Catalytic activity of palladium supported on single wall carbon nanotubes compared to palladium supported on activated carbon: Study of the Heck and Suzuki couplings, aerobic alcohol oxidation and selective hydrogenation

Article abstract of DOI:10.1016/j.molcata.2004.11.030

Nanoparticles (2-10 nm) of palladium have been deposited on single wall carbon nanotubes (SWNT) by spontaneous reduction from Pd(OAc)₂ or from oxime carbapalladacycle. These catalysts exhibit higher catalytic activity than palladium over activa

Full text of DOI:10.1016/j.molcata.2004.11.030

Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
Catalytic activity of palladium supported on single wall carbon
nanotubes compared to palladium supported on activated carbon
Study of the Heck and Suzuki couplings, aerobic alcohol oxidation
and selective hydrogenation
Avelino Corma∗∗, Hermenegildo Garcia∗, Antonio Leyva
Instituto de Tecnolog´ıa Qu´ımica CSIC-UPV, Universidad Polite´cnica de Valencia, Av. De los Naranjos s/n, 46022 Valencia, Spain
Received 30 September 2004; received in revised form 29 November 2004; accepted 29 November 2004
Available online 21 January 2005
Abstract
Nanoparticles (2–10 nm) of palladium have been deposited on single wall carbon nanotubes (SWNT) by spontaneous reduction from
Pd(OAc)2 or from oxime carbapalladacycle. These catalysts exhibit higher catalytic activity than palladium over activated carbon (Pd/C) for
the Heck reaction of styrene and iodobenzene and for the Suzuki coupling of phenylboronic and iodobenzene. This fact has been attributed
as reflecting the dramatic influence of the size particle on the activity of the palladium catalyst for C C bond forming reactions as compared
to other reaction types less demanding from the point of view of the particle size. Thus, in contrast to the Heck and Suzuki reactions, Pd/C
is more active than palladium nanoparticles deposited on SWNT for the catalytic oxidation by molecular oxygen of cinnamyl alcohol to
cinnamaldehyde and for the hydrogenation of cinnamaldehyde to 3-phenylpropionaldehyde.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Nanotubes; Palladium; Single wall carbon nanotubes; Catalysts; Hydrogenation
1. Introduction
necessary to assess the performance of SWNT, particularly
compared to widely used activated carbons. We have been
involved in the use of SWNT as solid supports for catalytic
reactions [19]. Our interest comes from the consideration
of the importance that activated carbons have as optimum
supports for many reaction types including catalytic hydro-
genations and oxidations. In both cases, the known ability
of SWNT to adsorb gases in even higher quantities than
activated carbons might be advantageous co-operating to the
success of the reaction. For instance, given that the interior
of the tubes is open to the exterior, there has also been a large
interest in the use of SWNT for hydrogen storage [20–27].
SWNTs are, however, structurally and chemically very
different from most activated carbons. SWNT have a well-
defined structure formed by thin (1.3 nm) and long (>100 nm)
carbon tubes of remarkable flexibility, mechanical strength
and high Young modulus that agglomerate through van der
Waals forces to form bundles. Purified SWNTs free from
Since their discovery by Iijima [1,2], there has been a
considerable interest in exploring the applications of the
remarkable structural, physical and chemical properties
of single wall carbon nanotubes (SWNT) in all areas of
physics and material chemistry. Most of the efforts have been
focused on nanotechnology trying to exploit the mechanical
strength [3–5] and metallic conductivity [5–10] of different
kinds of SWNT as well as the possibility to functionalize
covalently SWNT to obtain advanced materials [11–13].
SWNT have also been used in heterogeneous catalysis as
support for noble metals [14–18]. The reports on the use of
SWNT in catalysis are, however, considerably more scarce
than those focused on material science and more effort is still
∗
∗∗
Corresponding author. Tel.: +34 963877897; fax: +34 963877809.
Co-corresponding author.
1381-1169/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2004.11.030

98
A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
catalyst particles [14] are formed by an almost perfect
graphene wall terminated with carboxy groups. In contrast to
this remarkable periodicity, activated carbonshave ill-defined
structures constituted of a wide distribution of platelets of
condensed aromatic hydrocarbons of different sizes and com-
positions connected through methylene, oxygen, nitrogen or
sulfur bridges. Also depending on the source of the activated
carbons, significant percentages of heteroatoms (O, N, S) and
even transition metals can be present in activated carbons
forming a large variety of functional groups such as phenol,
quinones and sulphonic.
In the present work, we want to report the catalytic
activity of palladium metal-containing SWNTs prepared by
three different procedures for a series of palladium catalyzed
reactions, namely the Heck and Suzuki C C coupling,
alcohol aerobic oxidation and C C hydrogenation. Our goal
is to compare the results obtained with palladium-supported
SWNT with those achieved with Pd/C to establish differences
on the performance between both catalysts for a series of
catalytic reaction that might exhibit different demand from
the point of view of particle size. Our results show indeed
that the catalytic activity of palladium on SWNT exhibits
a characteristic pattern that is clearly different from that of
Pd/C, the former system being specially interesting from
C C cross-coupling reactions, for which palladium particle
size seems to play a crucial role compared to hydrogenations
and oxidations.
respect to the number of internal atoms [28,29]. From the re-
cent work on metal nanoparticles, it has become evident that
the catalytic activity of a palladium containing catalysts de-
pends on a large extent of the particle size, on the dispersion
of the nanoparticles and on the way in which nanoparticles
are stabilized against agglomeration [30–32]. Some of these
properties are determined by the preparation procedure and
controlled in a certain extent by the support. The prepara-
tion procedure is crucial to obtain particles of the nanometric
size with adequate stability and dispersion. In the present
work, we have prepared three types of SWNT supported pal-
ladium metal catalysts in where two precursors of the palla-
dium metal have been used or in where the SWNT has been
conveniently functionalized to introduce a covalently bonded
quaternary ammonium stabilizer [33].
A commercial sample of HiPCO single wall carbon nan-
otube normally contains rest of the solid catalyst particles
used for the synthesis of the SWNT from organic molecules.
The typical purification procedure consists on treating the
SWNT sample with concentrated aqueous solution of nitric
acid at 100 ◦C for long times. After purification, the absence
of catalyst particles can be simply assessed by thermogravi-
metric analysis (TGA) where the complete burning of the
material should be observed and also by elemental analy-
sis in where a high C content should be measured. In our
case, TGA after purification shows no residue and at the same
time the chemical analysis of ambient-equilibrated purified
SWNT (still containing some humidity) show a carbon con-
tent about 80%, the remaining percentage being attributed to
adsorbed water from the ambient. The simplest preparation
procedure consists in contacting SWNT with a solution of
palladium acetate. In the literature, it has been reported that
upon contacting platinum or gold salts with SWNT a spon-
taneous reduction occurs forming the corresponding metal
nanoparticles distributed along the carbon nanotubes [34]. A
electrochemical current can be measured during the reduc-
tion. Although we have not been able to find an analogous
report using palladium salts, it could be easily anticipated
2. Results and discussion
2.1. Preparation of palladium catalysts supported on
SWNT
Metal nanoparticles exhibit specific properties arising
from the large surface/bulk atomic ratio. These properties are
not observed in larger micrometric or submicrometric parti-
cles in where the fraction of external atoms is negligible with
Scheme 1. Synthetic route followed to obtain the SWNT-supported palladium catalyst Pd-SWNT-1 and Pd-R3N+-SWNT-1.

A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
99
also indicated in Scheme 1. The beneficial influence of
long-chain alkyl quaternary ammonium in the stabilization
of noble metal nanoparticles in colloidal solutions has
been established [35–37]. Stabilization of nanoparticles in
suspension may be probably different from stabilization of
nanoparticles on solid surfaces, our aim was to test whether
or not the incorporation of these quaternary ammonium
units on the SWNT plays an analogous role as in solution
stabilizing the Pd nanoparticles supported on the nanotube.
The stabilization of the particle size could be reflected in
the catalytic performance of the material. Although most
of the positive influence of the presence of long-chain alkyl
ammonium deals with colloids without any support, it has
been observed in other cases that the presence of quaternary
ammonium salts as additives play a beneficial influence on
the catalysts using supported metal particle.
Fig. 1. Transmission UV–vis spectrum of a solution of the oxime carbapal-
ladacycle complex (0.03 mmol) in THF (10 mL) before (a) and after treated
at reflux temperature with SWNT (50 mg) for 24 h. The decrease in intensity
is associated to decomposition of the complex and formation of palladium
metal particles on the SWNT.
In the third related catalyst based on SWNT (Pd-
SWNT-2), the precursor of palladium metal cluster
was an organometallic oxime carbapalladacycle complex
(Scheme 2). This complex is easily available by reacting
in methanol Li2PdCl4 with 4-hydroxyacetophenone oxime
[38–42]. This palladium complex has been used as a highly
active pre-catalyst to generate palladium nanoparticles. The
higher stability of palladium in the complex as compared
to free Pd2+ from Pd(OAc)2 may have a positive influence
on the palladium particle size and on the dispersion of the
clusters. As expected in view of the behaviour of Pd(OAc)2,
stirring a suspension of SWNT and the oxime carbapallada-
cycle leads to a gradual palladium reduction with formation
of black palladium as indicated by the progressive decrease
in the solution of the absorption band at 340 nm responsible
for the yellow colour of the oxime carbapalladacycle (Fig. 1).
The series of three palladium catalysts supported on
SWNT were studied by TEM. The corresponding TEM im-
ages show that the SWNT catalysts consist on bundles of thin
single-walled tubes in where palladium nanoparticles are dis-
persed. Selected images are shown in Fig. 2. The particle size
distribution was determined by a statistical analysis of a set
of the images with a surface area of 500 nm × 500 nm. In the
cases of Pd-SWNT-1 and Pd-R3N+-SWNT-1 TEM images
show that the dispersion of the palladium particles is rela-
tively high and that there is a broad particle size distribution.
Nevertheless, what is relevant for our study is the presence
of nanoparticles in the range 1–10 nm. Fig. 2 also presents
the particle size distribution analysis for Pd-SWNT-1 and Pd-
SWNT-2. As expected, this study shows that the particle size
distribution depends on a certain extent on the actual prepa-
that palladium salts would exhibit a similar behaviour. In
fact, we have been able to obtain palladium nanoparticles
upon contacting Pd(OAc)2 with SWNT, this simple proce-
dure leading to the first Pd-SWNT-1 catalyst used in this
work (Scheme 1). The following pieces of evidence support
that the palladium salt has been reduced upon contacting a
solution of palladium complex with SWNT: (i) the reduc-
tion peak potential of Pd2+ in THF (versus standard hydro-
gen electrode (SHE)) measured by us by cyclic voltammetry
was 0.466 V for Pd(OAc)2 and 0.509 V for carbapalladacy-
cle oxime; (ii) the reported work function of SWNT [34] is
−5.0 V in vacuum giving an energy of the Fermi level high
enough to effect the reduction of Pd2+ and (iii) the charac-
teristic UV-absorption band of Pd2+ in the oxime carbapal-
ladacycle complex disappears gradually upon contacting a
solution of this complex and SWNT (see Fig. 1). All these
facts are compatible with the spontaneous reduction of Pd2+
to Pd(0) with formation of metallic Pd particles, as it was
reported for platinum and gold.
Alternatively, spontaneous Pd2+ reduction starting from
Pd(OAc)2 was also accomplished on a modified SWNT in
which the carboxylic acid groups present on the tips and
defects of the tubes had been functionalized with long-chain
quaternary alkyl ammonium ions, namely dicetyl methyl
(2-hydroxyethyl) ammonium chloride (Pd-R3N+-SWNT-1).
The covalent linkage between the SWNT and the quaternary
ammonium ion was accomplished through esterification
of the 2-hydroxyethyl group of ammonium salt and the
carboxylic groups. The actual steps of the synthesis are
Scheme 2. Synthetic route followed for the preparation of the SWNT-supported palladium catalyst Pd-SWNT-2.

100
A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
Fig. 2. Trasmission electron microscopy images for Pd-SWNT-1 (top), Pd-R3N+-SWNT-1 (center) and commercial Pd/C (Aldrich, 1% w/w, bottom) together
with diameter size distribution of palladium nanoparticles for at least five different TEM images of Pd-SWNT-1, Pd-SWNT-2 and Pd/C. The palladium
nanoparticles (<2–3 nm) appear as little dark dots together with some palladium cluster (>10 nm).
ration procedure (see Fig. 2). This TEM study agrees with
the previous report on the spontaneous reduction of noble
metals on SWNT and overall confirms the nanometric size of
palladium metal particles formed on SWNTs. It is expected
that this different particle size distribution could be reflected
in differences in the catalytic activity for some reactions.
For the sake of comparison, we have carried out an analo-
gous TEM study of commercial Pd/C (Fig. 2). TEM analyses
showed that the particle size distribution for Pd/C is signif-
icantly broader with a considerable number of Pd particles
larger than in Pd-SWNTs. Although a more deep understand-
ing of the reduction process on SWNT is still needed, a likely
rationalisation of the smaller particle size using SWNT as
support can be put forward based on the presence of car-
boxylic acid groups. Carboxylate groups would establish a
Coulombic interaction with Pd2+ analogous to other known
metal ion–carboxylate complexes, followed by a subsequent
controlled reduction by the electrons of the SWNT walls. In
the case of activated carbons, the presence of a variety of dif-
ferent functional groups that interact differently with metal
ions would lead to a less homogeneous initial distribution of
Pd2+ with formation of larger agglomerates.

A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
101
Fig. 4. Time conversion plot for the oxidation of cinnamyl alcohol (1 eq.) in
ethanol (5 mL) under oxygen bubbling (flow: 25–30 mL min−1) at 80 ◦C in
the presence of Pd/C (a), Pd-SWNT-1 (b) and Pd-R3N+-SWNT-1 (c) (4 mg,
0.2 Pd mol%).
Fig. 3. Time conversion plot for the Heck reaction of iodobenzene (1 eq.),
styrene (1.5 eq.) and tributylamine (2 eq.) in DMF (5 mL) at 140 ◦C in
the presence of Pd-SWNT (a), Pd-R3N+-SWNT-1 (b) and Pd/C (c) (3 mg,
0.01 Pd mol%).
forming a reaction under the typical conditions, and filtering
the solid at 50% conversion while the mixture was still hot.
The clear solutions were allowed to react further in the ab-
sence of the solid, whereby the same temporal profile as in
the presence of the Pd-SWNT solids were observed. If the ac-
tivity would have been exclusively due to the presence of the
Pd-SWNT solid catalysts, the reaction would have stopped.
These experiments demonstrate that under the reaction con-
ditions in DMF, palladium nanoparticles become desorbed
from the support and suspended in the liquid phase.
The extent of leaching was significantly lower when the
Heck reaction is carried out in o-xylene using tributylamine
as base instead of DMF (20% in toluene as compared to 100%
inDMF). Obviously, thenatureofthesolventisknowntoplay
a crucial role in the occurrence and extent of leaching from
a solid to the liquid phase since the migration depends on
the partition coefficient between the two phases. In previous
studies on the extent of leaching on palladium complexes
supported on porous silicates, it has been established that
migration of palladium to the solvent is lower in toluene than
in DMF [50–52].
Obviously, leaching can be a major problem when deal-
ing with solid catalyst due to the eventual depletion of the
metal content with the consequent loss in activity. However,
in spite of the occurrence of the palladium leaching during
the reaction, no significant decrease in the catalytic activity of
the solids was observed. Maintenance of the catalytic activity
can be explained by the re-deposition of most of the palla-
dium nanoparticles on the SWNT as the reaction ends and the
system cools down. This behavior is also generally observed
in Pd/C systems and has been described as the “boomerang
effect” [53,54].
2.2. Catalytic activity
The palladium supported Pd-SWNT-1 and Pd-SWNT-
2 catalysts showed high activity for the Heck reaction of
iodobenzene with styrene in DMF at 150 ◦C using NaAcO
as base. Since the percentage of palladium metal varies in
each sample of SWNT catalyst, for the sake of having a valid
comparison all the reactions were carried out keeping the
substrate-to-palladium molar ratio constant at 0.01 mol%. In
all cases, t-stilbene was the only product detected (Eq. (1)).
The Pd-SWNT-1 catalysts prepared from Pd(AcO)2 exhibits
the highest activity in the series and no benefits were ob-
served when the catalysts was prepared starting from the
oxime carbapalladacycle complex (Pd-SWNT-2) or by in-
troducing covalently on the SWNT dicetylmethylammonium
(Pd-R3N+-SWNT-1). Fig. 3 showed the time conversion plot
for this Heck reaction in the presence of SWNT-supported
palladium catalysts. Importantly, the activity of Pd/C is con-
siderably lower than those catalysts obtained using SWNT
as support. This low activity of Pd/C as Heck catalyst is in
agreement with previous reports in the literature [43–45] and
exemplifies the influence of the support on the catalytic ac-
tivity. Most likely, these results reflect the high demand of
the Heck reaction for small particle size palladium particles
[46–49].
Palladiumsupportedonsinglewallcarbonnanotubeswere
also tested as catalysts for the aerobic oxidation of cinnamy-
lalcohol in ethanol (Eq. (2)) and the selective C C hydro-
genation of cinnamaldehyde to 3-phenylpropionaldehyde in
ethanol (Eq. (3)). The aim of this study was to determine
whether or not the reported superior ability of SWNT to ad-
sorb hydrogen and other gases with respect to active carbon
could play a positive role in heterogeneous catalysis using
gas reactants. Figs. 4 and 5 show selected time-conversion
(1)
When performing a reaction using a solid catalyst two
important issues that need to be addressed are to determine
whether or not the reaction is truly heterogeneous and the
reusability of the solid catalyst. In the case of Pd-SWNT-1
and Pd-SWNT-2 catalysts, leaching of the palladium from
the SWNT support to hot DMF was demonstrated by per-

102
A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
on single wall carbon nanotubes is higher than that of
palladium supported on activated carbon. On the other hand,
Pd/C is more effective when the reaction involves gases,
presumably because these reactions are less demanding from
the point of view of palladium particle size. The preparation
procedure and covalent functionalization of the SWNT
modifies the activity of the resulting palladium catalyst,
but no improvement has been achieved with respect to the
simplest procedure based on the spontaneous reduction
of Pd(OAc)2 upon contacting with the carbon nanotubes.
These catalytic results can be rationalised considering the
specific demands for C C bond forming reactions in terms
of small particle size as compared to hydrogenations and
oxidations.
Fig. 5. Time conversion plot for the selective hydrogenation of cin-
namaldehyde (1 eq.) in o-xylene (5 mL) under hydrogen bubbling (flow:
20 mL min−1) at 80 ◦C in the presence of Pd/C (a) Pd-R3N+-SWNT-1 (b)
(4 mg, 0.2 Pd mol%).
plots for these two reactions under some of the conditions
studied, showing that this turns out not to be the case.
3. Experimental section
The reagents and solvents were obtained from commer-
cial sources and were used without further purification. Gas
chromatographic analyses were performed on a HP 5890
instrument equipped with a 25 m capillary column of 5%
phenylmethylsilicone. GC/MS analyses were performed on
a Agilent 5973N spectrometer equipped with the same col-
(2)
1
umn and in the same conditions as GC. H and 13C NMR
were recorded in a 300 MHz Bruker Avance instrument us-
ing CD3OD or DMSO-d6 as solvents and TMS as internal
standard. IR spectra were recorded on a Jasko 460 plus spec-
trophotometer using sealed greaseless quartz cells provided
with CaF2 windows. Self-supported wafers (10 mg) were ob-
tained by mixing the solid with KBr and pressing the mix-
ture at 10 Ton cm−1 for 5 min. UV–vis spectra were recorded
on a Shimadzu scanning spectrophotometer using THF as
solvent. Elemental analysis of the solids was determined
by chemical combustion using a Perkin-Elmer CHNSO an-
alyzer. The Pd content of the SWNTs was determined by
dispersing the solid in a mixture 1:1:1 of HF:HCl:HNO3
concentration (2.5 mg in ca. 3 mL), diluting the solution
in water and measuring by quantitative atomic absorption
spectroscopy.
(3)
In contrast to the results observed for the C C cross-
coupling, Pd/C becomes far more active than the series of
palladium supported on SWNT for both reactions, the aerobic
oxidation of cinnamyl alcohol and for the C C hydrogena-
tion of cinnamaldehyde. Thus, although the Pd-SWNT was
also able to affect the aerobic oxidation or hydrogenation,
the initial reaction rate at the same substrate–to–palladium
ratio is considerably lower for the Pd-SWNT catalysts than
for Pd/C. In the case of hydrogenation, the difference in the
reaction mechanism and the occurrence of the known hy-
drogen spill-over effect renders the reaction less sensitive to
the particle size distribution in the length scale of the Pd-
SWNT and Pd/C tested in this work [55–58]. A trend worth
noting is, however, that the same relative activity in the se-
ries of Pd-SWNT was maintained and no beneficial influence
of quaternary ammonium ion functionalization or of the use
of an organometallic complex as precursor of the palladium
metal clusters was observed.
In summary, it is well known that the activity of
palladium-supported catalysts depends dramatically on the
metal dispersion and the particle size. However, there are
some reactions that should be more demanding than others
with respect to the particle size. In particular, C C bond
coupling reactions are very demanding from the point of
view of palladium particle size [46–49]. Probably for this
reason, the activity of palladium nanoparticles supported
3.1. Procedure for the synthesis of
4-hydroxyacetophenone oxime (2)
4-Hydroxy-acetophenone (1) (3 g, 0.022 mol) was added
to a solution of hydroxylamine hydrochloride (5.13 g,
0.074 mol) and sodium acetate (10.26 g, 0.125 mol) in wa-
ter (26 mL). The solution was stirred at reflux temperature
for 1 h. After that time, diethyl ether was added several times
for extraction. The organic phases were dried and the sol-
vent evaporated under vacuum. To the obtained residue, hex-
ane was added and 1-(4-hydroxyphenyl)ethanone oxime (2)
started to precipitate as a white solid (3.25 g, 0.0215 mol,
98%). Mp: 146–147 ◦C; IR (KBr, cm−1): 3324, 1642,
1603, 1514, 1444, 1316, 1240, 1176, 940, 825, 589. 1H
NMR δH (ppm, 300 MHz, CD3OD): 7.49 (2H, d, J = 5 Hz),

A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
103
6.77 (2H, d, J = 5 Hz), 2.18 (3H, s). 13C NMR δC (ppm,
300 MHz, CD3OD): 159.9, 156.7, 130.2, 128.9, 116.5, 12.55.
MS (FAB): m/z 151. Anal. Calcd for C8H9NO2 (151.15):
C, 63.5; H, 5.95; N, 9.26. Found: C, 63.19; H, 6.22;
N, 9.35.
ized SWNT was filtered through a PTFE membrane filter
(pore diameter: 0.2 nm) under vacuum and Soxhlet-extracted
with CH2Cl2 for 6 h. Raman peaks as purified SWNT. Loss
of weight by TGA: 88%. Elemental analysis of ambient-
equilibrated SWNT: C 79.62% H 3.21% N 1.39%. Ammo-
nium groups loading: 0.5 mmol g−1
.
3.2. Procedure for the synthesis of the oxime
carbapalladacycle
3.5. Deposition of palladium nanoparticles on purified
or functionalized SWNTs using Pd(OAc)2 as palladium
source
To a solution of Li2PdCl4 (524.1 mg, 2 mmol) in methanol
(4 mL), a methanolic solution (2 mL) of (2) (302 mg, 2 mmol)
and sodium acetate (0.164 g, 2 mmol) was added. The mix-
ture was stirred at room temperature for 72 h. The mixture
was filtered and after adding water (5 mL), the cyclopal-
ladated complex (3) starts to precipitate as a yellow solid
(75%). Mp > 250 ◦C; IR (KBr, cm−1): 3429, 1627, 1584,
1472, 1430, 1375, 1326, 1260, 1211, 1037, 873, 799, 608.
1H NMR δH (ppm, 300 MHz, DMSO-d6): 10.38(1H, s),
9.88(1H, s), 9.77(1H, s) 9.63(1H, s), 7.3 (2H, s), 7.15 (2H,
d, J = 8 Hz), 6.5 (2H, d, J = 8 Hz), 2.23 (6H, s). 13C NMR
δC (ppm, 300 MHz, DMSO-d6): 167.5, 157.3, 154.7, 133.1,
128.3, 122.1, 111.8, 11.6. MS (FAB): m/z 584. Anal. Calcd
for C16H16N2O4Pd2Cl2 (584.04): C, 32.87; H, 2.74; N, 4.79.
Found: C, 32.12; H, 2.87; N, 4.59.
Following the procedure reported by Reetz and co-
workers, Pd(OAc)2 (5.6 mg, 0.025 mmol) was dissolved in
anhydrous THF (9 mL) and 3 mL of this yellow solution
were filtered using a glass syringe coupled with a swinney
13 mm filter (Millipore) through a PTFE membrane filter
(pore diameter: 0.2 nm) to remove any insoluble palladium
species, and magnetically stirred in a pre-heated oil bath
at reflux temperature for 4 h under nitrogen atmosphere
in the presence of 50 mg of purified SWNT (Pd-SWNT-1)
or functionalized SWNT (Pd:ammonium salt molar ra-
tio 1:3, Pd-R3N+-SWNT-1). After this time the stirring
was stopped and a clear supernatant was observed. The
Pd/SWNTs were filtered through a PTFE membrane filter
(pore diameter: 0.2 nm), washed with anhydrous THF
(100 mL) and dried. Raman peaks as purified SWNT. Loss
of weight by TGA: 88% (Pd-SWNT-1). Palladium anal-
ysis: 0.149 mmol g−1 (Pd-SWNT-1) and 0.136 mmol g−1
(Pd-R3N+-SWNT-1).
3.3. Purification of the SWNT
HiPCO nanotube from Carbolex Co. (100 mg) was mag-
netically stirred in HNO3 (3 M, 10 mL) at 100 ◦C for 24 h.
Then, the dispersion was centrifuged and the obtained nan-
otube was washed with bidistilled water (two times) and
THF (two times) by centrifugation until the solution was
completely transparent. Raman spectrum of purified SWNT
3.6. Deposition of palladium nanoparticles on purified
SWNT using the carbapalladacycle oxime complex as
palladium source
showed the characteristic peaks at 1596, 1345 and 167 cm−1
,
A dark yellow solution of the carbapalladacycle complex
(8.7 mg, 0.03 mmol) anhydrous THF (10 mL) was magnet-
ically stirred in a pre-heated oil bath at reflux temperature
under nitrogen atmosphere in the presence of 50 mg of
purified SWNT (Pd-SWNT-2). The course of the reaction
was periodically followed by stopping the stirring and
taken aliquots (100 ␮L) that were diluted in THF (c.a.
3 mL) and monitored by UV spectroscopy. After 48 h,
the Pd/SWNT was filtered through a PTFE membrane
filter (pore diameter: 0.2 nm), washed with anhydrous THF
(100 mL) and dried. Raman peaks as purified SWNT. Loss of
weight by TGA: 82%. Palladium analysis: 0.537 mmol g−1
(Pd-SWNT-2)
the latter corresponding to the breathing mode being specific
of single walled nanotube. The thermogravimetric analysis
of purified SWNT measured under air flow shows a loss of
weightof98%, indicatingthannoinorganiccatalystispresent
in the sample.
3.4. Procedure for anchoring the
dicetylmethylammonium groups on SWNT
Purified SWNT (100 mg) were dispersed in thionyl chlo-
ride (25 mL) and DMF (1 mL). The mixture was placed in
a pre-heated oil bath and magnetically stirred at 65 ◦C for
24 h. Then, the chlorinated SWNT was isolated by filtration
through a PTFE membrane filter (pore diameter: 0.2 nm) un-
der vacuum, washed with dry CH2Cl2 (50 mL) and quickly
placed in a previously dry round-bottomed flask that con-
tains a solution of dicetyl-2-hydroxyethylmethylammonium
chloride (1 g, 2 mmol) in dry CH2Cl2 (5 mL) under argon
atmosphere at room temperature. Then, pyridine (0.5 mL,
6.45 mmol) was added and the dispersion was heated at
45 ◦C for 48 h under nitrogen atmosphere. The functional-
3.7. Procedure for the measurement of the reduction
potential peaks for palladium compounds
A solution of Pd(OAc)2 (6.7 mg, 0.03 mmol) or carbapal-
ladacycle complex (8.7 mg, 0.03 mmol) in dry THF (50 mL)
wasmagneticallystirredunderN2 atmosphereinthepresence
of the corresponding platinum working electrodes, using a
KCl saturated calomel electrode (SCE) as reference (E0(V)

104
A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
versus SHE = 0.2412) and a computer controlled potentiostat.
References
The voltammetry was performed at a rate of 50 mV min−1
.
[1] S. Iijima, T. Ichihashi, Nature 363 (1993) 603.
[2] P.M. Ajayan, T.W. Ebbesen, T. Ichihashi, S. Iijima, K. Tanigaki, H.
Hiura, Nature 362 (1993) 522.
3.8. Typical procedure for C C coupling reactions
[3] M.-F. Yu, B.S. Files, S. Arepalli, R.S. Ruoff, Phys. Rev. Lett. 84
(2000) 5552.
[4] D.A. Walters, L.M. Ericson, M.J. Casavant, J. Liu, D.T. Colbert,
K.A. Smith, R.E. Smalley, Appl. Phys. Lett. 74 (1999) 3803.
[5] H.Y. Yu, S.H. Jhang, Y.W. Park, A. Bittar, H.J. Trodahl, A.B. Kaiser,
Synth. Met. 121 (2001) 1223.
[6] T.V. Sreekumar, T. Liu, S. Kumar, L.M. Ericson, R.H. Hauge, R.E.
Smalley, Chem. Mater. 15 (2003) 175.
[7] N.R. Wilson, D.H. Cobden, J.V. Macpherson, J. Phys. Chem. B 106
(2002) 13102.
[8] M. Shiraishi, M. Ata, Synth. Met. 128 (2002) 235.
[9] N. Minami, S. Kazaoui, R. Jacquemin, H. Yamawaki, K. Aoki, H.
Kataura, Y. Achiba, Synth. Met. 116 (2001) 405.
[10] S.J. Tans, A.R.M. Verschueren, C. Dekker, Nature 393 (1998) 49.
[11] A. Hirsch, Angew. Chem. Int. Ed. 41 (2002) 1853.
[12] J.L. Bahr, J.M. Tour, J. Mater. Chem. 12 (2002) 1952.
[13] M. Sano, A. Kamino, J. Okamura, S. Shinkai, J. Incl. Phenom.
Macrocyccl. Chem. 41 (2001) 49.
[14] P. Serp, M. Corrias, P. Kalck, Appl. Catal. A 253 (2003) 337.
[15] V. Brotons, B. Coq, J.M. Planeix, J. Mol. Catal. A 116 (1997) 397.
[16] Y. Sun, B. Mayers, Y. Xia, Adv. Mater. 15 (2003) 641.
[17] X.R. Ye, Y. Lin, C.M. Wai, Chem. Commun. (2003) 642.
[18] J.P. Tessonnier, L. Pesant, C. Pham-Huu, G. Ehret, M.J. Ledoux,
Stud. Surf. Sci. Catal. 143 (2002) 697.
Pd-SWNT-1 (3 mg) was placed in a double-necked round-
bottomed flask and iodobenzene (496.8 ␮L, 4.44 mmol),
styrene (765.1 ␮L, 6.69 mmol), tributylamine (2.014 mL,
8.92 mmol) and DMF (5 mL) were added (0.01 Pd mol%).
Forothercatalysts, theamountofsolidwasmaintained(3 mg)
whiletheamountofregentsandsolventwasvariedtokeepthe
substrate–to–palladium molar ratio constant. The resulting
suspension was placed in a preheated oil bath at 140 ◦C with
magnetic stirring. After stopping the stirring, 50 ␮L samples
of the supernatant were periodically taken and analysed by
GC. At 1 h, the reaction was cooled for 15 min and CH2Cl2
(10 mL) was added. The mixture wasstirred for5 min, filtered
through a PTFE membrane filter (pore diameter: 0.2 nm) and
the solid washed (100 mL of CH2Cl2) and dried. All the prod-
ucts were characterized by comparison with pure samples and
GC–MS.
3.9. Typical procedure for oxidation of cinnamyl alcohol
A solution of cinnamyl alcohol (32.5 mg, 0.242 mmol)
in ethanol (5 mL) was magnitecally stirred in a pre-heated
oil bath at 80 ◦C in the presence of 4 mg of purified
SWNT (Pd-SWNT-1, 0.2 Pd mol%). Then, oxygen (flow:
25–30 mL min−1) was bubbled into the solution. The
course of the reaction was periodically followed by stop-
ping the stirring and taken aliquots (50 ␮L) that were
quickly cooled and analyzed by GC. All the products were
characterized by comparison with pure samples and by
GC–MS.
[19] C. Baleizao, B. Gigante, H. Garcia, A. Corma, J. Catal. 221 (2004)
77.
[20] J. Lawrence, G. Xu, Appl. Phys. Lett. 84 (2004) 918.
[21] M. Volpe, F. Cleri, Chem. Phys. Lett. 371 (2003) 476.
[22] M. Haluska, M. Hirscher, M. Becher, U. Dettlaff-Weglikowska, X.
Chen, S. Roth, AIP Conf. Proc. 633 (2002) 601.
[23] X. Chen, U. Dettlaff-Weglikowska, M. Haluska, M. Hulman, S. Roth,
M. Hirscher, M. Becher, Mater. Res. Soc. Symp. Proc. 706 (2002)
295.
[24] M. Haluska, M. Hulman, M. Hirscher, M. Becher, S. Roth, I.
Stepanek, P. Bernier, AIP Conf. Proc. 591 (2001) 603.
[25] A.C. Dillon, T. Gennett, M.J. Heben, in PCT Int. Appl., Midwest
Research Institute, Wo, USA, 2001, p. 28.
[26] Y. Ma, Y. Xia, M. Zhao, R. Wang, L. Mei, Phys. Rev. B 63 (2001),
115422/1.
[27] A.C. Dillon, K.M. Jones, M.J. Heben, Proc U.S. DOE Hydrogen
Prog. Rev., Miami May 1-219962, 1996, p. 747.
3.10. Typical procedure for hydrogenation of
cinnamaldehyde
A solution of cinnamaldehyde (35.9 mg, 0.272 mmol) in
o-xylene (5 mL) was magnitecally stirred (400 rpm.) in a
pre-heated oil bath at 80 ◦C in the presence of 4 mg of func-
tionalized SWNT (Pd-R3N+-SWNT-1, 0.2 Pd mol%). Then,
hydrogen (flow: 20 mL min−1) was bubbled into the solution.
The course of the reaction was periodically followed by stop-
ping the stirring and taken aliquots (50 ␮L) that were quickly
cooled and analyzed by GC. All the products were charac-
terized by comparison with pure samples and by GC–MS.
[28] L. Guczi, A. Beck, A. Horvath, D. Horvath, Top. Catal. 19 (2002)
157.
[29] K. Fugami, Organometallic News (2000) 25.
[30] M.-C. Daniel, D. Astruc, Chem. Rev. 104 (2004) 293.
[31] M. Moreno-Manas, R. Pleixats, Acc. Chem. Res. 36 (2003) 638.
[32] B. Corain, M. Kralik, J. Mol. Catal. A 173 (2001) 99.
[33] K.G. Thomas, P.V. Kamat, Acc. Chem. Res. 36 (2003) 888.
[34] H.C. Choi, M. Shim, S. Bangsaruntip, H. Dai, J. Am. Chem. Soc.
124 (2002) 9058.
[35] K.G. Thomas, J. Zajicek, P.V. Kamat, Langmuir 18 (2002) 3722.
[36] H. Yao, O. Momozawa, T. Hamatani, K. Kimura, Chem. Mater. 13
(2001) 4692.
[37] C.-B. Hwang, Y.-S. Fu, Y.-L. Lu, S.-W. Jang, P.-T. Chou, C.R.C.
Wang, S.J. Yu, J. Catal. 195 (2000) 336.
Acknowledgements
[38] L. Botella, C. Najera, Angew. Chem. Int. Ed. 41 (2002) 179.
[39] L. Botella, C. Najera, Tetraheddron Lett. 45 (2004) 1833.
[40] L. Botella, C. Najera, J. Org. Chem. 663 (2002) 46.
[41] D.A. Alonso, C. Najera, M.C. Pacheco, Adv. Synth. Catal. 344
(2002) 172.
Financial support by the Spanish Ministry of Science and
Technology (MAT-2003-1762) and Generalidad Valenciana
(grupos03-020) is gratefully acknowledged. A.L. also thanks
the Spanish Ministry of Science and Technology for a post-
graduate scholarship.
[42] C. Baleizao, A. Corma, H. Garcia, A. Leyva, J. Org. Chem. 69
(2004) 439.

A. Corma et al. / Journal of Molecular Catalysis A: Chemical 230 (2005) 97–105
105
[43] X. Xie, J. Lu, B. Chen, J. Han, X. She, X. Pan, Tetraheddron Lett.
[50] M. Dams, L. Drijkoningen, B. Pauwels, G. Van Tendeloo, D.E. De
Vos, P.A. Jacobs, J. Catal. 209 (2002) 225.
45 (2004) 809.
[44] J.G.E. Krauter, J. Pietsch, K. Koehler, R.G. Heidenreich, Chem. Ind.
[51] A. Corma, H. Garcia, A. Primo, A. Leyva, Appl. Catal. A. 247
(2003) 41–49.
89 (2003) 379.
[45] R.G. Heidenreich, K. Kohler, J.G.E. Krauter, J. Pietsch, Synlett
[52] A. Corma, H. Garcia, A. Leyva, Appl. Catal. A 236 (2002) 179.
[53] I.W. Davies, L. Matty, D.L. Hughes, P.J. Reider, J. Am. Chem. Soc.
123 (2001) 10139.
(2002) 1118.
[46] M. Moreno-Manas, R. Pleixats, S. Villarroya, Organometallics 20
(2001) 4524.
[54] B.H. Lipshutz, S. Tasler, W. Chrisman, B. Spliethoff, B. Tesche, J.
Org. Chem. 68 (2003) 1177.
[47] C. Rocaboy, J.A. Gladysz, New J. Chem. 27 (2003) 39.
[48] V. Calo, A. Nacci, A. Monopoli, S. Laera, N. Cioffi, J. Org. Chem.
[55] P.A. Sermon, G.C. Bond, Catal. Rev. 8 (1973) 211.
[56] G.C. Bond, Stud. Surf. Sci. Catal. 17 (1983) 1.
[57] T. Inui, Stud. Surf. Sci. Catal. 77 (1993) 17.
[58] G.M. Pajonk, Stud. Surf. Sci. Catal. 138 (2001) 165.
68 (2003) 2929.
[49] V. Calo, A. Nacci, A. Monopoli, J. Mol. Catal. A 214 (2004)
45.