Selective deposition of gold nanoparticles on or inside carbon nanotubes and their catalytic activity for preferential oxidation of CO

Article abstract of DOI:10.1002/ejic.201000643

Gold nanoparticles have been deposited on three kinds of carbon nanotubes (CNTs), including nitrogen-doped CNTs, by three different methods, namely, impregnation, organometallic decomposition, and deposition-precipitation. The choice of the gold precursor, the support, and the preparation procedure is critical for the control of the size and location (on or inside the nanotubes) of the gold nanoparticles. These catalysts were tested for the selective oxidation of CO in a hydrogen-rich atmosphere. We have shown that the use of nitrogen-doped CNTs as a support permits one to reach much higher activity and selectivity at low temperaturethan with the other CNT supports. This catalyst also shows a good stability under reaction conditions without detectable sintering. Gold nanoparticles have been selectively deposited on or inside different types of carbon nanotubes by using various synthetic strategies. The catalytic systems prepared on nitrogen-doped nanotubes show promising activity and selectivity for the selective oxidation of CO in a hydrogen-rich atmosphere.

Full text of DOI:10.1002/ejic.201000643

FULL PAPER
DOI: 10.1002/ejic.201000643
Selective Deposition of Gold Nanoparticles on or Inside Carbon Nanotubes and
Their Catalytic Activity for Preferential Oxidation of CO
Eva Castillejos,[a] Rubén Chico,[b] Revathi Bacsa,[a] Silverio Coco,[b] Pablo Espinet,[b]
María Pérez-Cadenas,[c] Antonio Guerrero-Ruiz,[c] Inmaculada Rodríguez-Ramos,[c] and
Philippe Serp*[a]
Keywords: Carbon / Gold / Nanocatalysis / Oxidation / Nanotubes
Gold nanoparticles have been deposited on three kinds of
carbon nanotubes (CNTs), including nitrogen-doped CNTs,
by three different methods, namely, impregnation, organo-
metallic decomposition, and deposition–precipitation. The
choice of the gold precursor, the support, and the preparation
procedure is critical for the control of the size and location
(on or inside the nanotubes) of the gold nanoparticles. These
catalysts were tested for the selective oxidation of CO in a
hydrogen-rich atmosphere. We have shown that the use of
nitrogen-doped CNTs as a support permits one to reach
much higher activity and selectivity at low temperature
than with the other CNT supports. This catalyst also shows a
good stability under reaction conditions without detectable
sintering.
(NPs) supported on oxides. Additionally, some researchers
have suggested that factors such as the choice of the sup-
port or the metal particle size, which might be influenced
by the catalyst preparation method, play an important role
in the CO oxidation reaction.[2]
Introduction
Hydrogen-containing gas mixtures for polymer-electro-
lyte-membrane fuel cells (PEMFCs) should be free of CO
to avoid poisoning of the Pt-based PEMFC catalyst. Prefer-
ential oxidation (PROX) of CO has been recognized as one
of the most direct and efficient methods to achieve accept-
able CO concentrations, typically below 100 ppm.[1] In an
H2-rich atmosphere, H2 and CO compete to react with O2,
thus leading to a decreased hydrogen yield. In addition, CO
can consume additional H2 by undergoing hydrogenation,
which should be avoided unless the CO concentration in the
reactant stream is low enough, because it consumes rela-
tively large amounts of hydrogen. To develop PEMFC for
practical applications, it is thus necessary to find catalysts
that can be used to remove CO selectively in an H2-rich
system by oxidizing CO to CO2, while at the same time
being inactive for hydrogen oxidation. Typical catalysts
used in this reaction are gold or platinum nanoparticles
Carbon nanotubes (CNTs) constitute a new class of sup-
port that, due to their remarkable properties, may play an
active role in catalytic activity (including oxidation reac-
tions),[3] which should permit one to reach higher catalytic
activity and stability than with conventional supports.[4]
Thus, CNTs constitute an interesting alternative to conven-
tional supports due to (i) their high purity that can avoid
self-poisoning; (ii) the improved mechanical properties,
high electrical conductivity, and thermal stability of these
materials; (iii) the high accessibility of the active phase and
the absence of any microporosity, which eliminates diffusion
and intraparticle mass transfer in the reaction medium;
(iv) the possibility of macroscopic shaping of this support;
(v) the specific metal–support interactions that can directly
affect catalytic activity and selectivity; (vi) the possibility of
chemical modification of their surface by doping (for exam-
ple, with nitrogen) or grafting; and (vii) the possibility of
confinement effect in their inner cavity.[5–7]
Since CNTs are chemically inert, a pretreatment of the
surface is needed for metal anchoring that involves oxi-
dation of the CNTs by a strong oxidizing acid, or tedious
metal-catalyst synthetic procedures.[8] Commonly, the syn-
thetic methods of metal catalysts involve citrate or borohyd-
ride reduction of higher-oxidation-state precursors, and
only a few methods produce particles of uniform size.[9–12]
Jiang et al.[13,14] selectively attached gold nanoparticles (by
[a] Laboratoire de Chimie de Coordination,
UPR CNRS 8241 composante ENSIACET, Toulouse
University,
4 Allée Monso, B. P. 44362, Toulouse Cedex 4, France
Fax: +33-5-34323596
E-mail: philippe.serpensiacet.fr
[b] IU CINQUIMA/Química Inorgánica,
Facultad de Ciencias, Universidad de Valladolid,
47071 Valladolid, Spain
[c] Instituto de Catálisis y Petroleoquímica,
Unidad Asociada “Group for Design Appl. Het. Catal.”
UNED-CSIC,
C/Marie Curie 2, Cantoblanco, 28049 Madrid, Spain
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201000643.
View this journal online at
wileyonlinelibrary.com
5096
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 5096–5102

Selective Deposition of Gold Nanoparticles
a cationic polyelectrolyte) to chemically functionalized
CNTs. Covalent coupling of the activated CNTs with ionic
liquids and subsequent anion exchange were used to
achieve the assembly of controlled gold NPs on modified
CNTs.[15,16] Covalent coupling with (3-aminopropyl)trie-
thoxysilane or polyethyleneimine was also reported.[17–19]
Gold NPs on CNTs present often a broad particle-size dis-
tribution, typically 12–20 nm.[20] Theoretical studies have
concluded that Au interactions with the graphite surface are
weaker when compared to other metals.[21]
Thus, to take advantage of all the potentiality of Au/
CNT catalysts, significant effort should be devoted to the
synthesis of gold nanoparticles supported on CNTs, to the
control of the morphology of nanostructures, and to factors
that address these needs. In this paper we present different
synthetic procedures for the design of high-performance
Au/CNTs catalysts including (i) liquid-phase impregnation
from preformed AuNPs; (ii) decomposition of a gold com-
plex in the presence of modified CNTs so as to favor the
dispersion of small AuNPs; and (iii) deposition–precipi-
tation (DP) on functionalized and nitrogen-doped CNTs.
Furthermore, we report the catalytic activity of AuNPs de-
posited on CNTs for selective CO oxidation in H2-rich stre-
ams. Characterization of the CNT supports and the Au/
CNT catalysts was performed by using nitrogen adsorption,
TEM, infrared spectroscopy, thermogravimetric analysis
(TGA), elemental analysis, chemical titration, and X-ray
photoelectron spectroscopy (XPS). TEM has also been used
to determine the spatial location (on or inside CNTs) of
AuNPs in the different samples. The syntheses, structure–
property correlations, and relative catalytic activity of the
systems are presented, with our attention focused on the
Scheme 1. Procedures for the preparation of the catalytic systems.
Table 1 gives the Brunauer–Emmett–Teller (BET) sur-
face-area characterization, mean pore diameter, mean dia-
meter, and ash-content results for the different CNT sup-
ports used. The elemental and XPS analyses of the as-re-
ceived and functionalized supports are given in Table 3.
Table 1. Textural properties of as-received CNTs.
Sample
BET
Mean pore diameter
[nm]
dex
din
Ash
[%]
[m2 g–1]
[nm] [nm]
CNTP
CNTL
CNTN
38
180
292
16
13
10
80
14
52
40
7
25
7
8.9
7.5
As-received CNTP and CNTL present a small amount of
influence of the size of the gold NPs on the catalytic activity surface oxygen functionalities (Table 2). Nitric acid treat-
and the interaction of these nanoparticles with the relevant ment allows the introduction of polar hydrophilic surface
support.
groups, mainly –COOH but also phenol, carbonyl, and quin-
one.[22] In the case of CNTP, the duration of this treatment
was limited to 3 h to avoid internal surface functionaliza-
tion; this treatment produced 1 mmol of –COOH/gCNT (de-
termined by chemical titration[23]) and an ash content of
7%, as evidenced by the results of elemental analysis
(Table 2). CNTL was functionalized with nitric acid treat-
ment for 8 h. The concentration of surface –COOH groups
in CNTL2 is 0.011 mmolg–1, which points to a lower reac-
Results and Discussion
Synthesis and Characterization of the Supports
Commercial nanotubes CNTP were segmented by ball- tivity of CNTL surface compared to CNTP. The bulk quan-
milling to open both their ends. The surface functionaliza- titative (elemental analysis) and surface semiquantitative
tion of segmented CNTP and CNTL (laboratory-grown (XPS) analyses confirm the functionalization of CNTP2,
CNTs) with carboxylic groups, so as to provide good sta- CNTP3, and CNTL2 and the presence of the incorporated
bility to the NPs, was achieved by nitric acid oxidation, nitrogen in doped CNTN.
thereby yielding CNTP2 and CNTL2 (Scheme 1). A simple
The effective functionalization of CNTP2 and CNTL2,
and efficient method for the selective confinement of NPs and the grafting of the long alkyl chain on the CNTP3 sur-
in the inner cavity of CNTs is to use preformed NPs and face were also checked by IR spectroscopy (Figure 1). The
CNTs functionalized with a long alkyl chain.[5] To graft band at 1580 cm–1 is due to CNT skeletal in-plane vi-
long-chain alkyl-terminated moieties onto the CNTP sur- bration, and that at 1721 cm–1 is assigned to νasym(COOH)
face, CNTP2 was treated with thionyl chloride to produce of the carboxylic groups (from hydrolysis of some acetyl
acetyl chloride, and further treated with hexadecylamine chloride groups in the case of CNTP3). CNTP3 shows a
(HDA) to produce the amide surface groups of CNTP3 band at 1635 cm–1 that is consistent with the amide func-
(Scheme 1). Nitrogen-doped CNTs (CNTN) were used with- tion, and the aliphatic C–H stretching bands located be-
out further functionalization.
tween 2850 and 2950 cm–1 are also present.
Eur. J. Inorg. Chem. 2010, 5096–5102
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5097

P. Serp et al.
FULL PAPER
Table 2. Elemental and XPS analyses of as-received and function-
alized CNTs.
XPS [atom-%]
C
Elemental analysis [wt.-%]
O
N
C
O
N
CNTP
CNTP2
CNTP3
CNTL
CNTL2
CNTN
98.6
86.1
89.2
95
87.3
91.8
1.4
13.9
9.8
0.6
4.6
5.5
–
–
1
–
–
93.3
83.2
88.8
94
91.4
89.5
0.4
8
3.3
–
7.6
2.1
0.6
0.5
1.1
–
–
2.6
3.7
Figure 2. TEM images of preformed Au1NPs and histogram with
the size distribution of the preformed AuNPs.
Synthesis and Characterization of the Supported Catalysts
Three synthetic routes were applied for the preparation
of Au/CNT samples (Table 3 and Scheme 1). The first
method was a simple impregnation (IP) procedure of pre-
formed NPs from complex Au1 on CNTP3 in THF, an or-
ganic solvent that presents
a low surface tension
(26 mNm–1) and that should wet and penetrate inside the
CNTP3. The sample prepared in this way was named Au1/
CNTP3. The second method consists of the thermal decom-
position (DC) at 373 K of the complexes Au2 or Au3 in the
presence of CNTP3, thereby giving rise to Au2/CNTP3 and
Au3/CNTP3. This method is known to favor the dispersion
of small AuNPs on and in CNTP3. The third route consists
of the synthesis by deposition–precipitation (DP) starting
from a water/methanol (15:1) solution of Na[AuCl4] that is
added to CNTL2 or CNTN; the resulting precipitate was
Figure 1. Infrared spectra (KBr pellets) of (a) CNTP and (b) CNTL.
Preformed Nanoparticles
Gold(I) complexes that bear alkynyl and isocyanide li- dried at 373 K in air overnight. The corresponding catalysts
gands with long alkyl or alkoxy chains were used as precur- were labeled Au/CNTL2 and Au/CNTN.
sors for the nanoparticles. The new complex [Au(CϵC–
C6H4–C14H29)(CϵN–C6H4–O–C2H5)] (Au1) was prepared
Table 3. Nomenclature and characteristics of the catalysts.
by reaction of [CϵN–C6H4–O–C2H5] with [Au(CϵC–
C6H4–C14H29)]n. As reported for similar systems,[24] it dis-
Support
CNTP3
Method
Sample
plays liquid-crystal behavior and shows a smectic A meso-
phase (SmA) in the temperature range 415–433 K.
IP
Au1CNTP3
Au2/CNTP3
Au3/CNTP3
Au/CNTL2
Au/CNTN
DC
DC
DP
DP
Gold nanoparticles were obtained by thermal decompo-
sition of Au1 in toluene at 383 K. Figure 2 shows the TEM
images of preformed AuNPs by using Au1 as precursor. The
nanoparticles average diameter is (18.7Ϯ0.5) nm. The IR
spectra of the nanoparticles show a weak ν(CϵN) absorp-
tion at 2195 cm–1, which has been reported as characteristic
CNTL2
CNTN
TEM was performed to determine dispersion, particle
of p-alkoxyphenyl isocyanides bonded to gold nanopar- size, and spatial location of gold NPs in the different CNT
ticles,[25] thereby suggesting that these nanoparticles are sta- samples. Figures 3, 4, 5, and 6 show TEM images of synthe-
bilized by coordination of the isocyanide ligand.
sized AuNPs on and in CNTs.
5098
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 5096–5102

Selective Deposition of Gold Nanoparticles
Figure 5. TEM micrograph and histogram with the size distribution
of Au1/CNTP3
.
Figure 3. TEM micrograph and histogram with the size distribution
of Au2/CNTP3
.
Figure 6. TEM micrograph and histogram with the size distribution
of Au/CNTN.
Table 4. Catalyst characterization.
Sample
Loading[a] Au NP size[b] Surface Au mass[c] Au 4f7/2
Figure 4. TEM micrograph and histogram with the size distribution
of Au3/CNTP3
(wt.-%)
[nm]
[%]
[eV]
.
Au2/CNTP3
Au3/CNTP3
Au1CNTP3
Au/CNTL2
Au/CNTN
7
5
6
1
5
1.25
7
18
3.5
11
3
5
84.4
85.2
nd[d]
nd[d]
84.2
Au2/CNTP3 compounds show very small AuNPs of
1.25 nm anchored on or inside CNTP3 (Figure 3). The con-
finement of AuNPs is clearly nonselective, and most of the
NPs are visible on the external CNTs surface. The presence
of gold nanoparticles inside CNTs cannot be excluded since
3D TEM has not been done to quantify precisely the ratio
of confined particles to those on the surface.[5,7] However,
nd[d]
nd[d]
2
[a] Determined by TG measurements. [b] Determined by TEM.
[c] Determined by XPS measurements. [d] Not detected.
For Au3/CNTP3, larger AuNPs (7 nm) are nicely dis-
XPS data (see Table 4) indicate that most of the gold is on persed on CNTP3 (Figure 4).
the external surface. The very small mean particle size can
The use of preformed AuNPs with controlled-surface
be attributed to stabilization by the support and by the or- CNTP3 lead to confined large NPs of dm = 18.1 nm inside
ganic species formed upon ligand dissociation.
the functionalized CNTP3 (Figure 5).
Eur. J. Inorg. Chem. 2010, 5096–5102
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5099

P. Serp et al.
FULL PAPER
Finally, for Au/CNTN (Figure 6), in which the bamboo
structure does not allow the confinement, AuNPs with a
mean diameter of 11 nm are located on the external surface.
Table 4 shows the characterization data of the gold cata-
lysts. Thermogravimetric analyses (SI1 in the Supporting
Information) performed in air provide the determination of
the gold content (Table 4). The CNTP compounds show the
lowest gasification temperature because of their low crystal-
linity and higher defect content. Surface semiquantitative
(XPS) analyses (Table 4) allowed us to determine the Au
loading on the CNT surface and the location of the
NPs.[5,6,26] The differences in the gold loading obtained
from TGA (bulk analysis) and XPS (surface analysis) can
be understood as a consequence of the confinement of
AuNPs inside CNTs. Since XPS is a surface-analysis tech-
nique, the XPS data confirm the nonselective confinement
of Au2NPs inside CNTP3 and the selective confinement of
Au1NPs in the inner cavity of CNTP3. The XPS spectra (SI2
in the Supporting Information) detect Au 4f7/2 with binding
energies over 84 eV (Table 4). These are typical values for
Au0,[27] thereby indicating the formation of metallic AuNPs
on the side walls of CNTs.
Figure 7. CO transformed (below) and selectivity to CO2 among
possible reactions of O2 (above) versus temperature.
The XRD diffractograms of Au/CNTs (SI3 in the Sup-
porting Information) show typical peaks at 2θ = 25° that
can be assigned to CNT (002 plane), and the peak 2θ values role in the high activity of AuNPs. CNTN contains nitrogen
at 38.4, 44.6, and 64.8° are assigned to gold. The peak in- in different forms including pyridinic, pyrrolic, and quater-
tensity of (002) decreases for CNTP, thus indicating lower nary species, which improve the electronic density and the
nanotube crystallinity. The broad peaks in the XRD pattern basicity of the carbon nanotubes. It was earlier shown that
in the case of Au2 indicate that AuNPs are too small to these nitrogen atoms act as an important promoter of Ru
yield XRD signals. In the case of Au3 (average size 7 nm as nanoparticles and increase their activity in the ammonia
determined by TEM), the broadening of the peaks could be decomposition reaction.[29] In this sense, modification of
due to the presence of the organic stabilizer (diffraction the electronic structure of CNTs to include electron-donor
peaks at low 2θ values). For higher particle size (Au1), the states near the conduction-band edge was reported for
presence of the ligand has no effect on the XRD diffracto- CNTN synthesized by pyrolyzing ferrocene/melamine mix-
grams.
tures.[30] Gong et al.[31] have recently observed that the in-
corporation of electron-accepting nitrogen atoms appears
to impart a relatively high positive-charge density. Dom-
mele et al.[32] demonstrated that nitrogen-containing carbon
nanotubes display basic properties. For Au/CNTN, even
though the mean particle size is quite large (11 nm), it con-
Catalytic Activity
The catalytic performance of Au1/CNTP3, Au2/CNTP3, tains a significant population of AuNPs Ͻ 6 nm that could
Au3/CNTP3, Au/CNTL2, and Au/CNTN was evaluated for explain the high activity.
the preferential oxidation of CO in an H2-rich stream.
Finally, the results of the catalytic oxidation of CO from
Figure 7 shows the CO oxidation performance of the cat- Au2/CNTP3 are surprising in the light of the very small par-
alysts at constant gas velocity. Au1/CNTP3 and Au/CNTL2 ticle size measured for this sample. Indeed, it is known that
exhibited almost no catalytic activity toward CO oxidation. very small gold nanoparticles are particularly active at low
For Au1/CNP3, this is attributed to its large particle size. temperature for this reaction.[2] This lack of reactivity is
For Au/CNTL2, in which AuNPs are mostly confined in the perhaps due to the presence of an organic stabilizer coating
CNT cavity, it is difficult to predict the reactivity,[6] and the gold particles. At higher temperatures, after thermal de-
its low activity might be due to the combined effects of composition of this stabilizer, the NPs should start to be-
confinement and the low Au loading (1% w/w). In fact, the come more active.
study of Au catalysts on different supports[28] has confirmed
TEM images (Figure 8) obtained for Au2/CNTP3 after
that the catalytic activity of supported Au depends on the reaction provide evidence of the sintering process between
size of AuNPs[2] and is influenced by the interaction be- neighboring NPs due to the thermal decomposition of the
tween Au and the support.
stabilizer during the reaction process. Interestingly, the sin-
Au/CNTN exhibited high catalytic activity and high sta- tering is more pronounced for AuNPs located inside CNTP3
bility; the catalytic performances remained unchanged after than on their external surface. In the case of AuNPs on
48 h. It is thus established that CNTN must play an essential CNTN, no sintering was observed after reaction.
5100
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 5096–5102

Selective Deposition of Gold Nanoparticles
Experimental Section
Preparation of Carbon Nanotube Supports: The pristine multiwalled
nanotubes CNTp were obtained from Applied Science (Pyro-
graph III, 98% purity, 2% remaining iron catalysts), which were
segmented (0.1–1 µm) by a ball-milling technique. Multiwalled
nanotubes CNTL (98% purity, 2% iron catalyst) were synthesized
by chemical vapor deposition according to previously reported pro-
cedures.[34] Nitrogen-doped nanotubes CNTN have been produced
in a fluidized bed reactor by catalytic chemical vapor deposition
on Fe/SiO2 catalysts from acetylene/ammonia mixtures at
1023 K;[29] they present a bamboo-like structure with graphene lay-
ers perpendicular to the CNT axis. CNTP and CNTL were func-
tionalized on the surface with carboxylic acids by nitric acid oxi-
dation to yield CNTP2 and CNTL2; according to reported pro-
cedures[5] CNTP were also functionalized with amide long-alkyl-
chain-terminated groups. A typical preparation is as follows: car-
bon nanotubes functionalized on the surface with carboxylic acid
groups (CNTP2) were treated with thionyl chloride to produce the
corresponding acetyl chloride derivative, and further treated with
hexadecylamine (HDA), thereby giving rise to the amide surface
groups of the carbon nanotubes (CNTP3).
Figure 8. TEM micrographs of Au2/CNTP2 after reaction.
Gold Precursors: H[AuCl4]·3H2O was purchased from Sigma–Ald-
rich. Literature methods were used to prepare [AuCl(tht)][34] (tht =
tetrahydrothiophene), [HCϵCC6H4–C14H29],[35] CϵNC6H4(4-
OC2H5), [Au(CϵC–C6H4–C8H17)(CϵN–C6H4–O–C10H21)] (Au2),
and [Au(CϵC–C6H4–C10H21)(CϵN–C6H4–O–C2H5)] (Au3).[24]
[Au(CϵC–C6H4–C14H29)(CϵN–C6H4–O–C2H5)] (Au1), not pre-
viously reported, was prepared as the analogous Au3, starting from
[HCϵCC6H4–C14H29]. It was isolated as a white solid. Yield:
0.55 g, 84%. 1H NMR (CDCl3): δ = 7.45 (d, J = 8.7 Hz, 2 H,
CϵNC6H4O–), 7.40 (d, J = 7.9 Hz, 2 H, CϵCC6H4), 7.06 (d, J =
7.9 Hz, 2 H, CϵCC6H4), 6.94 (d, J = 8.7 Hz, 2 H, CϵNC6H4O),
4.08 (c, J = 7.0 Hz, 2 H, CϵNC6H4OCH2–), 2.57 (t, J = 7.0 Hz, 2
H, CϵCC6H4CH2), 1.56 (m, 2 H, CϵCC6H4CH2CH2), 1.44 (t, J =
7.0 Hz, 3 H, CϵNC6H4OCH2CH3), 1.26 (m, 24 H, acetylide alkyl
chains), 0.89 [t, J = 6.6 Hz, 3 H, CH3 (acetylide)] ppm. IR (KBr):
At room temperature, the selectivity of Au/CNTN and
Au/CNTL2 reached 100%. At higher temperatures, both
conversion and selectivity decreased due to competition be-
tween CO and H2 oxidation. In contrast to this behavior,
the catalysts prepared with CNTP3 as support show higher
conversion and selectivity at 420 K. Thus, the oxidation rate
of H2 is not independent of the support materials and the
nature of the gold precursor. The organometallic precursors
that contain long alkyl or alkoxy chains, which produce
upon decomposition organic stabilizing species, yield cata-
lysts that show higher CO selectivity at higher temperature.
ν
= ν =
2215 cm–1 (CϵN); (CH2Cl2): 2214 cm–1 (CϵN).
˜
˜
C31H42AuNO (641.65): calcd. C 58.03, H 6.60, N 2.18; found C
58.06, H 6.43, N 2.20. DSC/transition, K [kJmol–1] (C = crystal
phase; Sm = smectic phase; I = isotropic liquid): C–C1, 352 (19.9);
C1–C2, 364 (3.0); C2–SmA, 415 (32.8); SmA–I (dec.), 433 (micro-
scopy data).
Conclusions
Synthesis of Preformed Gold Nanoparticles: Complex Au1 (0.046 g,
0.072 mmol) was dissolved in toluene (6 mL), and the colorless
solution was heated at 383 K for 14 h to give a red solution. Then
the solvent was removed under vacuum to give a brown solid. IR
Until now, the techniques used for depositing AuNPs on
activated carbon or CNTs required tedious procedures, and
only a few methods produce particles of uniform size. We
have investigated three methods (impregnation, decomposi-
tion, and deposition–precipitation) to control the loading
of gold NPs with a narrow particle-size distribution and to
conveniently attach AuNPs onto the walls of CNTs. The
(KBr): ν = 2195 cm–1 (CϵN).
˜
Preparation of Supported Gold Catalysts: Three methods were used
(see Table 3 and Scheme 1). Method A (Impregnation Procedure,
IP): A red solution of AuNPs (obtained from 0.046 g of Au1) in
use of preformed AuNPs with controlled surface leads to tetrahydrofuran (1 mL) was poured slowly into a flask that con-
tained CNTs (0.05 g). The resulting pasty mixture was stirred for
30 min and then sonicated for a further 30 min. Both steps, stirring
and subsequent sonication, were repeated three times. Then the
mixture was stirred overnight and the solvent evaporated under
vacuum. The sample prepared was labeled Au1/CNTP3. Method B
(Decomposition Procedure, DC): A mixture of Aun (n = 2, 3;
0.072 mmol) and CNTP3 (0.05 g) in tetrahydrofuran (1 mL) was
successively stirred and sonicated as described in Method A. After
3 h of alternating stirring and ultrasonic treatment at room tem-
perature, the mixture was stirred overnight and then dried at 373 K
in air to give Aun/CNTP3. Method C (Deposition–Precipitation Pro-
NPs confined inside the functionalized CNTs. This is an
efficient method for the selective confinement in CNTs.
Our results also show that bamboo-like nitrogen-doped
carbon nanotubes present interesting potential for the de-
position of AuNPs by the deposition–precipitation
method.[33] Additionally, we have shown that CNTs doped
with nitrogen are a promising support material for gold-
catalyzed low-temperature CO oxidation. Finally, the com-
petition between CO and H2 oxidation depends on the na-
ture of both the support and the gold precursor.
Eur. J. Inorg. Chem. 2010, 5096–5102
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
5101

P. Serp et al.
[6] P. Serp, E. Castillejos, ChemCatChem 2010, 2, 41–47.
[7] J.-P. Tessonnier, O. Ersen, G. Weinberg, C. Pham-Huu, D. Su,
R. Schlögl, ACS Nano 2009, 3, 2081–2089.
[8] C. H. Walker, J. V. St. John, P. Wisian-Neilson, J. Am. Chem.
Soc. 2001, 123, 3846–3847.
[9] R. Zhang, M. Hummelgard, H. Olin, Mater. Sci. Eng. B 2009,
158, 48–52.
FULL PAPER
cedure, DP): A solution of [HAuCl4] in a mixture of water/meth-
anol (15:1) was treated with an aqueous solution of Na2CO3 (0.2 )
until the pH reached 11. After stirring for 1 h, the solution was
added to CNTL2 or CNTN. After 4 h of alternating stirring and
ultrasonic treatment at room temperature as described above, the
precipitate was collected by filtration and washed with deionized
water until no Cl– was detected. The samples were dried at 373 K
in air overnight; the corresponding catalysts were labeled Au/
CNTL2 and Au/CNTN.
[10] S. Y. Moon, T. Kusunose, S.-I. Tanaka, T. Sekino, Carbon 2009,
47, 2924–2932.
[11] Y. Shi, R. Yang, P. K. Yute, Carbon 2009, 47, 1146–1151.
[12] R. Zhang, X. Wang, Chem. Mater. 2007, 19, 976–978.
[13] K. Jiang, A. Eitan, L. S. Schadler, P. M. Ajayan, R. W. Siegel,
Nano Lett. 2003, 3, 275–277.
[14] L. Jiang, L. Gao, Carbon 2003, 41, 2923–2929.
[15] Y. Gu, X. Hou, H. Hu, B. Yu, L. Wang, F. Zhou, Mater. Chem.
Phys. 2009, 116, 284–288.
Instrumentation and Measurements: The BET surface area of the
support was measured from the nitrogen-adsorption isotherm with
an ASAP instrument. TEM images were measured with a JOEL
1011 electron microscope operated at 110 kV. At least 200 particles
were used to determinate the mean diameter of the Au-particle
NPs. XRD patterns were obtained with a scanning Seifert XRD
3000P diffractometer. XPS analysis was performed with an ESCA-
PROBE P (Omicron) spectrometer by using nonmonochromatized
Mg-K radiation (1253.6 eV). IR spectra were recorded with a Per-
[16]
X. Hou, L. Wang, F. Zhou, L. Li, Z. Li, Mater. Lett. 2009, 63,
697–699.
[17] X. Hou, L. Wang, F. Zhou, F. Wang, Carbon 2009, 47, 1209–
1213.
[18] G.-H. Jeong, S. Suzuki, Y. Kobayashi, Nanotechnology 2009,
1
kin–Elmer FT BX instrument and H NMR spectra (room temp.,
20, 1–6.
CDCl3) with Bruker AC 300 or ARX 300 instruments. Combustion
analyses and mesophase identification were made as reported else-
where.[36] The catalytic experiments for CO oxidation were carried
out with a continuous-flow fixed-bed reactor with 0.05 g of catalyst
{contact time (w/F) [w = weight (catalyst); F = flow rate (feed
stream)] of 19 gcat/hmolCO}. The gas mixture consisted of 1% CO,
1% O2, and 40% H2 in He at 1ϫ10–2 cm3 min–1, and the products
were analyzed by gas chromatography using a carboxen 1000 col-
umn. The selectivity and conversion values were calculated accord-
ing to Equations (1) and (2), in which X and S are percentages of
conversion and selectivity, respectively, and F is the inlet and outlet
molar flow of the indicated gas.
[19] X. G. Hu, T. Wang, X. H. Qu, S. J. Dong, J. Phys. Chem. B
2006, 110, 853–857.
[20] N. R. Jana, L. Gearheart, C. J. Murphy, Langmuir 2001, 17,
6782–6786.
[21] Y. Matsuda, W. Q. Deng, W. A. Goddard, J. Phys. Chem. C
2007, 111, 11113–11116.
[22] T. I. T. Okpalugo, P. Papakonstantino, H. Murphy, J.
McLaughlin, N. M. D. Brown, Carbon 2005, 43, 153–161.
[23] M. L. Toebes, J. M. P. van Heeswijk, J. H. Bitter, A.
Jos van Dillen, K. P. de Jong, Carbon 2004, 42, 307–315.
[24] P. Alejos, S. Coco, P. Espinet, New J. Chem. 1995, 19, 799–805.
[25] S. Kim, K. Ihm, T.-H. Kang, S. Hwang, S.-W. Joo, Surf. Inter-
face Anal. 2005, 37, 294–299.
[26] F. Winter, G. Leendert Bezemer, C. vand der Spek, J. D. Meeld-
ijk, A. J. van Dillen, J. W. Geus, K. P. de Jong, Carbon 2005,
43, 327–332.
in
out
XCO = (F – F /gAu
)
(1)
(2)
CO
CO
in
out
SCO = [(F – F )/2(FiOn – FoOut)]100
CO
CO
2
2
2
[27] Z.-M. Qi, H.-S. Zhou, N. Matsuda, I. Honma, K. Shimada,
A. Takatsu, K. Kato, J. Phys. Chem. B 2004, 108, 7006–7011.
[28] I. Suarez-Martinez, C. Bittencourt, X. Ke, A. Felten, J. J. Pi-
reaux, J. Ghijsen, W. Drube, G. Van Tendeloo, C. P. Ewels, Car-
bon 2009, 47, 1549–1554.
Supporting Information (see footnote on the first page of this arti-
cle): Thermogravimetric, XPS and XRD analyses of selected sam-
ples.
[29] F. R. Garcia-Garcia, J. Alvarez-Rodríguez, I. Rodríguez-
Ramos, A. Guerrero-Ruiz, Carbon 2010, 48, 267–276.
[30] R. Czerw, M. Terrones, J.-C. Charlier, X. Blase, B. Foley, R.
Kamalakaran, N. Grobert, H. Terrones, D. Tekleab, P. M.
Ajayan, W. Blau, M. Rühle, D. L. Carroll, Nano Lett. 2001, 1,
457–460.
Acknowledgments
E. C. gratefully acknowledges financial support from the Spanish
Ministry of Science and Technology and the Institute of Catalysis
and Petroleochemistry (Spain). This work was partly supported by
the Spanish Government (contract CTQ2008-06839-C03-01/PPQ
and CTQ2008-06839-C03-03/PPQ, CTQ2008-03954/BQU and
MAT2008-06522-C02-02), INTECAT Consolider Ingenio 2010
(CSD2006-0003), and the Junta de Castilla y León (Project
VA012A08 and GR169).
[31] K. Gong, F. Du, Z. Xia, M. Durstock, L. Dai, Science 2009,
323, 760–763.
[32] S. van Dommele, K. P. de Jong, J. H. Bitter, Chem Commun.
2006, 4859–4861.
[33] K. R. Souza, A. F. F. de Lima, F. F. De Sousa, L. G. Appel,
Appl. Catal. A: Gen. 2008, 340, 133–139.
[34] R. Philippe, A. Morançais, M. Corrias, B. Caussat, Y. Kihn, P.
Kalck, D. Plee, P. Gaillard, D. Bernard, P. Serp, Chem. Vap.
Deposition 2007, 13, 447–457.
[35] R. Usón, A. Laguna, M. Laguna, J. Jiménez, M. P. Gómez, A.
Sainz, P. G. Jones, J. Chem. Soc., Dalton Trans. 1990, 3457–
3463.
[1] E.-Y. Ko, E. D. Park, H. C. Lee, D. Lee, S. Kim, Angew. Chem.
Int. Ed. 2007, 46, 734–737.
[2] M. Haruta, Chem. Rec. 2003, 3, 75–87.
[3] C.-Y. Lu, M.-Y. Wey, Fuel 2007, 86, 1153.
[4] P. Serp, in Carbon Materials for Catalysis (Eds.: P. Serp, J. L.
Figueiredo), Wiley, Hoboken, 2009, pp. 309–372.
[5] E. Castillejos, P. J. Debouttière, L. Roiban, A. Solhy, V. Marti-
nez, Y. Kihn, O. Ersen, K. Philippot, B. Chaudret, P. Serp,
Angew. Chem. Int. Ed. 2009, 48, 2529–2533.
[36] P. Herwig, C. W. Kayser, K. Miillen, H. W. Spiess, Adv. Mater.
1996, 8, 510–513.
Received: June 9, 2010
Published Online: September 29, 2010
5102
www.eurjic.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2010, 5096–5102