Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters to γ-valerolactone over ZrO2 catalyst supported on SBA-15 silica

Article abstract of DOI:10.1016/j.cattod.2016.05.016

A series of ZrO₂ catalysts supported on mesoporous SBA-15 silica were synthesized and examined as catalysts in the production of γ-valerolactone (GVL) from biomass-derived levulinic acid and its esters via a catalytic transfer hydrogenation (CTH) using several alcohols as hydrogen donors. Among the catalysts examined, ZrO₂ supported on high-surface-area SBA-15 silica bearing highly-dispersed zirconium species exhibited the highest catalytic activity, of which reaction rate was 1.7 times higher than that of the conventional bulk ZrO₂ catalyst. Zr K-edge XAFS analysis revealed that Zr⁴⁺-oxide species with a low-coordination state anchored on silica surface is the dominant active species. Such a Zr species efficiently converted levulinic acid and its esters to GVL (～95% yields) under mild reaction conditions (150 °C, 1.0 MPa Ar), and was reusable over multiple catalytic cycles without significant loss of catalytic performances. Comparative experiments, combined with detailed characterizations using NH₃/CO₂-TPD and in-situ FTIR, proposed a plausible reaction mechanism where basic Zr[sbnd]OH site triggers the CTH reaction involving a six-member ring transition state.

Full text of DOI:10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
Catalysis Today xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Catalytic transfer hydrogenation of biomass-derived levulinic acid
and its esters to ␥-valerolactone over ZrO2 catalyst supported on
SBA-15 silica
Yasutaka Kuwaharaa,b,∗, Wako Kaburagic, Yohsuke Osadac, Tadahiro Fujitanic,
Hiromi Yamashitaa,b,∗
a Division of Materials and Manufacturing Science, Graduate School of Engineering, Osaka University, 2-1 Yamada-oka, Suita, Osaka 565-0871, Japan
b Unit of Elements Strategy Initiative for Catalysts & Batteries (ESICB), Kyoto University, Katsura, Kyoto 615-8520, Japan
c Interdisciplinary Research Center for Catalytic Chemistry, National Institute of Advanced Industrial Science and Technology (AIST), 1-1 Higashi, Tsukuba,
Ibaraki, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of ZrO2 catalysts supported on mesoporous SBA-15 silica were synthesized and examined as
catalysts in the production of ␥-valerolactone (GVL) from biomass-derived levulinic acid and its esters
via a catalytic transfer hydrogenation (CTH) using several alcohols as hydrogen donors. Among the cata-
lysts examined, ZrO2 supported on high-surface-area SBA-15 silica bearing highly-dispersed zirconium
species exhibited the highest catalytic activity, of which reaction rate was 1.7 times higher than that
of the conventional bulk ZrO2 catalyst. Zr K-edge XAFS analysis revealed that Zr4+-oxide species with a
low-coordination state anchored on silica surface is the dominant active species. Such a Zr species effi-
ciently converted levulinic acid and its esters to GVL (∼95% yields) under mild reaction conditions (150 ◦C,
1.0 MPa Ar), and was reusable over multiple catalytic cycles without significant loss of catalytic perfor-
mances. Comparative experiments, combined with detailed characterizations using NH3/CO2-TPD and
in-situ FTIR, proposed a plausible reaction mechanism where basic Zr OH site triggers the CTH reaction
involving a six-member ring transition state.
Received 10 January 2016
Received in revised form 10 March 2016
Accepted 5 May 2016
Available online xxx
Keywords:
Biomass conversion
Levulinic acid
␥-Valerolactone
Catalytic transfer hydrogenation
Mesoporous silica
© 2016 Elsevier B.V. All rights reserved.
1. Introduction
ing platform molecules [8,9], since it can be used as a fuel additive,
solvent for biomass processing [10], as well as a versatile inter-
Due to the depletion of fossil fuel resources, rapid increase
of anthropogenic CO2 emissions and the associated deteriora-
tion of environmental quality, the search for renewable resources
is a matter of worldwide concern. Lignocellulosic biomass is a
promising feedstock for the production of fuels and chemicals,
since it is an abundant, inexpensive and carbon-neutral source and
is available worldwide. Recent efforts in the catalysis field have
found a number of promising approaches for the catalytic trans-
formation of lignocellulosic biomass to produce various chemicals
industrially valuable (so-called “biomass refinery”) [1–7]. Among
a wide range of chemicals obtained from lignocellulosic biomass,
␥-valerolactone (GVL) has been identified as one of the most allur-
mediate for the production of olefins [11], polymers [12,13], and
other value-added chemicals [14,15]. Moreover, GVL can be used
as a precursor to produce various liquid hydrocarbon fuels suitable
for gasoline, diesel and aviation kerosene etc. [16,17].
GVL is typically obtained through the reduction of levulinic
acid (LA) and its esters, which are intermediate molecules
produced by acid-catalyzed solvolysis of various carbohydrate
fractions of lignocelluloses, such as glucose, fructose, as well as
5-hydroxymethylfurfural (HMF), in water [18–20] and alcohols
[21–25], respectively. There are two possible pathways to produce
GVL from LA and its esters; (i) the hydrogenation (using molec-
ular H2) and (ii) the catalytic transfer hydrogenation (CTH) (using
alcohols). The former hydrogenation reaction is conventionally per-
formed over homogeneous/heterogeneous noble metal catalysts
(such as Pd [26–28], Pt [26], Ru [29–34], Ir [35]), which allow GVL
production with high yields. However, the noble metal catalysts are
expensive and inevitably suffer from deactivation, and the systems
require flammable, high-pressure H2 (>30 bar), limiting large-scale
operation requisite for biomass refinery. In contrast, the latter CTH
∗
Corresponding authors at: Division of Materials and Manufacturing Science,
Graduate School of Engineering, Osaka University, 2-1 Yamada-oka, Suita, Osaka
565-0871, Japan.
E-mail addresses: kuwahara@mat.eng.osaka-u.ac.jp (Y. Kuwahara),
yamashita@mat.eng.osaka-u.ac.jp (H. Yamashita).
http://dx.doi.org/10.1016/j.cattod.2016.05.016
0920-5861/© 2016 Elsevier B.V. All rights reserved.
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
2
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
process is a better alternative route for energy-saving and scalable
production of GVL, since the reaction proceeds under milder con-
ditions using low-cost metal oxides as catalysts and alcohols as
cheaper and safer H-donors [36,37]. In addition, in an ideal process
for the production of levulinate derivatives from lignocellulosic car-
bohydrates, use of alcohol media (methanol and ethanol etc.) is
preferred rather than water because of higher yields and an easier
isolation of the produced levulinate esters than LA [21–23]. Further-
more, the alcohols spent at this stage are supposed to be reused as
H-donors in the following CTH process. Thus, CTH process would
meet both the economic and technical requirements needed for
GVL production from biomass.
Regarding the production of GVL via a CTH process, several cat-
alytic systems have recently been reported. A pioneering work
was done by Chia and Dumesic et al. who reported a CTH pro-
cess to convert LA and its esters into GVL using bulk ZrO2 and
iso-propanol/iso-butanol as H-donors at 150 ◦C under pressurized
inert gas conditions (300 psig He) [38]. Sun and Lin et al. reported
that hydrated zirconia acts as an efficient heterogeneous catalyst
for a CTH reaction of ethyl levulinate in a supercritical ethanol at
around 250 ◦C [39,40]. However, these works revealed that bulk
zirconia catalysts hold critical drawbacks such as high catalyst load-
ing, relatively high reaction temperature/pressure, moderate GVL
selectivity and appreciable catalyst deactivation during the reac-
tion. Recently, Román-Leshkov et al. and Chuah et al. reported that
microporous molecular sieve bearing isolated Zr sites (Zr-beta)
is a more effective catalyst compared with bulk zirconia [41,42],
giving us an idea that dispersed Zr sites immobilized on oxide sup-
port can afford increased catalytic activities. According to these
works, we recently reported that low-cost and readily-prepared
silica-supported ZrO2 catalyst works as a better alternative for CTH
reaction of LA and its esters under relatively mild reaction condi-
tions (150 ◦C, 1.0 MPa of Ar) [43]. Although the catalytic activity
was found to be significantly influenced by altering type of the
oxide supports, a deep insight to the activity–structure relation-
ship has not been obtained, and the local structure of the active Zr
species and the associated reaction mechanism have not fully been
understood.
2.2. Catalyst preparation
Supported ZrO2 catalysts were prepared by in-situ hydroly-
sis of zirconium n-butoxide (Zr(OnBu)4, 70% in n-butanol, Tokyo
Chemical Industry Co., Ltd.) on oxide support in an organic solvent
according to a previously reported method [43]. In a typical synthe-
sis, SBA-15 silica was carefully treated at 150 ◦C under vacuum to
remove physisorbed water. To a glass flask containing dried SBA-15
support (2.0 g), 60 mL of 2-propanol containing desired amount of
Zr(OnBu)4 was added, followed by stirring at room temperature for
1 h in Ar atmosphere and further stirring at 80 ◦C for 5 h in air. The
obtained suspension was filtered, washed with ethanol, and dried
at 100 ◦C overnight, and was finally calcined in air at 400 ◦C for 6 h
to afford ZrO2(X)/SBA-15, where X represents the ZrO2 content in
the initial synthesis gel. Other types of silica (MCM-41 and fumed
SiO2) were also used as supports for ZrO2.
For comparison, pure ZrO2 was synthesized by a precipita-
tion method under basic conditions; briefly, a total of 13.5 g of
ZrO(NO3)2·2H2O (99.5%) was dissolved in 500 mL of deionized
water, followed by precipitation at pH 9.0 with 10% NH3 solution
with vigorous stirring and aging at room temperature for 48 h. The
obtained precipitate was filtered, thoroughly washed with deion-
ized water, dried at 100 ◦C overnight, and was finally calcined in air
at 400 ◦C for 6 h to give pure ZrO2 [25].
2.3. Characterization
X-ray diffraction (XRD) measurements were performed using a
Bruker AXS D8 Advance X-ray diffractometer with CuK␣ radiation
(␭ = 1.54056 Å). Nitrogen adsorption–desorption isotherms were
measured at −196 ◦C by using Micromeritics ASAP2020 system.
Prior to the measurements, the samples were degassed at 300 ◦C
under vacuum for 4 h to remove physisorbed water. The specific
surface area was calculated by the BET (Brunauer–Emmett–Teller)
method by using adsorption data ranging from p/p0 = 0.05
to 0.30. The pore size distributions were obtained from the
adsorption branches of the nitrogen isotherms by the BJH
(Barret–Joyner–Halenda) method. Transmission electron micro-
scope (TEM) images were taken with a FEI TITAN80-300 operated
at 200 kV. Thermal gravimetric analysis and differential thermal
gravimetric analysis (TG/DTG) were carried out on a MAC Science
TG-DTA2000 S system under a flow of air (50 mL/min) with a heat-
ing rate of 10 ◦C/min.
Zr K-edge X-ray absorption fine structure (XAFS) measurement
was performed in transmission mode at room temperature at the
BL01B1 beam line of SPring-8, Japan. A Si(311) single crystal was
used to obtain the monochromated X-ray beam. The Fourier trans-
formation of k3-weighted normalized EXAFS data (FT-EXAFS) from
k space to r space was performed over the range 3.0 < k (Å−1) < 13.0
to obtain the radial structure function (RDF). For the curve-fitting
In this study, an array of ZrO2 catalyst supported on SBA-15 silica
were prepared and examined in CTH reaction of LA and its esters to
produce GVL. The structure of the catalysts, including local infor-
mation concerning the overall porous properties and the state of
the supported Zr atoms, were investigated in detail to draw a rela-
tionship between the structure and catalytic performances and to
identify the real active Zr species. Furthermore, detailed charac-
terizations using NH3/CO2-TPD and in-situ FTIR were performed to
propose a plausible reaction mechanism. In addition, the scope of
substrates and alcohols as well as catalyst recyclability was also
investigated to demonstrate broad applicability of the catalyst.
2. Experimental
analysis, the empirical phase shift and amplitude functions for Zr
O
and Zr· · ·Zr were extracted from the data for monoclinic ZrO2. The
analysis of the XAFS data was performed by using Rigaku REX2000
software.
2.1. Materials
All commercially available chemical reagents for material
synthesis and catalytic tests were purchased from Wako Pure
Chemical Ind., Ltd. unless otherwise noted and used without fur-
ther purification. A pure siliceous SBA-15 was prepared by a
The acidity and basicity of the samples were quantified by
temperature programmed desorption of NH3 (NH3-TPD) and CO2
(CO2-TPD), respectively, using a BELCAT-B system (BEL Japan, Inc.).
Approximately 100 mg of sample mounted in a quartz tube was
pretreated at 600 ◦C in a He gas flow (50 mL/min) for 1 h, allowed
to cool to 50 ◦C, and subsequently exposed to either 5% NH3/He gas
or 100% CO2 gas (50 mL/min) for 1 h. After being flushed with He for
0.5 h to eliminate weakly-adsorbed NH3/CO2 molecules, TPD was
carried out between 50 and 600 ◦C under a He flow (30 mL/min)
with a ramping rate of 10 ◦C/min, and the desorbed NH3/CO2 was
detected by a thermal conductivity detector.
hydrothermal synthesis method using Pluronic P123® (PEO20
-
PPO70-PEO20, Aldrich, MW = 5800) as an organic template and
tetraethoxyorthosilicate (TEOS) as a silicon source according to the
previous literature [44]. A pure siliceous MCM-41 was prepared by
a hydrothermal synthesis method using cetyltrimethylammonium
bromide (CTAB) as an organic template and TEOS as a silicon source
with the molar ratio of Si: NH3: CTAB: H2O = 1.0: 3.3: 0.1: 66 [45].
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
3
In-situ FTIR spectroscopy measurement was performed by using
a Cary 670-IR instrument (Agilent technology, Inc.) in the spec-
tral range 4000–1000 cm−1. Approximately 5 mg of sample was
pressed into a self-supported wafer (␾10 mm). The wafer loaded
into a high-vacuum transmission FTIR cell was pretreated at 400 ◦C
under a He flow (100 mL/min) for 1 h, allowed to cool to 180 ◦C,
and then the baseline spectrum was collected. After exposing to a
flow of 25 mbar 2-propanol in He (100 mL/min) for 1 h, FTIR differ-
ence spectra of 2-propanol adsorbed on the sample were collected
under flushing in a He flow (100 mL/min) by subtracting the base-
line spectrum from the each collected spectrum.
The porous properties of the catalysts were investigated by N2
physisorption measurement. The parent SBA-15 displays a type
IV nitrogen adsorption-desorption isotherm with a clear capil-
lary condensation at around p/p0 = 0.7 due to the presence of
periodically-aligned mesopore channel systems (Fig. 2(A) (a)). The
isotherms of ZrO2/SBA-15 materials showed decreased quantities
of adsorbed N2 and less defined hysteresis loops (Fig. 2(A)–(d)),
indicating subpar porous qualities (for N2 physisorption data of
ZrO2(10)/MCM-41 and ZrO2(10)/SiO2, see Fig. B). The textural
properties obtained from XRD measurement and N2 adsorption
isotherms are tabulated in Table 1. Along with the increase of
ZrO2 content from 0 to 59.2 wt%, the specific surface area (SBET
)
and total pore volume (Vtotal) linearly reduced (SBET: from 1040
to 475 m2/g, Vtotal: from 1.34 to 0.39 cm3/g). Nonetheless, the
materials still had large surface areas and pore volumes as to
be regarded as mesoporous materials. The pore size distribution
curves calculated from the N2 adsorption data showed broader
pore size distributions as increasing the ZrO2 content (Fig. 2(B)).
The average pore diameter (Dp) linearly decreased from 8.5 to
5.8 nm as the ZrO2 content increases from 0 to 59.2 wt%, and con-
sequently the thickness of silica wall (Tw) increased from 5.5 nm
to 8.2 nm. The mesoporous structure of SBA-15 was confirmed
by TEM images. ZrO2(10)/SBA-15 showed a similar morphologi-
cal feature to that of bare SBA-15, representing the retention of the
original mesopore structure (Fig. 3(a) and (b)). On the other hand,
ZrO2(35)/SBA-15 and ZrO2(60)/SBA-15 showed extra-framework
aggregated ZrO2 particles (Fig. 3(c) and (d), respectively). These
combined characterization results demonstrate that Zr species are
dominantly deposited inside the pore channels with uniform distri-
bution at lower ZrO2 loading levels, but introduction of an excess
amount of Zr (>35 wt%) leads to a formation of extra-framework
aggregated ZrO2 particles and accordingly results in reduced meso-
porosity.
2.4. Procedures for catalytic reactions
In a typical reaction, a reaction mixture containing catalyst
(40 mg as ZrO2), substrate (2 mmol), and alcohol (10 mL) was
charged into a 60 mL cylindrical stainless steel high-pressure reac-
tor (EYELA, Inc.) equipped with a bourdon pressure gauge, which
was then sealed, purged and pressurized with 1.0 MPa of Ar
and then heated to 150 ◦C. During the reaction, magnetic stir-
ring at 600 rpm was continued. After the predetermined reaction
time, the reactor was cooled to room temperature and the liq-
uid products recovered from the reaction mixture were analyzed
by a gas chromatograph (Shimadzu GC-14B) with a frame ioniza-
tion detector equipped with a capillary column (ULBON HR-20 M;
0.53 mm × 30 m; Shinwa Chemical Ind., Ltd.). Conversion of sub-
strate and yields of products were quantified using biphenyl as an
internal standard. To assess the catalyst reusability, the spent cat-
alyst was retrieved from the reaction mixture by filtration, washed
with acetone, dried at 100 ◦C and then subjected to multiple cat-
alytic runs (for detailed procedures for catalyst reusability test, see
the Supplementary Information).
The thus synthesized catalysts were tested for CTH reaction of
methyl levulinate (ML) using 2-propanol as a H-donor at 150 ◦C
under 1.0 MPa of Ar. The amount of ZrO2 introduced to the reac-
tor was fixed to be the same in all cases to compare catalytic
activities per amount of ZrO2. When 2-propanol was used as a
H-donor, GVL was produced as a major product, and transesteri-
fication product (i.e., isopropyl levulinate ester) and hydrogenated
compounds (i.e., 4-hydroxypentanoic acid methyl/isopropyl esters
(4-HPE)) were produced as by-products. Appreciable amounts of
acetone and methanol were also detected, elucidating that the reac-
tion proceeds via a CTH reaction and the following dealcoholization
steps as illustrated in Scheme 1, where transesterifications of ML
and 4-HPE with alcohols take place as side reactions. As shown
in Table 2, a series of SBA-15-supported ZrO2 catalysts showed
higher catalytic activities (Entries 1–5) than bulk ZrO2 (Entry 8),
but increasing ZrO2 content from 9.8 to 49.7 wt% resulted in a grad-
ual activity decrease (for reaction kinetics data, see Fig. C). At a
fixed loading level (10 wt%) of ZrO2, SBA-15 was the most effective
support to achieve high GVL production rate among three types
of silica support (SBA-15, MCM-41 and fumed silica) (cf. Entries 1,
6 and 7). ZrO2(10)/SBA-15 (SBET = 810 m2/g, Dp = 7.7 nm) afforded
>99.5% conversion and 91% GVL yield at 3 h of reaction, of which
reaction rate was 1.7 times higher than that of the conventional
bulk ZrO2 (55% conversion and 53% GVL yield). ZrO2(10)/MCM-41
(SBET = 783 m2/g, Dp = 2.9 nm) and ZrO2(10)/SiO2 (SBET = 250 m2/g)
gave almost similar activities (98% conversions and 87–89% GVL
yield) under the same conditions, while they have totally differ-
ent porous structures and surface area. These experimental results
led us to a hypothesis that unique local structure of the Zr atoms
highly-dispersed on silica support might be responsible for achiev-
ing a high catalytic activity, rather than the textural properties of
silica supports such as topology, pore size and surface area.
3. Results and discussion
3.1. Structural analysis
For the synthesis of silica-supported ZrO2 catalysts, in-situ
hydrolysis method of zirconium n-butoxide was employed, so
as to obtain the catalysts with desired ZrO2 content. The
ZrO2 content was controlled by simply varying the amount
of zirconium n-butoxide precursor in the initial gel to obtain
SBA-15-supported ZrO2 catalysts with 9.8 wt% (ZrO2(10)/SBA-15),
19.4 wt% (ZrO2(20)/SBA-15), 32.7 wt% (ZrO2(35)/SBA-15), 49.7 wt%
(ZrO2(50)/SBA-15) and 59.2 wt% (ZrO2(60)/SBA-15) ZrO2 loading
(Table 1). Fig. 1 shows XRD patterns of ZrO2/SBA-15 materials
with varied ZrO2 contents. Low angle XRD patterns of ZrO2/SBA-15
materials all exhibited well-defined diffraction planes associated
with 2D hexagonal mesoporous structure (P6 mm symmetry) which
are similar to those of the parent SBA-15. The intensities of the
peak at around 2ꢁ = 0.9◦ ascribed to the (100) diffraction gradu-
ally decreased as increasing the ZrO2 content, while the lattice
paramater a0, which represents the distance between the meso-
pores, mostly remained unchanged (Table 1), indicating a blockage
of mesopores by the incorporation of ZrO2. In high angle XRD
patterns, no diffraction patterns were observed for the samples
with ZrO2 contents lower than 20 wt%, suggesting that Zr species
is existing as highly-dispersed Zr species or as amorphous zirco-
nia. A similar result was observed for ZrO2 supported on fumed
silica and MCM-41 silica (see Fig. A in the Supplementary Infor-
mation). On the other hand, the samples with higher ZrO2 content
(ZrO2 > 35 wt%) exhibited diffraction patterns indexed to tetrago-
nal ZrO2 phase (2ꢁ = 30.2, 35.2, 50.4 and 60.1◦), representing the
formation of aggregated ZrO2 species.
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ꢀ-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
4
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
Table 1
Textural properties of silica-supported ZrO2 catalysts.
Sample
ZrO2 contenta (wt%)
XRD
N2 physisorption
b
c
d
e
f
g
d100 (nm)
a0 (nm)
SBET (m2/g)
Vtotal (cm3/g)
Dp (nm)
Tw (nm)
SBA-15
–
9.9
9.9
9.6
9.6
9.6
9.9
5.6
5.6
14.0
14.0
13.6
13.6
13.6
14.0
7.9
1040
810
733
724
597
475
923
783
279
250
112
1.34
1.03
0.83
0.75
0.57
0.39
0.86
0.65
0.60
0.71
0.29
8.5
7.7
7.1
6.7
6.3
5.8
2.9
2.9
5.5
6.3
6.5
6.9
7.3
8.2
5.0
5.0
ZrO2(10)/SBA-15
ZrO2(20)/SBA-15
ZrO2(35)/SBA-15
ZrO2(50)/SBA-15
ZrO2(60)/SBA-15
MCM-41
ZrO2(10)/MCM-41
SiO2
ZrO2(10)/SiO2
ZrO2
9.8
19.4
32.7
49.7
59.2
–
9.7
–
9.6
100
7.9
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
N.D.
a
Molar ratio determined by EDX.
b
c
d
e
f
Calculated from Bragg’s equation: 2d100 sin ␪ = n␭ (␭ = 0.154056 nm).
√
Lattice parameter defined by a0 = 2d100
/
3 assuming hexagonal structure.
Specific surface area calculated by BET method from N2 adsorption data.
Total pore volume at p/p0 = 0.99.
Average pore diameter determined by BJH method.
Wall thickness defined by Tw = a0− Dp. N.D. = Not determined.
g
ی
Tetragonal ZrO2
(A)
(B)
d100 = 9.9 nm
(a)
(a)
d100 = 9.9 nm
(b)
(c)
(d)
(e)
(f)
(b)
d100 = 9.6 nm
(c)
ی
ی
ی
ی
d100 = 9.6 nm
(d)
ی
ی
ی
ی
d100 = 9.6 nm
ی
(e)
ی
d100 = 9.9 nm
(f)
0
1
2
3
4
5
10 20 30 40 50 60 70 80
/ degree
2
θ
/ degree
2θ
Fig. 1. (A) Small-angle and (B) wide-angle XRD patterns of (a) parent SBA-15 and ZrO2/SBA-15 with varied ZrO2 content; (b) 10, (c) 20, (d) 35, (e) 50 and (f) 60 wt% in the
initial gel.
Table 2
CTH reaction of ML to GVL with 2-propanol as a hydrogen donora.
Entry
Catalyst
ZrO2 content (wt%)
Conv. (%)
GVL yield (%)
GVL production rate (mmol/g-ZrO2/h)
1
2
3
4
5
6
7
8
9
ZrO2(10)/SBA-15
ZrO2(20)/SBA-15
ZrO2(35)/SBA-15
ZrO2(50)/SBA-15
ZrO2(60)/SBA-15
ZrO2(10)/MCM-41
ZrO2(10)/SiO2
ZrO2
9.8
>99.5
98
84
70
61
98
98
55
91
89
76
62
55
89
87
53
15.2
14.8
12.7
10.3
9.2
14.8
14.5
8.8
19.4
32.7
49.7
59.2
9.7
9.6
100
–
SBA-15
no reaction
a
Reaction conditions: catalyst (40 mg as ZrO2), ML (2 mmol), 2-PrOH (10 mL), 150 ◦C, Ar (1.0 MPa), 3 h.
3.2. Investigation of local structure of Zr species
(XANES) spectra at the Zr K-edge and the corresponding RDFs of
a series of ZrO2/SBA-15 samples, together with pure monoclinic
ZrO2 as a reference. As shown in Fig. 4(A), the peak positions of
the edge were invariant for all the samples. Focusing on the edge-
top region, the intensity of the peak at around 18018 eV increased
The local structure of the Zr atoms supported on silica sup-
ports were investigatedby X-ray absorption measurement. Fig. 4(A)
and (B) shows normalized X-ray absorption near-edge structure
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
5
2500
2000
1500
1000
500
3.5
(A)
(B)
8.5
3.0
+1500
(a)
(b)
(a)
(b)
2.5
7.7
+1200
2.0
7.1
+900
+600
(c)
(d)
(c)
(d)
(e)
(f)
1.5
6.7
1.0
6.3
+300
(e)
(f)
0.5
5.8
0
0.0
0.0
0.2
0.4
0.6
0.8
1.0
0
10
20
30
p/p0
Pore diameter / nm
Fig. 2. (A) Nitrogen adsorption-desorption isotherms and (B) the corresponding pore size distribution curves of (a) parent SBA-15 and ZrO2/SBA-15 with varied ZrO2 content;
(b) 10, (c) 20, (d) 35, (e) 50 and (f) 60 wt% in the initial gel. Filled and empty symbols in the left represent adsorption and desorption branches, respectively.
gradually and the peak at around 18025 eV adversely became less
pronounced as the ZrO2 loading in ZrO2/SBA-15 samples increased.
Interestingly, the latter peak was hardly observed in the spectrum
of pure ZrO2 with heptahedral coordination geometry [46]. The ori-
gin of the latter peak has been ascribed to the Zr atoms stabilized
in a distorted polyhedron with reduced Zr O bond distances [47].
These features indicate that the Zr species supported on SBA-15 sil-
ica are in the 4+ oxidation state in distorted coordination geometry.
In the RDFs, the peaks corresponding to Zr O and Zr· · ·Zr bonds
were observed at around r = 1.5–1.6 and 3.0 Å (the phase shift
Fig. 3. TEM images of ZrO2(10)/SBA-15 ((a) side-view, (b) top-view), (c) ZrO2(35)/SBA-15 and (d) ZrO2(60)/SBA-15.
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
6
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
Scheme 1. Catalytic transfer hydrogenation (CTH) of levulinic acid and its esters using alcohol as a H-donor.
Table 3
Zr K-edge EXAFS curve-fitting parameters of silica-supported ZrO2 catalystsa.
2d
Sample
ZrO2 content (wt%)
Shell
C.N.b
Rc (Å)
ꢂ␴ (Å2)
ZrO2(10)/SBA-15
ZrO2(20)/SBA-15
ZrO2(35)/SBA-15
ZrO2(50)/SBA-15
ZrO2(60)/SBA-15
ZrO2(10)/MCM-41
ZrO2(10)/SiO2
ZrO2
9.8
Zr OZr· · ·Zr
Zr OZr· · ·Zr
Zr OZr· · ·Zr
Zr OZr· · ·Zr
Zr OZr· · ·Zr
Zr OZr· · ·Zr
Zr OZr· · ·Zr
Zr OZr· · ·Zr
4.500.93
5.171.03
6.612.88
7.153.16
7.093.31
4.911.25
5.511.61
2.1213.401
2.1213.404
2.1283.425
2.1283.693
2.1203.691
2.1213.415
2.1223.431
2.1603.460
0.0620.043
0.0600.052
0.0810.088
0.0810.088
0.0720.086
0.0620.050
0.0600.059
N.D. N.D.
19.4
32.7
49.7
59.2
9.7
9.6
100
7
7
a
k3-weighted EXAFS, 3.0 ≤ k(Å−1) ≤ 13.0.
Average coordination number.
Average atomic distance.
b
c
d
Debye–Waller factor.
distance side with a significant decrease in the FT magnitude as the
ZrO2 content in the samples decreased. In fact, it was confirmed
by curve-fitting analysis that the coordination numbers (C.N.) and
the interatomic distance (R) of the first Zr O shell for the sup-
ported ZrO2 samples were apparently smaller than those for bulk
ZrO2 (Table 3); as for ZrO2(10)/SBA-15, the C.N. and R of the first
Zr O shell were determined to be 4.50 and 2.12 Å, respectively,
whereas those of bulk ZrO2 were 7 and 2.16 Å [46], elucidating
that the highly-dispersed Zr atoms supported on SBA-15 are in
low-coordination state with reduced Zr O bond distances. In com-
parison, the C.N. of the Zr O shell was found to be ca. 7.1 at higher
ZrO2 loading levels (50−60 wt%), which is almost similar to that
of bulk ZrO2, again indicating the formation of aggregated ZrO2
crystals at higher ZrO2 loading levels. Furthermore, the FT magni-
tude of the second shell (Zr· · ·Zr) for the supported ZrO2 samples
was considerably damped compared to that for bulk ZrO2; as the
ZrO2 content increased from 9.8 to 59.2 wt%, the C.N. for the Zr· · ·Zr
shell drastically changed from 0.93 to 3.31, which are quite smaller
than that for bulk ZrO2 (C.N. = 7). In particular, ZrO2(10)/SBA-15
provided the lowest C.N. value for the Zr· · ·Zr shell (0.93) com-
pared with other types of silica-supported ZrO2 samples (1.25 for
ZrO2(10)/MCM-41 and 1.61 for ZrO2(10)/SiO2) despite the same
ZrO2 content, which is probably due to the difference in sur-
face area effective for anchoring Zr atoms. It is also worth noting
that ZrO2(50)/SBA-15 and ZrO2(60)/SBA-15 exhibited the signals
assignable to Zr· · ·Zr shell at r = 3.34 Å (R = 3.69 Å), which coincided
with that of tetragonal ZrO2 (R = 3.63 Å) [46]. This result is in good
agreement with the XRD result (see Fig. 1). These XAFS results
can well explain the dispersion state of the Zr atoms stabilized on
(A)
(B)
Zr-O
Zr
͐Zr
(a)
(a)
(b)
(c)
(b)
(c)
(d)
(d)
(e)
(f)
(e)
(f)
17950
18000
Energy / eV
18050
18100
0
1
2
3
4
5
r / ύ
Fig. 4. (A) Zr K-edge XANES spectra and (B) radial distribution functions of
ZrO2/SBA-15 with varied ZrO2 content ((a) 10, (b) 20, (c) 35, (d) 50 and (e) 60 wt%
in the initial gel) and (f) monoclinic ZrO2 as a reference sample.
uncorrected), respectively (Fig. 4(B)). Distinct changes of RDFs were
found with the variation of the chemical composition of ZrO2/SBA-
15 samples; the position of the first shell (Zr O) shifted to shorter
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
7
might be the one of the reasons why ZrO2(10)/SBA-15 provided the
highest activity in the CTH reaction.
18
Poisoning experiments were also carried out by introducing two
different external additives, pyridine and benzoic acid, into the
reaction system in order to elucidate the catalytic role of the above-
mentioned acid/base properties. Fig. 7 shows the effects of the two
additives on the catalytic activity of ZrO2(10)/SBA-15 catalyst in
the CTH reaction of ML to GVL, where pyridine and benzoic acid
strongly interact with acidic and basic sites of the catalyst, respec-
tively. As shown in Fig. 7, the addition of benzoic acid drastically
reduced the activity, while the addition of pyridine had little impact
on the activity, being consistent with the result by Sun and Lin et al.
[39,40]. This result reveals that the CTH reaction of ML to GVL is pri-
marily dominated by basic property of the Zr-oxide species and acid
property have little involvement in the reaction.
16
14
12
10
8
1
6
7
2
3
1 ZrO2(10)/SBA-15
2 ZrO2(20)/SBA-15
3 ZrO2(35)/SBA-15
4 ZrO2(50)/SBA-15
5 ZrO2(60)/SBA-15
6 ZrO2(10)/MCM-41
7 ZrO2(10)/SiO2
4
5
0
0
1
2
3
4
C.N. of nearest-neighboring Zr
3.4. In-situ FTIR study
Fig. 5. Relationship between coordination numbers (C.N.) of nearest-neighboring
Zr atoms and catalytic activities in the CTH reaction of ML with 2-propanol over
silica-supported ZrO2 catalysts.
To obtain insights about the reaction mechanism, in-situ
FTIR measurement was performed using 2-propanol as a probe
molecule. The measurement was performed at the temperature
higher than the batch reaction conditions, 180 ◦C, to obtain suf-
ficient IR intensity. Fig. 8 shows in-situ FTIR difference spectra of
the adsorption complexes formed by the adsorption of 2-propanol
on ZrO2(10)/SBA-15 as a function of flushing time in He flow at
180 ◦C. The IR spectrum taken at 0 min showed a negative band
appeared at 3735 cm−1, which is attributed to the loss of sur-
face OH groups due to the interaction with 2-propanol, as well
as absorption bands typical of free 2-propanol molecules; ␯(OH):
silicas, i.e. the Zr atoms are present as highly-dispersed species with
low-coordination geometry at lower ZrO2 level, and are present as
agglomerated species at higher ZrO2 level.
The thus calculated C.N. values could be used as a good indi-
cator associated the catalytic activity in the CTH reaction. Fig. 5
shows the plot of the catalytic activities against the C.N. of nearest-
neighboring Zr atoms of the series of silica-supported ZrO2 samples,
which provided a more distinct correlation rather than the ones
plotted against the other textural parameters. This correlation cor-
roborates the idea that the highly-dispersed Zr4+-oxide species
with a low-coordination state anchored on silica surface is the dom-
inant active species for the CTH reaction, where high-surface-area
silica support provides a suitable environment for generating such
active species.
3653 cm−1, ␯(CH3)as: 2980 cm−1, ␯(CH3)s: 2890 cm−1, ␦(CH3)as
,
1462 cm−1, ␦(CH3)s: 1386 cm−1 [48]. The gradual decay of the
absorption bands at 3653, 2980, 2890, 1462 and 1386 cm−1 after
He flushing represents the desorption of 2-propanol molecules
weakly adsorbed on the surface. However, the bands assignable
to C H stretching/bending vibrations still remained even after the
He flow for 1 h at 180 ◦C. Unfortunately, the bands assignable to
C
O stretching vibration (typically seen at around 1163 cm−1) were
3.3. Investigation of acid-base properties
not observed due to the significant overlapping with the bands
of siloxane network of silica support. Furthermore, an absorption
band assignable to the C O stretching vibration mode appeared
at 1685 cm−1, which appreciably shifted to lower wavenumbers
compared to that of free acetone molecule, suggesting the pres-
ence of acetone molecules strongly adsorbed on the surface [48].
These results indicate that metal-bound 2-propoxide species are
formed via a deprotonation of 2-propanol, part of which is then
transformed to acetone via a hydride elimination from ␤-position.
Based on the in-situ FTIR results together with the results of the
above catalyst characterizations, a plausible reaction mechanism
of the CTH reaction is proposed as follows (Scheme 2). First step is
the interaction of alcohol with the Zr-related acid–base site (unsat-
urated Zr4+ O2− pairs), where alcohol is adsorbed on the basic Zr
site (Zr OH species) and dissociated to form the corresponding Zr-
bound alkoxide (Step 1). The carbonyl group of levulinate ester is
then coordinated to the Zr site, probably due to the electrostatic
interaction between electron-rich oxygen atoms in the carbonyl
group and the electron-deficient Zr-related Lewis acid site, while
this is not a rate-determining step, and subsequently forms a six-
membered ring transition state (Step 2). The ␤-hydride of the
Zr-bound alkoxide attacks the carbonyl group of levulinate ester
to yield Zr-bound 4-hydroxypentanoic acid ester (4-HPE) species
and new carbonyl compound corresponding to the alcohol (Step 3).
Subsequently, the carbonyl byproduct is released (Step 4) and the
Zr-bound 4-HPE species are displaced by another alcohol to regen-
erate the Zr-bound alkoxide species (Step 5), thus the catalytic cycle
is completed. The released 4-HPE is thermodynamically unstable as
rarely being detected by a GC analysis (see also Table 4), therefore,
To investigate the catalytic role of the Zr4+-oxide species,
TPD measurement was performed using NH3 and CO2 as probe
molecules. Normalized NH3 and CO2-TPD profiles of ZrO2(10)/SBA-
15, bulk ZrO2 and normal SBA-15 samples are shown in Fig. 6(A) and
(B), respectively. In NH3-TPD profiles, ZrO2(10)/SBA-15 exhibited a
desorption peak positioned at 140 ◦C which is attributed to weak
acid sites and a broad desorption peak in the range of 200–400 ◦C
which is assigned to medium acid sites, whereas normal SBA-
15 exhibited almost no signals. The total amount of acid sites
increased in the following order: SBA-15 (0.067 mmol/g) < ZrO2
(0.381 mmol/g) < ZrO2(10)/SBA-15 (0.784 mmol/g). A similar trend
was observed in CO2-TPD profiles. ZrO2(10)/SBA-15 exhibited
a desorption peak at around 225 ◦C which can be assigned to
medium basic sites, whereas normal SBA-15 exhibited almost
no TPD signals and bulk ZrO2 exhibited a peak at lower tem-
perature region (around 155 ◦C). The total amount of basic
sites increased in the order of SBA-15 (0.047 mmol/g) < ZrO2
(0.228 mmol/g) < ZrO2(10)/SBA-15 (0.259 mmol/g). These results
demonstrate that the Zr species in ZrO2(10)/SBA-15 catalyst
bear both acid and basic properties stronger than those in
bulk ZrO2. Topologically-different counterparts, ZrO2(10)/SiO2 and
ZrO2(10)/MCM-41, exhibited similar or even weaker acid/base
properties; however, the basic strength of ZrO2(10)/SBA-15 was,
interestingly, found to be the highest among the samples tested
(for TPD profiles of ZrO2(10)/MCM-41 and ZrO2(10)/SiO2, see Fig.
D). While the cause of the creation of such unusual stronger basic
site on the surface of SBA-15 silica has not fully understood yet, this
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
8
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
Fig. 6. (A) NH3 and (B) CO2 TPD profiles of SBA-15, ZrO2(10)/SBA-15 and bulk ZrO2 samples. TCD signals were normalized by the weight of the samples used in each
experiment.
Table 4
Scope of substrates and alcohols for the production of GVL over ZrO2(10)/SBA-15 catalysta.
Entry
Substratee
Alcohol
Conv. (%)
Selectivityf (%)
GVL
GVL yield (%)
Trans
Trans + H
Others
1b
2
3
ML
ML
ML
ML
ML
ML
ML
EL
EL
EL
BL
LA
MeOH
EtOH
4.7
98
98
>99.5
>99.9
99
97
15
65
97
30
42
40
90
95
90
91
93
92
94
88
90
0
0
0
4
4
4
6
6
0
0
6
6
5
70
7
2
0
0
0
2
7
8
0
0
0
1.4
41
39
91
95
89
88
14
60
91
71
90
51
54
6
1
4
1
0
0
0
1-PrOH
2-PrOH
2-PrOH
2-BuOH
CyOH
EtOH
EtOH
2-PrOH
2-PrOH
2-PrOH
4
5c
6
7
8
9d
10
11
12
81
>99.9
6
5
a
Reaction conditions: catalyst (40 mg as ZrO2), substrate (2 mmol), alcohol (10 mL), 150 ◦C, Ar (1.0 MPa), 3 h.
b
c
d
e
f
At 90 ◦C.
At 6 h of reaction.
At 21 h of reaction.
ML = methyl levulinate, EL = ethyl levulinate, BL = n-butyl levulinate, LA = levulinic acid.
Trans = transesterification product. Trans + H = transesterification and hydrogenated product. Others = hydrogenated product + unknown product.
2.5
100
v(CH3)as
2980
2.0
flushing in He flow
pyridine
80
60
40
20
0
0 min
1.5
1.0
1 min
δ(CH3)s
5 min
1386
10 min
20 min
30 min
40 min
60 min
v(CH3)s
2890
δ
(CH3)as
1462
v(C=O)
1685
0.5
0.0
benzoic acid
v(OH)
3653
-0.5
v(OH)
3735
-1.0
4000
0
100
200
300
400
500
3500
3000
1700 1600 1500 1400 1300
Wavenumber / cm-1
Amount of additives / mol%
Fig. 7. Effect of addition of pyridine and benzoic acid on the catalytic activity
of ZrO2(10)/SBA-15 catalyst. Reaction conditions: catalyst (40 mg as ZrO2), ML
(2 mmol), 2-PrOH (10 mL), additive (0–500 mol%), 150 ◦C, Ar (1.0 MPa), 3 h.
Fig. 8. In-situ FTIR difference spectra of the adsorption complexes formed by the
adsorption of 2-propanol on ZrO2(10)/SBA-15 as a function of flushing time in He
flow at 180 ◦C.
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
9
Scheme 2. Proposed reaction mechanism of the catalytic transfer hydrogenation (CTH) of levulinate esters over silica-supported ZrO2 catalyst using alcohol as a H-donor.
it readily undergoes intramolecular dealcoholization to give GVL
together with the formation of the equimolar amount of alcohol
(Step 6) [49]. Since each step in this reaction is reversible, the reac-
tion rate is controlled by the concentrations and thermodynamic
properties of the products and intermediates. Because the rate of
step 6 is very fast, the reaction equilibrium tends to incline to the
4-HPE formation, thereby a high GVL yield can be achieved. The
rate determining step in this process must be step 1 where the Zr-
related basic site (Zr OH species) plays a vital role for abstracting
protons from alcohols [40,50]. This assumption can be elucidated
by the experiments using ZrO2(10)/SBA-15 pre-treated at different
temperatures. The catalytic activity significantly decreased when it
was pre-treated at temperatures higher than 400 ◦C, and showed a
continuous decrease as the pre-treatment temperature increased
(Fig. E). This trend coincided well with that observed in the TG
analysis, which showed a distinct weight loss assignable to the
removal of surface OH groups over 330 ◦C (Fig. E). Considering the
fact that crystallinity and porous structures were hardly changed
upon pre-treatment up to 800 ◦C, the observed activity loss should
be attributed to the removal of Zr OH groups which triggers the
CTH reaction. One of the reasons why ZrO2(10)/SBA-15 provided
the highest activity in the CTH reaction might be attributed to the
creation of stronger basic sites as discussed above.
1-propanol resulted in moderate GVL yields (39–41%) (Entries 2, 3),
affording the corresponding transesterification products as main
products. This is because the ␤-hydride elimination from such Zr-
bound ethoxide/propoxide species is difficult to take place (for
reaction kinetics data, see Fig. F) [39,40].
GVL could be synthesized from other levulinate derivatives,
including ethyl levulinate (EL), butyl levulinate (BL) and levulinic
acid (LA), using 2-propanol as a H-donor (Entries 10–12). The GVL
production rate increased in the order of BL < EL < LA = ML (for reac-
tion kinetics data, see Fig. F), demonstrating a substantial impact of
the alkyl chain length of the ester groups, which might be associated
with the dealcoholization step (Step 6 in Scheme 2). In addition,
combination of EL and ethanol was possible for the production of
GVL, although the reaction rate was quite low; 14% and 60% GVL
yields were obtained after 3 and 21 h of reaction (Entries 8 and 9).
EL can be produced from biomass through alcoholysis of sugars in
ethanol media, which ideally affords an ethanol solution contain-
ing EL (as well as the equimolar amount of ethyl formate). A direct
use of the ethanol media as a H-donor in the following CTH process
will be a promising strategy for GVL production, because several
energy-consuming steps associated with isolation and purification
can be skipped. This result opens up a possibility for the establish-
ment of a simple and low-cost GVL production from biomass, for
which ZrO2(10)/SBA-15 catalyst can be a viable option, although
a further improvement of catalytic performance is needed before
substantiating this technology.
3.5. Scope of substrates and alcohols
The alcohol, which works as a H-donor as well as a reaction sol-
vent, plays an important role in the CTH process. Herein, various
primary and secondary alcohols were evaluated as the H-donors
in the CTH reaction of ML using ZrO2(10)/SBA-15 as a catalyst
(Table 4). Among various alcohols examined, secondary alcohols
such as 2-propanol, 2-butanol and cyclohexanol worked as effi-
cient H-donors, which afforded similar GVL yields (88–91%) at 3 h
of reaction with more than 90% selectivity (Entries 4, 6, 7). Pro-
longing reaction time to 6 h led to an increased GVL yield of 95%
(Entry 5). In contrast, use of primary alcohols such as ethanol and
3.6. Reusability of catalyst
To assess the reusability and stability of the catalyst,
ZrO2(10)/SBA-15 and bulk ZrO2 catalysts were subjected to five
successive catalytic cycles in the CTH reaction of ML with 2-
propanol under the standard reaction conditions (i.e., 150 ◦C, Ar
1.0 MPa, 3 h). As shown in Fig. 9, ZrO2(10)/SBA-15 was at least
five times reusable with keeping most of its activity (91% → 76%
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
10
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
developed in this study offers a simple, reusable, noble-metal-free
ZrO2(10)/SBA-15
bulk ZrO2
100
91
catalytic system that enables efficient production of a valuable plat-
form molecule, GVL, from biomass-derived levulinic acid and its
esters.
83
80
77
76
80
Acknowledgements
60
51
This study was financially supported by the National Institute
of Advanced Industrial Science and Technology (AIST). A part of
this work was also supported by the Grant-in-Aid from Frontier
Research Base for Global Young Researchers, Osaka University and
the Grant-in-Aid for Scientific Research (KAKENHI) from the Japan
Society for the Promotion of Science (JSPS) (No. 15K18270). The
synchrotron radiation experiments were performed at the BL01B1
beam line at SPring-8 with the approval of JASRI (No. 2014A1045).
40
20
0
25
19
19
9
1st
2nd
3rd
Cycles
4th
5th
Appendix A. Supplementary data
Fig. 9. Catalyst reusability tests of ZrO2(10)/SBA-15 and bulk ZrO2 catalysts.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.cattod.2016.05.
016.
GVL yields in fifth catalytic cycle), while bulk ZrO2 showed a
significant activity reduction (51% → 9% GVL yields in fifth cat-
alytic cycle). Structural analysis of the spent catalyst by means of
XRD, N2 physisorption and TEM measurements revealed that the
crystalline/porous structures of the catalyst were mostly retained
even after five catalytic cycles (see Fig. G). In addition, the loss of
active Zr species was not observed upon the reaction. These results
demonstrate that ZrO2(10)/SBA-15 catalyst acts as more efficient,
reusable heterogeneous catalyst for the CTH reaction of levuli-
nate esters to give GVL compared to the conventional bulk ZrO2.
Thermogravimetric (TG) analysis of the spent catalyst indicated
a deposition of appreciable amount (ca. 5 wt%) of carbonaceous
species on the catalyst surface after its use (see Fig. G). Consider-
ing the reaction mechanism proposed above, the detected organic
residue should originate from the surface-bound alkoxide and 4-
HPE species accumulated on the surface, which likely prevent the
access of levulinate esters and alcohols to the active Zr sites and
thus retard the overall reaction rate. The observed difference in the
catalyst reusability between ZrO2(10)/SBA-15 and bulk ZrO2 cata-
lysts might be attributed to the different chemical environment of
Zr species, i.e. the highly-dispersed Zr species stabilized on high-
surface-area SBA-15 may suppress over-accumulation of organic
species owing to the site separation effect, leaving spaces where
the reactants can access, whereas bulk ZrO2 with low surface area
causes an easy accumulation of organic species due to the proximity
of each active site.
References
[1] A. Corma, S. Iborra, A. Velty, Chem. Rev. 107 (2007) 2411–2502.
[2] J.N. Chheda, G.W. Huber, J.A. Dumesic, Angew. Chem. Int. Ed. 46 (2007)
7164–7183.
[3] D.M. Alonso, J.Q. Bond, J.A. Dumesic, Green Chem. 12 (2010) 1493–1513.
[4] J.J. Bozell, G.R. Petersen, Green Chem. 12 (2010) 539–554.
[5] M.J. Climent, A. Corma, S. Iborra, Green Chem. 13 (2011) 520–540.
[6] M.J. Climent, A. Corma, S. Iborra, Green Chem. 16 (2014) 516–547.
[7] J.C. Serrano-Ruiz, R. Luque, A. Sepúlveda-Escribano, Chem. Soc. Rev. 40 (2011)
5266–5281.
[8] I.T. Horváth, H. Mehdi, V. Fábos, L. Boda, L.T. Mika, Green Chem. 10 (2008)
238–242.
[9] D.M. Alonso, S.G. Wettstein, J.A. Dumesic, Green Chem. 15 (2013) 584–595.
[10] J.M.R. Gallo, D.M. Alonso, M.A. Mellmer, J.A. Dumesic, Green Chem. 15 (2013)
85–90.
[11] D. Wang, S.H. Hakim, D. Martin Alonso, J.A. Dumesic, Chem. Commun. 49
(2013) 7040–7042.
[12] L.E. Manzer, Appl. Catal. A 272 (2004) 249–256.
[13] J.-P. Lange, J.Z. Vestering, R.J. Haan, Chem. Commun. (2007) 3488–3490.
[14] J.C. Serrano-Ruiz, D. Wang, J.A. Dumesic, Green Chem. 12 (2010) 574–577.
[15] F.M.A. Geilen, B. Engendahl, A. Harwardt, W. Marquardt, J. Klankermayer, W.
Leitner, Angew. Chem. Int. Ed. 49 (2010) 5510–5514.
[16] J.Q. Bond, D.M. Alonso, D. Wang, R.M. West, J.A. Dumesic, Science 327 (2010)
1110–1114.
[17] D.J. Braden, C.A. Henao, J. Heltzel, C.C. Maravelias, J.A. Dumesic, Green Chem.
13 (2011) 1755–1765.
[18] X. Hu, C.-Z. Li, Green Chem. 13 (2011) 1676–1679.
[19] Z. Wu, S. Ge, C. Ren, M. Zhang, A. Yip, C. Xu, Green Chem. 14 (2012) 3336–3343.
[20] A. Démolis, N. Essayem, F. Rataboul, ACS Sustainable Chem. Eng. 2 (2014)
1338–1352.
[21] K.-i. Tominaga, A. Mori, Y. Fukushima, S. Shimada, K. Sato, Green Chem. 13
(2011) 810–812.
4. Conclusions
[22] L. Peng, L. Lin, J. Zhang, J. Shi, S. Liu, Appl. Catal. A 397 (2011) 259–265.
[23] X. Hu, Y. Song, L. Wu, M. Gholizadeh, C.-Z. Li, ACS Sustainable Chem. Eng. 1
(2013) 1593–1599.
[24] Y. Kuwahara, T. Fujitani, H. Yamashita, Catal. Today 237 (2014) 18–28.
[25] Y. Kuwahara, W. Kaburagi, K. Nemoto, T. Fujitani, Appl. Catal. A 476 (2014)
186–196.
[26] P.P. Upare, J.-M. Lee, D.W. Hwang, S.B. Halligudi, Y.K. Hwang, J.-S. Chang, J. Ind.
Eng. Chem. 17 (2011) 287–292.
[27] K. Yan, C. Jarvis, T. Lafleur, Y. Qiao, X. Xie, RSC Adv. 3 (2013) 25865–25871.
[28] K. Yan, T. Lafleur, G. Wu, J. Liao, C. Ceng, X. Xie, Appl. Catal. A 468 (2013) 52–58.
[29] L. Deng, J. Li, D.-M. Lai, Y. Fu, Q.-X. Guo, Angew. Chem. Int. Ed. 48 (2009)
6529–6532.
[30] Z.-p. Yan, L. Lin, S. Liu, Energy Fuels 23 (2009) 3853–3858.
[31] M.G. Al-Shaal, W.R.H. Wright, R. Palkovits, Green Chem. 14 (2012) 1260–1263.
[32] A.M.R. Galletti, C. Antonetti, V. De Luise, M. Martinelli, Green Chem. 14 (2012)
688–694.
[33] J.M. Nadgeri, N. Hiyoshi, A. Yamaguchi, O. Sato, M. Shirai, Appl. Catal. A 470
(2014) 215–220.
[34] Y. Kuwahara, Y. Magatani, H. Yamashita, Catal. Today 258 (Part (2)) (2015)
262–269.
[35] J. Deng, Y. Wang, T. Pan, Q. Xu, Q.-X. Guo, Y. Fu, ChemSusChem 6 (2013)
1163–1167.
In this study, a series of ZrO2 catalysts supported on mesoporous
SBA-15 silica were prepared and examined in the CTH reaction
of biomass-derived levulinic acid and its esters to produce GVL
using several alcohols as hydrogen donors. Catalytic results, com-
bined with structural analyses using XRD, N2 physisorption, TEM
and Zr K-edge XAFS, demonstrated that the GVL production rate
correlates well with the dispersion state of Zr atoms supported
on silicas rather than other textural properties, more specifically,
the highly-dispersed Zr4+-oxide species with low-coordination
geometry is the dominant active species. Of the catalysts exam-
ined, ZrO2(10)/SBA-15 with 10 wt% ZrO2 content was found to
be the most promising catalyst combining a high catalytic activ-
ity, reusability and stability, which far outperformed those of the
conventional bulk ZrO2 catalyst. Furthermore, based on the com-
parative experiments, combined with detailed characterizations
using NH3/CO2-TPD and in-situ FTIR, a plausible reaction mech-
anism was proposed, where basic Zr OH site triggers the CTH
reaction involving a six-member ring transition state. The catalyst
[36] R.S. Assary, L.A. Curtiss, J.A. Dumesic, ACS Catal. 3 (2013) 2694–2704.
[37] Y. Kuwahara, W. Kaburagi, T. Fujitani, RSC Adv. 4 (2014) 45848–45855.
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016

G Model
CATTOD-10207; No. of Pages11
ARTICLE IN PRESS
Y. Kuwahara et al. / Catalysis Today xxx (2016) xxx–xxx
11
[38] M. Chia, J.A. Dumesic, Chem. Commun. 47 (2011) 12233–12235.
[39] X. Tang, L. Hu, Y. Sun, G. Zhao, W. Hao, L. Lin, RSC Adv. 3 (2013) 10277–10284.
[40] X. Tang, H. Chen, L. Hu, W. Hao, Y. Sun, X. Zeng, L. Lin, S. Liu, Appl. Catal. B 147
(2014) 827–834.
[41] L. Bui, H. Luo, W.R. Gunther, Y. Román-Leshkov, Angew. Chem. Int. Ed. 52
(2013) 8022–8025.
[42] J. Wang, S. Jaenicke, G.-K. Chuah, RSC Adv. 4 (2014) 13481–13489.
[43] Y. Kuwahara, W. Kaburagi, T. Fujitani, Bull. Chem. Soc. Jpn. 87 (2014)
1252–1254.
[45] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli, J.S. Beck, Nature 359 (1992)
710–712.
[46] F. Meneghetti, E. Wendel, S. Mascotto, B.M. Smarsly, E. Tondello, H.
Bertagnolli, S. Gross, CrystEngComm 12 (2010) 1639–1649.
[47] H. Jung, S.-M. Paek, J.-B. Yoon, J.-H. Choy, J. Porous Mater. 14 (2007) 369–377.
[48] M.I. Zaki, M.A. Hasan, L. Pasupulety, Langmuir 17 (2001) 4025–4034.
[49] J. Geboers, X. Wang, A.B. de Carvalho, R. Rinaldi, J. Mol. Catal. A 388-389
(2014) 106–115.
[50] Y. Zhu, S. Liu, S. Jaenicke, G. Chuah, Catal. Today 97 (2004) 249–255.
[44] D. Zhao, J. Feng, Q. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D.
Stucky, Science 279 (1998) 548–552.
Please cite this article in press as: Y. Kuwahara, et al., Catalytic transfer hydrogenation of biomass-derived levulinic acid and its esters
to ␥-valerolactone over ZrO2 catalyst supported on SBA-15 silica, Catal. Today (2016), http://dx.doi.org/10.1016/j.cattod.2016.05.016