humins) was evident from the
yellow to brown color of the mixtures after reaction. The
reaction mixtures typically became darker when reactions
were carried out at higher base-to-acid ratios. This indi-
cates that humins formation arose during the dehydration
of fructose to 5-HMF in good agreement with literature
[31–34]. When the amount of Amberlyst-15 was
decreased to 25 mg, whilst keeping the amount of ZrSil at
200 mg, the 5-HMF yield and selectivity were lower.
Notably, the conversion was slightly lower, implying that
more by-products had formed.
Direct reuse of the mixed zirconosilicate/Amberlyst-15
catalyst after separation from the initial reaction mixture
led to a 5-HMF yield of only 4 %. From the observation
that the anhydroglucoses yield was significantly lowered,
we infer that the Amberlyst-15 catalyst was deactivated.
The deactivation is caused by ion-exchange of the acid
protons of Amberlyst-15 with Na cations originating from
the ZrSil solid base. Spent Amberlyst-15 contained about
1.5 wt% Na as determined by EDX analysis (the Na con-
tent of fresh Amberlyst-15 was negligible). Thermal deg-
radation of Amberlyst-15 at high temperature may also
contribute to deactivation. The ZrSil catalyst was still able
to isomerize glucose as the fructose yield was 8 % in this
case. Therefore, in another experiment the ZrSil powder
was separated from the larger Amberlyst-15 particles after
the first reaction and then combined with fresh Amberlyst-
15 (50 mg). This indeed led to higher anhydroglucoses
yield and also improved the 5-HMF yield slightly, but not
to the level of the fresh catalyst. Accordingly, we conclude
that the ZrSil catalyst deactivates during the glucose
isomerization in DMSO. Tentatively, we attribute this to
adsorption of high-molecular weight products on the cat-
alyst surface. It is difficult to assert this as oxidation of the
spent catalyst at elevated temperatures in order to remove
such products would lead to structural collapse of the
zirconosilicate.
Fig. 3 Temperature-programmed desorption of adsorbed CO2 for
a ZrSil and b ZrSil-reuse1
Table 2 Glucose isomerization in DMSO as a function of tempera-
ture (75 mg glucose, 30 mg ZrSil, 1.5 mL DMSO, reaction time 1 h)
Temperature (oC)
Xglucose (%)
Yfructose (%)
Sfructose (%)
100
120
140
160
22
23
44
63
6
11
10
13
29
46
24
21
X conversion, Y yield, S selectivity
3.2 Dehydration of Glucose to 5-HMF by Combining
Solid Base and Acid
DMSO has been reported to be one of the better solvents
when the aim is to obtain 5-HMF from carbohydrates [3,
28]. Accordingly, we evaluated the direct formation of
5-HMF from glucose by using a combination of the zir-
conosilicate ZrSil and the strongly acidic ion-exchange
resin Amberlyst-15 in DMSO. In order to suppress
dehydration of glucose to anhydroglucoses by the solid
acid catalyst [29, 30], the ratio of solid acid to base
catalysts was varied (Table 3). When the amount of solid
base was relatively small, anhydroglucoses were the main
products in DMSO. The fructose and 5-HMF yields were
very low. With increasing solid base amount, it is seen
that glucose conversion increased together with the
5-HMF yield with the yield of anhydroglucoses decreas-
ing. Optimum 5-HMF yield was observed for a combi-
nation of 50 mg Amberlyst-15 and 200 mg ZrSil. At a
higher ZrSil/Amberlyst-15 ratio the 5-HMF yield was
slightly lower. The fructose yield was always low in these
One-pot dehydration of glucose to 5-HMF by using solid
catalysts has been reported before [16, 29]. The highest
yield of 5-HMF was about 56 % by using Sn-Beta with
HCl and NaCl at a temperature of 180 °C. Addition of
NaCl improved the partitioning of 5-HMF into the organic
phase, increasing 5-HMF selectivity [16]. The Ebitani
group has combined a basic hydrotalcite with Amberlyst-
15 to convert glucose to 5-HMF [29]. The highest reported
yield of 5-HMF was 42 % in DMF after 9 h. The 5-HMF
yield in DMSO was 25 %. Compared to the present results
for the conversion of glucose to 5-HMF, the yields of
123

C. Yue et al.
Table 3 One-pot synthesis of 5-HMF from glucose using a combination of the zirconosilicate solid base (ZrSil) and Amberlyst-15 (75 mg
glucose, 1.5 mL DMSO, T = 160 °C, reaction time 0.5 h)
Amberlyst-15 (mg)
ZrSil (mg)
Base-to-acid ratio
Xglucose (%)
Yfructose (%)
Y5-HMF (%)
S5-HMF (%)
YaAHG (%)
50
50
50
50
50
50b
50c
50
25
a
0
30
–
98
80
76
78
78
73
83
73
73
0
2
3
3
3
8
3
3
3
0
4
0
62
46
21
16
15
4
0.6
1.2
2.0
4.0
4.0
4.0
6.0
8.0
5
60
7
9
100
200
200b
200b
300
200
8
10
24
6
18
4
7
8
30
14
14
18
12
22
14
AHG: anhydroglucoses (1,6-anhydro-b-glucopyranose and 1,6-anhydro-b-glucofuranose)
The spent ZrSil and Amberlyst-15 mixture was reused after dry at 110 °C for 10 h
Only the ZrSil solid base was reused and mixed with fresh Amberlyst-15
b
c
X conversion, Y yield, S selectivity
Table 4 Glucose conversion as a function of glucose concentration
by a combination of ZrSil and Amberlyst-15 in DMSO (1.5 mL
solvent, 200 mg ZrSil, 50 mg Amberlyst-15, 1.5 mL DMSO,
T = 160 °C, reaction time 1.0 h)
Glucose
concentration
(mg/mL)
Xglucose Yfructose Y5-HMF S5-HMF YAHG
(%)
(%)
(%)
(%)
(%)
20
82
83
81
66
Trace
Trace
2
19
21
21
17
23
25
26
26
10
10
11
14
30
50
100
Trace
X conversion, Y yield, S selectivity, AHG anhydroglucoses (1,6-
anhydro-b-glucopyranose and 1,6-anhydro-b-glucofuranose)
Fig. 4 Glucose dehydration as a function of the reaction time for an
optimized mixture of zirconosilicate ZrSil and Amberlyst-15: (black
filled square) glucose, (black filled circle) fructose, (filled triangle)
5-HMF yield and (black down triangle) 5-HMF selectivity (200 mg
ZrSil, 50 mg Amberlyst-15, 75 mg glucose, 1.5 mL DMSO,
T = 160 °C)
vary significantly with the initial glucose concentration. It
is observed that the anhydroglucoses yield decreased with
decreasing glucose concentration.
Finally, the effect of the solvent for the dehydration of
glucose to 5-HMF was investigated and the results are
shown in Table 5. In water, both the yield and selectivity of
5-HMF were only 8 %. The 5-HMF yield and selectivity in
a biphasic solvent system comprising 1-butanol/water were
similar to the results obtained when the reaction was car-
ried out in water at 160 °C. The higher fructose yield of
12 % in the biphasic system is due to the protection of this
reaction intermediate in the organic phase. The 5-HMF
yield in the organic solvent DMF is very similar to that
obtained for the reaction in DMSO. The use of aprotic
solvents limits the formation of fructose, which resulted in
lower 5-HMF yields. It is difficult to remove 5-HMF from
high-boiling point solvents such as DMF and DMSO. In
contrast, it has been found that 5-HMF produced by the
action of a Lewis acid Sn-Beta zeolite in combination with
a mineral acid can be easily extracted from a solvent
5-HMF by using a combination of ZrSil and acid are lower
than these previous works but comparable to result
obtained by combining hydrotalcites and solid acids in
DMSO.
It was then investigated how the performance changed
with reaction time for this optimized catalyst mixture.
Figure 4 shows that glucose conversion increased from
78 % after 30 min to 85 % after 2 h but remained similar
after 3 h. Fructose was found in very small amounts in all
cases. The maximum 5-HMF yield (selectivity) was 21 %
(25 %) after a reaction time of 1 h. Table 4 reports the
effect of the glucose concentration on these glucose
dehydration results. Notably, the 5-HMF yield does not
123

Combination of a Basic Zirconosilicate and a Solid Acid
Temperature (oC) Time (h) Xglucose (%) Yfructose (%) Y5-HMF (%) S5-HMF (%)
Table 5 One-pot synthesis of
5-HMF using a combination of
Solvent
ZrSil and Amberlyst-15 in
different solvents (50 mg
Amberlyst-15, 200 mg ZrSil,
1.5 mL solvent, 75 mg glucose)
H2O
160
1.0
1.0
1.0
1.0
1.0
3.0
1.0
1.5
2.0
91
77
79
78
46
76
78
87
91
5
8
5
8
10
24
24
34
34
38
45
41
1-BuOH/H2O (3)a 160
12
DMF
160
160
160
160
180
180
180
Trace
3
19
18
16
26
29
39
37
DMSO
THF/H2O (10)a
THF/H2O (10)a
THF/H2O (10)a
THF/H2O (10)a
THF/H2O (10)a
3
2
Trace
Trace
Trace
X conversion, Y yield,
S selectivity
a
Volume ratio
References
mixture of THF/H2O [16]. Therefore, we also evaluated the
combination of the zirconosilicate ZrSil and Amberlyst-15
in a THF/water mixture. Table 5 shows that after optimi-
zation of the temperature and reaction time (T = 180 °C
and t = 1.5 h) the 5-HMF yield (selectivity) was 39 %
(45 %) at a glucose conversion of 87 %. It is argued that
these promising results are due to stabilization of the
formed 5-HMF in the THF/water mixture [35]. It is rea-
sonable to mention that the basicity imparted in a THF/
water mixture is higher than in DMSO by ion-exchange of
the zirconosilicate. In the THF/water mixture, glucose was
efficiently isomerized into fructose (Table 1) and subse-
quently dehydrated by the Amberlyst-15 catalyst into
5-HMF. The dehydration of glucose to anhydroglucose did
not occur in the THF/H2O mixture, effectively increasing
the rate of 5-HMF formation.
1. Genti G, Van Santen RA (2007) Catalysis for renewables, Wiley-
VCH Weinheim 9
2. Dodds DR, Gross EA (2007) Science 318:1250–1251
3. Torres AI, Daoutidis P, Tsapatsis M (2010) Energy Environ Sci
3:1560–1572
4. Roman-Leshkov Y, Barrett CJ, Liu ZY, Dumesic JA (2007)
Nature 447:982–985
5. Wang TF, Nolte MW, Shanks BH (2014) Green Chem 16:5–48
6. Lima S, Dias AS, Lin Z, Brandao P, Ferreira P, Pillinger M,
Rocha J, Casilda VC, Valente AA (2008) Appl Catal A Gen
339:21–27
7. Roman-Leshkov Y, Moliner M, Labinger JA, Davis ME (2010)
Angew Chem Int Ed 49:8954–8957
8. Bhosale S, Rao M, Deshpande
60:280–300
9. Moreau C, Durand R, Razigade S, Duhamet J, Faugeras P, Ri-
valier P (1996) Appl Catal A Gen 145:211–224
10. Yong G, Zhang YG, Ying JY (2008) Angew Chem Int Ed
47:9345–9348
V (1996) Microbiol Rev
11. Li YN, Wang JQ, He LN, Yang ZZ, Liu AH, Yu B, Luan CR
(2012) Green Chem 14:2752–2758
4 Conclusions
12. Chen BF, Li FB, Huang ZJ, Lu T, Yuan Y, Yuan GQ (2014)
ChemSusChem 7:202–209
13. Huang R, Qi W, Su RX, He ZM (2010) Chem Commun
46:1115–1117
14. Takagaki A, Ohara M, Nishimura S, Ebitani K (2009) Chem
Commun 41:6276–6278
15. Osatiashtiani A, Lee AF, Brown DR, Melero JA, Morales G,
Wilson K (2014) Catal Sci Technol 4:333–342
16. Nikolla E, Leshkov YR, Moliner M, Davis ME (2011) ACS Catal
1:408–410
17. Zhao HB, Holladay JE, Brown H, Zhang ZC (2007) Science
316:1597–1600
18. Degirmenci V, Pidko EA, Magusin PCMM, Hensen EJM (2011)
ChemCatChem 6:969–972
19. Yue C, Magusin PCMM, Mezari B, Rigutto MS, Hensen EJM
(2013) Micropor Mesopor Mater 180:48–55
20. Jeong NC, Lee MH, Yoon KB (2007) Angew Chem Int Ed
46:5868–5872
21. Rocha J, Ferreira P, Lin Z, Agger JR, Anderson MW (1998)
Chem Commun 1269–1270
A basic layered zirconosilicate with the formula Na2-
ZrSi4O11 exhibits high catalytic activity in the isomeriza-
tion of glucose to fructose in water under relatively mild
conditions. The reused zirconosilicate has a lower activity
than the fresh sample. This is due to a homogeneous con-
tribution due to Na? ion exchange in the fresh catalyst.
Besides loss of sodium, the crystallinity of the zirconosi-
licate is also decreased after use as catalyst in water. The
solid acid to solid base ratio was optimized to achieve
optimum 5-HMF yield by combining the zirconosilicate
with Amberlyst-15. In a THF/water solvent system, the
yield and selectivity of 5-HMF from glucose were 39 and
45 %, respectively. The high yield of 5-HMF in THF/water
shows that this basic zirconosilicate can be used in com-
bination with an acid resin for the conversion of carbohy-
drates to 5-HMF.
22. Fletcher RA, Bibby DM (1987) Clays Clay Miner 35:318–320
23. Cybulski A, Kuster BFM, Marin GB (1991) J Mol Catal
68:87–103
24. Tewari YB (1990) Appl Biochem Biotechnol 23:187–203
25. Yang BY, Montgomery R (1996) Carbohydr Res 280:27–45
Acknowledgments The authors thank Shell Global Solutions for
financial support.
123

C. Yue et al.
26. Park SW, Cho SH, Ahn WS, Kim WJ (2011) Micropor Mesopor
Mater 145:200–204
31. Patil SKR, Heltzel J, Lund CRF (2012) Energy Fuels
26:5281–5293
27. Watanabe M, Aizawa Y, Iida T, Nishimura R, Inomata H (2005)
Appl Catal A Gen 295:150–156
32. Qi X, Watanabe M, Aida TM, Smith RL (2009) Green Chem
11:1327–1331
28. Musau RM, Munavu RM (1987) Biomass 13:67–74
29. Ohara M, Takagaki A, Nishimura S, Ebitani K (2010) Appl Catal
A Gen 383:149–155
33. Yang G, Pidko EA, Hensen EJM (2012) J Catal 295:122–132
34. De S, Dutta S, Saha B (2011) Green Chem 13:2859–2868
35. Roman-Leshkov Y, Dumesic JA (2009) Top Catal 52:297–303
30. Choudhary V, Bumett RI, Vlachos DG, Sandler SI (2012) J Phys
Chem C 116:5116–5120
123