Pt nanocatalysts supported on reduced graphene oxide for selective conversion of cellulose or cellobiose to sorbitol

Article abstract of DOI:10.1002/cssc.201301123

Pt nanocatalysts loaded on reduced graphene oxide (Pt/RGO) were prepared by means of a convenient microwave-assisted reduction approach with ethylene glycol as reductant. The conversion of cellulose or cellobiose into sorbitol was used as an application reaction to investigate their catalytic performance. Various metal nanocatalysts loaded on RGO were compared and RGO-supported Pt exhibited the highest catalytic activity with 91.5 % of sorbitol yield from cellobiose. The catalytic performances of Pt nanocatalysts supported on different carbon materials or on silica support were also compared. The results showed that RGO was the best catalyst support, and the yield of sorbitol was as high as 91.5 % from cellobiose and 58.9 % from cellulose, respectively. The improvement of catalytic activity was attributed to the appropriate Pt particle size and hydrogen spillover effect of Pt/RGO catalyst. Interestingly, the size and dispersion of supported Pt particles could be easily regulated by convenient adjustment of the microwave heating temperature. The catalytic performance was found to initially increase and then decrease with increasing particle size. The optimum Pt particle size was 3.6 nm. These findings may offer useful guidelines for designing novel catalysts with beneficial catalytic performance for biomass conversion. Support group: Pt nanocatalysts loaded on reduced graphene oxide are prepared by a microwave-assisted ethylene glycol reduction method, and present high activity and selectivity for the conversion of cellobiose or cellulose to sorbitol. The high catalytic activity is attributed to the synergistic effects of reduced graphene oxide and the supported Pt nanoparticles.

Full text of DOI:10.1002/cssc.201301123

CHEMSUSCHEM
FULL PAPERS
DOI: 10.1002/cssc.201301123
Pt Nanocatalysts Supported on Reduced Graphene Oxide
for Selective Conversion of Cellulose or Cellobiose to
Sorbitol
Ding Wang,[a, b] Wenqi Niu,[b] Minghui Tan,[b, c] Mingbo Wu,*[c] Xuejun Zheng,[a] Yanpeng Li,[c]
and Noritatsu Tsubaki*[b]
Pt nanocatalysts loaded on reduced graphene oxide (Pt/RGO)
were prepared by means of a convenient microwave-assisted
reduction approach with ethylene glycol as reductant. The
conversion of cellulose or cellobiose into sorbitol was used as
an application reaction to investigate their catalytic per-
formance. Various metal nanocatalysts loaded on RGO were
compared and RGO-supported Pt exhibited the highest catalyt-
ic activity with 91.5% of sorbitol yield from cellobiose. The cat-
alytic performances of Pt nanocatalysts supported on different
carbon materials or on silica support were also compared. The
results showed that RGO was the best catalyst support, and
the yield of sorbitol was as high as 91.5% from cellobiose and
58.9% from cellulose, respectively. The improvement of catalyt-
ic activity was attributed to the appropriate Pt particle size and
hydrogen spillover effect of Pt/RGO catalyst. Interestingly, the
size and dispersion of supported Pt particles could be easily
regulated by convenient adjustment of the microwave heating
temperature. The catalytic performance was found to initially
increase and then decrease with increasing particle size. The
optimum Pt particle size was 3.6 nm. These findings may offer
useful guidelines for designing novel catalysts with beneficial
catalytic performance for biomass conversion.
Introduction
Production of fuels or chemicals from renewable biomass re-
sources instead of fossil resources has attracted great attention
in recent years.[1] Cellulose, a polysaccharide mainly composed
of glucose via b-1-4 glycosidic linkage, exists widely in biomass
resources.[2] The conversion of cellulose into fuels or industrial
chemicals can avoid potential conflicts with the food supply,
a key issue in sustainable energy production.[3] The utilization
of cellulose is usually achieved through two steps: selective hy-
drolysis into glucose and further conversion into fuels and
chemicals.[4] Recently, the direct catalytic conversions of cellu-
lose to sugar alcohol, ethylene glycol (EG), and gluconic acid
have been reported.[5–7] The sugar alcohols, especially sorbitol,
are used not only as sweetener in diet foods, but also as an
important basic chemical for the production of sustainable
fuels and chemicals.[8] Many catalysts, for example, Pt/Al2O3,
Ru/carbon nanotubes, Ru/activated carbon, and NixP, Pt/Na(H)-
ZSM-5 have been used to convert cellulose or cellobiose into
sorbitol directly.[5,9–12] Nevertheless, few solid catalysts can
obtain sorbitol with high yield when microcrystalline cellulose
is used as feedstock except under harsh reaction condi-
tions.[9–11]
In the conversion of cellulose to sorbitol, the catalyst used
can accelerate hydrolysis or/and hydrogenation reactions
(Scheme 1). Carbon material is a good choice as the catalyst
support due to its excellent stability under hydrothermal con-
ditions and large surface area for dispersed active compo-
nents.[13] Carbon materials including activated carbon, mesopo-
rous carbon and carbon nanotubes were reported as supports
in the cellulose conversion reaction.[5,10,13] However, to the best
of our knowledge, graphene oxide, a novel carbon material,
has not been reported as a catalyst or as support material in
cellulose conversion. Compared with other carbon materials,
graphene oxide, a one-atom thick planar sheet of hexagonally
arrayed sp2 carbon atoms, has attracted tremendous attention
in recent years due to its relatively stable physical properties
and its unique two-dimensional planar structure.[14] In addition,
graphene oxide is slightly acidic due to a large amount of sur-
face functional groups.[15,16] Carbon catalyst based on graphene
oxide for facilitating oxidation and hydration reactions was re-
ported.[17] In addition, these abundant functional groups on
the surfaces of graphene oxide can also be utilized as anchor-
ing sites for loaded metal nanoparticles, which can increase
the dispersion of the supported nanocatalysts and tune their
catalytic performance. Furthermore, the graphene sheet has
a spillover effect under reaction atmospheres (such as H2 and
[a] Dr. D. Wang, Prof. X. Zheng
School of Materials Science and Engineering
University of Shanghai for Science and Technology
Shanghai 200093 (China)
[b] Dr. D. Wang, Dr. W. Niu, Dr. M. Tan, Prof. N. Tsubaki
Department of Applied Chemistry
Graduate School of Engineering, University of Toyama
Gofuku 3190, Toyama 930-8555 (Japan)
E-mail: tsubaki@eng.u-toyama.ac.jp
[c] Dr. M. Tan, Prof. M. Wu, Dr. Y. Li
State Key Laboratory of Heavy Oil Processing
China University of Petroleum
Shandong, Qingdao 266580 (PR China)
E-mail: wumb@upc.edu.cn
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cssc.201301123.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1398

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Results and Discussion
Catalytic performance of
carbon materials
The carbon materials such as GO
(preparation by modified Humm-
ers method from expanded
graphite),[25] AC (after acid treat-
ment), SiO2 (ID-type gel supplied
by Fuji Davison Chemical Ltd.,
specific surface area 270 m2 gÀ1
,
pore volume 1.22 cm3 gÀ1, aver-
age pore diameter 8.7 nm; after
calcination at 773 K),
G (pre-
pared by thermal exfoliation of
graphite oxide),[26] and CNTs
(after acid treatment)[27] were
Scheme 1. Representation of the catalytic conversion of cellulose or cellobiose into sorbitol and mannitol.
O2) especially when noble metal is the supported catalyst.[17]
These unique properties may be beneficial for the cellulose
conversion if graphene oxide or its derivative is used as cata-
lyst support.
The preparation method of catalysts directly influences their
catalytic performance due to the intrinsic relationship between
catalyst structure and reactivity. The nanoparticles loaded on
graphene oxide have been reportedly prepared by physical
methods such as electrostatic attraction, photoreduction, elec-
trophoretic, or electrochemical deposition methods.[18–20] These
catalysts have also been prepared by chemical methods in-
cluding hydrothermal/solvothermal method, ethylene glycol
reduction, solid-state reaction method etc.[21–23] Normally, these
methods are complicated and time-consuming. A rapid and
convenient method to prepare catalysts based on graphene-
oxide-supported metal nanoparticles is needed.
Figure 1. XRD patterns of graphite, graphite oxide (GO), reduced graphene
oxide (RGO), and graphene (G).
chosen as catalysts to catalyze cellobiose to sugar alcohols.
The XRD spectra in Figure 1 show the interlayer variations and
the crystalline properties of graphite, GO, G, and RGO. Graphite
shows a very sharp diffraction peak at 2q=26.268, correspond-
ing to the interlayer spacing of 0.34 nm. The diffraction peak
(2q) around 11.68 belongs to the reflection of GO, and the in-
terlayer spacing (0.80 nm) is much larger than that of pristine
graphite owing to the introduction of oxygen-containing func-
tional groups such as hydroxyl, epoxy, and carboxyl groups on
the surface of graphite sheet.[28] After reduction with the mix-
ture of EG and water, the intensity of the typical diffraction
peak (11.68) of GO decreases and the position of the peak
shifts to a higher angle, which is ascribed to the partial reduc-
tion of GO and the exfoliation of the layered GO nanosheets.
The G prepared with thermal exfoliation of GO does not show
any sharp peaks, demonstrating disorder and exfoliation. How-
ever, a broad and weak peak at 25.38 appears to show some
amount of reclustering.[29]
In this present work, the catalyst prepared by a microwave-
assisted reduction method was used in the conversion of cellu-
lose or cellobiose. Cellobiose, a glucose dimer connected by
a glycosidic bond, representing the simplest model molecule
of cellulose,[24] has been used in some experiments to investi-
gate the catalytic performance or reaction mechanism of cata-
lyst. Here, the catalytic performances of carbon materials [acti-
vated carbon (AC), graphite oxide (GO), reduced graphene
oxide (RGO), carbon nanotubes (CNTs) and graphene (G)] were
firstly studied. Then, various metal nanoparticles loaded on
RGO were prepared and their catalytic performances were
compared. Furthermore, the influence of supports on catalytic
performance was researched. As the best catalyst, Pt/RGO cata-
lyst was investigated in detail. Finally, the reaction mechanism
was explored.
Figure 2 shows the reaction results of cellobiose conversion
over various carbon materials. Without catalyst, cellobiose can
be hydrolyzed with 89.2% of total conversion and 32.7% of
glucose selectivity. The hydrolysis process in the conversion of
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1399

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
sion and the yield of sorbitol. All the catalysts were prepared
by the same EG reduction method with the assistance of mi-
crowave heating (Scheme S1). The advantage of microwave
heating over other conventional heating methods is that heat-
ing is uniform and rapid, which dramatically reduces process-
ing time, increases product yield, and enhances product purity
and properties.[35] Furthermore, the employment of microwave
radiation as a heat source has been demonstrated to provide
unique effects such as “superheating” of solvents above their
boiling point and the selective heating of strongly microwave-
absorbing materials, which can increase the metal loading effi-
Figure 2. Catalytic conversions of cellobiose into sorbitol with carbon mate-
rials as catalyst. Reaction conditions: catalysts weight=0.050 g; cellobiose
weight=0.171 g; reaction time=3 h; temperature=463 K;
pressure=5 MPa; H2 as reaction gas.
cellobiose to glucose can be carried out even in the absence
of catalyst, which may be due to H3O+ formed by water auto-
protolysis with the increase in temperature. These observations
are in agreement with a recent investigation of cellulose hy-
drothermal degradation.[30] The high conversion of cellobiose
may be due to its simple molecular structure with a glucose
dimer connected by a glycosidic bond. When carbon materials
are employed as catalysts in the conversion of cellobiose, the
conversion values are similar, but the product distribution is
dependent on the type of carbon catalyst used. In Figure 2
and Table S1 in the Supporting Information, the byproducts in-
clude 5-HMF, low molecular polyols, and others. For AC and G
catalysts, the primary by-product is 5-HMF, which is consistent
with a previous study in which the chemistry for the decompo-
sition pathways of cellulose and glucose in water indicates
that glucose undergoes isomerization to form fructose, which
can then undergo dehydration to form HMF.[31] For CNTs, GO,
and RGO catalysts, the yield of HMF is decreased whereas the
yields of hexitol and low molecular polyols including xylitol, ery-
thritol, glycerol, and ethylene glycol are increased. For GO or
RGO catalysts, the increased sorbitol yield may be due to the
oxygen-containing functionalities (alcohols, epoxides, and car-
boxylates), their unique structure, and defects in the graphene
sheets.[32] The GO and RGO catalysts for acid catalysis or hydro-
genation were also reported to be active for epoxides ring
Figure 3. Catalytic performance of different metal catalyst (10 wt% Ni,
10 wt% Cu, 5 wt% Rh, 5 wt% Ru, 5 wt% Pd, 5 wt% Pt) supported on RGO
catalyst for cellobiose conversion to sorbitol. Reaction conditions: catalyst
weight=0.050 g; cellobiose weight=0.171 g; reaction time=3 h; tempera-
ture=463 K; pressure=5 MPa; H2 as reaction gas.
ciency.[36] As shown in Figure 3, all the catalysts demonstrate
almost 100% conversion of cellobiose. The different product
distributions are influenced by the synergistic effect between
the loaded metal and RGO. The order of sorbitol yield is as fol-
lows:
Pt/RGO>Ru/RGO>Rh/RGO>Ni/RGO>Pd/RGO>Cu/
RGO. The highest sorbitol yield together with no production of
glucose is obtained for Pt/RGO, which indicates that Pt is the
most suitable catalyst with the highest sorbitol yield of approx-
imately 91.5%.
opening in the case of GO and nitrobenzene hydrogenation in
Catalytic performance of Pt nanocatalysts on different sup-
ports
[33,34]
the case of RGO.
As seen in Figure 2, the catalytic per-
formance of RGO is similar to that of GO, which indicates that
the catalytic activity is not influenced by the GO reduction pro-
cess. However, the yield of sorbitol on GO or RGO is also low.
Therefore, it is necessary to load hydrogenation catalysts onto
these carbon materials to improve their catalytic performance.
In order to clarify the effect of different supports on the con-
version of cellobiose to sorbitol, Pt, the best active metal as
above-mentioned, was also loaded on other supports. We
used TGA to determine the Pt loading amounts on various
carbon catalysts but there was no significant change, which
means that the catalytic activity of Pt loaded on different
carbon supports is comparable. As shown in Figure 4, the cel-
lobiose conversion is as high as 100%. The yield to sorbitol is
only 41.9% when Pt nanoparticles are used as catalyst.
Compared with Pt nanoparticles prepared by the same
Catalytic performance of different metal nanocatalysts sup-
ported on RGO
The typical hydrogenation catalysts such as Ni, Cu, Pd, Rh, Ru,
and Pt were loaded on RGO to improve the cellobiose conver-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1400

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
in Table 1, the Pt exposed surface area and dispersion of cata-
lyst also depend on the type of support. The Pt amounts of ex-
posed surface area and dispersion have the following order:
Pt/RGO>Pt/SiO2 >Pt/AC>Pt/CNTs>Pt/G. The catalyst with
greater exposed surface area and dispersion of Pt usually
shows higher conversion in hydrogenation reaction.[31,37]
Normally, the metallic catalyst with small metal particle size
possesses high dispersion and high catalytic activity. Why are
the Pt dispersion and the catalytic activity for Pt/G are so low?
The low catalytic activity of Pt/G catalyst may be due to the
high oxidation states of the metal catalyst on G support. As
shown in Figure S3, Pt4f spectra of Pt/G show that the main
valences of Pt are Pt2+ and Pt4+. However, the Pt0 is the active
site in the cellobiose conversion reaction (Scheme 1). The oxi-
dation state of Pt can be attributed to the reoxidation of
oxygen in air. The high oxidation state may be also the reason
that the Pt dispersion on Pt/G catalyst is so low. Compared
with Pt/AC, the Pt/CNTs catalyst with similar Pt dispersion has
better catalytic performance, which may be due to the follow-
ing reasons: Firstly, most pores in AC are micropores, and are
too small for cellobiose to enter.[32,38] Secondly, as in Figure 2,
CNTs with lots of surface functional groups formed by acid pre-
treatment have high sorbitol yield and low glucose yield, indi-
cating that the CNTs have higher catalytic activity than AC in
this reaction. With high Pt dispersion, Pt/SiO2 has the lowest
catalytic performance and this may be due to the same reason
as for Pt/AC. Therefore, the high catalytic activity is attributed
to the synergistic effects of supports and the supported Pt
nanoparticles. As in Figure 4, Pt loaded on RGO shows the
highest catalytic performance, which indicates that RGO is the
best support for the conversion of cellobiose to sorbitol.
Figure 4. Catalytic performance of Pt loaded on different supports for cello-
biose conversion to sorbitol. Reaction conditions: catalyst weight=0.050 g
(Pt nanoparticles, 0.0025 g); cellobiose weight=0.171 g; reaction time=3 h;
temperature=463 K; pressure=5 MPa; H2 as reaction gas.
method, the yield of sorbitol is improved from 41.9% to 91.5%
when RGO is used as the Pt support. Note that not all support
materials could improve the catalytic performance. From
Figure 4, it is clear that equal amounts of Pt loaded, the levels
of catalytic activity have the following order: Pt/RGO>Pt/
CNT>Pt nanoparticles> Pt/AC>Pt/SiO2 >Pt/G.
The XRD patterns of Pt nanoparticles loaded on different
supports are displayed in Figure S1. The strong diffraction
peaks at 2q=39.68, 46.28, and 67.58 can be assigned to the
characteristic (111), (200), and (220) crystalline planes of Pt. The
average crystallite sizes were calculated by the Scherrer equa-
tion at 39.68, which is shown in Table 1. Moreover, the mor-
phology of Pt nanoparticles and Pt nanoparticles loaded on
different supports was investigated by TEM. As in Figure S2,
the dispersion and particle sizes of Pt are dependent on sup-
port type. The particle size distribution in Figure S2 and Table 1
shows that the average size of Pt nanoparticles is about
3.5 nm for Pt nanoparticles, 3.6 nm for Pt/AC, 3.1 nm for Pt/
CNTs, 1.2 nm for Pt/G, 3.6 nm for Pt/RGO, and 3.1 nm for Pt/
SiO2. The dispersion and exposed surface area of the Pt nano-
particles were determined by chemisorption of CO. As shown
The effect of Pt particle size on the catalytic performance
To disclose the effect of Pt particle size on the catalytic activity,
Pt/RGO catalysts with different Pt particle sizes were prepared
through the same reduction method. In this microwave-assist-
ed method, Pt particle size can be controlled with different
temperature treatments. The XRD patterns of GO and Pt/RGO
after different heat treatments are shown in Figure 5. It can be
seen that the typical diffraction peak (002) of GO at 11.68 shifts
to higher angle after the loading of Pt nanoparticles on RGO
with a treatment temperature above 403 K; this indicates that
the GO is converted to RGO. The strong diffraction peaks at
2q=39.68, 46.28, and 67.58 can be assigned to the characteris-
tic (111), (200), and (220) crystalline planes of Pt, respectively.
Pt possesses face-centered-cubic (fcc) structure. The diffraction
peak of Pt (111) is used to estimate the Pt particle size by the
Scherrer equation. The calculated average particle sizes of Pt
on GO sheets are 2.0 nm for Pt/RGO-403 K, 2.7 nm for Pt/RGO-
413 K, 3.6 nm for Pt/RGO-433 K, and 4.3 nm for Pt/RGO-453 K.
Figure 6 shows the TEM images and Pt particle size distribu-
tions of Pt/RGO-T. All TEM images demonstrate that graphene
nanosheets are uniformly decorated by distributed Pt nanopar-
ticles with few aggregations, indicating a strong interaction be-
tween the graphene support and Pt particles. The mean size of
Table 1. The physicochemical properties of different catalysts.
Catalyst Exposed Pt surface Pt dispersion[a] Particle size[b] Crystallite sizes[c]
area[a] [m2 gÀ1
]
[%]
[nm]
[nm]
Pt/SiO2 65.2
26.4
13.7
12.8
3.5
3.1
3.6
3.1
1.2
3.6
2.9
7.7
4.0
–
3.6
Pt/AC 33.9
Pt/CNTs 31.6
Pt/G 8.7
Pt/RGO 79.2
[d]
32.3
[a] The surface area was determined by CO chemisorption. [b] The average
size of nanoparticles was evaluated from counting and averaging TEM
images. [c] The average crystallite sizes were calculated from XRD pattern
by using the Scherrer equation at 39.68. [d] The diffraction peak was too
low to estimate.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1401

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Figure 5. XRD patterns of GO and Pt/RGO-T (T=403, 413, 433, 453 K).
Figure 7. Conversions of cellobiose and cellulose into sorbitol and mannitol
by Pt/RGO-T catalysts with different particle size (T=403, 413, 433, 453 K).
Reaction conditions: catalyst weight=0.050 g; cellulose or cellobiose
weight=0.171 g; reaction time=24 h for cellulose and 3 h for cellobiose;
temperature=463 K; pressure=5 MPa; H2 as reaction gas.
Pt nanoparticles is very consistent with the value estimated by
the Scherrer equation from Figure 5.
As shown in Figure 7, the yield of sorbitol increases with the
mean size of Pt up to 3.6 nm and then decreases on further in-
crease in Pt particle size. The highest catalytic performance is
obtained by the Pt/RGO-433 catalyst with 91.5% sorbitol yield.
A similar trend is also observed when the cellobiose is replaced
by cellulose (see Scheme 1). The highest yield of hexitol is 69%
(58.9% of sorbitol and 10.1% of mannitol). The results show
that the Pt particle size is one of the critical factors affecting
the cellulose conversion. In addition, the catalytic activity may
also be affected by the different microwave temperatures at
which the RGO supports are processed.
Figure 6. TEM images and particle size distribution of Pt/RGO-T, at temperatures of (a) 403, (b) 413, (c) 433, and (d) 453 K.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1402

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
Characterization of Pt/RGO catalyst
The preparation, physicochemical properties, and the catalytic
performances of Pt/RGO catalyst with the highest catalytic ac-
tivity were systematically investigated. The morphology of
RGO and as-prepared Pt/RGO were characterized by means of
TEM. As shown in Figure 8a, the surface of RGO is smooth and
free from any contaminating particulate, which indicates that
few layered graphene oxides are formed, although the TEM
image cannot estimate the layer numbers of the graphene
oxide nanosheets exactly. As in Figure 8b, highly monodis-
persed Pt nanoparticles with a uniform size of 3.6 nm decorat-
ed on RGO surface are obtained by simple and rapid micro-
wave heating of H2PtCl6 and GO in the mixture of EG and
water. The most significant feature here is that the Pt nanopar-
ticles with a uniform size (3.6 nm) are monodispersed on the
surface of graphene oxide (as shown in Figure 8b and c). Fig-
ure 8d shows a high-resolution TEM image of Pt/RGO. The
measured interplanar spacing of the particle lattice is 0.23 nm,
which corresponds to the (111) crystal plane of Pt nanoparti-
cles shown in X-ray diffraction (XRD) data (Figure 5). These
TEM images confirm that the highly monodispersed Pt nano-
particles with uniform size can be successfully synthesized and
well dispersed on RGO sheets by the microwave-assisted re-
duction method. Compared with previous researches,[18–20] Pt
nanoparticles on graphene oxide sheets seem to be highly
monodispersed with smaller uniform size. Furthermore, the Pt/
RGO catalyst prepared by the microwave-assisted reduction
method is shown to possess a “fluffy” morphology and its sur-
face area is 180.2 m2gÀ1. The surface area is relatively low com-
pared to the surface area of the single graphene sheet,[20] indi-
cating a certain degree of restacking of the graphene sheets.
X-ray photoelectron spectra
Figure 8. TEM images of (a) RGO, (b, c) Pt/RGO, and (d) high-resolution trans-
mission electron microscopy (HR-TEM) image of Pt/RGO.
have been removed. These findings indicate that the addition
of H2PtCl6 plays an important role in the formation of RGO. Fig-
ure 9c shows the O1s XPS spectra of RGO and Pt/RGO nano-
composites. The decrease of oxygen-containing functional
groups can also be seen from the decreasing peaks of CÀO
and C=O. Pt4f spectra of Pt/RGO show the expected doublets
for Pt4f7/2 and Pt4f5/2, with Pt0, Pt2+, and Pt4+ oxidation states
(XPS) was used to investigate
the surface composition of RGO
and Pt/RGO catalysts. As shown
in Figure 9, the C1s core level for
RGO still shows a high degree of
oxidation, which consists of
three main components as-
signed to the CÀC (sp3 carbon,
285.8 eV), CÀC (sp2 carbon,
284.5 eV),
CÀO
(hydroxyl,
286.6 eV), and C=O (carbonyl,
288.4 eV) groups.[39] The large
amount of residual functional
groups including CÀO in RGO in-
dicate that GO was only partly
reduced by EG solution. After
the introduction of H2PtCl6, it is
clearly shown that the peak as-
sociated
with
CÀC
bond
(284.5 eV) becomes predominant
while the additional peak of CÀ
O (286.6 eV) decreases tremen-
dously, suggesting most oxygen-
Figure 9. Typical XPS spectra of RGO and Pt/RGO: (a) survey spectra and C1s of RGO, (b) survey spectra and C1s
containing functional groups of Pt/RGO, (c) O1s region XPS spectra, and (d) Pt4f region XPS spectra.
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1403

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
(Figure 9d). It is interesting to note that a significant percent-
age of Pt remains in its native state (Pt0) in Pt/RGO. In addition,
two different types of Pt (Pt2+ and Pt4+) can be assigned,
which suggests that oxygen linkages exist between the Pt
nanoparticles and the RGO surface, and between the Pt nano-
particles and the oxide layers formed on the Pt surface. The
small Pt nanoparticles are formed by the anchorage of Pt ions
on the graphene oxide surface and by their subsequent reduc-
tion to fine particles without obvious agglomeration.
Catalytic performances of Pt/RGO catalyst
The effect of the reaction temperature on the catalytic per-
formance of Pt/RGO catalyst is given in Figure 10. It can be
seen that the conversion of cellobiose increases from 75.2% to
100% on increasing the temperature from 423 K to 463 K. The
sorbitol yield increases simultaneously and reaches a maximum
at 463 K, then decreases on further increasing the temperature
to 483 K. It is though that a high temperature causes cello-
biose to be partially carbonized and that the produced sorbitol
can be converted further to other undesirable byproducts.
Consequently, the optimum reaction temperature is 463 K.
Using the milled cellulose as the raw material, the experi-
mental conditions such as the reaction temperature, the reac-
tion pressure, the reaction time, and the Pt concentration were
investigated. As shown in Figure 11a, the yield of sorbitol is
very low at 423 K. Compared with Figure 10, the low tempera-
ture might be not helpful to the hydrolysis of cellulose. The
yield of sorbitol and the conver-
Figure 10. Influence of the reaction temperature on the conversion of cello-
biose to sorbitol. Reaction conditions: catalyst weight=0.050 g; cellobiose
weight=0.171 g; reaction time=3 h; temperature=423-483 K; pressur-
e=5 MPa; H2 as reaction gas.
biose can be easily hydrolyzed compared with cellulose, indi-
cating that the hydrolysis of cellulose is the rate determining
step. The effect of level of Pt loading in the Pt/RGO catalyst is
also studied. The yield of sorbitol can be increased from 24.9%
to 58.9% on increasing Pt loading from 1 wt% to 5 wt%.
In order to study the stability of the Pt/RGO catalyst, the cat-
alyst was characterized with TEM after a 24 h reaction. As
shown in Figure S4, the average size of Pt nanoparticles on the
sion of cellulose increase with in-
creasing temperature. However,
the sorbitol yield is relatively low
when the reaction temperature
is 483 K, which might be due to
the partial decomposition or de-
hydration of products and the
generation of some byproducts,
such as low-molecular-weight
polyols, sorbitan, isosorbide, CO,
and CH4. The influence of reac-
tion pressure is shown in Fig-
ure 11b, from which it can be
seen that the conversion of cel-
lulose increases with increasing
pressure. Furthermore, the pres-
sure also influences the yield of
sorbitol. Higher pressure results
in an increase in sorbitol yield.
The reaction time is also found
to be important in this reaction,
as shown in Figure 11c. The con-
version of cellulose (36.3%) and
sorbitol yield are very low with
a reaction time of 1 h. However,
Figure 11. Influences of (a) reaction temperature, (b) reaction pressure, (c) reaction time, and (d) Pt loading
amount on the conversion of cellulose to sorbitol. Reaction conditions: catalyst weight=0.050 g; cellulose
weight=0.171 g; H2 as reaction gas; temperature=463 K (in (a) 423–483 K); pressure=3 MPa (in (b) 1–5 MPa); re-
the conversion of cellobiose is as
high as 100% when the reaction
time is 3 h (Figure 10). The cello- action time=24 h (in (c) 1–24 h); Pt loading amount=5% (in (d) 1–5%).
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1404

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
used Pt/RGO catalyst is similar to that on the fresh one. Particle
aggregation is not observed in the TEM image, suggesting
that the catalyst is very stable and can be reused. As shown in
Figure 12, the reusability is also indicated by the similar sorbi-
tol selectivity achieved on using the used Pt/RGO as catalyst
and fresh cellulose as reactant. The slight decrease of cellulose
conversion after several runs might be due to the loss of cata-
lysts during the transfer process.
Scheme 2. Conversion mechanism of cellulose to sorbitol catalyzed by Pt/
RGO.
sheets act as an excellent support and stabilizer for the very
finely dispersed Pt nanoparticles, preventing nanoparticle ag-
gregation. The sorbitol yield is as high as 91.5% or 58.9%
when cellobiose or cellulose are employed as the reactant, re-
spectively. The optimum particle size of Pt is 3.6 nm and the
optimum reaction temperature is 463 K. The improvement of
catalytic activity is attributed to the hydrogen spillover effects
of the Pt/RGO catalyst.
Figure 12. Reuse experiments of Pt/RGO catalyst for cellulose conversion.
Reaction conditions: catalysts weight=0.050 g; cellulose weight=0.171 g,
reaction time=24 h; temperature=463 K; pressure=5 MPa; H2 as reaction
gas.
Experimental Section
Catalyst preparation
Reaction mechanism of cellulose conversion
The support materials included GO,[25] AC (after acid treatment),
SiO2 (ID, after calcination at 773 K), G (prepared by thermal exfolia-
tion of graphite oxide.[26]), CNTs (after acid treatment).[27] For further
details, please see the Supporting Information.
The enhanced catalytic performance of the Pt/RGO catalyst for
the conversion of cellulose to sorbitol may be due to the fol-
lowing reasons: The increase in utilization efficiencies of plati-
num nanocatalysts on RGO supports can be attributed to the
enlarged surface area and the well-dispersion of the RGO sup-
ports and catalyst. The optimum size of Pt particles and the ex-
posed crystal plane are beneficial for the hydrogenation of re-
actants. Pt on RGO promotes the protonations of water, which
is due to the hydrogen molecules spilt over to create in situ
acid sites and to catalyze hydrolysis of cellulose or cellobiose
(Scheme 2). The intermediate glucose or other oligosaccharides
can be selectivity hydrogenated to sorbitol on the highly
active Pt nanoparticles. Therefore, in the conversion of sugars
to sorbitol, the RGO supports highly dispersed Pt nanoparticles
by means of an anchoring effect, maintains surface acidity
through H2 spillover from Pt, catalyzes hydration reaction
through its acidic sites, and suppresses the formation of by-
products.
EG reduction with the assistance of microwave radiation was ap-
plied to load metallic nanoparticles on RGO or on other support
materials. Metal salt solution, such as H2PtCl6, RuCl3, H2PdCl6,
Ni(NO3)2, or Cu(NO3)2 solution, was chosen as metal source to pre-
pare the metallic nanoparticles. For example, in a typical synthesis
of Pt/RGO catalyst, the GO and H2PtCl6 were dispersed in EG solu-
tion with the assistance of ultrasonic treatment. After being ultra-
sonically treated for 30 min, the mixture was then put into an au-
tomated focused microwave system and treated at 393–453 K for
30 min. The catalysts were obtained through a redox reaction
heated by microwave. The reduction reaction could be observed
by a color change from yellow GO in the mixture of EG and aque-
ous solution, to dark black after 30 min of microwave heating. The
solid products were then collected by filtration and washed with
deionized water as well as ethanol and were then dried. The metal
loading in all catalysts was 5 wt%. The obtained catalyst was de-
fined as M/RGO. RGO was prepared by the same method as above,
but without the addition of metal salt. Pure Pt nanoparticles were
prepared from the same EG reduction method without the exis-
tence of supports.
Conclusions
Pt/RGO catalyst is prepared by microwave-assisted EG reduc-
tion method, which presents high activity and selectivity for
the conversion of cellobiose or cellulose to sorbitol. The struc-
ture and the catalytic performance of Pt/RGO catalyst are sys-
tematically investigated. The high catalytic activity is attributed
to the synergistic effects of supports and the supported Pt
nanoparticles. The results show that the graphene oxide
Characterization of catalysts
The morphologies of the samples were characterized with a high
resolution TEM (JEOL JEM-2100 UHR) operated at 200 kV. The crys-
tal structure of the materials was confirmed by means of XRD
using a Rigaku D/max-2550 V diffractometer employing CuKa radia-
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1405

CHEMSUSCHEM
FULL PAPERS
www.chemsuschem.org
[6] D. An, A. Ye, W. Deng, Q. Zhang, Y. Wang, Chem. Eur. J. 2012, 18, 2938–
2947.
[7] N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang, J. G. Chen,
Angew. Chem. Int. Ed. 2008, 47, 8510–8513; Angew. Chem. 2008, 120,
8638–8641.
[8] L. N. Ding, A. Q. Wang, M. Y. Zheng, T. Zhang, ChemSusChem 2010, 3,
818–821.
[9] A. Fukuoka, P. L. Dhepe, Angew. Chem. Int. Ed. 2006, 45, 5161–5163;
Angew. Chem. 2006, 118, 5285–5287.
[10] H. Wang, L. Zhu, S. Peng, F. Peng, H. Yu, J. Yang, Renewable Energy
2012, 37, 192–196.
[11] P. Yang, H. Kobayashi, K. Hara, A. Fukuoka, ChemSusChem 2012, 5, 920–
926.
[12] S. You, I. Baek, Y. Kim, K.-E. Jeong, H.-J. Chae, T.-W. Kim, C.-U. Kim, S.-Y.
Jeong, T. Kim, Y.-M. Chung, S.-H. Oh, E. Park, Korean J. Chem. Eng. 2011,
28, 744–750.
[13] J. Pang, A. Wang, M. Zheng, Y. Zhang, Y. Huang, X. Chen, T. Zhang,
Green Chem. 2012, 14, 614–617.
tion (l=1.54 ꢁ; scanning rate: 0.028/s). XPS was conducted using
an ESCALAB 250Xi spectrometer equipped with a pre-reduction
chamber. The position of the C1s peak (284.5 eV) was used to cor-
rect the XPS binding energies for all catalysts for possible charging
effects. Thermal analysis was carried out on a DTG-60 (Shimadzu)
in air flow. The loading amount of Pt was analyzed by means of an
inductive coupled plasma emission spectrometer (Shimadzu ICPE-
9000). The metal dispersion and metal surface area was deter-
mined by CO chemisorption using a pulse chemisorption mode
(BELCAT, BEL Japan, Inc.). Samples were treated in He flow by in-
creasing the temperature to 423 K at 5 KminÀ1 and holding at
423 K for 1 h. Samples were then heated to 673 K at 10 KminÀ1
and reduced in 10% H2/Ar at 673 K for 1 h. Chemisorbed hydrogen
was removed by treatment in He flow at 493 K for 1 h. Samples
were then exposed to pulses of 5% CO/He. CO concentration was
measured using a thermal conductivity detector. Dispersions, de-
fined as the fraction of Pt atoms exposed at surfaces, were deter-
mined by assuming a 1:1 ratio of CO molecules to surface Pt
atoms.
[14] C. Huang, C. Li, G. Shi, Energy Environ. Sci. 2012, 5, 8848–8868.
[15] X. Chu, Q. Zhu, W.-L. Dai, K. Fan, RSC Adv. 2012, 2, 7135–7139.
[16] A. Dhakshinamoorthy, M. Alvaro, P. Concepcion, V. Fornes, H. Garcia,
Chem. Commun. 2012, 48, 5443–5445.
[17] D. R. Dreyer, H.-P. Jia, C. W. Bielawski, Angew. Chem. Int. Ed. 2010, 49,
6813–6816; Angew. Chem. 2010, 122, 6965–6968.
Catalytic reaction
[18] J. Zhu, Y. K. Sharma, Z. Zeng, X. Zhang, M. Srinivasan, S. Mhaisalkar, H.
Zhang, H. H. Hng, Q. Yan, J. Phys. Chem. C 2011, 115, 8400–8406.
[19] Z. Sun, X. Lu, Ind. Eng. Chem. Res. 2012, 51, 9973–9979.
[20] J. Wang, T. Tsuzuki, B. Tang, X. Hou, L. Sun, X. Wang, ACS Appl. Mater. In-
terfaces 2012, 4, 3084–3090.
[21] M.-S. Wu, Y.-P. Lin, C.-H. Lin, J.-T. Lee, J. Mater. Chem. 2012, 22, 2442–
2448.
[22] S. Wu, Q. He, C. Zhou, X. Qi, X. Huang, Z. Yin, Y. Yang, H. Zhang, Nano-
scale 2012, 4, 2478–2483.
[23] Y. Li, W. Gao, L. Ci, C. Wang, P. M. Ajayan, Carbon 2010, 48, 1124–1130.
[24] N. Yan, C. Zhao, C. Luo, P. J. Dyson, H. Liu, Y. Kou, J. Am. Chem. Soc.
2006, 128, 8714–8715.
[25] W. S. Hummers, R. E. Offeman, J. Am. Chem. Soc. 1958, 80, 1339–1339.
[26] A. Kaniyoor, T. T. Baby, S. Ramaprabhu, J. Mater. Chem. 2010, 20, 8467–
8469.
[27] D. Wang, G. Yang, Q. Ma, M. Wu, Y. Tan, Y. Yoneyama, N. Tsubaki, ACS
Catal. 2012, 2, 1958–1966.
[28] J. Li, H. Lin, Z. Yang, J. Li, Carbon 2011, 49, 3024–3030.
[29] Y. Guo, C. Bao, L. Song, B. Yuan, Y. Hu, Ind. Eng. Chem. Res. 2011, 50,
7772–7783.
The conversion of cellobiose or cellulose was performed with
a batch-type high-pressure autoclave reactor. Typically, 0.050 g cat-
alyst and 0.171 g cellulose (after ball milling for 4 days) or cello-
biose were loaded into the reactor precharged with 20 mL H2O.
The reactor H2 pressure was maintained at 5 MPa at room temper-
ature, and then the reaction was conducted at 463 K for 3 h. After
reaction, the solid phase was separated by centrifugation, and the
liquid products were analyzed with HPLC (Waters Assoc, USA).
Cellulose conversion was determined as reported by the weight
difference of dried cellulose before and after the reaction.[5] The
cellobiose conversion was calculated by means of the HPLC results
before and after the reaction. The product selectivity was calculat-
ed based on the carbon basis (calculated by dividing the amount
of carbon found in a particular product by the amount of carbon
found in the products). The product yield was calculated as fol-
lows: yield(%)=(weight of product)/(weight of cellulose or cello-
biose charged in reactor).
[30] V. Jollet, F. Chambon, F. Rataboul, A. Cabiac, C. Pinel, E. Guillon, N. Es-
sayem, Green Chem. 2009, 11, 2052–2060.
[31] W. Deng, R. Lobo, W. Setthapun, S. T. Christensen, J. W. Elam, C. L. Mar-
shall, Catal. Lett. 2011, 141, 498–506.
Acknowledgements
[32] X. Huang, X. Qi, F. Boey, H. Zhang, Chem. Soc. Rev. 2012, 41, 666–686.
[33] Y. B. Huang, Y. Fu, Green Chem. 2013, 15, 1095–1111.
[34] Y. J. Gao, D. Ma, C. L. Wang, J. Guan, X. H. Bao, Chem. Commun. 2011,
47, 2432–2434.
M.T. thanks the China Scholarship Council (CSC) for financial sup-
port.
[35] M. Baghbanzadeh, L. Carbone, P. D. Cozzoli, C. O. Kappe, Angew. Chem.
Int. Ed. 2011, 50, 11312–11359; Angew. Chem. 2011, 123, 11510–11561.
[36] C. O. Kappe, Chem. Soc. Rev. 2008, 37, 1127–1139.
[37] R. S. Oosthuizen, V. O. Nyamori, Platinum Met. Rev. 2011, 55, 154–169.
[38] S. Barrientos-Ramꢂrez, G. Montes de Oca-Ramꢂrez, A. Sepffllveda-Escriba-
no, M. M. Pastor-Blas, A. Gonzµlez-Montiel, F. Rodrꢂguez-Reinoso, Catal.
Today 2010, 150, 42–48.
Keywords: biomass · graphene oxide · hydrogen spillover
effect · hydrogenation · sorbitol
[1] A. M. Ruppert, K. Weinberg, R. Palkovits, Angew. Chem. Int. Ed. 2012, 51,
2564–2601; Angew. Chem. 2012, 124, 2614–2654.
[2] M. S. Mettler, S. H. Mushrif, A. D. Paulsen, A. D. Javadekar, D. G. Vlachos,
P. J. Dauenhauer, Energy Environ. Sci. 2012, 5, 5414–5424.
[39] D. Sun, X. Yan, J. Lang, Q. Xue, J. Power Sources 2013, 222, 52–58.
[3] K. Kohse-Hçinghaus, P. Oßwald, T. A. Cool, T. Kasper, N. Hansen, F. Qi,
C. K. Westbrook, P. R. Westmoreland, Angew. Chem. Int. Ed. 2010, 49,
3545–3545; Angew. Chem. 2010, 122, 3625–3625.
[4] H. Kobayashi, H. Ohta, A. Fukuoka, Catal. Sci. Technol. 2012, 2, 869–883.
[5] C. Luo, S. Wang, H. Liu, Angew. Chem. Int. Ed. 2007, 46, 7636–7639;
Angew. Chem. 2007, 119, 7780–7783.
Received: October 21, 2013
Published online on March 19, 2014
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemSusChem 2014, 7, 1398 – 1406 1406