Synthesis of sugar alcohols by hydrolytic hydrogenation of cellulose over supported metal catalysts

Article abstract of DOI:10.1039/c0gc00666a

Cellulose is converted into sorbitol and related sugar compounds over water-tolerant and durable carbon-supported Pt catalysts under aqueous hydrogenation conditions. Pre-treatment of cellulose with ball-milling effectively reduces the crystallinity and particle size of cellulose, which results in high conversion of cellulose to sorbitol and mannitol. The selectivity of sorbitol increases by using Cl-free metal precursors in the catalyst preparation as residual Cl on the catalysts promotes the side-reactions. The transformation of cellulose to sorbitol consists of the hydrolysis of cellulose to glucose via water-soluble oligosaccharides and the successive hydrogenation of glucose to sorbitol. The hydrolysis of cellulose is the rate-determining step, and the Pt catalysts promote both the hydrolysis and the hydrogenation steps.

Full text of DOI:10.1039/c0gc00666a

View Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2011, 13, 326
www.rsc.org/greenchem
PAPER
Synthesis of sugar alcohols by hydrolytic hydrogenation of cellulose over
supported metal catalysts†
Hirokazu Kobayashi, Yukiko Ito, Tasuku Komanoya, Yuto Hosaka,a,b Paresh L. Dhepe,‡ Koji Kasai,a
a
a
a,b
a
a
a
Kenji Hara and Atsushi Fukuoka*
Received 10th October 2010, Accepted 15th November 2010
DOI: 10.1039/c0gc00666a
Cellulose is converted into sorbitol and related sugar compounds over water-tolerant and durable
carbon-supported Pt catalysts under aqueous hydrogenation conditions. Pre-treatment of
cellulose with ball-milling effectively reduces the crystallinity and particle size of cellulose, which
results in high conversion of cellulose to sorbitol and mannitol. The selectivity of sorbitol
increases by using Cl-free metal precursors in the catalyst preparation as residual Cl on the
catalysts promotes the side-reactions. The transformation of cellulose to sorbitol consists of the
hydrolysis of cellulose to glucose via water-soluble oligosaccharides and the successive
hydrogenation of glucose to sorbitol. The hydrolysis of cellulose is the rate-determining step, and
the Pt catalysts promote both the hydrolysis and the hydrogenation steps.
Introduction
In amylose of starch, glucose units are linked by a-1,4-glycosidic
bonds, while in cellulose they are joined together through b-
,4-glycosidic bonds (Fig. 1). Although both polymers form
Catalytic conversion of biomass to fuels and chemicals has
attracted great attention as one of the future technologies for
inhibiting the global warming and for building up the low-
1
crystalline structures with extensive intra- and inter-molecular
hydrogen bonds, those of amylose are relatively weak, and
amylose is gelatinised in hot water. On the other hand, those
of cellulose are remarkably more rigid and impenetrable even
1
–3
carbon society. In the photosynthesis reaction, a huge amount
of carbohydrates and oxygen are produced from carbon dioxide
and water in plants using sunlight, and the plant biomass is
a potentially abundant and renewable resource. Recently, the
use of edible biomass for producing bioethanol has resulted in a
competition between food and fuel, and this lesson has given us a
consensus that we should use non-food biomass as a precursor to
fuels and chemicals. A typical nonfood biomass is lignocellulose
in cell walls of plants. Lignocellulose is composed of cellulose,
hemicellulose and lignin, in which cellulose is a main component
5,6
by hot-compressed water at <590 K. As a result, cellulose is
greatly more resistant to depolymerisation than starch.
4
in woods. Therefore, the first target among non-food biomass
5
is cellulose.
Cellulose is a polymer of D-glucose, and starch is also a
polymer of glucose. They have different properties derived from
the contrastive configuration of C–O bonding in polymer chains.
a
Catalysis Research Centre, Hokkaido University, Kita 21 Nishi 10,
Kita-ku, Sapporo, 001-0021, Japan.
E-mail: fukuoka@cat.hokudai.ac.jp; Fax: +81 11 706 9139
b
Division of Chemical Sciences and Engineering, Graduate School of
Chemical Sciences and Engineering, Kita 13 Nishi 8, Kita-ku, Sapporo,
0
60-8628, Japan
†
Electronic supplementary information (ESI) available: Hydrogenation
Fig. 1 Structures of cellulose and amylose.
of sugars, yields of products and time course of temperature in Fig. 9,
G
EDX analyses, curve fitting of the cellulose conversion to determine k .
See DOI: 10.1039/c0gc00666a
An extensive number of works have been devoted to the
degradation of cellulose using enzymes, mineral acids and
‡
Present address: National Chemical Laboratory, Pune 411008, India
3
26 | Green Chem., 2011, 13, 326–333
This journal is © The Royal Society of Chemistry 2011

View Online
4,7–9
supercritical water,
but these methods have drawbacks such
physisorbed water was determined by using a total organic
as slow reaction rate, difficult separation of products and
catalysts, corrosion hazards, harsh conditions and generation of
large amount of neutralisation waste. Heterogeneous catalysis
has advantages to overcome these drawbacks. We reported
the first catalytic degradation of cellulose into sugar alcohols
such as sorbitol and mannitol by supported metal catalysts
carbon analyser (TOC; Shimadzu SSM-5000A) and weight loss
in vacuum drying at 373 K, and both experiments indicated that
the water content was 5.0 ± 1.0 wt% in the cellulose ball-milled
for 2 days.
Hydrogenation of cellulose
10
under hydrogenolysis conditions. The sugar alcohols are useful
precursors for the production of plastics and pharmaceuticals.
Since then, cellulose conversion by heterogeneous catalysis
The conversion of cellulose under H
hydrogenation of cellulose, was conducted in a stainless steel
SUS316) high-pressure reactor (OM Lab-Tech MMJ-100,
00 mL). The milled cellulose (324 mg, 1.90 mmol glucose units,
containing 5.0 wt% physisorbed water), a supported Pt catalyst
195 mg, Pt 20 mg-atom) and water (40 mL) were charged in
the reactor (molar ratio of glucose unit/Pt = 95), and then
pressurised with H of 5.0 MPa. The reactor was heated to
63 K and maintained at the temperature for 24 h with stirring
2
pressure, denoted as
(
1
1
1
has rapidly developed for the synthesis of sugar alcohols,
12
13
ethylene glycol and glucose. The hydrogenolysis reaction
using supported metal catalysts has also been investigated on
(
14
the conversion of lignin and its model compounds. However,
there still remains a need to improve the productivity in the
synthesis of sugar alcohols. In this work, we studied the several
factors controlling the reactivity of cellulose and the catalytic
activity to discuss the detail of the reaction mechanism.
2
4
at 600 rpm. Products were separated by centrifugation and
decantation, and water-soluble products were analysed by high-
performance liquid chromatography (HPLC; Shimadzu LC10-
ATVP, refractive index detector). The columns used in this
work were a Phenomenex Rezex RPM-Monosaccharide Pb++
Experimental
Catalyst preparation
-
1
column (ø7.8 ¥ 300 mm, mobile phase: water 0.6 mL min ,
353 K) and a Shodex Sugar SH-1011 column (ø8 ¥ 300 mm,
Catalyst supports used in this study were as follows: carbon
blacks BP2000 (Black Pearls 2000) and VULCAN XC72 were
supplied from Cabot. Activated carbon denoted as AC (Norit
-
1
mobile phase: water 0.5 mL min , 323 K).
SX Ultra) was purchased from Aldrich. Al
and ZrO (JRC-ZRO-2) were obtained from Catalysis Society
of Japan, and TiO (P-25) from Degussa.
The supported metal catalysts prepared from Pt(NH
NO , Pt(NH Cl , H PtCl , Ru(NO)(NO and RuCl were
2
O (JRC-ALO-2)
3
Results and discussion
2
Pre-treatment of cellulose
2
3
2
) -
Cellulose has crystalline and amorphous parts in its structure.
It is known that the amorphous region is first attacked by acids
(
2
)
2
3
)
4
2
2
6
3
)
3
3
denoted as Pt(N), Pt(Cl, N), Pt(Cl), Ru(N) and Ru(Cl), respectively. Other
1
6
a
in the hydrolysis reaction as cellulose I crystal, which is a
metal precursors used in this study were Rh(NO
Pd(NO . The typical catalyst, Pt(N)/BP2000 (Pt metal loading
wt%) was prepared by a conventional impregnation method
as follows: Pt(NH (NO aq. (0.202 mmol in 5 mL of water)
was dropped into a mixture of carbon black BP2000 and water
3
)
3
, Ir(acac)
3
and
typical crystal phase of natural cellulose, has 8 hydrogen bonds
for each glucose unit, while the number of those in amorphous
)
3 2
2
17
cellulose is 5.3. Likewise, it has been reported that amorphous
cellulose is hydrolysed more rapidly than crystalline one by
3
)
2
)
2 2
18
enzyme mixtures, although different cellulase enzymes have
(
20 mL), and the mixture was stirred for 16 h. After drying
19
different preferential reaction sites on cellulose. The conversion
of the crystalline region into an amorphous one gives higher
reactivity to cellulose. Various methods have been reported for
the pre-treatment of cellulose to reduce the crystallinity, and
in our study ball-milling was selected for this purpose. Fig. 2
depicts the change of XRD diffractions of the cellulose after
-
1
in vacuo, the solid was reduced with H (30 mL min ) at 673 K
2
for 2 h. In the preparation using oxide supports, the catalyst
-
1
precursors were calcined in O
before the H -reduction, and in the impregnation of Ir(acac)
CH Cl was used as the solvent. The catalyst was analysed by
2
flow (30 mL min ) at 673 K
2
3
,
2
2
using energy dispersive X-ray spectroscopy (EDX; Shimadzu
EDX-720), X-ray diffraction (XRD; Rigaku Miniflex, Cu-Ka
radiation) and CO adsorption measurement (Quantachrome,
ChemBET-3000).
◦
the pre-treatment. The intensity of (002) diffraction at 22.5
clearly decreased by the treatment. The crystallinity index (CrI)
of each cellulose was 81% (non-treated), 22% (2-day-milled)
and 14% (4-day-milled), estimated from the intensity difference
7a,20
of the (002) diffraction and that of amorphous part.
The
Pre-treatment of cellulose
CrI of cellulose significantly lowered during the pre-treatment
for 2 days, but further milling gave no improvement. Thus,
we compared the non-treated and 2-day-milled cellulose in the
further characterisation.
Cellulose was pre-treated using a ball-mill (Nitto Kagaku,
ANZ-51S). Cellulose (Merck, Avicel microcrystalline, 10 g) was
charged in a rotary ceramic container with ZrO
2
balls (1 kg, 1 cm
diameter). The container was rotated at a speed of 60 rpm for
days (48 h) or 4 days (96 h). The ball-milled cellulose was
15
The CrI can also be estimated by solid-state NMR. Fig. 3
1
3
2
exhibits CP/MAS C NMR spectra of non-treated and 2-day-
milled cellulose. C1–C6 carbons of pyranose ring resonate at
60–110 ppm, and the peaks at 82–92 ppm are assigned to C4.
The C4-peak is allowed to be divided into two parts; the lower-
field peak is assigned to crystalline part and the higher-field one
characterised by XRD, cross polarisation/magic angle spinning
nuclear magnetic resonance (CP/MAS NMR; Bruker MSL-
1
3
4
00, C 100.6 MHz, rotation rate 4 kHz, external standard:
glycine), optical microscope (Keyence VHX-1000) and laser
diffraction (Nikkiso Microtrac MT3300EXII). The amount of
21
is derived from amorphous part. The CrI calculated from the
This journal is © The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 326–333 | 327

View Online
The primary particles with spherical or ellipsoidal structures in
the range of 2–10 mm were observed in a higher magnification
image (Fig. 4(c)). The ball-milling treatment changes both the
particle size and the morphology.
Fig. 5 shows the distribution of the secondary particle diame-
ter of cellulose in water. We assume that this analysis reflects the
valid specific surface area of cellulose to be contacted with bulk
solid catalysts. The diameter range of the non-treated cellulose
was 10–300 mm and the median diameter (the diameter that the
cumulative frequency is 50%) was 67 mm, which conforms to the
particle diameter observed by the microscope. The distribution
shifted lower and the median diameter diminished to 42 mm by
the pre-treatment for 2 days. Although the ball-milled cellulose
is rather aggregated in water as the secondary particles are one-
order of magnitude larger than the primary particles, the size is
still smaller than that of non-treated cellulose.
Fig. 2 XRD patterns of ball-milled cellulose.
Fig. 5 Distribution of the particle diameter of cellulose in water.
Crosses (¥) and dashed line: non-treated cellulose. Circles (᭹) and solid
line: 2-day-milled cellulose.
1
3
Fig. 3 CP/MAS C NMR of cellulose with or without the pre-
treatment. Crosses (¥): crystalline part, filled circles (᭹): amorphous
part.
From these results, it is demonstrated that the ball-milling
treatment decreases the CrI and the particle size of cellulose,
thus enabling us to expect the improvement of the reactivity of
cellulose.
area ratio of the two peaks were 65% for the non-treated cellulose
and 10% for the 2-day-milled cellulose. These CrI values are
similar to those determined by XRD. Accordingly, we confirmed
that the crystallinity is effectively decreased by the ball-milling
treatment.
Non-treated cellulose had various shapes as shown in
Fig. 4(a). A typical dimension of needle-type one was ca. 20 ¥
0 mm and some parts were aggregated to make large secondary
particles. After the ball-milling for 2 days, cellulose turned to
smaller particles and their major part was aggregated (Fig. 4(b)).
Catalytic conversion of cellulose
Effect of the cellulose pre-treatment. The reactivity of cel-
lulose with or without the pre-treatment was compared in the
hydrogenation reaction. For this purpose, we used Pt(N)/BP2000
catalyst in this experiment because of the highest activity and
good durability among the catalysts used in this work (see
Optimisation of catalysts). Table 1 summarises the results of the
5
Fig. 4 Optical microscope images of (a) non-treated cellulose and (b, c) cellulose ball-milled for 2 days.
28 | Green Chem., 2011, 13, 326–333
3
This journal is © The Royal Society of Chemistry 2011

View Online
a
2
Table 1 Degradation of cellulose with H by supported metal catalysts
Yield based on carbon/%
Sum of sorbitol
and mannitol
b
c
Entry
Catalyst
Sorbitol
Mannitol
1,4-Sorbitan
C
2
–C
3
polyols
Glucose
Cellulose conversion /%
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
Pt(Cl)/Al
Pt(Cl, N)/Al
2
O
3
31.6
36.7
46.0
35.2
45.1
37.6
38.8
48.7
52.5
52.7
53.4
44.3
41.9
16.8
20.0
35.7
36.1
32.5
53.2
36.9
5.2
6.6
7.7
4.8
8.2
8.1
9.6
4.3
9.0
11.0
12.1
6.3
11.1
9.7
5.3
6.2
10.0
9.1
5.4
4.1
12.7
3.2
2.3
0.2
38.2 (43)
44.4
15.8
7.3
3.0
1.2
1.5
13.6
5.3
1.9
2.2
3.0
10.4
1.9
1.3
0
1.3
2.7
1.9
15.9
3.2
1.9
0.3
0.9
0.1
8.3
9.2
0.8
0.6
2.0
3.7
0.9
1.5
1.8
2.0
1.0
1.4
1.7
4.3
1.6
1.7
0.7
7.8
0.5
0.5
6.2
8.9
2.2
N.d.
N.d.
0.2
0.5
0.7
0.7
0
0.1
0.1
0.1
0
0
0
0.2
0
0
0.2
N.d.
0.1
0.1
0.5
0.3
0.3
89.2
N.d.
72.3
78.9
101
98.0
65.6
81.9
101
2
O
3
3
Pt(N)/Al
Pt(N)/Al
Pt(N)/Al
Pt(N)/Al
2
2
2
2
O
O
O
O
50.8 (70)
43.4 (55)
53.2 (53)
47.2 (48)
43.1 (66)
57.7 (70)
63.5 (63)
64.8 (64)
59.7 (65)
55.4 (63)
51.6 (59)
22.1 (29)
26.2 (33)
45.7 (62)
45.2 (56)
37.9 (40)
57.3 (71)
49.6 (57)
8.4 (10)
d
3
3
3
d
e
f
Pt(N)/BP2000
Pt(N)/BP2000
Pt(N)/BP2000
Pt(N)/BP2000
Pt(N)/BP2000
Pt(N)/XC72
Pt(N)/XC72
Pt(N)/AC
Pt(N)/TiO
Pt(N)/ZrO
Pt(N)/ZrO
Ru(Cl)/Al
Ru(N)/Al
Ru(N)/BP2000
Rh/BP2000
Ir/BP2000
d
d
0
1
2
3
4
5
6
7
8
9
0
1
2
3
101
g
92.3
87.8
86.9
75.3
78.8
74.1
80.6
95.5
81.1
86.4
86.8
77.3
76.6
d
2
2
d
2
2
O
3
2
O
3
20.7
0.3
23.0 (30)
0.5 (1)
Pd/BP2000
a
Reaction conditions; cellulose 324 mg (ball-milled for 2 days), catalyst 20 mg-atom-metal (metal 2 wt%), water 40 mL, reaction time 24 h, T =
b
4
63 K, P(H ) = 5.0 MPa at r.t. Selectivity based on the substrate conversion is shown in parenthesis. Sum of glycerine, propylene glycol and ethylene
2
c
d
glycol. Based on the loss of weight during the reaction. The reuse experiment using the residue of the experiment of the previous entry number and
e
f
fresh cellulose of 324 mg. The conversion and yields are based on the fresh cellulose. HCl of 59 mmol was added. Cellulose was used without the
pre-treatment. 4-Day-milled cellulose was used.
g
hydrogenation under pressurised H
major products were sugar alcohols (sorbitol 39% and mannitol
%, total 43%, entry 7) and turnover number (TON) based
2
conditions at 463 K. The
the effect of Cl in catalyst precursors on the conversion of the pre-
treated cellulose. In this work, we used a Cl-free alumina support
4
(JRC-ALO-2). Pt(Cl)/Al
2
3
O gave sorbitol (32%) and mannitol
on bulk Pt for the production of the sugar alcohols was 42.
The conversion of the non-treated cellulose, determined by the
weight loss in the reaction, was 66%. Therefore, the selectivity
for the sugar alcohols based on the converted cellulose was 66%.
Other identified products were 1,4-sorbitan, ethylene glycol,
propylene glycol and glycerine. By ball-milling cellulose for
(6.6%), and the conversion of cellulose was 89% (Table 1,
entry 1). The selectivity for the sugar alcohols based on the
conversion was 43%. Large amounts of 1,4-sorbitan and C–C
and/or C–O dissociated compounds such as ethylene glycol and
propylene glycol were also generated (total 50% yield). The yield
of the sugar alcohols was improved to 44% by using a less Cl-
2
days, the conversion of cellulose was raised from 66% to 82%,
containing precursor [Pt(NH
free precursor [Pt(NH (NO
sorbitan and other by-products significantly decreased, thereby
raising the selectivity for the sugar alcohols up to 70% (entry
3
)
2 2
)
4
Cl
2
, entry 2] and to 51% by a Cl-
and the total yield of the sugar alcohols reached 58% with 70%
selectivity and 54 TON of Pt (entry 8). From these results, it is
clearly shown that the reduction of the CrI and the particle size
of cellulose by the ball-milling treatment enhances the reactivity
of cellulose. No further improvement was observed for the 4-
day-milled cellulose (60% yield, entry 11) as suggested in the
XRD measurement shown in Fig. 2.
Accordingly, we conclude that the 2-day-ball-milling is ef-
fective for the pre-treatment of cellulose under our catalytic
conditions, and hereafter this pre-treatment is used as the
standard method.
3
)
2
, entry 3]. In the latter case, 1,4-
3). Ru catalysts also showed the same tendency; Ru(Cl)/Al
prepared from RuCl produced the sugar alcohols in 38% yield
and 40% selectivity (entry 18), whereas Ru(N)/Al provided
2
O
3
3
2
O
3
57% yield and 71% selectivity (entry 19). These results indicate
that Cl-free precursors are suitable for the transformation of
cellulose to sugar alcohols.
We performed EDX analysis of the Pt catalysts to determine
the content of Cl (Fig. S1†). Residual Cl in Pt(Cl)/Al
.1 wt%, but Cl in Pt(N)/Al was less than 0.01 wt%. Fig.
(6b, c) shows the XRD patterns of these catalysts, and both
catalysts showed the diffraction lines of g-Al with no distinct
2
O
3
was
1
2
O
3
Optimisation of catalysts. In the previous report, we used
Pt(Cl)/Al
2
O
3
and Ru(Cl)/Al
2
O
3
catalysts for the hydrogenation of
2
O
3
non-treated cellulose; however, the yields of the sugar alcohols
peaks of Pt, indicating that Pt is highly dispersed. The mean
diameters of Pt particles, determined from CO-uptake, were
10
(
sorbitol 25%, mannitol 6%) need to be improved. It is known
that a small amount of Cl on supported metal catalysts may
affect the hydrogenation activity, which prompted us to study
3.3 nm for Pt(N)/Al
which represents that the structural properties of the catalysts
2
O
3
and 4.5 nm for Pt(Cl)/Al
2
O , respectively,
3
22
This journal is © The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 326–333 | 329

View Online
Fig. 6 XRD patterns of (a) Al
after the hydrogenation of cellulose. Circles (᭹): g-Al
triangles (ꢀ): boehmite.
2
O
3
, (b) Pt(N)/Al
2
O
3
, (c) Pt(Cl)/Al
2
O
3
, and
O ,
2 3
Fig. 7 XRD patterns of (a) BP2000, (b) Pt(N)/BP2000 catalyst before
the hydrogenation of cellulose, and (c) Pt(N)/BP2000 after the reaction.
Circles: Pt metal.
(
d) Pt(N)/Al
2
O
3
Pt(N)/BP2000 catalyst before and after the reaction [Fig. 7(b,c)],
and the mean Pt diameters calculated with Scherrer’s equation
were both 2.6 nm. These results indicate the water-tolerant
property of this catalyst. Other supports such as XC72 and
are analogous. Therefore, a clear difference is the content of Cl,
and the residual Cl may adsorb on the metal surface to cause
the C–C and C–O cleavage reactions by inhibiting the selec-
tive hydrogenation sites and/or promoting the hydrogenolysis
reactions. In addition, since the yield of 1,4-sorbitan, which is
generated by the cyclodehydration of sorbitol, increased with the
Cl-content in the catalyst precursors, another possible function
of Cl is to accelerate acid-catalysed reactions by the formation of
ZrO
2
were less effective than BP2000, and AC and TiO were not
2
suitable as the support (entries 12–17). It is therefore concluded
that BP2000 is a useful support for Pt to produce the sugar
alcohols in water. BP2000-supported Ru, Rh, Ir and Pd catalysts
were prepared from Cl-free precursors, but they were less active
than the Pt catalyst (entries 20–23). This finding led us to choose
Pt(N)/BP2000 as the typical catalyst to study the kinetics of this
reaction.
HCl or AlCl
x
species. In fact, the addition of HCl (Cl 1.1 wt%)
, which corresponded to the same Cl amount as
that contained in Pt(Cl)/Al (entry 1), reduced the selectivity
to Pt(N)/Al
2
O
3
2
O
3
for the sugar alcohols (48%) and increased the formation of by-
products like 1,4-sorbitan (entry 6). It is thus shown that the
use of Cl-free precursors is one of the key factors to gain higher
yields of sugar alcohols.
Mechanistic study using Pt(N)/BP2000 catalyst. Fig. 8(a)
shows the time course of the hydrogenation of cellulose using
Pt(N)/BP2000 catalyst. The “reaction time” denotes the duration
time at 463 K with excluding the heating and the cooling time
The durability of Pt(N)/Al
2
3
O catalyst was tested in reuse
(
Fig. S2†), thereby showing that the reaction slightly proceeded
experiments; the catalyst was reused after drying at 383 K by
adding fresh cellulose of 324 mg, where the yields and the
conversion were calculated based on the fresh cellulose. Total
yields of the sugar alcohols in the reuse experiments were 43%
even at 0 h. The amount of cellulose gradually decreased with the
reaction time and reached 14% at 16 h. Although the conversion
rate of cellulose is not expressed by a simple kinetic equation, we
applied the first-order approximation in eqn (1) to evaluate the
reaction rate, where k is the pseudo first-order rate constant.
Eqn (1) is transformed to an integral form as eqn (2), where X
is the conversion of cellulose. The plot showed good linearity in
(
2nd run, entry 4) and 53% (3rd run, entry 5), which were
23
comparable to that in the first run (51%). The conversion of
cellulose increased from 72% to 101% in the reuses because of
the consumption of remaining cellulose in the previous runs.
Therefore, the selectivity for the sugar alcohols decreased from
C
the first 6 h (7–76% conversion), and k was determined to be
0
C
-
1
.21 h [Fig. 8(b)].
7
0% to 53%, indicating the reduction of the hydrogenation
activity of the catalyst. In the XRD measurement of the
spent catalyst after the first reaction, diffraction patterns of g-
d[Cellulose]/dt = -k
C
[Cellulose]
(1)
(2)
Al
2
O
3
completely disappeared, and sharp peaks of AlO(OH)
ln([Cellulose]/[Cellulose]
0
) = ln(1-X) = -k t
C
(
boehmite) were observed (Fig. 6(d)). The reduction of the
catalytic activity is raised by the hydration of the support,
perhaps causing burial of some Pt particles by the morphology
change.
In focusing on the products, glucose was formed in 6.7% yield
at 0.2 h as a major product [Fig. 8(a) inset], whereas water-
soluble oligosaccharides were negligible (less than 1% yield in
total, data not shown). The yield of glucose decreased to 0.4%
at 16 h, while that of sorbitol smoothly increased after a short
induction period and reached 45% at 16 h. These facts show that
the conversion of cellulose to sorbitol consists of the hydrolysis
of cellulose to glucose via water-soluble oligosaccharides, which
is quickly hydrolysed to glucose, and the hydrogenation of
glucose to sorbitol (Scheme 1 steps A–C). The maximum yield
We carried out the screening of water-tolerant supports such
as carbons, TiO
2
and ZrO (entries 8–10, 12–17). Among these
2
supports, a carbon black BP2000 was found to provide the
highest catalytic activity and good durability; the yields of the
sugar alcohols in the reuse tests were 58%, 64% and 65%,
respectively, and the total TON of Pt was 175 (entries 8–10).
Furthermore, no change was seen for the XRD patterns of
3
30 | Green Chem., 2011, 13, 326–333
This journal is © The Royal Society of Chemistry 2011

View Online
hydrogenated by the Langmuir-Hinshelwood mechanism with
the adsorption of glucose and the dissociative adsorption of
24
H
2
.
Assuming a pseudo first-order reaction (eqn (3)) because
of the sufficiently low concentration of glucose (<4 mM) and the
large excess amount of H , the rate constant (k ) was calculated
to be 2.8 h by the curve fitting using the experimental results
Fig. S3†), which is 13 times higher than that of the hydrolysis
2
G
-
1
(
-
1
step (0.21 h ). From these results, it is concluded that the rate-
determining step is the hydrolysis of cellulose, as expected.
d[Sorbitol]/dt = k
G
[Glucose]
(3)
A small amount of mannitol formed along with sorbitol. We
also investigated the formation route of mannitol by assuming
two possible pathways: (1) the hydrogenation of fructose and/or
mannose and (2) the isomerisation of sorbitol.
Regarding the first hypothesis, small amounts of fructose
(0.4%) and mannose (0.3%) were detected at 0.5 h, which were
generated from glucose by Lobry de Bruyn and van Ekenstein
25
transformation (Scheme 2), but they decreased with reaction
time. In our experiments, the hydrogenation of fructose pro-
ceeded 2.2 times more rapidly than that of mannose, determined
by an initial rate method, and produced sorbitol and mannitol in
a ratio of 1 : 1 by using Pt(N)/BP2000 catalyst (Scheme 1 steps E,
F, Table S1†). As to the second hypothesis, if the isomerisation
of sorbitol to mannitol is the major pathway, the formation rate
of mannitol is correlated with the concentration of sorbitol, but
it is not the case in our reactions as shown in Fig. 8(a). Therefore,
the main intermediate for mannitol formation can be ascribed
to fructose (steps D, F).
Fig. 8 (a) Time course of the hydrogenation of cellulose by
Pt(N)/BP2000 catalyst and (b) the first-order plot of the remaining
cellulose. The inset shows the initial period of 2 h. Reaction conditions:
cellulose 324 mg (1.9 mmol, 2-day-milled), Pt(N)/BP2000 195 mg (Pt
2
0 mg-atom), water 40 mL, P(H ) 5.0 MPa at r.t., T = 463 K. Each plot
2
is the result of each batch reaction with different reaction times.
Scheme 2 Lobry de Bruyn and van Ekenstein transformation of
glucose.
Fig. 9(a) represents the time course of the cellulose degra-
dation with no catalyst under H
2
pressure. The amount of
cellulose smoothly decreased with the reaction time and the
+
unreacted quantity was 29% at 16 h. It is proposed that H
3
O
from water catalyses the hydrolysis of cellulose under these
11a,11c,26
reaction conditions,
because the ionic product of water
Scheme 1 Reaction pathway of the hydrogenation of cellulose using
Pt(N)/BP2000 catalyst.
(pK ) decreases with the temperature from 14.0 (298 K) to 11.3
(463 K) as shown in Fig. S2†. Therefore, higher temperatures
w
27
have an advantage to accelerate the hydrolysis of cellulose in the
+
of glucose was only 6.7%, indicating that the hydrogenation
of glucose is faster than the formation of glucose. To discuss
the reaction rates of these steps quantitatively, we estimated
the hydrogenation rate by using following hypothesis: glucose is
reaction without catalysts in terms of the concentration of H
3
O .
The first-order plot of the remaining cellulose showed a linear
-
1
line with a gradient of 0.082 h [Fig. 9(b)], which was lower than
-
1
the rate constant of the reaction using the Pt catalyst (0.21 h )
This journal is © The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 326–333 | 331

View Online
Table 2 Hydrolysis of cellulose under aerial conditions by
a
Pt(N)/BP2000 catalyst
Yield based on carbon (%)
Cellulose
b
c
Entry
Catalyst
Glucose
Oligosaccharides
conv. (%)
2
2
2
4
5
6
No catalyst
BP2000
Pt(N)/BP2000
4.6
4.7
12.9
14.1
12.3
12.1
24.2
18.5
38.0
a
Reaction conditions; cellulose 324 mg (ball-milled for 4 days), catalyst
13e
5
0 mg, water 40 mL, T = 503 K, reaction time <1 min, under air.
b
c
Dimer-octamer. Based on the loss of weight during the reaction.
Scheme 3 Reaction pathway of the degradation of cellulose without
catalysts.
of various isomerisation and dehydration by-products such as
fructose, mannose, 5-hydroxymethylfurfural (5-HMF), furfural,
levoglucosan and organic acids, whose detail is shown in Table
S2†, because of high chemical reactivity of glucose and the
reducing terminals of the oligosaccharides (steps B¢, G, H).
Among the by-products, 5-HMF was obtained as a major
component after passing 4 h in Fig. 9. No formation of the
hydrogenated products shows that supported metal catalysts are
essential for the reduction reactions.
Conclusions
To achieve the effective production of sorbitol and mannitol
from cellulose, we studied the pre-treatment of cellulose and
developed new catalysts for the hydrogenation of cellulose.
The reduction of the CrI and the particle size of cellulose
by ball-milling treatment for 2 days increased the reactivity
and gave higher yields of the sugar alcohols with keeping
the selectivity. Secondly, we found that Pt(N) supported on
a carbon black BP2000 is a selective and durable catalyst,
compared to the previous Pt catalysts such as Pt(Cl)/Al O .
A key point is the Cl-free catalyst precursors, as residual Cl
on the catalysts derived from precursors induces side-reactions
such as the C–C cleavage and the dehydration. Another point
is the water-tolerant supports which allow catalyst reuse with
good durability. The hydrogenation of cellulose consists of
the hydrolysis of cellulose to glucose via oligosaccharides and
hydrogenation of glucose to sorbitol. The kinetic constant of
each reaction was determined, which demonstrated that the Pt
catalyst promotes both the hydrolysis and the hydrogenation
steps and that the rate-determining step is the hydrolysis step.
2
Fig. 9 (a) Time course of the degradation of cellulose under H pressure
by no catalysts and (b) the first-order plot of the remaining cellulose.
Reaction conditions: cellulose 324 mg (1.9 mmol, 4-day-milled), water
4
0 mL, P(H ) 5.0 MPa at r.t., T = 463 K. Each plot is the result of each
2
batch reaction with different reaction times.
in the conversion range of 10–70% (Fig. 8). These results suggest
that Pt(N)/BP2000 enhances the hydrolysis of cellulose; i.e., the
supported Pt catalyst could hydrolyse the glycosidic bonds, and
other possibility might be that Pt generates protons by the
2
3
10,11c
heterolysis of H
2
.
As shown in Table 2, Pt(N)/BP2000 catalyst
gave a higher yield of glucose (13%) than BP2000 (5% yield)
and no catalyst (5% yield) in the hydrolysis of cellulose without
H pressure under a rapid heating and cooling conditions at
2
28
5
03 K. This result also indicates that Pt(N)/BP2000 itself has
the hydrolysis activity.
The reaction mixture without catalysts turned brownish with
the reaction time, and black solids precipitated at ≥16 h, which is
in contrast to the colourless solution in the reactions using the Pt
catalyst. In the HPLC analysis of the solution, oligosaccharides
were found as the principal products in the initial stage [Fig. 9(a),
Scheme 3 Step A¢], which is different from the reaction using
the Pt catalyst. This result also supports that the Pt catalyst
hydrolyses the glycosidic bonds. After the formation of the
oligosaccharides, glucose was generated with large amounts
Acknowledgements
This work was supported by Grant-in-Aid for Scientific Re-
search (KAKENHI, 20226016) from the Japan Society for the
Promotion of Science (JSPS).
3
32 | Green Chem., 2011, 13, 326–333
This journal is © The Royal Society of Chemistry 2011

View Online
Notes and references
12793; (b) D. Yamaguchi, M. Kitano, S. Suganuma, K. Nakajima,
H. Kato and M. Hara, J. Phys. Chem. C, 2009, 113, 3181–3188; (c) A.
Onda, T. Ochi and K. Yanagisawa, Green Chem., 2008, 10, 1033–
1037; (d) R. Rinaldi, R. Palkovits and F. Sch u¨ th, Angew. Chem.,
Int. Ed., 2008, 47, 8047–8050; (e) H. Kobayashi, T. Komanoya, K.
Hara and A. Fukuoka, ChemSusChem, 2010, 3, 440–443; (f) S. Van,
de Vyver, L. Peng, J. Geboers, H. Schepers, F. de Clippel, C. J.
Gommes, B. Goderis, P. A. Jacobs and B. F. Sels, Green Chem.,
2010, 12, 1560–1563; (g) K. Matsumoto, H. Kobayashi, K. Ikeda, T.
Komanoya, A. Fukuoka and S. Taguchi, Bioresour. Technol., 2010,
DOI: 10.1016/j.biortech.2010.09.098.
1
G. W. Huber, S. Iborra and A. Corma, Chem. Rev., 2006, 106, 4044–
4
2
098; A. Corma, S. Iborra and A. Velty, Chem. Rev., 2007, 107,
411–2502.
2
3
B. Kamm, P. R. Gruber and M. Kamm, Biorefineries - Industrial
Processes and Products, Wiley-VCH, Weinheim, 2006.
J. N. Chheda, G. W. Huber and J. A. Dumesic, Angew. Chem., Int.
Ed., 2007, 46, 7164–7183.
4
5
Y. Sun and J. Cheng, Bioresour. Technol., 2002, 83, 1–11.
D. Klemm, B. Heublein, H.-P. Fink and A. Bohn, Angew. Chem., Int.
Ed., 2005, 44, 3358–3393.
14 N. Yan, C. Zhao, P. J. Dyson, C. Wang, L.-T. Liu and Y. Kou, Chem-
SusChem, 2008, 1, 626–629; C. Zhao, Y. Kou, A. A. Lemonidou,
X. Li and J. A. Lercher, Angew. Chem., Int. Ed., 2009, 48, 3987–
3990; J. Zakzeski, P. C. A. Bruijnincx, A. L. Jongerius and B. M.
Weckhuysen, Chem. Rev., 2010, 110, 3552–3599.
15 H. Zhao, J. H. Kwak, Y. Wang, J. A. Franz, J. M. White and J. E.
Holladay, Energy Fuels, 2006, 20, 807–811.
16 E. Kontturi and T. Vuorinen, Cellulose, 2009, 16, 65–74.
17 K. Mazeau and L. Heux, J. Phys. Chem. B, 2003, 107, 2394–
2403.
18 M. Hall, P. Bansal, J. H. Lee, M. J. Realff and A. S. Bommarius,
FEBS J., 2010, 277, 1571–1582.
19 S. D. Mansfield, C. Mooney and J. N. Saddler, Biotechnol. Prog.,
1999, 15, 804–816.
20 L. Segal, J. J. Creely, A. E. Martin and C. M. Conrad, Text. Res. J.,
1959, 29, 786–794.
21 H. Kono, S. Yunoki, T. Shikano, M. Fujiwara, T. Erata and M. Takai,
J. Am. Chem. Soc., 2002, 124, 7506–7511.
22 T. Nakamura, M. Ohshima, H. Kurokawa and H. Miura, Chem.
Lett., 2010, 39, 62–63.
6
7
S. Deguchi, K. Tsujii and K. Horikoshi, Chem. Commun., 2006,
3
293–3295.
(a) Y.-H. P. Zhang and L. R. Lynd, Biotechnol. Bioeng., 2004, 88,
7
97–824; (b) Y.-H. P. Zhang, J. Cui, L. R. Lynd and L. R. Kuang,
Biomacromolecules, 2006, 7, 644–648.
8
R. Rinaldi and F. Sch u¨ th, ChemSusChem, 2009, 2, 1096–1107; C.
Li and Z. K. Zhao, Adv. Synth. Catal., 2007, 349, 1847–1850; K.
Shimizu, H. Furukawa, N. Kobayashi, Y. Itaya and A. Satsuma,
Green Chem., 2009, 11, 1627–1632.
M. Sasaki, Z. Fang, Y. Fukushima, T. Adschiri and K. Arai, Ind.
Eng. Chem. Res., 2000, 39, 2883–2890.
9
1
1
0 A. Fukuoka and P. L. Dhepe, Angew. Chem., Int. Ed., 2006, 45,
5
9
161–5163; P. L. Dhepe and A. Fukuoka, ChemSusChem, 2008, 1,
69–975.
1 (a) C. Luo, S. Wang and H. Liu, Angew. Chem., Int. Ed., 2007,
4
6, 7636–7639; (b) W. Deng, X. Tan, W. Fang, Q. Zhang and Y.
Wang, Catal. Lett., 2009, 133, 167–174; (c) V. Jollet, F. Chambon,
F. Rataboul, A. Cabiac, C. Pinel, E. Guillon and N. Essayem, Green
Chem., 2009, 11, 2052–2060; (d) S. Van, de Vyver, J. Geboers, M.
Dusselier, H. Schepers, T. Vosch, L. Zhang, G. Van Tendeloo, P. A.
Jacobs and B. F. Sels, ChemSusChem, 2010, 3, 698–701; (e) L. Ding,
A. Wang, M. Zheng and T. Zhang, ChemSusChem, 2010, 3, 818–821.
2 N. Ji, T. Zhang, M. Zheng, A. Wang, H. Wang, X. Wang and J. G.
Chen, Angew. Chem., Int. Ed., 2008, 47, 8510–8513; M. Zheng, A.
Wang, N. Ji, J. Pang, X. Wang and T. Zhang, ChemSusChem, 2010,
23 N. S. Mosier, C. M. Ladisch and M. R. Ladisch, Biotechnol. Bioeng.,
2002, 79, 610–618.
24 E. Crezee, B. W. Hoffer, R. J. Berger, M. Makkee, F. Kapteijn and
J. A. Moulijn, Appl. Catal., A, 2003, 251, 1–17.
25 J. C. Speck Jr., Adv. Carbohydr. Chem., 1958, 13, 63–103.
26 T. Minowa, F. Zhen and T. Ogi, J. Supercrit. Fluids, 1998, 13, 253–
259.
1
1
3
4
, 63–66; Y. Zhang, A. Wang and T. Zhang, Chem. Commun., 2010,
6, 862–864.
27 W. L. Marshall and E. U. Franck, J. Phys. Chem. Ref. Data, 1981,
3 (a) S. Suganuma, K. Nakajima, M. Kitano, D. Yamaguchi, H. Kato,
S. Hayashi and M. Hara, J. Am. Chem. Soc., 2008, 130, 12787–
10, 295–304.
28 The detail of the hydrolysis method is described in reference 13e.
This journal is © The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 326–333 | 333