Production of gamma-valerolactone from sugarcane bagasse over TiO2-supported platinum and acid-activated bentonite as a co-catalyst

Article abstract of DOI:10.1039/c5ra06180f

Nowadays, biomass utilization has become the center of attention for researchers worldwide and is driven by the depletion of global petroleum supplies for the production of energy and valuable chemicals while easing the atmospheric CO₂ burden. We propose here a green strategy for transforming sugarcane bagasse into gamma-valerolactone (GVL), an attractive platform molecule that can be further converted into a variety of chemical derivatives for wide use in industrial applications. Our recent strategy involves the solid acid-catalyzed hydrothermal conversion of cellulose and hemicellulose derived from biomass to give an aqueous solution comprising levulinic acid (LA), followed by catalytic hydrogenation of LA to GVL. Native and acid-activated bentonites were used as solid acid catalysts to promote hydrothermal conversion of cellulose and hemicellulose. The maximum achievable yield of LA was 159.17 mg per gram of oven-dried biomass for 60 min reaction at 473.2 K in the presence of a 2% acid-activated bentonite catalyst. Catalytic hydrogenation reactions of LA to GVL over 1% Pt@TiO₂ and acid-activated bentonite as a co-catalyst were performed at temperatures of 393.2-473.2 K and residence times of 120-360 min. The combined solid catalyst gave an attractive performance with respect to LA conversion (～100%) and GVL selectivity (95%) under milder reaction conditions in comparison to 1% Pt@TiO₂ without an acid co-catalyst. The spent catalyst could be reused for five consecutive hydrogenation cycles with a marginal decrease in the catalytic activity and GVL selectivity. Coke formation was believed to be the main cause of catalyst poisoning and calcination of the spent catalyst under a stream of pure oxygen at 723.2 K was applied for removing coke deposits from the active catalyst sites, thus restoring the catalytic performance.

Full text of DOI:10.1039/c5ra06180f

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Production of gamma-valerolactone from
sugarcane bagasse over TiO -supported platinum
2
Cite this: RSC Adv., 2015, 5, 41285
and acid-activated bentonite as a co-catalyst
a
b
a
b
Jindrayani N. Putro, Alfin Kurniawan, Felycia E. Soetaredjo,* Shi-Yow Lin,
b
a
Yi-Hsu Ju and Suryadi Ismadji*
Nowadays, biomass utilization has become the center of attention for researchers worldwide and is driven
by the depletion of global petroleum supplies for the production of energy and valuable chemicals while
2
easing the atmospheric CO burden. We propose here a green strategy for transforming sugarcane
bagasse into gamma-valerolactone (GVL), an attractive platform molecule that can be further converted
into a variety of chemical derivatives for wide use in industrial applications. Our recent strategy involves
the solid acid-catalyzed hydrothermal conversion of cellulose and hemicellulose derived from biomass
to give an aqueous solution comprising levulinic acid (LA), followed by catalytic hydrogenation of LA to
GVL. Native and acid-activated bentonites were used as solid acid catalysts to promote hydrothermal
conversion of cellulose and hemicellulose. The maximum achievable yield of LA was 159.17 mg per gram
of oven-dried biomass for 60 min reaction at 473.2 K in the presence of a 2% acid-activated bentonite
2
catalyst. Catalytic hydrogenation reactions of LA to GVL over 1% Pt@TiO and acid-activated bentonite as
a co-catalyst were performed at temperatures of 393.2–473.2 K and residence times of 120–360 min.
The combined solid catalyst gave an attractive performance with respect to LA conversion (ꢀ100%) and
2
GVL selectivity (95%) under milder reaction conditions in comparison to 1% Pt@TiO without an acid co-
Received 7th April 2015
Accepted 30th April 2015
catalyst. The spent catalyst could be reused for five consecutive hydrogenation cycles with a marginal
decrease in the catalytic activity and GVL selectivity. Coke formation was believed to be the main cause
of catalyst poisoning and calcination of the spent catalyst under a stream of pure oxygen at 723.2 K was
applied for removing coke deposits from the active catalyst sites, thus restoring the catalytic performance.
DOI: 10.1039/c5ra06180f
www.rsc.org/advances
to nd an alternative route to convert the hardly decomposed
cellulose and hemicellulose into ve- and six-carbon sugars that
Introduction
Over the past decade, the 2nd generation of liquid biofuels has can be run economically at the scale required. The existing
become the present global challenge in the energy eld. The thermochemical routes which are considered more economi-
primary focus of such research is to develop sustainable liquid cally and technically feasible than conventional biochemical
biofuels from lignocellulosic biomass in a competitive market route featuring enzymatic hydrolysis and fermentation for the
and environmentally sustainable route, which can be used to conversion of lignocellulosic biomass to fuels and commodity
replace nite fossil fuel resources and boost the economics of chemicals include the Fischer–Tropsch synthesis (FTS) and
1,2
biofuel production. Lignocellulosic biomass, in the form of bione processes, catalytic hydrotreatment, catalytic pyrolysis
5–11
agricultural crops, forestry residues or waste from whole food and hydrothermal liquefaction.
However, one the major
supply chains is being paid more attention worldwide as a obstacles faced in these thermochemical conversion routes is
potential resource for the production of liquid biofuels, thanks associated with the production of bio-crude containing high
to its renewability, high abundance and inherently carbon oxygen, nitrogen and sulfur, which renders the product with
3
–5
neutral nature. Lignocellulosic biomass consists of 35–50% undesirable properties as fuel including thermally and chemi-
5
cellulose, 20–35% hemicellulose and 10–25% lignin. All are cally unstable, corrosive and low energy density. Further phys-
natural polymeric materials which constitute the structural ical or chemical upgrading strategies are required in order to
integrity and much of the mass of plants. The challenge now is improve the quality of as-produced biocrude, which oen
involves a series of unit operations and extensive energy input
requirements for the processes performed under high temper-
ature conditions. Furthermore, the implementation of most, if
a
Department of Chemical Engineering, Widya Mandala Surabaya Catholic University,
Kalijudan 37, Surabaya 60114, Indonesia. E-mail: suryadiismadji@yahoo.com
b
not all, of the aforesaid thermochemical technologies for
industrial process scale-up is now under-performing.
Department of Chemical Engineering, National Taiwan University of Science and
Technology, No. 43, Sec. 4, Keelung Rd, Taipei 106, Taiwan, Republic of China
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41285

View Article Online
RSC Advances
Paper
There are two critical steps in the conversion of cellulosic steam and electricity for factory operation and stockpiling
components of lignocellulosic biomass into fuels and chemicals: becomes a common practice for handling the excess bagasse.
hydrolysis and dehydration processes. In the hydrolysis process, However, long-term storage of large quantities of bagasse in the
hemicellulose is converted into ve- and six-carbon sugars under uncovered stockpiles continues to be a menace to the environ-
catalysis by mineral acids (mostly sulfuric acid) while cellulose ment and surrounding communities through hazards associ-
ber is enzymatically hydrolyzed into glucose by commercial ated with spontaneous combustion, groundwater seepage and
cellulase and cellobiase enzymes. The resulting aqueous mixture generation of contaminated leachates. It is required to nd a
of fermentable sugars containing xylose, arabinose and glucose suitable use for the excess bagasse as a means of managing
then undergoes dehydration reaction, leading to formation of these hazards and minimizing potential environmental
levulinic acid (LA) and furan derivatives. In the next process, the impacts. SB is composed of cellulose, hemicellulose and lignin
aqueous solution containing LA undergoes catalytic reduction and more importantly this lignocellulosic biomass has signi-
with molecular hydrogen to form gamma-valerolactone (GVL), cant advantage that it does not interfere with food resources for
22,23,26
which serves as a high-value chemical intermediate for the chemicals and fuels production. Several studies
have
production of straight and branched alkanes and cycloalkanes demonstrated different methods for transforming sugars in
with appropriate molecular weight and structure for use in lignocellulosic biomass into sustainable platform chemicals
gasoline and jet fuel applications. Intensive research activities such as LA, GVL and furanic aldehydes (e.g., 5-HMF and
have sought to develop a simple yet highly efficient catalyst for furfural). For instance, the high sugar-containing juice derived
the hydrogenation of LA to GVL. Heterogeneous catalysts based from sweet sorghum can be utilized as a non-food source to
on noble metals such as Pd, Ru, Rh, Pt, Au and Ir and base metals produce platform molecules namely LA and 5-HMF by a
27
such as Fe, Cu, Co and Ni on carbon, zeolite or metal oxide microwave dielectric heating method. The maximum yield of
support materials have received much attention for practical LA (31.4%) was achieved from the sorghum sample treated for
8
,10,12–18
use,
mainly associated with their advantages such as ease 20 min centrifugation time in the presence of 2 M sulfuric acid
28
of catalyst separation, reusability, less damage to reactor and under 30 min microwave irradiation. Horvath and co-workers
relatively high turnover numbers (TONs). Homogeneous catalytic investigated several intermediates and different reaction path-
systems have also shown great potential for the hydrogenation of ways in the acid-catalyzed dehydration of fructose to 5-HMF in
LA to GVL in terms of catalyst activity and selectively under dimethyl sulfoxide solvent with or without sulfuric acid as
relatively mild reaction conditions. For instance, ruthenium/ catalyst. They found from isotopic labeling results that the
phosphine complexes, especially complexes with the facially reversible formation of 2,6-anhydro-b-D-fructofuranose inter-
coordinating tridentate ligand (TriPhos) have found application mediate played a crucial role in controlling two irreversible
as a highly active catalyst toward carbonyl compounds such as pathways toward desirable product (5-HMF) via a furanose route
esters and amides and even free carboxylic acids in the homo- or toward the unwanted side products via a pyranose route. In
19
geneous hydrogenation reaction. Robust iridium trihydride this study, we utilize non-food SB biomass as a potential raw
complexes of pincer ligands with TONs as high as 71 000 have material for the production of GVL via a two-step process
been tested for the synthesis of GVL from hydrogenation of involving a rst step for hydrothermal conversion of biomass to
biomass-derived LA under relatively mild reaction conditions by LA and a second step for hydrogenation of LA to GVL. In the rst
20
Zhou and his group and the GVL yield could reach 99%. Very step, subcritical water (hereaer called sub-CW) is used as a
recently, Tukacs and co-workers have reported the application of reaction medium, which is safe, non-toxic, inexpensive and
ruthenium catalysts modied with bidentate phosphine ligands readily available. The native and acid-activated bentonites are
and they showed that the proposed catalytic system could be used as heterogeneous solid acid catalysts in the rst step and
reused for ten consecutive runs while achieving full conversion of the catalyst performance study for given reaction is presented.
ꢁ
1
LA with a representative turnover frequency (TOF) of 21 233 h
On the other hand, titania-supported platinum catalysts
2
1
in solvent-, chlorine- and additive-free reaction environments.
(Pt@TiO ) in the absence or in combination with acid-activated
2
GVL has proved to demonstrate many of the important properties bentonite as a co-catalyst are utilized in the second step. The
of an ideal liquid fuel including renewable, easy and safe to store catalyst performance evaluation in the second step is based on
and transport, very low vapor pressure even at high temperature LA conversion, GVL selectivity and GVL yield. Furthermore, the
ꢂ
ꢂ
(
3.5 kPa at 80 C), high boiling point (207 C) and open cup ash effects of temperature and residence time on the extent of
ꢂ
22
point (96 C) and miscible with water. Therefore, the produc- hydrogenation of LA to GVL are investigated and discussed in
tion of GVL from lignocellulosic biomass has moved globally in this paper. Finally, studies were carried out to assess the
large numbers and very recent potential applications include a potential reuse of the hydrogenating catalyst in at least ve
green reaction medium for the conversion of various carbohy- consecutive reaction cycles.
22,23
drates to 5-hydroxymethylfurfural (5-HMF) and LA,
ionic
liquid phase catalytic system for the hydrogenation of different
Experimental
Materials
24
olens, precursors in manufacturing bio-based polymers and
25
23
resins or chemical additives for gasoline and diesel fuel.
Sugarcane bagasse (SB) is a major agro-industrial byproduct SB was collected from a cane sugar mill located at Klaten district
generated from the sugar milling. In the sugar milling process, in Central Java. The collected biomass was repeatedly washed
a part of this brous residue is normally burned to generate with tap water to remove surface impurities and dried in an
41286 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
ꢂ
oven at 80 C for 12 h. The dried biomass was pulverized using a Sigma Aldrich), sulfuric acid (98%, Merck), chloroplatinic acid
bench-scale hammer mill to particle sizes ranging from 0.15– hexahydrate (8 wt% in H O, Aldrich), acetonitrile (LC-MS
2
0.18 mm (80/100 mesh, U.S. Sieve Series). Moisture content of CHROMASOLV®, 99.9%, Fluka), hexyl alcohol (99.9%, Fluka)
the pulverized biomass was determined by using an Ohaus® and standards of D(+)glucose (99.5%, Sigma), D(+)xylose (99%,
MB-35 moisture analyzer. Delignication step was conducted by Aldrich), L(+)arabinose (98%, Aldrich), D(+)galactose (99%,
soaking powdered biomass into a 20 wt% sodium hydroxide Fluka), furfural (98.5%, Fluka), 5-HMF (98%, Fluka), LA (98%,
solution with a solid to solution ratio of 1 : 10 under continuous Aldrich) and GVL (98.5%, Fluka) were directly used without
stirring (500 rpm) at room temperature for 24 h. The treated further purication. Ultra-high purity nitrogen and hydrogen
biomass was separated by vacuum ltration, washed with warm gases (99.9%) were supplied by a local company. The ultrapure
distilled water until the pH of the washing solution ranged water obtained from a Milli-Q water ltration station (Millipore
ꢂ
between 6.5 and 7 and oven-dried at 110 C for 12 h. The nal Corp., Milford, MA) was used in all experiments unless other-
product (DSB) was stored in sealed plastic bags for further use. wise stated.
The determination of lignocellulose contents of SB and DSB
samples was carried out following the standardized wet chem-
ical methods used in the wood industry sector. The percentage
weights of the elements C, H, N, O and S of SB and DSB samples
were determined by a CHNS/O analyzer model 2400 from
Perkin-Elmer. The proximate analysis was conducted following
the established ASTM procedure. The results from three
repeated analyses are presented as average values ꢃ standard
deviations (Table 1).
Preparation of catalyst materials
Natural Ca-bentonite was utilized as a heterogeneous solid
catalyst in this study due to its mechanical and hydrothermal
stability, low swelling capacity, high cation exchange capacity
ꢁ1
(
CEC, 58.3 meq. 100 g clay according to methylene blue test
method designed in ASTM C837-99), environmentally benign
and easily obtainable material. The acid treatment of bentonite
aims to enhance its adsorption and catalytic properties for the
hydrolysis of cellulosic components of biomass into water-
soluble saccharides. The acid-activated bentonite was
prepared based on the following procedure: 50 g of puried
bentonite was immersed in a 500 ml solution of sulfuric acid
Natural Ca-bentonite with grayish white color was collected
from a mining site located at Pacitan district in East Java. Prior
to its use, the bentonite was puried according to the procedure
29
described by Soetaredjo and colleagues. The chemical
composition of the puried bentonite was SiO (63.5%), Al O
2 3
2
(
2 N) and transferred into a round-bottom ask. The resulting
(
17.8%), Fe O (2.2%), MnO (0.5%), TiO (0.2%), MgO (1.2%),
2 3 2 2
ꢂ
suspension was heated at 80 C under reux and stirring for 4 h.
Then, the solid was separated by vacuum ltration, washed with
K O (1.6%), Na O (1.5%), CaO (2.8%) and LOI (loss on ignition)
2
2
(8.7%). The montmorillonite content of the bentonite was 78%.
warm distilled water until free of sulphate ions (tested by BaCl
solution) and dried in an oven at 110 C for 6 h. The acid-
activated bentonite was crushed and sieved to particle sizes of
2
Anatase titanium(IV) oxide (TiO ) powder (99%) was purchased
2
ꢂ
from Sigma Aldrich, Singapore and used as-received. For
chemical analysis and preparation of catalysts, the following
reagents include anhydrous sodium hydroxide pellets (98%,
1
25–150 mm. Here, the untreated and acid-activated bentonite
samples were designated as UB and AAB, respectively. The
titania-supported 1 wt% Pt catalyst (1% Pt@TiO ) was prepared
2
by incipient wetness impregnation method based on Zhang and
co-workers procedure.
Table 1 Characteristics of SB and DSB lignocellulosic materials
30
Components
SB
DSB
Characterizations of biomass precursors and catalyst
materials
Hemicellulose
Cellulose
Lignin
24.7 ꢃ 2.5
35.4 ꢃ 1.8
21.8 ꢃ 1.2
18.8 ꢃ 0.9
52.7 ꢃ 2.1
10.3 ꢃ 0.8
The thermal stability of SB and DSB samples was determined by
thermal gravimetric analysis (TGA) using a Mettler-Toledo
TGA/DTA 1 thermal analyzer with a heating and cooling rate
Ultimate compositions, wt% (on dry basis)
Carbon
Hydrogen
Nitrogen
Sulphur
Oxygen (by difference)
53.2
6.1
0.5
1.8
38.4
53.8
6.2
0.6
1.7
37.7
ꢁ1
of 10 K min up to 1073 K under continuous nitrogen gas ow
ꢁ1
of 150 ml min . The mass of the samples in each experiment
was about 10 mg. The surface morphology of biomass precur-
sors and catalyst materials was visualized by scanning electron
microscope (SEM), on a JEOL JSM-6390F operated at an accel-
erating voltage of 10 kV. Prior to SEM imaging, an ultra-thin
layer of conductive platinum was sputter-coated on the speci-
mens using an auto ne coater (JFC-1200, JEOL, Ltd., Japan) for
120 s in an argon atmosphere. The powder X-ray diffraction
Proximate compositions, wt% (on dry basis)
Volatile matter
Fixed carbon
Moisture content
Ash content
66.1 ꢃ 2.4
15.8 ꢃ 0.5
5.9 ꢃ 0.7
12.2 ꢃ 1.2
64.8 ꢃ 1.7
17.0 ꢃ 1.2
6.5 ꢃ 0.3
11.7 ꢃ 0.7
(XRD) patterns were acquired on a Philips PANalytical X'Pert
ꢁ1
Sugar compositions, mg g (on dry basis)
Glucan
Xylan
Galactan
Arabinan
powder X-ray diffractometer with a monochromated high-
intensity Cu Ka1 radiation (l ¼ 0.15406 nm) operating at
348.7 ꢃ 5.9
507.8 ꢃ 2.7
165.6 ꢃ 3.5
1.9 ꢃ 0.2
216.0 ꢃ 4.4
2.5 ꢃ 0.3
ꢂ
ꢁ1
40 kV, 30 mA and a step size of 0.05 s in the 2-theta scan
ꢂ
27.6 ꢃ 2.5
20.1 ꢃ 1.3
range of 5–90 . The peak locations corresponding to the
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41287

View Article Online
RSC Advances
Paper
crystalline phases were compared with standard JCPDS les. temperatures of 393.2–473.2 K and residence times were varied
The specic surface areas and pore structures of the catalysts between 120 and 360 min. During the course of the reaction,
were characterized by nitrogen adsorption–desorption mixing was achieved through an internal 4-blade impeller
isotherms. The adsorption and desorption isotherms were operating at 1200 rpm. Preliminary experiments showed that
conducted at the boiling temperature of liquid nitrogen (77.2 K) this stirring speed was sufficient to avoid external mass transfer
using an automated Micromeritics ASAP2010 sorption appa- limitation, thus ensuring the reaction operates in the kinetic
ratus. Before measurements, the samples were outgassed under regime. Aer the predetermined reaction time, the reactor was
vacuum for 6 h at 473.2 K. The Brunauer–Emmett–Teller (BET) rapidly cooled down to room temperature and the residual
equation was used to calculate the specic surface area from gases were discharged. The reaction mixtures containing liquid
o
adsorption data in the relative pressure (P/P ) range of 0.01– and solid fractions were centrifuged at 3000 rpm for 15 min and
0
.30. Total pore volume (V ) was determined from the amount then vacuum-ltered to obtain a clear solution.
T
of adsorbed gas at a relative pressure of 0.995.
HPLC analysis of the product mixture from hydrothermal
reaction
Catalytic hydrothermal conversion of DSB biomass
The hydrothermal conversion of DSB biomass was carried out in The aqueous phase containing various water-soluble reaction
a 150 ml lab-scale high pressure batch reactor. The high pres- products from hydrothermal liquefaction of DSB was analyzed
sure batch reactor was equipped with a pressure gauge, an using a Jasco chromatographic separation module consisting of
external electrical heating system, a Type K thermocouple as a model PU-2089 quaternary low pressure gradient pump, a
internal temperature control device and M8 screws for tight- model RI-2031 refractive index detector and a model LC-NetII/
ening the reactor with its cap. The maximum allowable oper- ADC hardware interface system. All solution samples were
ating pressure and temperature of the reactor are 10 MPa and ltered through a 0.22 mm PVDF syringe lter prior to injection
523.2 K, respectively. Typical hydrothermal conditions were in the HPLC. The analysis of monomeric sugars including
described as follows: temperatures 393.2–473.2 K, catalyst glucose, fructose, xylose, galactose and arabinose was con-
loading 0–2% (w/v), DSB to water ratio 1 : 10 (w/v), stirring ducted with an Aminex HPX-87P sugar column (Bio-Rad, 300 ꢄ
speed (300 rpm) and internal pressures (3–5 MPa). The hydro- 7.8 mm) using degassed HPLC-grade water isocratically owing
ꢁ
1
thermal reaction was conducted at desired temperatures for at a rate of 0.60 ml min . The column was operated at 358.2 K
0 min and aer the completion of reaction, the reactor was and detection of effluent sugars was performed by using a RI-
6
rapidly cooled down to room temperature by quenching in a 2031 detector. For the analysis of simple organic acids (i.e.,
cold water bath and the gas pressure was recorded and vented. formic, acetic and levulinic acids) and furan derivatives (i.e., 5-
The solid fraction was separated by centrifugation at 3000 rpm HMF and furfural), the separation of compounds was achieved
for 15 min. An aliquot of liquid samples obtained aer the with a Bio-Rad Aminex HPX-87H column (300 ꢄ 7.8 mm) using
reaction was withdrawn and the concentrations of 5-HMF, isocratic elution of sulfuric acid aqueous solution (5 mM) as the
ꢁ
1
furfural, LA, organic acids and monomeric sugars were quan- mobile phase at a ow rate of 0.60 ml min . The RI-2031
tied by using high performance liquid chromatography detector was used to identify the compounds and the column
(
HPLC) as described later.
oven was set to 328.2 K. The concentrations of each compound
in the product mixture were determined using the calibration
curves obtained from the injection of solutions of authentic
samples with known concentrations.
Catalytic hydrogenation of hydrothermal liquor containing LA
The catalytic hydrogenation reactions were performed in a 250
ml Parr 4576A bench top stirred HP/HT reactor equipped with a
temperature controller unit (model P.I.D. 4843), a pressure
transducer and gage and overhead stirring. Before use, the
Gas chromatography analysis of the reaction products
containing GVL
2
catalyst was activated under owing H at 573.2 K and atmo- The reaction products containing GVL were quantitatively
spheric pressure for 3 h. In a typical procedure, the aqueous analyzed by using a Shimadzu GC-2014 gas chromatograph
solution containing LA obtained from hydrothermal reaction at equipped with a Phenomenex Zebron™ ZB-Wax capillary
4
53.2 K with 2% AAB catalyst was introduced into the reactor, column (30 m length ꢄ 0.32 mm i.d., 0.25 mm lm thickness)
followed by the addition of 0.5 g of catalyst comprising 0.4 g 1% and a ame ionization detector (FID). The carrier gas was ultra-
Pt@TiO and 0.1 g AAB as an acid co-catalyst. An equivalent high purity (UHP) grade helium (99.99%) with a constant ow
amount of 1% Pt@TiO
2
ꢁ
1
2
hydrogenation catalyst (0.4 g) was also rate of 1.0 ml min . The injection volume was 1.0 ml in splitless
used in the reaction. In order to remove air inside the reaction mode. The following temperature program was used in the
ꢂ
vessel, the reactor was slowly ushed with argon three times analysis: an isothermal step at 60 C (held for 3 min), which was
ꢂ
ꢁ1
ꢂ
before hydrogen gas was charged. Aerward, the reactor was ramped at a rate of 10 C min to 230 C and held for 5 min.
pressurized with hydrogen gas to 3 MPa and sufficient The injector and detector temperatures were kept at 250 and
ꢂ
temperature ramping time was allowed before the reaction step 300 C, respectively. The identication of reaction products
at the chosen temperature. Zero reaction time was dened when was accomplished by comparison of sample peak retention
the temperature inside the reactor reached the chosen times with those of authentic standards analyzed under
temperature. The hydrogenation reactions were performed at identical conditions. Standard solutions encompassing the
41288 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
concentration range of the samples were used to construct the methods and the established protocols of National Renewable
calibration curves for quantifying the concentrations of the Energy Laboratory (NREL). The thermogravimetric curve of DSB
compounds of interest.
sample shows a similar characteristic to that of SB sample with
higher onset of degradation temperatures for hemicellulose
and cellulose at 306 and 364 C, respectively. This indicated that
ꢂ
Results and discussion
the treated biomass exhibited higher thermal stability due to
the lignin and hemicellulose removal. Maximum rates of weight
The characteristics of SB and DSB samples include elemental
and proximate compositions and their lignocellulose contents
are presented in Table 1. In order to examine the thermal
stability of biomass precursor, TGA was conducted under
nitrogen environment by raising the temperature gradually
from ambient to an upper limit of 1073.2 K, where it is safely
assumed that all the elemental oxygen and hydrogen are liber-
ated to gaseous compounds, thus leaving only elemental
carbon. TGA result indicates that SB is a potential feedstock for
energy and chemical production due to high cellulose (35.4%)
and hemicellulose (24.7%) contents, which are the main sour-
ces of fermentable sugars. From TGA curves shown in Fig. 1, it is
clearly seen that SB sample undergoes a gradual thermal
decomposition process with a signicant weight loss happening
ꢂ
loss occur between 270 and 380 C for delignied biomass.
Again, the amount of weight loss was insignicant at high
ꢂ
temperatures between 500 and 800 C, yielding residual carbon
of 17.2%. The individual monosaccharides include glucose,
xylose, galactose and arabinose present in the hydrolysate
solution obtained from dilute acid hydrolysis (5% HCl under
ꢂ
reux at 96 C for 2 h) of SB and DSB samples were quantied
34
with a chromatographic procedure of du Toit et al. and the
ꢁ
1
monomeric sugar yields, in the unit of mg g dry biomass,
were given as follows: glucose (387.4) for SB and (564.2) for DSB,
xylose (245.5) for SB and (188.2) for DSB, galactose (2.8) for SB
and (2.1) for DSB and arabinose (31.4) for SB and (22.8) for DSB.
In this respect, the concentrations of the polymeric sugars
corresponding to its monosaccharide units (Table 1) are
calculated according to NREL conversion formula.
ꢂ
at temperatures between 260 and 360 C. The weight loss of
about 60% corresponds to the rapid breakdown of hemi-
cellulose and cellulose fractions with the DTG peak tempera-
Lignin can act as a physical barrier, encapsulating and
conning cellulose and this complex three-dimensional
aromatic polymer architecture is highly recalcitrant toward
enzymatic, chemical and biological degradation. Therefore,
pretreatment strategy to promote efficient hydrolysis is an
essential step in the lignocellulosic biomass-processing bio-
reneries to reduce the inherent lignocellulosic cell wall recal-
citrance for cost-effective downstream processing. Since the
biomass feedstock contains a quite high portion of lignin
ꢂ
tures at 295 and 352 C, respectively. The degradation of lignin
might also happen over this temperature range, which is
primarily caused by the breakdown of chemical bonds with low
31
activation energy. A minor weight loss about 5% occurring
ꢂ
between 400 and 500 C is attributed to the lignin degradation
reaction involving the breakdown of more stable bonds. The
temperature regions where depolymerization of hemicellulose
and the random cleavage of the glycosidic linkages in cellulose
structure occur are in good agreement with the studies of Yang
(21.8%), delignication pretreatment with NaOH solution is
31
32
33
employed to effectively remove lignin from the biomass, at the
expense of parts of the structural carbohydrates are decom-
posed. Aer delignication, the lignin content signicantly
decreased from 21.8% in the native biomass to 10.3% in the
treated biomass. In line with this, the concentrations of poly-
meric hemicellulose sugars including xylan, galactan and ara-
et al., Garcia-Perez et al. and Cao and Aita. At high
ꢂ
temperatures between 500 and 800 C, the weight loss signal
was not very evident, approaching a value for the charcoal yield
of 15.7%. The cellulose, hemicellulose and lignin contents (on a
dry basis) of SB sample obtained from TGA method are in good
agreement with the values determined from “wood-industry”
ꢁ
1
binan in the treated biomass were 165.6, 1.9 and 20.2 mg g
dry biomass, respectively, indicating that alkaline delignica-
tion brought hemicellulose dissolution. On the other hand,
most of the glucan portion was preserved during delignication
pretreatment, which could be ascribed to the robust crystalline
structure and low reactivity of cellulose with alkali. For
comparison purposes, the alkaline delignication conditions of
various lignocellulosic biomass resources are listed in Table 2.
From this table, it can be shown that delignication efficiency
varies greatly with process conditions and types of the biomass
feedstocks. In general, alkali pretreatment is more effective for
removing a substantial portion of lignin from hardwood
biomass and agriculture waste products. Moreover, the
delignication efficiency under conditions described in this
study (52.8%) compares well with other studies performed at
higher temperatures (e.g., 368.2, 394.2 and 433.2 K) and shorter
residence time (e.g., 1, 1.5 and 3 h) in the absence of stirring.
The main chemical mechanism by which the lignin polymer
degrades in alkaline and alkaline-oxidative environments has
Fig. 1 Thermogravimetric curves of sugarcane bagasse before (SB)
and after NaOH delignification pretreatment (DSB).
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41289

View Article Online
RSC Advances
Paper
Table 2 Summary of alkaline delignification conditions of various lignocellulosic biomass resources
Delignication conditions
Solid–liquid
ratio (w/v)
Temperature
(K)
Residence
time
Delignication
efficiency (%)
Lignocellulosic resources
Alkali
NH OH
Stirring speed (rpm)
3
3
Sugarcane bagasse – Aita et al.
4
1 : 20
1 : 15
1 : 2
433.2
333.2
328.2
433.2
394.2
358.2
393.2
433.2
328.2
368.2
Room
1 h
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
500
14
60
87.5
44
65.6
76.1
67.8
28.8
23.1
24.4
52.8
35
Corn stalk – Yu et al.
Corn stover – Kim and Holtzapple
NaOH
1.5 h
4 weeks
1 h
1.5 h
24 h
24 h
Not available
3 h
36a
Ca(OH)
2
37a
Sorghum – Salvi et al.
Cotton stalk – Silverstein et al.
NH
4
OH
2 : 1
38
NaOH
NaOH
1 : 10
1 : 10
1 : 10
1 : 4
5 : 1
10 : 1
1 : 10
3
9
Switchgrass – Gupta and Lee
NH
GL
NaOH
Ca(OH)
NaOH
4
OH
40b
Mixed hardwood chips – Yu et al.
Rice straw – Cheng et al.
41c
2
3 h
24 h
Sugarcane bagasse – this study
a
b
The loading ratios of alkali to raw biomass are given on a mass basis. GL (green liquor) is an alkaline solution prepared by mixing sodium
c
ꢁ1
carbonate and sodium sulde with a suldity of 25%. The loading ratio is expressed as g H
of 4 wt% for NaOH and 10 wt% for Ca(OH)
2
O g oven-dried rice straw with an alkali loading
2
.
been mechanistically studied and is through the cleavage of size distribution of the solid aggregates ranged from tens of
hydrolysable a- and b-aryl ether bonds in lignin and glycosidic microns to submicron. The electron micrograph clearly reveals
bonds in polysaccharides. Meanwhile, hemicellulose dissolu- that spherical particles were not likely formed in the hydro-
42
tion and cellulose swelling are a consequence of hydrogen bond thermal processing at low temperatures. Titirici and group
weakening. The post-pretreatment biomass washing step by obtained sphere-shaped and monodisperse hydrothermal
warm or hot distilled water aims to remove the alkali residues, carbons derived from pure glucose at 533.2 K while the overall
soluble phenolics and other degradation products formed during morphology of hydrothermal carbons derived from cellulose
lignin depolymerization. In this regard, the concentration of and rye straw (a model of lignocellulosic biomass) was not as
alkali metal ion remained with the washed biomass was very low homogeneous as from pure glucose, with some structural
or absent aer repeated washing step. The pretreatment liquor components of the precursor material still remained intact
from delignication step containing dissolved lignin, lignin- and arranged together with spherical carbon particles. The
derived soluble phenolics, acetic acid, hemicellulose-derived elemental composition analysis was conducted by an INCA
poly/oligosaccharides, proteic compounds and other extractives energy dispersive X-ray microanalysis system integrated with
can be reclaimed for further applications such as production of the hardware of the SEM and result showed that the solid
bioethanol or biohydrogen.
Alkaline delignication pretreatment had a notable effect on
the biomass morphology. Images obtained by SEM show some
visible changes associated with the biomass wall disruption
aer delignication. The rod-like morphology of native biomass
(SB) was easily recognizable in Fig. 2(a) and the untreated SB
had a relatively smooth and intact surface. Delignication
pretreatment by NaOH resulted in the disrupted surface struc-
ture due to the breakdown of lignin as shown in Fig. 2(b) and
therefore the surface of delignied biomass (DSB) appears
rougher when compared to the native biomass. Moreover, the
structure of DSB biomass was opened up as a result of lignin
removal, indicating the more available or exposed surface area
of cellulose for hydrolysis. Fig. 2(c) depicts the morphological
features of DSB biomass wall aer sub-CW treatment. The
surface structure of sub-CW treated DSB biomass changed
signicantly when compared with DSB sample, showing a
sponge-like structure. The small pores/cavities formed on the
wall of sub-CW treated DSB sample might be the result of
release of hemicellulose or cellulose decomposition products
from the interior of the biomass particle. Fig. 2(d) shows the
aggregates of irregularly-shaped, non-uniform solid particles
obtained from hydrothermal conversion of DSB sample. The
Fig.
bagasse – SB (a) and after NaOH delignification pretreatment – DSB
b), morphological structure of biomass (DSB) wall following sub-CW
treatment (c) and solid ‘humin’ byproduct obtained from hydrothermal
conversion of DSB biomass (d).
2 Scanning electron micrographs of untreated sugarcane
(
41290 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
ꢂ
aggregates were comprised of elemental carbon (64.8 at%) and phases. The reection at 2q ¼ 6.7 is correlated to the basal
oxygen (35.2 at%). spacing (001) of montmorillonite, corresponding to a value of
The morphological features of the catalysts are depicted in 1.32 nm. Compared to UB catalyst, the position of the (001)
Fig. 3. From SEM images, it can be seen that UB catalyst is basal reection in AAB catalyst remains essentially unchanged
ꢂ
aluminosilicate clay with a layered structure and smooth (2q ¼ 6.8 , 1.30 nm), showing the minor effect of dilute acid
surface. The specic surface area of UB catalyst calculated from treatment on the lattice structure of clay. As it can be seen from
2
ꢁ1
ꢂ
the BET method was 61.8 m g and its total pore volume was Fig. 4(b), the bare TiO exhibits reections at 2q angles of 25.5 ,
8
2
ꢁ
2
3
ꢁ1
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
ꢂ
.3 ꢄ 10 cm g . Nitrogen adsorption–desorption isotherms 37.2 , 38.0 , 48.1 , 54.4 , 55.2 , 62.7 and 68.9 in which all
revealed a classical combination of micro/mesoporous struc- these reection peaks can be well indexed to the (101), (103),
tures in the catalyst particle. Aer chemical modication by (004), (200), (105), (211), (204) and (116) crystal planes of TiO2
sulfuric acid, some evident changes in the platelet morphology with anatase phase according to JCPDS database card no. 21-
of UB catalyst were observed. Generally speaking, an acid 1272. The diffraction lines for (111), (200) and (220) of Pt cubic
ꢂ
ꢂ
ꢂ
treatment results in the replacement of exchangeable cations phases were located at about 2q angles of 39.8 , 46.4 and 67.5
+
+
2+
+
that are between the layers (i.e., Na , K or Ca ) with H ions of (JCPDS database card no. 04-0802). Result showed that the
the acid and depending on the treatment conditions, partial diffraction pattern of 1% Pt@TiO was essentially the same as
delamination of Al-octahedral and Si-tetrahedral layers, fol- that of diffraction pattern from bare anatase TiO . Similar
lowed by subsequent dissolution of structural cations (typical observations have been noted by Zhang et al. and Moonsiri
, Mg or Al in the Al-octahedral layer and et al. in their studies. The absence of diffraction signals from
Al or Si in the Si-tetrahedral layer). The small difference in metallic Pt species could be attributed to the small size and
2
2
30
2
+/3+
2+
3+
45
examples are Fe
3
+
4+
43
CEC between UB (58.3) and AAB (57.5) samples suggests that the highly distributed Pt particles supported on TiO . Zhang and co-
2
30
structural cations in the Al-octahedral and Si-tetrahedral layers workers observed that Pt particles were highly dispersed on
are marginally affected by dilute acid attack. The surface acidity TiO support into a size smaller than 1 nm from high-resolution
or catalytic acid sites) of the bentonite, as determined by a transmission electron microscope (HRTEM) image. Hence, we
2
(
44
2
simple acid–base titration method, increased dramatically believe that in case of fresh 1% Pt@TiO catalyst, the average Pt
ꢁ
1
from 30.7 (UB) to 118.2 meq. 100 g clay aer treatment by 2 N particle sizes are around 1 nm (or even smaller) since the
sulfuric acid. Results of nitrogen adsorption–desorption catalyst preparation in our study was adopted from Zhang et al.
measurements for AAB catalyst show an increase in the BET procedure. The high dispersion of nano-sized Pt particles on
specic surface area and total pore volume. The BET specic anatase TiO support could be very important for high activity
2
surface area and total pore volume of AAB catalyst are and selectivity of this catalyst in the hydrogenation of LA.
2
ꢁ1
3
ꢁ1
1
22.4 m
higher compared to those values of UB catalyst. The scanning building block chemicals for clean liquid fuels, the cellulose
electron micrograph of 1% Pt@TiO in Fig. 3(b) shows that the and hemicellulose components of biomass were hydrolyzed to
g
and 0.12 cm
g
, which is ꢀ2- and ꢀ1.5-fold
In the hydrothermal reaction with the aim of producing
2
catalyst is spherical-like irregular particles with average diam- six and ve carbon sugar monomers, followed by dehydration
eter sizes are in the submicrometer range (150–200 nm). The reactions of these monomeric sugars to produce 5-HMF and
BET specic surface area of 1% Pt@TiO was lower compared to furfural in water under subcritical condition. Given the fact that
2
2
ꢁ1
bare TiO (48.4 vs. 51.3 m g ), likely due to the deposition of sub-CW possesses the catalytic role of an acid or base and has
2

ne Pt particles on the surface of the support that might block been widely used as an alternative medium for the conversion of
pores of TiO
2
.
lignocellulosic biomass into platform chemicals, Table 3 pres-
The wide angle XRD patterns of AAB and 1% Pt@TiO
2
cata- ents the conversion yields of cellulose and hemicellulose
lysts are shown in Fig. 4. From the XRD spectrum of AAB cata- components of DSB biomass into a variety of reaction products
ꢂ
ꢂ
ꢂ
ꢂ
lyst, the montmorillonite (2q ¼ 6.7 , 20.2 , 21.8 , 35.1 and including sugars, 5-HMF, furfural and organic acids in the
ꢂ
ꢂ
ꢂ
ꢂ
6
1.8 ) and quartz (2q ¼ 26.8 , 50.4 and 68.3 ) are the major absence or presence of solid acid catalysts. From this table, it is
phases with other minor reections corresponding to mineral easily noticed that the performance of non-catalyzed, mildly
impurities of crystobalite, calcite, K-feldspar, mica and/or illite sub-CW treatment for cellulose hydrolysis is not so impressive
ꢁ
1
with glucose yield ranged between 32.74 and 54.18 mg g dry
biomass (3.27–5.42 wt%) in the temperature range of 393.2–
473.2 K for 60 min reaction. The incomplete and slow conver-
sion rate of cellulose into glucose under non-catalyzed, mildly
sub-CW treatment might be due to the fact that hydrolysis
reaction was limited at the surface region without cellulose
swelling and dissolution compared to the reaction taking place
46
in high-density regions (near-critical and supercritical water)
as well as insufficiently high temperature to disrupt the robust
structure of cellulose. Hemicellulose-derived monosaccharides
were also present in the reaction product mixture, accounting
for 3.3–4.1 wt% yields. The quantities of 5-HMF and LA were
barely detectable in the reaction product mixture, affording
Fig. 3 Scanning electron micrographs of AAB clay (a) and 1% Pt@TiO
catalysts (b).
2
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41291

View Article Online
RSC Advances
Paper
2 2
Fig. 4 XRD patterns of AAB (a) and bare TiO and 1% Pt@TiO catalysts (b).
Table 3 Non-catalytic and catalytic results of sub-CW treatment of DSB biomass in the presence of 2% (w/v) UB or AAB clay-based catalysts at
various temperatures for reaction time of 60 min
ꢁ1
Yields (mg g dry biomass)
a
a
a
Sub-CW treatments
Non-catalytic
T (K)
Glucose
Xylose
Arabinose
Galactose
5-HMF
LA
FA
Furfural
AA
393.2
32.74
35.56
42.33
54.18
50.69
38.38
46.28
54.19
89.41
78.56
92.00
115.71
145.06
216.75
190.46
28.36
29.70
32.41
35.69
33.25
41.47
43.01
47.45
63.08
54.25
58.64
64.82
72.33
95.30
86.71
4.24
4.44
4.83
5.33
5.02
6.20
6.43
7.09
9.42
8.18
8.76
9.68
10.81
14.35
12.62
0.28
0.30
0.32
0.35
0.37
0.42
0.44
0.49
0.67
0.63
0.62
0.67
0.75
0.93
0.91
0.76
0.90
1.16
1.67
1.49
2.18
2.82
3.60
6.79
0.08
0.11
0.16
0.28
0.23
0.44
0.65
0.89
3.14
0.01
0.02
0.04
0.09
0.12
0.14
0.23
0.34
1.38
1.59
9.10
14.30
23.68
55.09
61.72
0.52
0.61
0.74
0.95
0.88
5.49
6.64
8.03
14.38
13.15
28.90
32.90
37.78
53.87
46.48
0.58
0.71
0.87
1.02
1.14
0.80
0.93
1.09
1.17
1.32
1.41
1.64
1.90
2.23
2.56
4
4
4
4
13.2
33.2
53.2
73.2
UB-catalyzed
AAB-catalyzed
393.2
4
4
4
4
13.2
33.2
53.2
73.2
5.54
2.76
393.2
23.18
30.78
41.63
80.41
65.69
30.52
43.64
74.78
159.17
125.44
4
4
4
4
13.2
33.2
53.2
73.2
a
LA, FA and AA represent levulinic acid, formic acid and acetic acid, respectively.
yield less than 0.2 wt%. In pure sub-CW medium, the dehy- fragmentation and cross-polymerization reactions involving
47
dration of glucose into 5-HMF is non-selective, leading to sugars, water molecules and various reactive intermediates. The
formation of multiple side-products and insoluble carbona- use of heterogeneous solid acid catalysts for enhanced cellulose
ceous residue ‘humin’. The reaction mixture aer sub-CW conversion and also selective and high-yielding production of 5-
treatment also turned into dark brown, which could be HMF and LA has been coming up to a promising direction of
ascribed to the formation of soluble and insoluble polymeric research. In this regard, clays and clay minerals can catalyze
side-products. These polymeric side-products were not charac- chemical reactions via Bronsted and/or Lewis acid sites on their
terized in this study due to several analytical issues. The pres- surfaces although native or natural clays generally show low or
ence of furfural was detected in the reaction product mixture almost no catalytic activity prior to acid or cation exchange
ꢁ1
and a maximum yield of 0.95 mg g was achieved at 453.2 K. treatments. As expected, the cellulose conversion catalyzed with
The furfural yield for non-catalyzed, mildly sub-CW treatment UB solid slightly enhanced compared with blank reaction (no
ꢁ
1
in this study compares well with results in the literature for a catalyst) and the glucose yield reached only 89.41 mg g dry
one-stage process where hydrolysis and dehydration reactions biomass for 60 min reaction at 453.2 K. Low catalytic activity of
occur simultaneously within the same reactor.
untreated bentonite might lead to speculation that the solid
In the absence of catalysis, the hydrothermal degradation acid catalyst is not active enough for promoting hydrolysis of
rate of cellulose to glucose is not signicant. It was later cellulose, possibly attributed to: (1) the surface of solid acid
48
revealed by Kuster that the enolization rate of glucose, a rate- catalyst cannot readily adsorb or attach to cellulose and b-1,4
determining step for 5-HMF formation, is slow in water and glucan in water or (2) the absence of freely available, strong
intermediate 5-HMF is more likely to undergo a series of Bronsted acid sites for decomposing hydrogen bonds and
unwanted side reactions such as condensation, rehydration, hydrolyzing b-1,4 glycosidic bonds in cellulose. Under the same
41292 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
reaction condition, 5-HMF and LA with the corresponding
ꢁ
1
yields of 6.79 and 3.14 mg g dry biomass were produced.
Based on the results obtained, it could be inferred that the
product distribution was affected by addition of UB catalyst
while minor improvement on the cellulose conversion in water
was observed compared to blank reaction. Moreover, the yield
values of 5-HMF and LA are higher than those obtained from
blank reaction, showing that water, in the presence of acid
catalyst, is a favorable medium for promoting rehydration of 5-
49
HMF to LA. Very recently, Li and his team demonstrated that
the decomposition of 5-HMF to LA was the preferred reaction
pathway in water and an unexpectedly high LA yield could be
obtained in the presence of Amberlyst™ 70 as the solid acid
catalyst. Under our experimental conditions, fructose was
present at trace level in the reaction product mixture either due
to limited extent of glucose isomerization to fructose or once
Fig. 5 Effect of AAB solid acid catalyst load amounts on the yields of
reaction products obtained from hydrothermal conversion of DSB
biomass at 453.2 K.
50
formed, fructose was rapidly dehydrated to 5-HMF. The
temperature variation experiments for UB-catalyzed hydro-
thermal treatment showed an upward trend of cellulose
depolymerization rate with temperature. The actual yields of
glucose, 5-HMF and LA increased when the reaction tempera-
ture was raised from 393.2 to 453.2 K and a decrease in yield was
noted at 473.2 K. Such results imply that an increase in
temperature not only improves the extent of cellulose depoly-
merization, but also enhances the rate of side reactions.
The actual yields of pentoses increased to the range of 47.67–
of high purity sugars or microcrystalline cellulose (MCC,
Avicel®) as model compounds to produce 5-HMF or furfural as
end products. For this case, the product selectivity term is
dened as the mol% yield of targeted product divided by mol%
of reactant consumed or mathematically described by: selec-
tivity, S (%) ¼ product yield (mol%)/conv. (mol%). Using simple
carbon sources such as glucose, xylose or other mono/
disaccharides, it is reasonable to achieve moderate to very
high selectivity toward targeted product due to simpler and
better control in reaction within one reactor. Following this
statement, the determination of product selectivity term by the
above mathematical expression for direct conversion pathways
such as from glucose or fructose into 5-HMF, from MCC into LA
or from xylose or hemicellulose into furfural is deemed to be
valid. On the other hand, there still exists conicting denitions
for the product yield and selectivity terms in the study utilizing
real lignocellulosic biomass. In such typical cases, the conver-
sion of lignocellulosic biomass into platform chemicals
involves a complex multi-step route. Another potential issue to
consider is the determination of initial amount (moles) of
biomass reactant fed into the reactor may not be possible unless
a thorough characterization study (e.g., gel permeation chro-
ꢁ1
72.50 mg g dry biomass, indicating that UB catalyst was active
for hemicellulose conversion. Over the temperature range
studied, ca. 0.55–1.44 wt% furfural yields were achieved with UB
catalyst having a Si/Al molar ratio of 5.3. A similar increasing
trend in pentoses and furfural yields from hemicellulose
conversion over synthetic zeolites (HBeta, HMOR and HUSY)
and K10 montmorillonite catalysts has been demonstrated as
51
well. The increase in furfural yield implies that the Bronsted
acid sites located at the external surface of the Si-tetrahedral
+
layer (one may refer to the terminal silanol groups, Si–OH )
2
are particularly active in catalyzing the dehydration of xylose to
furfural. The effects of catalyst loading on the yields of sugars, 5-
HMF, furfural and LA were investigated by varying the catalyst
amounts from 0 to 2% (w/v) for 60 min reaction at 473.2 K.
Experimental results depicted in Fig. 5 show that higher yields
of LA and furfural could be obtained by increasing the catalyst
loading from 1 to 2%, due to a higher number of catalytic acid
sites are available for reaction. Further increase in the catalyst
loading may or may not adversely affect the formation of LA and
furfural and optimization of the catalyst loading is beyond the
scope of this study.
Table 4 summarizes comparison of hexoses (e.g., glucose and
fructose), xylose and other carbonaceous feedstocks (cellulose
and hemicellulose) conversion into 5-HMF, LA and furfural over
solid acid catalysts having only Bronsted acid sites or rationally
designed by varying the Bronsted-to-Lewis ([B]/[L]) acid site
ratios such as H-mordenite, ion-exchange polymeric resins
13
matography or cross-polarization magic angle spinning C-
NMR spectroscopy) is performed to interpret the molecular
mass of biomass. Thus, the product selectivity term for hydro-
thermal conversion step is not evaluated in this paper due to a
set of main and side reactions are involved during the indirect
conversion of DSB biomass to LA and several reaction products
remained unidentied by HPLC analysis, which may over-
estimate true selectivity values. Another reason was due to the
instrumentation issues for conducting a thorough character-
ization of biomass feedstock in this study.
On the contrary, the cellulose conversion to glucose and
rehydration of 5-HMF to LA was quite signicant with AAB as a
solid acid catalyst (Si/Al molar ratio ¼ 6.8). AAB catalyst also
possesses dual Bronsted and Lewis acidity similar to the
untreated bentonite; however the number of Bronsted acid sites
(
Amberlyst™ 15 and Amberlyst™ 70), Sn-exchanged montmo-
rillonite, acid-functionalized mesoporous SBA-15, K10 mont-
morillonite and metal oxides catalysts (TiO and ZrO ). In this
2 4
appears to be higher than the Lewis acid sites aer H SO
treatment. In this regard, the presence of Bronsted acid sites
2
2
table, it can be seen that most studies used commercial sources
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41293

View Article Online
RSC Advances
Paper
Table 4 Conversion and selectivity in the aqueous-phase dehydration reaction of hexoses, xylose and other carbonaceous sources into 5-HMF,
LA and furfural catalyzed by various solid acid catalysts
Selectivity (%)
Conv.
Entry
Catalyst
Reaction pathway
5-HMF
LA
Furfural
(%)
Reference
1
2
3
4
5
6
7
8
9
LZY zeolite
HUSY zeolite
Fructose / LA
Hemicellulose / furfural
Hemicellulose / furfural
6.5
—
—
91.0
—
6/72
8.9
16.2/54.4
0
84.4
—
—
—
—
—
—
—
—
—
—
—
—
—
53.9
6.4
—
96.0
36.1
24.6
76.0
62.3 a
45/79
78.7
40.2/98.4
48.0
8.0
85.0
79.0
85.0
90.0
100
50
51
51
52
53
54
55
55
55
53
56
56
57
57
58
59
45.7
67.0
—
22.7
—
—
—
—
99.8
—
K10 montmorillonite
H-mordenite (Si/Al ¼ 11)
HY zeolite
Sn-beta/HCl zeolite
Amberlyst™ 15
Sn-exchanged montmorillonite
Ca-montmorillonite
Amberlyst™ 70
a
Fructose / 5-HMF
Xylose / furfural
Glucose / 5-HMFa
Glucose / 5-HMF
Glucose / 5-HMFa
Glucose / 5-HMF
Xylose / furfural a
Fructose / 5-HMF
a
a
a
1
0
—
60.0
66.0
16.5
15.6
2
a
11
12
13
14
15
Acid-functionalized SBA-15
Anatase TiO
Zirconia
Zirconia
Tungstated alumina (AlW)
Fructose / 5-HMF
—
2
Glucose / 5-HMF
Fructose / 5-HMF
Cellulose / LA
1.2
1.1
8.0
—
Cellulose / LA
—
47
a
2 2
Dehydration reactions were carried out in biphasic systems such as in water/methyl isobutyl ketone (Entry 4), H O/THF with 0.35 g salt per g H O
(Entry 6), THF/DMSO (Entries 7–9), aqueous and methyl isobutyl ketone/2-butanol phases (Entry 10: fructose / 5-HMF, Entry 11).
located at the external surface of the Si-tetrahedral layer and in also take part in the catalytic reaction. The Lewis acid sites on
the interlayer region can serve as the catalytic acid sites for clays can also be generated due to octet vacancies that occur as a
2
+
3+
hydrolysis of cellulose to glucose and rehydration of 5-HMF to result of structural metal cations (e.g., Mg or Al ) being
LA. The Bronsted acid sites in the acid-activated clay might also carried away during acid activation. However, the extent of
3
+
coordinate with hydroxyl groups bonded with Al cations in the catalysis by the Lewis acid sites toward the rates of cellulose and
Al-octahedral layer based on FTIR spectroscopic study of Tyagi hemicellulose depolymerization remains unknown in our
60
and his group. The selective hydrolysis of cellulose to glucose experiments. The polymerization reactions leading to formation
over various solid acid catalysts include a robust sulfonated of soluble and insoluble polymeric side-products did occur in
activated carbon (AC-SO H) catalyst at 423.2 K has been con- the sub-CW treatment catalyzed by AAB solid, but to a lesser
3
44
ducted by Onda and co-workers. Further, they found that the extent compared to blank and UB-catalyzed sub-CW treatments.
AC-SO H catalyst displayed signicant catalytic activity and a It is anticipated that the catalyst mesoporosity might enhance
3
remarkably high selectivity toward glucose production, which diffusion of furan and LA products away from the active sites
was attributed to high hydrothermal stability, strong acidic where hydronium-catalyzed side and secondary reactions could
SO
3
H groups and the hydrophobic graphene planes of the occur. In consequence, higher LA yield was obtained in the AAB-
61
catalyst. Tong and co-workers conducted catalytic hydrolysis catalyzed sub-CW treatment as shown in Table 3. The yield of LA
ꢁ
1
of cellulose to reducing sugar in water over acid-activated ranged between 30.52 and 159.17 mg g dry biomass in the
montmorillonite catalysts and their study showed that the AAB-catalyzed sub-CW treatment, representing a signicant
higher concentrations of the Bronsted acid sites in the inter- increase in yield over the range of values obtained from UB-
layer region contributed to greater conversion of cellulose. Since catalyzed sub-CW treatment. Therefore, an aqueous solution
AAB possesses a micro-mesoporous structure with slit-shaped containing LA (15.92 wt%) obtained from AAB-catalyzed sub-
pores of 3–4 nm in diameter, it can be argued that the CW treatment at 453.2 K and 2% catalyst loading was further
internal diffusion limitation may not be signicant. The used in the catalytic hydrogenation process to produce GVL.
molecular dimensions of monomeric sugars and their degra-
Saturated lactone (GVL) can be synthesized from an aqueous
dation products are found to be similar to pore sizes of micro- solution containing LA by catalytic hydrogenation reaction. For
62
porous materials (0.4–0.9 nm), thus the 3 to 4 nm-sized pores this reaction, it is well recognized that supported catalysts
are large enough for monomeric sugars, furans (5-HMF and featuring noble metals such as platinum and ruthenium in
furfural) and simple organic acids (LA) to diffuse in and out of nano-sizes and highly dispersed form are the most active and
the catalyst particle. Once monomeric sugars or intermediate product selective toward GVL. In the catalytic hydrogenation of
oligomers (e.g., dimers, trimers or tetramers) are formed within LA to GVL, the reaction pathways can be distinguished into two
pores, these molecules can enter narrow channels in the cata- main routes, depending on the order in which dehydration and
63
lyst particles and interact with internal acid sites to give furans hydrogenation occur: (i) acid-catalyzed dehydration of LA into
and LA. The Lewis acidity associated with the clay edge sites, angelica lactones (ALs, a or b isomers) as intermediate followed
although less prominent than the Bronsted acid sites, might by hydrogenation of ALs to GVL (i.e., dehydration–
41294 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
hydrogenation route) or (ii) direct synthesis through 4-hydroxy- high GVL selectivity (81.8–97.6 mol%, 343.2 K) for the hydro-
pentanoic acid (4-HPA) which subsequently undergoes highly genation of LA in water over anatase or anatase/rutile mixed-
favorable acid-catalyzed intramolecular esterication (ring phase TiO -supported platinum catalysts. Under our reaction
2
closure) to ve member ring GVL (i.e., hydrogenation–dehydra- conditions, the resulting byproducts from Pt@TiO -catalyzed
2
tion route). Results for the effects of reaction time and temper- hydrogenation of LA were 2-MTHF and VA, both formed with 0.6
ature on the conversion and GVL selectivity in Pt@TiO
2
-
and 4.2% selectivity at the upper and lower temperature limits
catalyzed hydrogenation of LA are shown in Table 5. In this for 360 min reaction. ALs was untraceable in the reaction
table, it can be seen that the conversion of LA increased with the product mixture, suggesting that this species was rapidly
2 2
extension of reaction time, from 59.3% (120 min) to 88.5% (360 transformed into GVL under H atmosphere and Pt@TiO
min) at 433.2 K. The same promoting effect was noticed for catalyst. Due to the fact that GVL is the dominant product, it
temperature, in which higher LA conversion was achieved by seems that the role of TiO support in the reaction pathway is
2
conducting the reaction at higher temperatures. The best cata- limited to its interaction with platinum metal (adhesion and
lytic activity shown by 1% Pt@TiO
73.2 K for 360 min reaction, corresponding to 99.3% GVL yield sites on its surface. In order to investigate the role of AAB as a
99.3% selectivity at 100% conversion). The selectivity of 1% heterogeneous acid co-catalyst, the catalytic performance
Pt@TiO catalyst toward GVL in this work was considerably between 1% Pt@TiO and 1% Pt@TiO combined with AAB co-
2
catalyst was observed at dispersion properties) and inuencing the hydrogenation active
4
(
2
2
2
64
higher than that reported by Upare and co-workers for 5% catalyst was compared in terms of conversion and GVL selec-
Pt@C catalyst (30% selectivity at 100% conversion). The differ- tivity. When it comes to the selectivity, in the case of 1%
ence in hydrogenation activity between Pt@C and Pt@TiO2 Pt@TiO2 catalyst, the primary product was GVL with >90%
catalysts might be attributed to the nature of the support selectivity. Fig. 6 shows the compositions of reaction products
2 2
material in which TiO support could adsorb and activate the obtained from catalytic hydrogenation of LA with 1% Pt@TiO
carbonyl group of LA and consequently enhance the catalytic in combination with AAB co-catalyst. The results reveal that a
activity several times with respect to conventional Pt supported similar product distribution consisting mainly of GVL (>92%
65
on carbon that does not perform such activation. Another selectivity) was obtained and only very few amounts of byprod-
explanation proposed in the literature for high GVL selectivity ucts such as 2-MTHF and VA were detected. Full conversion of LA
over titania-supported catalysts is the absence of strong acid could be achieved at 413.2 K aer 240 min reaction by utilizing
sites that are responsible for formation of GVL decomposition combined catalyst system. Hence, AAB was demonstrated to be
products, for example, 2-methyl tetrahydrofuran (2-MTHF), an effective acid co-catalyst, which contributed a remarkable
66
valeric acid (VA) or 1,4-pentanediol (PD). An excellent perfor- positive effect on the catalytic performance. A similar nding has
68
2
mance of 1% Pt@TiO catalyst in the hydrogenation of LA was been reported by the group of Galletti et al. for commercial
67
noticed in the study of Lange and his group, yielding >90% ruthenium supported catalysts in combination with different
conversion and >95% GVL selectivity for reaction at 473.2 K. A heterogeneous acid co-catalysts include Amberlyst resins and
66
more recent study conducted by Ruppert et al. also showed niobium phosphate or oxide. Further studies on the rational
design and optimization of the acid co-catalyst dosage and
reaction tests under industrial viable conditions are still being
carried out by our research group (Table 6).
The catalyst stability and its possible reuse are both crucial
issues with respect to economic and green chemistry consid-
Table 5 Effects of reaction time and temperature on the conversion
and product yields in the catalytic hydrogenation of LA with 1%
Pt@TiO catalyst
2
erations. Moreover, catalyst reuse considerably reduces both
production costs and chemical waste, which is another critical
issue of sustainability and environmental impact. The
Yield (%)
Conv.
Time (h) (%)
a
T (K)
GVL AL 2-MTHF VA
Selectivity
combined 1% Pt@TiO and AAB co-catalyst was recovered from
2
the decanted reaction mixture and reused ve times. In detail
procedure, a fresh charge of reactant was introduced into the
reactor containing spent catalyst from the rst reaction test and
the subsequent catalytic run was continued. The recyclability
results of the spent catalyst are shown in Fig. 7. It is clearly
shown that there is a gradual decrease in the catalytic activity
(conversion) and GVL selectivity aer ve consecutive catalytic
runs, reaching a value of 87.4% conversion and 86.3% selec-
tivity. This indicates that deactivation of the catalyst occurs
during the subsequent reaction cycles, presumably due to coke
formation from the polymerization of ALs over acidic AAB co-
catalyst surface. The coke formation renders the catalyst less
active, due to a decrease in the number of active surface sites
available for hydrogenation. We also examined the extent of Pt
metal leaching by performing inductively coupled plasma-mass
3
4
4
4
4
93.2
13.2
33.2
53.2
73.2
2
4
6
2
4
6
2
4
6
2
4
6
2
4
6
35.3
32.0
39.9
52.1
43.5
55.3
69.7
55.9
68.7
85.9
74.3
94.7
98.4
90.8
98.2
99.3
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
1.6
1.4
1.0
1.6
1.5
1.1
1.5
1.4
1.2
1.7
1.2
0.7
1.3
0.7
0.2
1.7 90.7
42.8
54.4
47.1
58.6
72.2
59.3
71.7
88.5
77.8
97.4
100.0
93.6
100.0
100.0
1.5 93.2
1.3 95.8
1.9 92.4
1.8 94.4
1.4 96.5
1.8 94.3
1.6 95.8
1.4 97.1
1.8 95.5
1.4 97.2
0.9 98.4
1.5 97.0
1.1 98.2
0.4 99.3
a
Selectivity was dened as the product yield (mol%) divided by LA
conversion (mol%).
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41295

View Article Online
RSC Advances
Paper
Fig. 6 Time and temperature effects on the hydrogenation of LA to GVL in the presence of 1% Pt@TiO
2
(0.25 g) and AAB clay (0.25 g) as an acid
co-catalyst (LA: levulinic acid; GVL: gamma-valerolactone; 2-MTHF: 2-methyl tetrahydrofuran; VA: valeric acid).
spectrometry (ICP-MS) analysis in the liquid phase aer 240 the catalyst surface as possible. By performing calcination at
min reaction. The loss of the metal phase to the liquid phase this temperature, the phase transformation of TiO support, for
2
was found to be very limited under applied reaction conditions. example from anatase to rutile phase and the sintering or
Hence, lower catalytic activity (LA conversion) and selectivity to collapse of the 2 : 1 layer lattice structure of AAB clay do not
GVL in the consecutive catalytic hydrogenation cycles points at occur and this appears to result in comparable catalytic
deactivation of the spent catalyst by coking. In this regard, the performance of the calcined sample to that of fresh catalyst. The
spent catalyst aer the h catalytic run was subjected to severe deformation of the layer lattice structures of clays and
calcination at 723.2 K under a stream of pure oxygen for 180 clay minerals was observed upon calcination at high tempera-
70
min in order to remove as much carbon residue deposited on tures (i.e., 973.2–1473.2 K).
41296 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
Table 6 Comparison of the performance of various heterogeneous solid catalysts with or without co-catalysts in the hydrogenation of LA to
GVL
Hydrogenation Catalyst
a
Entry
Metal
Particle size (nm)
Support
Acid Co-catalyst
Conv. (%)
Selectivity to GVL (%)
Reference
64
Without acid co-catalyst
1
Ru (5%)
Pd (5%)
Pt (5%)
Ru (0.6%)
Pt (1%)
2–5
17–18
9
2
4.5
Carbon
Carbon
—
—
—
—
—
—
—
—
—
—
—
—
—
—
100
100
100
100
54
27
38
95
>90
100
92
19
100
58.6
98.6
90
30
Carbon
b
2
3
A-TiO
A-TiO
2
93
65
66
2
92.6
85.2
81.6
87.4
>95
90.8
99
32
100
94.4
b
2
3.2
R-TiO
A-TiO
R-TiO
Ru (1%)
—
2
2.2
2
4
5
6
Pt (1%)
Cu–Cr
Ru (5%)
Raney Ni
Cu
NA
NA
NA
NA
10–14
ꢀ1
A-TiO
2
67
8
69
—
Carbon
—
7
8
Zirconia
17
Pt (1%)
A-TiO
2
This study
With acid co-catalyst
9
Ru (5%)
NA
Al
2
O
3
Niobium oxide
Niobium phosphate
Amberlyst™ 70
32
36
57
100
97.5
97.8
98.6
99.2
68
1
0
Pt (1%)
ꢀ1
A-TiO
2
Acid-activated bentonite
This study
a
The metal particle sizes dispersed on the support materials were determined by using TEM/HRTEM, H
2
chemisorption or XRD techniques (NA ¼
b
not available). A-TiO
2
: anatase TiO
2
and R-TiO
2
: rutile TiO
2
(Entry 3: the R-TiO
2
support for Pt was a mixed phase of 20% rutile and 80% anatase).
AAB co-catalyst for 240 min reaction at 413.2 K. The favorable
formation of furfuryl alcohol from hydrogenation of furfural in
49
water has been recently reported by Li and his group in the
presence of Pd@C and Amberlyst™ 70 dual catalysts. Details of
the effects of reaction time and temperature on the conversion
and selectivity toward furfuryl alcohol in the hydrogenation of
furfural over 1% Pt@TiO and 1% Pt@TiO in combination with
2
2
AAB co-catalyst are not given because our study focused on the
production of GVL as desirable end product.
Conclusions
We demonstrate herein a green approach for production of GVL
from sugarcane bagasse involving a rst step for hydrothermal
conversion of biomass to LA and a second step for catalytic
hydrogenation of LA to GVL; both steps worked in a green
solvent such as water. The temperature variation experiments in
the hydrothermal conversion step showed an upward trend of
cellulose and hemicellulose conversion with reaction tempera-
ture; however LA yields did not follow the same trend. A very low
amount of LA was detected in the reaction product mixture from
non-catalyzed hydrothermal reaction, which was attributed to
incomplete, slow and non-selective conversion of cellulose in
pure sub-CW medium. It has been demonstrated that the
presence of heterogeneous solid acid catalyst having strong
Bronsted and Lewis acidity such as acid-activated bentonite
increases not only the hydrothermal degradation rates of
cellulose and hemicellulose into monosaccharides, but also
favors the formation of furan ring-opening products such as LA.
Fig. 7 LA conversion and GVL selectivity after 4 h reaction in the
presence of 1% Pt@TiO and AAB co-catalyst (fresh and consecutive
hydrogenation cycles of the spent catalyst).
2
We also examined hydrogenation of furfural to furfuryl
alcohol, an important industrial chemical intermediate and
possible formation of reaction side-products such as tetrahy-
drofurfuryl alcohol (THFA) and 2-methylfuran (2-MF) over 1%
Pt@TiO
The resultant 1% Pt@TiO
2
and 1% Pt@TiO
2
in combination with AAB co-catalyst.
and combined 1% Pt@TiO and AAB
2
2
co-catalyst both display good catalytic performance with actual
yields of furfuryl alcohol of 33.3% (65.7% selectivity at 50.7%
conversion) in the case of 1% Pt@TiO2 and 45.4% (68.4%
selectivity at 66.4% conversion) in the case of 1% Pt@TiO with
2
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41297

View Article Online
RSC Advances
Paper
The TiO
2
-supported platinum metal (Pt@TiO
2
) in combination 17 A. M. Hengne and C. V. Rode, Green Chem., 2012, 14, 1064–
with acid-activated bentonite as a co-catalyst was utilized in the
catalytic hydrogenation of LA to GVL under moderate temper- 18 X. L. Du, Q. Y. Bi, Y. M. Liu, Y. Cao and K. N. Fan,
atures. It was found that the presence of acid-activated ChemSusChem, 2011, 4, 1838–1843.
bentonite co-catalyst could improve the reaction rate while 19 F. M. A. Geilen, B. Engendahl, M. Holscher, J. Klankermayer
maintaining high selectivity toward GVL, thus enabling the and W. Leitner, J. Am. Chem. Soc., 2011, 133, 14349–14358.
adoption of even milder conditions to perform the reaction. 20 W. Li, J. H. Xie, H. Lin and Q. L. Zhou, Green Chem., 2012, 14,
Recyclability study showed that the spent 1% Pt@TiO 2388–2390.
combined with AAB co-catalyst could retain its catalytic 21 J. M. Tukacs, M. Novak, G. Dibo and L. T. Mika, Catal. Sci.
performance up to ve consecutive hydrogenation cycles. Technol., 2014, 4, 2908–2912.
Catalyst poisoning was primarily caused by coke deposition on 22 I. T. Horvath, H. Mehdi, V. Fabos, L. Boda and L. T. Mika,
1072.
2
the active catalyst sites, presumably from the polymerization of
ALs over solid AAB surface.
Green Chem., 2008, 10, 238–242.
23 L. Qi, Y. F. Mui, S. W. Lo, M. Y. Lui, G. R. Akien and
I. T. Horvath, ACS Catal., 2014, 4, 1470–1477.
2
4 A. Stradi, M. Molnar, M. Ovari, G. Dibo, F. U. Richter and
L. T. Mika, Green Chem., 2013, 15, 1857–1862.
Acknowledgements
J. N. Putro would like to express her sincere appreciation to the 25 M. Chalid, H. J. Heeres and A. A. Broekhuis, J. Appl. Polym.
department of chemical engineering, National Taiwan Univer-
Sci., 2012, 123, 3556–3564.
sity of Science and Technology (NTUST) for providing all the 26 L. Hu, G. Zhao, W. Hao, X. Tang, Y. Sun, L. Lin and S. Liu,
research facilities under the undergraduate internship RSC Adv., 2012, 2, 11184–11206.
program. The partial nancial support from the Directorate 27 G. Novodarszki, N. Retfalvi, G. Dibo, P. Mizsey, E. Csefalvay
General of Higher Education (DIKTI) through Undergraduate and L. T. Mika, RSC Adv., 2014, 4, 2081–2088.
Student Research Grant Programs is also gratefully 28 G. R. Akien, L. Qi and I. T. Horvath, Chem. Commun., 2012,
acknowledged.
48, 5850–5852.
2
3
9 F. E. Soetaredjo, A. Ayucitra, S. Ismadji and A. L. Maukar,
Appl. Clay Sci., 2011, 53, 341–346.
0 C. Zhang, H. He and K. Tanaka, Appl. Catal., B, 2006, 65, 37–
43.
References
1
2
3
4
R. E. H. Sims, W. Mabee, J. N. Saddler and M. Taylor,
Bioresour. Technol., 2010, 101, 1570–1580.
G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106,
31 H. Yang, R. Yan, H. Chen, D. H. Lee and C. Zheng, Fuel, 2007,
86, 1781–1788.
32 M. Garcia-Perez, A. Chaala, J. Yang and C. Roy, Fuel, 2001, 80,
1245–1258.
4
044–4098.
S. Liu, L. P. Abrahamson and G. M. Scott, Biomass Bioenergy,
012, 39, 1–4.
2
33 S. Cao and G. M. Aita, Bioresour. Technol., 2013, 131, 357–364.
H. Schwaiger, A. Tuerk, N. Pena, J. Sijm, A. Arrasto and 34 P. J. du Toit, S. P. Olivier and P. L. van Biljon, Biotechnol.
C. Kettner, Biomass Bioenergy, 2012, 38, 102–108.
Bioeng., 1984, 26, 1071–1078.
J. Wang, J. Xi and Y. Wang, Green Chem., 2015, 17, 737–751. 35 H. Yu, W. Du, J. Zhang, F. Ma, X. Zhang and W. Zhong,
S. W. Fitzpatrick, Production of levulinic acid from Bioresour. Technol., 2010, 101, 6728–6734.
carbohydrate-containing materials, US Pat., 5608105, 36 S. Kim and M. T. Holtzapple, Bioresour. Technol., 2005, 96,
Bione Incorporated, March 04 1997. 1994–2006.
S. S. Toor, L. Rosendahl and A. Rudolf, Energy, 2011, 36, 37 D. A. Salvi, G. M. Aita, D. Robert and V. Bazan, Appl. Biochem.
5
6
7
2328–2342.
Biotechnol., 2010, 161, 67–74.
8
9
K. Yan and A. Chen, Energy, 2013, 58, 357–363.
J. M. Tukacs, R. V. Jones, F. Darvas, G. Dibo, G. Lezsak and
L. T. Mika, RSC Adv., 2013, 3, 16283–16287.
38 R. A. Silverstein, Y. Chen, R. R. Sharma-Shivappa,
M. D. Boyette and J. Osborne, Bioresour. Technol., 2007, 98,
3000–3011.
10 K. Shimizu, S. Kanno and K. Kon, Green Chem., 2014, 16, 39 R. Gupta and Y. Y. Lee, Bioresour. Technol., 2010, 101, 8185–
3899–3903.
8191.
1
1
1 A. V. Bridgwater, Catal. Today, 1996, 29, 285–295.
2 M. Selva, M. Gottardo and A. Perosa, ACS Sustainable Chem.
Eng., 2013, 1, 180–189.
40 Z. Yu, H. Jameel, H. M. Chang and S. Park, Bioresour.
Technol., 2011, 102, 9083–9089.
41 Y. S. Cheng, Y. Zheng, C. W. Yu, T. M. Dooley, B. M. Jenkins
and J. S. VanderGheynst, Appl. Biochem. Biotechnol., 2010,
162, 1768–1784.
13 S. G. Wettstein, J. Q. Bond, D. M. Alonso, H. N. Pham,
A. K. Datye and J. A. Dumesic, Appl. Catal., B, 2012, 117–
1
18, 321–329.
4 O. A. Abdelrahman, A. Heyden and J. Q. Bond, ACS Catal.,
014, 4, 1171–1181.
5 K. Yan and A. Chen, Fuel, 2014, 115, 101–108.
42 C. Falco, N. Baccile and M. M. Titirici, Green Chem., 2011, 13,
3273–3281.
43 D. Doulia, C. Leodopoulos, K. Gimouhopoulos and F. Rigas,
J. Colloid Interface Sci., 2009, 340, 131–141.
1
2
1
1
6 W. R. H. Wright and R. Palkovits, ChemSusChem, 2012, 5, 44 A. Onda, T. Ochi and K. Yanagisawa, Green Chem., 2008, 10,
657–1667. 1033–1037.
1
41298 | RSC Adv., 2015, 5, 41285–41299
This journal is © The Royal Society of Chemistry 2015

View Article Online
Paper
RSC Advances
45 M. Moonsiri, P. Rangsunvigit, S. Chavadej and E. Gulari, 60 B. Tyagi, C. D. Chudasama and R. V. Jasra, Spectrochim. Acta,
Chem. Eng. J., 2004, 97, 241–248.
Part A, 2006, 64, 273–278.
4
4
6 Y. Yu, X. Lou and H. Wu, Energy Fuels, 2008, 22, 46–60.
7 J. N. Chheda, Y. Roman-Leshkov and J. A. Dumesic, Green
Chem., 2007, 9, 342–350.
61 D. S. Tong, X. Xia, X. P. Luo, L. M. Wu, C. X. Lin, W. H. Yu,
C. H. Zhou and Z. K. Zhong, Appl. Clay Sci., 2013, 74, 147–
153.
4
4
8 B. F. M. Kuster, Starch-Starke, 1990, 42, 314–321.
9 X. Hu, R. J. M. Westerhof, L. Wu, D. Dong and C. Z. Li, Green
Chem., 2015, 17, 219–224.
62 J. S. Kruger, V. Nikolakis and D. G. Vlachos, Curr. Opin.
Chem. Eng., 2012, 1, 312–320.
63 J. C. Serrano-Ruiz, R. M. West and J. A. Dumesic, Annu. Rev.
Chem. Biomol. Eng., 2010, 1, 79–100.
50 J. Jow, G. L. Rorrer, M. C. Hawley and D. T. A. Lamport,
Biomass, 1987, 14, 185–194.
1 P. L. Dhepe and R. Sahu, Green Chem., 2010, 12, 2153–2156.
2 C. Moreau, R. Durand, S. Razigade, J. Duhamet, P. Faugeras,
64 P. P. Upare, J. M. Lee, D. W. Hwang, S. B. Halligudi,
Y. K. Hwang and J. S. Chang, J. Ind. Eng. Chem., 2011, 17,
287–292.
5
5
P. Rivalier, P. Ros and G. Avignon, Appl. Catal., A, 1996, 145, 65 A. Primo, P. Concepcion and A. Corma, Chem. Commun.,
11–224. 2011, 47, 3613–3615.
3 R. Weingarten, G. A. Tompsett, W. C. Conner Jr and 66 A. M. Ruppert, J. Grams, M. Jedrzejczyk, J. Matras-Michalska,
2
5
5
5
5
5
5
5
G. W. Huber, J. Catal., 2011, 279, 174–182.
4 E. Nikolla, Y. Roman-Leshkov, M. Moliner and M. E. Davis,
ACS Catal., 2011, 1, 408–410.
5 J. Wang, J. Ren, X. Liu, J. Xi, Q. Xia, Y. Zu, G. Lu and Y. Wang,
Green Chem., 2012, 14, 2506–2512.
N. Keller, K. Ostojska and P. Sautet, ChemSusChem, 2015,
DOI: 10.1002/cssc.201403332.
67 J. P. Lange, R. Price, P. M. Ayoub, J. Louis, L. Petrus, L. Clarke
and H. Gosselink, Angew. Chem., Int. Ed., 2010, 49, 4479–
4483.
6 A. J. Crisci, M. H. Tucker, M. Y. Lee, S. G. Jang, J. A. Dumesic 68 A. M. R. Galletti, C. Antonetti, V. De Luise and M. Martinelli,
and S. L. Scott, ACS Catal., 2011, 1, 719–728. Green Chem., 2012, 14, 688–694.
7 M. Watanabe, Y. Aizawa, T. Iida, T. M. Aida, C. Levy, K. Sue 69 Z. P. Yan, L. Lin and S. Liu, Energy Fuels, 2009, 23, 3853–
and H. Inomata, Carbohydr. Res., 2005, 340, 1925–1930. 3858.
8 S. S. Joshi, A. D. Zodge, K. V. Pandare and B. D. Kulkarni, Ind. 70 Y. Sarikaya, M. Onal, B. Baran and T. Alemdaroglu, Clays
Eng. Chem. Res., 2014, 53, 18796–18805.
Clay Miner., 2000, 48, 557–562.
9 F. Chambon, F. Rataboul, C. Pinel, A. Cabiac, E. Guillon and
N. Essayem, Appl. Catal., B, 2011, 105, 171–181.
This journal is © The Royal Society of Chemistry 2015
RSC Adv., 2015, 5, 41285–41299 | 41299