Preparation of Pt/CeO 2 -CNT through spontaneous adsorbing Pt

Through a two-step strategy, Pt/CeO₂-CNTs and Pt/CNTs were prepared by first microwave heating H2PtCl6 in NaOH ethylene glycol solution and then depositing Pt nanoparticles onto CeO₂–CNTs and CNTs, respectively [41]. Figure 2-12 shown the HRTEM of Pt/CeO2-CNTs. In Fig. 2-12 (a) three Pt nanoparticles were found to be surrounded by CeO₂ nanoparticles, which can be clearly distinguished from the lattice distance (~ 0.305 nm for CeO2 (111), ~ 0.225 nm for Pt (111). From Fig. 2-12 (b) we can

Figure 2-12 HRTEM image of CNT–CeO2/Pt catalyst.

18

clearly see the Pt nanoparticles on sidewall of CeO₂-CNTs adjacent to CeO₂ particles.

Because CO electro-oxidation is viewed as a rate-determining step during methanol electro-oxidation, CeO₂ would promote methanol electro-oxidation on Pt-based catalysts according to the bi-functional mechanism. To verify the role of CeO2 in Pt/CeO2-CNTs, Wang et al. performed the potentiodynamic and potentiostatic electro-oxidation of CO adlayer on Pt, as shown in Fig. 2-13 [41]. The CO stripping on Pt/CeO2-CNTs starts at 0.41 V, 0.08 V lower than 0.49 V on Pt/CNTs. In addition, the peak potential of CO stripping lies at 0.52 V on Pt/CeO2-CNTs while lying at 0.6 V on Pt/CNTs. Such results indicate that CO can be electro-oxidized more easily on Pt/CeO₂-CNTs than on Pt/CNTs.

In Fig. 2-13, two peaks on Pt/CeO2–CNTs, possibly because that there are different contact states between Pt and CeO₂. The results in Fig. 2-13 indicate that CO can be electro-oxidized on Pt/CeO2-CNTs more easily than on Pt/CNTs. This is possibly because CeO₂ can provide oxygen-containing group to electro-oxidize CO, the major intermediate during methanol electro-oxidation. Therefore, CeO2 can be viewed as promising cocatalyst for methanol electro-oxidation on Pt-based catalysts.

Figure 2-13 CO stripping curves on Pt/CNTs and Pt/CeO₂–CNTs recorded in 1 M HClO₄.

19

2.3.5 Pt nanoparticles on SnO

nanowire-based electrodes

Because carbon nanotubes (CNTs) and nanofibers (CNFs) have unique properties, such as high surface area, good electronic conductivity, strong mechanical properties, and chemical stability, they have attracted great attention as promising catalyst supports. A number of research groups have demonstrated the advantages of using CNTs or CNFs as supports to better disperse Pt and its alloys for oxygen reduction and methanol oxidation reaction. In particular, the nanotube-based 3D electrode structure has shown very promising implications. Unlike CNTs, nanowires (NWs) can be made of various

Figure 2-14 SEM and TEM micrographs of SnO₂ NWs grown on carbon fibers of carbon paper by thermal evaporation method. (a) SEM image showing full coverage of SnO₂ NWs on fibers of carbon paper. (Inset) Fibers of bare carbon paper. (b) TEM image showing individual SnO₂ NWs. (c) TEM images showing Pt nanoparticles electrochemically deposited onto SnO₂ NWs (Inset) Pt nanoparticles deposited onto a single SnO2 NW

20

compositions of materials and they have solid cores. NWs have demonstrated superior electrical, optical, mechanical, and thermal properties. Metal oxide NWs have unique advantages as supports for dispersing noble metal nanoparticles such as Pt for practical applications.

Saha et al. prepared the composite electrodes by electrochemical deposition of Pt nanoparticles onto the surface of SnO₂ nanowires directly grown on the carbon fibers of a carbon paper [22]. In the comparison to a standard Pt/C electrode, the nanowire-based electrode exhibited higher electrocatalytic activity both for oxygen reduction reaction and methanol oxidation reaction. Figure 2-14 shows the SEM and TEM images of the SnO2

NWs grown on a commercially available carbon paper backing used in fuel cell applications. As shown in fig. 2-14 (a), a thin layer of high-density SnO2 NWs completely cover the surface of the carbon fibers in the carbon paper. Figure 2-14 (b) shows that the NWs have a straight-line morphology. Figure 2-14 (c) presents the TEM image of the Pt nanoparticles electrochemically deposited on the SnO₂ NWs. The successful deposition of

Figure 2-15 CVs for methanol oxidation reaction in 1 M H2SO4 aqueous solution with 2 M MeOH at Pt/SnO2

NW. carbon paper with 0.12 mg/cm² Pt loading and standard 30 wt % Pt/C electrode with 0.1 mg/cm² Pt loading. Potential scan rate 50 mV/s. The current normailized on the basis of Pt loading.

21

Pt nanoparticles indicates a good electrical contact between the SnO2 NWs and the carbon fibers. Furthermore, the electrochemical performance of the Pt/SnO2 NW/carbon paper electrode for methanol oxidation was also examined and the corresponding results are shown in Fig. 2-15. Compared with the Pt/C electrode, the oxidation peak current for the Pt/SnO2 NW/ carbon paper electrode is about 71.5 mA/mgPt, which is higher than that of the Pt/C electrode, suggesting a higher utilization of Pt for methanol oxidation reaction.

The significant improvement in catalytic activites of the Pt/SnO2 NW/carbon paper composite electrode may be attributed to the following: (1) the unique 3D structure and electronic properties of SnO2 NWs, (2) strong interaction between Pt catalyst particles and the SnO₂ NW surface, (3) synergies resulting from the combined properties of Pt nanoparticles and SnO2 NW supports, and (4) low impurities of SnO2 NW compared to Vulcan carbon XC-72 which contains a significant amount of organosulfur impurities, which can poison the Pt metal.

2.5 Adsorbed CO on Pt surface

CO adlayers of different coverages were produced by a dosing preocedure and their electrooxidative removal was studied by Koponen et al [42]. Figure 2-16 shown that the CO stripping voltammetry for pure catalyst ink electrode in 0.5 M H2SO4 for submono and saturation adlayers which is created by dosing of CO containing electrolyte for various periods of time. CO-oxidation processes present at 0.79 and 0.85-0.88 V. For CO on Pt(111), two main oxidation peaks are found, one minor and stationary at intermediate coverages, and one that shifts to the higher potential as a function of increasing CO coverage. The potential values of the stationary are 0.75 V and 0.83 - 0.88 V for the shifting peak. The stationary peak takes place at saturation coverage for the Pt(111) electrode. The stationary low potential peak may be contributed by (110) sites on the

22

crystallite; for saturated adlayers of CO on Pt(110) a single oxidation peak at 0.75 V is found. Another possibility may result from the oxidation of CO on edge sites on Pt (111) planes of the crystallite. The behavior of CO adsorbed on edge sites is similar to the CO present in loosely packed submonolayers on Pt(111) electrodes and this results oxidized at a lower potential.

2.6 Electrochemical behavior of Palladium electrode

Palladium presents high catalytic activity towards several electrochemical processes, such as the methanol oxidation reaction (MOR), hydrogen absorption and adsorption and the hydrogen evolution reaction (HER). The crucial difference characteristic between Pd and Pt is the capability of adsorbing the hydrogen [43]. The hydrogen absorption into Pd takes place in the potential range of the under potential deposition of hydrogen (UPD H).

However, desorption of hydrogen occurs at potentials higher that the onset potential of HER. On the other hand, the potential range of hydrogen absorption and adsorption is depended on the different metallic alloys and should be carried out by CV measurement.

Figure 2-16 CO stripping voltammetry for Pt catalyst ink electrode in 0.5 M H₂SO₄ for submono and saturation adlayers created by dosing of CO containing electrolyte for various periods of time. CO was dosed at 0.2 V and the scan rate was 50 mVs^-1.

23

Hydrogen adsorptions also one of the factors influencing the shape of CV profiles.

Grden et al. exhibited a typical CV profile of Pd(poly) electrode in aqueous H2SO4 [43].

Figure 2-17 reveals that the hydrogen absorption and adsorption take place at the same potential range and commence at ca. 0.3 V vs. RHE. It also exhibits that the oxidation of Pd occurs at ca. 0.65 V vs. RHE and the reduction of Pd oxide starts at ca. 0.85 V vs. RH E.

In the case of Pd metal, the potential region of the metal oxidation and the oxide reduction is not the same, as shown in figure 2-17. The electrochemical behavior is the nature of the surface compound formed and then reduced. Its chemical composition and structure depend on the oxidation conditions. Palladium oxides are prevailing materials and act as catalysts in various electrochemical studies but sometimes also as inhibitors in some processes, such as oxygen reduction reaction (ORR). The electro-oxidation of Pd and reduction of Pd oxide lead to changes in the surface morphology because soluble Pd compounds or oxide may generate during the electro-dissolution of Pd.

Figure 2-17 Cyclic voltammograms for Pd(poly) is in 0.5 M aqueous H2SO4 solution. The temperature was at 298 K and the scan rate is 50 mVs^-1. The CV profiles exhibit the oxidation of Pd begins at 0.65 V vs. RHE and graduate onset of hydrogen absorption upon decrease of the potential limit from 0.4 to 0.16 V, with the process commencing at 0.3 V.

24

2.7 Bi-functional mechanism

2.7.1 Langmuir-Hinshelwood mechanism and Eley-Rideal mechanism

In the bi-functional mechanism model, hydroxyl surface groups can oxidize adjacent CO adspecies on the Pt catalyst, thereby avoiding CO poisoning. The reaction mechanism can be described by the Langmuir-Hinshelwood mechanism. In an Eley-Rideal surface reaction, the reaction product is formed by direct collision of a reactant species from the solution phase with an adspecies. In the case of CO electro-oxidation in the KOH solution, the carbonaceous adspecies on the Pt surface can be readily oxidized by the abundant OH -ions in the electrolyte [23]. The CO electro-oxidation reaction does not only lie in the surface concentration of OHad species and COad (L-H mechanism), but is strongly affected between the CO_ad and OH^－ ions from the electrolytes (E-R mechanism).

Poisoning of the Pt surface by CO-like species produced during methanol oxidation is the major reason for the low rate of reaction. In order to solve the problem, Pt-based alloys or Pt/metal oxide composites are employed as a catalyst to increase the MOR activity and CO tolerance of electrode, based on a bi-functional mechanism. Song et al.

Figure 2-18 Schematic diagram illustrating the Langmuir-Hinshelwood mechanism and the Eley-Rideal mechanism.

25

synthesized a new carbon nanotube-supported sulfated TiO2 and Pt (Pt-S-TiO2/CNT) by improved sol-gel and ethylene glycol reduction methods [44]. Careful structural design allowed the Pt nanoparticles to homogeneously disperse on a sulfated TiO2 layer, which means that all of the Pt nanoparticles could be in direct contact with sulfated TiO₂. This special catalyst structure, as shown in Fig. 2-19, and the high proton conductivity of sulfated TiO₂ increased the catalytic activity and CO tolerance of Pt for methanol electro-oxidation. The CO stripping voltammograms showed that addition of sulfated TiO₂ is more beneficial for CO electro-oxidation. This can be attributed to the presence of more absorbed OH (OHad) on the S-TiO2/CNT electrode and the carefully designed structure of Pt-S-TiO₂/CNT electrode. These two factors are favorable for a bi-functional mechanism (L-H mechanism).

2.7.2 Surface electrochemistry of CO on Pt

Although the nature and adsorption site occupancy of CO_ad are strongly dependent on the applied potential, the mechanism for COad oxidation on Pt (111) is independent of electrode potential; e.g., CO_ad reacts with oxygen containing species through either a non-competitive or competitive Langmuir-Hinshelwood type reaction to form CO2. The physical state of oxygen-containing species is still uncertain (bulk H₂O, adsorbed H₂O or adsorbed OH). Markovic et al. have proposed that COad is oxidized by OHad, the latter

Figure 2-19 Schematic diagram of the structure of the novel catalyst Pt-S-TiO2/CNT.

26

species resulting from oxidative water decomposition in acid solution or from OH -discharge in alkaline solution [45]:

COad + OHad = CO2 + H⁺ +e^- (2-7) Nevertheless, regardless of the true nature of the oxygen-containing species it was suggested that the kinetics of reaction do not depend only on the surface concentration of CO_ad and OH_ad species, but are strongly affected by a delicate balance between the coverage of and anions from electrolytes.

Recall that although the nature of CO_ad changes with electrode potential, the mechanism for COad oxidation on Pt (hkl) follows the L-H reaction mechanism in the entire potential range [45]. Bergelin et al. suggested that in the preignition potential region CO_ad oxidation in CO free solution cannot proceed through the L-H mechanism, but rather through an Eley-Rideal mechanism, i.e. reaction between CO_ad and “activated” water molecules in the electrical double-layer [46]. Markovic et al. proposed that CO oxidation on Pt(111) will be considered to proceed through the L-H mechanism, in which the kinetics are strongly dependent on the delicate balance between the surface coverage of COad, OHad

and anions from supporting electrolytes [35].

2.8 Charge transfer between Pt particles and metal oxide support

Croy et al. presented the decomposition of methanol over Pt nanoparticles supported on a series of oxide powder [47]. The samples tested may be roughly grouped in two categories consisting of large and small Pt particles deposited on reducible (CeO₂, TiO₂) and non-reducible (SiO2, ZrO2, Al2O3) support. Figure 2-20 shows XPS spectra of Pt deposited on the different oxide powder supports measured after annealing at 500^oC. In fig. 2-13 (a), the solid lines indicate the positions of the main core-level peaks of metallic Pt at 71.1 eV (4f_7/2) and 74.3 eV (4f_5/2), the dashed lines Pt²⁺ in PtO (73.3 and 76.6 eV), and

27

the dotted lines Pt⁴⁺ in PtO₂ (75.0 and 78.8 eV). For the Pt/TiO₂ sample, it is predominantly metallic with the 4f7/2 appearing at ~ 70.5 eV. This corresponds to a negative binding energy shift of ~ 0.6 eV with respect to the bulk value of 71.1 eV. Such negative energy shifts can be explained by charge transfer to the particle from the support due to delocalized electron distributions arising from oxygen vacancies, or small particles with a large number of surface atoms having reduced coordination number. The Pt/CeO2

sample appears highly oxidized (mainly Pt⁴⁺) and the higher binding energies indicate a string interaction between the CeO2 support and the Pt particles. The possible formation

Figure 2-20 (a) Pt-4f core level XPS spectra of Pt nanoparticles supported on: (from top to bottom) TiO₂, ZrO₂, SiO₂, CeO₂. (b) Pt 4d_5/2 from Pt/Al₂O₃. All spectra were measured after removal of the encapsulating polymer by annealing in air at 500 ^oC

28

of Pt-Ce alloys might explain the anomalously large binding energies observed in the XPS data of these samples.

The XPS data indicate that for similarly sized particles that state of oxidation of Pt depends on the support. Because all samples underwent identical thermal treatments, the stability of Pt^δ+ species can be affected by the choice of support. This suggest that for MeOH decomposition, or perhaps in general, for reactions not involving the dissociation of O2, the reducibility of the support plays a secondary role to the more important parameters of particle size and oxidation state of Pt. The role of the support is that of a stabilizer, a provider of preferential/additional sites of interaction, and a mediator among the different oxides of Pt.

29

Chapter 3 Experimantal

3.1 Experimental flowchart

Figure 3-1 illustrates the experimental flowchart for the fabrications and analyses of the Pt/porous TiO2, Pt/karst-Ni and Pt/PdO electrodes. The porous TiO2 and the karst-Ni electrodes were fabricated on p-type Si wafer; the PdO nanoflakes electrode was fabricated on graphite carbon cloth. Pt nanoparticles were electro-deposited on the three electrodes by different galvanostatic pulse plating condition, respectively, for the study on methanol oxidation reaction (MOR). Measurements of electrocatalytic activity of the Pt/porous TiO₂ and the Pt/PdO electrodes were in acidic system, and it of the Pt/karst-Ni electrode was in alkaline system. The cyclic voltammetry (CV) measurement and CO stripping

Figure 3-1 Experimental flowchart for the fabrications and analyses of the Pt/porous TiO2, Pt/karst-Ni and Pt/PdO electrodes.

30

measurement were used to study the ability toward methanol electro-oxidation and CO tolerance, respectively, of electrodes. The chronoamperometry measurement was utilized to estimate the electro-activity stability of the electrode in the MOR electro-activity with long time. The surface morphology was examined by scanning electron microscopy (SEM). The crystallinity and the chemical composition of electrodes were studied by x-ray diffractomery (XRD) and x-ray photoelectron spectroscopy (XPS), respectively.

The microstructure and particle size distribution of Pt nanoparticles were analyzed by transmission electron microscopy (TEM).

3.2 Preparation of Pt nanoparticles deposited on porous TiO

support

Figure 3-2 Experimental flowchart for the fabrications the Pt/TiO₂ electrode.

3.2.1 Preparation of the porous TiO

support

To prepared the porous titania support, a Ti thin film 100 nm thick was first deposited on a p-type 6-inch Si wafer of low resistivity (0.002 Ω-cm) by electron beam evaporation (e-beam) deposition. The Ti thin film coated silicon wafer was then immersed in the aqueous solution of 10 M NaOH at 80^oC for 25 minute, followed by a rinse with 0.1 M HNO₃. Before the Pt nanoparticle deposition on the porous TiO₂ support, the as-synthesized thin film was annealed in vacuum (10^-7 torr) at various temperatures to improve the electrical conductivity of the porous support.

31

3.2.2 Pulse electrodeposition of Pt nanoparticles on the TiO

support

Pt nanoparticles were electrodeposited on the TiO₂ support in the aqueous solution of 0.001 M H2PtCl6 (200 ml)/ 0.33 M HCl (2 ml) at room temperature by galvanostatic pulse plating in a three electrode cell system with a saturated calomel reference electrode (SCE).

The vacuum annealed porous TiO2 support was used as the working electrode and a Pt plate as the counter electrode. The particle size and dispersion of Pt nanoparticles can be controlled by tuning the pulse height of the applied current and the pulse duration. The time durations for the high current pulse (4 mA) and the low current pulse (-20 mA) were 2 and 1 ms, respectively. A total of 3000 pulse cycles was performed to deposit Pt nanoparticles on the porous TiO₂ support. For comparison, Pt particles of larger size were also electrodeposited on a blanket TiO2 surface, which was prepared by annealing a Ti thin film 10 nm thick on the Si substrate in oxygen ambient at 600^oC for 5 min., followed by vacuum anneal at 600^oC for 1 h. A Pt island film and a blanket Pt thin film of 5 nm thickness were also deposited by pulse electrodeposition and e-beam deposition, respectively, on a metallic Ti thin film, which was e-beam deposited on the Si wafer. The mass loading of the Pt catalyst on the support was determined by inductively coupled plasma mass spectroscopy (Thermo X Series II).

3.3 Preparation of Pt nanoparticles deposited on PdO nanoflake support

Figure 3-3 Experimental flowchart for the fabrications the Pt/PdOelectrode.

32

3.3.1 Preparation of the PdO nanoflake support

PdO nanoflakes were deposited on carbon cloths at room temperatures by reactive sputter deposition (RSD) in a radio frequency magnetron sputter deposition system. The RSD condition for the PdO nanoflake thin film has been described in detail previously [48].

The palladium target was 2 in. in diameter with a purity of 99.99 %. The RSD was performed with a gas mixture of Ar (20 sccm) and O₂ (20 sccm) at the working pressure of 9×10^-3 torr and an RF power of 50 W. Before the PdO deposition, the carbon cloth was immersed in the aqueous solution of 5 M HNO₃ at 25 ^oC for 5 hours, followed by a rinse with deionized (DI) water.

3.3.2 Pulse electrodeposition of Pt nanoparticles on the PdO support

Pt nanoparticles were electrodeposited on the PdO nanoflake support in a mixed aqueous solution of 0.001 M H2PtCl6 at room temperature by galvanostatic pulse plating in a two-electrode cell system. The PdO nanoflake support was used as the working electrode and a Pt disk (1.5 cm in diameter) as the counter electrode. During the pulse electrodeposition, the time durations for the positive current pulse (5 mA) and the negative current pulse (-10 mA) were 1 and 2 ms, respectively. A total of 400 pulse cycles were performed to deposit Pt nanoparticles on the PdO nanoflake support. For comparison, we also pulse-electrodeposited Pt particles on the carbon cloth support (denoted by Pt/C), and prepared a 10 nm thick Pt thin film on the carbon cloth by e-beam deposition (denoted by blanket-Pt). For the preparation of the Pt/C electrode, following Pt electrodeposition conditions were used: the DC pulse current: -1.0 mA, the pulse duration: 100 ms, the cycle period: 200 ms, and the total number of pulse cycles: 200. The Pt loading of the electrodes was determined by inductively coupled plasma mass spectrometer. Under the sample preparation conditions described above, the Pt/PdO, the Pt/C and the blanket -Pt

Pt nanoparticles were electrodeposited on the PdO nanoflake support in a mixed aqueous solution of 0.001 M H2PtCl6 at room temperature by galvanostatic pulse plating in a two-electrode cell system. The PdO nanoflake support was used as the working electrode and a Pt disk (1.5 cm in diameter) as the counter electrode. During the pulse electrodeposition, the time durations for the positive current pulse (5 mA) and the negative current pulse (-10 mA) were 1 and 2 ms, respectively. A total of 400 pulse cycles were performed to deposit Pt nanoparticles on the PdO nanoflake support. For comparison, we also pulse-electrodeposited Pt particles on the carbon cloth support (denoted by Pt/C), and prepared a 10 nm thick Pt thin film on the carbon cloth by e-beam deposition (denoted by blanket-Pt). For the preparation of the Pt/C electrode, following Pt electrodeposition conditions were used: the DC pulse current: -1.0 mA, the pulse duration: 100 ms, the cycle period: 200 ms, and the total number of pulse cycles: 200. The Pt loading of the electrodes was determined by inductively coupled plasma mass spectrometer. Under the sample preparation conditions described above, the Pt/PdO, the Pt/C and the blanket -Pt