A facile strategy to support palladium nanoparticles on carbon nanotubes, employing polyvinylpyrrolidone as a surface modifier

Article abstract of DOI:10.1002/ejic.201301585

This paper describes the development of a facile and environmentally friendly strategy for supporting palladium nanoparticles (Pd NPs) on multiwalled carbon nanotubes (MWCNTs) with ethanol as the solvent/reducing agent, potassium tetrachloropalladate(II) (K₂PdCl₄) as the Pd precursor, and polyvinylpyrrolidone (PVP) as the surface modifier. More specifically, our approach was based on the nonbonding interaction between MWCNTs and PVP, which is a cheap, nontoxic, and commercial polymer. As PVP can serve not only as a surface modifier but also as a stabilizing agent for Pd NPs, the utilization of additional functionalization steps, reducing agents, and stabilizers was not required to achieve uniform Pd deposition over the MWCNTs. Our results demonstrate that Pd NPs below 5 nm in diameter can be directly supported on MWCNTs by this route. Also, sequential Pd-reduction steps can be employed to improve the coverage of Pd NPs at the MWCNT surfaces, although this can also lead to the formation of larger Pd particles or aggregates. The electrocatalytic activity for ethanol oxidation was investigated as a function of the composition and structure of the materials produced, in which MWCNTs decorated with Pd NPs of smaller sizes and lower coverages displayed the highest activities. The results described herein suggest that our approach may serve as a simple platform for the synthesis of MWCNTs decorated with metal NPs with well-defined morphologies and uniform dispersion for electrochemical and catalytic applications. The nonbonding interaction between polyvinylpyrrolidone (PVP) and multiwalled carbon nanotubes (MWCNTs) was employed to achieve deposition of palladium nanoparticles less than 5 nm in size over the surface of the MWCNTs; with ethanol as the solvent/reducing agent, potassium tetrachloropalladate(II) as the precursor, and PVP as the surface modifier. Copyright

Full text of DOI:10.1002/ejic.201301585

FULL PAPER
DOI:10.1002/ejic.201301585
A Facile Strategy to Support Palladium Nanoparticles on
Carbon Nanotubes, Employing Polyvinylpyrrolidone as a
Surface Modifier
Mariana B. T. Cardoso,[a] Edward Lewis,[b] Pollyana S. Castro,[a]
Luiza M. F. Dantas,[a] Caio C. S. De Oliveira,[a] Mauro Bertotti,[a]
Sarah J. Haigh,*[b] and Pedro H. C. Camargo*[a]
COVER PICTURE
Nanoparticles / Nanotubes / Polymers / Palladium / Electrocatalysis
This paper describes the development of a facile and envi-
ronmentally friendly strategy for supporting palladium nano-
particles (Pd NPs) on multiwalled carbon nanotubes
(MWCNTs) with ethanol as the solvent/reducing agent, po-
tassium tetrachloropalladate(II) (K2PdCl4) as the Pd precur-
sor, and polyvinylpyrrolidone (PVP) as the surface modifier.
More specifically, our approach was based on the nonbond-
ing interaction between MWCNTs and PVP, which is a
cheap, nontoxic, and commercial polymer. As PVP can serve
not only as a surface modifier but also as a stabilizing agent
for Pd NPs, the utilization of additional functionalization
steps, reducing agents, and stabilizers was not required to
achieve uniform Pd deposition over the MWCNTs. Our re-
sults demonstrate that Pd NPs below 5 nm in diameter can
be directly supported on MWCNTs by this route. Also, se-
quential Pd-reduction steps can be employed to improve the
coverage of Pd NPs at the MWCNT surfaces, although this
can also lead to the formation of larger Pd particles or aggre-
gates. The electrocatalytic activity for ethanol oxidation was
investigated as a function of the composition and structure of
the materials produced, in which MWCNTs decorated with
Pd NPs of smaller sizes and lower coverages displayed the
highest activities. The results described herein suggest that
our approach may serve as a simple platform for the synthe-
sis of MWCNTs decorated with metal NPs with well-defined
morphologies and uniform dispersion for electrochemical
and catalytic applications.
ties.[7–10] These features make them attractive as supports
for the immobilization of Pd NPs for catalytic/electrocata-
lytic applications. In general, the synthesis of carbon nano-
tubes (CNTs) decorated with Pd NPs by solution phase
methods requires either covalent or noncovalent function-
alization of the CNTs with specific groups to enable the
manipulation/dispersion of CNTs in certain solvents as well
as to provide sites for the immobilization of Pd ions, fol-
lowed by Pd nucleation and growth after a reduction
step.[5,11–14] However, covalent modification strategies can
not only damage the CNT surfaces (by the inclusion of sur-
face defect sites that lead to detrimental effects on the prop-
erties of the CNTs), but also lead to a nonuniform distribu-
tion of functional groups along the CNT surfaces.[11,12]
Moreover, many procedures based on covalent modification
involve toxic reagents and/or harsh conditions.[11–15]
Introduction
Among noble metals, palladium is a key catalyst compo-
nent in several industrial processes and organic synthesis
protocols.[1,2] For instance, palladium nanoparticles (Pd
NPs) display exceptional catalytic activities for carbon–
carbon (Suzuki and Heck) coupling reactions.[3,4] Other ex-
amples include the electrochemical alcohol and formic acid
oxidation reactions, which are relevant in the context of fuel
cells.[5,6] In these applications, the utilization of solid sup-
ports is often employed to facilitate the recovery and im-
prove durability of the catalysts.[5]
Single-walled and multiwalled carbon nanotubes
(SWCNTs and MWCNTs, respectively) display high surface
areas, high thermal/electrical conductivities, excellent me-
chanical properties, and chemical/electrochemical stabili-
On the other hand, methodologies based on noncovalent
interactions may preserve the structural integrity of the
CNTs and enable their homogeneous functionaliza-
tion.[12,16] Despite these advantages, the synthesis of CNTs
decorated with Pd NPs by this approach, in which the com-
position and structure can be tightly controlled, is still lim-
ited and remains challenging. Moreover, many procedures
based on noncovalent modifications involve the use of extra
reducing agents, organic and toxic solvents, and/or rela-
[a] Departamento de Química Fundamental, Instituto de Química,
Universidade de São Paulo
Av. Prof. Lineu Prestes, 748, 05508-000, São Paulo-SP, Brazil
E-mail: camargo@iq.usp.br
http://www.pedrocamargo.org
[b] School of Materials, University of Manchester,
Manchester M13 9PL, UK
E-mail: sarah.haigh@manchester.ac.uk
http://www.materials.manchester.ac.uk
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejic.201301585.
Eur. J. Inorg. Chem. 2014, 1439–1445
1439
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
tively sophisticated surface modifiers.[13,16–19] Thus, the de- PVP–Pd1). We then decided to perform successive deposi-
velopment of simple strategies for the synthesis of CNTs tion steps in order to investigate whether this procedure
decorated with a high density of Pd NPs less than 10 nm would enable us to control the coverage and size of the Pd
in size displaying controlled compositions and structures, NPs over the MWCNT surfaces. In this case, the CNT–
obtained under mild conditions by employing nonhazard- PVP–Pd1 sample was employed as a template for further
2–
ous solvents/reactants, is central to meet the requirements Pd deposition by the addition of another PdCl4 aliquot
of green chemistry.[20]
to the reaction mixture, to yield the sample CNT–PVP–Pd2
This paper describes the development of a facile and en- (Scheme 1D). Similarly, sample CNT–PVP–Pd2 was em-
vironmentally friendly strategy for preparing Pd NPs sup- ployed as a template in a third deposition step by the ad-
2–
ported on MWCNTs, employing ethanol as the solvent/re- dition of another PdCl4 aliquot to the reaction mixture,
ducing agent, potassium tetrachloropalladate(II) (K2PdCl4) to yield the sample CNT–PVP–Pd3 (Scheme 1E).
as the Pd precursor, and polyvinylpyrrolidone (PVP) as the
Figures 1A–F show SEM and high angle annular dark
surface modifier. More specifically, our approach is based field scanning transmission electron microscopy (STEM-
on the nonbonding interaction between MWCNTs and HAADF) images for samples CNT–PVP–Pd1 (Figures 1A
PVP,[21–24] which is a cheap, nontoxic, and commercial and B), CNT–PVP–Pd2 (Figures 1C and D), and CNT–
polymer. As PVP can serve not only as a surface modifier PVP–Pd3 (Figures 1E and F), that confirm that the em-
for MWCNTs but also as a stabilizing agent for Pd NPs, ployed procedure led to Pd NPs supported over the surfaces
the utilization of additional functionalization steps, reduc- of the MWCNTs. SEM images show that the deposited Pd
ing agents, and stabilizers was not required to achieve the NPs were uniformly dispersed over the MWCNT surfaces
uniform deposition of Pd NPs over the MWCNT surfaces. in all samples. Also, it is clear that the successive reduction
steps led to an increase in the amount of Pd deposited at
the MWCNT surfaces. This result is in agreement with the
increase in Pd content in the hybrid materials after the suc-
Results and Discussion
Our approach for the decoration of MWCNTs with Pd cessive reduction steps. Inductively coupled plasma mass
NPs started with the preparation of an ethanolic suspension spectrometry (ICP-MS) analyses determined the Pd weight
containing 35 mg of PVP and 0.056 g of MWCNTs (the percentages in the CNT–PVP–Pd materials to be 2.8, 9.0,
PVP/MWCNT molar ratio was 66.5:1 based on the PVP and 28.7% for samples CNT–PVP–Pd1, CNT–PVP–Pd2,
repeating unit), which yielded the material labeled as CNT– and CNT–PVP–Pd3, respectively. It is plausible that this Pd
PVP, as shown in Scheme 1A and B. It has been established increase may be due to a boost in the autocatalytic re-
that PVP, as well as other polymers and polyelectrolytes, duction of PdCl42–, which can be favored in the presence of
can interact with the MWCNT surfaces by means of non- an increased amount of Pd0 in the MWCNTs during the
bonding interactions.[16,18,19,21–24] In the case of PVP, it has successive deposition steps.[25]
been proposed that this procedure leads to the wrapping
In order to gain further insight into the morphological
of the PVP over the MWCNT surfaces, making the CNTs features of the materials after the successive reduction steps,
hydrophilic and thus enabling their dispersion in eth- the materials CNT–PVP–Pd1, CNT–PVP–Pd2, and CNT–
anol.[21–24] As PVP can also act as a reducing/stabilizing PVP–Pd3 were studied by STEM-HAADF (Figures 1B, D,
agent in the synthesis of Pd NPs, we were particularly inter- and F, respectively). STEM-HAADF analysis for the sam-
ested in investigating whether the PVP present on the ple obtained after the first deposition step (Figure 1B) re-
MWCNT surfaces could favor the uniform nucleation and vealed that MWCNTs were decorated with a large number
growth of Pd NPs on the MWCNT surfaces. To this end, of spherical Pd NPs 3.3Ϯ0.7 nm in size. Interestingly, the
the CNT–PVP material was directly employed as a template coverage of NPs over the MWCNT surfaces was not com-
for Pd deposition with PdCl42– as the Pd precursor and eth- plete, which suggested that Pd nucleation and growth took
anol as the solvent at 75 °C (Scheme 1C, sample CNT– place at specific sites that were uniformly distributed along
Scheme 1. Approach for the synthesis of MWCNTs decorated with NPs by employing PVP as a surface modifier. After the MWCNTs
(A) were functionalized with PVP (CNT–PVP, B), they serve as seeds for the deposition of Pd NPs less than 5 nm in size at their surface,
with PdCl42– (aq) as the Pd precursor, PVP as the stabilizer, and ethanol as the solvent, at 75 °C (C). Then, second and third Pd-deposition
steps were carried out by the addition of new aliquots of PdCl42– (aq) to the reaction mixture (D and E, respectively).
Eur. J. Inorg. Chem. 2014, 1439–1445
1440
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
STEM-HAADF images (Figure 1F) for the sample ob-
tained after the third reduction step (CNT–PVP–Pd3,
Scheme 1E) are in agreement with these observations. The
presence of Pd NPs less than 5 nm (3.8Ϯ2.0 nm) in dia-
meter as well as larger particles/aggregates up to 35 nm in
size could be detected at the MWCNT surfaces. These
larger particles/aggregates increased in size relative to the
sample obtained after the second reduction step (CNT–
PVP–Pd2) as a result of further Pd deposition. In this case,
the larger Pd particles/aggregates formed during the second
reduction step could serve as sites for further Pd deposition.
It is important to note that some Pd NPs less than 5 nm in
size could also serve as nucleation sites for Pd nucleation
and growth at this stage, increasing the size distribution of
the Pd NPs at the CNT–PVP surfaces.
HRTEM images for the samples obtained after the first
(CNT–PVP–Pd1, Figures 2A and B) and third (CNT–PVP–
Pd3, Figures 2C and D) reductions confirmed the presence
of Pd NPs less than 5 nm in size at the MWCNT surfaces.
Phase contrast images (Figures 2B and D) and the fast Fou-
rier transform (FFT) patterns (insets) revealed the 0.22 and
0.19 nm d spacings assigned to Pd {111} and {002} with a
face-centered cubic structure. The 0.34 nm d spacing due to
the MWCNTs could also be observed in the FFT pattern
shown as an inset in Figure 2B. Elemental maps obtained
Figure 1. SEM (A, C, and E) and STEM-HAADF (B, D, and F)
images of CNT–PVP–Pd materials that were obtained after the
first (A and B), second (C and D), and third (E and F) Pd–deposi-
tion steps (samples CNT–PVP–Pd1, CNT–PVP–Pd2, and CNT–
PVP–Pd3, respectively).
the CNT–PVP surface, but that there was insufficient Pd to by energy-dispersive X-ray (EDX) spectroscopy for the
cover the entire surface at this stage of growth. Figure 1D sample obtained after the third deposition step (CNT–
shows the STEM-HAADF image for the material after the PVP–Pd3, Figure 3A–D) are in agreement with the afore-
second Pd-deposition step (CNT–PVP–Pd2, Scheme 1D). mentioned results. It can be observed that Pd was clearly
As observed for sample CNT–PVP–Pd1, the presence of Pd present at the larger particles/aggregates as well as over the
NPs less than 5 nm (3.3Ϯ1.7 nm) in size at the MWCNT entire MWCNT surfaces. The iron content shown inside the
surface could also be detected in this material. After the MWCNTs in Figure 3D was due to its use as a catalyst
second reduction step, an increase in the coverage of the during the MWCNT synthesis.
MWCNTs with Pd NPs relative to the sample obtained af-
ter the first deposition step could be observed. In fact, Pd
NPs seemed to completely cover the MWCNT surfaces. In
addition to the presence of Pd NPs less than 5 nm in size
(3.3Ϯ1.7 nm), larger particles/aggregates approximately 5–
14 nm in diameter were also detected over the MWCNTs.
This observation reveals that, during the second reduction
step, nucleation and growth of Pd NPs at the surfaces of
the MWCNTs continued to take place until they formed
close to a complete layer at the MWCNT surface. This ob-
servation suggests that Pd nucleation at the CNT–PVP sur-
face was favored over nucleation at, or growth of, the prede-
posited Pd NPs (from the first deposition step). Then, after
the formation of a complete layer of Pd NPs at the
MWCNT surface, further Pd nucleation and growth took Figure 2. (A–D) HRTEM images for the samples obtained after
the first (A and B, CNT–PVP–Pd1) and third (C and D, CNT–
place at the deposited Pd NPs, leading to the formation of
PVP–Pd3) Pd-deposition steps. Phase-contrast images (B and D)
larger particles and aggregates distributed along the sur-
and FFT patterns (insets) revealed the 0.22 and 0.19 nm d spacings
assigned to Pd {111} and {002} and the 0.34 nm d spacing from
faces of the MWCNTs (second deposition step). Here,
rather than the uniform growth of predeposited Pd NPs
(that would be expected if further Pd nucleation and growth
took place uniformly along the predeposited Pd NPs), the
the MWCNTs (B).
We then employed Raman spectroscopy in order to in-
formation of larger particles/aggregates distributed along vestigate the structural integrity of the MWCNTs after their
the MWCNT surfaces was observed, indicating that further functionalization with PVP and subsequent Pd deposition
Pd deposition took place only at specific sites over the sur- after the various deposition steps. Raman spectroscopy is a
faces (Scheme 1D).
powerful tool to investigate the structure of MWCNTs as it
Eur. J. Inorg. Chem. 2014, 1439–1445
1441
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Figure 4. Raman spectra recorded for the starting CNT sample
(black trace), CNT after nonbonding functionalization with PVP
(CNT–PVP, red trace), CNT–PVP–Pd1 (after first Pd deposition,
blue trace), CNT–PVP–Pd2 (after second Pd deposition, green
trace), and CNT–PVP–Pd3 (after third Pd deposition, magenta
trace).
Figure 3. (A) STEM-HAADF image for sample CNT–PVP–Pd3
and elemental maps (B–D) obtained by EDX spectroscopy, show-
ing the distribution of Pd (B), C (C), and Fe (D). The scale bar in
(B) applies to all images.
towards ethanol oxidation in alkaline media. All electrocat-
alytic tests employed the same mass of Pd loaded onto the
is well established that the intensity ratio between the D and electrode. Figure 5A shows the cyclic voltammograms for
G bands in the Raman spectrum can be used to evaluate the samples CNT–PVP–Pd1, CNT–PVP–Pd2, and CNT–PVP–
density of defects (and structural integrity).[26] In this case, Pd3 as black, red, and blue traces, respectively, recorded in
while the D band is associated with the presence of disorder a 1.0 m KOH solution. The characteristic features of Pd in
(structural defects), the G band is assigned to planar KOH solution could be clearly identified in the voltammog-
vibrations of carbon atoms.[26,27] Therefore, the inclusion of ram.[5,30–33] During the potential sweep in the supporting
defect sites usually leads to an increase of the D band inten- electrolyte solution, the forward scan showed a very broad
sity in relation to the G band.[26,27] Figure 4 shows the oxidation wave starting at approximately –0.2 V, while a
Raman spectra recorded for the starting CNT sample, much narrower and more intense reduction peak was ob-
CNT–PVP (CNT after nonbonding functionalization with served in the reverse scan with an onset potential of approx-
PVP), CNT–PVP–Pd1 (after the first Pd-deposition step), imately –0.2 V and a peak at –0.4 V. The position and shape
CNT–PVP–Pd2 (after the second Pd-deposition step), and of this reduction wave are consistent with the reduction of
CNT–PVP–Pd3 (after the third Pd-deposition step) as PdII species, either to PdO or Pd(OH)2.[5,30–33] By compar-
black, red, blue, green, and magenta traces, respectively. It ing the cyclic voltammograms of CNT–PVP–Pd1, CNT–
can be observed that the Raman spectra for all of the PVP–Pd2, and CNT–PVP–Pd3 samples, we noticed that the
samples were practically identical and presented similar D sample CNT–PVP–Pd1 displayed the highest current densi-
to G ratios. The ID/IG corresponded to 0.65, 0.64, 0.65, ties, while the sample CNT–PVP–Pd3 displayed the lowest,
0.69, and 0.78 for samples CNT, CNT–PVP CNT–PVP– indicating a decrease in the electrochemically active surface
Pd1, CNT–PVP–Pd2, and CNT–PVP–Pd3, respectively; in- areas for the samples obtained after the successive deposi-
dicating that the nonbonding interaction with PVP as well tion steps. The cyclic voltammograms for ethanol oxidation
as the sequential Pd-deposition steps did not lead to any recorded in a 1.0 m ethanol/1 m KOH solution are presented
significant structural changes (inclusion of surface defects) in Figure 5B and are in agreement with results obtained in
at the MWCNT surfaces. This is supported by the regular the ethanol-free supporting electrolyte solutions (Fig-
and undamaged appearance of the MWCNT walls observed ure 5A). The electrooxidation of ethanol in alkaline media
as the parallel lattice fringes in Figure 2B. In order words, may be a complex multistep reaction that involves adsorbed
our approach enabled the controlled deposition of Pd NPs intermediate species like acetyl and CO, which inactivate
without causing any structural changes at the MWCNT and poison the catalysts and lead to the production of acet-
surfaces.
aldehyde and acetic acid instead of CO2 as the final prod-
After the synthesis of the CNT–PVP–Pd materials with uct. The backward scan peak represents either the removal
well-defined compositions and morphological features, we of intermediate species incompletely oxidized in the forward
turned our attention to the investigation of their electrocat- scan, or the oxidation of numerous ethanol molecules sur-
alytic applications. We employed the CNT–PVP–Pd materi- rounding the active electrode surface. The complete oxi-
als as electrocatalysts for the ethanol oxidation reaction, dation of ethanol to CO2 is described by the following
which is relevant in the context of direct-alcohol fuel cells equation:
(DAFCs).[28,29] In particular, we were interested in investi-
gating the activities of the CNT–PVP–Pd materials as a
C2H5OH + 3H2O Ǟ 2CO2 + 12H+ + 12e–
function of their compositions and morphologies (samples
It can be observed that the current densities for ethanol
obtained after the first, second, and third reduction steps) oxidation decrease for the samples obtained after the se-
Eur. J. Inorg. Chem. 2014, 1439–1445
1442
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Table 1. Current peak and potential peak values for the electrocata-
lysts.
Sample
Current peak Potential peak Onset potential
[mAcm–2]
[V]
[V]
CNT–PVP–Pd1
CNT–PVP–Pd2
CNT–PVP–Pd3
2.65
1.36
0.05
–0.19
–0.17
–0.15
–0.60
–0.59
–0.42
electrode surface due to diffusion was expected. Figure 6
shows the chronoamperometric curves recorded at –0.1 V
in a solution containing 1 m ethanol and 1 m KOH, with
different electrodes. Firstly, control experiments were per-
formed with a bare glassy carbon surface before and after
coating with CNT–PVP (red and blue curves, respectively)
and, as expected, no faradaic current was observed. Then,
the experiment was repeated with an electrode whose sur-
face was coated with CNT–PVP–Pd1 (black line). A sharp
current decay at very short times was noticed because of
the charging of the double layer. At longer times a steady
state faradaic current was readily reached, confirming that
Pd plays a role in the ethanol anodic oxidation process. It
should also be pointed out that the current was maintained
at the same level for more than 1 h for the ethanol electrox-
idation process in alkaline medium, indicating that no sig-
nificant leaching of Pd NPs took place during this period.
Figure 5. (A) CV curves and (B) CV curves for ethanol oxidation
for samples CNT–PVP–Pd1, CNT–PVP–Pd2, and CNT–PVP–Pd3
(black, red, and blue traces, respectively). The CVs were recorded
at room temperature in an alkaline electrolyte with a sweep rate of
5 mV/s. The electrolytes consisted of 1 m KOH (A) and 1 m ethanol/
1 m KOH (B).
quential Pd-deposition steps (CNT–PVP–Pd1 Ͼ CNT–
PVP–Pd2 Ͼ CNT–PVP–Pd3). Specifically, the peak current
densities for ethanol oxidation were 2.65, 1.36, and
0.05 mAcm2 for samples CNT–PVP–Pd1, CNT–PVP–Pd2,
and CNT–PVP–Pd3, respectively (Table 1). These results
indicate that the sample obtained after the first deposition
step was the most electrocatalytically active. This can be
attributed to the lower degree of aggregation between the
Pd NPs less than 5 nm in size at the surfaces of the
MWCNTs as well as the absence of larger particles and ag-
gregates (as observed for the samples obtained after the sec-
ond and third reduction steps), which would lead to a de-
crease in the electrochemically active surface areas in the
materials. It is noteworthy that the electrooxidation of eth-
anol in basic media was characterized by two well-defined
Figure 6. Chronoamperometric curves for ethanol oxidation re-
corded in a 1 m ethanol/1 m KOH solution at –0.1 V on constant
stirring, with a bare glassy carbon electrode before (red trace) and
after immobilization of CNT–PVP (blue trace) and CNT–PVP–
Pd1 (black trace).
Conclusions
We have demonstrated a facile and environmentally
current peaks on the oxidation forward and the reverse friendly strategy for supporting Pd NPs on MWCNTs to
scans. In the oxidation forward scan, the oxidation peak obtain systems displaying controlled compositions and
corresponds to the oxidation of freshly chemisorbed species structures. Our approach did not require any prior chemical
(ethanol adsorption).[5,30–33] On the other hand, the reverse modification steps to include functional groups on the
scan peak is primarily associated with removal of carbon- MWCNT walls to achieve Pd deposition. In fact, our ap-
aceous species that are not completely oxidized in the for- proach was based on the nonbonding interaction and thus
ward scan.[30–33]
the noncovalent functionalization of MWCNTs with PVP.
The electrochemical stability of the CNT–PVP–Pd1 ma- As PVP can act as a stabilizing agent in the synthesis of Pd
terial was investigated by chronoamperometry. Experiments nanomaterials, the PVP present at the MWCNT surfaces
were performed under constant stirring (50 rpm), hence no also serve as sites for the nucleation and growth of Pd NPs.
significant influence of the depletion of material from the Our results demonstrated that Pd NPs below 5 nm in dia-
Eur. J. Inorg. Chem. 2014, 1439–1445
1443
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
Pd NPs Supported on MWCNTs: An ethanol suspension (3 mL)
containing MWCNTs (0.056 g) was added to a solution of PVP
(35 mg) dissolved in ethanol (6 mL). This mixture was heated to
75 °C for 15 min under magnetic stirring, which led to the uniform
dispersion of the MWCNTs in the suspension. Then an aqueous
solution of PdCl42– (5 mm, 1 mL) was added dropwise and the reac-
tion allowed to proceed for 3 h. This was the first reduction step.
The second reduction step was performed by the addition of fur-
meter could be directly deposited on MWCNTs by this
route. Also, sequential reduction steps could be employed
to improve the coverage of Pd NPs at the MWCNT sur-
faces, but this also led to the formation of larger Pd par-
ticles/aggregates. The electrocatalytic activity for ethanol
oxidation was investigated for the MWCNTs decorated with
Pd NPs as a function of their composition and structure.
In this case, MWCNTs decorated with Pd NPs of smaller
sizes and with lower coverages displayed the highest electro-
2–
ther PdCl4 (5 mm, 1 mL) to the reaction mixture obtained at the
end of the first reduction step, which was stirred at 75 °C for an-
catalytic activities. We believe the results presented herein other 1 h. Similarly, the third reduction step was carried out by the
2–
addition of further PdCl4 (5 mm, 1 mL) to the reaction mixture
obtained at the end of the second reduction step, which was stirred
at 75 °C for 1 h. The solids obtained after the first, second, and
third reduction steps were designated as CNT–PVP–Pd1, CNT–
PVP–Pd2, and CNT–PVP–Pd3, respectively. The reaction was
stopped at the end of each corresponding reduction step to isolate
the solids. In all cases, the products were harvested by centrifuga-
tion, washed several times with ethanol, and resuspended in eth-
anol for further use.
can have important implications for the design of
MWCNTs supported with metal NPs for electrochemical
and catalytic applications.
Experimental Section
Materials and Instrumentation
Analytical grade K2PdCl4 (98%, Aldrich) and PVP (Sigma–Ald-
rich, M.W. 55.000 g/mol) were used as received. The MWCNTs
(Figure S1) were kindly donated by the Laboratório de Nanomater-
iais, Departamento de Física, Universidade Federal de Minas
Gerais. The MWCNTs were synthesized by chemical vapor deposi-
tion (CVD) with ferrocene as a catalyst and ethylene gas as the
carbon source. The process temperature ranged from 700 to 800 °C.
All solutions were prepared with deionized water (18.2 MΩ) or eth-
anol.
Electrocatalytic Oxidation of Ethanol
For the preparation of the working electrode, a glassy carbon disk
electrode (with an area of 0.071 cm2) was firstly polished with an
alumina slurry (1.0 and 0.5 μm), washed with deionized water, and
sonicated for 10 min to remove any possible alumina residues.
Then, a specific mass of the MWCNTs supported with Pd NPs
(CNT–PVP–Pd1, CNT–PVP–Pd2, and CNT–PVP–Pd3) was dis-
persed into a solution of Pd in dimethylformamide/Nafion (1:1,
0.113 mg of Pd/mL) and sonicated for 20 min. Finally, CNT–PVP–
Pd1, CNT–PVP–Pd2, and CNT–PVP–Pd3 (3.5 μL) were each drop
casted onto the working electrode, then dried in vacuo for 12 h.
The Pd loading onto the working electrode for each sample was
0.392 μg and was calculated based on ICP-MS results.
The scanning electron microscopy images were obtained with a
JEOL microscope FEG-SEM JSM 6330F operated at 5 kV. We pre-
pared the samples for SEM by drop-casting an ethanolic suspen-
sion of the nanostructures over a Si wafer, followed by drying them
under ambient conditions. HRTEM imaging was carried out with
a probe-side aberration-corrected FEI Titan G2 operated at 200 kV
with an X-FEG electron source. STEM-HAADF imaging was car-
ried out with the same instrument with a convergence angle of
26 mrad, a HAADF inner angle of 52 mrad, and a probe current
of approximately 200 pA. EDX spectroscopy was carried out with
the Titan and a Super-X EDX detector with a collection solid angle
of 0.7 srad. We prepared the samples for (S)TEM imaging by drop-
casting an ethanolic suspension of the nanostructures over a holey
carbon-coated copper mesh support grid, followed by drying them
under ambient conditions. The Raman spectra were acquired with
a Renishaw Raman InVia (Renishaw, New Mills, Wotton-under-
Edge, UK) equipped with a charge-coupled device detector and
coupled to a Leica microscope (BTH2, Leica Microsystems GmbH,
Ernst-Leitz-Strasse, Germany) that allows the rapid accumulation
of Raman spectra with a spatial resolution of about 1 μm (micro-
Raman technique). The laser beam was focused on the substrate
by a ϫ50 lens. We prepared the substrate by drop casting the ethan-
olic-suspension-containing samples over an aluminum-foil-coated
glass slide. Laser power was always kept below 0.7 mW at the sam-
ple. The experiments were performed under ambient conditions
with a backscattering geometry. The samples were irradiated with
the 632.8 nm line of a Renishaw RL633 laser (Renishaw, New Mills,
Wotton-under-Edge, UK). Inductively coupled plasma mass spec-
trometry was performed in a Spectro Ciros CCD ICP optical emis-
sion spectrometer at the Analytical Instrumentation Center of Uni-
versity of Sao Paulo. Electrochemical experiments were carried out
to evaluate the activity of the synthesized CNT–PVP–Pd materials
with an Autolab PGSTAT 128N (Eco-Chemie, The Netherlands)
potentiostat equipped with the GPES software (Eco-Chemie, The
Netherlands).
The electrochemical cell was assembled with a standard three-elec-
trode setup consisting of the glassy carbon disk electrode modified
with CNT–PVP–Pd1, CNT–PVP–Pd2, or CNT–PVP–Pd3 as the
working electrode, Ag/AgCl (saturated KCl) as the reference elec-
trode, and a platinum wire as the counter electrode. Cyclic voltam-
metry (CV) experiments were carried out at room temperature in
1 m KOH as the supporting electrolyte. The scan rate was 5 mVs–1
and experiments were performed in the potential range of –0.9–
0.6 V. Experiments involving the anodic oxidation of ethanol were
carried out under similar conditions. Chronoamperometry was
used to investigate the long-term response of the supported materi-
als to ethanol oxidation and these experiments were performed at
–0.1 V under constant stirring (50 rpm) during 5000 s.
Supporting Information (see footnote on the first page of this arti-
cle): SEM image of MWCNTs and size histograms.
Acknowledgments
This work was supported by the Fundação de Amparo à Pesquisa
do Estado de São Paulo (FAPESP) (grant numbers 2011/06847-
0 and 2013/19861-6), the Conselho Nacional de Desenvolvimento
Científico e Tecnológico (CNPq) (grant number 471245/2012-7),
and start-up funds from the Universidade de São Paulo (grant
numbers 11.1.25042.1.0 and 2012-145). P. H. C. C. and M. B. thank
the CNPq for research fellowships. M. B. T. C. and P. S. C. thank
the FAPESP for the fellowships. C. C. S. D. O. and L. M. F. D.
thank the CNPq for the fellowships. S. J. H. and E. L. thank the
UK Engineering and Physical Sciences Research Council and the
US Defence Threat Reduction Agency for funding support.
Eur. J. Inorg. Chem. 2014, 1439–1445
1444
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

www.eurjic.org
FULL PAPER
[19] R. J. Chen, S. Bangsaruntip, K. A. Drouvalakis, N. W. S. Kam,
M. Shim, Y. M. Li, W. Kim, P. J. Utz, H. J. Dai, Proc. Natl.
Acad. Sci. U. S. A. 2003, 100, 4984–4989.
[20] P. T. Anastas, J. C. Warner, Green Chemistry: Theory and Prac-
tice, Oxford University Press, New York, 1998.
[21] M. J. O’Connell, P. Boul, L. M. Ericson, C. Huffman, Y. H.
Wang, E. Haroz, C. Kuper, J. Tour, K. D. Ausman, R. E. Smal-
ley, Chem. Phys. Lett. 2001, 342, 265–271.
[22] J. Hua, Z. Wang, J. Zhao, L. Xu, J. I. Zhang, R. Li, X. Sun, J.
Macromol. Sci. B 2011, 50, 679–687.
[23] Y. Huang, Y. Zheng, W. Song, Y. Ma, J. Wu, L. Fan, Compos-
ites Part A 2011, 42, 1398–1405.
[24] T. Hasan, V. Scardaci, P. Tan, A. G. Rozhin, W. I. Milne, A. C.
Ferrari, J. Phys. Chem. C 2007, 111, 12594–12602.
[25] P. J. Cappillino, K. M. Hattar, B. G. Clark, R. J. Hartnett, V.
Stavila, M. A. Hekmaty, B. W. Jacobsc, D. B. Robinson, J. Ma-
ter. Chem. A 2013, 1, 602–610.
[26] M. S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio, Phys.
Rep. 2005, 409, 47–99.
[1]
[2]
K. C. Nicolaou, P. G. Bulger, D. Sarlah, Angew. Chem. Int. Ed.
2005, 44, 4442–4489; Angew. Chem. 2005, 117, 4516.
D. Astruc, F. Lu, J. R. Aranzaes, Angew. Chem. Int. Ed. 2005,
44, 7852–7872; Angew. Chem. 2005, 117, 8062.
L. Yin, J. Liebscher, Chem. Rev. 2007, 107, 133–173.
J. Le Bras, J. Muzart, Chem. Rev. 2011, 111, 1170–1214.
C. Bianchini, P. K. Shen, Chem. Rev. 2009, 109, 4183–4206.
H. Zhang, M. Jin, Y. Xiong, B. Lim, Y. Xia, Acc. Chem. Res.
2013, 46, 1783–1794.
P. Avouris, Z. Chen, V. Perebeinos, Nat. Nanotechnol. 2007, 2,
605–615.
J. Wang, Electroanalysis 2005, 17, 7–14.
P. Serp, M. Corrias, P. Kalck, Appl. Catal. A 2003, 253, 337–
358.
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10] P. Avouris, Z. H. Chen, V. Perebeinos, Nat. Nanotechnol. 2007,
2, 605–615.
[11] B. Wu, Y. Kuang, X. Zhang, J. Chen, Nano Today 2011, 6, 75–
90.
[27] M. S. Dresselhaus, P. C. Eklund, Adv. Phys. 2000, 49, 705–814.
[28] A. S. Arico, S. Srinivasan, V. Antonucci, Fuel Cells 2001, 1,
133–161.
[29] E. Antolini, J. Power Sources 2007, 170, 1–12.
[30] C. W. Xu, H. Wang, P. K. Shen, S. P. Jiang, Adv. Mater. 2007,
19, 4256–4259.
[12] N. Karousis, N. Tagmatarchis, D. Tasis, Chem. Rev. 2010, 110,
5366–5397.
[13] D. Tasis, N. Tagmatarchis, A. Bianco, M. Prato, Chem. Rev.
2006, 106, 1105–1136.
[14] D. Eder, Chem. Rev. 2010, 110, 1348–1385.
[15] V. Georgakilas, D. Gournis, V. Tzitzios, L. Pasquato, D. M.
Guldi, M. Prato, J. Mater. Chem. 2007, 17, 2679–2694.
[16] D. A. Britz, A. N. Khlobystov, Chem. Soc. Rev. 2006, 35, 637–
659.
[31] X.-m. Chen, Z.-j. Lin, T.-t. Jia, Z.-m. Cai, X.-l. Huang, Y.-q.
Jiang, X. Chen, G.-n. Chen, Anal. Chim. Acta 2009, 650, 54–
58.
[32] K. Wu, X. Mao, Y. Liang, Y. Chen, Y. Tang, Y. Zhou, J. Lin,
C. Ma, T. Lu, J. Power Sources 2012, 219, 258–262.
[33] H. T. Zheng, Y. Li, S. Chen, P. K. Shen, J. Power Sources 2006,
163, 371–375.
[17] H. Li, L. Han, J. Cooper-White, I. Kim, Green Chem. 2012,
14, 586–591.
[18] Y. P. Sun, K. F. Fu, Y. Lin, W. J. Huang, Acc. Chem. Res. 2002,
35, 1096–1104.
Received: December 17, 2013
Published Online: March 5, 2014
Eur. J. Inorg. Chem. 2014, 1439–1445
1445
© 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim