Chemical conversion of biomass-derived hexose sugars to levulinic acid over sulfonic acid-functionalized graphene oxide catalysts

Article abstract of DOI:10.1039/c3gc40353j

Heterogeneous graphene oxide (GO)-based catalysts with sulfonic acid (SO₃H) functional groups (GO-SO₃H) were used for the selective decomposition of the hexose sugars, glucose and fructose into levulinic acid (LA), which has been used as a platform chemical for various value-added derivatives. The GO-SO₃H catalysts gave high yields of around 78% for LA and showed good reuse compatibility with reliable performance. The chemical transformation patterns for hexose sugar decompositions are affected by the temperature, the density of acid sites, and the type of catalyst. The high catalytic performance of GO-SO₃H was shown to result from the higher density of Bronsted acid sites in the GO, compared with the Lewis acid sites in other AC-SO₃H catalysts. The morphology, surface characteristics, and other physiochemical properties were evaluated using several characterization techniques.

Full text of DOI:10.1039/c3gc40353j

Green Chemistry
View Article Online
View Journal | View Issue
PAPER
Chemical conversion of biomass-derived hexose sugars
to levulinic acid over sulfonic acid-functionalized
graphene oxide catalysts†
Cite this: Green Chem., 2013, 15, 2935
Pravin P. Upare,a,b Ji-Woong Yoon,a Mi Yeon Kim,a,b Hyo-Yoon Kang,a,b
Dong Won Hwang,a Young Kyu Hwang,*a,b,c Harold H. Kungc and
Jong-San Chang*a,b,d
Heterogeneous graphene oxide (GO)-based catalysts with sulfonic acid (SO3H) functional groups
(GO–SO3H) were used for the selective decomposition of the hexose sugars, glucose and fructose into
levulinic acid (LA), which has been used as a platform chemical for various value-added derivatives. The
GO–SO3H catalysts gave high yields of around 78% for LA and showed good reuse compatibility with
reliable performance. The chemical transformation patterns for hexose sugar decompositions are
affected by the temperature, the density of acid sites, and the type of catalyst. The high catalytic per-
formance of GO–SO3H was shown to result from the higher density of Brønsted acid sites in the GO,
compared with the Lewis acid sites in other AC–SO3H catalysts. The morphology, surface characteristics,
and other physiochemical properties were evaluated using several characterization techniques.
Received 19th February 2013,
Accepted 7th August 2013
DOI: 10.1039/c3gc40353j
www.rsc.org/greenchem
γ-valerolactone (GVL), 1,4-pentanediol (1,4-PDO), angelica
lactone, diphenolic acid, δ-aminolevulinic acid, and 5-nona-
1 Introduction
The increasingly rapid consumption of non-renewable fossil none (Scheme 1).6,7
resources to satisfy global energy needs1 has threatened their
The chemical conversion of hexose feedstocks to LA and
long-term viability.2 In order to avoid potentially unpleasant formic acid using homogeneous or heterogeneous catalysts in
consequences, the use of renewable biomass or bio-based a biorefinery scheme has been investigated,7,8 but there are
feedstocks is strongly recommended.3 Cellulose derived from complications associated with catalyst separation, corrosion,
biomass is a key source of monosaccharide sugars such as toxicity (in the case of homogeneous catalysts), low product
glucose, fructose, xylose, and starch,4 and it can be converted yields, and selectivity (in the case of heterogeneous catalysts).
chemically to monosaccharides using acids, enzymes, alkalis, Rackemann et al. summarized the literature related to the
or supercritical systems with known hydrolysis techniques.5 chemical conversions of lignocellulose to LA.7 Bozell et al.
These biomass-derived sugars are promising resources for plat- reviewed different pathways for LA production using different
form chemicals, especially carboxylic acids such as levulinic feedstocks and their commercial limitations. They also noted
acid (LA), formic acid (FA), succinic acid, and fumaric acid. LA, that the highest LA yield of more than 60% was obtained from
which possesses ketonic and carboxylic active functional hexose-containing carbohydrate as a starting material using
groups, is one of the top ten most attractive platform chemi- what is termed the Biofine process. This process was actually
cals for the production of value-added chemicals and trans- first demonstrated for commercial-scale production in Italy
portation fuels, such as 2-methyltetrahydrofuran (MTHF), (La Caserta).8
Recently, selective syntheses of LA from hexose, sugars and
cellulose using homogeneous catalysts have been reported,
aCatalysis Center for Molecular Engineering, Korea Research Institute of Chemical
showing relatively high molar yields of 35–81%.7–9 There are
Technology (KRICT), 141 Gajeong-Ro, Yuesong, Daejeon 305-600, Korea.
also reports of eco-friendly routes using solid-acid catalysts,
E-mail: ykhwang@krict.re.kr, jschang@krict.re.kr; Fax: (+)82-42-860-7676
bDepartment of Green Chemistry, University of Science and Technology (UST),
but the yields are often low, even if an excess of catalyst is
used.10 For example, Hara et al. attempted the chemical con-
version of cellulose and the production of biodiesel from oleic
acid using sulfonated graphene-like solid acid catalysts,11
Fu et al. reported the use of biomass char functionalized by
sulfonic acid for the hydrolysis of crystalline cellulose in a
217 Gajeong-Ro, Yuseong, Daejeon 305-350, Korea
cDepartment of Chemical and Biological Engineering, Northwestern University,
Evanston, IL 60208-3120, USA
dDepartment of Chemistry, Sungkyunkwan University, Suwon 440-476, Korea
†Electronic supplementary information (ESI)
available. See DOI:
10.1039/c3gc40353j
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2935–2943 | 2935

View Article Online
Paper
Green Chemistry
Scheme 1 Chemical conversion of hexose sugars to LA and LA-derived chemicals.
microwave reaction system,3 and Onda et al. investigated the
selective synthesis of glucose from cellulose catalyzed by sul-
fated carbon.12 Instead of carbonized materials, some studies
2 Experimental
2.1 Materials
have used SO3H-functionalized silica, tungstated zirconia, zir- All chemicals (analytical grade) were purchased from Sigma-
conia, titania, and SBA-15-type silica materials for the hydro- Aldrich (Darmstadt, Germany) and used without further purifi-
lysis of cellulose and the decomposition of carbohydrates.12 In cation. Graphene oxide (BET surface area 248 m2 g−1; N-BARO
many cases, Brønsted acids were used for the chemical conver- TECH, South Korea) and activated carbon (BET surface area
sion of cellulose or carbohydrates.9,13 However, as summarized 1150 m2 −1; Duksan Chemicals, South Korea) were used for
g
by Weingarten et al., both Brønsted and Lewis acids may play a our study.
role in solid catalysts for the chemical conversion of carbo-
2.2 Catalyst preparation
hydrates in water.14 Indeed, Davis et al. showed that Lewis acid
catalysts can successfully isomerize glucose in an aqueous The sulfonated graphene (GO–SO3H) catalyst was prepared by
phase.14 The use of water as a solvent for biomass conversion the sulfonation of GO using chlorosulfonic acid in chloroform.
is very attractive from both an economical and an environ- In a typical procedure, 1.0 g of GO was dispersed in 50 ml of
mental point of view,14 because water is a cheap, non-toxic, chloroform for 1 h in an ultrasonic cleaning apparatus. After
and non-flammable solvent. In general, heterogeneous cata- sonication, a black dispersed GO suspension was obtained.
lysis is advantageous over homogeneous catalysis in terms of Then, 0.5 g of chlorosulfonic acid was added carefully to the
separation, catalyst recycling, and environmental impact. To dispersed GO solution in a round-bottomed flask. The solution
the best of our knowledge, however, there has to date been no was then immediately stirred and refluxed in a round-bot-
report positing a higher yield of LA from carbohydrates using tomed flask equipped with a cold-water condenser for 4 h at
heterogeneous catalysts with reusability of the catalyst.
70 °C. It should be noted that the catalyst preparation steps
Graphene, which consists of hexagonal networks of sp2 were done in a vacuum hood and that some precautions were
carbon, has attracted much interest owing to its outstanding taken during the chlorosulfonic acid additions for its proper
electronic and mechanical properties.15 Its chemistry has been mixing into the GO suspension. After 4 h of sulfonation, the
summarized in several reviews, along with that of related resulting suspension was cooled, and then filtered out and
materials.16,17 Graphene oxide (GO) is graphene that contains washed with excess ethanol to remove organic moieties on the
a high density of hydrophilic functional groups including surface and water until the pH of the filtrate became neutral.
hydroxyl, carboxyl, and epoxy groups. These hydrophilic The sample was then dried at 110 °C for 12 h in an oven, and
groups can be used for sulfonation to incorporate Brønsted this material was used for the chemical conversion of
acid sites, and sulfonated GO is a promising candidate as a biomass-derived carbohydrates. The same procedure was used
solid-acid catalyst for the hydrolysis of cellulose or carbo- to prepare AC–SO3H using activated carbon.
hydrate to make industrially important chemicals, or for other
useful chemical reactions.11,18,19
To confirm the effect of the surface hydroxyl groups in GO–
SO3H on the catalytic performance, these groups were pro-
In this work, we report the chemical conversion of carbo- tected by the grafting of methoxytrimethylsilane (MTS) onto
hydrates to LA over a new solid-acid catalyst, GO–SO3H, pre- the surface. In a typical procedure, 1.0 g of GO–SO3H was
pared by the facile sulfonation of GO. A high LA yield (up to initially dispersed in 50 ml of toluene in a 100 ml round-bot-
78%) was obtained using catalytic amounts of GO–SO3H catalyst tomed flask and the resulting mixture was stirred to make the
(1.7% of initial feed) from glucose in an aqueous medium at a suspension. Then, MTS of 1.65 mmol (three times the amount
relatively high concentration. Such a high catalytic performance of the residual OH group density present in GO–SO3H) was
during glucose decomposition was explained in terms of the added to the suspension and used for the removal of surface
synergetic effects between surface hydroxyl groups and Brønsted hydroxyl groups from GO through methanolysis. The resulting
acid sites in the GO–SO3H. For comparison, the activity of a sul- mixture was treated by refluxing at around 70 °C for 12 h and
fonated activated carbon (AC–SO3H) is also reported.
then cooled to room temperature. It was further filtered in a
2936 | Green Chem., 2013, 15, 2935–2943
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
glass funnel with Whatman filter paper, after which the fil- following formula:
tered solid was washed three times with anhydrous ethanol,
X toluene toluene adsorption capacity ðg=gÞ
HI ¼
¼
and finally dried overnight at 100 °C. Through these steps,
MTS-grafted GO–SO3H (MTS/GO–SO3H) was obtained. Its for-
mation was confirmed by FT-IR (not shown).
X water
water adsorption capacity ðg=gÞ
From the breakthrough measurement, the HI of GO–SO3H
was estimated to be 2.54, which is higher than that of the AC–
2.3 Catalyst characterization
The structural characteristics and morphologies of the cata- SO3H catalyst (1.64).
lysts were determined by X-ray diffraction (Rigaku), TEM
2.5 Catalytic measurements
(Tecnai G2 Retrofit), SEM, (XL30S, Philips, Eindhoven, The
Netherlands), and Raman spectroscopy (SENTERRA) with an
excitation wavelength of 534 nm. The textural properties of the
catalysts were determined using BET N2 sorption (Tristar,
Micromeritics) experiments, and the same instrument
equipped with a TCD detector was used for NH3-TPD. Determi-
nation of different functional groups in the catalyst was per-
formed using FT-IR spectroscopy (Nicolet-6700) with a diffuse
reflectance attachment at temperatures ranging from room
temperature to 300 °C. The surface acidities of the catalysts
were determined by FT-IR analyses after pyridine adsorption.
Surface analyses of carbon and sulfur on the catalysts were per-
formed by X-ray photoelectron spectroscopy (XPS, Escalab MK
II spectrometer) using an Al X-ray excitation source (Kα =
1487.6 eV). Acid–base back titrations3 using molar solutions of
NaOH, HCl, and NaCl and elemental analyses of C, H, N, and
S (Flash 2000 organic EA, Thermo Scientific) were used for
quantitative determination of the sulfur present in the
catalysts.
The catalytic decomposition of glucose was performed in a
high pressure stainless-steel reactor (200 ml) equipped with
a temperature controller and a magnetic drive stirrer. In a
typical experiment, 30.0 g of glucose and a catalytic amount
(0.5 g) of GO–SO3H in 200 ml of deionized water were loaded
into the reactor. All reactions were carried out at 200 °C for 2 h
at a stirring speed of 1000 rpm. After the reaction, the catalyst
was separated by filtration and then washed with ethanol
(100 ml) at 60 °C, followed by 1,4-dioxane (100 ml) at 70 °C for
2 h to reduce the carbonized char which formed during the
reaction. Next, the washed catalyst was used for the next run or
cycle. The product sample was diluted with distilled water, fil-
tered using a syringe membrane filter (13 mm to 0.45 µm,
Whatman), and then used for qualitative and quantitative ana-
lyses. Identification of the products and their quantitative ana-
lyses were performed by HPLC and LC/mass spectrometry
(MS), using an RI detector and a Sugar-pack™ water column
(6.5 mm × 13 mm, no. WAT085188, Waters, USA). Gaseous pro-
ducts were analyzed by GC equipped with a TCD detector and
a Carbosphere GC column. The same procedure was repeated
for the AC–SO3H, HY, Amberlyst-36 and 0.1 M H2SO4 catalysts.
2.4 Glucose adsorption studies
For the liquid adsorption experiments, 5 wt% of a 100 ml
aqueous glucose solution was prepared. Typically, 0.50 g of
samples (GO, AC–SO3H and GO–SO3H) were added to the
glucose solution while keeping the temperature constant at
25 °C. For the successive glucose adsorption step, the samples
3 Results and discussion
were sonicated for 3 h. After sonication, the samples were with- The GO and AC were obtained from commercial sources, and
drawn, and then separated with a centrifuge at 5000 rpm. The the procedure to sulfonate them is described in the Experi-
supernatant liquids were analyzed using a high pressure liquid mental section. The XRD patterns of GO, AC, and their sulfo-
chromatograph (HPLC) equipped with a refractive index (RI) nated counterparts GO–SO3H and AC–SO3H are shown in
detector and a sugar pack column. Distilled water solvent was Fig. S3.† They all showed a broad peak at around 26° 2θ,
used as a mobile phase. The column temperature was main- suggesting the presence of a poorly crystalline graphite struc-
tained at 70 °C. The HPLC chromatograms are shown in ture in all of these samples. For the AC and AC–SO3H samples,
Fig. S1.† Among all samples, GO–SO3H showed a higher there were additional sharp peaks that indicated the presence
adsorption amount of glucose (6.5 wt%) compared to the of well-crystalline 3D domains of graphite.
others (3.4 wt% at GO, 2.7 wt% at AC, and 4.3 wt% at AC–
The Raman spectra (Fig. 1) showed two peaks in the region
SO3H). To characterize the surface hydrophobicity of the cata- between 1000 cm−1 and 1700 cm−1: a G-band at around
lysts, breakthrough measurements during the competitive 1580 cm−1 due to the Raman-active vibration of E2g symmetry
sorption of water and toluene were carried out in a fixed-bed in graphite, and a D-band at 1350 cm−1 due to the disordered
reactor at 35 °C by passing these vapors (5600 Pa for water and boundaries of the graphene layers,20,21 indicating the presence
6200 Pa for toluene) through a catalyst bed (0.1 g). Before the of aromatic carbon sheets in all four samples.22 For AC–SO3H,
test, the catalysts were pretreated at 150 °C for 3 h under a flow the G-band was shifted to a higher frequency of 1590 cm−1
of He gas (15 ml min−1). Qualitative and quantitative adsorp- from 1580 cm−1, indicating the presence of isolated nanocrys-
tion trials were monitored using a mass spectrometer con- talline domains in amorphous carbon.23 The broad intense
nected to a fixed-bed reactor. The resulting breakthrough D-band in AC–SO3H also indicated the presence of amorphous
curves as a function of the elution time are shown in Fig. S2.† material in the sample, consistent with the presence of the
The hydrophobicity index (HI) was calculated using the broad peak in the XRD pattern. It is interesting that the
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2935–2943 | 2937

View Article Online
Paper
Green Chemistry
(Fig. 2b). Similar to the SEM images, the TEM image of GO
(Fig. 2c) showed that the sheets were relatively well aligned
compared with those of GO–SO3H, which shows randomly
oriented and crumpled sheets (Fig. 2d).
The diffuse reflectance (DR) FT-IR (Fig. 3 and Fig. S4†). and
X-ray photoelectron spectroscopic (XPS) (Fig. 4) results con-
firmed the presence of SO3H groups. In the IR spectrum of
GO–SO3H, the presence of vibrational bands at 1040, 1080,
and 1171 cm−1 (arising from OvSvO stretching) and at 1125,
1277, and 1320 cm−1 (from the SO3H stretching) in the FT-IR
spectrum was a clear indication of the presence of SO3H in
GO–SO3H. These bands were also observed on SO3H-functiona-
lized materials.11,18 The above-mentioned bands were also
observed for AC–SO3H although they were less intense than
those of GO–SO3H. We also performed a temperature-depen-
dent FT-IR study (Fig. 3, Fig. S4†) for GO–SO3H and AC–SO3H,
to check whether the SO3H groups are bonded to the carbon
in GO or simply adsorbed on the surface. The above-men-
tioned bands for OvSvO and SO3H stretching vibrations were
observed across the full temperature range from 50 to 300 °C
in the spectra of both the GO–SO3H and AC–SO3H samples. The
temperature-dependent DR FT-IR spectra showed that the
SO3H group is robustly bonded on the GO surface, without a
loss of intensity, even after the sample was treated at 300 °C
for 12 h (Fig. S5†).
Fig. 1 Raman spectra of carbon-based catalysts: (a) GO, (b) GO–SO3H, (c) AC,
and (d) AC–SO3H.
Raman peaks of GO were very broad, becoming narrower after
sulfonation as well. Thus, the sulfonation procedure, most
likely the drying steps, resulted in the re-ordering of the struc-
ture. This is consistent with the XRD results, which also
showed a sharper peak for GO–SO3H than for GO. Moreover,
the 2D-peak from the overtone of the D-band peak at around
2670 cm−1 in GO–SO3H can be attributed to a double reso-
nance transition in multilayered graphene, which is only
Raman active in the presence of defects, i.e. surface
functionalization.24 Ji et al.25 reported that the intensity ratio
of graphene, ID/IG, can be used as an indicator of structural
changes. In this work, the ID/IG ratio of the sulfonated GO is
higher than that of GO as a result of the chemical modification
of the GO surface with SO3H groups.
Fig. 2 depicts SEM and TEM images of GO and GO–SO3H.
The SEM images clearly indicate the multilayer nature of the
exfoliated GO (Fig. 2a). The sulfonation of GO led to further
exfoliation of the clusters of sheets into a random pile
Fig. 3 (a) Temperature-dependent DR FT-IR spectra of GO–SO3H and room
temperature DR FT-IR spectra of (b) GO–SO3H and (c) GO. From the second to
the top, the temperature increases from RT to 300 °C in increments of 100 °C.
Fig. 4 X-ray photoelectron spectroscopy (XPS) of the S2p binding energy of (a)
graphene oxide (GO), (b) sulfonated graphene oxide (GO–SO3H), and (c) carbon
(AC–SO3H).
Fig. 2 SEM (a and b) and TEM (c and d) images of the catalysts: (a) and (c) GO,
and (b) and (d) GO–SO3H.
2938 | Green Chem., 2013, 15, 2935–2943
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
Table 1 Catalytic activities on the decomposition of glucose over acid catalysts and their physicochemical properties
Molar yields (%)
Catalyst
BET surface area (m2 g−1
)
Total aciditya (mmol g−1
)
SO3H densityb (mmol g−1
)
Conv. (%) LAc
FAd
FTe TON f
GO
AC
GO–SO3H
AC–SO3H
Am-36g
0.1 M H2SO4
HY
248
1136
217
753
33
1.9
2.14
2.7
3.8
4.5
20.4
—
2.7
—
19
21
89
47
7
98
46
31
91
13
2.3
78
4.2
0
63
14
11
74
1.2
1.5 15
8
0
—
1.2
2.5
4.5
—
—
1.16
1.2
0
124
31
5
1.4 35
0
26
19
5
5
—
2
—
8
622
213
217
—
43
126
MTS/GO–SO3H
13
GO–SO3Hh
2.7
16
Reaction conditions: catalyst 0.5 g, glucose 30.0 g in 200 ml distilled water, 200 °C (the temperatures for Amberlyst-36 and for the H2SO4 catalyst
were 130 °C and 180 °C, respectively), reaction time of 2 h. a Total acidity by titration. b SO3H density measured by EA. c LA: levulinic acid.
d FA: formic acid. e FT: fructose. f TON: turnover number based on the S content. g Am-36: ion-exchange resin Amberlyst-36. h GO–SO3H: results
for fructose decomposition. MTS: methoxytrimethylsilane.
The surface functionalization of GO with SO3H was also COOH groups in GO.29 COOH, being electron-withdrawing,
confirmed through an XP spectroscopic analysis of the S2p reduces the electron densities of the neighboring carbons and
and C1s levels. The XP spectrum of the S2p orbital is shown in leads to strong bonding between that carbon and SO3H, in
Fig. 4. Blue lines indicate the deconvoluted S2p binding agreement with the other reported results.11,18,19,25–31
energy of GO–SO3H and AC–SO3H. The binding energy of the
The Brunauer–Emmett–Teller (BET) surface areas were
sulfur compounds is well established, e.g., 166 eV for the sulf- determined by N2 adsorption to be 248 m2 g−1 and 217 m2 g−1
oxides, 168 eV for sulfones and 169 eV for both sulfonic acid for fresh GO and GO–SO3H, respectively (Table 1). Functionali-
and sulfates.27,28 The binding energy of 168.7 eV is close to the zation with SO3H only changed the surface area slightly. The
anticipated values for sulfonic acid and sulfates. The S2p gravimetric density of the SO3H groups in GO–SO3H (pore size
signals of GO–SO3H and AC–SO3H around 169 eV (Fig. 4) were = 15 nm) and AC–SO3H (pore size = 4.1 nm) was calculated
doublets due to the spin–orbit coupling of S2p orbitals.27 The through C, H, O, and S elemental analyses (Table S2†) and
splitting of the XP spectrum of the deconvoluted S2p binding acid–base titrations (Table S3†). The results are listed in
energy was 1.1 eV in agreement with published data.28 The Table 1. AC–SO3H had 2.5 mmol g−1 of SO3H groups, which
C1s XP spectra of carbon-based materials are commonly used was nearly double that of GO–SO3H (1.2 mmol g−1). Consider-
to evaluate the formation of the chemical bonds of carbon ing its much larger surface area, AC–SO3H had a lower areal
with the different functional groups in the literature.16,28 As density of its sulfonate groups. The total acidities of the
illustrated in Fig. S6,† the XP spectra of C1s for GO and GO– samples were calculated through back titration with NaOH,
SO3H showed peaks at 284.5, 285.5, 286.6, 287.8 and 289.0 eV, using a previously reported technique.3 Interestingly, the total
which are assigned to sp2 (CvC) or sp3 (C–C) carbon, hydroxyl acidity of the GO–SO3H catalyst was found to be 2.7 mmol g−1
.
(C–OH), epoxy (C–O–C), carbonyl (CvO), and carboxylate The difference between this value and the SO3H density is
(CO2H) carbons, respectively.26 The intensity of the peak associated with other oxygenate groups, such as OH and
assigned to carbons of the C–C/CvC at 284.5 eV for GO COOH.31
increased after sulfonation. These results correspond to the
removal of oxygenated functional groups by sulfonation determined using a stainless-steel, high-pressure batch reactor
(Fig. S6 and Table S1†). at 200 °C, and then compared to that of other catalysts. The
The glucose decomposition activity of the GO–SO3H was
The rather large uncertainties in the distribution of oxy- results are shown in Table 1. Under similar reaction con-
genated functional groups preclude any attempt to identify ditions, GO exhibited poor activity, achieving 13% LA yield
which functional group is sulfonated. In contrast, the relative (molar yield) at a glucose conversion rate of 19%. GO–SO3H
peak intensity of the C–C and CvC bonds did not change was much more active, showing a distinctively higher LA yield
much in AC–SO3H, suggesting a lower areal density of the sul- of 78% after 2 h of reaction, 8% formic acid, and traces of
fonated groups for this sample compared to that of GO–SO3H HMF (1–2%) at a glucose conversion rate of 89%. There was
due to the high surface area of AC–SO3H compared to GO– also decomposition of FA into CO2, as confirmed by an ana-
SO3H. This was confirmed by titration.
lysis of gaseous products after the reaction using a GC
From the IR and XPS data, it was concluded that the SO3H equipped with a TCD detector. Very recently, Dumesic et al.
species form strong bonds with the carbon in GO–SO3H and reported the direct conversion of the initial feedstock cellulose
AC–SO3H. Such a strong bonding should help minimize leach- selectively to LA over solid-acid catalysts,32 obtaining
a
ing in an aqueous reaction medium. It has been suggested maximum LA yield of 69% with the Amberlyst-70 catalyst. They
that the strong bonding is mainly due to the presence of also mentioned the decomposition of FA during the reaction,
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2935–2943 | 2939

View Article Online
Paper
Green Chemistry
which was given as the reason for the lower FA yield. GO–SO3H
The acid strengths of the catalysts were also evaluated
was also more active than AC–SO3H, Amberlyst-36 resin, and through NH3-temperature-programmed desorption (TPD);
HY zeolite, although it had a lower SO3H content (1.2 mmol these profiles are shown in Fig. S7.† From the NH3-TPD analy-
g−1) than those of AC–SO3H and the Amberlyst-36 resin cata- sis, the GO–SO3H catalyst had a total of 2.17 mmol g−1 of acid
lyst. The decomposition activity of the homogeneous Brønsted sites, of which 1.89 mmol g−1 was attributed to medium acid
acid catalyst 0.1 M H2SO4 was also checked, as presented in sites (270–400 °C) and 0.28 mmol g−1 to strong acid sites
Table 1; greater amounts of char were obtained as compared (above 400 °C). However, the catalyst did not show any de-
to GO–SO3H, even at 180 °C. In the Biofine process, the sorption peaks in the weak acidic region (below 270 °C) in the
reported LA yield was around 70–80% of the theoretical LA TPD profile. The AC–SO3H samples had 2.72 mmol g−1 of
yield value (71.6%) from cellulose-containing hexose by weak acid sites and 3.70 mmol g−1 of strong acid sites for a
1.5–3% of sulfuric acid at 210 °C–220 °C, with FA and tars as total of 6.42 mmol g−1. However, they did not have many
the remaining carbon balance in the mass.8 Our GO–SO3H medium acid sites. It is often found that Lewis acid sites are
catalyst also produced an appreciable LA yield of around 78% strong acid sites and Brønsted acid sites are medium ones.35
of the theoretical value from C6 sugars using similar reaction Thus, these observations are consistent with the pyridine
conditions (200 °C). adsorption data, showing that there is a much higher density
The surface acidities of the catalysts were evaluated by of Brønsted acid sites in GO–SO3H.
monitoring the degree of pyridine adsorption with FTIR. After
The difference in activity between GO–SO3H and AC–SO3H
the removal of physically adsorbed pyridine by evacuation, can be attributed to the difference between their structures
absorption bands at 1633, 1602, 1531, 1490, and 1459 cm−1 and types of acid sites. Although AC–SO3H has a higher total
were observed on GO–SO3H (Fig. 5a). The bands at 1633 cm−1 gravimetric acid density, most of the acid sites are Lewis acid
and 1531 cm−1 were characteristic peaks of Brønsted acid sites sites. On the other hand, GO–SO3H has a higher density of
resulting from the formation of pyridinium ions with proton Brønsted acid sites. It also has much larger pores. Conse-
sites.33 Another very broad band at around 1460 cm−1 was quently, it is much easier for the sizable glucose and levulinic
attributed to the Lewis acid site. These originated in the SvO acid to diffuse in and out of them.36 The difference in product
double bonds of the sulfate complex,33 due to the electron selectivity is due to their different densities of Lewis and
inductive effect of the SvO double bonds in the surface Brønsted acid sites, as it is known that Lewis acid sites are
sulfate complex, as shown in sulfated zirconia.33,34 Tanabe important for the isomerization of glucose to fructose,
et al. demonstrated the generation of this strong Lewis acidity whereas Brønsted acid sites are important for the formation of
by sulfonation or due to the dehydration of Brønsted acid HMF and LA.30,37,38
sites. Gao et al. also agreed for similar reasons.34 These bands
The catalytic properties of HY can be explained similarly.
were thermally stable at 300 °C, suggesting that both the acid Its relatively low catalytic activity is due to its small pore size
sites and the adsorbed pyridine remained intact. For AC– (0.65 nm) and its low selectivity for fructose but high selec-
SO3H, the intensities of the IR peaks assigned to Lewis acid tivity for LA are due to the fact that most of its acid sites are
sites at 1457, 1503 and 1610 cm−1 34 were relatively much Brønsted acid sites, as typically found in zeolites.14 In addition
stronger than those for Brønsted acid sites. The bands were to the explanation described above, possible synergistic effects
also rather stable thermally, consistent with the small changes of –SO3H with other functional groups (COOH and OH) may
of the IR peaks for the solid (Fig. 3). In addition to these contribute to the high activity of GO–SO3H. Recently, Hara
bands, combined bands at 1583 and 1490 cm−1 remained after et al. discussed the adsorption-enhanced hydrolysis of β-1,4-
the thermal treatment.
glucan by graphene-based carbon bearing SO3H and hydro-
philic functional groups.19,25 They reported that the syn-
ergistic effects of both Brønsted acid sites (i.e., SO3H, and
hydrophilic OH and COOH groups present on the surface) are
responsible for the high level of activity during the hydrolysis
of glucan to glucose. That is, adsorption takes place through
glucosidic bonds with β-1,4-glucan by hydrophilic –OH
groups, assisting the hydrolysis of glucan to glucose. It is poss-
ible that a similar synergistic effect of the SO3H and hydro-
philic functional groups in the GO–SO3H contributes to the
high glucose isomerization and decomposition activity. In
order to test this synergistic effect, we treated a GO–SO3H
sample with methoxytrimethylsilane (see the Experimental
section) to protect the hydroxyl and carboxylic groups and
render them inactive for hydrogen bonding. The resulting
material was less active for glucose decomposition, showing a
31% conversion rate and a poor LA selectivity of 41% (11%
molar yield).
Fig. 5 FTIR spectra of adsorbed pyridine on (a) GO–SO3H, and (b) AC–SO3H.
From the bottom to top, the sample is shown before pyridine adsorption, after
exposing to pyridine and then treating under vacuum for 3 h, and after thermal
heating at 300 °C for 1 h.
2940 | Green Chem., 2013, 15, 2935–2943
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
The lower activity and lower LA selectivity observed for AC– during the reaction. These results confirm that the de-
SO3H, despite the fact that it also possessed both hydrophilic activation of the catalyst is mainly caused by char formation
functional groups and SO3H, suggest that there is an optimum during the reaction.
ratio of these groups for high activity and selectivity. Acid site
We also tested the GO–SO3H catalyst for the hydrolysis/
titration results (Table S3†) showed a much lower (hydroxyl + decomposition of the initial α-cellulose feedstock (Table S5†)
carboxylic acid)/SO3H ratio in AC–SO3H, implying an insuffi- and of the intermediate fructose, using the same reaction
cient number of these carbon–oxygen groups for the optimal system. Fructose is an intermediate during the formation of
contribution of hydrogen bonding with the reactant to facili- LA from glucose, as suggested in the literature (Scheme S1†).40
tate the reaction at the Brønsted acid sites.
An LA yield of 23% was obtained from cellulose with 36%
One consequence of the synergistic effect is that it glucose as the major product from hydrolysis and the
enhances glucose adsorption. This was tested by taking a sus- decomposition of alfa cellulose. The low yield of LA by the
pension of 5 wt% aqueous glucose solution containing known heterogeneous catalyst GO–SO3H is mainly due to limited
amounts of GO–SO3H, AC–SO3H, or GO with HPLC to deter- cellulose solubility in water and lower sulfur content as com-
mine the concentration of glucose remaining in the solution pared with homogeneous sulfuric acid catalysts.
qualitatively (details are in the Experimental section). The
The time-dependent product distributions over GO–SO3H
results are shown in Fig. S1.† GO–SO3H adsorbed more (Fig. 7) are shown in Scheme S1.† As the reaction time
glucose from the solution than the other samples. The higher increases, the yield of LA is linearly increased from 10 min to
amount of adsorption likely contributes to the higher activity 2 h after a small amount of fructose is initially formed. By con-
of the catalyst. The hydrophobicity index (HI), which is the trast, HMF starts forming after 20 min of the reaction and
ratio of the toluene adsorption capacity to the water adsorption then maintains a constant yield until 90 min. After 2 h, the
capacity, for each catalyst was calculated using breakthrough HMF disappears in the product mixture. These results indicate
curve experiments (Fig. S2†). The HI of the GO–SO3H catalyst that HMF is the intermediate leading to the production of LA
(2.54) was higher than that of AC–SO3H (1.64).
and FA, as depicted in Scheme S1.† Interestingly, the yield of
Fig. 6 shows the regeneration of the GO–SO3H catalyst up to FA increases linearly up to 90 min and then decreases,
the fifth cycle. In the regeneration study, a small amount of because the decomposition of FA into gaseous CO2 is facili-
char was obtained with the catalysts formed from water-insolu- tated after 90 min. The highest glucose conversion rate of
ble humins,39 after each recycling reaction. We separated the 96.8% and levulinic acid yield of 84% were obtained after
catalyst used and the char from the reaction mixture through 150 min, but the amount of char was also found to be higher
simple filtration, but it is difficult to separate catalysts and in this case (1.06 g). Therefore, we chose a condition of 2 h for
char in a homogeneous mixture. The SO3H group density catalyst screening. The effect of temperature on glucose
decreased marginally after each cycle, and this slight decrease decomposition is also consistent with this mechanistic
in sulfur density was confirmed by EA. After each cycle, a scheme (Fig. S8†).
sulfur density loss of around 0.08–0.1 mmol g−1 was observed
Fructose is one of the intermediates for the formation of LA
(Table S4†) and the glucose conversion decreased slightly (by via glucose decomposition40 as well as an important starting
3% to 4% after each of the four cycles used). However, an ICP model feedstock for the chemical conversion of biomass, due
analysis of the product mixture indicated no leaching of sulfur to its easy availability from corn starch.41 The GO–SO3H cata-
lyst showed good performance for fructose decomposition into
Fig. 6 Regeneration test of glucose decomposition over the GO–SO3H catalyst.
Reaction conditions: catalyst 0.5 g, glucose 30.0 g in 200 ml distilled water,
temperature 200 °C, and reaction time 2 h.
Fig. 7 The product distribution profile of the glucose decomposition reaction
on GO–SO3H as a function of time. Reaction conditions: catalyst 0.5 g, glucose
30 g, at 200 °C in 200 ml distilled water.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2935–2943 | 2941

View Article Online
Paper
Green Chemistry
other chemicals. As illustrated in Table 1, the catalyst also had
a high FT conversion rate (91%) with a 74% LA yield and a 5%
hydroxy-methylfurfural (HMF) yield. Interestingly, the amount
of char present in the catalyst after the reaction was found to
be very low compared to that from glucose decomposition,
which may have been due to the fast conversion of fructose
stemming from the exclusion of an isomerization step when
using glucose as a starting material.42 These results further
demonstrate that GO–SO3H can be used for the chemical con-
version of other bio-based feedstocks or in other solid acid
hydrolysis/decomposition reactions.
6 P. P. Upare, J.-M. Lee, Y. K. Hwang, D. W. Hwang, J. H. Lee,
S. B. Halligudi, J. S. Hwang and J.-S. Chang, ChemSusChem,
2011, 4, 1749–1752; 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; A. S. Mamman, J. M. Lee,
Y. C. Kim, I. T. Hwang, N. J. Park, Y. K. Hwang, J. S. Hwang
and J. S. Chang, Biofuels, Bioprod. Biorefin., 2008, 2, 438–
454; J. C. Serrano-Ruiz, D. Wang and J. A. Dumesic, Green
Chem., 2010, 12, 574–577.
7 D. W. Rackemann and W. O. S. Doherty, Biofuels, Bioprod.
Biorefin., 2011, 5, 198–214.
8 J. J. Bozell, L. Moensa, D. C. Elliott, Y. Wang,
G. G. Neuenscwander, S. W. Fitzpatrick, R. J. Bilski and
J. L. Jarnefeld, Resour., Conserv. Recycl., 2000, 28, 227–239;
D. J. Hayes, S. W. Fitzpatrick, M. H. B. Hayes and
J. R. H. Ross, The Biofineprocess Production of Levulinic Acid,
Furfural, and Formic Acid from Lignocellulosic Feedstocks,
Biorefineries-Industrial Processes and Products: Status Quo
and Future Directions, ed. B. Kamm, P. R. Gruber,
M. Kamm, Wiley-VCH, Weinheim, 2008, vol. 1, p. 139;
S. Ritter, Italian levulinic acid facility is first to make tar-
geted biomass-based chemical feedstocks, Chem. Eng.
News, 2006, 84, 47.
4 Conclusion
In conclusion, the present study showed that GO–SO3H is a
promising catalyst for the selective synthesis of LA during the
single-step hydrolysis of glucose. The active sites are the
Brønsted acid SO3H sites, but the presence of other functional
groups, such as carboxyl and hydroxyl, is also important in
enhancing the adsorption of glucose (and presumably other
biomass chemicals) for the reaction. The layered morphology
of GO–SO3H is beneficial for the rapid diffusion of the reac-
tants and products, and the sulfonate groups are both ther-
mally stable and do not leach into the reaction mixture. The
catalyst can be recycled with minimal loss of activity, and can
be used as an initial bio-based cellulose feedstock or other bio-
based sugars. Future studies will aim to determine strategies
with which to avoid char formation.
9 L. Deng, J. Li, D. M. Lai, Y. Fu and Q. X. Guo, Angew. Chem.,
Int. Ed., 2009, 48, 6529–6532; C. Chun, M. A. Xiaojian and
C. Peilin, Chin. J. Chem. Eng., 2006, 14, 708–712.
10 B. C. Redmon, US patent, 2738367, 1956; J. Jow,
G. L. Rorrer, M. C. Hawley and D. T. A. Lamp, Biomass,
1987, 14, 185–194; H. Heeres, R. Handana, D. Chunai,
C. B. Rasrendra, B. Girisuta and H. J. Heeres, Green
Chem., 2009, 11, 1247–1255; J. Hegner, K. C. Pereira,
B. DeBoef and B. L. Lucht, Tetrahedron Lett., 2010, 51,
2356–2358; F. Chambon, F. Rataboul, C. Pinel, A. Cabiac,
E. Guillon and N. Essayem, Appl. Catal., B, 2011, 105,
171–181.
Acknowledgements
This work was supported by the Institutional Research
Program (SK-1301) and partially funded by the Korea Research
Council for Industrial Science & Technology (ISTK, SK-1311).
Mr P. P. Upare is grateful to KRICT for financial support under
the Ph.D. research program of UST. We also thank Dr. Jong
Yeol Jeon for his helpful discussion.
11 M. Hara, Top. Catal., 2010, 53, 805–810; M. Hara, Chem-
SusChem, 2009, 2, 129–135; L. Geng, Y. Wang, G. Yu and
Y. Zhu, Catal. Commun., 2011, 13, 26–30.
12 A. Onda, T. Ochi and K. Yanagisawa, Green Chem., 2008, 10,
1033–1037; D. Lai, L. Deng, J. Li, B. Liao, Q. X. Guo and
Y. Fu, ChemSusChem, 2011, 4, 55–58; J. A. Melero, J. Iglesias
and G. Morales, Green Chem., 2009, 11, 1285–1308;
A. cabiac, E. Guillon, F. Chambon, C. Pinel, F. Rataboul
and N. Eassayem, Appl. Catal., A, 2011, 402, 1–10.
13 X. L. Du, L. He, S. Zhao, Y. M. Liu, Y. Cao, H. Y. He and
K. N. Fan, Angew. Chem., Int. Ed., 2011, 123, 1–6; H. Mehdi,
V. Fabos, R. Tuba, A. Bodor, L. T. Mika and I. T. Horvath,
Top. Catal., 2008, 48, 49–54.
14 R. Weingarten, G. A. Tompsett, W. C. Conner and
G. W. Huber, J. Catal., 2011, 279, 174–182; Y. Roman-
Leshkov, M. Moliner, J. A. Labinger and M. E. Davis, Angew.
Chem., Int. Ed., 2010, 49, 8954; M.-O. Simon and C.-J. Li,
Chem. Soc. Rev., 2012, 41, 1415–1427.
References
1 A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
2411–2502; D. Martin-Alonzo, J. Q. Bond and J. A. Dumesic,
Green Chem., 2010, 12, 1493–1513.
2 J. J. Bozell and G. R. Petersen, Green Chem., 2010, 12, 539–
554.
3 Y. Wu, Z. Fu, D. Yin, Q. Xu, F. Liu, C. Lu and L. Mao, Green
Chem., 2010, 12, 696–700.
4 A. Fukuoka and P. L. Dhepe, Angew. Chem., Int. Ed., 2006,
45, 5161–5163; P. L. Dhepe and A. Fukuoka, ChemSusChem,
2008, 1, 969–975.
5 L. T. Fan, M. M. Gharpuray and Y. H. Lee, Cellulose Hydro-
lysis, Springer, Berlin, 1987, p. 5; B. Yang, D. M. Willies and 15 K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang,
C. E. Wyman, Biotechnol. Bioeng., 2006, 94, 1122–1128;
M. L. Rabinovich, Appl. Biochem. Microbiol., 2006, 42, 1–26.
Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov,
Science, 2004, 306, 666–669.
2942 | Green Chem., 2013, 15, 2935–2943
This journal is © The Royal Society of Chemistry 2013

View Article Online
Green Chemistry
Paper
16 D. R. Dreyer, S. Park, C. W. Bielawski and R. S. Ruoff, 28 H. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomden, in
Chem. Soc. Rev., 2010, 39, 228–240; Y. Zhu, S. Murali,
W. i. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv.
Handbook of X-ray Photoelectron Spectroscopy, ed. J. Chas-
tain, 1992.
Mater., 2010, 22, 3906–3924; M. J. Allen, V. C. Tung and 29 M. Kitano, D. Yamaguchi, S. Suganuma, K. Nakajima,
R. B. Kaner, Chem. Rev., 2010, 110, 132–145;
H. Kato, S. Hayashi and M. Hara, Langmuir, 2009, 25, 5068–
K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang,
5075.
Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, 30 K. Shimizu and A. Satsuma, Energy Environ. Sci., 2011, 4,
Science, 2004, 306, 666–669.
3140–3153; M. Hara, T. Yoshida, A. Takagaki, T. Takata,
J. N. Kondo, S. Hayashi and K. Domen, Angew. Chem., 2004,
116, 3015–3018.
17 M. Englert, C. Dotzer, G. Yang, M. Schmid, C. Papp,
J. M. Gottfried, H. P. Steinruck, E. Spiecker, F. Hauke and
A. Hirsch, Nat. Chem., 2011, 3, 279–286.
18 S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi,
H. Kato, S. Hayashi and M. Hara, J. Am. Chem. Soc., 2008,
130, 12787–12793; W. Y. Lou, Q. Guo, W. J. Chen,
31 J. A. Maciá-Agulló, M. Sevilla, M. A. Diez and A. B. Fuertes,
ChemSusChem, 2010, 3, 1352–1354; O. C. Compton, B. Jain,
D. A. Dikin, A. Abouimrane, K. Amine and S. T. Nguyen,
ACS Nano, 2011, 5, 4380–4391.
M. H. Zong, H. Wu and T. J. Smith, ChemSusChem, 2012, 5, 32 D. M. Alonso, J. M. R. Gallo, M. A. Mellmer, S. G. Wettstein
1533–1541; X. Qi, H. Guo, L. Li and R. L. Smith Jr., Chem-
SusChem, 2012, 5, 2215–2220.
19 M. Hara, Energy Environ. Sci., 2010, 3, 601–607.
and J. A. Dumesic, Catal. Sci. Technol., 2013, 3, 927–931.
33 S. Hamoudi and S. Kaliaguine, Microporous Mesoporous
Mater., 2003, 59, 195–204.
20 D. Roy, E. Angeles-Tactay, R. J. C. Brown, S. J. Spencer, 34 K. Tanabe, H. Hattori and T. Yamaguchi, Crit. Rev. Surf.
T. Fry, T. A. Dunton, T. Young and M. J. T. Mi, Chem. Phys.
Lett., 2008, 465, 254–257.
Chem., 1990, 1, 1–25; W. Hua, Y. Yue and Z. Gao, J. Mol.
Catal. A: Gen., 2001, 170, 195–202.
21 L. M. Malarda, M. A. Pimentaa, G. Dresselhaus and 35 N. Katada, T. Tsubaki and M. Niwa, Appl. Catal., A, 2008,
M. S. Dresselhaus, Phys. Rep., 2009, 473, 51–87. 340, 76–86.
22 K. Fukuhara, K. Nakajima, M. Kitano, H. Kato, S. Hayashi 36 C. Moreau, R. Durand, S. Razigade, J. Duhamet,
and M. Hara, ChemSusChem, 2011, 4, 778–784.
23 A. C. Ferrari and J. Robertson, Phys. Rev. B: Condens.
Matter, 2000, 61, 14095–14107.
24 G. Eda and M. Chhowalla, Adv. Mater., 2010, 22, 2392–
2415.
25 J. Ji, G. Zhang, H. Chen, S. Wang, G. Zhang, F. Zhang and
X. Fan, Chem. Sci., 2011, 2, 484–487.
26 F. Liu, J. Sun, L. Zhu, X. Meng, C. Qib and F. Xiao, J. Mater.
Chem., 2012, 22, 5495–5502.
P. Faugeras, P. Rivalier, P. Ros and G. Avignon, Appl. Catal.,
A, 1996, 145, 211–224.
37 W. Zeng, D.-G. Cheng, H. Zhang, F. Chem and X. Zhan,
React. Kinet., Mech. Catal., 2010, 100, 377–384.
38 K. Lourvanij and G. L. Rorrer, J. Chem. Technol. Biotechnol.,
1997, 69, 35–44.
39 B. G, irisuta, L. P. B. M. Janssen and H. J. Heeres, Chem.
Eng. Res. Des., 2006, 84(A5), 339–349.
40 S. Saravanamurugan, O. N. V. Buu and A. s. Riisager, Chem-
SusChem, 2011, 4, 723–726; B. Girisuta, L. P. B. M. Janssen
and H. J. Heeres, Ind. Eng. Chem. Res., 2007, 46, 1696–1708.
27 A. J. Crisci, M. H. Tucker, M.-Y. Lee, S. G. Jang,
J. A. Dumesic and S. L. Scott, ACS Catal., 2011, 1, 719–728;
S. D. Evans, K. E. Goppert-Beraducci, E. Urankar, 41 K. Parker, M. Salas and V. C. Nwosu, Biotechnol. Mol. Biol.
L. J. Gerenser, A. Ulman and S. G. Snyder, Langmuir, 1991, Rev., 2010, 5, 71–78.
7, 2700–2709; N. Samarendra, N. Dutta and D. C. Frost, 42 R. O. Marshall, E. R. Kooi and G. M. Moffett, Science, 1957,
Fuel, 1983, 62, 840–841.
125, 648–649.
This journal is © The Royal Society of Chemistry 2013
Green Chem., 2013, 15, 2935–2943 | 2943