Electrocatalytic activity

The electrochemical surface area (ESA) of the Pt film/Si and the Pt/ACNC electrodes was evaluated by the cyclic voltammetry (CV), which was performed in argon saturated 1M H2SO4 aqueous solution using a Pt wire as the counter electrode and a saturated calomel electrode (SCE) as reference electrode.

Figure 5-7 shows the room temperature cyclic voltammograms of (a) Pt/ACNC and (b) Pt film/Si electrodes measured at 25 mV s^-1 in 1 M H2SO4.

Typical CV curves of the both electrodes are shown in Fig. 5-7. The ESA values obtained from hydrogen adsorption/desorption peaks are 412 and 39 m² g^-1 for Pt/ACNC and Pt film/Si, respectively. The ESA for the Pt/ACNC higher than that of Pt film/Si, most likely due to the smaller size and much better dispersion of the Pt nanoparticles on ACNC arrays.

Figure 5-8 shows the CO stripping cyclic voltammograms of the Pt/ACNCs electrode in a CO saturated 1 M H2SO4 solution. The scan rate was 25 mV s^-1.

The electroactive surface areas (ESAs) of the Pt /ACNC were further reinvestigated by CO-stripping CV, performed in a CO saturated 1 M H2SO4 aqueous solution using a Pt wire as the counter electrode and SCE as the reference electrode. For comparison, the ESA of a Pt film/Si electrode, prepared by electrodepositing a continuous Pt film on the flat Si substrate, was also measured. Figure 5-8 shows a representative CO-stripping cyclic voltammogram for the Pt-ACNC electrode. A very high ESA of 327m²/g was calculated for the Pt/ACNC electrode by integrating the electro-oxidation peak of adsorbed CO molecules, assuming an

electro-oxidation charge of 420 μC-cm⁻² for a monolayer of CO adspecies on a smooth Pt surface [90]. The ESA of the Pt/ACNC electrode was about eight times higher than that of the Pt film/Si electrode and much higher than that of many previously reported electrodes of carbon materials [85, 111, 112]. The high ESA ascribes to the large surface area of the nanocone array and the highly dispersed Pt nanoparticles. The discrepancy between surface area by hydrogen adsorption/desorption and from CO-stripping might be due to experimental error.

Efficient removal of the CO adspecies from the Pt catalyst surface is crucial for minimizing the poisoning effect in DMFCs. The CO stripping voltammograms of Fig. 5-8 also provide information about electro-oxidation of CO adspecies on the electrocatalyst. The onset potential of CO electrooxidation on the Pt/ACNC and the Pt film /Si electrodes was about ~0.36 and ~0.51 V, respectively. Improved electro-oxidation activity of CO adspecies on the Pt/ACNC electrode may be ascribed to the bi-functional mechanism, which describes that effective electrooxidation of CO adspecies can be achieved by reaction with neighboring oxygen containing functional groups. The XPS analysis discussed above showed plenty of oxygen containing adspecies on the α-C layer. The nanometer size of Pt catalyst particles made it easier for surrounding oxygen-containing groups on the ACNC support to access and interact with CO adspecies on the Pt nanoparticles, and allowed fast migration of adspecies over the particle surface. Therefore electro-oxidation of CO adspecies, in particular-C-OH adspecies, could readily proceed via the Langmuir-Hinshelwood mechanism. The lower onset potential of CO electro-oxidation and the higher ESA clearly demonstrate that the Pt/ACNC electrode effectively improves electro-activity of the Pt catalyst toward MOR.

MOR activity of these two electrodes in the 1 M CH3OH + 1 M H2SO4 aqueous solution at room temperature was studied by CV. Figure 5-9 (A) shows the current density and mass activity of the Pt film/Si, and Pt/ACNC electrode. As Fig. 5-9 (A) shows, the Pt/ACNC had a higher mass activity and current density than the Pt film /Si electrode. The current ensity and

mass activity of the Pt/ACNC electrode were ~4 and ~10 times that of the Pt film/Si electrode, respectively. Even compared with many reported electrodes composed of carbon materials, such as carbon-coated anatase TiO2 composites [85] and nanoporous materials synthesized by electrochemical lithiation [48],

Figure 5-9 (A) Current density and mass activity of the Pt/ACNC electrode (solid rectangle) and the Pt film/Si electrode (open rectangle). (B) Loss of the Pt electrochemical surface area of (a) the Pt/ACNC electrode and (b) the Pt film/Si electrode as a function of the number of CV cycles in argon saturated 1 M H2SO4 aqueous solution at room temperature (scan rate: 25 mV s^-1).

the Pt/ACNC electrode also exhibited better MOR performance in terms of catalyst mass activity and current density. The excellent MOR activity of the Pt/ACNC electrode can ascribe to the large electrode surface area and the nanometer size of Pt nanoparticles. The geometric shape of ACNC support provided a surface area for Pt loading five times larger than that for the flat Si surface. The highly dispersed Pt nanoparticles also further increased the active surface area on the Pt/ACNC electrode, thus increasing mass activity. Furthermore, the nanometer-sized Pt catalyst particles enhanced electro-oxidation of CO adspecies via the bifuncitonal mechanism as discussed above, creating more active sites for MOR and thus improving MOR efficiency.

Electrocatalytic stability of the Pt film/Si and the Pt/ACNC electrodes for MOR was evaluated by CV within the potential range -0.4–1.2V (vs. SCE) at a scan rate of 25 mVs^-1 in an argon saturated 1 M H2SO4 aqueous solution for more than 1000 cycles. Figure 5-9(B) shows the ESA loss as a function of the number of CV cycles for the two electrodes. The Pt/ACNC electrode showed a much more moderate reduction in the ESA during the repeating CV scans. The ESA of the Pt/ACNC electrode decreased by ~12% after 1000 CV cycles in the acid solution, while the Pt films/Si electrode showed a decrease of more than 50%. Loss of the ESA of the Pt catalyst might result from two pathways, Pt particles detachment from the support and Pt catalyst dissolution during electrochemical measurements. We did not observe a noticeable decrease in Pt nanoparticle density after electrocatalytic stability measurement according to TEM analysis (not shown), indicating that Pt nanoparticle detachment from the ACNC support did not take place or was insignificant. Thus, the gradual decrease in the ESA of the Pt/ACNC during the stability test was likely due to Pt nanoparticle dissolution into the electrolyte solution at the high potential range during the CV sweep, which had a maximum of 1.2 V vs. SCE [113, 114]. The much smaller ESA loss for the Pt/ACNC electrode implied that Pt nanoparticles were relatively stable in the acidic electrolyte during the voltage scan.

The nanocrystalline graphitic structure in the α-C layer as revealed by Raman scattering

spectroscopy might play a crucial role in making the Pt /ACNC electrode much more stable toward MOR than the Pt film/Si electrode. In the broad Pt(4f) doublet XPS spectrum shown in Fig. 5-5(C), the Pt(4f7/2) peak reached a maximum at 70.5 eV, which was ~0.8 eV smaller than that of bulk Pt. A red shift of 0.5 eV for the Pt(4f) doublet peaks have been reported for nanometer-sized Pt colloids supported on highly oriented pyrolytic graphite [115]. The negative shift of the Pt(4f) doublet peak for Pt nanoparticles is generally ascribed to the reduced coordinated number of surface atoms in a nanometer sized cluster [116, 117]. Charge transfer between Pt nanoparticles and the support can also positively or negatively shift the Pt(4f) binding energies, depending on the chemical property of the support [118]. The large red shift observed in the study might be a combination effect of the particle size effect and the π-electron donation from the support to Pt nanoparticles. It has been widely reported that increasing the graphitization degree of carbon supports strengthens the metal-support interaction and enhances resistance to support oxidation and catalyst agglomeration [99-101, 119]. Charge transfer between π-sites of the α-C layer and Pt nanoparticles significantly alters the electronic structure of nanosized Pt catalyst particles, thereby enhancing electrochemical stability of the Pt catalyst and thus mitigating Pt dissolution due to electrooxidation. In addition to chemically stabilizing Pt nanoparticles, the charge transfer interaction also enhances the adhesion of electrodeposited Pt nanoparticles to the ACNC support, thus preventing Pt nanoparticles from separating from the ACNC electrode.

The slow ORR rate on the cathode is always a concern for DMFC applicability. We have also studied electrocatalytic activity of the Pt/ACNC electrode toward ORR by the CV measurement. Figure 5-10(A) shows typical ORR polarization curves of the Pt film/Si and the Pt/ACNC electrode, obtained in an O2 saturated 1 M H2SO4 aqueous solution at room temperature. The ORR half-wave potentials for the Pt film/Si and Pt/ACNC electrodes were

~0.52 and ~0.60 V, respectively. The peak ORR current densities of the Pt film/Si and Pt/ACNC electrodes were -1.2 and -7.4 mA cm^-2, respectively. The current density of the

Figure 5-10 (A) Cyclic voltammograms of ORR in O2 saturated 1 M H2SO4 aqueous solution at 25^oC for (a) the Pt/ACNC electrode and (b) the Pt film/Si electrode; and (B) ORR current density loss as a function of the number of CV cycles in oxygen saturated 1 M H2SO4 at room temperature for (a) the Pt/ACNC electrode and (b) the Pt film/Si electrode (scan rate: 25 mV s^-1).

Pt/ACNC electrode was six times higher than that of the Pt film/Si electrode, indicating that the Pt/ACNC electrode significantly enhanced the ORR rate. Compared with many previous reported electrodes, such as Pt catalysts with gold clusters [120], Pt3Ni(111) catalysts [121], supportless Pt and Pt-Pd nanotubes [122] and dealloyed Pt-Cu-Co nanoparticles [123], the Pt/ACNC electrode showed a higher peak current density as well. The great improvements on ORR electroactivity of the Pt catalyst on the ACNC support can be attributed to similar factors enhancing MOR electroactivity. The well dispersed Pt nanoparticles on the ACNC support provided a large ESA leading to a faster reaction rate of ORR. The size of Pt nanoparticles affects the potential distribution within the double layer around the catalyst and also enhances the mass transport of electroactive species on the Pt catalyst [124].

The Pt/ACNC electrode also demonstrated excellent electrocatalytic stability for ORR.

The electrocatalytic stability test was performed in an O2 saturated 1 M H2SO4 aqueous solution at 25^oC. The CV measurement of more than 1000 cycles was conducted in the potential range between 0.2-0.8 V vs. SCE at a scan rate of 25 mV s^-1. Figure 5-10 (B) shows the ORR current density loss for the Pt film/Si and the Pt/ACNC electrodes as a function of the number of CV cycles. The decrease in ORR activity of the Pt/ACNC and the Pt film/Si electrodes after 1000 cycles was about 17% and >60%, respectively. The better ORR activity stability of the Pt/ACNC electrode may also be due to the unique chemical properties of the ACNC support, composed of nanocrystalline graphitic structures and full of oxygen containing adspecies.

5.6 Summary

This study has shown the excellent electrochemical performance of Pt nanoparticles pulse-electrodeposited on the highly ordered ACNC array, fabricated by the AAO templation.

Because Pt nanoparticles were well dispersed on the ACNC support with a large surface area,

the Pt/ACNC electrode had a large ESA. The Pt-ACNC electrode exhibited high electrocatalytic activity and stability toward both ORR and MOR. According to Raman scattering and XPS analyses, the ACNC array was composed of nanocrystalline graphitic structures, and full of oxygen containing adspecies. The uniform dispersion of Pt nanoparticles on the α-C layer and the high resistance of the Pt/ACNC electrode against CO poisoning resulted from the presence of oxygen-containing adspecies. The nanocrystalline graphitic structure in the α-C layer was able to electronically stabilize the Pt nanoparticles via charge transfer, thereby enhancing electroactivity stability of the Pt/ACNC electrode.

Chapter 6

Novel New Approach to the Synthesis of Nanoporous Graphitic Carbon as a Unique Electrocatalyst Support

for Methanol Oxidation 6.1 Introduction

Recently, the synthesis of nanostructure carbon materials has great potential for applications in electronic, [125, 126] catalytic, [127, 128] hydrogen-storage, [129, 130]

electrochemical double-layer capacitors, [131] and lithium ion batteries, [132]. In an effort to improve catalyst activity and stability, Pt-Ru based catalysts, which are the most practical catalysts for fuel cell applications, have been synthesized for the purpose of reducing catalyst particle size (nanolevel) and increasing the surface area of the catalyst deposited on support.

The one best way to accomplish this is to create 3D nanoporous graphitic carbon (g-C) structure. Several methods have been reported for the synthesis of carbon nanostructure with high surface area by using templates, stabilizers, or surfactants [96, 133, 134]. Therefore, preparation of ordered porous carbon with graphitic pore walls is of importance. To prepare ordered graphitic porous carbons, some unconventional carbon precursors such as mesophase pitch [131], acenaphthene [134], polyvinyl chloride [135], naphthalene, anthracene, pyrene [136], and polypyrrole [137] have been employed to infiltrate the pores of porous templates.

However, infiltration and polymerization using these liquid carbon precursors are either difficult or time consuming because repeated infiltration and polymerization are required in order to obtain an ordered carbon replica. On the other hand, polymerization and pyrolysis of the carbon precursors during high-temperature carbonization often lead to the emission of a large amount of small molecules such as H2O, which can deteriorate the pore structure of the templates [138], thus the structural ordering of the resultant carbon [139, 140]. However, the

synthesis of graphitic carbon with high surface area has been a great challenge and method for economically large scale production is still the vital task.

Herein, we describe new, simple and efficient method for the synthesis of a 3D nanoporous g-C structure by adamantane flame, which was utilized as a support for Pt and Pt50–Ru50 alloy catalysts at low temperature. Moreover, the electrochemical performance of the Pt50–Ru50/3D nanoporous g-C shows the potential application in liquid feed fuel cells.

6.2 Synthesis of 3D nanoporous g-C, Pt/3D nanoporous g-C and Pt

-Ru

/3D nanoporous g-C

The synthesis process of the 3D nanoporous g-C structure is illustrated in Fig. 6-1. The synthesis of 3D nanoporous g-C on the Si surface consists of three consecutive steps. First, the flat Si substrate was washed by acetone and deionized water (>18 MΩ), then it was treated with a 10 wt% HF at room temperature for 5 min to remove the thin native oxide layer and dry blown with nitrogen. Then, the polished side of Si substrate was immediately placed in front of adamantane (C10H16) flame for 7 min to deposit of thick layer of 3D nanoporous g-C.

Further, for the better adhesion between Si and 3D nanoporous g-C, the sample was heated at 300^○C for one hour. A precursor solution containing 10 ml acetone and H2PtCl6 was prepared.

And then, the solution was poured drop by drop onto 3D nanoporous g-C at 100^○C for 2 min for the deposition of Pt. To deposit Pt50–Ru50 alloy catalyst on the 3D nanoporous g-C, 10 ml acetone, H2PtCl6 and RuCl3 solution was used, following the similar procedure as of Pt. The Pt/3D nanoporous g-C, and Pt50-Ru50/3D nanoporous g-C electrodes was rinsed thoroughly with deionized water to remove residual chlorine ions after Pt and Pt50-Ru50 deposition, and allowed to dry before using for electrochemical measurements.

Figure 6-1 shows the synthesis scheme of 3D nanoporous g-C: (a) the Si wafer was cut into 2 x 2 cm² dimension (b) SiO2 layer was removed by treating with 10 % HF for 5 min and dried by N2 gas (c) the cleaned sample was immediately placed in front of flame adamantane for 7 minute, and (d) finally, sample was heated at 300^○C for one hour to obtain a good depositing of 3D nanoporous g-C over the silicon substrate.

6.3 Physico-chemical characterization

Scanning electron microscopy (SEM) was used to characterize the surface morphology of the synthesized 3D nanoporous g-C. Figure 6-2 shows a typical SEM image of 3D nanoporous g-C where the Pt and Ru content is zero. As is evident from the image, the Si substrate is well covered by 3D nanoporous carbon with pore sizes of 100–900 nm.

Figure 6-2 shows the SEM image of the resulting 3D nanoporous g-C.

The graphitization of the synthesized 3D nanoporous carbon was characterized by Raman spectroscopy. The Raman spectra of the 3D nanoporous carbon in the frequency range between 1200 and 1800 cm^-1 are shown in Fig. 6-3. The two main peaks at 1594 and 1359 cm

-1 illustrated in the spectra corresponds to G and D bands, respectively. The G band represents the E2g vibration of g-C with a sp² electronic configuration and D band represents the A1g

mode of diamond-like carbon with a sp³ electronic configuration. To determine the degree of graphitization, we estimated the intensities of D and G peaks from their peak heights and hence their ratio (ID/IG). As shown in Fig. 6-3, the ID/IG ratio less than 1, indicate the better graphitization of 3D nanoporous carbon [141]. Raman spectra suggest formation of g-C in 3D nanoporous carbon structure.

Figure 6-3 Raman spectrum of 3D nanoporous g-C deposited on Si substrate.

SEM was further used to characterize the surface morphologies of the synthesized Pt/3D nanoporous g-C, and Pt50-Ru50/3D nanoporous g-C. Fig. 6-4 (A) shows a typical SEM image

of Pt/nanoporous g-C where 3D nanoporous g-C surface is deposited by nanoparticles Pt. The pore size distribution of the synthesized 3D nanoporous g-C is 100-800 nm.

Figure 6-4 shows the plane-view SEM images (A) Pt/3D nanoporous g-C, and (B) typical Pt50-Ru50/3D nanoporous g-C.

(A)

(B)

Figure 6-5 shows the TEM images (A) Pt/3D nanoporous g-C, and (B) typical Pt50-Ru50/3D nanoporous g-C.

It is evident that the some pores are filled due to the large size of the Pt nanoparticles.

Figure 6-4 (B) shows a typical SEM image of Pt50-Ru50/nanoporous g-C, where 3D nanoporous g-C surface deposited by nanoparticles Pt and Ru. From the SEM image, we can see that the Pt50-Ru50/3D nanoporous g-C has more porous surface than the Pt/3D nanoporous g-C, showing has higher surface area than Pt/3D nanoporous g-C. Because the Pt and Pt50 -Ru50 alloy catalysts on the g-C support were very small in size, Pt and Pt-Ru alloy catalyst was hardly observed from the SEM images of Fig. 6-4 (A and B).

The transmission electron microscopy (TEM) measurements were performed to reinvestigate the presence of Pt and Pt-Ru nanoparticles. A TEM image of the Pt/3D nanoporous g-C is shown in Fig. 6-5 (A). As seen from TEM images, Pt nanoparticles showing the agglomeration. And the Pt nanoparticles were ~ 20 nm in size. Figure 6-5 (B) shows the TEM image of the g-C with the deposited Pt-Ru alloy catalysts. It can be clearly seen that Pt-Ru alloy catalyst was dispersed on g-C. In addition, the composition of individual particles was probed using energy-dispersive X-ray analysis (EDX). An EDX spectrum of Pt

(B)

(A)

and Pt50-Ru50 alloy catalysts is shown in Fig. 6-6 and 6-7. The EDX indicated that Pt and Pt-Ru alloy catalysts was deposited on the g-C.

Figure 6-6 shows the EDX spectrum of the Pt catalyst on the g-C.

Figure 6-7 shows the EDX spectrum of the Pt50-Ru50 alloy catalyst on the g-C, showing the presence of Pt and Ru nanoparticles on the 3D nanoporous g-C.

X-ray photoelectron spectroscopic (XPS) system is a convenient tool to investigate the presence of Pt and Pt-Ru alloy catalysts on the g-C. The wide-scan XPS spectrum of the Pt/3D nanoporous g-C and Pt50-Ru50/3D nanoporous g-C are shown in Fig. 6-8. From the spectrum it is clear that the Pt and Pt-Ru alloy catalysts was deposited on the 3D nanoporous

g-C. The relatively strong O(1s) XPS signal suggests that the oxygen surface group was present on the 3D nanoporous g-C. This is greatly affects the nanoparticles deposition and electrocatalytic characteristics of the Pt and Pt-Ru nanoparticles. On the other hand, the weak signals of Si(2s) and Si(2p) were detected from XPS, indicates nanoporous g-C completely covered the Si surface. This is consistent with our SEM observation as shown in Fig. 6-4 (A and B).

Figure 6-8 shows the XPS survey spectrum of the 3D nanoporous g-C supported Pt and Pt-Ru alloy catalysts.

6.4 Electrocatalytic activity

The hydrogen adsorption and desorption was used to investigate the electrochemical active surface area (ECSA) of these electrodes. Figure 6-9 shows the CV curves of the Pt/3D nanoporous g-C and Pt50-Ru50/3D nanoporous g-C electrodes in a 1 M H2SO4 aqueous solution. The potential was scanned between -0.4 and 1.2 V (vs. SCE) at the potential scan rate of 25 mV s^-1.

Figure 6-9 shows the cyclic voltammograms of the Pt50-Ru50/3D nanoporous g-C and Pt/3D nanoporous g-C electrodes in aqueous solution in N2 saturated aqueous solution of 1 M H2SO4. The scan rate was 25 mV s^-1.

Figure 6-9 shows the cyclic voltammograms of the Pt50-Ru50/3D nanoporous g-C and Pt/3D nanoporous g-C electrodes in aqueous solution in N2 saturated aqueous solution of 1 M H2SO4. The scan rate was 25 mV s^-1.