Comprehensive Understanding of the Role of Br?nsted and Lewis Acid Sites in Glucose Conversion into 5-Hydromethylfurfural

Article abstract of DOI:10.1002/cctc.201601203

The conversion of glucose and selectivity into 5-hydromethylfurfural (HMF) were investigated over various silica–alumina composite (AlSiO) catalysts. The type, amount, and strength of the acidic sites were characterized by using NH₃ temperature-programmed desorption and FTIR spectroscopy and then correlated to the catalytic conversion of glucose into HMF to provide a quantitative relationship between the acidity and product selectivity. Lewis acid sites played an important role in glucose conversion, which can enhance the isomerization of glucose to fructose, whereas Br?nsted acid sites had a detrimental effect. HMF selectivity had an almost linear relationship with the weak/total Lewis acid ratio (L*/L), indicating that weak Lewis acids could promote formation of HMF. The medium-to-strong Lewis acid sites can enhance the formation of undesired byproducts (levulinic acid, humins). The Br?nsted to Lewis acid ratio (B/L) had an influence on the HMF selectivity; at similar L*/L ratios, volcano curves were obtained with the increase of the B/L ratio, but the influence was not as great as that of the L*/L ratio. Nb-doped AlSiO catalysts were prepared and used in the conversion of glucose into HMF, which also confirmed the above findings. Under the optimized conditions, the HMF selectivity can reach 71 % at 92.6 % conversion of glucose with no clear decline after four catalytic cycles.

Full text of DOI:10.1002/cctc.201601203

Accepted Article
Title: Comprehensive understanding the role of Brønsted and Lewis
acid sites in glucose conversion to 5-hydromethylfurfural
Authors: Yanqin Wang, Xiangcheng Li, Kaihao Peng, Xiaohui Liu, and
Qineng Xia
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201601203
Link to VoR: http://dx.doi.org/10.1002/cctc.201601203
A Journal of
www.chemcatchem.org

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Comprehensive understanding the role of Brønsted and Lewis
acid sites in glucose conversion to 5-hydromethylfurfural
Xiangcheng Li, Kaihao Peng, Xiaohui Liu*, Qineng Xia, Yanqin Wang*[a]
Abstract: In this paper, the conversion of glucose and the selectivity
to 5-hydromethylfurfural (HMF) are investigated over various silica-
alumina composite (AlSiO) catalysts. The type, amount and strength
of the acidic sites were characterized by NH3-TPD and Py-FTIR
techniques, and then correlated with the catalytic conversion of
glucose to HMF for providing a quantitative relationship between the
acidity and products selectivity. The results showed that Lewis acid
sites played important role in glucose conversion, which can
enhance the isomerization of glucose to fructose, while Brønsted
acid sites had detrimental effect on glucose conversion. Very
importantly, it is shown that the HMF selectivity had a nearly linear
relationship with the weak/total Lewis acid (L*/L) ratio, indicating that
the weak Lewis acid could promote the formation of HMF. Further
analysis showed that the medium to strong Lewis acid sites can
enhance the formation of the undesired by-products (levulinic acid,
humins). It is also found that the B/L ratio had an influence on HMF
selectivity, at every similar L*/L ratio, volcano curves were obtained
with the increase of B/L ratio, but the change was not as great as the
influence of L*/L ratio. Furthermore, niobium incorporated Nb-AlSiO
catalysts were prepared and used in the conversion of glucose to
HMF, which also confirmed the above findings. Under optimized
conditions, HMF selectivity can reach as high as 71% at 92.6%
conversion of glucose and no obvious decline after four catalytic
cycles.
dehydrated to HMF. During these processes, side reactions
always accompany, such as the further hydration of HMF to
levulinic acid and the polymerization of HMF itself or with
fructose to humins.[5] These reaction pathways are drawn in
scheme 1. The unwanted reactions have been identified as the
main obstacle for the production of HMF commercially from
glucose, so the selective conversion of glucose to HMF
represents currently one of the key challenges in biorefining
processes. Understanding the reaction of glucose conversion
and controlling the side reactions is a big challenge. To
overcome this challenge, many groups have explored the
mechanism for the isomerization of glucose and dehydration of
fructose to HMF.[4,6-10]
Vlachos et al.[6] investigated the mechanisms of glucose
isomerization in aqueous media over various Lewis acids,
metal(III) chlorides, and found that the metal chlorides
isomerized glucose to fructose via a C2−C1 intramolecular
hydride transfer. Recently, elucidating how the surface acidities
of solid acid catalysts work on the Lewis acid-catalyzed glucose
isomerization reaction has received much attention. A key
development in glucose isomerization by heterogeneous
catalytic systems was made by Davis and co-workers, it was
well established that the Sn-beta was an effective Lewis acid
catalyst for the glucose isomerization to fructose.[4] Furthermore,
Bell et al.[7] investigated the isomerization mechanism consisting
of a sequence of ring-opening, isomerization, and ring-closing
processes over metal substituted beta zeolite and it was found
the Sn and Zr which acted as Lewis acid sites showed the low
reaction barrier for glucose isomerization.
Introduction
Ordomsky et al.[8] found high selectivity to HMF (55–60%) was
achieved over silylated phosphates, due to a decrease in the
amount of Lewis acidity which would lead to unselective glucose
transformation into humins, and the selectivity to HMF increased
with the increase of B/L ratio (ca. 0.09–0.27). But this result was
conflicted with our previous work[10] that the selectivity to HMF
decreased with the increase of B/L ratio (ca. 0.72–1.85) over
NbPO catalysts. Then, a recent study showed that HMF
formation rate exhibited volcano type curve vs. the B/L ratio, and
the maximum HMF yield approached 60% was achieved in
biphasic phase.[9] The above reports indicated that the Lewis
acid sites played important role in the isomerization of glucose,
and the HMF selectivity had correlation with B/L ratio; but at best
of our knowledge, no clear correlations between the types and
strength of acid sites with products selectivity had been
reported.
An unprecedented international effort has been made to explore
alternative renewable lignocellulosic biomass for the large-scale
production of fuels and chemicals due to the rapid depletion of
fossil fuels. 5-Hydroxymethylfurfural (HMF), derived from C6
carbohydrates, is one of the most versatile and important
building blocks, because it can be upgraded into a large number
of chemicals.[1]
HMF is formed by the loss of three water molecules from a
hexose by acid catalysis. It can be easily obtained from fructose
by using a wide variety of homogeneous and heterogeneous
catalysts,[2] but fructose is produced from glucose in industry
catalyzed by enzyme, making glucose
a more desirable
substrate. Glucose can be converted to HMF via a tandem
steps: first glucose is isomerized to fructose in the presence of
an enzyme, a Lewis acid or base catalyst,[3,4] then fructose is
It was shown recently that mesoporous silica-alumina
composite (AlSiO) could catalyze the glucose conversion to
HMF efficiently.[11] Here, the conversion of glucose and the
selectivity to HMF are well correlated with the acidic properties
of various silica-alumina composite (AlSiO) catalysts, which
offered a possibility for deeply understanding the nature of the
active sites on glucose isomerization, dehydration reactions and
[a]
X.C. Li, K. H. Peng, Dr. X.H. Liu, Dr. Q.N. Xia, Prof. Y. Q. Wang
Shanghai Key Laboratory of Functional Materials Chemistry and
Research Insitute of Industrial Catalysis, School of Chemistry and
Molecular Engineering
East China University of Science and Technology
No. 130 Meilong Road, Shanghai (PR China)
Fax: (+86)21-64253824
E-mail: xhliu@ecust.edu.cn; wangyanqin@ecust.edu.cn
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Scheme 1. The reaction pathway for the conversion of glucose to HMF.
further for the design of catalysts. Moreover, we continue our
study to investigate the acidic properties of AlSiO by modyfing
with niobium and further verify the findings on glucose
conversion.
resusts are plotted in Figure 2. The AlSiO samples (except for
Na-AlSiO-20 sample) and NbPO samples exhibited the
ammonia desorption at two tempertures: a low-temperture peak
at 180 C and
a high-temperture peak at 300-500 C,
corresponding to the weak and medium to strong acid sites. For
Na-AlSiO-20, the peak is rather narrow and concentrated at 180
C, indicating its weak acidity. The total amounts of the acid
sites determined by NH3-TPD increases in the following order:
Na-AlSiO-20 < AlSiO-20 < AlSiO-20/B < AlSiO-20/L < AlSiO-10,
and the total acid sites of NbPO-2 is close to that of Al-NbPO-2.
Results and Discussion
Characterization of texture properties and surface acidity of
AlSiO and NbPO catalysts
A series of AlSiO catalysts were prepared by a sol-gel method
as discribed before[11] and modified with post-treatments
(silylation, ionic-exchange). The NbPO-pH2 and Al-NbPO-pH2
catalysts were also used for expanding the scope of acidity. The
X-ray powder diffraction patterns of AlSiO-20, AlSiO-10, AlSiO-
20/B, AlSiO-20/L, Na-AlSiO-20, NbPO-pH2 and Al-NbPO-pH2
catalysts are shown in Figure 1. It can be seen that all the AlSiO
samples are amorphous, as only one broad band is detected
from 2θ = 15° to 40° for each sample, which implies the
amorphous characteristic of these materials. NbPO-2 and Al-
NbPO-2 samples display two broad peaks in the 2θ range of 15-
40° and 40-60°, indicating the amorphous characteristic and no
crystallized NbPO or Nb2O5 phase are detected.
(a)
30
AlSiO-10
AlSiO-20
20
AlSiO-20/B
10
AlSiO-20/L
Na-AlSiO-20
0
The acidic properties of different catalysts were investigated
by temperature-programmed desorption of ammonia and the
100
200
300
400
500
600
T/oC
(b)
20
AlSiO-10
AlSiO-20
NbPO-pH2
Al-NbPO-pH2
AlSiO-20/B
10
0
AlSiO-20/L
Na-AlSiO-20
NbPO-pH2
Al-NbPO-pH2
100
200
300
400
500
600
T/oC
10
20
30
40
50
60
70
80
2degree
Figure 2. NH3-TPD profiles of various catalysts. (a) AlSiO and (b) NbPO
samples. (The TCD signal of AlSiO-20/L, AlSiO-20/B, AlSiO-20, AlSiO-10
shifted up 5, 10, 15, 20, respectively).
Figure 1. XRD patterns of various catalysts.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
To investigate the different types and strength of acidic sites
on various catalysts, the Py-FTIR spectra were recorded at
different temperatures and drawn in Figure S1. The intense
bands at 1448 cm-1 and 1540 cm-1 are due to Lewis and
Brønsted acid sites, respectively, and the band at 1490 cm-1 is
attributed to the adsorption of pyridine on both Brønsted and
Lewis acid sites.[12] The weak sites are defined as the ones from
which pyridine is removed by evacuation at between 100 C and
200 C, the medium strength corresponds to evacuation
between 200 C and 400 C and the strong sites remains
adsorbing pyridine after evacuation at 400 C.
Table 1 summarized the concentration of the Lewis and
Brønsted acid sites calculated on the basis of the extinction
coefficients from bands at 1448 cm-1 and 1540 cm-1, respectively,
at different temperature. The intensity of the band assigned to
pyridine adsorbed on Brønsted acid sites increases in the
following order: Na-AlSiO-20 < AlSiO-20 < AlSiO-20/L < AlSiO-
10 ≈ AlSiO-20/B < Al-NbPO-pH2 < NbPO-pH2. In order to
reduce Brønsted acid sites, AlSiO-20 catalyst was treated with
NaOH and resulted in a significant decrease in the amount of
Brønsted acid sites from 106 to 40 μmol g-1 (Na-AlSiO-20
catalyst, Table 1). While treatment of the AlSiO-20 with TEOS
(AlSiO-20/B) leads to a significant increase in the amount of
Brønsted acid sites and decrease in the Lewis acid sites, this
procedure is well known for deactivation of acid sites on the
external surface of zeolites.[13,14] Lewis acid sites were located at
the external surface, thus, having a higher accessibility for
TEOS compared to the bridging hydroxyl groups,[14] the
introducing of silanol groups and bridging hydroxyl groups which
came from new Si-O-Al bonds formed leaded to the increase of
Brønsted acid sites on the surface (Scheme 2).
Scheme 2. Possible mechanism of deactivation Brønsted and Lewis acid sites
formation by silylation over silica-alumina composite.
The intensity of the band assigned to pyridine adsorbed on
Lewis acid sites increases in the following order: AlSiO-20/B <
Na-AlSiO-20 ≈ NbPO-pH2 < AlSiO-20 ≈ Al-NbPO-pH2 <
AlSiO-20/L < AlSiO-10. The AlSiO-20/L catalyst was obtained by
calcining AlSiO-20/B catalyst at 550 C, resulting in the increase
of Lewis acid sites (550 μmol g-1) with a low effect on the
Brønsted acid sites (Table 1), a possible explanation is the loss
of bridging hydroxyl groups and silanol groups after calcination
(Scheme 2), then more three-coordinated Al atoms formed
which acted as Lewis acid sites by accepting the lone pair of
electrons. Doping aluminium in NbPO-pH2 can enhance Lewis
acid because aluminium is always considered as a strong Lewis
acid site. As shown in Table 1 and Figure S1, the Lewis acid
sites increased from 346 to 469 μmol g-1, which indicates that
aluminum is indeed a good strong Lewis acid center. In
summarize, we have prepared several catalysts which contain
various strength of acid with different B/L acid ratio, following the
glucose dehydration over these catalysts are investigated and
discussed in the next section.
Table 1. Surface area and amounts of Brønsted and Lewis acid sites on various catalysts by Py-FTIR.
Acid from Py-FTIR (μmol g-1)
Lewis
SBET
Brønsted
Entry
Cata
/m2g-1
Total
B/L
B*/B
L*/L
W
M
S
Total
W
M
S
Total
1
2
3
4
5
AlSiO-20
AlSiO-10
717.6
722.9
360.4
662.0
360.4
12
102
81
14
129
132
83
80
66
106
297
337
232
40
151
172
125
44
185
362
67
81
104
75
417
638
267
550
316
523
935
604
782
357
0.26
0.47
1.27
0.30
0.13
0.11
0.34
0.24
0.16
0.48
0.35
0.27
0.47
0.08
0.57
AlSiO-20/B
AlSiO-20/L
Na-AlSiO-20
124
112
11
37
385
41
121
95
19
10
180
6
7
NbPO-pH2
Al-NbPO-pH2
AlSiO-10
290.1
77
122
102
81
241
136
129
132
136
323
234
535
124
414
641
492
766
337
672
69
32
242
119
362
67
35
346
469
638
501
469
986
962
1.86
1.05
1.20
0.67
1.43
0.12
0.25
0.13
0.24
0.18
0.20
0.07
0.27
0.25
0.07
50.4
318
104
309
318
8[a]
9[b]
10[c]
-
-
-
172
125
32
1404
838
AlSiO-20/B
Al-NbPO-pH2
122
119
1141
W: weak acid sites; M: medium acid sites; S: strong acid sites;
B*/B: weak / total Brønsted acid sites;
L*/L: weak / total Lewis acid sites;
0.2g catalyst added: [a] 2 ml 0.05 mol/L HCl; [b] 30 mg yttrium(III) triflate; [c] 2 ml 0.02 mol/L HCl.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Effect of the acidity on the conversion of glucose
(Table 2, Entry 2,3), this might due to the more Lewis acid sites
of AlSiO-10 sample, which would promote glucose conversion.
The conversion of glucose and the selectivity to HMF, levulinic
acid and humins during dehydration of glucose over various
catalysts are investigated and all data are collected in Table 2.
Besides the main products were detected in the reaction system
by HPLC. Fructose was not detected and the selectivity to
furfural was below 1mol %. Humins could not be observed by
HPLC or GC-MS, therefore the amount of humins produced was
estimated from the carbon balance.
The conversion of glucose over various catalysts increased in
the order: AlSiO-20/B < Na-AlSiO-20< NbPO-pH2 < AlSiO-20 <
Al-NbPO-pH2 < AlSiO-20/L < AlSiO-10, this order did not
correspond to the increase of the total amounts of acid sites. As
shown in Figure S2, the glucose conversion exhibited a volcano
curve with the total amounts of acid sites, it initially increased
with the total acidity, then decreased slightly when the total
acidity was larger than 935 μmolg-1.
Further detailed analysis showed that glucose conversion also
exhibited a volcano relationship with the Brønsted acid sites
(Figure S2), suggesting that more Brønsted acid sites might
suppress glucose conversion, which agreeed with our previous
results over NbOPO4 catalysts.[10] To further verify this, AlSiO-
10 sample with additional Brønsted acid, such as hydrochloric
acid was used to evaluate this reaction under the same
conditions (Table 2, Entry 8) and found that it really decreased
the conversion of glucose from 98.9% to 90.9%, which
confirmed that Brønsted acid sites had detrimental effect on the
glucose isomerization process. Then for AlSiO-20/B sample,
which had similar Brønsted acid sites (337 μmol g-1) to that of
AlSiO-10 sample (297 μmol g-1), but lower glucose conversion
(68.6%) was obtained than that (98.8%) over AlSiO-10 sample
Further analysis showed a positive correlation between
glucose conversion and Lewis acid sites (Figure 4), meaning
that Lewis acid sites really promoted glucose conversion since
glucose isomerization was a Lewis acid catalyzed reaction.[4,10]
This was further confirmed by using AlSiO-20/B sample with
additional Lewis acid, Yttrium trifluoroacetate [(TfO)3Y] to
evaluate this reaction under the same conditions (Table 2, Entry
11), it was indeed found that glucose conversion increased from
68.6% to 92.4%. The above findings showed that Lewis acid
sites played important role in the glucose conversion via
isomerization of glucose into fructose, whereas Brønsted acid
sites had detrimental effect on this isomerization process.
100
90
80
70
400
600
Lewis acid, μmol/g
Figure 4. Effect of the Lewis acid sites on the conversion of glucose over
various catalysts under the same reaction condition.
Table 2. Summary of the glucose conversion, HMF yield and selectivity, LA yield and selectivity and humins selectivity catalyzed
by various catalysts.
Entry
Catalyst
Cglucose
/%
YHMF
/%
YLA
/%
SHMF
/%
SLA
/%
Shumins
/%
1
AlSiO-20
91.7
63.1
1.1
68.8
1.2
30.0
2
3
AlSiO-10
AlSiO-20/B
AlSiO-20/L
Na-AlSiO-20
NbPO-pH2
Al-NbPO-pH2
AlSiO-10
98.8
68.6
96.3
70.6
88.2
95.9
90.9
88.5
90.2
92.4
91.8
47.9
46.0
28.9
48.5
21.5
38.2
41.7
58.6
60.5
47.0
33.4
5.6
-
48.5
67.1
30.0
68.7
44.6
39.8
45.9
66.2
67.1
50.9
36.4
5.7
-
45.8
32.9
47.3
31.3
50.2
47.6
52.1
33.8
32.9
44.3
50.7
4
21.9
-
22.7
-
5
6
4.6
12.1
1.8
-
5.2
12.6
2.0
-
7
8[a]
9[b]
10[b]
11[c]
12[d]
AlSiO-20/B
Na-AlSiO-20
AlSiO-20/B
Al-NbPO-pH2
-
-
4.4
11.8
4.8
12.9
Reaction conditions: catalyst (0.2 g), glucose (0.2 g), water saturated with NaCl (2 ml), THF (6 ml), 160 °C, 1.5h;
[a] 2 ml 0.05 mol/L HCl instead of water; [b] reaction time: 2 h; [c] add 30 mg yttrium(III) triflate; [d] 2 ml 0.02 mol/L HCl instead of water.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Effect of the acidity on the selectivity to HMF, levulinic acid
and humins
where the a low B/L acid molar ratio (ca. 0.09–0.27) would
promote the formation of HMF from glucose and inhibit the
formation of humins, but also in agreement with the Zhang’s
work,[10] where a high B/L acid molar ratio (ca. 0.72–1.85) would
decrease the selectivity to HMF. At the L*/L ratio of 0.1, the
maxium HMF selectivity (39.8%) was obtained with the B/L ratio
of 1.05; as the L*/L ratio increased to 0.2-0.3, the optimum B/L
ratio decreased to 0.67 for achieving the maxium HMF
selectivity (50.9%). This was because as L*/L ratio decreased,
more medium to strong Lewis acids accelerated the glucose
isomerization,[9] thus high B/L ratio was desirable to speed up
dehydration for achieving the maximum HMF selectivity.
While detailed analysis for by-products showed that the
selectivity toward levulinic acid (LA) production from glucose
decreased with the increase of L*/L ratio, as shown in Figure 6 a,
and the LA selectivity exhibited volcano type curve vs B/L ratio
in Figure 6 b. These results indicated the medium to strong
Lewis acid sites was favourable for the selectivity to LA, which
was consistent with the previous studies[5,10] that medium to
strong Lewis acid sites could drive Brønsted acid-catalyzed
rehydration of HMF to formic and levulinic acids.
Figure 6 c shows the relationship of humins selectivity to L*/L
ratio and it can be found that the high L*/L ratio leads to less
formation of humins, although there was no clear quantitative
correlation between the amount of Lewis acid sites and humins
selectivity. Therefore the selectivity to humins was plotted
against weak and medium to strong Lewis acidity and shown in
Figure 6 d. It was surprisingly found that the medium to strong
Lewis acidity nearly had a linear correlation with the humins
selectivity under the similar B/L acid ratio (Figure S4) except at
high L*/L ratio, indicating that the medium to strong Lewis acid
sites were mainly responsible for the the transformation of
sugars and HMF into humins via cross-condensation.[13]
Moreover, high B/L ratio would promoted the formation of
humins, revealing that the humins also generated over Brønsted
acid sites, and this observation might be explained that Brønsted
acid sites caused aldol addition and condensation reactions from
HMF via 2,5-dioxo-6-hydroxy-hexanal(DHH).[15]
In order to address the relationship between the selectivity to
HMF and by-products with the acidity of catalysts, all reactions
were compared at the similar conversion level (Table 2).
Detailed analysis showed that neither Lewis nor Brønsted acid
sites had correlation with HMF selectivity (Figure S3), but it is
very interesting to note that the weak Lewis acid sites had
positive correlation with HMF selectivity; on the contrary,
medium to strong Lewis acid sites had negative correlation with
HMF selectivity. Therefore, in order to better understand the
variation of HMF selectivity with weak Lewis acidity, HMF
selectivity was plotted against the ratio of weak Lewis acidity to
total Lewis acidity, a nearly linear relationship between the
weak/total Lewis acid (L*/L) ratio and the HMF selectivity was
observed except at high L*/L ratio, as present in Figure 5 a.
Minimum HMF selectivity (30.0%) was achieved at the L*/L ratio
of 0.08 (Table 2, Entry 4), this was because the medium to
strong Lewis acid sites could promote the formation of levulinic
acid by rehydration of HMF and condensation of sugars, HMF
into humins.[13] Whereas high HMF selectivity (66.2%) was
achieved over AlSiO-20/B catalyst for its high L*/L ratio of 0.47
(Table 2, Entry 9). In order to verify the low L*/L ratio led to the
decrease of HMF selectivity, a strong Lewis acid, Yttrium
trifluoroacetate [(TfO)3Y] was added, which decreased L*/L ratio
from 0.47 to 0.25 and a significant decrease in the HMF
selectivity was observed (Table 2, Entry 11), these results
implied that the L*/L ratio played important role in HMF
selectivity.
Previous study
[8-10]
also indicated that B/L ratio had effect on
the HMF selectivity, thus, the B/L ratio with various L*/L ratio
was calculated and the correlation was present in Figure 5 b. At
the similar L*/L ratio, the HMF selectivity exhibited volcano type
curve vs B/L ratio, it meaned that there was a balance of Lewis
and Brønsted acids for the glucose isomerization to fructose and
fructose dehydration to HMF, not the higher B/L ratio, a better
selectivity to HMF. Moreover, this surprising result was not only
consistent with the studies reported by Ordomsky et al.,[8,13]
(b)
(a)
L*/L≈0.4-0.6
L*/L≈0.2-0.3
70
70
L*/L≈0.10
60
50
40
30
60
50
40
30
0.0
0.5
1.0
1.5
2.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
B/ L acid ratio
L*/L ratio
Figure 5. Effect of the L*/L (a) and B/L (b) ratio on the selectivity to HMF. (L*/L: weak/total Lewis acid ratio)
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
(a)
(b)
L*/L≈0.4-0.6
L*/L≈0.2-0.3
L*/L≈0.10
20
20
10
0
10
0
0.0
0.5
1.0
1.5
2.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
B/ L acid ratio
L*/L ratio
(c)
(d)
B/L≈1.2-1.9
B/L≈0.1-0.7
50
40
30
50
40
30
100
200
300
400
500
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Lewis(M+S) acid, μmol/g
L*/L ratio
Figure 6. Effect of the L*/L (a) and B/L ratio (b) on the selectivity to LA; Effect of the L*/L (c) and medium-strong Lewis acid sites (d) on the selectivity to humins.
(L*/L: weak/total Lewis acid ratio; M+S: medium+strong Lewis acid sites)
To summarize, HMF selectivity nearly had a linear correlation
with the L*/L ratio, but exhibited a volcano type curve vs B/L
ratio. In another hand, the medium to strong Lewis acid sites
were favourable for the formation of LA and humins, thereby
reducing the HMF selectivity. Similar conclusion was also drawn
in lower conversions and the results are shown in Table S1 and
Figure S5-7. In order to verify the above findings, NbOx sites
were added to the silica-alumina composite AlSiO for acidic
adjustment because of the hydrothermal stibity of Nb2O5,[16] then
used in conversion of glucose to HMF.
weak, medium to strong acid sites. The total amount of the acid
sites determined by NH3-TPD increases in the following order:
Nb0.15Al0.85Si25O < Nb0.30Al0.70Si25O < Nb0.40Al0.60Si25O, revealing
the increase of the Nb/Al ratio lead to the increase of the total
acid sites.
To investigate the types and strength of acidic sites, the Py-
FTIR spectra were recorded in Figure S11. The amounts of
Brønsted and Lewis acid sites on various mesoporous Nb-AlSiO
catalysts were quantified and summarized in Table 4. Both the
amounts of Brønsted and Lewis acid sites increased with
increasing Nb/Al ratio, especially for Brønsted acid sites, the B/L
acid ratio increased from 0.68 to 0.87, while the L*/L acid ratio
increased from 0.40 to 0.49. The Nb0.40Al0.60Si25O sample had
the larger acidity (718 μmol g-1) than other samples, which was
in agreement with the results of NH3-TPD (Figure S10).
Niobium doped AlSiO for high selectivity to HMF with acidic
adjustment.
The X-ray powder diffraction patterns of various Nb-AlSiO
materials indicated that all the Nb-AlSiO samples were
amorphous, as shown in Figure S8, and no crystalline Nb2O5
phase was detected in Nb-AlSiO samples, indicating that the
amorphous niobium oxide species were well dispersed. All Nb-
AlSiO materials were mesoporous from Nitrogen adsorption–
desorption isotherms (Figure S9), and the respresentative
sample of Nb0.15Al0.85Si25O had the surface area of 766.2 m2g-1
and pore volume of 0.50 cm3g-1 (Table S2). To study the amount
of the acid sites of Nb-AlSiO materials, the NH3-TPD was
measured and shown in Figure S10. All samples had broad
peaks in the range of 100-500 C, indicating the presence of
The Nb-AlSiO catalysts had similar L*/L acid ratio with that of
AlSiO-20 catalyst, and the B/L ratio increased in the following
order: AlSiO-20
<
Nb0.15Al0.85Si25O
<
Nb0.30Al0.70Si25O
<
Nb0.40Al0.60Si25O, then the HMF selectivity decreased in the
following order: Nb0.15Al0.85Si25O > AlSiO-20 > Nb0.30Al0.70Si25O >
Nb0.40Al0.60Si25O under the similar glucose conversion (Table 2
and Table 5). This result can be explained from Figure 5 b, the
B/L ratio (0.68) of Nb0.15Al0.85Si25O is more suitable than others
in terms of the selectivity to HMF, and higher B/L ratio (0.87) of
Nb0.40Al0.60Si25O may lead to more unselective glucose, fructose
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Table 4. The amounts of Brønsted and Lewis acid sites on various catalysts by Py-FTIR.
Acid from Py-FTIR (μmol g-1)
Entry
Cata
Brønsted
S
Lewis
Total
B/L
B*/B
L*/L
W
M
Total
W
M
S
Total
1
2
3
Nb0.15Al0.85Si25
Nb0.30Al0.70Si25
Nb0.40Al0.60Si25
O
O
O
28
39
70
63
79
92
98
189
232
334
112
146
188
93
80
89
74
92
279
318
383
468
550
717
0.68
0.73
0.87
0.15
0.17
0.21
0.40
0.46
0.49
114
172
106
W: weak acid sites; M: medium acid sites; S: strong acid sites;
B*/B: weak /total Brønsted acid sites;
L*/L: weak /total Lewis acid sites;
Table 5. Summary of the glucose conversion, HMF yield and selectivity, LA yield and selectivity and humins selectivity catalyzed by various catalysts
Cglucose
/%
YHMF
/%
YLA
/%
SHMF
/%
SLA
/%
Shumins
/%
Entry
Catalyst
1
2
Nb0.15Al0.85Si25
O
O
80.8
82.2
57.8
55.9
0.3
0.7
71.5
68.0
0.4
0.9
28.1
31.1
Nb0.30Al0.70Si25
3
Nb0.40Al0.60Si25
Nb0.15Al0.85Si25
Nb0.30Al0.70Si25
Nb0.40Al0.60Si25
O
O
O
O
87.7
92.6
95.4
97.7
51.4
65.7
65.3
56.0
0.7
0.7
1.0
1.3
58.6
71.0
68.5
57.3
0.8
0.8
1.0
1.3
40.6
28.2
30.5
41.4
4[a]
5[a]
6[a]
Reaction conditions: catalyst (0.2 g), glucose (0.2 g), water saturated with NaCl (2 ml), THF (6 ml), 160 °C, 1.5 h;
[a] reaction time: 2 h.
or HMF condensation into humins, resulting in a low HMF
selectivity.[15]
Conv.Glu
Yield.HMF
Selec.HMF
80
60
40
20
0
The results obtained over the Nb-AlSiO samples well
confirmed above findings, the excellent performance of the solid
acid catalysts with high selectivity to HMF requires the higher
L*/L ratio and suitable B/L ratio. The recyclability and stability of
Nb0.15Al0.85Si25O catalyst was investigated by reusing the catalyst
in consecutive catalytic runs and the results are presented in
Figure 7. It can be seen that the glucose conversion and HMF
yield were still maintained at 61.4% and 41.3%, respectively,
indicating its good stability.
1
2
3
4
5
Conclusions
Run numbers
Figure 7. Cycle usage of Nb0.15Al0.85Si25O catalyst. (Reaction conditions: 0.2 g
Nb0.15Al0.85Si25O catalyst, 0.2 g glucose, 2 ml water saturated with NaCl, 6 ml
THF, 160 °C, 1.5 h.)
The selective formation of HMF from glucose has been studied
using silica-alumina composite (AlSiO) catalysts and the effect
of acidity on the catalytic activity and selectivity were
systematically investigated. The results shows that Lewis acid
sites promotes glucose conversion, whereas Brønsted acid sites
not only has detrimental effect on glucose conversion, but also
promotes aldol addition and condensation reaction from HMF;
HMF selectivity nearly had a linear correlation with the L*/L ratio,
and exhibited a volcano type curve vs B/L ratio; weak Lewis acid
sites promotes HMF production, while the medium to strong
Lewis acid sites is responsible for the undesired by-products (LA,
humins) formation. Then, niobium-doped mesoporous AlSiO as
a catalyst, showed good recycle performance and well verified
the above findings for its excellent performances on catalytic
activity and selectivity, and highest HMF selectivity (71%) was
achieved over Nb0.15Al0.85Si25O catalyst. This work can help to
design acid catalysts with suitable amount and strength to
achieve the ideal goal for producing HMF from glucose or other
carbohydrates.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
thermal conductivity detector (TCD) at 110 C. Besides, a blank
measurement was conducted without adsorption of ammonia
and after deducted the blank measurement, the NH3-TPD curve
of catalysts were obtained.
Experimental Section
Materials
5-Hydroxymethylfurfural (HMF) were purchased from Alfa Aesar
Chemical Reagent Company. All chemicals were purchased
from Sinopharm Chemical Reagent Co. Ltd. Nb tartrate was
prepared in our laboratory according to the literature.[10]
Tetrahydrofuran (THF) was chemical pure solvent and
purchased from shanghai chemicals company. All bought
chemicals were of analytical grade and used without further
purification.
Infrared (IR) spectra of pyridine adsorption were recorded on
Nicolet NEXUS 670 FT-IR spectrometer. The samples were
pressed into self-supporting disks and placed in an IR cell
attached to a closed glass-circulation system. The disk was
dehydrated by heating at 400 C, for 1 h under vacuum in order
to remove physic-sorbed water. After the cell was cooled to
room temperature, the IR spectrum was recorded as
background. Pyridine vapor was then introduced into the cell at
room-temperature until equilibrium was reached, and then a
second spectrum was recorded. Subsequent evacuations were
performed at 100、200 and 400 C, for 10 min followed by
spectral acquisitions. The spectra presented were obtained by
subtracting the spectra recorded before and after pyridine
adsorption.
Catalyst preparation
AlSiO-10, AlSiO-20 samples was synthesized according to the
literature.[11] The preparion of Nb-AlSiO was similar to that of
AlSiO except extra addition of niobium tartrate before the
addition of TEOS.
The quantification of acid sites was performed using the same
AlSiO-20/B was obtained by using TEOS as the silylation
agent. 1g of dried AlSiO-20 catalyst was added to a solution of 1
g of TEOS in 25 mL of n-hexane. The mixture was stirred at
room temperature for 12 h. AlSiO-20/L prepared by the above
and calcination (823K) for further 4 h. Na-AlSiO-20 sample was
obtained by ion-exchange of AlSiO-20 sample with NaCl. In
details, 3.0 g of AlSiO-20 sample was stirred in 300 mL of 0.2 M
NaCl solution, and the mixture was maintained at pH = 5.5 by
adding 0.05 M NaOH solution. After 24 h, the collected sample
was washed repeatedly with distilled water until Na+ and Cl- ions
were no longer detected, and then dried at 373 K for 24 h.
Mesoporous niobium phosphate catalysts NbPO and Al-NbPO
were prepared according to our previous work.[10]
expressions as those in the research article of Rocha et al.: [12]
( )
and
( )
CL and CB are the concentrations of Lewis acid sites and
Brønsted acid sites in μmol g-1;
and
are the
integrated areas of bands at 1450 and 1540 cm-1 in the original
data of FTIR spectra, as shown in the Results and discussion
section. and are molar extinction constants for Lewis and
Characterization
Brønsted acid sites; and are integration molar
extinction coefficients, 2.22 and 1.67 cm μmol-1 for Lewis and
Brønsted acids, respectively; r is the wafer radius in cm and w is
the wafer weight in g of the self-supporting catalyst disk.
Powder XRD patterns were recorded on a Bruker diffractometer
(D8 Focus) by using Cu Kα (λ=0.154056 nm) radiation. Nitrogen
adsorption–desorption isotherms were measured at -96 C on a
NOVA 4200e analyzer (Quantachrome Co. Ltd). Before the
measurements, all samples were outgassed at 180 C for 12 h
under vacuum to remove moisture and volatile impurities. The
BET method was used to calculate the specific surface area.
Pore size distribution curves were derived from the desorption
branches of the isotherms and calculated by the Barrett–Joyner–
Halanda (BJH) method. The total pore volume (Vt) was
estimated at a relative pressure of 0.975.
Catalytic reactions
The influence of external diffusion can be totally eliminated by
changing stirring speed. It is found that stirring speed had no
influence on the catalytic performance if stirring speed was
above 300 rpm. The influence of the internal diffusion can be
eliminated by decreasing the particle size of catalyst, the
catalyst size had no effect on the conversion of glucose if the
size was lower than 150 µm.[11] All the dehydration reaction
experiments were conducted in a Teflon-lined stainless steel
Ammonia temperature-programmed-desorption (NH3-TPD) was
carried out in
a PX200 apparatus (Tianjin Golden Eagle
Technology Limited Corporation) with a thermal conductivity
detector (TCD). The catalyst (100 mg) was charged into the
quartz reactor, and the temperature was increased from room
temperature to 600 C at a rate of 10 C min-1 under a flow of N2
(40 ml min-1), and then the temperature was decreased to 90 C.
Finally, NH3 was injected into the reactor at 90 C under a flow of
N2 (40 ml min-1). When the adsorption saturation was reached,
followed by a flow of N2 for 1h at 90 C, then the temperature
was increased from 90 C to 600 C at a rate of 10 C min-1 and
the amount of desorbed ammonia was detected by using
autoclave (30 ml) equipped with
a temperature-controlled
heating-jacket and a magnetic stirrer, then the mixture was
stirred at 700 rpm to guarantee that it is properly stirred. The
mixture of carbohydrate, solvent (6 mL organic phase and 2 mL
H2O saturated with NaCl) and 0.2 g catalyst (the particle size of
catalysts are the same and lower than 150 µm) were put in the
sealed autoclave. 0.5 M Pa N2 gas was used for purging air
outside the reactor and keeping solvent in liquid phase. When
the reactor was raised to the desired temperature, zero time was
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Rataboul, D. Cardoso, N. Essayem. Catalysis Today. 2012, 195, 114–119.
[4] Y. Román-Leshkov, M. Moliner, J. A. Labinger, and M. E. Davis, Angew.
Chem. Int. Ed. 2010, 49, 8954-8957.
recorded. Then, the reactor was held at this temperature for a
given period of time. After reaction, the autoclave was cooled to
room temperature by loading cooling water. The solid catalyst
was removed by filtration, and the filtrate was taken for analysis.
During the recycling, the upper organic phase was removed
after the reaction and the catalyst kept in the aqueous phase
was used directly in the next cycle with separation, washing and
drying.
[5] a) T. M. C. Hoang, E. R. H. van Eck, W. P. Bula, J. G. E. Gardeniers, L.
Lefferts and K. Seshan. Green Chem. 2015, 17, 959-972; b) B. Saha and
M. M. Abu-Omar. Green Chem. 2014, 16, 24–38.
[6] H. Nguyen, V. Nikolakis, D. G. Vlachos. ACS Catal. 2016, 6, 1497−1504.
[7] Y. P. Li, M. H. Godon, and A. T. Bell, ACS Catal. 2014, 4, 1537−1545.
[8] V. V. Ordomsky, V. L. Sushkevich, J. C. Schouten, J. van der Schaaf, T. A.
Nijhuis, J. Catal. 2013, 300, 37–46.
[9] T. Dallas Swift, H. Nguyen, Z. Erdman, J. S. Kruger, V. Nikolakis, D. G.
Vlachos. J. Catal. 2016, 333, 149–161.
Product analysis
[10] a) Y. Zhang, J. J. Wang, J. W. Ren, X. H. Liu, X. C. Li, Y. J. Xia, G. Z. Lu,
Y. Q, Wang, Catal. Sci. Technol. 2012, 2, 2485-2491; b) Y. Zhang, J. J.
Wang, X. C. Li, Y. J. Xia, B. C. Hu, G. Z. Lu, Y. Q. Wang, Fuel. 2015, 139,
301-307; c) D. Q. Ding, J. X. Xi, J. J. Wang, X. H. Liu, G. Z. Lu, Y. Q.
Wang, Green Chem. 2015, 17, 4037–4044.
The analysis of HMF was carried out by means of an HPLC
apparatus (Agilent 1200 Series) equipped with an XDB-C18
column (Eclipse USA) and quantified with an ultraviolet detector
(Agilent G1314B) at 254 nm based on the external standard.
The eluent with a flow rate of 0.6 mL min-1 was a mixture of
methanol and water with volumetric ratio of 20:80. An auto-
sampler (Agilent G1329A) was used to enhance the
reproducibility. Glucose, fructose, and levulinic acid (LA) were
analyzed with a refractive index detector (Agilent G1362A) and
using a Biorad Aminex HPX-87H sugar column at 328K. The
mobile phase was 0.004 M H2SO4 with a flowing rate of 0.45 ml
min-1.
[11] X. C. Li, Q. N. Xia, V. C. Nguyen, K. H. Peng, X. H. Liu, N. Essayem and
Y. Q. Wang, Catal. Sci. Technol. DOI: 10.1039/c6cy01628f.
[12] A. S. Rocha, A. M. S. Forrester, M. H. C. de la Cruz, C. T. da Silva and E.
R. Lachter, Catal. Commun. 2008, 9, 1959-1965.
[13] a) V. V. Ordomsky, J. V. D. Schaaf, J. C. Schouten, and T. A. Nijhuis,
ChemSusChem. 2013, 6, 1697–1707; b) V. V. Ordomsky, J. V. D.
Schaaf, J. C. Schouten, and T. A. Nijhuis. ChemSusChem. 2012, 5,
1812-1819.
[14] S. Zheng, H. R. Heydenrych, A. Jentys, J. A. Lercher, J. Phys. Chem. B.
2002, 106, 9552–9558.
[15] S. K. R. Patil, C. R. F. Lund, Energy Fuels. 2011, 25, 4745–4755.
[16] K. Nakajima, Y. Baba, R. Noma, M. Kitano, J. N. Kondo,S. Hayashi and
M. Hara, J. Am. Chem. Soc. 2011, 133, 4224-4227.
Conversion of glucose and yield of HMF and levulinic acid
were defined as follows:
moles of glucose reacted
Glucose conversion
moles of starting glucose
moles of HMF produced
HMF yield
moles of starting glucose
moles of levulinic acid produced
Levulinic acid yield
moles of starting glucose
Acknowledgements
This work was supported financially by the NSFC of China (No.
91545103, 21273071 and 21403065) and the Science and
Technology
Commission
of
Shanghai
Municipality
(10dz2220500).
Keywords: glucose conversion; 5-hydromethylfurfural; Lewis
acid sites; B/L ratio
[1] a) Y. Román-Leshkov, C. J. Barrett, Z. Y. Liu and J. A. Dumesic, Nature.
2007, 447, 982-985; b) P. Gallezot. Chem. Soc. Rev. 2012, 41, 1538–
1558; c) C. H. Zhou, C. X. Lin, D. S. Tong and J. beltramini. Chem. Soc.
Rev. 2011, 40, 5588–5617; d) J. J. Bozell and G. R. Petersen, Green
Chem. 2010, 12, 539-554.
[2] a) J. J. Wang, J. W. Ren, X. H. Liu, G. Z. Lu and Y. Q. Wang, AICHE
Journal. 2013, 59, 2558-2566; b) J. J. Wang, W. J. Xu, J. W. Ren, X. H.
Liu, G. Z. Lu and Y. Q. Wang, Green Chem. 2011, 13, 2678-2681; c) P.
Gallezot, Chem. Soc. Rev. 2012, 41, 1538–1558; d) A. Rosatella, S. P.
Simeonov, R. F. M. Frade and C. A. M. Afonso, Green Chem. 2011, 13,
754–793.
[3] a) H. Huang, C. A. Denard, R. Alamillo, A. J. Crisci, Y. Miao, J. A.
Dumesic,S. L. Scott, and H. Zhao, ACS Catal. 2014, 4, 2165-2168; b) M.
Moliner, Y. Román-Leshkov and M. E. Davis, Proc. Natl. Acad. Sci. U. S.
A. 2010, 107, 6164-6168; c) R. O. L. Souza, D. P. Fabiano, C. Feche, F.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201601203
ChemCatChem
FULL PAPER
Entry for the Table of Contents (Please choose one layout)
Layout 1:
FULL PAPER
Text for Table of Contents
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
((Insert TOC Graphic here: max.
width: 5.5 cm; max. height: 5.0 cm))
Layout 2:
FULL PAPER
Xiangcheng Li, Kaihao Peng, Xiaohui
Liu*, Qineng Xia, Yanqin Wang*
Page No. – Page No.
Comprehensive Understanding the
Role of Brønsted and Lewis Acid
Sites in Glucose Conversion to 5-
Hydromethylfurfural
The selective formation of HMF form glucose has been studied using silica-alumina
composite (AlSiO) catalysts, and the effect of Brønsted and Lewis acid sites on the
catalytic activity and selectivity was systematically investigated.
This article is protected by copyright. All rights reserved.