Production of lactic acid derivatives from sugars over post-synthesized Sn-Beta zeolite promoted by WO3

Article abstract of DOI:10.1016/j.foodchem.2019.03.039

Various metal oxides were used as co-catalysts to improve the production of alkyl lactate over Sn-Beta-P. WO₃ exhibited the best promotion effect. The yield of MLA increased from 25% (6.5 g L^?1) over Sn-Beta-P (0.2 g) to 52% (13.4 g L^?1) over WO₃ (0.1 g) and Sn-Beta-P (0.1 g) at 160 °C for 5 h and 3.1 wt% of glucose concentration. MLA yield of 38% was attained even at glucose concentration of 10 wt% and the space-time yield reached 7.1 g L^?1 h^?1. The action mechanism of WO₃ was investigated. Fine WO₃ particles adsorbed on surface of Sn-Beta-P in reaction media and decreased the silanol defects of Sn-Beta-P. This promotes retro-aldol of fructose, the rate-determining step of whole reaction, thus facilitated the formation of MLA. Kinetic studies indicate that the presence of WO₃ decreased the activation energy of the retro-aldol of fructose. The binary solid WO₃ and Sn-Beta-P is recyclable.

Full text of DOI:10.1016/j.foodchem.2019.03.039

FoodChemistry289(2019)285–291
Contents lists available at ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Research Article
Production of lactic acid derivatives from sugars over post-synthesized Sn-
Beta zeolite promoted by WO3
Xiaomei Yanga, Yali Zhanga, Lipeng Zhoua, Beibei Gaoa, Tianliang Lub,⁎, Yunlai Sua, Jie Xuc
a College of Chemistry and Molecular Engineering, Zhengzhou University, 100 Kexue Road, Zhengzhou 450001, China
b School of Chemical Engineering and Energy, Zhengzhou University, 100 Kexue Road, Zhengzhou 450001, China
c State Key Laboratory of Catalysis, Dalian National Laboratory for Clean Energy, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan
Road, Dalian 116023, China
A R T I C L E I N F O
A B S T R A C T
Keywords:
Various metal oxides were used as co-catalysts to improve the production of alkyl lactate over Sn-Beta-P. WO3
exhibited the best promotion effect. The yield of MLA increased from 25% (6.5 g L−1) over Sn-Beta-P (0.2 g) to
52% (13.4 g L−1) over WO3 (0.1 g) and Sn-Beta-P (0.1 g) at 160 °C for 5 h and 3.1 wt% of glucose concentration.
MLA yield of 38% was attained even at glucose concentration of 10 wt% and the space-time yield reached
7.1 g L−1 h−1. The action mechanism of WO3 was investigated. Fine WO3 particles adsorbed on surface of Sn-
Beta-P in reaction media and decreased the silanol defects of Sn-Beta-P. This promotes retro-aldol of fructose, the
rate-determining step of whole reaction, thus facilitated the formation of MLA. Kinetic studies indicate that the
presence of WO3 decreased the activation energy of the retro-aldol of fructose. The binary solid WO3 and Sn-
Beta-P is recyclable.
Lactic acid derivatives
Biomass-derived sugars
Glucose
Post-synthesized Sn-Beta
Metal oxide
1. Introduction
Taarning, 2010; Zhou et al., 2014); moreover, the redro-aldol of fruc-
tose to trioses is considered to be the rate-determining step (Holm et al.,
Lactic acid (LA) and its derivatives have wide applications in food,
cosmetic, pharmaceutical and chemical industries (Lu et al., 2017;
Mäki-Arvela, Simakova, Salmi, & Murzin, 2014; Panesar, Kennedy,
Gandhi, & Bunko, 2007). In food industry, lactic acid can be applied as
acidifying agent, flavouring, antimicrobial substance, and preserver in
production of food (Djukić-Vuković et al., 2012). Meanwhile, the de-
rivative poly(lactic acid), a promising biodegradable polymer, can be
used to produce active food packaging which can prolong the shelf life
and/or improve the sensory or safety properties while preserving the
quality of the food (Altan, Aytac, & Uyar, 2018). Sugars including
glucose, sucrose, cellobiose and so on are usually used to produce LA
and its derivatives by fermentation method. Besides microbial fer-
mentation, catalytic conversion of sugars to lactate derivatives is also
attracts much attention in recent years (Mäki-Arvela et al., 2014).
Glucose is the basic unit in many sugars. Conversion of glucose to LA or
alkyl lactate was selected as the model reaction for sugars conversion in
many reports. It is generally accepted that the catalytic conversion of
glucose to LA or alkyl lactate proceeds through the isomerization of
glucose to fructose, the redro-aldol of fructose to trioses, and then the
isomerization of triose to LA in water or alkyl lactate in alcohol
(Scheme S1) (de Clippel et al., 2012; Holm, Saravanamurugan, &
2010; Orazov & Davis, 2015).
Solid Lewis acids and solid bases are two types of efficient hetero-
geneous catalysts for transformation of sugars to LA and alkyl lactates
(de Clippel et al., 2012; Guo et al., 2013; Holm et al., 2010; Liu et al.,
2011; Orazov & Davis, 2015; Pang et al., 2017; Zhang et al., 2017).
Lewis acids are more active and selective than solid bases. Among Lewis
acids, Sn-Beta zeolite is the most studied one, which can be prepared by
the bottom-up (i.e. hydrothermal synthesis and dry gel conversion)
(Chang, Cho, Wang, Wang, & Fan, 2015; Holm et al., 2010; Kang,
Zhang, Liu, Qiu, & Yeung, 2013; Tolborg et al., 2014; Zhu, Xu, Jiang,
Guan,
& Wu, 2017), and the top-down method (post-synthesis)
(Dijkmans et al., 2013; Hammond, Conrad, & Hermans, 2012; Li et al.,
2011; Tang et al., 2014; Zhang et al., 2017). The post-synthesis method
is simple and timesaving than the hydrothermal synthesis method;
moreover, Sn-Beta zeolite with higher Sn content can be obtained.
However, the post-synthesized Sn-Beta (Sn-Beta-P) zeolite usually has
considerable silanol defects and thus is hydrophilic, which render it less
selective for production of alkyl lactate from hexose. The yield of me-
thyl lactate (MLA) from glucose, the most cheap and abundant C6
monosaccharide, is about 20% over Sn-Beta-P, which is much lower
than over Sn-Beta hydrothermally synthesized in F− media (Tolborg
⁎ Corresponding author.
E-mail address: lutianliang@zzu.edu.cn (T. Lu).
https://doi.org/10.1016/j.foodchem.2019.03.039
Received 1 November 2018; Received in revised form 7 March 2019; Accepted 9 March 2019
Availableonline12March2019
0308-8146/©2019ElsevierLtd.Allrightsreserved.

X. Yang, et al.
FoodChemistry289(2019)285–291
et al., 2015; Yang et al., 2017). In view of the virtues of the post-
synthesis method, it is highly desirable that Sn-Beta-P can give high
selectivity for alkyl lactate from hexose. Tolborg et al. incorporated
alkali salts to Sn-Beta-P or to the reaction medium to increase MLA
yield from hexose (Elliot, Tolborg, Madsen, Taarning, & Meier, 2018;
Tolborg et al., 2015). It is speculated that the improvement in MLA
selectivity could arise from the neutralization of Brønsted acidity of
defects in the framework, thereby promoting the retro-aldol of hexose
and limiting the formation of byproducts. Our previous work indicates
that recrystallization of Sn-Beta-P in the presence of low concentration
of tetraethyl ammonium hydroxide can heal silanol defects and at the
same time generate hierarchical porosity. Thus, the retro-aldol of
fructose to trioses was facilitated and the yield of MLA improved (Yang
et al., 2017). Orazov and Davis combined MoO3 with Sn-MFI and rea-
lized the conversion of ketohexoses to alkyl lactates at moderate tem-
perature (about 100 °C) (Orazov & Davis, 2015). But, the action me-
chanism of MoO3 was not revealed; moreover, the yield of alkyl lactate
decreased drastically with increasing the substrate concentration with
this catalyst system. When the content of fructose in ethanol increased
from 1 wt% to 5 wt%, the yield of ethyl lactate reduced from 65.7% to
21.0%.
emission spectroscopy (ICP-AES) was applied to determine the W con-
tent in methanol. The acid properties of the catalyst were measured by
FT-IR spectra with pyridine adsorption on a spectrometer of Bruker
Tensor II. First, the catalyst was treated at 450 °C for 3 h under eva-
cuation condition. By using air as the background, FT-IR spectrum in
the hydroxyl stretching vibration region was collected after cooling
down to room temperature. After collection of another IR spectrum for
using as background for the following test, pyridine gas was charged
into the pool. After adsorption saturation, pyridine molecules were
desorbed at designated temperature for 30 min, and pyridine-FT-IR
spectrum was collected.
2.4. Reaction tests
Retro-aldol reaction of sugars was performed in a stainless auto-
clave (80 mL) with PTFE liner. Methanol (15 mL), sugar and catalyst
were charged into the reactor in turn under stirring. After the autoclave
was sealed, the atmosphere in the reactor was replaced four times with
N2 and then 0.5 MPa N2 was charged. The reactor was heated to the
desired temperature with a heating rate of ∼4 °C min−1 and kept the
temperature for a designed time. When the reaction was over, the re-
actor was quenched in ice-water. The reaction mixture was centrifuged
and the obtained liquid was collected and adjusted to constant volume
with methanol for GC and HPLC analyses. GC analysis was carried out
on a Shimadzu GC-2014C equipped with an FID detector and naph-
thalene was used as the internal standard. LC analysis was performed
on a Shimadzu LC-20AT HPLC analysis system equipped with a re-
fractive index detector (RID-10A) and an Aminex HPX-87H column
(300 mm × 7.8 mm). H2SO4 (0.005 M, 0.5 mL min−1) was used as the
mobile phase and the column temperature was 60 °C. It is noted that the
yields of retro-aldol products are carbon yields.
In the present work, metal oxide is used to modify the selectivity
and activity of Sn-Beta-P for retro-aldol of hexose to alkyl lactate which
is the precursor for producing high quality LA by hydrolysis. Various
metal oxides were tested for the retro-aldol of glucose to MLA, and WO3
exhibited the best promotion effect on the production of MLA. The
action mechanism of WO3 was revealed through kinetic studies and
controlled experiments. Additionally, various reaction conditions were
optimized to improve the productivity of MLA.
2. Materials and methods
2.1. Materials
3. Results and discussion
Metal oxides were prepared by calcination of the corresponding
precursors in muffle furnace. The used materials and the detailed cal-
cination conditions were listed in Table S1. The structure of the metal
oxides was determined by XRD (Fig. S1).
3.1. Characterization of Sn-Beta-P
XRD pattern (Fig. S2a) shows that Sn-Beta-P has a typical BEA
structure and no other crystalline phase is observed. SEM image in-
dicates that Sn-Beta-P is lump-like aggregates in micrometer size. The
result of N2 physisorption (Fig. S2b, S2c, S3 and Table S2) shows that
Sn-Beta-P has micropore diameter of 0.78 nm; additionally, some me-
sopores with volume of 0.15 mL g−1 present due to dealumination.
UV–vis DRS (Fig. S2d) of Sn-Beta-P has a strong absorption at 200 nm,
which is assigned to the charge transfer of O2− to tetrahedrally co-
ordinated Sn4+ in the framework positions (Dijkmans et al., 2013; Yang
et al., 2017). The framework Sn4+ sites generated strong Lewis acid
sites as shown in Fig. S2e, which interact with pyridine giving the ab-
sorption band at ca. 1450 cm−1 (Tang et al., 2014). Sn-Beta-P has also a
very small amount of Brønsted acid sites (ca. 1550 cm−1) deriving from
residual framework Al (Yang et al., 2017). The band at around
1490 cm−1 is ascribed to Lewis and Brønsted acid sites (Tang et al.,
2014). These characterization results confirm that Sn-Beta-P zeolite was
successfully prepared.
2.2. Synthesis of Sn-Beta-P
Sn-Beta-P was prepared by solid-state ion-exchange method from
Al-Beta (nSi/nAl of 13.8) purchased from Nankai University. According
to the literature (Hammond et al., 2012; Yang et al., 2017), Al-Beta was
first dealuminated with HNO3 (13 mol L−1) at 100 °C for 20 h. The
dealuminated Beta (deAl-Beta) was filtered, washed with deionized
water and dried at 100 °C overnight. The dried deAl-Beta was ground
with SnCl4·5H2O in agate mortar for 1 h; then the mixture was dried at
100 °C overnight and calcined at 550 °C for 6 h to obtain Sn-Beta-P. The
nominal Sn content is 2 wt%.
2.3. Characterization
X-ray powder diffraction (XRD) patterns were obtained with a
PANalytical X’pert PRO instrument with Cu Kα radiation
(λ = 0.15418 nm). Scanning electron microscopy (SEM) was performed
on JEOL JSM-6700F. The adsorption/desorption isotherms were mea-
sured with a Quantachrome Autosorb using N2 as adsorbate at −196 °C.
Sn-Beta-P was outgassed at 300 °C for 3 h before measurements. Total
surface area was calculated on the basis of Brunauer-Emmet-Taller
(BET) method. Micropore size distribution was calculated based on
Horvath-Kawazoe (HK) method, and mesopore size distribution was
calculated from the desorption branch of the isotherm with Barret-
Joyner-Halenda (BJH) method. Ultraviolet–visible diffuse reflectance
spectra (UV–vis DRS) were carried out on an Agilent Technologies Cary
Series UV–vis-NIR spectrophotometer. Induced coupled plasma-atomic
3.2. Screening metal oxides as promoter
Metal oxides including MgO, TiO2, ZrO2, NiO, CuO, ZnO, MoO3 and
WO3 were selected as promoter for the conversion of glucose to MLA
(Fig. S4). The products of glucose conversion were identified by HPLC
and GC–MS. Besides the main product MLA and the addition products
of the intermediate pyruvaldehyde (PA) with methanol which are de-
noted as other C3 products (Yang et al., 2017), low concentration of
glycolaldehyde (C2 product) and erythrose (C4 product) was also de-
tected. In the blank experiment, the total yield of C3 products is no more
than 10% though 39% of glucose was converted. The yield of MLA and
other C3 products both increased significantly to 25% under the
286

X. Yang, et al.
FoodChemistry289(2019)285–291
Table 1
Isomerization of glucose in methanol.1
Catalyst
Temperature (oC)
Time (h)
Conversion (%)
Yield (%)
Fructose
Mannose
MG
MLA
Other C3
WO3
110
130
130
160
90
110
130
90
1
1
10
1
1
1
1
1
1
1
12
19
41
46
65
95
99
40
75
98
0
0
0
0
7
0
0
0
0
7
8
10
5
12
10
2
0
0
0
2
3
9
2
9
0
0
4
3
0
6
11
2
10
20
13
14
16
0
0
0
0
0
0
Sn-Beta-P
39 (60)
58 (61)
42 (42)
26 (65)
30 (40)
21 (21)
2
Sn-Beta-P + WO3
110
130
28
1
Reaction conditions: glucose (0.37 g), catalyst (0.2 g), methanol (15 mL), N2 (0.5 MPa). The value in the parenthesis is the selectivity of fructose.
The binary solids are composed of Sn-Beta-P (0.1 g) and WO3 (0.1 g).
2
catalysis of Sn-Beta-P, confirming the ability of aldose-ketose iso-
P combined with WO3 is less active than that combined with MoO3 for
the conversion of glucose to MLA. But, the activity of Sn-Beta-P com-
bined with WO3 at low temperature gradually approaches to that of Sn-
Beta-P combined with MoO3 with the increase of glucose concentration.
When the concentration of glucose is 8.5 wt% (Fig. S6c), the activity of
Sn-Beta-P combined with WO3 at ∼130 °C already exceeds that of Sn-
Beta-P combined with MoO3. At glucose concentration of 8.5 wt%, 43%
yield of MLA was obtained at 160 °C for 5 h over WO3 promoted Sn-
Beta-P, which is much higher than over Sn-Beta-P promoted by MoO3
(24%). WO3 promoted Sn-Beta-P catalyst gave 38% MLA yield even
when the concentration of glucose was 10 wt% and the corresponding
space-time yield reached 7.1 g L−1h−1. The high space-time yield is
highly desirable and very favorable for practical application. In addi-
tion, it should be mentioned that MoO3 is unstable and partially re-
duced by glucose during the reaction as revealed by the XRD results in
Fig. S1g.
merization and retro-aldol condensation of this Lewis acid catalyst. It is
reported that the formation of other C3 products is catalyzed by
Brønsted acid sites (Pescarmona et al., 2010; Saravanamurugan &
Riisager, 2013; Taarning et al., 2009). Sn-Beta-P had a few Brønsted
acid sites (Fig. S2e) generated from the residual Al species (Yang et al.,
2017). So, the production of other C3 was attributed to the presence of
these Brønsted acid sites. Over metal oxide alone (Fig. S5), the yields of
MLA and other C3 products were much lower than over Sn-Beta-P.
However, when combining metal oxide with Sn-Beta-P, the systems
presented different behaviors. Addition of MgO as a promoter, the yield
of MLA was unchanged, but the yield of other C3 decreased remarkably.
MgO is a typical solid base. During the reaction, a small amount of MgO
would leach into the solution (Liu et al., 2011), which interacted with
the Brønsted acid sites of Sn-Beta-P and thereby limited the formation
of other C3 products. The addition of TiO2 and ZrO2 reduced both MLA
and other C3 products. The other metal oxides including NiO, CuO,
ZnO, MoO3 and WO3 all promoted the formation of MLA, but the yield
of other C3 is similar with that over sole Sn-Beta-P except for MoO3. In
the presence of MoO3, the formation of other C3 products was re-
strained markedly. The yield of MLA increased in the presence of dif-
ferent promoters following the order of NiO < CuO < ZnO <
MoO3 < WO3. The yields of MLA and total C3 products reached 52%
and 74% over WO3 combined with Sn-Beta-P catalyst, respectively,
which are higher than over other binary solids. Besides the C3 products,
low yields of glycolaldehyde (2.3%) and erythrose (0.7%) were de-
tected over WO3 combined with Sn-Beta-P. Glycolaldehyde and ery-
throse were formed via direct retro-aldol condensation of glucose, while
C3 products mainly generated from the retro-aldol condensation of
fructose intermediate (Orazov & Davis, 2015). The catalyst showed
high selectivity for C3 products, which is attributed to the excellent
performance of Sn-Beta-P for aldose-ketose isomerization. It is reported
that the other C3 products, that is the addition products of PA with
methanol, can be transformed to MLA through back reaction (West
et al., 2010). If it can be realized, the yield of the target product MLA
can be further increased. The focus of this work is the C3 products, and
the concentrations of C2 and C4 products are very low in the presence of
Sn-Beta. So, the yields of C2 and C4 products are not given in the fol-
lowing section.
3.3. Action mechanism of WO3
As mentioned in literature (de Clippel et al., 2012; Holm et al.,
2010; Zhou et al., 2014), the conversion of glucose to alkyl lactate is a
tandem reaction and comprises three main steps (Scheme S1). In order
to find the exact action mechanism of WO3 in the reaction, isomeriza-
tion of glucose, retro-aldol of fructose and isomerization of triose (1,3-
dihydroxyacetone, DHA) to MLA were performed over WO3, Sn-Beta-P
and binary solids (WO3 and Sn-Beta-P), respectively. For isomerization
of glucose in methanol (Table 1), WO3 alone shows low activity,
moreover, glucose does not isomerize to fructose but epimerize to
mannose. This behavior of WO3 is similar with molybdate that is a
typical catalyst for epimerization of aldose (Bílik, 1972). Sn-Beta-P was
much more active for glucose conversion, which was converted quickly
at moderate temperature and fructose was a main product. The ex-
cellent performance of Sn-Beta-P for glucose isomerization to fructose is
ascribed to the abundant Si-OH defects (Jiang et al., 2018). This point is
important for production of alkyl lactate, because two trioses can be
obtained from retro-aldol of fructose. Increasing temperature to 130 °C,
considerable retro-aldol products formed and the amount of fructose
decreased. This confirmed the ability of Sn-Beta-P for retro-aldol of
fructose. In addition, a certain amount of methyl glucoside (MG) was
also detected due to that Sn-Beta-P has a few Brønsted acid sites. The
Sn-Beta-P combined with WO3 is less active than solo Sn-Beta-P for
glucose conversion. Fructose is still a main product at lower tempera-
ture, but its selectivity is lower than over solo Sn-Beta-P at comparable
conversion of glucose. It means that the presence of WO3 is unfavorable
for isomerization of glucose to fructose. In order to further confirm this
point, the amount of WO3 was varied in the isomerization of glucose to
fructose. Fig. S7 indicates that the conversion of glucose and the yield of
fructose decreased with increasing the amount of WO3. Contrarily, the
In a work reported by Orazov and Davis, MoO3 combined with a Sn-
containing silicate molecular sieve was reactive for retro-aldol reaction
of fructose even at 100 °C, whereas tungstate species performed poorly
(Orazov & Davis, 2015). It is speculated that reaction conditions, such
as reaction temperature and substrate concentration, have important
influence on the catalytic performance of Sn-Beta-P combined with
MoO3 and WO3. So, we investigated further the conversion of glucose in
methanol over Sn-Beta-P promoted by MoO3 and WO3, respectively.
The results are shown in Fig. S6. Indeed, at lower temperature, Sn-Beta-
287

X. Yang, et al.
FoodChemistry289(2019)285–291
Table 2
Retro-aldol of fructose in methanol.1
Catalyst
Temperature (oC)
Time (h)
Conversion (%)
Yield (%)
MLA
Other C3
MLE
MG
WO3
140
150
160
120
120
160
120
120
160
5
5
5
1
5
5
1
5
5
31
49
80
46
69
99
50
70
99
3
4
4
4
4
4
6
6
11
21
11
17
24
4
5
5
0
4
6
0
0
3
0
0
3
7
10
0
9
8
0
Sn-Beta-P
8
25
11
18
50
2
Sn-Beta-P + WO3
1
Reaction conditions: fructose (0.34 g), catalyst (0.2 g), methanol (15 mL), N2 (0.5 MPa).
The binary catalyst is composed of Sn-Beta-P (0.1 g) and WO3 (0.1 g).
2
selectivity for retro-aldol products increased significantly over the
binary solids. At 130 °C for 1 h, the yield of total C3 products reached
48%.
isomerization of DHA to MLA was also investigated (Fig. 1). The ac-
tivities of the three catalysts follow the order of Sn-Beta-P > the binary
solids > WO3. It shows that the addition of WO3 has an adverse effect
on the conversion of DHA to MLA.
The performance of WO3, Sn-Beta-P and the binary solids for the
retro-aldol reaction of fructose in methanol is given in Table 2. The
activity of WO3 for the retro-aldol reaction of fructose is very low, al-
though it is an acid material (Fig. S8); the yield of total C3 products was
no more than 10% even at 160 °C for 5 h. A small amount of methyl
levulinate (MLE) was also detected, which derived from fructose de-
hydration, a competitive reaction of the retro-aldol of fructose. This
implies that WO3 in methanol can afford Brønsted acid sites because
fructose dehydration is generally catalyzed by Brønsted acid (Jiang
et al., 2018). Sn-Beta-P alone shows much higher activity for the retro-
aldol reaction of fructose than WO3 alone. 46% yield of total C3 pro-
ducts was achieved at 160 °C for 5 h. For the Sn-Beta-P combined with
WO3, the conversion of fructose was comparable to that over Sn-Beta-P
alone, but the selectivity for the retro-aldol products increased re-
markably. At 160 °C for 5 h, 74% yield of total C3 products was ob-
tained. The above results proved that Sn-Beta-P combined with WO3
promoted the retro-aldol reaction of fructose, the rate-determining step
of glucose to alkyl lactate.
The kinetics of the retro-aldol of fructose over Sn-Beta-P combined
with WO3 was studied and compared with Sn-Beta-P alone (Fig. S9 and
Fig. S10). The retro-aldol of fructose fitted well with the pseudo-first-
order reaction kinetics, which is consistent with our previous report
(Yang et al., 2017). Over Sn-Beta-P combined with WO3, the apparent
activation energy is 91 kJ mol−1 (Fig. S11a), which is lower than that
over Sn-Beta-P alone (131 kJ mol−1) (Fig. S11b). This means that WO3
as a promoter changed the formation pathway of MLA, lowered the
activation energy, and thus facilitated the formation of MLA.
The performance of the Sn-Beta-P, WO3, and the binary solids for
Based on the above results, it can be concluded that: (1) WO3 alone
shows low activity in all the three steps including isomerization of
glucose, retro-aldol of fructose and isomerization of DHA to MLA. (2)
Compared to Sn-Beta-P alone, addition of WO3 has a significant pro-
motion effect on the retro-aldol of fructose, while has adverse effects on
both the isomerization reactions of glucose to fructose and DHA to
MLA. Considering the poor performance of WO3 alone in the retro-aldol
of fructose to C3 products, the promotion effect of WO3 might be rea-
lized through the interaction between WO3 and Sn-Beta-P. In other
words, WO3 might modify some properties of Sn-Beta-P. To investigate
the promotion mechanism, control experiments were carried out. First,
WO3 was added to methanol. After treatment at 160 °C for 5 h, WO3 was
removed by centrifugation. Then, Sn-Beta-P and glucose were added to
the treated methanol for reaction. The result is listed in Table 3 (entry
1). 49% of MLA yield and 20% of other C3 yield were obtained, which
are comparable to that over the Sn-Beta-P combined with WO3 (52%
MLA yield and 22% other C3 yield as shown in Fig. S4). This indicates
the promotion effect also existed when WO3 treated methanol and Sn-
Beta-P were combined. Similar phenomenon also presents in MoO3
catalytic system (Table 3, entry 2). ICP results indicate 99.9 mg L−1 of
W element and 503 mg L−1 of Mo element exist in the reaction solution.
Is the W species in treated methanol homogeneous or heterogeneous? If
the W species in treated methanol is homogeneous, there would be a
precipitate-dissolution equilibrium in the system. In this case, the
concentration of W species in methanol would be unrelated to the ad-
dition amount of WO3 as long as the presence of solid WO3 in methanol.
However, as shown in Fig. S12, there is an obvious relationship
100
b
100
a
WO3
80
90
80
Sn-Beta-P
WO3 + Sn-Beta-P
60
40
20
0
70
WO3
Sn-Beta-P
WO3 + Sn-Beta-P
60
50
40
50
60
70
80
90
40
50
60
70
80
90
Reaction temperature (oC)
Reaction temperature (oC)
Fig. 1. Isomerization of DHA to MLA over WO3, Sn-Beta-P and Sn-Beta-P combined with WO3. Reaction conditions: DHA (0.34 g), catalyst (0.2 g), methanol (15 mL),
N2 (0.5 MPa), 1 h. The binary solids are composed of Sn-Beta-P (0.1 g) and WO3 (0.1 g).
288

X. Yang, et al.
FoodChemistry289(2019)285–291
Table 3
vibration region. The amount of silanols on the surface of Sn-Beta-P
decreased significantly after interaction with ultra-small WO3 particles.
The interaction between metal oxides, such as WO3 and MoO3, with
–OH groups is a common phenomenon (Li, Pidko, & Hensen, 2016;
Miliordos, Caratzoulas, & Vlachos, 2017). It has been reported that si-
lanols are favorable for the isomerization of glucose to fructose
(Bermejo-Deval, Orazov, Gounder, Hwang, & Davis, 2014; Jiang et al.,
2018; Li, Pidko, & Hensen, 2014; Rai, Caratzoulas, & Vlachos, 2013),
and DHA to MLA (Yang et al., 2016), but disadvantageous for the retro-
aldol of hexose to alkyl lactate (Elliot et al., 2018; Tolborg et al., 2015;
Yang et al., 2017, 2018). Josephson has reported that silanols facilitate
the fructose etherification which is undesirable in conversion of glucose
to MLA (Josephson et al., 2018). Thus, the isomerization of glucose to
fructose and DHA to MLA was restrained over WO3 combined Sn-Beta-P
catalyst due to the decrease in Lewis acidity and silanols; whereas for
the retro-aldol of fructose the decrease of silanols is favorable, and the
yield of MLA was improved. The promotion effect of WO3 on the for-
mation of MLA from glucose over Sn-Beta-P was summarized in Scheme
S2.
To further prove the role of WO3 in conversion of glucose to MLA,
alkali ions, which were reported to be good promoters for production of
MLA from C6 over Sn-containing silicate molecular sieves (Elliot et al.,
2018; Tolborg et al., 2015), were also applied as promoters in this
study. As shown in Fig. S13, KBr can improve the yield of MLA mark-
edly over solo Sn-Beta-P; the yield of MLA was improved from 25%
without KBr to the maximum of 56% at KBr concentration of
6 mmol L−1. However, for binary solids of WO3 and Sn-Beta-P, the
promotion effect of KBr on the selectivity of MLA was moderate. The
yield of MLA only increased from 52% without KBr to the maximum of
60% with 10 mmol L−1 KBr. This result was caused by the overlapping
role of WO3 and KBr, both of which decrease silanol defects of Sn-Beta-
P.
Results of control experiments.1
Entry
Catalytic system
Yield (%)
MLA
Other C3
1
2
3
4
WO3 treated methanol + Sn-Beta-P
MoO3 treated methanol + Sn-Beta-P
Treated Sn-Beta-P + fresh methanol
WO3/Sn-Beta-P
49
41
47
49
20
11
22
14
1
Reaction conditions: glucose (0.37 g), catalyst (0.2 g), methanol (15 mL),
N2 (0.5 MPa), 160 °C, 5 h.
between WO3 amount and MLA yield. Thus, the W species in methanol
should be heterogeneous. Probably, some ultra-small particles that
stripped off from the bulk WO3 adsorbed on Sn-Beta-P, and modified
the properties of Sn-Beta-P, which led to the promotion effect on the
catalytic performance. Another control experiment was carried out to
prove this inference. First, methanol was treated with WO3 at 160 °C for
5 h. After removing WO3, Sn-Beta-P was added to the treated methanol.
The mixture was treated at 160 °C for 5 h, and then Sn-Beta-P was se-
parated out. Finally, the treated Sn-Beta-P and glucose were added to
fresh methanol for reaction. The result is listed in entry 3 in Table 3.
47% of MLA and 22% of other C3 yield were obtained. This result is
comparable to that of catalytic system of WO3 treated methanol and Sn-
Beta-P in entry 1 in Table 3. It suggests that the promotion effect of
WO3 was realized by modification of Sn-Beta-P through adsorption of
heterogeneous ultra-small WO3 particles on Sn-Beta-P. Inspired by the
result, we prepared a sample with highly dispersed W species. W species
was loaded on Sn-Beta-P by incipient wetness impregnation method
with an aqueous (NH4)10W12O41·xH2O solution. After drying, the solid
was calcined at 450 °C for 2 h. The sample was denoted as WO3/Sn-
Beta-P and W/Sn molar ratio was controlled at unit. As expected, the
catalytic performance of WO3/Sn-Beta-P (Table 3, entry 4) is similar
with the controlled experiments and comparable to that of Sn-Beta-P
combined with bulk WO3. The result confirmed the promotion effect of
heterogeneous small particles of WO3 on the catalytic performance of
Sn-Beta-P.
3.4. Product evolution with reaction time
The effects of reaction temperature and time on the conversion of
glucose over WO3 promoted Sn-Beta-P catalyst are depicted in Fig. S14.
The yield of MLA is highly dependent on the temperature, which in-
creased significantly with the increase of the temperature. Nevertheless,
the yield of other C3 products only slightly enhanced with the tem-
perature. The yields of MLA and other C3 products increased gradually
with the reaction time, but too long time reduced the yields of MLA and
other C3 products slightly.
To investigate the effect of WO3 on the acid properties of Sn-Beta-P,
FT-IR spectra of pyridine adsorbed on Sn-Beta-P with/without small
particles of WO3 were collected (Fig. 2a). For treated Sn-Beta-P
(Table 3, entry 3), the absorption bands at 1451 cm−1 and 1612 cm−1
corresponding to strong Lewis acid sites shifted to lower wavenumber
(1447 cm−1 and 1607 cm−1) compared to Sn-Beta-P, which indicates
the presence of WO3 weakened the strength of Lewis acid. This is dis-
advantageous for all the three steps of glucose to MLA. Fig. 2b shows
the FT-IR spectra of the two samples in the hydroxyl stretching
3.5. Scope of WO3 promoted Sn-Beta-P catalyst
WO3 promoted Sn-Beta-P catalyst was also tested for transformation
b
Treated Sn-Beta-P
Sn-Beta-P
1451
a
1447
Sn-Beta-P
1607
1612
Treated Sn-Beta-P
1650
1600
1550
1500
1450
1400
4000
3800
3600
3400
3200
3000
Wavenumber (cm-1)
Wavenumber (cm-1)
Fig. 2. FT-IR spectra of Sn-Beta-P and treated Sn-Beta-P (a) after pyridine adsorption and evacuation at 150 °C for 30 min, and (b) in the hydroxyl stretching vibration
region.
289

X. Yang, et al.
FoodChemistry289(2019)285–291
yield of MLA. The promotion mechanism of WO3 in this study is also
suitable for other metal oxides, such as MoO3, in similar reaction
system. Moreover, the binary catalyst of WO3 and Sn-Beta-P is easily
separated and recyclable.
Table 4
Conversion of various sugars to alkyl lactates over WO3 promoted Sn-Beta-P
catalyst.1
Entry
Substrate
Solvent
Yield of alkyl lactate (mol %)
1
2
3
4
5
6
7
Glucose
Fructose
Mannose
Sucrose
Cellobiose
Glucose
Glucose
Methanol
Methanol
Methanol
Methanol
Methanol
Ethanol
52
50
50
60
9
Acknowledgements
This work was supported by the National Natural Science
Foundation of China (21871236); the China Postdoctoral Science
Foundation (2017M612418); the Natural Science Foundation of Henan
Province (182300410122); the Outstanding Young Talent Research
Fund of Zhengzhou University (1521316006); and the State Key
Laboratory of Catalysis in DICP (N-16-02).
34
23
n-Butanol
1
Reaction conditions: carbohydrate (containing 1.87 mmol C6 units), Sn-
Beta-P (0.1 g), WO3 (0.1 g), methanol (15 mL), N2 (0.5 MPa), 160 °C, 5 h.
Declarations of interest
None.
of other C6 carbohydrates to MLA. The results are presented in Table 4.
Glucose, fructose and mannose (entries 1–3) give similar MLA yield,
implying the binary catalyst can catalyze the mutual isomerization of
the three carbohydrates. Sucrose (entry 4) composed of one glucose
unit and one fructose unit linked by α,β-1,2-glycosidic bond gives
higher MLA yield than C6 monosaccharide. The reason is probably that
the gradual degradation of sucrose made the concentration of C6
monosaccharide in low level, which reduced the side reactions (Pang
et al., 2017; Yang et al., 2018). Cellobiose (entry 5), a disaccharide,
made of two glucose units through β-1,4-glycosidic bond, provides
much lower MLA yield than sucrose due to the difficult depolymer-
ization of β-1,4-glycosidic bond. With ethanol and n-butanol as sol-
vents, the yields of alkyl lactates reduced. The change trend is similar
with our previous reports (Yang et al., 2017, 2018). The reason can be
ascribed to larger diffusion limitation in higher alcohols over Sn-Beta
zeolite.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.foodchem.2019.03.039.
References
Altan, A., Aytac, Z., & Uyar, T. (2018). Carvacrol loaded electrospun fibrous films from
zein and poly(lactic acid) for active food packaging. Food Hydrocolloids, 81, 48–59.
Bermejo-Deval, R., Orazov, M., Gounder, R., Hwang, S., & Davis, M. E. (2014). Active
sites in Sn-Beta for glucose isomerization to fructose and epimerization to mannose.
ACS Catalysis, 4(7), 2288–2297.
Bílik, V. (1972). Reactions of saccharides catalyzed by molybdate ions. II.* epimerization
of D-glucose and D-mannose. Chemical Papers, 26(2), 183–186.
Chang, C., Cho, H. J., Wang, Z., Wang, X., & Fan, W. (2015). Fluoride-free synthesis of a
Sn-BEA catalyst by dry gel conversion. Green Chemistry, 17(5), 2943–2951.
de Clippel, F., Dusselier, M., Rompaey, R. V., Vanelderen, P., Dijkmans, J., Makshina, E.,
et al. (2012). Fast and selective sugar conversion to alkyl lactate and lactic acid with
bifunctional carbon–silica catalysts. Journal of the American Chemical Society,
134(24), 10089–10101.
Dijkmans, J., Gabriëls, D., Dusselier, M., de Clippel, F., Vanelderen, P., Houthoofd, K.,
et al. (2013). Productive sugar isomerization with highly active Sn in dealuminated β
zeolites. Green Chemistry, 15(10), 2777–2785.
Djukić-Vuković, A. P., Mojović, L. V., Vukašinović-Sekulić, M. S., Rakin, M. B., Nikolić, S.
B., Pejin, J. D., et al. (2012). Effect of different fermentation parameters on L-lactic
acid production from liquid distillery stillage. Food Chemistry, 134(2), 1038–1043.
Elliot, S., Tolborg, S., Madsen, R., Taarning, E., & Meier, S. (2018). Effects of alkali and
counter ions in Sn-Beta catalyzed carbohydrate conversion. ChemSusChem, 11(7),
1198–1203.
Guo, Q., Fan, F., Pidko, E. A., van der Graaff, W. N. P., Feng, Z., Li, C., et al. (2013).
Highly active and recyclable Sn-MWW zeolite catalyst for sugar conversion to methyl
lactate and lactic acid. ChemSusChem, 6(8), 1352–1356.
Hammond, C., Conrad, S., & Hermans, I. (2012). Simple and scalable preparation of
highly active Lewis acidic Sn-β. Angewandte Chemie International Edition, 51(47),
11736–11739.
Holm, M. S., Saravanamurugan, S., & Taarning, E. (2010). Conversion of sugars to lactic
acid derivatives using heterogeneous zeotype catalysts. Science, 328(5978), 602–605.
Jiang, L., Zhou, L., Chao, J., Zhao, H., Lu, T., Su, Y., et al. (2018). Direct catalytic con-
version of carbohydrates to methyl levulinate: Synergy of solid Brønsted acid and
Lewis acid. Applied Catalysis B: Environmental, 220, 589–596.
Josephson, T. R., DeJaco, R. F., Pahari, S., Ren, L., Guo, Q., Tsapatsis, M., et al. (2018).
Cooperative catalysis by surface Lewis acid/silanol for selective fructose etherifica-
tion on Sn-SPP zeolite. ACS Catalysis, 8(10), 9056–9065.
Kang, Z., Zhang, X., Liu, H., Qiu, J., & Yeung, K. L. (2013). A rapid synthesis route for Sn-
Beta zeolites by steam-assisted conversion and their catalytic performance in Baeyer-
Villiger oxidation. Chemical Engineering Journal, 218, 425–432.
3.6. Reusability of WO3 promoted Sn-Beta-P catalyst
The reusability of WO3 promoted Sn-Beta-P catalyst was studied
(Fig. S15). From the 1st to the 3rd run, the catalyst was recovered by
centrifugation, washed with methanol for three times, dried at ambient
temperature and then used directly. The yields of MLA and other C3
products unchanged in the 2nd run, but decreased dramatically in the
3rd run. The catalyst recovered from the 3rd run was calcined at 550 °C
for 1 h besides washing with methanol and then used for the 4th run.
The yields of MLA and other C3 products regained. This indicates the
organics deposited during the reaction restrained the performance of
the catalyst. After that, the recovered catalyst was all calcined before
used for the next run. The yields of MLA and other C3 products were
kept, proving the high recyclability of the binary solids. In addition, a
hot filtration experiment was performed. The binary solids were sepa-
rated from the reaction mixture after reaction at 160 °C for 1 h, and
then the reaction proceeded for another 4 h. The yield of MLA only
slightly increased from 42% to 45%, confirming the heterogeneous
character of the binary catalyst. Sn-Beta-P used here is prepared
without toxic fluoride and organic template, and the catalyst of Sn-
Beta-P and WO3 is recyclable. Besides, the product of MLA can be easily
separated from the reaction mixture by distillation. In view of these
advantages, the method developed here can be regarded as a green
catalytic approach.
Li, P., Liu, G., Wu, H., Liu, Y., Jiang, J., & Wu, P. (2011). Postsynthesis and selective
oxidation properties of nanosized Sn-Beta zeolite. The Journal of Physical Chemistry C,
115(9), 3663–3670.
4. Conclusions
Li, G., Pidko, E. A., & Hensen, E. J. M. (2014). Synergy between Lewis acid sites and
hydroxyl groups for the isomerization of glucose to fructose over Sn-containing
zeolites: A theoretical perspective. Catalysis Science & Technology, 4(8), 2241–2250.
Li, G., Pidko, E. A., & Hensen, E. J. M. (2016). A periodic DFT study of glucose to fructose
isomerization on tungstite (WO3·H2O): Influence of group IV–VI dopants and co-
operativity with hydroxyl groups. ACS Catalysis, 6(7), 4162–4169.
Liu, Z., Li, W., Pan, C., Chen, P., Lou, H., & Zheng, X. (2011). Conversion of biomass-
derived carbohydrates to methyl lactate using solid base catalysts. Catalysis
Communications, 15(1), 82–87.
WO3 shows significant promotion effect for increasing the yield and
selectivity of MLA from glucose over Sn-Beta-P; the yield of MLA in-
creased from 25% to 52%. The space-time yield of 7.1 g L−1 h−1 was
obtained. Fine WO3 particles were adsorbed on surface of Sn-Beta-P in
reaction media and decreased the silanol defects on Sn-Beta-P. This
improves the selectivity of retro-aldol of fructose, which is the rate-
determining step of hexose conversion to alkyl lactate, and thus the
Lu, T., Fu, X., Zhou, L., Su, Y., Yang, X., Han, L., et al. (2017). Promotion effect of Sn on
Au/Sn-USY catalysts for one-pot conversion of glycerol to methyl lactate. ACS
290

X. Yang, et al.
FoodChemistry289(2019)285–291
Tolborg, S., Katerinopoulou, A., Falcone, D. D., Sádaba, I., Osmundsen, C. M., Davis, R. J.,
et al. (2014). Incorporation of tin affects crystallization, morphology, and crystal
composition of Sn-Beta. Journal of Materials Chemistry A, 2(47), 20252–20262.
Tolborg, S., Sádaba, I., Osmundsen, C. M., Fristrup, P., Holm, M. S., & Taarning, E. (2015).
Tin-containing silicates: Alkali salts improve methyl lactate yield from sugars.
ChemSusChem, 8(4), 613–617.
West, R. M., Holm, M. S., Saravanamurugan, S., Xiong, J., Beversdorf, Z., Taarning, E.,
et al. (2010). Zeolite H-USY for the production of lactic acid and methyl lactate from
C3-sugars. Journal of Catalysis, 269(1), 122–130.
Yang, X., Bian, J., Huang, J., Xin, W., Lu, T., Chen, C., et al. (2017). Fluoride-free and low
concentration template synthesis of hierarchical Sn-Beta zeolites: Efficient catalysts
for conversion of glucose to alkyl lactate. Green Chemistry, 19(3), 692–701.
Yang, X., Liu, Y., Li, X., Ren, J., Zhou, L., Lu, T., et al. (2018). Synthesis of Sn-containing
nano-sized Beta zeolite as efficient catalyst for transformation of glucose to methyl
lactate. ACS Sustainable Chemistry & Engineering, 6(7), 8256–8265.
Yang, X., Wu, L., Wang, Z., Bian, J., Lu, T., Zhou, L., et al. (2016). Conversion of dihy-
droxyacetone to methyl lactate catalyzed by highly active hierarchical Sn-USY at
room temperature. Catalysis Science & Technology, 6(6), 1757–1763.
Zhang, J., Wang, L., Wang, G., Chen, F., Zhu, J., Wang, C., et al. (2017). Hierarchical Sn-
Beta zeolite catalyst for the conversion of sugars to alkyl lactates. ACS Sustainable
Chemistry & Engineering, 5(4), 3123–3131.
Zhou, L., Wu, L., Li, H., Yang, X., Su, Y., Lu, T., et al. (2014). A facile and efficient method
to improve the selectivity of methyl lactate in the chemocatalytic conversion of
glucose catalyzed by homogeneous Lewis acid. Journal of Molecular Catalysis A:
Chemical, 388–389, 74–80.
Catalysis, 7(10), 7274–7284.
Mäki-Arvela, P., Simakova, I. L., Salmi, T., & Murzin, D. Y. (2014). Production of lactic
acid/lactates from biomass and their catalytic transformations to commodities.
Chemical Reviews, 114(3), 1909–1971.
Miliordos, E., Caratzoulas, S., & Vlachos, D. G. (2017). A periodic-DFT study of retro-aldol
fragmentation of fructose on MoO3. Applied Catalysis A: General, 530, 75–82.
Orazov, M., & Davis, M. E. (2015). Tandem catalysis for the production of alkyl lactates
from ketohexoses at moderate temperatures. Proceedings of the National Academy of
Sciences of the United States of America, 112(38), 11777–11782.
Panesar, P. S., Kennedy, J. F., Gandhi, D. N., & Bunko, K. (2007). Bioutilisation of whey
for lactic acid production. Food Chemistry, 105(1), 1–14.
Pang, J., Zheng, M., Li, X., Song, L., Sun, R., Sebastian, J., et al. (2017). Catalytic con-
version of carbohydrates to methyl lactate using isolated tin sites in SBA-15.
ChemistrySelect, 2(1), 309–314.
Pescarmona, P. P., Janssen, K. P. F., Delaet, C., Stroobants, C., Houthoofd, K.,
Philippaerts, A., et al. (2010). Zeolite-catalysed conversion of C3 sugars to alkyl
lactates. Green Chemistry, 12(6), 1083–1089.
Rai, N., Caratzoulas, S., & Vlachos, D. G. (2013). Role of silanol group in Sn-Beta zeolite
for glucose isomerization and epimerization reactions. ACS Catalysis, 3(10),
2294–2298.
Saravanamurugan, S., & Riisager, A. (2013). Brønsted acid ionic liquid catalyzed for-
mation of pyruvaldehyde dimethylacetal from triose sugars. Catalysis Today, 200,
94–98.
Taarning, E., Saravanamurugan, S., Holm, M. S., Xiong, J., West, R. M., & Christensen, C.
H. (2009). Zeolite-catalyzed isomerization of triose sugars. ChemSusChem, 2(7),
625–627.
Tang, B., Dai, W., Wu, G., Guan, N., Li, L., & Hunger, M. (2014). Improved postsynthesis
strategy to Sn-Beta zeolites as Lewis acid catalysts for the ring-opening hydration of
epoxides. ACS Catalysis, 4(8), 2801–2810.
Zhu, Z., Xu, H., Jiang, J., Guan, Y., & Wu, P. (2017). Sn-Beta zeolite hydrothermally
synthesized via interzeolite transformation as efficient Lewis acid catalyst. Journal of
Catalysis, 352, 1–12.
291