XANES and EXAFS analysis

Since the activity for Pt complexes and their associated redox potentials are determined by the nature and number of ligands, the solution pH value therefore becomes very critical because the hydrolysis of Pt complexes is highly pH-dependant. The nature and number of ligand around the Pt cation in the electrolyte, as well as the oxidation state for Pt and Ru in the Ru@Pt nanoparticles can be inferred from XANES and EXAFS analysis. The Ru K-edge XANES spectra for the reference group (Ru/XC72/CC in pH 1, pH 2, and pH 8 solution) are demonstrated in Fig. 4.5, along with Ru and RuO₂ serving as the reference. The metallic Ru is established to have a K-edge absorption around 22,117 eV and its position is shifted to higher energy in oxidized state, as confirmed by the RuO₂. Spectra from the Ru/XC72/CC after immersing in HClO₄ (pH 1), de-ionized water (pH 7), and KOH (pH 8) suggested that some of the Ru existed in oxidized forms and their oxidation states

68

were rather similar. However, the exact oxidation state was still unknown but its value was estimated between 0 and +4 as indicated by the absorption edge between Ru and RuO2. According to literature, the Ru was prone to form surface oxide and hydroxide when it was immersing in a liquid electrolyte. Hence, presence of Ru in an oxidized state was not entirely unsupported.[124]

22110 22140 22170 22200

Normalized absorbance (a.u.)

Energy (eV)

Ru/XC72/CC pH 1 Ru/XC72/CC pH 7 Ru/XC72/CC pH 8 Ru powder

RuO2

Figure 4.5. The Ru K-edge XANES spectra of Ru, RuO2, and Ru/XC72/CC from reference group of pH 1, pH 7, and pH 8.

Fig. 4.6 exhibits the Ru K-edge XANES spectra for samples from group A and group B, respectively. In general, their spectra were similar to those obtained in Fig. 4.5. Hence, we concluded that the oxidation state of Ru remained unchanged regardless the electrolyte they encountered was HClO₄, KOH, or H₂PtCl₆. However, after further exploration, as shown in the inset of Fig. 4.6, it was found that the Ru from group A was slightly more oxidized than that in group B. This behavior was not unexpected as the samples from group B underwent a hydrogen reduction treatment leading to their reduced oxidized state. It is noted that from XRD in Fig. 4.2 and Fig. 4.3, the samples from group A and group B did not reveal the presence of RuO_x or Ru(OH)_x. However, in the XANES spectra, the Ru existed in an oxidized state anyway. We attributed this

69

discrepancy to the sensitivity of XANES that took into consideration of both metallic Ru at the core and oxidized Ru on the surface. Besides, the oxidized Ru was likely amorphous that obscured the XRD signal.

22110 22140 22170 22200

2 2 1 1 6 2 2 1 1 8

Normalized absorbance (a.u.)

Energy (eV)

Group A Ru@Pt/XC72/CC pH 1 Group A Ru@Pt/XC72/CC pH 2.2 Group A Ru@Pt/XC72/CC pH 8 Group B Ru@Pt/XC72/CC pH 1 Group B Ru@Pt/XC72/CC pH 2.2 Group B Ru@Pt/XC72/CC pH 8

Figure 4.6. The Ru K-edge XANES spectra of Ru@Pt/XC72/CC from group A and group B.

Similarity in Fig. 4.5 and Fig. 4.6 confirmed that the Ru maintained an identical oxidation state with or without displacement reaction. Therefore, during displacement reaction, we believed that the oxidized Ru left the surface in the form of dissolved complexes and freshly-exposed Ru adopted a similar oxidation state. To confirm our premise, it is necessary to carry out XANES analysis on

70

the electrolyte to verify the identity of dissolved Ru cations. Unfortunately, the Ru cations in the electrolyte after displacement reaction was so dilute that validation of their identity became rather difficult. From samples of group B, the oxidation state for Ru was not completely reduced to metallic form which suggested that the oxidized Ru state we observed was intrinsic to Ru after the Ru was exposed to electrolyte or ambient moisture.

The Ru K-edge Fourier-transformed EXAFS spectra for the reference group, group A, and group B are provided in Fig. 4.7. The peaks at 1.6 Å and 2.3 Å (without phase correction) were associated with Ru-O bond and Ru-Ru bond at the first shell coordination. The EXAFS fitting results are summarized in Table 4.3. For the reference group, the coordination number for Ru-O in all baths was around 2. This bonding between Ru and O was attributed to the formation of oxide or hydroxide on the Ru surface. On the other hand, the Ru-Ru coordination number was around 4-5, which was expected because earlier studies by Huang et al. reported a similar coordination number for Ru nanoparticles.[25] For the Ru@Pt/XC72/CC (sample of group A), we obtained a reduced Ru-O coordination number for pH 1 and pH 2.2 bath. On the other hand, the coordination number of Ru-Ru remained unchanged suggesting that the Ru core was likely intact. However, the EXAFS fitting did not reveal any Ru-Pt bond which excluded the possibility of Pt sitting next to Ru in the Ru@Pt nanoparticles. Nevertheless, once the hydrogen reduction treatment was imposed (as shown in group B), the Ru-Pt coordination number of 1.33, 1.08, and 0.82 was obtained for pH 1, pH 2.2, and pH 8 bath, respectively. At the same time, the number of Ru-O bond became smaller due to the hydrogen reduction treatment. Likewise, the Ru-Ru bond remained relatively unchanged because the Ru core was mostly unaffected. Based on previous work by Hu et al.[125], the Ru might form a Ru-O-Ru bonding near its surface. In our case, the bridged oxygen was likely to connect both Pt and Ru forming a Ru-O-Pt. As a result, the Ru-O-Pt was formed instead of Ru-Pt on the surface of Ru nanoparticles. Hence, it became reasonable that the Ru-Pt coordination was established as long as the bridged oxygen was removed during hydrogen reduction treatment.

71

1 2 3 4 5 6

FT magnitude (a.u.)

R (Angstrom)

Ru/XC72/CC pH 1 Ru/XC72/CC pH 7 Ru/XC72/CC pH 8

Group A Ru@Pt/XC72/CC pH 1 Group A Ru@Pt/XC72/CC pH 2.2 Group A Ru@Pt/XC72/CC pH 8

Group B Ru@Pt/XC72/CC pH 1 Group B Ru@Pt/XC72/CC pH 2.2 Group B Ru@Pt/XC72/CC pH 8

Figure 4.7. The Ru K-edge Fourier-transformed EXAFS spectra from Figs. 4.4 and 4.5.

Fig. 4.8 demonstrates the Pt L_III-edge XANES spectra for H₂PtCl₆ at pH 1, pH 2.2, and pH 8, respectively. Also shown is the XANES for Pt foil. The purpose for this measurement is to determine the nature of complexing ions for the Pt cations in different environments. From literature, electronic transitions from 2p3/2 to 5d is responsible for the Pt LIII-edge which is also known as white line.[126, 127] In general, the white line intensity is able to provide information on the oxidation state of Pt due to its relevance to the d-band vacancy. A larger white line intensity often infers more vacant d-band orbitals. From the XANES spectra, there appeared a high intensity white line for the Pt cations in H2PtCl6 solution at various pH values. These patterns were expected

72

as the Pt cations were possibly present in Pt⁴⁺. For the white line of Pt foil, its low intensity confirmed its metallic nature and its magnitude corresponded to the oxidation state of “0”.

11560 11570 11580 11590 11600 H2PtCl

6 pH 1 H2PtCl

6 pH 2.2 H2PtCl

6 pH 8 Pt foil

Normalized absorbance (a.u.)

Energy (eV)

Figure 4.8. The Pt L_III-edge XANES spectra of Pt foil and H₂PtCl₆ solution of pH 1, pH 2.2, and pH 8.

73

Table 4.3. EXAFS fitting parameters at the Ru K-edge for Ru/XC72/CC and Ru@Pt/XC72/CC under various conditions.

Path Coordination

74

For the identity of Pt atoms which were deposited onto Ru after the displacement reaction, the Pt LIII-edge XANES spectra in Fig. 4.9 can provide insightful information. In general, for samples from group A and group B, the XANES spectra clearly demonstrated a notable distinction in which signals from group A revealed a larger d-band vacancy than those from group B. This trend was expected as samples from group B underwent a hydrogen reduction treatment. For samples of group A, obviously the pH 8 sample revealed the largest d-band vacancy with its magnitude close to the H₂PtCl₆ solution in Fig. 4.9. This suggested that the Pt existed in +4 oxidation state on the Ru particle after the displacement reaction which inferred a physical adsorption process without involving the oxidation loss of Ru. In addition, for baths of pH 1 and pH 2, it appeared that the Pt still existed in the oxidized state albeit with a slightly reduced state. This trend was reversed from samples of group B in which the pH 8 sample demonstrated the lowest oxidative state with magnitude close to the metallic Pt foil shown in Fig. 4.8. However, the samples from pH 1 and pH 2.2 indicated that the Pt was present at a slightly oxidized state. We surmised that at pH 8 the Pt atoms clustered around themselves revealing a metallic behavior while at pH 1 and pH 2.2, the Pt atoms were intermixed with Ru forming a quasi-alloying state instead. Earlier, it was suggested by Sasaki et al. that in alloying of PtRu the Pt exhibits a larger d-band vacancy because its electronic structure is influenced by nearby Ru[4]. This might be another possible reason accountable for the recorded oxidized Pt in Fig. 4.9. An alternative explanation is that the RuOx is able to withdraw electrons partially from the Pt nearby leading to relatively stronger white line intensities.

75

11560 11570 11580 11590 11600

Normalized absorbance (a.u.)

Energy (eV)

Group A Ru@Pt/XC72/CC pH 1 Group A Ru@Pt/XC72/CC pH 2.2 Group A Ru@Pt/XC72/CC pH 8

Group B Ru@Pt/XC72/CC pH 1 Group B Ru@Pt/XC72/CC pH 2.2 Group B Ru@Pt/XC72/CC pH 8

Figure 4.9. The Pt LIII-edge XANES spectra of Ru@Pt/XC72/CC from group A and group B.

The Fourier-transformed EXAFS spectra for H2PtCl6, group A, and group B are exhibited in Fig. 4.10. As shown, the peaks at 1.7 Å and 2.0 Å (without phase correction) were associated with the Pt-O bond and Pt-Cl bond at the first shell coordination, respectively. On the other hand, the peaks at 2.1 Å and 2.7 Å (without phase correction) corresponded to the Pt-Ru bond and Pt-Pt bond at the first shell coordination, respectively. The EXAFS fitting results are summarized in Table 4.4.

In H2PtCl6 solution, the coordination number of Pt-O was 1.31, 1.65, and 3.58 for pH 1, pH 2.2, and pH 8 bath, respectively. In addition, the EXAFS results determined that the coordination number of Pt-Cl was 4.69, 4.35, and 2.42 for pH 1, pH 2.2, and pH 8 bath, respectively. Previously, Spieker et al., in their careful study of dilute H2PtCl6 acid with various pH values, observed that the chloride ion ligands associated with the Pt complexes can be exchanged by hydroxide ligand (OH) or aquo ligand (OH2) due to the hydrolysis reaction.[114] Similar phenomena were observed in this work in which the EXAFS fitting for samples of different pHs indicated a significant ligand changeup between the Pt-Cl and Pt-O coordination number. In other words, the number of chloride ligands on the Pt complexes was reduced due to the hydrolysis reaction when the pH value was increased. The variation of Pt ligand species was believed to be responsible for the notable difference toward the behavior of displacement reaction.

76

0 1 2 3 4 5 6

FT magnitude (a.u.)

R (Angstrom) H2PtCl

6 pH 1 H2PtCl

6 pH 2.2 H2PtCl

6 pH 8

Group A Ru@Pt/XC72/CC pH 1 Group A Ru@Pt/XC72/CC pH 2.2 Group A Ru@Pt/XC72/CC pH 8

Group B Ru@Pt/XC72/CC pH 1 Group B Ru@Pt/XC72/CC pH 2.2 Group B Ru@Pt/XC72/CC pH 8

Figure 4.10. The Pt LIII-edge Fourier-transformed EXAFS spectra from Figs. 4.8 and 4.9.

For group A, the sum of coordination number for Pt-O and Pt-Cl was decreased after the deposition/attachment to the Ru. This indicated that some of the ligands were removed from the Pt complexes when the Pt cations were anchored to Ru. Nevertheless, a direct Pt-Ru bond was not established which was consistent with what we found in Table 4.4. Therefore, the bridged oxygen structure was believed to be the linkage between the Pt and Ru. For samples in group B, the coordination environment was changed dramatically with the disappearance of Pt-O and Pt-Cl.

Instead, we witnessed the formation of Pt-Ru and Pt-Pt. The coordination number for Pt-Ru was 5.75, 4.23, and 3.21 for pH 1, pH 2.2, and pH 8 bath. These results suggested that at low pH value, the Pt atoms were embedded in the Ru core occupying lattice sites vacated by dissolving Ru atoms.

At a high pH value, the Pt cations were merely adsorbed physically on the Ru surface which led to a decrease in the Pt-Ru coordination number after the hydrogen reduction treatment. On the other hand, the coordination number for Pt-Pt was 1.71, 2.43, and 4.24 for pH 1, pH 2.2, and pH 8 bath,

77

respectively. At a low pH value, the Pt atoms were assumed to mix with Ru forming a quasi-alloying state that resulted in a small Pt-Pt coordination number. On the contrary, the Pt clusters on the Ru surface were expected to exist for the pH 8 bath. These results were consistent with XRD data in Fig. 4.3, leading to a larger Pt-Pt coordination number.

78

Table 4.4. EXAFS fitting parameters at the Pt L_III-edge for Ru/XC72/CC and Ru@Pt/XC72/CC under various conditions.

Path

79

We have also examined the oxidation state of Ru ions after galvanic displacement reaction by XAS. During the displacement reaction, the metallic Ru atoms were oxidized by the PtCl62-

ions, and became Ru ions. The carbon-supported Ru nanoparticles, with additional treatment of hydrogen reduction at 100 ˚C for the removal of surface oxide, were immersed in the aqueous 80 mM H₂PtCl₆ solution at 30 ˚C for 24 hours, followed by filtering of the suspension. The electrolyte contained dissolved Ru ions and excess H2PtCl6 were studied by XAS measurements. As shown in Fig. 4.11, by comparing the Ru K-edge XANES spectra of Ru ions, with those of Ru metal, Ru/C, RuCl3(aq) and RuO2(s), it is noted that the inflection point of absorption curve for the Ru ions was very close to that of RuCl3(aq) but not of RuO2(s). This suggested that the valence of Ru ions is smaller than +4 . Also it is known that Ru²⁺ ions are not able to exist in a stable state in aqueous solution. We determine that the valence of Ru ions after displacement reaction is +3. And therefore the PtRu galvanic displacement reaction for Ru nanoparticles in aqueous H2PtCl6 solutions can be expressed as

Figure 4.11 The Ru K-edge XANES spectra of Ru ions after displacement reaction, along with Ru metal, Ru/C, RuCl3(aq) and RuO2(s) serving as references.

80

By summarizing the analytical results from XRD, ICP-MS, TEM, EDX, and XAS spectra, a mechanism for the Pt displacement reaction on the Ru nanoparticles is provided in the schematic diagrams shown in Fig. 4.12. Fig. 4.12(a) depicts the scenario for pH 1 and pH 2 in which the Pt cations were partially reduced accompanied by the oxidative dissolution of Ru on the surface. The dissolution of Ru was likely resulted from Pt reduction. Meanwhile, the formation of Ru oxide layer on the surface took place slowly during the immersion into aqueous baths which eventually terminated the displacement reaction as the surface oxide layer prevented the underneath Ru from further dissolving and consequently inhibiting the reduction of Pt ions. Fig. 4.12(b) illustrates the mechanism of PtRu displacement reaction in pH 8 scenario. The Pt cations were physically adsorbed on the Ru surface mostly, and after the hydrogen reduction treatment, the Pt cations were reduced and agglomerated as clusters on the Ru surface. In short, the pH values of H2PtCl6 acid played a critical role for the extent of displacement reaction between the Pt and Ru and its resulting alloy state. Furthermore, two factors should also be considered. They were the various ligands around the Pt complexes and extent of Ru dissolution in acidic/basic aqueous solvents. The deposition for Pt on the Ru was formed through the bridged oxygen instead of direct Pt-Ru bonding.

81

Figure 4.12. Schematic diagrams for PtRu displacement reaction occurring at (a) low pH and (b) high pH conditions.

82