Chemically modified cyclodextrins as supramolecular tools to generate carbon-supported ruthenium nanoparticles: An application towards gas phase hydrogenation

Article abstract of DOI:10.1016/j.apcata.2010.07.006

A series of carbon-supported ruthenium catalysts was synthesized from zerovalent ruthenium nanoparticles stabilized by randomly methylated cyclodextrins (α-, β- and γ-CD) followed by their adsorption onto the carbon support. The catalysts were characterized by N₂ physisorption and thermal analyses. The deposited ruthenium nanoparticles were characterized by transmission electron microscopy, which has highlighted predominantly spherical shapes with a mean diameter of 2.4 nm. Catalytic activity was investigated in the gas phase hydrogenation of o-, m- and p-xylene at 85 °C, both separately and in a two-component mixture (o- and p-xylene). The catalyst prepared by a 1:3 concentration ratio of RuCl₃ to randomly methylated β-cyclodextrin exhibited the highest hydrogenation activity and stereoselectivity toward the formation of trans- dimethylcyclohexane. The β-cyclodextrin appeared as multifunctional molecular receptors enabling the stabilization and dispersion of the metallic nanoparticles onto the support and the promotion of the catalytic reaction through host-guest interactions.

Full text of DOI:10.1016/j.apcata.2010.07.006

Applied Catalysis A: General 391 (2011) 334–341
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Chemically modified cyclodextrins as supramolecular tools to generate
carbon-supported ruthenium nanoparticles: An application towards gas
phase hydrogenation
F. Wyrwalskia,b,d, B. Légera,b,d, C. Lancelota,c,d, A. Roucouxe,f, E. Monfliera,b,d, A. Ponchela,b,d,∗
a Univ Lille Nord de France, F-59000 Lille, France
b UArtois, UCCS, F-62300 Lens, France
c USTL, UCCS, F-59650 Villeneuve d’Ascq, France
d CNRS, UMR 8181, France
e Ecole Nationale Supérieure de Chimie de Rennes, CNRS, UMR 6226, Avenue du General Leclerc, CS 50837, 35708 Rennes Cedex 7, France
f
Université Européenne de Bretagne, France
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of carbon-supported ruthenium catalysts was synthesized from zerovalent ruthenium nanopar-
ticles stabilized by randomly methylated cyclodextrins (␣-, ␤- and ␥-CD) followed by their adsorption
onto the carbon support. The catalysts were characterized by N2 physisorption and thermal analyses.
The deposited ruthenium nanoparticles were characterized by transmission electron microscopy, which
has highlighted predominantly spherical shapes with a mean diameter of 2.4 nm. Catalytic activity was
investigated in the gas phase hydrogenation of o-, m- and p-xylene at 85 ◦C, both separately and in a
two-component mixture (o- and p-xylene). The catalyst prepared by a 1:3 concentration ratio of RuCl3 to
randomly methylated ␤-cyclodextrin exhibited the highest hydrogenation activity and stereoselectivity
toward the formation of trans-dimethylcyclohexane. The ␤-cyclodextrin appeared as multifunctional
molecular receptors enabling the stabilization and dispersion of the metallic nanoparticles onto the
support and the promotion of the catalytic reaction through host–guest interactions.
Received 14 March 2010
Received in revised form 20 May 2010
Accepted 5 July 2010
Available online 15 July 2010
Keywords:
Cyclodextrins
Metallic nanoparticles
Activated carbon
Hydrogenation
Gas phase reactions
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
and macro-pore ranges. Further interesting aspects of the use of
carbon supports in catalysis over inorganic oxides are related to
The development of efficient heterogeneous catalytic systems
involving transition metal catalysts constitutes an active area of
both academic and industrial research. Among heterogeneous cata-
lysts, carbon-supported metal catalysts are widely used in practical
processes and the increasing importance of carbon materials in
catalytic processes is now widely recognized [1,2]. Their uses as
supports are directly related to their wide availabilities and supe-
rior physical and chemical properties, including electrical and
thermal conductivity, hydrogen storage ability, chemical stabil-
ity, high surface area, low density and easy recovery of the metal
phase by burning off the spent catalyst. Thus, conventional car-
bons such as activated carbon can be synthesized by pyrolysis and
physical or chemical activation of organic precursors, such as coal,
wood, fruit shell or polymer at elevated temperature. These carbons
present relatively broad pore size distributions in micro-, meso-
the fact that their surface favors the metallic reduced state and
decreases the coking propensity. However, the inert nature of the
carbon surface makes sometimes the metal deposit difficult, lead-
ing to mediocre dispersions of the metallic phase without any
narrow size distribution [3].
To overcome these difficulties, carbon-supported nanoparticles
prepared by the deposition of metal nanoparticles onto carbons
have received an increasing attention in recent years [4–11]. Thus,
narrow dispersed metal nanoparticles have been successfully stabi-
lized on carbon supports and have often revealed superior catalytic
behaviors to conventional supported systems due to their large
surface areas available for catalysis. A further interesting aspect of
the use of carbon-supported nanoparticles lies in the fact that the
carbon pores could act as a physical barrier that prevents the coales-
cence of nanoparticles and limits the phenomenon of aggregation,
thus generating robust and reusable catalysts, as recently high-
lighted in the Suzuki–Miyaura [4] and nitrobenzene hydrogenation
[6] reactions carried out in liquid media.
∗
Corresponding author at: UArtois, UCCS, F-62300 Lens, France.
Fax: +33 3 2179 1755.
For the rational design of supported metal nanoparticles, the
choice of the metal precursors and capping materials are key
E-mail address: anne.ponchel@univ-artois.fr (A. Ponchel).
0926-860X/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2010.07.006

F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
335
Scheme 1. Chemical representations of the native cyclodextrins.
parameters to control the size and shape of the nanoparticles and
obtain reproducible catalytic activities. Thus, the stabilization of
colloidal metallic suspensions often requires the addition of a pro-
tective agent (e.g., polymers, oligomers or surfactants) to prevent
the aggregation of the colloids into larger particles. Several articles
have offered highly detailed reviews of specific synthesis tech-
niques for preparing nanoparticles by chemical methods [12–16].
However, concerning the preparation of the carbon-supported
nanocatalysts, the efficiency of the approach needs to take into
account, not only the stability of the metal colloids but also the
metal-support interactions that will exert a dominant effect during
the deposition step.
Among the various protective agents of metallic nanoparti-
cles reported in the literature, biomolecular templates such as
cyclodextrins (CDs) provide an excellent potential tool for bottom-
up device fabrication. CDs are water-soluble cyclic oligosaccharides
formed of 6 (␣), 7 (␤) or 8 (␥) glucopyranose units (Scheme 1),
exhibiting a hydrophobic internal cavity that can host a large vari-
ety of organic molecules of appropriate size and shape, such as
aromatic compounds. In the presence of active metallic species, this
property has been widely exploited in the fields of homogeneous
catalysis as artificial enzymes [17,18] and biphasic catalysis as mass
transfer promoters [19,20]. In recent years, they have appeared
to be ideal multifunctional agents for the production of original
nanometer-sized metal catalysts. Since the pioneering works of
Komiyama and Hirai [21], it has been widely demonstrated that
the CDs could stabilize active metal nanoparticles for the liquid
phase catalysis. These authors reported that refluxing of an aque-
ous solution of rhodium(III) and ␣-CD or ␤-CD, followed by further
refluxing in the presence of ethanol gave a colloidal dispersion of
rhodium particles that could effectively catalyze the hydrogena-
tion of olefins. In another approach, Kaifer et al. have extensively
described that per-6-thio-␤-CD can effectively stabilized platinum
[22], palladium [22–24], silver [25] and gold [26] nanoparticles
through multiple sulfur bonds between the metal core and the
oligosaccharide. Interestingly, it has been found that these per-6-
In an extension to these works, Roucoux and Monflier have
recently demonstrated that the randomly methylated CDs (RaMe-
CDs) can efficiently stabilize ruthenium nanoparticles in water and
that these CD-capped nanoparticles could catalyze the hydrogena-
tion of ␣- or ␤-pinene and monofunctionalized arene derivatives
under biphasic conditions [29,30]. Interesting chemoselectivities
have been observed according to the relevant choice of cavity and
methylation degree of the CD. Recently, we have also found that
the RaMe-CDs could also be used to synthesize heterogeneous
nanocatalysts. The synthesis strategy was based on the sponta-
neous adsorption of CD-capped ruthenium nanoparticles thanks to
the strong affinity of CDs towards the carbon supports [31]. Notably,
these carbon-supported Ru(0) catalysts allowed the hydrogenation
of aromatic substrates in gas phase under mild reaction conditions.
Starting from these observations, we report here the prepa-
ration and the optimization of the supported catalytic systems,
based on the ability of RaMe-CDs to stabilize Ru(0) nanoparti-
cles and interact with carbons (Scheme 2). The influence of the
size and initial amount of RaMe-CD on the adsorption of CD-
capped Ru(0) nanoparticles on the support has been studied by BET
measurements and thermal analyses. The particle size and shape
have been characterized by transmission electron microscopy. The
catalytic performances of these original nanocatalysts containing
supramolecular entities have been evaluated in the hydrogenation
of o-, m- and p-xylene, both separately and in a two-component
mixture (o- and p-xylene).
2. Experimental
2.1. General
The starting activated carbon, i.e. Nuchar®WV (denoted AC-WV)
was supplied by MeadWestvaco Corporation (Covington, USA),
produced from wood and activated by phosphorous acid. Ran-
domly methylated ␤-cyclodextrin (RaMe-␤-CD) was purchased
from Aldrich Chemicals. Randomly methylated ␣-cyclodextrin
(RaMe-␣-CD) and randomly methylated ␥-cyclodextrin (RaMe-␥-
CD) were prepared by adapting a procedure reported by Kenichi
thio-␤-CDs were effective catalysts for the hydrogenation of C
C
or C N double bonds [23,27] and C–C coupling reactions [24,28].
Scheme 2. Preparation of the carbon-supported Ru(0) nanoparticles stabilized by randomly methylated cyclodextrins.

336
F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
Table 1
Textural characteristics of the starting carbon support and Boehm titrations.
a
b
c
f
SBET (m2 g−1
)
Vtot (cm3 g−1
)
Vmicro (cm3 g−1
)
% microd
dP (nm)e
Titration results (␮mol g−1
)
Carboxylic
4
Lactonic
11
Phenolic
19
Acidic
34
Basic
5
Total
39
1570
1.25
0.54
43
3.3
a
The specific area is calculated from the BET equation using P/P0 values in the 2.5 × 10−3 and 1.7 × 10−1 range.
The total pore volume (Vtot) is evaluated at P/P0 = 0.95.
The micropore volume (Vmicro) is determined by the Dubinin–Radushkevich (DR) plot of the N2 adsorption isotherm.
The relative microporosity percentage is defined as the ratio of the micropore volume to the total pore volume.
The average pore size is calculated from the ratio of the total pore volume to the BET surface area by using the Gurvich rule assuming that the pore is of cylindrical
b
c
d
e
geometry [33].
f
The determination of the surface composition of AC-WV is determined by the Boehm’s titrations [34].
et al. [32]. These cyclodextrins were partially methylated. Methy-
lation occurred at positions 2, 3, or 6 and 1.8 OH groups per
glucopyranose unit were statistically modified. Ruthenium(III)
chloride hydrate purum and organic compounds were purchased
from Aldrich Chemical Co. and Acros Organics in their highest purity
and used without further purification. Double distilled water was
used in all experiments.
2.3.3. Transmission electron microscopy
Transmission electron microscopy (TEM) was performed on a
Tecnai microscope (200 kV). Before the analysis, the solid powder
was ultrasonically dispersed in ethanol and two drops of ethanol
containing the solid were deposited on a carbon coated copper grid.
The size distribution histogram was obtained with the program
SCION Image from the measurement of about 700 particles.
2.4. Catalytic procedure
2.2. Catalyst synthesis
o-, m- and p-xylene were purchased from Aldrich (99.99%) and
gaseous hydrogen from Air Liquide (99.995%). Hydrogenation activ-
ities were measured at 85 ◦C, without additional activation steps,
in a fixed-bed flow reactor at atmospheric pressure. The hydro-
gen feed was saturated with vapors (1 kPa) of pure component (o-,
m- or p-xylene at 27, 23 and 22 ◦C, respectively) or binary mix-
ture (o- and p-xylene at 25 ◦C) with a ratio 0.42:0.58. The total flow
rate of the reaction mixture (H2/xylene = 100) was 60 mL min−1. In
order to keep the conversions low (<30%), small amounts of cat-
alyst in the powder form are used, i.e. 25 or 50 mg diluted with
1 mL of inert carborundum particles of average size of 100 ␮m. The
reaction products were analyzed by gas chromatography (Perkin-
Elmer Clarus 500) equipped with a 5% diphenyl 95% dimethyl
silicon capillary column (25 m × 0.25 mm) and a FID detector. Cat-
alytic activities were evaluated in the steady-state, obtained after
ca. 10 min. Each catalytic run was performed in duplicate and the
reported results are the average between the two runs.
The carbon-supported ruthenium nanocatalysts were prepared
by adsorption of CDs-stabilized ruthenium nanoparticles. The pro-
cedure is described as follows. In an aqueous solution, 90 mL of
RuCl3·xH2O (76 mg, 0.3 mmol) was added to 60 mL of an aqueous
solution containing x equiv. of RaMe-␣-, RaMe-␤- or RaMe-␥-CD
(x × 0.3 mmol). The solution was stirred for 15 min. Sodium boro-
hydride (0.75 mmol) in 50 mL of double distilled was added all at
once under vigorous stirring. The solution was kept under agita-
tion for 2 h at room temperature. Then, the AC-WV carbon (1 g),
previously ground to powder in a planetary ball mill (model Retsch
S100) at 500 rpm for 1 h, was added to the Ru(0) aqueous suspen-
sion and stirred for 2 supplementary hours. Finally, the catalyst
was filtered, washed thoroughly with water (3 × 40 mL) and dried
at 100 ◦C for 48 h. The solids are designated as Ru-x-CD/C where x
is the initial molar ratio of RaMeCD to RuCl3 used to stabilize the
colloidal suspension. The ruthenium loadings were determined by
elemental analyses at the Vernaison CNRS center and can be aver-
aged to 1.4 0.2 wt.%. It is also important to note that a reference
Ru/C catalyst has been synthesized in the same conditions, except
that the aqueous solution of cyclodextrins was replaced by pure
water.
3. Results and discussion
3.1. Characterization of the AC-WV support
2.3. Characterization techniques
Prior to studying the ability of AC-WV to adsorb CD-capped Ru
nanoparticles, the porous texture of the starting carbon material
was characterized by nitrogen sorption measurements (Table 1).
From these results, it appears that AC-WV is a good candidate to
support the ruthenium nanoparticles since it develops a combined
micro- and meso-porosity, high surface area (1570 m2 g−1) and
large average pore size (3.3 nm) [33]. The Boehm titration method
was applied to determine the surface chemistry of the carbon mate-
rial [34]. The main results are gathered in Table 1. Thus, we can see
that the AC-WV carbon presents a variety of species mainly acid,
likely due to its origin and method of activation by phosphoric acid.
However, the total amount of sites remains low, i.e. 39 ␮mol g−1
corresponding to 0.02 function per nm2, consistent with a carbon
surface almost free of oxygen functional groups [35].
2.3.1. Specific surface area measurements
Porosity measurements were obtained from nitrogen adsorp-
tion/desorption isotherms at T = −196 ◦C using
a Nova 2200
apparatus from Quantachrom Corporation, carried out on pow-
ders previously outgassed overnight at 100 ◦C. Specific areas
were calculated from the BET equation using P/P0 values in the
2.5 × 10−3 and 1.7 × 10−1 range. The total pore volume (Vtot) was
estimated at P/P0 = 0.95 and micropore volume (Vmicro) by the
Dubinin–Radushkevich (DR) plots of the N2 adsorption isotherm.
The average pore size was estimated from the ratio of the total
pore volume to the BET surface area by using the Gurvich rule and
assuming that the pore is of cylindrical geometry [33].
2.3.2. Thermal analyses
3.2. Characterization of the carbon-supported ruthenium
catalysts
Differential thermal and thermogravimetric analyses (DTA
and TGA) were performed using a SDT 2960 analyzer from TA
Instrument. All analyses were conducted under nitrogen flow
(75 mL/min) with a heating rate of 2 ◦C/min from room temperature
to 400 ◦C.
3.2.1. BET measurements
Fig. 1 reports the BET specific area measured on the series of
catalysts prepared by adsorption of aqueous colloidal solutions

F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
337
number of glucose units of the CD and size of the carbon pores
[31].
Further evidence of the adsorption of RaMe-CD-stabilized
ruthenium nanoparticles onto the support was obtained by carry-
ing out a thermal treatment at 400 ◦C under N2 of the Ru-3␤-CD/C
solid. Interestingly, such a treatment allows to thermally decom-
pose the oligosaccharide molecules adsorbed in the pores and
recover, almost completely, the initial surface (1524 vs 1548 m2 g−1
for the Ru/C).
3.2.2. Thermal analyses
Fig. 2 shows the thermogravimetric and differential thermal
analysis curves of the catalyst series prepared by adsorption of
metallic nanoparticles stabilized by RaMe-CD (␣-, ␤- or ␥-) and
adsorbed onto AC-WV. Whatever the nature and amount of CD,
the TGA curves reveal important weight losses in the tempera-
ture range 250–350 ◦C, corresponding to the thermal degradation
of the organic capping agent [36]. Moreover, the examination of
the TDA signals shows that the temperature, for which the rate of
decomposition is the most significant, slightly varies as a function
of the number of glucose units in the oligosaccharide molecule,
from 299 ◦C for the RaMe-␣-CD (n = 6) to 287 ◦C for the RaMe-␥-CD
(n = 8). This size effect on the CD decomposition is in agreement
with the study of Camino et al. [36] dealing with the thermal
degradation of native ␣-, ␤- and ␥-CDs and mono-substituted ␤-
cyclodextrins. Thus, it was demonstrated by these authors that the
thermal decomposition of cyclodextrins was dependent on the sub-
stituent and cavity size.
Furthermore, these TGA experiments allow also the determi-
nation of the amount of RaMe-CD, which is still anchored onto
the metal and carbon surfaces after the deposition of the ruthe-
nium nanoparticles and their washing with water (Table 2). With
increasing initial molar ratio CD:Ru an increase in the amount of
adsorbed RaMe-␤-CD in the final catalyst is observed, with the
Ru-1.5␤-CD/C sample achieving 114 ␮mol g−1 of CD whereas the
Ru-5␤-CD/C sample reaches 213 ␮mol g−1. Note that this last value
seems to indicate that the coverage of the carbon support by RaMe-
␤-CD is approaching the saturation since a further twofold increase
in the initial amount of CD results in only a small increase in
the number of adsorbed RaMe-␤-CD, i.e. 236 ␮mol g−1. Further-
more, at a constant molar ratio CD:Ru of 3, we can see that the
amount of adsorbed organic species of the Ru-CD/C samples sig-
nificantly increases when the size of the oligosaccharide decreases.
This behavior can be explained by the effect of the pore structure of
the AC-WV material on the adsorption rate of methylated cyclodex-
trins and nanoparticles, in line with what has been demonstrated
in the specific surface area measurements.
Fig. 1. BET specific surface areas measured on the ruthenium carbon-supported
catalysts. aThe Ru-3␤-CD/C(400) sample was obtained by a thermal treatment of
the Ru-3␤-CD/C under N2 (60 mL min−1) at 400 ◦C for 4 h.
of zerovalent ruthenium. In comparison with the value obtained
with the starting AC-WV, no significant decrease in the specific
surface area is measured with the reference Ru/C catalyst, pre-
pared by adsorption of Ru(0) after the chemical reduction of the
ruthenium salt with sodium borohydride in water. On the other
hand, it is clearly shown that the textural properties of the sup-
port are greatly affected after the adsorption of Ru(0) nanoparticles
protected by randomly methylated cyclodextrin; thus, the higher
the initial amount of RaMe-␤-CD, the lower the specific surface
area of the resulting catalyst. For instance, the Ru-10-␤-CD/C sam-
ple shows the lowest value of the series (i.e. 715 vs 1570 and
1548 m2 g−1 for the AC-WV and Ru/C, respectively). The decrease
in the specific surface area is ascribed to the adsorption of the
CD-stabilized nanoparticles on the support, and this phenomenon
preferentially occurs in the mesopores and on the external sur-
face of the support by blocking the entrance of the micropores
(≤2 nm). This assumption can be connected to the size of the CDs
(i.e. 1.46, 1.54 and 1.75 nm for the native ␣-, ␤- and ␥-CD, respec-
tively), which is of the same order of magnitude than the size of
micropores (dp < 2 nm). Furthermore, it should be emphasized that
the specific surface areas depend on the nature of the CDs used to
stabilize the Ru(0) nanoparticles, according to: Ru-3␣-CD/C < Ru-
3␤-CD/C < Ru-3␥-CD/C. This tendency indicates that the adsorption
process seems to be governed by a size-exclusion mechanism,
which may limit the diffusion and adsorption of the oligosaccha-
rides in the microporous structure of AC-WV by steric hindrance.
These results concur with those already obtained by Abe et al. who
showed that the activated carbons can adsorb native and modi-
fied cyclodextrins and that this adsorption was dependent on the
Finally, thermogravimetric analyses also prove that the sup-
ported nanocatalysts are thermally stable up to 250 ◦C under
nitrogen atmosphere, and also under hydrogen atmosphere, which
is of fundamental importance for catalysis. In fact, we have found
Table 2
Characteristics of the carbon-supported ruthenium nanoparticles deduced from the thermal analysis.
b
Entry
Sample
Ru (wt.%)
TM (◦C)a
Weight loss (%)
NCD (␮mol g−1
)
1
2
3
4
5
6
7
8
9
C (AC-WV)
Ru/C
–
–
–
0.4
1.0
20
24
22
15
28
31
25
–
–
1.8
1.2
1.2
1.2
1.2
1.5
1.4
–
Ru-3␥-CD/C
Ru-3␣-CD/C
Ru-3␤-CD/C
Ru-1.5␤-CD/C
Ru-5␤-CD/C
Ru-10␤-CD/C
Ru-3␤-CD/Cc
287
299
295
296
297
296
284
128
210
167
114
213
236
–
a
Temperature corresponding to the maximum of the differential thermal analysis curve.
Calculations with M = 1140, 1315 and 1562 g mol−1 for RaMe-␣-CD, RaMe-␤-CD and RaMe-␥-CD, respectively.
Thermal decomposition carried out under H2 atmosphere instead of N2.
b
c

338
F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
Fig. 2. DTA and TGA curves obtained under N2 for Ru-3␥-CD/C (a), Ru-3␤-CD/C (b), Ru-3␣-CD/C (c), Ru-1.5␤-CD/C (d), Ru-5␤-CD/C (e) and Ru-10␤-CD/C (f).
that the thermal behavior of the Ru-3␤-CD/C catalyst under H2
were similar to that measured under N2, as evidenced by the weight
loss and TM value (entry 9, Table 2).
The successful adsorption of the Ru(0) nanoparticles onto the
support is clearly supported by the TEM analyses as they show
the presence of monodispersed CD-capped Ru nanoparticles cov-
ering the AC-WV support (Fig. 3). As illustrated by the TEM
images presented in Fig. 3a, we can observe mainly isolated Ru-CD
nanoparticles randomly distributed on the AC-WV support even if
some aggregates are also present, probably due to random assem-
blies of nanoparticles at the surface during the preparation (Fig. 3b).
However, it is worth noting that the Ru nanoparticles are near
spherical (Fig. 3c) and that, notably, these spheres have similar
morphological characteristics compared to those of colloidal CD-
stabilized Ru nanoparticles grown by the same procedure but in
the absence of any carbon support. As shown in Fig. 4, the sup-
ported nanoparticles display a Gaussian-like distribution (solid
line, Fig. 4) and thus a mean diameter of 2.4 nm with 90% between
1.5 and 3 nm is obtained. These results are in line with those
obtained in the absence of carbon support and reported by our
group, showing Ru(0) nanoparticles in the size range of 2.5 nm with
80% between 2 and 3 nm [30]. Finally, as the average particle size of
3.2.3. Transmission electron microscopy (TEM)
In order to get further information on the shape, size and dis-
persion of the ruthenium nanoparticles stabilized by methylated
cyclodextrins after the deposition step onto the support, TEM char-
acterization has been performed. As an illustrative example, Fig. 3
shows the TEM pictures of the Ru-3␤-CD/C sample for which the
Ru(0) nanoparticles were synthesized with a 1:3 ratio of Ru to
RaMe-␤-CD followed by adsorption onto the AC-WV. The metal
particle size distribution has been estimated from the measure-
ment of ca. 700 particles found in arbitrarily chosen area of the
microphotographs and is given in Fig. 4. It is worth adding here
that the measurement is only related to the size of the metal
core. In fact, TEM analysis provides direct information about the
metal particles without any visualization of the oligosaccharides
(stabilizers).

F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
339
Fig. 3. Transmission electron microscopy images of the Ru-3␤-CD/C catalyst synthesized from ruthenium nanoparticles with a concentration ratio of 1:3 ruthenium salt to
RaMe-␤-CD and adsorbed onto the carbon support.
3.3. Catalytic hydrogenation
3.3.1. Hydrogenation of xylenes
As the previous results have provided evidence of the adsorption
of the ruthenium particles protected by RaMe-CDs, their catalytic
behavior has been investigated in the hydrogenation of xylene
(o-, m- and p-xylene) in gas phase at 85 ◦C. Table 3 reports the
steady-state catalytic activities, expressed as turn-over frequen-
cies (TOFs) and selectivities. The reaction products are the cis- and
trans-dimethylcyclohexane (DMCH).
When comparing the steady-state TOFs reported in Table 3,
carbon-supported Ru(0) nanocatalysts stabilized by randomly
cyclodextrins (␣, ␤ and ␥) appear to be more efficient than the
reference Ru/C catalyst whatever the substrate. However, catalytic
activity is found to be dependent on the nature of the methylated CD
used, in the following order RaMe-␤-CD > RaMe-␣-CD > RaMe-␥-
CD. Thus, it is worth emphasizing that the Ru-3␤-CD/C solid is by far
the most active and that the TOF value is 1.5–17 times higher than
values reported in the literature for gas phase hydrogenation of o-
xylene at 100 ◦C over ruthenium catalysts [37]. The reproducibility
of this maximum has been confirmed by using two different Ru-
3␤-CD/C samples obtained from two synthesis batches (entries 4
and 5, Table 3). With further increasing initial molar ratio CD:Ru
a decrease in activity is observed and this behavior could indicate
that the active metal surface is less accessible for the gaseous reac-
tants when excess of RaMe-␤-CD is used. This assumption concurs
with our recent results obtained in aqueous phase catalysis, which
Fig. 4. Size distribution of the Ru-3␤-CD/C catalyst synthesized from ruthenium
nanoparticles with a concentration ratio of 1:3 ruthenium salt to RaMe-␤-CD and
adsorbed onto the carbon support.
the supported ruthenium nanoparticles is smaller than the average
carbon pore aperture (3.1 nm), it can be concluded that the Ru(0)
nanoparticles are not located inside the micropores of AC-WV but
rather in the mesopores and on the external surface of the carbon
particles.
Table 3
TOFs and selectivities in xylene hydrogenation at 85 ◦C measured on the ruthenium catalysts.
a
Entry
Sub.
Sample
TOF × 103 (s−1
)
Sel. trans (%)
Sel. cis (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
o-xyl.
o-xyl.
o-xyl.
o-xyl.
o-xyl.
o-xyl.
o-xyl.
o-xyl.
o-xyl.
m-xyl.
m-xyl.
m-xyl.
m-xyl.
p-xyl.
p-xyl.
p-xyl.
p-xyl.
Ru/C
4.2
5.7
8.6
14.4
14.9
8.6
7.4
9.9
5.0
6.6
11.3
20.5
30.5
6.7
10.8
20.0
29.8
1.9
2.1
1.8
3.9
3.9
3.4
2.6
2.8
2.1
7.5
8.3
8.8
11.4
15.0
15.3
16.7
21.9
98.1
97.9
98.2
96.1
96.1
96.6
97.4
97.2
97.9
92.5
91.7
91.2
88.6
85.0
84.7
83.3
78.1
Ru-3␥-CD/C
Ru-3␣-CD/C
Ru-3␤-CD/Cb
Ru-3␤-CD/Cc
Ru-1.5␤-CD/C
Ru-5␤-CD/C
Ru-10␤-CD/C
Ru-3␤-CD/C (400)d
Ru/C
Ru-3␥-CD/C
Ru-3␣-CD/C
Ru-3␤-CD/Cc
Ru/C
Ru-3␥-CD/C
Ru-3␣-CD/C
Ru-3␤-CD/Cc
Conditions: reaction temperature = 85 ◦C, carrier gas = H2, H2:xylene = 100:1, total flow rate = 60 mL min−1
.
a
Turn-over frequency defined as number of moles of converted xylene per mole of ruthenium per second.
First preparation of Ru-3␤-CD/C.
Second preparation of Ru-3␤-CD/C.
b
c
The Ru-3␤-CD/C(400) sample was obtained by a thermal treatment of the Ru-3␤-CD/C under N2 (60 mL min−1) at 400 ◦C for 4 h.
d

340
F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
Table 4
TOFs, cumulate TOFs and trans-DMCH selectivities measured in the hydrogenation of mixture of o-xylene and p-xylene at 85 ◦C over carbon-supported ruthenium catalysts.
a
Sample
TOFo-xylene × 103 (s−1
)
TOFp-xylene × 103 (s−1
)
Cumulate TOF × 103 (s−1
)
Sel. trans 1,2-DMCH (%)
Sel. trans 1,4-DMCH (%)
Ru/C
1.5
2.3
4.1
2.8
3.0
5.3
4.3
5.3
9.4
1.7
2.0
2.0
4.1
15.4
16.9
16.3
24.8
Ru-3␥-CD/C
Ru-3␣-CD/C
Ru-3␤-CD/Cb
11.3
16.2
27.5
Conditions: reaction temperature = 85 ◦C, carrier gas = H2, feed composition = 420 ppm o-xylene and 580 ppm p-xylene in H2, total flow rate = 60 mL min−1
.
a
Turn-over frequency defined as number of total moles of converted xylene per mole of ruthenium per second.
b
Second preparation of Ru-3␤-CD/C.
have shown that the efficiency of the Ru(0) nanoparticles in the
hydrogenation of unsaturated compounds was affected by steric
hindrance at high CD:Ru ratios [30]. Note that the Ru-5␤-CD/C and
Ru-10␤-CD/C samples give similar activities and the 15% difference
is likely due to experimental uncertainties occurring during the
preparation of the Ru-CD/C catalysts by the adsorption method.
In terms of stereoselectivity, the product distribution in o-
xylene hydrogenation shows the preferential formation of the
cis-1,2-DMCH product over the reference Ru/C catalyst (entry 1,
Table 3). This effect has been confirmed with m- and p-xylene,
even if a slight increase in the proportion of the trans-product
from 1,3- and 1,4-dimethylbenzene is recorded (entries 10 and 14,
Table 3). This stereoselectivity tendency is in line with what is gen-
erally observed for the ruthenium supported catalysts in xylene
hydrogenation [38]. In fact, even if the hydrogenation of xylene is
often considered as a structure insensitive reaction, Rahaman and
Vannice have reported that different factors, such as the nature
of the metal, surface acidity and metal particle dispersion could
affect the selectivity [39]. On the basis of the roll-over model
proposed by Inoue et al. [40], it is generally assumed that the for-
mation of the trans-product is more energetically demanding and
requires a longer stay of the aromatic precursor on the surface,
and therefore a stronger adsorption. This could be achieved by
strengthening the interactions of the intermediate with the active
site.
aromatic substrate, and especially the para isomer, is better rec-
ognized in the cavity of RaMe-␤-CD through weak supramolecular
interactions.
In order to further underline the importance of RaMe-␤-
CD during the catalytic process, an additional catalytic test has
been performed over the Ru-3␤-CD/C(400) sample, for which the
RaMe-␤-CD has been thermally decomposed at 400 ◦C under N2.
Interestingly, the resulting catalytic TOF and 1,2-DMCH selectivity
are very similar as those obtained with the Ru/C sample prepared
without cyclodextrin (entry 9, Table 3) and confirms the promo-
tional effect of the RaMe-␤-CD in catalysis.
3.3.2. Hydrogenation of a two-component mixture (o- and
p-xylene)
In order to investigate the scope and limitations of the Ru-CD/C
catalysts, the catalytic hydrogenation of a mixture of o- and p-
xylene (420 and 580 ppm, respectively) has been also examined.
The results in terms of cumulate TOF (TOFo-xylene + TOFp-xylene) at
85 ◦C are gathered in Table 4 and Fig. 5 and show that replacing
single compounds with mixtures lead to considerable changes in
the hydrogenation of xylenes.
Thus, as clearly shown over the reference Ru/C catalyst in Fig. 5,
an inhibiting effect on the hydrogenation of xylenes in the two-
component mixture is obtained. This lower rate of hydrogenation
seems to indicate that the competitive adsorption of the aromatic
reactants occurs on the same active sites of the catalyst, and that
o-xylene could act as a pseudo-poison in the hydrogenation of
p-xylene (taking into account that p-xylene is more easily hydro-
genated than o-xylene when tested separately). However, it should
be noticed that the cis- and trans-stereoselectivities remain practi-
cally unchanged in comparison with those obtained with the single
substrates.
Notably, we see that the use of RaMe-␤-CD as the stabiliz-
ing agent allows the increase in the trans to cis ratio. Thus, the
use of Ru-3␤-CD/C sample shows a twofold increase in trans-1,2-
DMCH selectivity from 1,2-dimethylbenzene (3.9% vs 1.9% for the
Ru/C) whereas no significant change in the trans to cis ratio is
observed with the catalysts prepared via the nanoparticles stabi-
lized by RaMe-␣- and RaMe-␥-CD. This behavior has been further
confirmed in the hydrogenation of m- and p-xylene.
The above results can be rationalized by the promoter effect of
cyclodextrins on the dispersion of the active species, as previously
evidenced by TEM. In fact, it is well known that the particle sizes and
the dispersion state influence the catalytic activity of a supported
metal catalyst. Moreover taking into consideration that the RaMe-
CDs are intimately linked to the metal particles, the CDs can play
an additional role during the catalytic process, through host–guest
interactions by promoting the adsorption of the aromatic sub-
strate on the catalyst surface. The ability of bulk cyclodextrins to
interact selectively with gaseous o-, m- and p-xylene compounds
has already been demonstrated for gas phase chromatography
applications through the use of CD-packed columns [41,42]. It is
worth mentioning that the effect of inclusion on the retention fol-
lowed the same tendency as in reverse liquid chromatography.
Although the association constants (Kf) between the xylenes and
CDs were only determined in aqueous medium, it was clearly
evidenced that the higher values of Kf were obtained with the
cyclodextrins possessing a ␤-type cavity, such as the RaMe-␤-
CD. For instance, Fenyvesi et al. reported the Kf values between
the RaMe-␤-CD and o-xylene (225 M−1), m-xylene (216 M−1) and
p-xylene (300 M−1) [43]. Finally, the best catalytic properties of
the Ru-3␤-CD/C samples could also be linked to the fact that the
Fig. 5. Cumulate TOF of o-xylene and p-xylene in mixture over carbon-supported
ruthenium catalyst (reaction temperature = 85 ◦C, feed composition = 420 ppm o-
xylene and 580 ppm p-xylene in H2, total flow rate = 60 mL min−1). aSecond
preparation of Ru-3␤-CD/C.

F. Wyrwalski et al. / Applied Catalysis A: General 391 (2011) 334–341
341
In line with the previous results obtained with the Ru/C catalyst,
the same inhibition phenomenon of the p-xylene hydrogenation is
observed with the Ru-3-␣CD/C and Ru-3-␥CD/C samples. In con-
trast, when using the chemically modified RaMe-␤-CD, it has been
found that the simultaneous hydrogenation of o- and p-xylene is
of the same order of magnitude than the one measured with the
most reactive substrate, i.e. p-xylene, showing that the inhibition
effect is here drastically reduced. This behavior confirms the pre-
vious results obtained with the xylenes tested separately and can
be explained by a higher reactant surface concentration close to
the metal active sites thanks to the best ability of RaMe-␤-CD to
form inclusion complexes with the aromatic substrate. In terms of
stereoselectivity, the formation of the trans-1,4-DMCH is also more
favored over the Ru-3-␤CD/C sample as the selectivity achieves
24.8% against 15.4% without cyclodextrin.
[3] A. Guerrero-Ruiz, P. Badenes, I. Rodriguez-Ramos, Appl. Catal. A Gen. 173 (1998)
313–321.
[4] R. Narayanan, M.A. El-Sayed, J. Catal. 234 (2005) 348–355.
[5] C.L. Hui, X.G. Li, I.M. Hsing, Electochim. Acta 51 (2005) 711–719.
[6] S. Ikeda, S. Ishino, T. Harada, N. Okamoto, T. Sakata, H. Mori, S. Kuwabata, T.
Torimoto, M. Matsumura, Angew. Chem. 118 (2006) 7221–7224;
S. Ikeda, S. Ishino, T. Harada, N. Okamoto, T. Sakata, H. Mori, S. Kuwabata, T.
Torimoto, M. Matsumura, Angew. Chem. Int. Ed. 45 (2006) 7066–7069.
[7] H.S. Oh, J.G. Oh, H.G. Hong, H. Kim, Electochim. Acta 52 (2007) 7278–7285.
[8] C.Y. Lu, M.N. Wey, L.I. Chen, Appl. Catal. A Gen. 325 (2007) 163–174.
[9] J.M. Nadgeri, M.M. Telkar, C.V. Rode, Catal. Commun. 9 (2008) 441–446.
[10] B.A. Kakade, S. Sahoo, S.B. Halligudi, V.K. Pillai, J. Phys. Chem. C 11 (2008)
13317–13319.
[11] Y. Motoyama, M. Takasaki, S.-H. Yoon, I. Mochida, H. Nagashima, Org. Lett. 11
(2009) 5042–5045.
[12] A. Roucoux, J. Schulz, H. Patin, Chem. Rev. 102 (2002) 3757–3778.
[13] B.L. Cushing, V.L. Kolesnichenko, C.J. O’Connor, Chem. Rev. 104 (2004)
3893–3946.
[14] U. Drechsler, B. Erdogan, V.M. Rotello, Chem. Eur. J. 10 (2004) 5570–5579.
[15] D. Astruc, F. Lu, J.R. Aranzaes, Angew. Chem. 117 (2005) 8062–8083;
D. Astruc, F. Lu, J.R. Aranzaes, Angew. Chem. Int. Ed. 44 (2005) 7852–7872.
[16] A. Roucoux, A. Nowicki, K. Philippot, in: D. Astruc (Ed.), Nanoparticles and
Catalysis, Wiley–VCH, Weinheim, 2008, pp. 349–388.
[17] M.T. Reetz, S. Waldvogel, Angew. Chem. 109 (1997) 870–873;
M.T. Reetz, S. Waldvogel, Angew. Chem. Int. Ed. 36 (1997) 865–867.
[18] R. Breslow, S.D. Dong, Chem. Rev. 98 (1998) 1997–2011.
[19] H. Bricout, F. Hapiot, A. Ponchel, S. Tilloy, E. Monflier, Sustainability 1 (2009)
924–945.
[20] M. Ferreira, H. Bricout, A. Sayede, A. Ponchel, S. Fourmentin, S. Tilloy, E. Monflier,
Adv. Synth. Catal. 350 (2008) 609–618.
[21] M. Komiyama, H. Hirai, Bull. Chem. Soc. Jpn. 56 (1983) 2833–2834.
[22] J. Alvarez, J. Liu, E. Roman, A.E. Kaifer, Chem. Commun. (2000) 1151–
1152.
[23] J. Liu, J. Alvarez, W. Ong, E. Roman, A.E. Kaifer, Langmuir 17 (2001) 6762–
6764.
[24] L. Strimbu, J. Liu, A.E. Kaifer, Langmuir 19 (2003) 483–485.
[25] J. Liu, W. Ong, A.E. Kaifer, Langmuir 18 (2002) 5981–5983.
[26] J. Liu, S. Mendoza, E. Roman, M.J. Lynn, R.L. Xu, A.E. Kaifer, J. Am. Chem. Soc. 121
(1999) 4304–4305.
[27] S.C. Mhadgut, K. Palaniappan, M. Thimmaiah, S.A. Hackney, B. Török, J. Liu,
Chem. Commun. (2005) 3207–3209.
[28] C. Xue, K. Palaniappan, G. Arumugam, S.A. Hackney, J. Liu, H. Liu, Catal. Lett. 116
(2007) 94–100.
[29] A. Nowicki, Y. Zhang, B. Léger, J.P. Rolland, H. Bricout, E. Monflier, A. Roucoux,
Chem. Commun. (2006) 296–298.
[30] A. Nowicki-Denicourt, A. Ponchel, E. Monflier, A. Roucoux, J. Chem. Soc., Dalton
Trans. (2007) 5714–5719.
[31] A. Denicourt-Nowicki, A. Roucoux, F. Wyrwalski, E. Monflier, A. Ponchel, Chem.
Eur. J. 14 (2008) 8090–8093.
[32] Y. Kenichi, M. Atsushi, T. Yukio, S. Mitsukatsu, Y. Yoshiaki, I. Tomoyuki, JP patent
8333406 (1996).
4. Conclusion
Ruthenium(0) nanoparticles were synthesized by using RuCl3
in the presence of randomly methylated cyclodextrins (␣, ␤ and
␥) as capping materials and deposited onto the carbon support via
adsorption. The TEM images showed that the nanoparticles have
a spherical shape with an average diameter of 2.4 nm, randomly
dispersed on the external surface and in the mesopores of the car-
bon. The carbon-supported ruthenium(0) nanocatalysts stabilized
by randomly methylated cyclodextrins were evaluated in the gas
phase hydrogenation of xylene and appeared to be more efficient
than the reference Ru/C catalyst. Thus, our results demonstrated
that the randomly methylated ␤-cyclodextrin can be considered as
an original tool to develop heterogeneous catalysts for gas phase
reactions thanks to its ability to stabilize and disperse the metal-
lic nanoparticles and to concentrate the aromatic reactants in the
vicinity of the catalytic active sites through molecular recogni-
tion phenomena. Thus, the optimization of the catalyst preparation
showed that the Ru-3␤-CD/C sample allowed the achievement of
the highest hydrogenation activity and stereoselectivity towards
the formation of trans-dimethylcyclohexane. Most important, we
have demonstrated that these catalytic performances in terms of
activity and stereoselectivity remain high in the case of a two-
component mixture of aromatics. This key result could answer to a
great industrial catalytic challenge, in particular with the increasing
industrial demand for low aromatic diesel fuels.
[33] S. Lowell, J. Shields, M. Thomas, M. Thommes, Characterization of Porous Solids
and Powders: Surface Area, Pore Size and Density, Kluwer Academy Publishers,
2004.
[34] H.P. Boehm, Carbon 32 (1994) 759–769.
[35] S. Hermans, C. Diverchy, O. Demoulin, V. Dubois, E.M. Gaigneaux, M. Devillers,
J. Catal. 243 (2006) 239–251.
[36] F.T. Trotta, M. Zanetti, G. Camino, Polym. Degrad. Stabil. 69 (2000) 373–
379.
Acknowledgments
The authors are grateful to Roquette Frère (Lestrem, France)
for the generous gift of cyclodextrins. MeadWestvaco Corporation
(Covington, USA) is also gratefully acknowledged for the gener-
ous gift of the activated carbon. The TEM facility in Lille (France)
is supported by the Conseil Regional du Nord-Pas de Calais, and the
European Regional Development Fund (ERDF).
[37] P. Reyes, M.E. König, G. Pecchi, I. Concha, M. Lopez Granados, J.L.G. Fierro, Catal.
Lett. 46 (1997) 71–75.
[38] R. Gomez, G. Del Angel, V. Bertin, React. Kinet. Catal. Lett. 44 (1991) 517–
522.
[39] M.V. Rahaman, M.A. Vannice, J. Catal. 127 (1991) 251–266.
[40] Y. Inoue, J.M. Herrmann, H. Schmidt, R.L. Burwell, J.B. Butt, J.B. Cohen, J. Catal.
53 (1978) 401–413.
[41] T. Kos´cielski, D. Sybilska, L. Feltl, E. Smolková-Keulemansová, J. Chromatogr.
286 (1984) 23–30.
References
[42] E. Smolková-Keulemansová, L. Feltl, S. Krysl, J. Incl. Phenom. 3 (1985) 183–
196.
[1] F. Rodriguez-Reinoso, Carbon 36 (1998) 159–175.
[2] H. Marsh, F. Rodriguez-Reinoso, Activated Carbon, Elsevier Ltd., Amsterdam,
2006.
[43] N. Szaniszlo, É. Fenyvesi, J. Balla, J. Incl. Phenom. Macrocycl. Chem. 53 (2005)
241–248.