Electrocatalytic activity of the Pt/PdO electrode toward MOR

Figure 5-6(a) shows the SEM image of the PdO thin film after the pulse-electrodeposition of Pt nanoparticles. The Pt nanoparticle-loaded PdO thin film has a flake shape less distinct than the as-deposited PdO thin film shown in Fig. 5-1 because of the accumulation of Pt nanoparticles on the ridge of the nanoflake. The XRD spectrum of

Figure 5-6 SEM image of the PdO thin film after the pulse-electrodeposition of Pt nanoparticles; (b) XRD spectrum of the Pt/PdO thin film shown in (a).

62

the Pt/PdO thin film is shown in Fig. 5-6(b). The two diffraction peaks situated at 40.2^o and 68.3^o correspond to the Pt (111) and the (220) lattice planes, respectively. The other diffraction peaks shown in the XRD spectrum are due to the hexagonal lattice structure of graphite [114]. The average particle size of Pt nanoparticles on the PdO nanoflakes thin film is 5.5 nm according to the Debye-Scherrer calculation based on the peak width of the Pt (111) diffraction peak [32, 105].

Figure 5-7(a) shows the bright field TEM image of electrodeposited Pt nanoparticles on several PdO nanoflakes, which were separated from the Pt/PdO electrode by ultrasonic agitation in the ethanol solution. The PdO nanoflakes are decorated with dense spots of dark contrast. The high resolution TEM (HRTEM) images of a selected area on the edge

Figure 5-7 (a) The bright field TEM image of electrodeposited Pt nanoparticles on PdO nanoflakes separated from the Pt/PdO electrode; (b) and (c) HRTEM images of a selected area on the edge of a nanoflake.

63

of a nanoflake (Figs. 5-7(b) and (c)) indicate that the dark spots are due to Pt nanocrystals;

the two lattice spacings labelled by A and B correspond to the Pt (111) and (20 0) lattice planes, respectively. The size of most Pt nanoparticles is in the range between 3.0 and 5.5 nm, which is reasonably in agreement with the average size derived from the XRD analysis.

The Pt nanoparticles uniformly spread over the PdO nanoflake although particle aggregation occurs in some areas. Because of the nanoscale size and the good dispersion of Pt nanoparticles on the PdO nanoflake, a large amount of surface Pt atoms must be located in the border area with the PdO substrate. As a consequence, numerous Pt electrochemical active sites are subject to strong local chemical and electronic modifications induced by the PdO substrate. These modifications may greatly influence the electrocatalytic activity of the Pt nanoparticles toward MOR.

The cyclic voltammogram of the Pt/PdO electrode in the aqueous solution of 1 M CH3OH and 0.5 M H2SO4 is shown in Fig. 5-8. For comparison, the figure also presents the CV curves of an electron-beam deposited Pt thin film and electrodeposited Pt particles;

Figure 5-8 Cyclic voltammograms of the Pt/PdO, the Pt/C and the blanket-Pt electrodes in the aqueous solution of 1 M CH3OH + 0.5 M H2SO4. The CV curves shown in the figure are for the fourth CV cycle.

The scan rate was 20 mV s^-1.

64

Table 5-1 The onset and peak potentials of the CO electro-oxidation reaction on the Pt/PdO, the Pt/C, the blanket-Pt and the PdO electrodes. The electrochemical active surface area (ESA) derived from the CO stripping cyclic voltammograms is also listed.

Pt/PdO Pt/C Blanket-Pt PdO nanoflakes

Onset potential (V) 0.49 0.52 0.52 0.59

Onset potential (V) 0.55 0.58 0.63 0.67

ESA (cm²/mg) 189.6 82.2 52.2

both types of the Pt electrocatalyst were deposited directly on the carbon cloth, and we will refer to the Pt thin film and the Pt particle electrodes as blanket-Pt and Pt/C, respectively.

The MOR peak of the Pt/PdO electrode has the peak maximum around 0.62 V with the onset potential at 0.45 V. The Pt/PdO electrode has both the MOR peak potential and the onset potential less positive than the Pt/C and the blanket-Pt electrodes, indicating that the Pt/PdO electrode has a lower overpotential for methanol electro-oxidation. Moreover, the Pt/PdO electrode has a much larger intensity ratio of the MOR peak (denoted by I_f) to the anodic peak in the reverse scan (denoted by Ib) than the other two electrodes. The If/Ib

ratio of the Pt/PdO electrode is 1.87, and the ratio for the Pt/C and the blanket-Pt electrodes is 1.28 and 1.22, respectively. We may use the If/Ib ratio as an index to describe the resistance of the electrocatalyst against the CO poisoning in MOR. The large I_f/I_b ratio of the Pt/PdO electrode indicates that Pt nanoparticles on the PdO support have a higher CO tolerance. The efficient removal of CO adspecies is crucial for improving the electrochemical performance in DMFCs. To examine the CO tolerance of the Pt/PdO electrode, we performed the CO stripping measurement for the electrode and the result is presented in Fig. 5-9. Also shown in the figure are the CO stripping CV voltammograms for the Pt/C, the blanket Pt and the PdO thin film electrodes. The onset and the peak potentials of the CO stripping peak are listed in Table 1. The Pt/PdO electrode has the CO oxidation peak potential smaller than the PdO nanoflake electrode by 0.12 V; this large

65

potential difference suggests that the CO stripping peak of the Pt/PdO electrode has little contribution directly from the metal Pd sites, which are created on the PdO support at the polarization potential (0.1 V) for the CO adsorption treatment. The Pt/PdO electrode has the CO stripping peak potential and the onset potential smaller than the Pt/C and the blanket-Pt electrodes, indicating that CO adspecies on the Pt/PdO electrode are easier to be removed from Pt electrocatalytic sites.

The electrochemical surface area (ESA) of a Pt based catalyst is often determined by the total charge in the hydrogen adsorption (or desorption) region in a CV voltammogram.

However, metal Pd has a great capacity of hydrogen absorption, the ESA of the Pt catalyst on the PdO support will be overestimated if the hydrogen region is used to calculate the ESA. We thus used the CO stripping peak to estimate the ESA of the Pt/PdO, the Pt/C and the blanket Pt electrodes by the following equation:

CO

0.484

ESA Q

 M



Figure 5-9 CO stripping CV curves for the Pt/PdO, the PdO thin film, the Pt/C and the blanket Pt electrodes. The scan rate was 20 mV s^-1.

66

where QCO (in mC) is the total charge calculated from the CO stripping peak, and MPt

represents the mass loading (mg) of the Pt catalyst on the electrode. The value, 0.484 (mC cm^-2), corresponds to the charge density required for the electro-oxidation of one monolayer of CO adspecies on the Pt surface [115] . The calculated ESA values of the three Pt-based electrodes are also listed in Table 1. The Pt/PdO electrode has an ESA of 189.6 cm²/mg, which is much larger than the ESA of the other two electrodes. The large ESA of the Pt/PdO electrode is ascribed to the enormous surface area of the PdO nanoflake support and the good dispersion of Pt nanoparticles on the support. Because of the large ESA, the Pt/PdO electrode exhibits a very high electrocatalytic activity toward MOR as shown in the CV voltammogram of Fig. 5-8.

Compared with the two electrodes without the PdO support, the Pt/PdO electrode has a low overpotential for MOR and a high CO tolerance, suggesting that the PdO nanoflake must play an important role in the enhancement of the electrocatalytic performance of the Pt nanoparticles. In general, the enhancement in the electrocatalytic activity of Pt catalysts toward MOR is explained by the bi-functional mechanism and the electronic effect model [4, 5, 8, 10, 20, 23]. The bi-functional mechanism is widely used to describe how hydroxyl surface groups oxidize and remove CO adspecies from neighboring Pt adsorption sites, thus avoiding the CO poisoning. For the Pt/PdO electrode, Pt nanoparticles are surrounded by abundant Pd-OH surface groups, which are produced when metal Pd sites are oxidized in the acidic electrolyte via Eq. 5-2. CO adspecies in the rim of a Pt nanoparticle are readily oxidized by neighboring Pd-OH surface groups via the Langmuir-Hinshelwood reaction mechanism because of immediate interactions of the CO adspecies with Pd-OH groups surrounding the Pt nanoparticle [20, 23]. Pd-OH surface groups may also spill over a Pt nanoparticle and progressively interact with CO adspecies in the inner surface area of the nanoparticle, resulting in the recovery of Pt free sites, which allow further methanol adsorption and oxidation.

67

Figure 5-10 (a) Pt 4f spectra of the Pt/PdO, the Pt/C and the blanket-Pt electrodes before the electrochemical measurement; (b) the Pd 3d XPS spectrum of the as-prepared Pt/PdO electrode.

Besides the bi-functional mechanism, electronic interactions between Pt nanoparticles and the PdO support may also improve the electrocatalytic activity of the Pt catalyst toward MOR. The electronic interaction can modify the electronic structure of the P t nanoparticle and thus alter the chemisorption behavior of CO adspecies on the nanoparticle [20, 23]. If the adsorption strength of CO adspecies becomes weaker as a result of the electronic structure modification on the Pt nanoparticle, the oxidation of CO adspecies may be kinetically favored, leading to a better catalytic activity of the Pt nanoparticle toward MOR.

68

Figure 5-10(a) shows Pt 4f spectra of the Pt/PdO, the Pt/C and the blanket-Pt electrodes before the electrochemical measurement. The Pt 4f_7/2 peak of the Pt/PdO electrode is situated at 70.8 eV, which negatively shifts from that of the blanket Pt thin film (71.4 eV) by ~0.6 eV. Previous studies have shown that the binding energy of core level electrons of metal nanoparticles shifts from that of the bulk counterpart as a function of the particle size [116-118]. The binding energy shift due to the nanosize effect generally increases with decreasing the size of metal nanoparticles, and the size effect on the energy shift becomes unimportant when the particle size is larger than 3 nm [116]. Therefore, the large negative shift in the Pt 4f7/2 binding energy is unlikely to result from the nanosize effect but is caused by an electronic interaction between the Pt nanoparticle and the PdO support. The negative binding energy shift indicates that negative charge transfer from the PdO support to Pt nanoparticles must occur. Surface atoms of a smaller Pt nanoparticle should experience a larger electronic modification because of their close proximity to the interface between the Pt nanoparticle and the PdO support, at where the charge transfer takes place [20, 23]. The XPS analysis clearly shows the importance of the PdO substrate to the electronic effect, which may modify adsorption properties of methanol and CO adspecies on Pt nanoparticles and thus enhance the electrocatalytic activity of the Pt nanoparticles. The Pt/C electrode has the Pt 4f peak maximum at 71.2 eV; the less negative Pt 4f energy shift compared with the Pt/PdO electrode indicates that Pt particles on the Pt/C electrode are subject to a smaller electronic modification. Moreover, most Pt particles on the Pt/C electrode have a size in the order of a few tens of nanometers, and, as a result, the number of surface Pt atoms that experience the electronic modification is only a very small portion of the total Pt atoms on the particle. The electrochemical performance of the Pt/C electrode is consequently less efficient that of the Pt/PdO electrode.

Pd is liable to anodic dissolution in acidic electrolytes [43, 91, 93] and, therefore, is considered an unsuitable electrocatalyst in DMFCs. Since the PdO nanoflake has a CV

69

behavior similar to metallic Pd, the anodic dissolution is also expected to occur to the PdO nanoflake electrode. During a CV test in acidic solutions, anodic dissolution of Pd takes place when the surface PdO (or Pd(OH)2), which is produced via the Pd oxidation in the forward sweep, is reduced in the reverse sweep, producing Pd²⁺ ions in the solution [43, 93].

Figure 5-11(a) shows the SEM image of a PdO nanoflake electrode after 10 CV cycles in the solution of 1M CH₃OH + 0.5 M H₂SO₄. The nanoflake morphology of the PdO/C electrode was nearly destroyed after 10 CV cycles, indicating severe anodic dissolution of Pd occurred. In contrast, the nanoflake feature was well preserved on the Pt/PdO electrode after 10 cycles of the CV measurement (Fig. 5-11(b)). This observation suggests that electrodeposited Pt nanoparticles can chemically stabilize the PdO support, and thus

Figure 5-11 Images of (a) the PdO nanoflake electrode and (b) the Pt/PdO electrode after 10 CV cycles in the solution of 1M CH3OH + 0.5 M H2SO4.

70

suppress the anodic dissolution of the PdO nanoflake support. It has been reported that the change in the surface chemical composition of Pd-containing alloys can affect the electrodissolution behavior of the alloys [43, 119, 120]. Figure 5-10(b) shows a Pd 3d XPS spectrum of the as-prepared Pt/PdO electrode. After the Pt electrodeposition, the Pt/PdO electrode exhibits a broad Pd 3d peak, which comprises two component peaks or more; the Pd(II) and the Pd(IV) oxidation states can be easily assigned in the XPS spectrum.

We do not know presently if the chemical phase with the Pd(IV) state is on the open area or covered by Pd nanoparticles, but the XPS analysis clearly demonstrates the dramatic change in the chemical composition on the surface of the PdO support after the Pt electrodeposition. The distinct chemical change on the nanoflake surface may modify the electrochemical behavior of the PdO support and thus alleviate the anodic Pd dissolution during the CV measurement in acidic media.

To study the electrocatalytic stability of the Pt/PdO electrode in the solution of 1 M CH3OH + 0.5 M H2SO4, we carried out chronoamperometric test at 0.5V. The

Figure 5-12 Chronoamperograms of the Pt/PdO electrode, the Pt/C, the blanket-Pt and the PdO electrodes in the aqueous solution of 1 M CH3OH + 0.5 M H2SO4 at room temperature (~25 ^oC) for 1 h. The oxidation potential was kept at 0.5 V vs. SCE.

71

chronoamperograms shown in Fig. 5-12 clearly demonstrates that the Pt/PdO electrode has a much better chronoamperometric performance than the Pt/C and the blanket -Pt electrodes.

The Pt/PdO electrode has a gradual decrease in the current density within the first 30 min, and then its current density keeps steady at ~0.8 mA/cm². On the other hand, the Pt/C and the blanket-Pt electrodes show a low current density of 0.27 mA/cm² and 0.02 mA/cm², respectively, after one hour of the chronoamperometric test. For comparison, the chronoamperometric result of the PdO nanoflake electrode is also shown in Fig. 5-12; the electrode is completely inactive under the experimental condition. The much better chronoamperometric performance of the Pt/PdO electrode is ascribed to the large ESA and the better CO tolerance of Pt nanoparticles. Moreover, Pt nanoparticles can suppress the anodic dissolution of the PdO support and hence improve the electrochemical stability of the Pt/PdO electrode.

5.4 Summary

We prepared PdO nanoflake thin films on carbon cloths by reactive sputtering deposition, and pulse-electrodeposited Pt nanoparticles on the PdO thin films. The nanoflake morphology of the PdO support provides a large surface area for Pt nanoparticle loading, resulting in a large ESA. The electrocatalytic activity of the PdO thin film and the Pt/PdO electrode toward MOR was studied in acidic media. The PdO nanoflake thin film has a cyclic voltamperometric behavior similar to a metallic Pd electrode. According to the XPS analysis, a thin metallic Pd surface layer is produced on the PdO nanoflake thin film after the first cycle of the CV test in the aqueous H2SO4 solution. Because of the repetitive PdO reduction and Pd oxidation in the acidic solution, methanol electro-oxidation on the PdO nanoflake thin film exhibits a CV feature that is closely related to the Pd/PdO redox reaction. We proposed a reaction mechanism scheme to explain the observed CV

72

features for the methanol electro-oxidation on the PdO nanoflake electrode in the acidic solution. The nanoflake morphology on the PdO electrode is seriously damaged after 10 cycles of the CV test because of the anodic dissolution of metal Pd in acidic medi a.

However, the anodic dissolution is greatly alleviated when Pt nanoparticles are electrodeposited on the PdO nanoflake thin film. The XPS study shows that negative charge transfer occurs from the PdO support to the Pt nanoparticles. The electronic interaction may modify adsorption properties of adspecies on the Pt nanoparticles, and thus affect electrochemical properties of the Pt/PdO electrode. The Pt/PdO electrode has a higher electrocatalytic activity toward MOR than the Pt/C and the blanket-Pt electrodes.

We ascribe the much better electrocatalytic performance of the Pt/PdO electrode to a high CO tolerance and the large ESA. The high CO tolerance of the Pt-PdO electrode is a result of the synergism of the bi-functional mechanism and the electronic effect operating on the electrode.

73