Introduction

Membrane electrode assembly (MEA) is the key component of direct fuel/or liquid fuel cells. A MEA is generally fabricated by coating the anode catalyst and cathode catalyst particles on the respective side of a proton exchange membrane and then pressing a gas diffusion layer (normally, a porous carbon paper) on both the anode catalyst and cathode catalyst. In other words, carbon paper (CP), anode catalyst, membrane, cathode catalyst, and CP are sequentially integrated to form a coherent whole. MEA has been extensively applied to various fuel cells like direct H2/O2 fuel cell [1,2,3,4,5], direct HCOOH/O2 fuel cell [6,7,8,9,10], and direct alcohol/O2 fuel cell [11,12,13,14,15]. As for the direct liquid fuel cells, ORR takes place in a three-phase region: electrolyte/oxygen gas/catalyst. CP is not only used as the catalyst support, but also used as the channel of transferring gas and discharging water formed from ORR.

Regarding the construction steps of the MEA, catalysts need to be prepared separately, which includes several steps of operation [16,17,18,19,20,21]. Firstly, a support like carbon powder and carbon nanotubes and metallic salts such as the metal precursors were mixed in a solvent under stirring and a reduction agent solution like NaBH4 dissolved in water or ethylene glycol was subsequently added dropwise under stirring to form the black suspension. Followed by filtration and washing, the obtained catalyst particles were dried in a vacuum and put aside. Then, the catalyst particles, a solvent (water or alcohol), and a binder (generally Nafion solution) were mixed under sonication to form a uniform suspension (called ink). This ink was finally coated on a proton exchange membrane or on CP to fabricate a MEA. It is quite clear that there are lots of deficiencies for the conventional preparation of the catalyst particles: (1) the preparation process is cumbersome, (2) the amount of the catalyst particles may be reduced a little due to the filtration and washing steps, and (3) some catalyst particles may be aggregated together during their filtration and storage. It is therefore significant to develop a new preparation method of catalyst nanoparticles. Rego and coworkers reported the preparation of Pd nanoparticles directly deposited on porous CP by electroless deposition [22]. CP was firstly subjected to the necessary treatment of sensibilization/activation step by successively immersing the CP in SnCl2 and PdCl2 solutions. The as-treated CP was then immersed in a Pd electroless solution (27 mmol L−1 N2H4, 28 mmol L−1 Pd2+, 0.1 mol L−1 EDTA, 600 mL NH4OH) at room temperature to let the formed Pd nanoparticles be deposited on the CP. In this electroless deposition, the formation of Pd nanoparticles and their deposition on CP were carried out separately. In other words, Pd nanoparticles were firstly produced in places outside CP, and subsequently, some particles were immobilized on CP. This would of course lead to the loss of the Pd nanoparticles. In addition, larger-size Pd particles were formed with this electroless deposition.

It is generally known that Pd-based catalysts present excellent electrocatalytic activity for ethanol oxidation in alkaline media [23, 24]. Among them, Pd-Ni [10, 17, 20, 21], Pd-Sn [21, 25,26,27], Pd-Pt [28,29,30], and Pd-metal oxide [31,32,33,34] have been paid considerable attention. In addition, binary Pd-Ni catalysts present efficient electroactivity for ethanol oxidation. Recently, we synthesized Pd x Ni1/multiwalled carbon nanotube (CNT) catalysts by in situ deposition of Pd nanoparticles on Ni particles [17]. Results revealed that Pd3.7Ni1/MWCNT catalyst exhibits a peak current density of 130.9 mA cm−2 μg−1 (Pd), showing the excellent electrocatalytic activity of the Pd-Ni catalyst for ethanol oxidation.

In this work, a novel preparation method of Pd/PdNi nanoparticles by in situ deposition onto a porous CP using metallic salt as the precursor has been proposed and investigated. The metal precursor was firstly immobilized onto the porous CP by Nafion solution, and after drying, the metal salt-loaded CP was immersed in aqueous NaBH4 solution, leading to the formation of metal nanoparticles which were in situ deposited on CP. This new method of preparation of catalyst nanoparticles is very simple and can be easily used to synthesize the CP-loaded metal nanoparticles uniformly dispersed on CP with small sizes. The as-fabricated CP-loaded Pd (PdNi) nanocatalysts were characterized by TEM and X-ray diffractogram (XRD) techniques, and their electroactivity for ethanol oxidation was studied in alkaline media.

Experimental

All chemicals used in this work, including sodium hydroxide, palladium chloride, nickel(II) chloride, sodium borohydride, and ethanol, were of analytical grade and used as received without further purification. Pure water (18.2 MΩ cm) was obtained by double distillation of deionized water. CP (thickness = 0.3 mm, porosity = 75%) and Nafion solution (5%, Dupont) were purchased from Sinopharm Group Chemical Reagent Co. Ltd. Synthesis of CP-supported Pd/PdNi nanocatalysts was carried out according to the following steps. The CP was firstly immersed in a 0.1% (w/w) Triton X-100 solution for 24 h to reduce the paper hydrophobicity, according to [22]. The wet paper was then immersed in a large amount of pure water to remove the Triton X-100 and was dried in air for use. Meanwhile, x mg of PdCl2, y mg of NiCl2·6H2O, and 3 mg of multiwalled carbon nanotube (CNT) were dissolved in 0.24 mL of water under sonication, and then, 60 μL of Nafion solution was added to form a uniform suspension. Next, 20 μL of the suspension was added dropwise onto the one side of the CP treated by Triton X-100. Addition of CNT particles leads to the well-dispersed metal salts when the suspension was dried and thus facilitates the reaction between NaBH4 and the metal ions, as indicated in the next step. After drying in air, the CP was immersed in 10 mL of 5% (wt%) NaBH4 solution and held on for 5 h. The CP was afterwards removed from the solution and put to a large amount of water until neutral water was obtained. At last, the CP was dried on a vacuum at 40 °C for 10 h to obtain the CP-supported Pd/PdNi nanocatalysts, which were marked by Pd-situ/CP at x = 1 mg and y = 0 mg, Pd9Ni1-situ/CP at x = 1 mg and y = 0.15 mg, and Pd8Ni2-situ/CP at x = 1 mg and y = 0.34 mg. Figure 1 shows the schematic diagram of the preparation process of the samples. As indicated in Fig. 1, formation and deposition of the metal nanoparticles take place on the surface of the CNT and CP. For a comparison, a conventional CNT-supported Pd nanoparticles (Pd/CNT) with the Pd mass loading of 17% were firstly synthesized by NaBH4 reduction reaction according to our recent report [21], and then, 5 mg of the Pd/CNT particles was added to 0.96 mL of ethanol containing 40 μL of Nafion solution. After being subjected to sonication for 1 h, 50 μL of the as-obtained suspension was transferred onto the one side of the CP treated by Triton X-100, leading to the Pd/CNT particles dispersed on CP, named by Pd/CP. For the four prepared samples Pd-situ/CP, Pd9Ni1-situ/CP, Pd8Ni2-situ/CP, and Pd/CP, their Pd mass loading on CP was controlled to the same value (0.17 mg cm−2).

Fig. 1
figure 1

Schematic diagram for in situ deposition of metal nanoparticles onto carbon paper

Full size image

Crystalline phase compositions of the samples were studied with XRD obtained in a D/MAX2500X diffractometer (Japan), operating with Cu K alpha radiation generated at 40 kV and 250 mA (l = 0.15418 nm). The surface structures of the samples were characterized by TEM images recorded in a JEM-2100F transmission electron microscopy.

Electrochemical measurements were carried out with an AutoLab PGSTAT30/FRA (Eco Chimie, The Netherlands) electrochemical workstation. A conventional three-electrode system was applied, where the Pd or PdNi nanocatalyst-modified glassy carbon (GC), Pt foil, and Ag/AgCl (in saturated KCl) were used as working, counter, and reference electrodes, respectively. All potentials reported in this paper were against the Ag/AgCl. For the EIS measurements, the amplitude of modulation potential was 10 mV, and the range of frequency was controlled from 40 kHz to 40 mHz. Geometrical area of the CP on which the catalyst particles were dispersed was controlled to be 0.25 cm2 and was used to calculate the current density. All experiments were performed in room temperature.

Results and discussion

Figure 2a shows large aggregates of Pd particles, which may be related to the preparation process of the Pd/CP [21]. Once the Pd2+ ions are reduced in solution, the as-formed Pd particles start to aggregate before they are deposited on CNT, leading to the larger aggregates on CNT. This can be clearly seen from TEM image of the Pd/CP shown in Fig. 2a, where a number of large aggregates, composed of smaller particles with sizes of 4.5∼7.1 nm, are formed. Thus, it is reasonable to infer that according to the fabrication of the Pd/CP, if the formation of Pd nanoparticles and their deposition on CNT take place on the same place (called “in situ synthesis”), the formation of the aggregates will be avoided or greatly reduced. TEM images of the Pd-situ/CP well support this hypothesis, as indicated in Fig. 2b. The Pd nanoparticles of the Pd-situ/CP sample are uniformly dispersed on CNTs, and no obvious aggregation between the Pd particles is observed. The sizes of the Pd nanoparticles of the Pd-situ/CP are 15–24 nm. As for the Pd9Ni1-situ/CP and Pd8Ni2-situ/CP samples (Fig. 2c, d), their particle sizes are 7∼11 and 8∼14 nm, respectively, and the catalyst particles exhibit relatively uniform distribution except for a small number of aggregates. This indicates that addition of Ni may decline the reduction of Pd ions, leading to the decrease of the size of Pd particles [35]. It is seen from the preparation process of the Pd-situ/CP, Pd9Ni1-situ/CP, and Pd8Ni2-situ/CP samples (Fig. 1) that reduction of metal precursors (metallic salts) is directly carried out on the surface of CNTs. The as-formed metal nanoparticles are therefore in situ immobilized on CNTs, reducing the chance for the particles to contact with each other. Figure 2c, d also shows that the boundary between the catalyst particles is relatively fuzzy, which may be caused by the presence of Ni oxides.

Fig. 2
figure 2

TEM images of the Pd/CP (a), Pd-situ/CP (b), Pd9Ni1-situ/CP (c), and Pd8Ni2-situ/CP (d) samples

Full size image

X-ray diffraction patterns of the Pd/CP, Pd-situ/CP, Pd9Ni1-situ/CP, and Pd8Ni2-situ/CP samples were recorded and are shown in Fig. 3. All samples exhibit a maximum peak located at 26.4° corresponding to the graphite (0 0 2) plane of CNT support. The weak diffraction peak at 54.5° is also ascribed to the (003) plane of CNTs. The diffraction peaks at 39.9°, 46.3°, 67.8°, and 81.5° are indexed to the (111), (200), (220), and (311) planes of Pd face-centered cubic (fcc) crystal structure. Note in Fig. 3 that there are no obvious fraction peaks involved in Ni due to the low contents of Ni in Pd9Ni1-situ/CP and Pd8Ni2-situ/CP samples. Indeed, XRD is mass sensitive, and if the reflection intensities of Ni are much lower than Pd, Ni will not be detected in the diffractograms. The Pd diffraction peaks of the binary PdNi catalysts maintain the same diffraction angle as the Pd/CP and Pd-situ/CP, showing that no alloy between Pd and Ni is formed in Pd9Ni1-situ/CP and Pd8Ni2-situ/CP catalysts. The Pd crystalline size of the as-synthesized samples can be estimated from the XRD data according to Scherrer’s equation. And the (111) reflection of the Pd face-centered cubic (fcc) structure at 2θ of 39.9° is used to calculate the crystalline size from the following Scherrer’s equation:

$$ d\left({A}^o\right)=\frac{0.89{\lambda}_{CuK}}{\beta_{2\theta} \cos {\theta}_{\max }} $$
(1)

where β is the width of half peak, λ is the incident wavelength, d is the particle diameter, and θ is the diffraction angle at peak maximum. The average crystalline size is thus reckoned to be 4.9 nm for Pd/CP, 5.8 nm for Pd-situ/CP, 5.1 nm for Pd9Ni1-situ/CP, and 5.2 nm for Pd8Ni2-situ/CP. These calculated data are less than those observed from TEM images, showing that the particles observed from TEM images are composed of a number of smaller crystal grains.

Fig. 3
figure 3

X-ray diffraction patterns of the prepared samples

Full size image

Cyclic voltammetric (CV) data on the prepared samples in 1 mol L−1 NaOH solution were recorded as indicated in Fig. 4. It is observed from Fig. 4 that the samples show a characteristic reduction peak of polycrystalline Pd electrode at potentials of −0.39 to −0.42 V, which can be attributed to the reduction of Pd oxides produced during the forward potential scan. In contrast to the Pd/CP and Pd-situ/CP, Pd9Ni1-situ/CP and Pd8Ni2-situ/CP samples display a well-defined oxidation peak r p at ca. 0.48 V and a reduction peak r n1 at 0.28∼0.29 V, which are attributed to the formation and reduction of Ni(III) species, respectively. The real surface areas involved to Pd active sites (S activePd) for the prepared samples can be reckoned according to the coulombs consumed for the reduction of Pd oxides (cathodic peak r n2) formed during the forward scan on the smooth polycrystalline Pd electrode (CV not shown). S activePd value of the Pd-situ/CP, Pd9Ni1-situ/CP, Pd8Ni2-situ/CP, and Pd/CP samples is 7.1, 7.8, 9.3, and 5.1 cm2, respectively, showing that the Pd and PdNi nanoparticles prepared by this in situ deposition method present larger active Pd surface area. This is consistent with the TEM images where highly dispersed metal nanoparticles of the Pd-situ/CP, Pd9Ni1-situ/CP, and Pd8Ni2-situ/CP samples were observed. It can be also seen that the addition of Ni enhances the synergistic effect of metals between Pd and Ni and thus improves the catalytic activity.

Fig. 4
figure 4

Cyclic voltammograms of the samples in 1.0 mol L−1 NaOH at a scan rate of 50 mV s−1

Full size image

Figure 5 reveals CV data of the prepared samples in NaOH solution containing 0.5 mol L−1 ethanol. All samples present a typical anodic peak r p for ethanol oxidation during the forward scan. The r p peak current density on the Pd/CP, Pd-situ/CP, Pd9Ni1-situ/CP, and Pd8Ni2-situ/CP is 88.7, 112.7, 99.0, and 93.4 mA cm−2, respectively. Also, the Pd-situ/CP, Pd9Ni1-situ/CP and Pd8Ni2-situ/CP exhibit a respective onset potential of −0.62, −0.57, and −0.55 V for ethanol oxidation, which present more negative shifts compared to that of −0.51 V on the Pd/CP. Results show that for Pd9Ni1-situ/CP and Pd8Ni2-situ/CP catalysts, the synergistic effect between Pd and Ni is not obvious due to the absence of the alloying of Pd and Ni, leading to their lower current densities and more positive onset potentials compared to the Pd-situ/CP catalyst. It is further found from Fig. 5 that the current density of ethanol oxidation displays a rapid rise from the onset potential to the peak r p potential. In this potential range, the four samples also present different anodic current densities as indicated in the zoomed main panel, where the current density at the potential of −0.25 V on the Pd/CP, Pd-situ/CP, Pd9Ni1-situ/CP, and Pd8Ni2-situ/CP is 53.0, 53.4, 54.1, and 50.7 mA cm−2, respectively. After the peak r p, the anodic current density declines nearly to zero, which is caused by the formation of the Pd oxides at higher potentials. The peak r p potentials on the Pd-situ/CP, Pd9Ni1-situ/CP, and Pd8Ni2-situ/CP, however, shift to more positive values, compared to that on the Pd/CP. This indicates that the formation of Pd oxides at higher anodic potentials on these in situ synthesized Pd samples is suppressed. The rapid rise of the anodic current on the Pd9Ni1-situ/CP and Pd8Ni2-situ/CP commences at a potential of 0.4 V, caused by the formation of Ni(III) species. Subsequent cathodic peaks at 0.28∼0.3 V are related to the reduction of the Ni(III) species.

Fig. 5
figure 5

Cyclic voltammograms of the samples in 1.0 mol L−1 NaOH in the presence of 0.5 mol L−1 ethanol at a scan rate of 50 mV s−1. Inset is the zoomed main panel of the CVs

Full size image

The mechanism of ethanol oxidation on Pd in alkaline media can be described according to the following reactions (2)–(5) [36]:

$$ \mathrm{Pd}+{\mathrm{CH}}_3{\mathrm{CH}}_2\mathrm{OH}\to \mathrm{Pd}-{\left({\mathrm{CH}}_3{\mathrm{CH}}_2\mathrm{OH}\right)}_{\mathrm{ads}} $$
(2)
$$ \mathrm{Pd}-{\left({\mathrm{CH}}_3{\mathrm{CH}}_2\mathrm{O}\mathrm{H}\right)}_{\mathrm{ads}}+{3\mathrm{OH}}^{-}\to \mathrm{Pd}-{\left({\mathrm{CH}}_3\mathrm{CO}\right)}_{\mathrm{ads}}+{3\mathrm{H}}_2\mathrm{O}+{3\mathrm{e}}^{-} $$
(3)
$$ \mathrm{Pd}-{\left({\mathrm{CH}}_3\mathrm{CO}\right)}_{\mathrm{ads}}+\mathrm{Pd}-{\mathrm{OH}}_{\mathrm{ads}}\to \mathrm{Pd}-{\mathrm{CH}}_3\mathrm{CO}\mathrm{OH}+\mathrm{Pd} $$
(4)
$$ \mathrm{Pd}-{\mathrm{CH}}_3\mathrm{COOH}+{\mathrm{OH}}^{-}\to \mathrm{Pd}+{\mathrm{CH}}_3{\mathrm{COO}}^{-}+{\mathrm{H}}_2\mathrm{O}\kern1em \left(\mathrm{fast}\right) $$
(5)

As for the binary Pd-Ni catalysts, Ni(OH)2 and NiO are easily formed in alkaline media. In addition, ethanol oxidation is related to both (CH3CO)ads and OHads. Addition of Ni can increase the concentration of OH− on the surface of catalyst, which could accelerate the ethanol oxidation to a certain extent. Therefore, the ethanol oxidation mechanism on Pd x Ni y -situ/CP catalysts can be indicated by reactions (2), (3), and (6) [37]:

$$ \mathrm{Pd}-{\left({\mathrm{CH}}_3\mathrm{CO}\right)}_{\mathrm{ads}}+{\mathrm{OH}}^{-}\overset{\mathrm{Ni}\left(\mathrm{OH}\right)2}{\to}\mathrm{Pd}-{\mathrm{CH}}_3\mathrm{CO}\mathrm{OH}+{\mathrm{e}}^{-} $$
(6)

Chronoamperometric data of the prepared samples were recorded by maintaining the electrode at a constant potential of −0.25 V in a solution containing 0.5 mol L−1 ethanol and 1.0 mol L−1 NaOH, as indicated in Fig. 6. The Pd-situ/CP and Pd9Ni1-situ/CP present the highest current density compared to the Pd/CP and Pd8Ni2-situ/CP samples. It is clearly seen from Fig. 6 that the maximum current density obtained is over 70 and 60 mA cm−2 on Pd9Ni1-situ/CP and Pd8Ni2-situ/CP, respectively. Although almost the same maximum current density arises on Pd/CP and Pd-situ/CP, after that, the current density on the Pd-situ/CP is always significantly larger than that on the Pd/CP. Figure 6 also reveals that the Pd-situ/CP and Pd9Ni1-situ/CP exhibit a very close current density after holding the step time of 600 s, and the current densities from the samples synthesized by this “in situ method” are higher than the current density measured for the Pd/CP sample. Moreover, the Pd-situ/CP or Pd9Ni1-situ/CP sample delivers a current density at 1800 s that is over 2.6 times larger than that for the Pd/CP sample. This reveals the highly stable current density of ethanol oxidation on the samples prepared by this in situ deposition.

Fig. 6
figure 6

Chronoamperometric responses of the samples in 1.0 mol L−1 NaOH in the presence of 0.5 mol L−1 ethanol at a potential of −0.25 V

Full size image

Electrochemical reaction resistances of ethanol oxidation on the prepared samples were further studied through electrochemical impedance spectroscopic measurements. Figure 7 reveals the Nyquist plots of the samples at a potential of −0.25 V in 1 mol L−1 NaOH containing 0.5 mol L−1 ethanol. All samples exhibit a well-defined semicircle Nyquist plot in the range of the frequency studied, which is caused by the ethanol oxidation. The impedances on both the imaginary and real axes follow the order Pd9Ni1-situ/CP < Pd-situ/CP < Pd/CP < Pd8NI2-situ/CP. The fitting curves shown in Fig. 7 are obtained from the equivalent electric circuit compatible with the corresponding Nyquist diagrams. It is clearly seen from the fitting curves that the equivalent circuit is well compatible with the electrochemical processes. The charge transfer resistance (Rct) values for ethanol oxidation on Pd9Ni1-situ/CP, Pd-situ/CP, Pd/CP, and Pd8NI2-situ/CP samples are 33.5, 49.6, 61.2, and 67.3 Ω, respectively. Results are consistent with the anodic current densities of ethanol oxidation at a potential of −0.25 on the prepared samples as indicated in CVs (Fig. 5).

Fig. 7
figure 7

a Electrochemical impedance spectra of the samples in 1.0 mol L−1 NaOH in the presence of 0.5 mol L−1 ethanol at −0.25 V. Dots and corresponding lines present the experimental data and fitting values, respectively. b The equivalent electric circuit compatible with the Nyquist diagrams. Rct charge transfer resistance, Rs electrolyte resistance, Q constant phase element (CPE), C capacitance, R3 resistance of the oxidation of the intermediates produced from alcohol oxidation

Full size image

Electrocatalytic stabilities of the as-synthesized samples for ethanol oxidation were investigated by cyclic voltammetry. Figure 8a–c shows 400 consecutive cyclic voltammograms of the Pd/CP, Pd-situ/CP, and Pd9Ni1-situ/CP samples in 1.0 mol L−1 NaOH + 0.5 mol L−1 ethanol at a scan rate of 50 mV s−1. It is clearly seen from Fig. 8 that the anode current density on the forward-going scan declines with the increasing of the number of successive cycle sweep. According to the changes of the anode peak (r p) current density on the forward-going scan, Pd9Ni1-situ/CP sample exhibits relatively higher electrocatalytic stability than Pd/CP and Pd-situ/CP. The r p current densities for the 400th cycle on the Pd/CP, Pd-situ/CP, and Pd9Ni1-situ/CP samples are respectively 67, 73, and 80% of those for the first cycle. That means that the r p peak current density on the Pd9Ni1-situ/CP sample keeps a slower falling with the cycle sweeping number, while those on the Pd/CP and Pd-situ/CP samples display a relatively rapid decline with the cycle sweeping number. Results show that an appropriate adding amount of Ni improves the stability of Pd catalyst.

Fig. 8
figure 8

Four hundred consecutive sweeps of cyclic voltammograms on the Pd/CP (a), Pd-situ/CP (b), and Pd9Ni1-situ/CP (c) in 1 mol L−1 NaOH containing 0.5 mol L−1 ethanol at a potential scan rate of 50 mV s−1

Full size image

Electrocatalytic stabilities of the prepared samples can be further seen from their CV profiles recorded at the 400th cycle. Figure 9 shows the CV curves at the 400th cycle on the Pd/CP, Pd-situ/CP, and Pd9Ni1-situ/CP catalysts in 1 mol L−1 NaOH containing 0.5 mol L−1 ethanol at a potential scan rate of 50 mV s−1. It can be clearly seen from Fig. 9 that Pd-situ/CP and Pd9Ni1-situ/CP have higher r p current densities than the Pd/CP. Results show that the samples prepared by the in situ deposition method possess higher electrocatalytic stability for ethanol oxidation.

Fig. 9
figure 9

The cyclic voltammetric curves at the 400th cycle obtained from Fig. 8. Others same as Fig. 8

Full size image

The different electrocatalytic behaviors of Pd/CP and Pd-situ/CP samples can be attributed to their different deposition mechanisms. It is clearly seen from Fig. 1 that both production and deposition of metallic nanoparticles take place on the surface of CNTs. Immobilization of nanoparticles on CNTs reduces their surface energy and is beneficial to improve their stability, leading to the great reduction of the aggregation between the nanoparticles. By contrast, conventional synthesis of metallic nanoparticles undergoes such steps as reduction of metal precursor, migration of nanoparticles to a substrate, and immobilization on the substrate, as indicated in Fig. 10. The aggregation between the particles occurs during their migration to the surface of CNTs. The in situ deposition method proposed in this work is of therefore great significance for the synthesis of a substrate-supported catalyst with high electroactivity and electrostability.

Fig. 10
figure 10

Schematic diagram for conventional synthesis of metallic nanoparticles

Full size image

Conclusions

In summary, we have proved the feasibility of in situ deposition of metal nanoparticles onto CP by immersing the CP, on which metal salts as the metal precursors have been loaded, into the NaBH4 solution. Compared to the conventional synthesis of metal nanoparticles (Pd/CP), this new in situ deposition method eliminates the filtration step and thus avoids the loss of catalyst particles. TEM images of the Pd and PdNi nanoparticles prepared by this new method (Pd-situ/CP and Pd-Ni-situ/CP) show that they are uniformly dispersed on CP without obvious aggregation between the particles. The as-fabricated Pd-situ/CP and Pd9Ni1-situ/CP samples reveal higher electrocatalytic activity and stability for ethanol oxidation than the Pd/CP. This study opens up an avenue to explore a new synthesis method of metal nanoparticles immobilized on a support for fuel cell applications.