One-Pot Transformation of Cellobiose to Formic Acid and Levulinic Acid over Ionic-Liquid-based Polyoxometalate Hybrids

Article abstract of DOI:10.1002/cssc.201402157

Currently, levulinic acid (LA) and formic acid (FA) are considered as important carbohydrates for the production of value-added chemicals. Their direct production from biomass will open up a new opportunity for the transformation of biomass resource to valuable chemicals. In this study, one-pot transformation of cellobiose into LA and FA was demonstrated, using a series of multiple-functional ionic liquid-based polyoxometalate (IL-POM) hybrids as catalytic materials. These IL-POMs not only markedly promoted the production of valuable chemicals including LA, FA and monosaccharides with high selectivities, but also provided great convenience of the recovery and the reuse of the catalytic materials in an environmentally friendly manner. Cellobiose conversion of 100 %, LA selectivity of 46.3 %, and FA selectivity of 26.1 % were obtained at 423 K and 3 MPa for 3 h in presence of oxygen. A detailed catalytic mechanism for the one-pot transformation of cellobiose was also presented.

Full text of DOI:10.1002/cssc.201402157

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201402157
One-Pot Transformation of Cellobiose to Formic Acid and
Levulinic Acid over Ionic-Liquid-based Polyoxometalate
Hybrids
Kaixin Li, Linlu Bai, Prince Nana Amaniampong, Xinli Jia, Jong-Min Lee,* and Yanhui Yang*[a]
Currently, levulinic acid (LA) and formic acid (FA) are consid-
ered as important carbohydrates for the production of value-
added chemicals. Their direct production from biomass will
open up a new opportunity for the transformation of biomass
resource to valuable chemicals. In this study, one-pot transfor-
mation of cellobiose into LA and FA was demonstrated, using
a series of multiple-functional ionic liquid-based polyoxometa-
late (IL-POM) hybrids as catalytic materials. These IL-POMs not
only markedly promoted the production of valuable chemicals
including LA, FA and monosaccharides with high selectivities,
but also provided great convenience of the recovery and the
reuse of the catalytic materials in an environmentally friendly
manner. Cellobiose conversion of 100%, LA selectivity of
46.3%, and FA selectivity of 26.1% were obtained at 423 K and
3 MPa for 3 h in presence of oxygen. A detailed catalytic mech-
anism for the one-pot transformation of cellobiose was also
presented.
Introduction
Biomass has been extensively discussed as one of the alterna-
tive feedstocks for the production of energy and chemicals be-
cause of its abundance, high energy content, sustainability,
and biodegradability. Compared to existing techniques (enzy-
matic fermentation, high-temperature pyrolysis, and thermal
gasification), aqueous-phase catalytic transformation of noned-
ible biomass resources (e.g., cellulose) into valuable chemicals
or biofuels at high efficiencies and selectivities is more desira-
ble.[1–3] Although various routes for the catalytic conversion of
biomass resources have been explored,[4–7] the transformation
of cellulose into valuable platform molecules in a one-pot ap-
proach is undoubtedly one of the most attractive routes for its
utilization. Yan et al. reported the one-pot transformation of
cellobiose (a model compound for cellulose) into C6 alcohols
over a ruthenium-based catalyst, combining hydrolysis and hy-
drogenation in aqueous medium.[8] A high yield of 62.3%
toward 5-hydroxymethyl furfural (5-HMF) from lignocellulosic
biomass materials was obtained by using an ionic liquid as sol-
vent and the inorganic salt CrCl3 as homogeneous catalyst.[9]
Shunmugavel and co-workers also developed a catalytic trans-
formation of disaccharides into ethyl levulinate, which is the
final product of 5-HMF rehydration in the presence of a func-
tionalized ionic liquid with strong Brønsted acidity.[10]
alternative routes for the efficient production of valuable plat-
form molecules. Through catalysis by gold nanoparticles
loaded onto nitric-acid-pretreated carbon nanotubes, selective
oxidation of cellobiose to gluconic acid proceeded efficiently
in aqueous solution with molecular dioxygen.[11] Recently, Was-
serscheid et al. reported a catalytic route to produce formic
acid (FA) with extremely high selectivity by converting bio-
mass-based carbohydrates in the presence of H5PV2Mo10O40 as
a bifunctional catalyst and molecular dioxygen. Although this
catalytic system required a long reaction time (22 h) and the
detailed mechanism remained unclear, it provided a novel
route for the sustainable production of FA, which is considered
an important commodity chemical and an energy carrier offer-
ing high energy density.[12] Inspired by this progress in catalytic
biomass transformation, we believe that sustainable produc-
tion of energy carriers and chemicals via one-pot synthetic
procedures can be realized by exploring multifunctional cata-
lysts.
Ionic liquid-heteropolyanion (IL-HPA) salts have been report-
ed to show excellent catalytic performance and self-separation
behavior in an esterification process.[13] Because of its high
melting point and polarity, IL-HPA salt dissolves in the polar
solvents and can act as a homogeneous catalyst at high tem-
peratures and precipitate as a gelatinous solid at low tempera-
tures, which makes catalyst recovery remarkably convenient.[13]
In our previous work, IL-HPA salts comprising Keggin-type HPA
and functionalized Brønsted-acidic IL cation were found to ex-
hibit similar catalytic performance and separation behavior in
the transesterification of trimethylolpropane.[14] Nonetheless, all
these results about IL-HPA demonstrate that HPA is primarily
responsible for self-separation, rather than a catalytic active
site carrier.
In recent years, one-pot syntheses combining hydrolytic and
oxidative degradation of biomass resources have emerged as
[a] Dr. K. Li, L. Bai, P. N. Amaniampong, Dr. X. Jia, Prof. J.-M. Lee, Prof. Y. Yang
Chemical and Biomedical Engineering
School of Chemical and Biomedical Engineering
Nanyang Technological University
Singapore 637459 (Singapore)
Fax:(+65)67947553
E-mail: jmlee@ntu.edu.sg
yhyang@ntu.edu.sg
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2670

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Results and Discussion
Herein, a series of multifunctional ionic liquid-based poly-
oxometalate (IL-POM) hybrids comprising IL cations with
strongly acidic groups and POM anions containing various
metallic atoms (tungsten, molybdenum, and vanadium) are de-
signed, fabricated, and used in one-pot catalytic transforma-
tions of cellobiose into valuable chemicals in the presence of
molecular dioxygen (Scheme 1). Both the Brønsted-acidic sites
in the IL cations and the redox catalytic sites in the POM
anions are involved in the catalytic reaction. Upon hydrolysis
Structural characterization of as-synthesized IL-POM
catalysts
FTIR spectra of some representative as-synthesized IL-POM cat-
alysts are collected and shown in Figure 1. The IR bands at
Figure 1. FTIR spectra of different catalyst: (a) [H5PV2Mo10O40
(b) [PyBS]5PV2Mo10O40 (c) [PyBS]4SiW12O40 (d) [PyBS]6P2W18O62 (e) reused
[PyBS]5PV2Mo10O40 after 4 runs.
1633 and 1486 cmÀ1, which are observable for all [PyBS]+
cation-based IL-POM catalysts, are attributed to stretching vi-
bration modes characteristic of CÀH and C=N bonds of pyri-
dine rings, respectively. Characteristic IR bands at 794, 914,
976, and 1033 cmÀ1 were observed for [PyBS]4SiW12O40 (Fig-
ure 1c), indicating that the Keggin structure is well preserved
in the [PyBS]4SiW12O40 catalyst. Similarly, the as-synthesized
[PyBS]6P2W18O62 (Figure 1d) possesses a typical Dawson struc-
ture, as evidenced by the characteristic IR bands at 790,909,
966, and 1091 cmÀ1 attributable to n(WÀOcÀW), n(WÀObÀW),
nas(W=Ot), and n(PÀO), respectively.
The IR spectrum of Keggin-structured vanadium-containing
[PyBS]5PV2Mo10O40 (Figure 1b) shows slight differences in the
characteristic bands for the POM anions compared to its
parent
molybdovanadophosphoric
heteropoly
acid
(H5PV2Mo10O40, Figure 1a). These differences are due to struc-
tural distortion, induced by the donation of electron density
from the conjugated p system from the organic cations to the
inorganic anions.[16,17] Moreover, the IR band at 1078 cmÀ1 for
n(PÀO) in H5PV2Mo10O40 is split into two bands at 1085 and
1041 cmÀ1 after the protons are completely substituted by
[PyBS]+ cations, indicating the formation of extra hydrogen
bonds between terminal oxygens of POM anions and sulfonate
functional groups in the IL cations.[18,19] Accordingly, the solid
nature of these as-synthesized IL-POM composites at room
temperature (melting points>373 K) mainly arises from the ex-
istence of numerous extra hydrogen bonds and strong electro-
static interaction between the IL cations and the POM anions.
X-ray diffraction (XRD) patterns for H5PV2Mo10O40, H4SiW12O40,
Scheme 1. Catalytic conversion of cellobiose to formic acid and levulinic
acid.
of cellobiose, subsequent oxygenation and rehydration leads
to selective CÀC bond cleavage of the sugar intermediates. As
a result, FA, formed by both oxygenation and rehydration, and
levulinic acid (LA), formed by rehydration, are found to be the
dominant products. Moreover, separation of the IL-POM cata-
lyst from the reaction mixture is achieved by simple filtration.
H6P2W18O62,
[PyBS]4SiW12O40,
[PyBS]5PV2Mo10O40,
and
[PyBS]6P2W18O62 were also collected and are shown in Figure 2.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2671

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
symmetry of the vanadium site. According to earlier results,[21]
there are five possible Keggin structures for [PV2Mo10O40]5À,
and they are identified by different degrees of rotation around
the XÀO axis of edge-sharing M3O13 groups in the Keggin
structure. The 51V-NMR chemical shifts of these five positional
isomers significantly vary with different pH values of the
medium in the range of 0–2, whereas they change smoothly
when the pH value is above 2. According to the abundance
and chemical shift (d3s) in a 51 V-NMR spectrum reported by
Pettersson et al.,[21] the chemical shift of ca. À532 ppm shown
in Figure 3 can be assigned to isomer a-1,2 and the chemical
Figure 2. XRD patterns of (a) [PyBS]4SiW12O40 (b) [PyBS]5PV2Mo10O40
(c) H5PV2Mo10O40 (d) H4SiW12O40 (e) H6P2W18O62 (f) [PyBS]6P2W18O62.
The measured XRD patterns of both Keggin-structured
H5PV2Mo10O40 and H4SiW12O40 and Dawson-structured
H6P2W18O62 agree well with diffraction patterns reported in lit-
erature.[18] Moreover, the diffraction peaks for these three
proton-type heteropoly acids are distinct and sharp, indicating
the high percentage of crystalline material (78%, 89%, and
82% for H5PV2Mo10O40, H4SiW12O40, and H6P2W18O62 respective-
ly). After the protons in Keggin-structured H5PV2Mo10O40 and
H4SiW12O40 were completely replaced by [PyBS]+ cations, char-
acteristic XRD peaks in the range 158~408 almost completely
disappeared, suggesting a low percentage of crystalline materi-
al (12% and 46% for [PyBS]4SiW12O40 and [PyBS]5PV2Mo10O40 re-
spectively). Introducing the organic [PyBS]+ cations greatly in-
fluences the crystallization of POM anions and then results in
the formation of a semi-amorphous structure. However, some
diffraction peaks in the range 58~108 emerged that were not
observed in the parent heteropoly acids (H5PV2Mo10O40 and
H4SiW12O40). These low-angle diffractions indicate the existence
of facets with large interplanar spacing, according to the Bragg
equation. Therefore, a unique inorganic–organic hybrid materi-
al with relatively large cavities is formed as Keggin structured
POM anions combine with organic [PyBS]+ cations. The
Dawson-structured POM based IL-POM ([PyBS]6P2W18O62) shows
clear XRD diffractions in the range of 158~408. The peaks in
the range of 5~108, not found in the XRD pattern of the
parent proton-typed Dawson heteropoly acids, are also clearly
observed.
Figure 3. 51V NMR spectrum of [PyBS]5PV2 Mo10O40 before and after reaction.
shift at ca. À534 ppm can be assigned to isomer a-1,6 and
isomer a-1,11. Moreover, these two chemical shifts are within
2 ppm of each other, which agrees well with the data reported
by Pettersson and co-workers. Quantitative analysis shows that
the molar ratio of isomer a-1,2 to isomer a-1,6/isomer a-1,11 is
approximately 1:1 for as-synthesized catalyst before the reac-
tion (Figure 3a).
In brief, the characterization reveals that Keggin- or Dawson-
type IL-POM catalysts synthesized in this study retain the struc-
tural features typical of POMs in the particular combination of
oxygen with metals and nonmetals when POM anions bond
with the organic IL cations, as confirmed by the preservation
of characteristic IR bands. However, various factors, such as su-
percritical electrostatic interactions between organic cations
and inorganic anions, hydrogen bonding, Van der Waals forces,
and others, greatly influence the crystallization behavior of the
IL-POM compounds in atypical XRD patterns.[20]
One-pot transformation of cellobiose over different IL-POM
catalysts
Cellobiose is a disaccharide that consists of two glucose mole-
cules linked by a glycosidic bond, and represents the simplest
model molecule for cellulose. Accordingly, cellobiose was
chosen as the model substrate to facilitate the evaluation of
different IL-POM catalysts. The catalytic performances of these
IL-POM catalysts with different IL cations and POM anions are
listed in Table 1. The blank experiment, in the absence of any
The vanadium-containing IL-POMs were further character-
ized by 51V-NMR spectrometry to gain information on the local
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2672

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
The structures of POM anions
may also influence the hydrolysis
of cellobiose and subsequent re-
actions. As compared to Keggin-
structured IL-POMs (Table 1, en-
tries 2–6), Dawson-structured IL-
POMs (entries 12–13) afford
higher cellobiose conversions
with of glucose and subsequent
transformations of glucose to
fructose, 5-HMF as primary prod-
ucts. The formation of LA and FA
over these Dawson-structured IL-
POM catalysts is significantly in-
hibited, implying superior cata-
lytic activity of Keggin-type
POMs to Dawson-type POMs in
the isomerization of glucose to
fructose and subsequent hydro-
Table 1. Catalytic performances of different IL-POM catalysts in cellobiose degradation in the presence of mo-
lecular dioxygen.[a]
Entry Catalyst
Conver- Selectivity [%]
sion [%] glucose sorbitol fructose levulinic
acid
formic
acid
5-
gaseous product
HMF (mainly CO2)
1[b]
2
3
4
5
6
7
8
–
7.8
83.6
88.8
90.1
90.5
98.9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3.6
0
0
0
0
7.8
2.1
0.6
0.2
5.9
3.0
17.7
15.9
16.8
3.6
3.7
0.6
9
39.7
52.0
32.5
[TEABS]3PW12O40
[MIMBS]3PW12O40
[PyBS]3PW12O40
[PyBS]3PMo12O40
[PyBS]4SiW12O40
[PyBS]5 PV2 Mo10O40 100
[PyBS]4HPV2 Mo10O40 99.7
[PyBS]3H2PV2 Mo10O40 98.8
47.1
45.2
44.1
41.3
36.9
6.6
11.6
13.9
39.1
34.4
82.4
79.0
56.8
47.0
30.7
17.6
16.0
10.4
11.7
10.0
3.1
4.5
5.9
13.4
11.4
3.1
25.6
29.2
35.8
32.1
40.8
46.3
40.7
34.8
31.5
38.9
10.2
8.5
5.1
5.9
7.3
6.6
8.2
26.1
24.1
23.6
9.3
10.8
2.0
1.6
0.5
0
2.5
3.1
2.2
2.4
1.1
0.2
3.2
5.0
3.1
0.8
1.7
0.1
0
9
10[c] [PyBS]5PV2 Mo10O40
11[d] [PyBS]5PV2 Mo10O40
89.1
88.2
12
13
14
15
[PyBS]6P2W18O62
[PyBS]6P2Mo18O62
H3PW12O40
100
100
95.0
100
34.6
1.8
0.4
1.0
2.3
2.6
0
21.5
[PyBS]HSO4
0
0.8
16[e] [PyBS]5PV2 Mo10O40
8.6
[a] Reaction condition: 0.125 g cellobiose,423 K, 3 MPa O2,catalyst amount=5 mol% (based on the mole of cel-
lobiose), 10 mL H2O, 3 h. The “0” value was defined as “not detected.” [b] Blank experiment. [c] Under 3 MPa H2.
[d] Under 3 MPa N2. [e] Reaction conditions: 0.1 g cellulose, 423 K, 3 MPa O2, 3 h, catalyst amount=5 wt%
(base on the total weight of mixture).
lytic
degradations.
Notably,
under similar reaction condi-
tions, the selectivity toward FA
sharply increases in the degrada-
tion of cellobiose catalyzed by
vanadium-containing IL-POMs,
catalyst, reveals that there is nearly no conversion of cellobiose
into valuable hydrolytic or oxidized products, such as saccha-
rides or carboxylic acids (entry 1), implying that acid and redox
sites on the catalyst are important in the catalytic transforma-
tion of cellobiose. Adding IL-POMs significantly promotes the
degradation of cellobiose, showing remarkably high conver-
sions in the range of 83.6%~100% under relatively mild reac-
tion conditions. CÀC bond cleavage of sugar intermediates is
demonstrated by the appearance of FA and LA as primary
products. These results demonstrate the multiple roles of IL-
POM in accelerating the transformation of cellobiose, including
the hydrolysis of cellobiose and CÀC bond cleavage of sugar
intermediates.
concurrent with a significant decline of selectivity toward glu-
cose and increase of selectivity of gaseous products (entries 7–
9). Introducing vanadium into the IL-POMs markedly promotes
the catalytic oxidation of sugar intermediates with molecular
oxygen after hydrolyzing cellobiose, leading to FA formation
and deep oxidation products. A stepwise Mars–van Krevelen-
type mechanism is proposed to be involved in such catalytic
oxidation reactions, in which the valence alternation of vanadi-
um between V5+ and V4+ in catalytic cycles is responsible for
the redox transformations of substrates.[23] In order to confirm
the hypothesis of oxidative transformation of cellobiose, two
comparative experiments on [PyBS]5PV2Mo10O40 were conduct-
ed by changing the reaction atmosphere from dioxygen to hy-
drogen and nitrogen (entries 10–11). No significant increase of
FA and a similar product distribution to that under IL-POMs
without vanadium (entry 2–6) were found in the absence of
molecular dioxygen as the oxidant for formaldehyde, which is
the first product due to the initial oxidative CÀC bond scis-
sion.[24] Furthermore, three vanadium-containing IL-POMs with
different exchange degrees between proton and [PyBS]+
afford dissimilar cellobiose conversion and selectivities toward
LA, FA, and gaseous product. As the molar ratio of cellobiose/
IL-POM is kept at a constant of 20:1 for all catalytic runs, the
amount of [PyBS]+ in the reaction mediums varied in the order
[PyBS]5PV2Mo10O40 >[PyBS]4HPV2Mo10O40 >[PyBS]3H2PV2Mo10O40.
This decrease in the concentration of [PyBS]+ substantially hin-
ders the reaction rates of cellobiose catalytic degradation and
reduces the yields of LA and FA, indicating that the protons
from the sulfonic groups in [PyBS]+ are more beneficial for
promoting cellobiose transformation via hydrolytic pathways
than that from parent heteropolyacids. A pure heteropolyacid
catalyst of H3PW12O40 is examined in the reaction (entry 14). A
According to the mechanism proposed for LA and FA forma-
tion via hydrolytic pathways,[10,22] the rehydration of 5-HMF af-
fords equimolar FA and LA simultaneously (i.e., n(Catoms of
FA)/n(Catoms of LA)=1:5). Evidence for the hydrolytic transfor-
mation of cellobiose in this study is the presence of trace
amounts of fructose and 5-HMF in the reaction products. In
the acidic environment, mainly formed by the sulfonic groups
in the IL cations, the catalytic hydrolysis of cellobiose to glu-
cose and the isomerization of glucose to fructose proceeds
more rapidly than in a neutral environment without catalyst,
consequently leading to almost complete conversion of cello-
biose and rehydration of 5-HMF to LA and FA. The strong
Brønsted acidic sites of these IL-POMs are responsible for their
catalytic activity in the hydrolytic degradation of cellobiose,
which is consistent with our previous results.[14] The influence
of the cation of IL-POM on the catalytic activity follows the se-
quence: [PyBS]>[MIMBS]>[TEABS] (entry 2–4), as evidenced
by the conversion of cellobiose and selectivity toward LA and
FA.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2673

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
small amount of 0.5% FA and 2.6% LA are found, implying
that the cation of IL-POM is responsible for the rehydration of
5-HMF to LA and FA. The IL without heteropolyanion is also
tested (entry 15). Surprisingly, 47.0% selectivity of glucose is
obtained and no other side-product is detected, implying that
the acid strength of the IL is insufficient to rehydrate 5-HMF to
form LA and FA.
the molar ratio of isomer a-1,2 to isomer a-1,6/isomer a-1,11
was 2:1 according to their relative intensities in the NMR spec-
trum shown in Figure 3b. The alternation of relative intensities
is probably due to the decreased pH of the reaction mixture as
LA and FA are accumulated during the reaction, which agrees
well with the conclusion proved by Pettersson and co-workers
that the relative abundance of these isomers changes when
part of them become monoprotonated owing to a decline of
the pH of the system.[21]
In order to explore the practical application of IL-POM cata-
lysts in the transformation of biomass into valuable chemicals,
the optimal catalyst [PyBS]5PV2Mo10O40 was employed in the
cellulose degradation at 423 K. Upon 3 MPa pressure and
5 wt% catalyst amount, 34.6% conversion is obtained with
considerable selectivities toward FA and LA (Entry 16). The
lower conversion compared with the cellobiose transformation
is due to the robust crystalline structure and water-insolubility
of cellulose.
Exploring the catalytic reaction mechanism
The catalytic degradation of cellobiose catalyzed by IL-POM
catalyst in the presence of molecular dioxygen comprises
a series of successive hydrolysis, isomerization, and oxidation
reactions. In order to probe the catalytic mechanism several
compounds including glucose, sorbitol, xylitol, erythritol, glyc-
erol and ethylene glycol, which are possible reaction inter-
mediates, were employed as substrates in the presence of
[PyBS]5PV2Mo10O40 under identical reaction conditions, and the
results are shown in Table 2. As compared to the glucose con-
Recovery and reusability of IL-POM catalysts
One of the outstanding advantages of as-synthesized IL-POMs
as catalysts for the one-pot transformation of cellobiose is the
convenient recovery of these catalysts. Because the IL-POMs
have relatively high melting points and low solubility in water,
they can recrystallize into the solid state, and then be easily re-
cycled for the next catalytic run. The results of a four-run recy-
cling test of [PyBS]5PV2Mo10O40 for the transformation of cello-
biose are shown in Figure 4. There is only a slight decline in
Table 2. Selectivity of final product in reactions of different sugar inter-
mediates.[a]
Entry Substrate
Selectivity [%]
fructose levulinic formic acid gaseous product
acid
(mainly CO2)
1[b]
2
Glucose
Glucose
Glucose
Sorbitol
5.7
4.8
1.8
0
0
0
47.1
50.8
5.4
0
0
0
9.6
27.3
48.1
65.8
61.2
49.5
67.8
99.1
36.8
17.1
44.5
34.2
38.8
50.5
32.2
0.9
3[c]
4
5
6
7[d]
8[d]
Xylitol
Erythritol
Glycerol
Ethylene glycol
0
0
0
0
[a] Reaction conditions:423 K, 3 h, 3 MPa O2, 10 mL H2O, 0.125 g sugars
intermediates, catalyst amount of [PyBS]5PV2Mo10O40 =3 mol% (based on
the mole of substrate), sugar intermediates conversions are all 100%. The
“0” value was defined as “not detected.” [b] The reaction was catalyzed
by [PyBS]4SiW12O40 instead of [PyBS]5PV2Mo10O40. [c] The reaction was cata-
lyzed by H5PV2 Mo10O40 instead of [PyBS]5PV2Mo10O40. [d] Conversions of
glycerol and ethylene glycol were 49.2% and 58.3%, respectively.
Figure 4. Cellobiose and cellulose conversion and selectivity for levulinic
acid and formic acid for four recycling runs on [PyBS]5PV2Mo10O40 (Reaction
conditions: 0.125 g cellobiose, 423 K, 3 MPa O2,catalyst amount=5 mol%
base on the mole of substrate, 10 mL H2O, 3 h).
version catalyzed by [PyBS]4SiW12O40 in the absence of vanadi-
um active sites incorporated in the POM anion (entry 1), the
selectivity toward FA increases sharply over the
[PyBS]5PV2Mo10O40 catalyst. The amount of gaseous products
significantly decreases while the LA yield only slightly changes
(entry 2). Glucose is converted to LA and FA with relatively
high selectivities of 50.8% and 27.3%, respectively. The oxida-
tive CÀC bond cleavage of glucose, which is the first product
derived from the hydrolysis of cellobiose, dominates the trans-
formation of cellobiose to FA and CO2 over the active species
of vanadium,[12,24] the FA/CO2 ratio in the products of glucose
degradation should be close to that for cellobiose degradation,
after subtracting the FA yielded from rehydration of 5-HMF.
The oxidative CÀC bond cleavage of organics in the presence
cellobiose conversion and FA selectivity after four successive
recycling tests. The stable performance of this catalyst is con-
sistent with the characterization results. Compared to the fresh
catalyst, the recovered one shows similar IR bands at 1087,
951, 904, and 791 cmÀ1 for the POM anion along with the
bands at 1170, 1631, and 1423 cmÀ1 for the IL-cation (curve e
in Figure 1). Results of 51VNMR spectroscopy on the spent cata-
lyst (Figure 3b) did not reveal any evidence of impurities and
degradation,
further
suggesting
the
durability
of
[PyBS]5PV2Mo10O40 during the catalytic cycle. After the reaction,
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2674

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
active structure. This is a two-
electron oxidation coupled with
one molecular oxygen transfer
and release of two protons.
According to this mechanism
and the study on oxidation of
glucose and C2 substrates to FA
and CO2 by Wasserscheid
et al.,[12] the ratios of FA/CO2 in
products for glucose, sorbitol,
xylitol, erythritol, glycerol, and
ethylene glycol should be 1:1,
2:1, 3:2, 1:1, 2:1, and 1:0, respec-
tively. In this study, the FA/CO2
ratios in all these substrates are
consistent with the proposed
molar ratios except glucose
transformation (entry 2). This ex-
ception for the glucose reaction
becomes more understandable
when taking into account that
part of FA coming from rehydra-
tion of 5-HMF. H5PV2Mo12O40 was
employed to confirm this hy-
pothesis because the rehydra-
tion of 5-HMF to LA and FA is
significantly inhibited in the ab-
sence of IL-cation. A 1:1 ratio of
FA/CO2 was obtained (entry 3),
which is similar to the result in
the glucose reaction reported by
Wasserscheid.
These
results
strongly support the validity of
degradation of polyols and sac-
charides via ET-OT oxidation.
5-HMF and LA cannot be
found in the transformation of
C2 ~C5 alcohols (sorbitol, xylitol,
erythritol, glycerol, and ethylene
glycol), which further supports
the standpoint that LA originates
from the rehydration of 5-HMF
via a series of hydrolysis and iso-
Scheme 2. Pathway for the oxidation of glucose to formic acid and carbon dioxide.
merization steps of saccharides.
Accordingly, a possible reaction
pathway for Brønsted-acid-cata-
of molecular dioxygen and vanadium-containing POMs has
been proposed to follow an electron–oxygen transfer (ET–OT)
mechanism,[12,23–25] as depicted in Scheme 2, for oxidative deg-
radation of glucose. Initially, a POM–substrate complex is
formed after electron and proton transfer from glucose to IL-
POM, and then this complex undergoes CÀC bond cleavage
after the oxygen transfer from the IL-POM to the substrate, fol-
lowed by an additional electron and proton transfer. Finally,
the ET-OT reactions lead to the formation of final products and
the reduced and deoxygenated IL-POM, which can be easily re-
oxidized by molecular dioxygen to restore their catalytically
lyzed rehydration of 5-HMF is proposed in Scheme 3.[11] The re-
hydration of 5-HMF undergoes a series of hydrolysis and iso-
merization steps, leading to a variety of intermediates, and fi-
nally forms LA and FA. It is worth noting that 1) LA and FA
molecules, which are all terminal carboxylic acids without
oxygen-containing functional group on the a-carbon atom, are
remarkably more stable than glucose, fructose, and 5-HMF in
the presence of vanadium-containing POM-IL catalyst and mo-
lecular dioxygen; 2) equimolar LA and FA, which corresponds
to a 5:1 ratio of n(C atoms of LA)/n(C atoms of FA), is always
obtained through hydrolysis and isomerization pathways.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2675

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Preparation of different POMs
and IL-POM catalysts
H6P2W18O62: Na2WO4·2H2O was first
dissolved in hot deionized water
and then mixed with a 85 wt%
H3PO4 solution under vigorous stir-
ring for 6 h at 413 K. After cooling
to room temperature, the resultant
mixture was acidified with an ade-
quate amount of 37 wt% hydro-
chloric acid solution. Product ex-
traction by using diethyl ether as
extraction solvent lasted for no
less than 5 h until the liquid was
obviously stratified. The bottom
part was then carefully separated
out and heated to remove the di-
ethyl ether solvent to yield the
yellow solid powder H6P2W18O62.
Scheme 3. Pathway for the rehydration of 5-HMF to formic acid and levulinic acid.
Further purification via aqueous
dissolution and distillation at 393 K
was also performed to remove hy-
drophilic residues.
Conclusions
H5PMo10V2O40: Following a procedure described in a previous litera-
ture,[15] 0.10 mol of MoO3 and 0.01 mol of V2O5 were mixed in
250 mL of deionized water under vigorous stirring at 393 K for 2 h,
prior to introducing a 85 wt% H3PO4 aqueous solution (0.01 mol).
After continuous stirring for 24 h, the resultant product was filtered
and dried at 323 K in vacuum for 24 h. The obtained orange-col-
ored fine powder H5PMo10V2O40 was further purified via an aque-
ous recrystallization procedure.
[PyBS]3PW12O40: 0.11 mol of prydine and 0.10 mol of 1,4-butane sul-
fone were dissolved in 20 mL of toluene and the mixture was con-
tinuously stirred for 24 h at 313 K under the protection of a nitro-
gen atmosphere. The obtained white precipitate (denoted as PyBS)
was then filtered, washed with diethyl ether three times and dried
in vacuum at room temperature. In a round-bottomed flask,
0.06 mol of dry PyBS was added into an aqueous solution contain-
ing 0.02 mol of H3PW12O40, followed by vigorous stirring at room
temperature for 24 h. Continuous evacuation at room temperature
was conducted to completely remove the water moisture and the
final product of [PyBS]3PW12O40 was carefully collected in a powder
form.
A series of multifunctional ionic-liquid-based polyoxometalate
(IL-POM) catalysts is synthesized and applied in the one-pot
transformation of cellobiose/cellulose to valuable chemicals,
such as formic acid (FA) and levulinic acid (LA). A vanadium-
containing POM-IL catalyst is found to give a significantly
higher selectivity of LA (46.3%) and FA (26.1%) than the other
catalysts, together with 100% conversion of cellobiose. All
these catalysts are easily recyclable without any significant loss
of activity, as demonstrated by a four-run recycling test. This
approach successfully combines two different reaction path-
ways: (i) the formation of LA and FA via acid-catalyzed hydroly-
sis and isomerization under the IL functional groups, and
(ii) the oxidative degradation of carbonoxygenates to FA and
CO2 under the active vanadium species, to provide a green
and efficient synthesis of LA and FA from biomass materials.
Other IL-POM catalysts with Keggin and Dawson structured POM
anions including [TEABS]3PW12O40, [ImBS]3PW12O40, [PyBS]3PW12O40,
[PyBS]3PMo12O40,
[PyBS]4HPV2Mo10O40,
[PyBS]5PV2Mo10O40,
[PyBS]4SiW12O40,
[PyBS]5PV2Mo10O40,
[PyBS]6P2W18O62,
[PyBS]5PV2Mo10O40,
[PyBS]3H2PV2Mo10O40,
Experimental Section
[PyBS]6P2Mo18O62
and
[PyBS]5PV2Mo10O40 were all synthesized following the procedure
similar to that for the [PyBS]3PW12O40 preparation.
Chemicals and characterization
All chemicals including pyridine (99%), triethylamine (99%), 1-
methylimidazole (99%), 1,4-butane sulfone (99%), silicotungstic
acid (AR), phosphotungstic acid (AR), and phosphomolybdic acid
(AR) were purchased from Sigma–Aldrich or Fluka and used with-
out further purification. FTIR measurements were performed on
a Nicolet 510P absorption spectrometer by dropping a sample
liquid onto a KBr window to form a film suitable for transmittance
spectrum collection. XRD patterns of different IL-POM catalysts
were collected on a Rigaku Ultima IV diffractometer using CuKa ra-
diation source (l=1.54056 ꢁ) at 40 kV and 20 mA with a scan rate
of 0.58minÀ1. 51 V-NMR spectra were recorded on a JEOL ECX-
400 MHz in D2O and were calibrated with VOCl3 as the internal
standard.
Catalytic evaluation of cellobiose conversion
Catalytic conversion of cellobiose was carried out in a stainless-
steel autoclave reactor with Teflon liner (Parr 4950, 4843 Control-
ler). Typically, a certain quantity of cellobiose substrate, IL-POM cat-
alyst, and deionized water were added into a 50 mL of Teflon con-
tainer. High-purity oxygen was introduced to reach an initial reac-
tion pressure of 3 MPa after expelling residue air by pressurizing
and releasing oxygen for three times. After the catalytic reaction
carried out at 423 K for 3 h, the autoclave containing unreacted re-
actants, products and catalysts was cooled in an ice-bath for no
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2676

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[4] S. Biella, L. Prati, M. Rossi, J. Catal. 2002, 206, 242–247.
less than 6 h until the used catalyst emerged at the bottom as the
transparent solid. Especially, for the catalyst with Dawson-struc-
tured POM anions, a cooling time for complete precipitation ex-
ceeded 10 h. The products as well as unreacted reactants were all
reserved in the liquid phase and could be conveniently separated
from the solid catalysts by decantation or filtration. The recycled
IL-POM catalyst was washed with ethanol for three times, dried in
vacuum at 333 K for 8 h prior to the next catalytic run. The quanti-
tative analysis of the products including glucose, sorbitol, fructose
and FA was carried out on an Agilent 1100 HPLC using a differential
refraction detector and was determined by an internal standard
method. A Hi-Plex H column (300ꢂ6.5 mm) was employed and the
[5] Y. ꢃnal, S. Schimpf, P. Claus, J. Catal. 2004, 223, 122–133.
[6] X. S. Tan, W. P. Deng, M. Liu, Q. H. Zhang, Y. Wang, Chem. Commun.
2009, 7179–7181.
[7] C. Baatz, U. Prꢄße, J. Catal. 2007, 249, 34–40.
[8] N. Yan, C. Zhao, C. Luo, P. J. Dyson, H. Liu, Y. Kou, J. Am. Chem. Soc.
2006, 128, 8714–8715.
[9] P. Wang, H. Yu, S. Zhan, S. Wang, Bioresour. Technol. 2011, 102, 4179–
4183.
[10] S. Saravanamurugan, O. Nguyen Van Buu, A. Riisager, ChemSusChem
2011, 4, 723–726.
[11] J. Zhang, X. Liu, M. N. Hedhili, Y. Zhu, Y. Han, ChemCatChem 2011, 3,
1294–1298.
[12] R. Wçlfel, N. Taccardi, A. Bçsmann, P. Wasserscheid, Green Chem. 2011,
13, 2759–2763.
[13] Y. Leng, J. Wang, D. R. Zhu, X. Q. Ren, H. Q. Ge, L. Shen, Angew. Chem.
Int. Ed. 2009, 48, 168–171; Angew. Chem. 2009, 121, 174–177.
[14] K. X. Li, L. Chen, H. L. Wang, W. B. Lin, Z. C. Yan, Appl. Catal. A 2011, 392,
233–237.
[15] Y. Leng, J. Wang, D. R. Zhu, L. Shen, P. P. Zhao, M. J. Zhang, Chem. Eng. J.
2011, 173, 620–626.
mobile phase was 0.01m H2SO4 buffer at
a flow rate of
0.3 mLminÀ1. The column temperature was 313 K. Gaseous prod-
ucts including CO2 and CO were collected into a gas pocket and
then analyzed on an Agilent 7890A GC with a TCD detector and
a TDX-01 column. Cellobiose conversion was determined by the
weight difference before and after the reaction.
[16] L. San Felices, P. Vitoria, J. M. Gutierrez-Zorrilla, S. Reinoso, J. Etxebarria,
L. Lezama, Chem. Eur. J. 2004, 10, 5138.
[17] A. M. Douvas, K. Yannakopoulou, P. Argitis, Chem. Mater. 2010, 22, 2730.
[18] K. Y. Lee, N. Mizuno, T. Okuhara, M. Misono, Bull. Chem. Soc. Jpn. 1989,
62, 1731.
[19] S. Bareyt, S. Piligkos, B. Hasenknopf, P. Gouzerh, E. Lacote, S. Thorimbert,
M. Malacria, J. Am. Chem. Soc. 2005, 127, 6788.
[20] X. Yan, P. Zhu, J. Fei, J. Li, Adv. Mater. 2010, 22, 1283–1287.
[21] L. Pettersson, I. Andersson, A. Selling, J. H. Grate, Inorg. Chem. 1994, 33,
982–993.
Acknowledgements
We acknowledge the National Environment Agency, the Environ-
ment Technology Research Program (ETRP, grant 1202-109), and
the Ministry of Education, Singapore, the Academic Research
Fund (AcRF Tier 1 grants, RG 48/12).
[22] W. Deng, Q. Zhang, Y. Wang, Dalton Trans. 2012, 41, 9817.
[23] R. Neumann, Inorg. Chem. 2010, 49, 3594–3601.
[24] A. M. Khenkin, R. Neumann, J. Am. Chem. Soc. 2008, 130, 14474–14476.
[25] A. M. Khenkin, G. Leitus, R. Neumann, J. Am. Chem. Soc. 2010, 132,
11446–11448.
Keywords: biomass · hydrolysis · polyoxometalates · reaction
mechanism · rehydration
[1] D. Klemm, B. Heublein, H. Fink, A. Bohn, Angew. Chem. Int. Ed. 2005, 44,
3358–3393; Angew. Chem. 2005, 117, 3422–3458.
[2] J. Zakzeski, B. M. Weckhuysen, ChemSusChem 2011, 4, 369.
[3] S. van de Vyver, J. Geboers, P. A. Jacobs, B. F. Sels, ChemCatChem 2011,
3, 82–94.
Received: March 10, 2014
Published online on August 8, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 2670 – 2677 2677