Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose to polyols on carbon-supported WO3 and Ru catalysts

Article abstract of DOI:10.1016/j.cattod.2015.09.056

Efficient conversion of cellulose, the most abundant biomass on Earth, to chemicals in high yields remains a formidable challenge. Here, we report the marked change in the distribution of polyol products in the cellulose reaction on Ru/C and WO₃/C, strongly depending on the competitive reactions of the glucose intermediate. WO₃ crystallites not only promote, as a solid acid, the efficient hydrolysis of cellulose to glucose, but also catalyze the selective cleavage of the C-C bonds in glucose and other C₆ sugar intermediates, leading to the formation of ethylene glycol and propylene glycol, in competition with the sugar hydrogenation to the corresponding C₆ polyols (e.g. sorbitol) on Ru/C. The basic C support, behaving similar to other solid bases (i.e. MgO), catalyzes the isomerization of glucose into fructose, leading to the favored formation of propylene glycol instead of ethylene glycol. Such strong dependence of the product distribution on the catalytic functions is clarified by the kinetic analysis of the three competitive reactions of glucose, including its hydrogenation, isomerization and C-C bond cleavage. Importantly, such kinetic analysis can predict the maximum selectivity ratio of propylene glycol to ethylene glycol, which is 2.5, for example, at 478 K under the reaction conditions in this work, corresponding to a maximum yield of propylene glycol of ～71%. These understandings shed new insights into the selective conversion of cellulose, which provides guidance for the rational design of catalyst functions and tuning of reaction parameters towards the controllable synthesis of specific products from cellulose.

Full text of DOI:10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
Catalysis Today xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Kinetic insight into the effect of the catalytic functions on selective
conversion of cellulose to polyols on carbon-supported
WO3 and Ru catalysts
Yue Liu, Haichao Liu∗
Beijing National Laboratory for Molecular Sciences, State Key Laboratory for Structural Chemistry of Stable and Unstable Species,
College of Chemistry and Molecular Engineering, Peking University, Beijing 100871, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Efficient conversion of cellulose, the most abundant biomass on Earth, to chemicals in high yields remains
a formidable challenge. Here, we report the marked change in the distribution of polyol products in the
cellulose reaction on Ru/C and WO3/C, strongly depending on the competitive reactions of the glucose
intermediate. WO3 crystallites not only promote, as a solid acid, the efficient hydrolysis of cellulose to
glucose, but also catalyze the selective cleavage of the C–C bonds in glucose and other C6 sugar interme-
diates, leading to the formation of ethylene glycol and propylene glycol, in competition with the sugar
hydrogenation to the corresponding C6 polyols (e.g. sorbitol) on Ru/C. The basic C support, behaving
similar to other solid bases (i.e. MgO), catalyzes the isomerization of glucose into fructose, leading to the
favored formation of propylene glycol instead of ethylene glycol. Such strong dependence of the product
distribution on the catalytic functions is clarified by the kinetic analysis of the three competitive reac-
tions of glucose, including its hydrogenation, isomerization and C–C bond cleavage. Importantly, such
kinetic analysis can predict the maximum selectivity ratio of propylene glycol to ethylene glycol, which
is 2.5, for example, at 478 K under the reaction conditions in this work, corresponding to a maximum
yield of propylene glycol of ∼71%. These understandings shed new insights into the selective conversion
of cellulose, which provides guidance for the rational design of catalyst functions and tuning of reaction
parameters towards the controllable synthesis of specific products from cellulose.
Received 5 July 2015
Received in revised form
10 September 2015
Accepted 22 September 2015
Available online xxx
Keywords:
Cellulose
Ethylene glycol
Propylene glycol
Sorbitol
Tungsten trioxide
Solid base
© 2015 Elsevier B.V. All rights reserved.
1. Introduction
hydrogenolysis to polyols [10–21]. The polyol products generally
have a wide distribution from C2 to C6 components, accord-
Cellulose is
a
naturally occurring biopolymer of glucose
ingly with low selectivity to specific products. To address this
problem, many efforts have been focusing on the exploration of
highly selective catalysts [11–22]. For example, Ni, Ru and Pt-
SnOx-based catalysts exhibit high activity for hydrogenation of
the glucose intermediate to hexitols with improved selectivity,
which would otherwise undergo degradation [3,4,11–16]. Zhang
and co-workers used Ni-W2C/C and H2WO4 to achieve efficient
conversion of cellulose to ethylene glycol [17,18]. By taking advan-
tage of the basicity of SnOx species, Deng and Liu achieved the
direct conversion of cellulose to acetol on Pt-SnOx/Al2O3 and Ni-
SnOx/Al2O3 catalysts [19]. Recently, we reported the selective
synthesis of sorbitol, ethylene glycol and propylene glycol in the
cellulose reaction by tuning the structure of WO3 domains and acid-
basicity of the underlying supports, and particularly in the presence
of basic carbon, the dominant formation of propylene glycol
[20].
with 1,4-␤-glycosidic linkages. As the most abundant non-edible
biomass in nature, cellulose is considered to be a viable alterna-
tive to fossil resources for the sustainable production of fuels and
chemicals [1–4]. Several efficient and green routes of cellulose con-
version to different chemicals or intermediates, such as monomeric
sugars [5–9], polyols [10–22], 5-hydroxymethylfurfural [23,24] and
organic acids [25–27], have been explored in recent years. The
routes to polyols have attracted particular attention due to their
high atom-economy and energy-efficiency as well as the versatile
uses of polyols as fine chemicals directly and as precursors for the
synthesis of fuels and especially value-added compounds [28–31].
Catalytic conversion of cellulose to polyols involves cellu-
lose hydrolysis to glucose and its subsequent hydrogenation or
Herein, we extend our preliminary work, and present a detailed
study on the effects of the catalytic functions from WO3 and basic
∗
Corresponding author. Tel.: +86 10 62754031; fax: +86 10 62754031.
E-mail address: hcliu@pku.edu.cn (H. Liu).
http://dx.doi.org/10.1016/j.cattod.2015.09.056
0920-5861/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
2
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
carbon on the distribution of polyol products, especially propylene
glycol and ethylene glycol, in the cellulose reaction. We also exam-
ine the effects of other solid bases, such as Mg-Al hydrotalcite and
MgO, on the C–C bond cleavage of the glucose intermediate. These
effects, together with the discussion on the reaction pathways and
kinetics for the cellulose conversion to polyols, lead to the in-depth
understanding on the cellulose conversion to ethylene glycol and
especially propylene glycol.
2.3. Cellulose and sugar reactions
Cellulose (microcrystalline, Alfa Aesar) reactions were carried
out in 40 mL water in a stainless steel autoclave (100 mL), typically
at 478 K and 6 MPa H2 for 30 min with vigorous agitation. Cellulose
conversions were determined according to the difference in the
weight of cellulose before and after the reactions. The products (e.g.
polyols) were analyzed by high-performance liquid chromatogra-
phy (Shimadzu LC-20A) equipped with an RID detector (RID-20A),
using a Bio-rad HPX-87C column. Gas products were analyzed by GC
(Shimadzu 2010GC) using an HJ-OV-101 column connected to an
FID detector. Product selectivity was calculated on a carbon basis.
2. Experimental
2.1. Catalyst synthesis
3. Results and discussion
Activated carbon (C) supported Ru catalyst (Ru/C) was prepared
by incipient wet impregnation of typically 1.0 g carbon (AR, Beijing
Dali Fine Chemical) with 5.0 mL aqueous solution of RuCl3 (0.03 g
Ru). The impregnated sample was dried at 383 K for 2 h in air and
then reduced in a H2/N2 (1/4) flow (at a rate of 50 mL/min) at
623 K for 5 h. The as-synthesized 3 wt% Ru/C catalyst was kept in
a desiccator before its use.
C-supported WO3 catalysts (WO3/C) were prepared by impreg-
nation of 1.0 g carbon with 5 mL aqueous solution of ammonium
metatungstate (AR, Sinopharm Chemical). Then the samples were
dried at 383 K for 2 h in air and calcined in a N2 flow (50 mL/min)
at 773 K for 3 h. By varying the concentrations of ammonium
metatungstate, WO3/C catalysts of different loadings in the range
0.5–20 wt% were obtained, which were denoted as 0.5 wt% WO3/C,
1.0 wt% WO3/C, 2.0 wt% WO3/C, 6.0 wt% WO3/C, 12 wt% WO3/C and
20 wt% WO3/C. Following the same procedure, Al2O3-supported
WO3 with a 50 wt% loading was prepared for comparison, denoted
as 50 wt%WO3/Al2O3.
3.1. Effect of catalyst components on cellulose conversion to
polyols
Table 1 shows the cellulose conversion and polyol selectiv-
ity on Ru/C as a function of the amount of 6.0 wt% WO3/C. At
478K and 6 MPa H2, cellulose conversion was 12.5% after 30 min
in the absence of 6.0 wt%WO3/C, as reported previously [20]. The
major products were hexitols with a combined selectivity of 64.7%,
including sorbitol (54.4%) and mannitol (10.3%). Other products
included ethylene glycol (7.5%), propylene glycol (3.3%), glycerol
(3.0%), tetritols (3.2%) and pentitols (10.3%). Upon addition of 0.05 g
6.0 wt%WO3/C, as a result of the WO3 acidity, cellulose conver-
sion increased from 12.5% to 17.6%. The selectivities to ethylene
glycol, propylene glycol and glycerol increased to 20.6%, 6.0% and
6.5%, respectively, with the concurrent decline of the selectivities
to pentitols and hexitols to 2.2% and 46.5%. Further increase in the
amount of 6.0 wt%WO3/C led to a gradual increase in the cellu-
lose conversion to the maximum value of ca. 25% and evolution
of the major products from hexitols to ethylene glycol and pro-
pylene glycol. For example, at the largest amount of 6.0 wt%WO3/C
(i.e. 0.35 g), the selectivities to ethylene glycol and propylene glycol
reached 58.0% and 14.7%, respectively, while the combined selec-
tivity to hexitols dramatically declined to 3.1%. Such change in the
conversions and product distributions reflects the bifunctional role
of WO3 in promoting the cellulose hydrolysis to the sugar interme-
diates and subsequent cleavage of their C–C bonds to ultimately
form ethylene glycol and propylene glycol [20].
The cellulose conversion and product distributions also depend
on the WO3 loadings on C support, as discussed below, due origi-
nally to the corresponding change in the amount of basic C support.
As shown in Fig. 1, at a given amount of WO3 (0.016 g), with increas-
ing WO3 loadings from 0.5 wt% to 2.0 wt%, the cellulose conversion
increased slightly from 20.1% to 25.0%, and then did not change
at the higher WO3 loadings up to 20 wt%. In contrast, the product
distributions changed significantly with the WO3 loadings. As the
WO3 loading increased from 0.5 wt% to 6.0 wt%, the ethylene glycol
selectivity increased largely from 25.7% to 55.8% while the propyl-
ene glycol selectivity decreased from 37.6% to 12.3%. By further
increasing the WO3 loadings to 20 wt%, their selectivities remained
essentially constant. Notwithstanding such changes, the selectivi-
ties to hexitols were always below 10% while the selectivities for
the lower carbon (C2 + C3) polyols were above 60%, indicating that
the activities for the C–C bond cleavage are independent of the
WO3 loadings on the C support. By referring to our previous report
that the C–C bond cleavage occurs on the WO3 crystallites while
the dispersed WO3 domains behave only as acid to promote the
cellulose hydrolysis [20], we tentatively conclude that WO3 exists
dominantly in the form of crystallites on the WO3/C catalysts in the
whole range of the WO3 loadings (0.5–20 wt%), as characterized by
XRD (Fig. 2).
2.2. Catalyst characterization
The BET surface areas of catalysts were measured on
a
Micromeritics ASAP 2010 BET Surface Area Analyzer by N2 physi-
cal adsorption at 77 K. Samples were outgassed at 473 K in vacuum
before measurement.
X-ray diffraction (XRD) profiles of supported WO3 cata-
lysts were taken on a Rigaku D/Max-2000 X-ray diffractometer
˚
using Cu K␣1 radiation (ꢀ = 1.5406 A), operated at 40 kV and
100 mA, in the 2ꢁ range of 10–60◦ at
a
scanning rate of
4◦/min. The crystal sizes of WO3 species in the samples were
calculated by the Scherrer equation: d = 0.90ꢀ/(ˇ cos ꢁ), where
ꢁ
is the diffraction angle and ˇ is the full width at half-
maximum (FWHM) of given diffraction peaks, typically an average
value calculated from peaks of (0 0 2), (0 2 0) and (2 0 0) lattice
planes.
Temperature programmed desorption (TPD) of CO2 was per-
formed on a Quantachrome ChemBET-Pulsar TPR/TPD system.
Before measurement, catalyst was pretreated at 500 ◦C in a N2 flow
(30 mL/min) for 30 min. Adsorption of CO2 was done at 100 ◦C in a
CO2 flow (30 mL/min) for 15 min, followed by purging in N2 flow
for 30 min. Then the temperature was raised at a ramping rate of
10 ◦C/min, and desorbed CO2 was detected on a TCD detector. Cal-
ibration was done using CO2 pulses with varying volumes from 20
to 100 ␮L.
Point of zero charge (PZC) of catalysts was determined using
mass titration method [32]. Typically, sample powders were grad-
ually added into 40 mL deionized water under agitation at 400 rpm
until the pH of the suspension reached constant, which was moni-
tored by a Metller Toledo Delta 320 pH Meter equipped with LE438
electrodes. The final constant pH was just the PZC of the given
catalyst sample.
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
3
Table 1
Cellulose conversion and product selelctivity on Ru/C and 6 wt% WO3/C.a
6.0 wt%WO3/C
(g)
Conversion
(%)
Selectivity (%)
Ethylene
glycol
C2
Propylene
glycol
C3
Glycerol
C3
Tetritolsb
C4
Pentitolsc
C5
Mannitol
C6
Sorbitol
C6
0
12.5
17.6
20.5
22.5
24.6
24.8
25.1
7.5
20.6
34.6
47.0
53.2
55.8
58.0
3.3
6.0
6.8
3.0
6.5
4.9
3.3
3.0
2.7
2.6
3.2
5.6
7.1
6.8
4.5
4.9
3.2
10.3
2.2
1.2
0.7
0.1
0.0
0.1
10.3
5.5
4.2
3.1
1.5
1.3
0.8
54.4
41.0
30.1
13.8
7.2
0.05
0.10
0.15
0.20
0.27
0.35
8.4
12.0
12.3
14.7
5.3
2.3
a
Reaction condition: 478 K, 6 MPa H2, 30 min, 40 mL H2O, 1.0 g cellulose, 0.02 g 3 wt% Ru/C.
Erythritol and threitol.
Xylitol, and arabitol and ribitol.
b
c
Fig. 2. XRD profiles of supported WO3 catalysts. (1) 50 wt% WO3/Al2O3, (2) C, (3)
0.5 wt% WO3/C, (4) 1.0 wt% WO3/C, (5) 2.0 wt% WO3/C, (6) 6.0 wt% WO3/C, (7) 12
wt% WO3/C, (8) 20 wt% WO3/C.
Fig. 1. Cellulose conversion and product selectivity as a function of WO3 loadings
on C. Reaction condition: 478 K, 6 MPa H2, 30 min, 40 mL H2O, 1.0 g cellulose, 0.02 g
3 wt% Ru/C, 0.016 g WO3.
diffraction peaks is only due to the small number of the crystalline
particles and thus their low diffraction intensities.
The structure of the WO3 species on supported WO3 cata-
lysts depends on the identity of the supports and their surface
densities. WO3 is highly dispersed on the oxide surfaces such as
Al2O3 and ZrO2 below the theoretical monolayer coverage (ca.
5 W/nm2). Above this coverage, WO3 tends to aggregate into crys-
tallites [33,34]. However, C as support can hardly disperse WO3.
Alverez-Merino and co-workers found that WO3 forms crystallites
on C even at a surface density of as low as 0.4 W/nm2 [35]. To under-
stand the structure of our WO3/C catalysts, we characterized them
by XRD. As shown in Fig. 2, 50 wt% WO3/Al2O3 (7.2 W/nm2) exhibits
broad diffraction peaks of typical crystalline m-WO3 at 23◦, 34◦
with an estimated size of 13 nm. On C, when WO3 loadings are
below 6.0 wt% (0.2 W/nm2), only diffraction peaks of C are observed.
Weak diffraction peaks of m-WO3 appear at the loading of 6.0 wt%.
With increasing the WO3 loading, m-WO3 diffraction peak becomes
more intense while the half-peak width remains unchanged. This
reveals that increasing the WO3 loading on C support leads to an
increase in the number of WO3 crystalline particles but without
significant effect on the particle size. We infer that WO3 exists in
the form of crystallites on the WO3/C catalysts even at low load-
ings (0.5–2.0 wt%), and the absence of the corresponding m-WO3
It has been reported that in the cellulose reaction, ethylene
glycol and propylene glycol are derived from glucose and fruc-
tose intermediates, respectively, through the selective cleavage of
the C–C bonds in the sugar molecules on WO3 and subsequent
hydrogenation of the corresponding degraded intermediates on
Ru/C [20]. This reaction mechanism led to our proposition that
the observed higher propylene glycol selectivities at the lower
WO3 loadings for the WO3/C catalysts (Fig. 1) is due to their
larger available C surfaces when the total amount of WO3 loaded
in the cellulose reaction was kept constant, thus providing more
basic sites for the isomerization of glucose to fructose, a known
base-catalyzed reaction [36,37]. The basicity of the C support was
characterized by CO2-TPD and point of zero charge (PZC). Its surface
basic site density was 11 ␮mol/g, and PZC in water under ambient
conditions was 8.1. Such basic C support is capable to catalyze the
glucose isomerization into fructose. As shown in Fig. 3, the rate for
the isomerization glucose into fructose in pure water at 393 K was
0.15 mg/(mL min), which increased to 0.44 mg/(mL min) by about 3
folds upon addition of 1% equivalent amount of C. However, treat-
ment of the C support with 0.5% HNO3 for 2 h led to neutralization
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
4
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
below 20%, and the concurrent increase in the selectivities to the
C3 polyols from 18.1% to around 35%. These results reflect the favor-
able formation of the C3 polyols, especially propylene glycol, on the
catalysts containing both the basic component for the isomeriza-
tion of glucose into fructose and the WO3 component for the C–C
bond cleavage in the glucose and cellulose conversions.
3.2. Kinetic analysis of effect of catalyst components
Based on the above results, together with the previous under-
standing [20], the cellulose reaction pathways are tentatively
illustrated in Scheme 1. Cellulose is firstly hydrolyzed to the glu-
cose intermediate with the acidity of WO3. Glucose can be directly
hydrogenated to hexitols on Ru/C or break its C–C bond on WO3
to glycolaldehyde which is subsequently hydrogenated to ethyl-
ene glycol. In the presence of solid bases, glucose isomerizes into
fructose. Fructose can be then hydrogenated to hexitols on Ru/C
or undergo degradation to form glyceraldehyde and dihydroxyace-
tone on WO3 which are sequentially dehydrated and hydrogenated
to propylene glycol. Therefore, the selectivities of hexitols, ethylene
glycol and propylene glycol depend sensitively on the competitive
reactions of the glucose intermediate including its hydrogenation,
C–C bond cleavage and isomerization, catalyzed respectively by
Ru/C, WO3 and bases. Such dependence can be seen more clearly
from the kinetics of the three glucose reactions.
Fig. 3. Formation rate of fructose in glucose isomerization in the absence and pres-
ence of basic C and HNO3-treated C. Reaction condition: 393 K, 20 mg/mL glucose,
0.02 mg/mL C or HNO3-treated C.
For the glucose hydrogenation, several kinetic models have
been reported depending on the catalysts and reaction conditions
[38–41]. By referring to the previous studies, we propose that the
reaction in this work follows the Langmuir–Hinshelwood mecha-
nism and is limited by the surface reaction between the H atom
from the dissociated H2 and glucose competitively adsorbed on
Ru/C. Then the glucose hydrogenation rate, rHydrogenation, can be
expressed as:
rHydrogenation = k1mRuꢁH2 ꢁGlu-Ru
(1)
in which k1 is the rate constant for glucose hydrogenation on Ru,
mRu is the amount of Ru loaded for the reaction, ꢁH and ꢁGlu-Ru are
the coverages of H atoms and glucose. For the C–C bond cleavage
of glucose and its isomerization, the corresponding reaction rates
rCleavage and rIsomerization can be expressed as
rCleavage = k2mWO ꢁGlu-W
(2)
(3)
3
rIsomerization = kimBꢁGlu-B
Fig. 4. Polyol selectivity in glucose reaction on Ru/C in the presence of WO3 at pH of
7–9 (adjusted using buffer solutions of Na2HPO4 and NaH2PO4) and of WO3/Al2O3
with solid bases (MgO and Mg-Al hydrotalcite). Reaction condition: 478 K, 6 MPa
H2, 10 min, 40 mL H2O, 0.10 g glucose, 0.02 g 3 wt% Ru/C, 1.0 g WO3 or 0.08 g 50 wt%
WO3/Al2O3, 0.02 g Hydrotalcite (HT) or MgO. Glucose conversion 100%.
where k2 and ki are the rate constants for C–C bond cleavage of
glucose on WO3 and its isomerization on a solid base, respectively,
mWO and mB are the weight of WO3 and solid base, respectively,
3
and ꢁGlu-W and ꢁGlu-B are the glucose coverage on WO3 and solid
bases, respectively. Thus, the relative rate for the three glucose
reactions is expressed as
of its basic sites (as partly reflected by its lower PZC of 4.4) and
consequently the negligible isomerization rate (0.01 mg/(mL min)),
which unequivocally confirmed the role of the basicity on the C
support in catalyzing the isomerization of glucose into fructose.
To further understand the role of the basicity of C, several other
basic catalysts were examined in the conversion of glucose to poly-
ols in the presence of Ru/C and WO3. As shown in Fig. 4, with
increasing the pH values of the reaction solutions from 7 to 9,
achieved by using buffer solutions of Na2HPO4 and NaH2PO4, the
selectivities to ethylene glycol gradually declined from 47.3% to
21.9%, concurrently with the increase in the selectivities to the C3
polyols (propylene glycol + glycerol) from 14.0% to 29.0%. Using 50
wt%WO3/Al2O3 instead of unsupported WO3, the effects of basic-
ity on the product distributions were similar. In the presence of 50
wt% WO3/Al2O3, addition of solid bases, hydrotalcite and MgO, led
to the decrease in the ethylene glycol selectivities from 45.0% to
rHydrogenation : rCleavage : rIsomerization
= k1mRuꢁH2 ꢁGlu-Ru : k2mWO ꢁGlu-W : kimBꢁGlu-B
(4)
3
which indicates that the glucose reactions can be controlled by
modulating the relative amounts of the three catalyst components.
Next, we discuss the aforementioned effects of the relative amounts
of WO3 and C in Table 1 and Fig. 1, following these kinetic consid-
erations.
The rate ratio of the C–C bond cleavage of glucose to its hydro-
genation (rCleavage/rHydrogenation) represents the selectivity ratio of
ethylene glycol to hexitols (SC /SC6 ). Based on Eq. (4), the increase
in the WO3 amount in Table 21 would lead to an increase in the
SC /SC ratio, due to the change in both catalyst amount and sur-
2
6
face coverage. This is shown in Fig. 5a. Under the conditions in this
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
5
Scheme 1. Proposed reaction pathways of cellulose conversion to polyols.
rC = k3mRuꢁH2 ꢁGlycol
work, the coverage of H atom, glucose and glycoladehyde on Ru/C
and the coverage of glucose on WO3 are expressed as:
2
KHPHKGlycolCGlycol
ꢀ
= k3mRu
(11)
ꢁ
ꢂ
3
ꢀ
KHPH
1 + KGlu−RuCGlu + KGlycolCGlycol
+
KHPH
ꢁH
=
(5)
(6)
ꢀ
1 + KGlu−RuCGlu + KGlycolCGlycol
+
KHPH
in which, k3 is the rate constant for glycolaldehyde hydrogenation
on Ru. Considering the glucose C–C bond cleavage as the rate deter-
mining step for ethylene glycol formation, rCleavage should be equal
ꢀ
KGlu−RuCGlu
ꢁGlu−Ru
=
ꢀ
1 + KGlu−RuCGlu + KGlycolCGlycol
+
KHPH
to rC
:
2
ꢀ
KGlycolCGlycol
rCleavage = rC
(12)
2
ꢁGlycol
=
(7)
(8)
ꢀ
1 + KGlu−RuCGlu + KGlycolCGlycol
+
KHPH
Solving Eqs. (9)–(12), the selectivity ratio SC /SC is finally
2
6
expressed as
K
Glu-WCGlu
ꢁGlu-W
=
1 + KGlu-WCGlu
SC
SC
rCleavage
A
2
6
=
=
ꢁ
ꢂ
rHydrogenation
mWO
where KH, KGlu-Ru and KGlycol are the adsorption equilibrium
constants of H2, glucose and glycolaldehyde on Ru/C, respectively,
and KGlu-W is the adsorption equilibrium constant of glucose on
WO3 while PH, cGlu and cGlycol are the H2 pressure and the concen-
trations of glucose and glycolaldehyde in the reaction. Thus, the
reaction rates of glucose hydrogenation to hexitols (rHydrogenation),
glucose cleavage to glycolaldehyde (rCleavage) and glycolaldehyde
hydrogenation to ethylene glycol (rC2 ) are expressed as:
3
B ·
−1 − 1
mRu
ꢁ
ꢂ
ꢀ
(13)
k1 1 + 3KGlu-RuCGlu + 3 KHPH
3k3KGlu-RuCGlu
A =
B =
k3KHPH(1 + KGlu-WCGlu
3k3KGlu-WCGlu
)
rHydrogenation = k1mRuꢁH2 ꢁGlu-Ru
The derivation details of this Eq. (13) are included in Supplemen-
tary materials (Derivation 1). This Eq. (13) fits well the experimental
data points in Fig. 5a, as represented by the black curve.
For the selectivities to C3 polyols (propylene glycol and glycerol)
and ethylene glycol, their ratio (SC /SC2 ) can be represented by the
rate ratio of the C–C bond cleavag3e of glucose to its isomerization
(rCleavage/rIsomerization). In this case, the competitive adsorption of
fructose with glucose on WO3 must be considered because fructose
KHPHKGlu-RuCGlu
= k1mRu
(9)
ꢁ
ꢂ
3
ꢀ
1 + KGlu−RuCGlu + KGlycolCGlycol
+
KHPH
K
Glu-WCGlu
rCleavage = k2mWO ꢁGlu-W = k2mWO ꢁGlu-W
=
(10)
3
3
1 + KGlu-WCGlu
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
6
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
cleavage to glycolaldehyde (rCleavage = rC2 ) and fructose cleavage to
glyceraldehyde and dihydroxyacetone (rC3 ) are expressed as
KGlu-BCGlu
rIsomerization = kimBꢁGlu−B = kimB
1 + KGlu-BCGlu
(17)
rC = rCleavage = k2mWO ꢁGlu−W
2
3
K
Glu-WCGlu
= k2mWO
(18)
(19)
3 1 + KGlu-WCGlu + KFru-WCFru
KFru-WCFru
rC = k4mWO ꢁFru−W = k4mWO
3 1 + KGlu-WCGlu + KFru-WCFru
3
3
in which, k4 is the rate constant for C–C bond cleavage of fructose
on WO3. If the isomerization of glucose is the rate determining step
for C3 polyol formation, rIsomerization should be equal to rC3
:
rIsomerization = rC
(20)
3
Solving Eqs. (17)–(20), the selectivity ratio SC /SC (derivation
3
2
details are included in Supplementary materials, see Derivation 2)
is finally expressed as
SC
SC
rC
rC
k4 1 + KGlu-WCGlu
(
)
3
3
=
=
k2KGlu-WCGlu
2
2
ꢃ
ꢄ−1
mWO
mB
k4 1 + KGlu-BCGlu
(
)
3
·
·
− 1
(21)
kiKGlu-BCGlu
Considering the glucose as a highly reactive intermediate under
reaction condition, its concentration must be very low. Then Eq.
(21) is further simplified as
SC
SC
kiKGlu-B
mB
3
=
·
(22)
k2KGlu-W mWO
2
3
This indicates a linear relationship between the SC /SC ratio
2
and the amount of the basic catalyst, i.e. basic C support3, as indeed
observed in Fig. 5b (plotted on the basis of the results in Fig. 1). The
intercept on y-axis is not 0 but 0.2, which means that the isomeriza-
tion of glucose into glucose can occur under the reaction conditions
even in the absence of any bases, notwithstanding its very low rate.
Such effect of the basicity on the SC /SC ratio is also observed in
3
2
the glucose reaction on 50 wt%WO3/Al2O3 using MgO as a basic
catalyst. As shown in Fig. 6, the SC /SC ratio increased linearly
3
2
with the amount of MgO below 0.02 g. However, when the MgO
amount was above 0.02 g, the ratios deviated from the linearity and
finally reached a constant value of around 2.5 (at 0.05 g MgO). Such
constant value was also obtained when using abundant amount
of other solid bases including Mg-Al hydrotalcite (Mg/Al = 2), ZnO,
Fe2O3 and basic C instead of MgO. We propose that this phe-
nomenon is due to the rapid equilibrium for the isomerization of
glucose into fructose in the presence of the excessive amount of
basic catalyst, which led to the change in the rate-determining step
of the C3 polyol formation from the glucose isomerization to the
cleavage of the fructose C–C bond. In this case, based on Eqs. (18)
and (19), the SC /SC ratio is expressed as
Fig. 5. (a) Selectivity ratio of ethylene glycol to hexitols (SC /SC6 ) as a function of
the amount of 6.0 wt%WO3/C. Reaction condition: 478 K, 6 MPa H2, 30 min, 40 mL
H2O, 1.0 g cellulose, 0.02 g 3 wt% Ru/C. (b) Selectivity ratio of C3 polyols to ethylene
2
glycol (SC /SC2 ) as a function of the amount of C support. Reaction condition: 478 K,
3
6 MPa H2, 30 min, 40 mL H2O, 1.0 g cellulose, 0.02 g 3 wt% Ru/C, 0.06 g WO3.
is significantly formed in the presence of base catalyst. Therefore
the coverage of glucose on WO3 and on solid base, and the coverage
of fructose on WO3 (ꢁFru-W) are expressed as
3
2
KGlu−WCGlu
ꢁGlu−W
ꢁGlu−B
ꢁFru−W
=
(14)
(15)
(16)
SC
SC
rC
rC
k4mWO KFruu-WCFru
k4KFruu-W
k2KGlu-W
3
3
3
1 + KGlu−WCGlu + KFru−WCFru
=
=
=
(23)
k2mWO KGlu-WCGlu
2
3
3
KGlu−BCGlu
1 + KGlu−BCGlu
=
where Ki is the thermodynamic equilibrium constant between glu-
cose and fructose. From this equation, the SC /SC ratio depends
3
2
only on the relative rates for the C–C bond cleavage of fructose
and glucose (k4KFru-W/k2KGlu-W) and the thermodynamic equilib-
rium constant between them (Ki). At 478 K, Ki was calculated to be
KFru−WCFru
1 + KGlu−WCGlu + KFru−WCFru
=
where the KFru-W is the adsorption equilibrium constant of fructose
on WO3, and cFru represents fructose concentration. Thus the reac-
tion rates of glucose isomerization to fructose (rIsomerization), glucose
3.2 0.9 [42–44]. In this work, the SC /SC ratio was stabilized at 2.5
2
(Fig. 6). Therefore, the k4KFru-W/k2K 3
ratio is approximately to
Glu-W
be 0.8 0.3, indicating that glucose and fructose undergo the C–C
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
7
Fig. 6. Selectivity ratio of C3 polyols to ethylene glycol as a function of MgO with varying amounts and of Hydrotalcite, ZnO, Fe2O3 and C with excess amounts. Reaction
condition: 478 K, 6 MPa H2, 10 min, 40 mL H2O, 0.10 g glucose, 0.02 g 3 wt% Ru/C, 0.08 g 50 wt%WO3/Al2O3, 0.10 g HT, 0.30 g ZnO, 0.50 g Fe2O3, 5.0 g.
bond cleavage on WO3 at a similar rate. We have reported previ-
ously that the cleavage of C–C bond in glucose and fructose occurs
selectively at the ␤-position to their carbonyl groups [20]. Here, we
found that they have comparable cleavage rates, implying that they
have similar adsorption conformations on the WO3 surface and
react through the same mechanism as a result of their analogous
molecular structures. Notably, based on the maximum SC /SC ratio
of ∼2.5, the maximum yield of propylene glycol directly from the
cellulose and glucose reactions is estimated to be as high as ∼71% at
478 K, which can be apparently achieved upon further optimization
of the reaction parameters and catalysts.
and the National Natural Science Foundation of China (nos.
21433001, 21173008 and 21373019).
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2015.09.
056.
3
2
References
[1] G.W. Huber, S. Iborra, A. Corma, Chem. Rev. 106 (2006) 4044–4098.
[2] P. Gallezot, Catal. Today 121 (2007) 76–91.
[3] S. Van de Vyver, J. Geboers, P.A. Jacobs, B.F. Sels, ChemCatChem 3 (2011)
82–94.
4. Conclusions
Cellulose conversion to sorbitol, propylene glycol and ethyl-
ene glycol strongly depends on the relative amounts of Ru, WO3
and solid bases, as a result of their effect on the major parallel
reactions of glucose intermediate. WO3 crystallites can efficiently
improve cellulose hydrolysis to glucose and selectively cleave the
C–C bonds of glucose and fructose to ethylene glycol and propyl-
ene glycol at the expense of sorbitol and other C6 polyols. Solid
bases, such as the C support, catalyze isomerization of glucose into
fructose, which consequently facilitates the formation of propylene
glycol over ethylene glycol. Such effects of the catalyst functions
are consistent with the kinetic effects of the glucose reactions of
hydrogenation, isomerization and C–C bond cleavage. Kinetic anal-
ysis predicts that the maximum selectivity ratio of propylene glycol
(assuming complete conversion of glycerol to propylene glycol) to
ethylene glycol is 2.5 at 478 K under the reaction conditions in this
work, mainly limited by the thermodynamic equilibrium between
glucose and fructose, corresponding to a maximum yield of pro-
pylene glycol of ∼71%. These results would be helpful for rational
design of multi-functional catalysts and tuning of reaction parame-
ters for controlling the conversion of cellulose to specific products.
[4] H. Kobayshi, A. Fukuoka, Green Chem. 15 (2013) 1740–1763.
[5] R. Rinaldi, R. Palkovits, F. Schüth, Angew. Chem. Int. Ed. 47 (2008) 8047–8050.
[6] A. Onda, T. Ochi, K. Yanagisawa, Green Chem. 10 (2008) 1033–1037.
[7] S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi, H. Kato, S. Hayashi, M.
Hara, J. Am. Chem. Soc. 130 (2008) 12787–12793.
[8] S. Vyver, L. Peng, J. Geboers, H. Schepers, F. Clippel, C.J. Gommes, B. Goderis,
P.A. Jacobs, B.F. Sels, Green Chem. 12 (2010) 1560–1563.
[9] P. Lanzafame, D.M. Temi, S. Perathoner, A.N. Spadaro, G. Centi, Catal. Today
179 (2012) 178–184.
[10] A. Fukuoka, P.L. Dhepe, Angew. Chem. Int. Ed. 45 (2006) 5161–5163.
[11] C. Luo, S. Wang, H. Liu, Angew. Chem. Int. Ed. 46 (2007) 7636–7639.
[12] W. Deng, X. Tan, W. Fang, Q. Zhang, Y. Wang, Catal. Lett. 133 (2009) 167–174.
[13] J. Geboers, S. Van de Vyver, K. Carpentier, P.A. Jacobs, B.F. Sels, Chem.
Commun. 47 (2011) 5590–5592.
[14] M. Liu, W. Deng, Q. Zhang, Y.L. Wang, Y. Wang, Chem. Commun. 47 (2011)
9717–9719.
[15] G. Liang, H. Cheng, W. Li, L. He, Y. Yu, F. Zhao, Green Chem. 14 (2012)
2146–2149.
[16] T. Deng, H. Liu, Green Chem. 15 (2013) 116–124.
[17] N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J.G. Chen, Angew. Chem.
Int. Ed. 47 (2008) 8510–8513.
[18] Z. Tai, J. Zhang, A. Wang, M. Zheng, T. Zhang, Chem. Commun. 48 (2012)
7052–7054.
[19] T. Deng, H. Liu, J. Mol. Catal. A 388 (2014) 66–73.
[20] Y. Liu, C. Luo, H. Liu, Angew. Chem. Int. Ed. 51 (2012) 3249–3253.
[21] Y. Li, Y. Liao, X. Cao, T. Wang, L. Ma, J. Long, Q. Liu, Z. Xua, Biomass Bioenerg.
74 (2015) 148–161.
[22] J. Zhang, B. Hou, A. Wang, Z. Li, H. Wang, T. Zhang, AIChE J. 61 (2015) 224–238.
[23] Y. Su, H.M. Brown, X. Huang, X. Zhou, J.E. Amonette, Z.C. Zhang, Appl. Catal. A:
Gen. 361 (2009) 117–122.
Acknowledgements
[24] J.B. Binder, R.T. Raines, J. Am. Chem. Soc. 131 (2009) 1979–1985.
[25] J. Zhang, X. Liu, M. Sun, X. Ma, Y. Han, ACS Catal. 2 (2012) 1698–1702.
[26] D. An, A. Ye, W. Deng, Q. Zhang, Y. Wang, Chem. Eur. J. 28 (2012) 2938–2947.
This work was supported by the National Basic Research Pro-
gram of China (973 Programs, 2011CB201400 and 2011CB808700)
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056

G Model
CATTOD-9870; No. of Pages8
ARTICLE IN PRESS
8
Y. Liu, H. Liu / Catalysis Today xxx (2015) xxx–xxx
[27] R. Weingarten, W.C. Conner, G.W. Huber, Energy Environ. Sci. 5 (2012)
7559–7574.
[37] R.O.L. Souza, D.P. Fabian, C. Feche, F. Rataboul, D. Cardoso, N. Essayem, Catal
Today 195 (2012) 114–119.
[28] R.D. Cortright, R.R. Davda, J.A. Dumesic, Nature 418 (2002) 964–967.
[29] S. Carrettin, P. McMorn, P. Johnston, K. Griffin, C.J. Kiely, G.A. Attand, G.J.
Hutchings, Top. Catal. 27 (2004) 131–136.
[38] P.H. Brahme, L.K. Doralswamy, Ind. Eng. Chem. Proc. Des. Dev. 15 (1976) 130.
[39] N. Déchamp, A. Gamez, A. Perrard, P. Gallezot, Catal. Today 24 (1995)
29–34.
[30] J. Sun, H. Liu, Green Chem. 13 (2011) 135–142.
[40] E. Crezz, B.W. Hoffer, R.J. Berger, M. Makkee, F. Kapteijin, J.A. Moulijin, Appl.
Catal. A: Gen. 251 (2003) 1–17.
[41] J. Zhang, B. Hou, A. Wang, Z. Li, H. Wang, T. Zhang, AIChE J. 60 (2014)
3804–3813.
[42] R.L. David, CRC Handbook of Chemistry and Physics, 84th ed., CRC Press, Boca
Raton, Florida, 2003.
[43] M. Moliner, Y. Román-Leshkov, M.E. Davis, Proc. Natl. Acad. Sci. U.S.A. 107
(2010) 6164–6168.
[31] S. Wang, K. Yin, Y. Zhang, H. Liu, ACS Catal. 3 (2013) 2112–2121.
[32] J.S. Noh, J.A. Schwarz, J. Colloid Interface Sci. 130 (1989) 157–164.
[33] D.G. Barton, M. Shtein, R.D. Wilson, S.L. Soled, E. Iglesia, J. Phys. Chem. B 103
(1999) 630–640.
[34] I.E. Wachs, T. Kim, E.I. Ross, Catal. Today 116 (2006) 162–168.
[35] M.A. Alvarez-Merino, F. Carrasco-Marín, J.L.G. Fierro, C. Moreno-Castilla, J.
Catal. 192 (2000) 363–373.
[36] S. Lima, A.S. Dias, Z. Lin, P. Brandão, P. Ferreira, M. Pillinger, J. Rocha, V.
Dalvin-Casilda, A.A. Valente, Appl. Catal. A: Gen. 339 (2008) 21–27.
[44] Y. Tewari, Appl. Biochem. Biotechnol. 23 (1985) 17–24.
Please cite this article in press as: Y. Liu, H. Liu, Kinetic insight into the effect of the catalytic functions on selective conversion of cellulose
to polyols on carbon-supported WO3 and Ru catalysts, Catal. Today (2015), http://dx.doi.org/10.1016/j.cattod.2015.09.056