4.1 Introduction
PtRu is an anode catalyst for methanol oxidation reaction (MOR). It is known that Ru substantially promotes the CO-resistance of Pt during MOR process. To date, both bifunctional mechanism and ligand effect are proposed to explain the behavior that Ru prevents Pt from CO-poisoning [27-29]. Since both Pt and Ru are precious metals, it becomes an issue of reducing their loading while maintaining necessary catalytic activity. To achieve this objective, a core-shell structure has been explored recently. Of many routes to prepare core-shell structure, the displacement reaction is a novel preparation technique that allows the fabrication of noble metal for electrochemical active materials with ultra low content of Ru or Pt [30-32].
Earlier, fundamental studies related to the displacement reaction conducted by Adzic et al.
suggest that the driving force is the potential difference between PtCl6
reduction and Ru0 oxidation [33-35]. However, the detailed mechanism for Pt deposition on Ru still requires further clarification.
Possibilities of chemical reactions occurring on the Ru surface consist of either the oxidation of Ru to Ru(OH), or a higher oxidation state of Ru. Alternatively, the possible dissolution of Ru, which also contributes to the displacement reaction, should be taken into consideration. Manandhar et al.
emphasized that the spontaneously-deposited Pt on Ru was Pt cations in intermediate oxidation states by X-ray photoelectron spectroscopy (XPS) [36-37]. Pt chloride, oxide, and hydroxide species were observed on Ru substrates depending on various ligand environments of Pt. The oxidation states for the as-deposited Pt were between the state of solvated Pt precursors and that of metallic Pt.
Previous research by Spieker et al. demonstrated that the hexachloroplatinic acid is a strong acid with rapid hydrolysis confirmed by extended X-ray absorption fine structure (EXAFS). A thorough study provides the elucidation regarding the Pt coordination environment for its precursor,
20
hexachloroplatinic acid, in the form of aqueous solution under various pH values and ion concentrations. The bonding of Pt-oxygen and Pt-chloride play crucial role throughout the reaction of galvanic displacement reaction due to the transition to Pt-Ru bonding. However, a direct and further observation covering from liquid-phased ionic Pt to solid-phased metallic Pt should make a contribution for building up the detailed and complete mechanism. In addition, the acidic aqueous environment in which the displacement reaction occurs is feasible to dissolve Ru. The dissolved Ru and displaced Ru are substantially mixed together and both results are received during the analysis [38].
From the discussion above, a sophisticated analytical method should be developed to provide a further investigation. The Pt deposited with displacement reaction on Ru/C and corresponding liquid phase Pt precursor ought to be taken as correlated analysis to establish the mechanism for PtRu displacement reaction. In this work, X-ray absorption spectroscopy was employed to provide a precise investigation to determine the oxidation states and coordination environment including solid and liquid samples. Inductively coupled plasma mass spectroscopy (ICP-MS) is a quantitative instrument that reveals the exact amount of Pt and Ru atom participating in the displacement process. X-ray diffraction confirms the slight lattice expansion for Pt-deposited Ru compared with pure Ru nanoparticles. Finally, we attempt to establish the detailed reaction involved in the displacement process.
4.2 Results and discussion
Figure 6, 7 and 8 demonstrate the XRD patterns for samples from group A, B and C, respectively. The diffraction peaks at the top and bottom of Figure 6(A), 7(A) and 8(A) exhibit patterns for Ru/C, in which the hcp Ru lattice was confurmed, albeit with reduced intensity caused by low loading and interference from carbon support. The diffraction peak at 38.4 degree of Ru (100) phase is selectively shown in Figure 6(B), 7(B) and 8(B) for group A, B and C, respectively.
21
In Figure 6(B), the diffraction peaks of group A reveal relatively broad compared with the rest groups. In Figure 7(B), peaks are sharpened due to hydrogen reduction and slightly shifted toward lower angles with the sequence of pH 2.2 and pH 1. This indicates that the expansion for the lattice of Ru was caused by the displacement of Pt atoms and their incorporation into Ru lattice. However, the peak position of pH 8 remained the same as pure Ru/C, suggesting that the adsorption of Pt atoms was taking place only on the surface of Ru nanoparticles. In Figure 8(B), diffraction peak at 39.5 degrees for pH 8 was observed that probably was attributed to Pt (111). The presence of Pt diffraction peak suggested that the Pt atoms were present after hydrogen reduction. However, the absence of Pt diffraction peak for pH 1 inferred that the feasibility of forming Pt clusters was reduced due to the incorporation of Pt atoms into Ru lattice, leaving a sub-monolayer Pt on the Ru surface.
30 40 50 60 70 80 90
Carbon
A-group pH8.0 A-group pH2.2 A-group pH1.0
In te n si ty / a .u .
2
/ degree
Ru/C
(A)
22
35 36 37 38 39 40 41 42
2
/ degree
2
/ degree
In te n si ty / a .u .
A-group pH1.0
A-group pH2.2
A-group pH8.0
Figure 6(A). The XRD patterns for carbon supported Ru (Ru/C), group A pH 1, pH 2.2, pH 8 and pure carbon.
Figure 6(B). The XRD patterns at 38 degrees for group A pH 1, pH 2.2, and pH 8.
30 40 50 60 70 80 90
In te n si ty / a .u
2
/ degree
Ru/C
Carbon
B-group pH8.0 B-group pH2.2 B-group pH1.0
(B)
(A)
(B)
23
36 38 40 42
2
/ degree
In te n si ty / a .u .
B-group pH1.0
B-group pH2.2
B-group pH8.0
Figure 7(A). The XRD patterns for carbon supported Ru (Ru/C), group B pH 1, pH 2.2, pH 8 and pure carbon.
Figure 7(B). The XRD patterns at 38 degrees for group B pH 1, pH 2.2, and pH 8.
30 40 50 60 70 80 90
In te n si ty / a .u .
2
/ degree
Ru/C
C-group pH1.0 C-group pH2.2 C-group pH8.0 Carbon
24
Figure 8(A). The XRD patterns for carbon supported Ru (Ru/C), group C pH 1, pH 2.2, pH 8 and pure carbon.
Figure 8(B). The XRD patterns at 38 degrees for group C pH 1, pH 2.2, and pH 8.
Table 1 presents the ICP-MS results for the as-prepared samples of group A, the corresponding immersion baths of group A, and the immersion baths for reference group. The results indicated Pt loadings were 1.26, 1.17 and 0.62 mmole for pH 1, pH 2.2 and pH 8, respectively. Lower pH value allowed more Pt loading deposited on Ru. On the other hand, the Ru loadings were 2.02, 2.94, and 4.86 mmole for pH 1, pH 2.2 and pH 8, respectively. Apparently, Ru dissolution caused by Pt displacement and acid corrosion was observed. For the corresponding hexachloroplatinic acid immersion baths, the amount of Pt remained around 20 mmole. Generally, more Ru content was detected in the immersion baths of hexachloroplatinic acids rather than that of pure perchloric acid, DI water and potassium hydroxide. These behaviors suggested that a greater amount of Ru atoms was replaced by Pt ions. The dissolution of Ru by acidic or basic solvents still existed but it’s
(A)
(B)
Pt (111)
25 amount was slightly varied.
Table 1. Results from material characterizations on group A, the immersion baths of group A and the immersion baths of reference group by ICP-MS.
Pt (μmole) Ru (μmole)
Nanoparticles pH 1 1.26 2.02
pH 2.2 1.17 2.94
pH 8 0.62 4.86
Solutions H2PtCl6 pH 1 20.01 1.88
H2PtCl6 pH 2.2 21.61 1.37 H2PtCl6 pH 8 22.57 0.05
Solutions HClO4 N/A 0.31
DI water N/A 0.01
KOH N/A 0.03
TEM images for carbon supported Ru nanoparticles are exhibited in Figure 9. Ru nanoparticles were well-dispersed with notable size uniformity on the carbon supports. Figure 10 to 12 display the TEM images of group A with various immersion baths of pH 1, pH 2.2 and pH 8. As shown, the PtRu/C particles were uniformly dispersed and their sizes were approximately 2~3 nm. The slight variations of sizes between each groups suggested that the Ru nanoparticles were rather stable in various environments against corrosive dissolution. In addition, possible particle size variation
26
influenced by acidic immersion or displaced by Pt atoms was negligible. TEM/EDX confirmed the presence of Pt species on the surface of Ru nanoparticles with quantitative results in table 2. The immersion bath of pH 1 possessed a larger amount of deposited Pt relatively compared with that of other pH values. This result suggested that the displacement reaction showed more tendencies to allow the deposition of Pt on Ru nanoparticles in acidic environment than in basic one. The detailed mechanism is explained in the following parts of the discussion.
Table 2 Results from material characterizations on Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 (group A)by EDX.
Pt (at%) Ru (at%)
pH 1 11.08 88.92
pH 2.2 6.65 93.35
pH 8 3.38 96.62
27
Figure 9. The TEM images for carbon supported Ru.
28
Figure 10. The TEM images for group A, Ru/C immersed with pH 1 H2PtCl6.
29
Figure 11. The TEM images for group B, Ru/C immersed with pH 2.2 H2PtCl6.
30
Figure 12. The TEM images for group C, Ru/C immersed with pH 8 H2PtCl6.
The Ru K-edge XANES spectra for reference group are shown in Figure 13. After the immersion of perchloric acid (pH 1), de-ionized water (pH7.0) and potassium hydroxide (pH 8), Ru nanoparticles were oxidized based on the energy shift of the edge jump. The oxidation states of as-immersed Ru nanoparticles under various baths were similar. In addition, those oxidation states were between 0 to +4 according to the energy shifts which were located between that of RuO2 and metallic Ru powder (reference). The slight elevation for the oxidation states of Ru nanoparticles during the immersion was reasonably attributed to both contribution of water adsorption and oxide formation at the surface of Ru nanoparticles occurring together [18]. On the other hand, the oxide formation at the surface of Ru nanoparticles provided moderate protection against further
31
dissolution. After extended immersion, the Ru nanoparticles maintained their stability and sizes.
Figure 14 exhibits the Ru K-edge XANES spectra of group A. The oxidation states of Ru nanoparticles were similar for those immersed by various pH values of hexachloroplatinic acids.
Interestingly, the oxidation states of group A Ru nanoparticles were approximately identical to that of reference group. These results confirmed that Ru maintained identical oxidation state during the displacement reaction by Pt atoms. Therefore, we assumed that the Ru species with higher oxidation states somehow left the surface in the form of dissolved complexes into the immersion baths.
Similarly, the reduction process of Pt ions from the immersion baths onto the Ru nanoparticles could be assumed to include the formation of oxidized Ru and its subsequent dissolution. The Ru K-edge XANES for group B is exhibited in Figure 15. Similar results were obtained as group A even with a hydrogen reduction on the Ru nanoparticles. However, due to the reduced surface and the removal of surface oxides of Ru nanoparticles, a relatively stronger reduction driving force was provided for group B than group A. Therefore, the Pt species deposited on group B revealed lower oxidation states compared with group A. The Ru K-edge XANES for group C is shown in Figure 16.
The hydrogen reduction treatment was applied after the immersion of Ru nanoparticles into hexachloroplatinic acids with various pH values. The oxidation states were not completely reduced to 0 like metallic Ru. This inevitable oxidation of Ru might be attributed to the oxygen-containing groups on the surface of functionalized carbon supports [19].
32
Figure 13. Ru K-edge XANES spectra of Ru metallic powder, RuO2, Ru/C immersed with HClO4, DI water and KOH.
Figure 14. Ru K-edge XANES spectra of Ru metallic powder, RuO2, Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 (group A).
33
Figure 15. Ru K-edge XANES spectra of Ru metallic powder, RuO2, hydrogen reduced Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 (group B).
Figure 16. Ru K-edge XANES spectra of Ru metallic powder, RuO2, Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 and followed by hydrogen reduction (group C).
34
The Ru K-edge Fourier transform EXAFS spectra for reference group, group A, group B, and group C are shown in Figure 17 to 20, respectively. The peaks at 1.6 Å and 2.3 Å (without phase correction) corresponded to the Ru-O bond and Ru-Ru bond at the first shell coordination. The EXAFS fitting results were summarized in table 3. For reference group, the coordination numbers of Ru-O of all immersion baths are around 2. The bonding between Ru and O was attributed to the formation of oxide layer on Ru nanoparticles caused by the immersion. Unfortunately, the fitting results of Ru which were immersed into H2PtCl6 baths, group A, and group B, did not reveal any Ru-Pt coordination. As a result, the deposition of Pt on Ru was inconclusive from these EXAFS fittings. Nevertheless, once the hydrogen reduction was applied after the immersion as group C, Ru-Pt coordination numbers of 1.33, 1.08 and 0.82 were obtained for pH 1, pH 2.2 and pH 8 hexachloroplatinic acid immersion baths. EXAFS fitting results for group C also confirmed the presence of Ru-Pt bonding. This suggested that during the displacement reaction, Pt atoms were not deposited to Ru directly. Based on previous research by Hu et al. [46], the Ru might existed as bridged-oxygen oxide. The bridged-oxygen was expected to connecting Pt to Ru. This may explain why the Ru-O was the predominant bonding specie after the displacement reaction of Ru by Pt.
This is because the Ru-O-Pt was formed instead of Ru-Pt on the surface of Ru nanoparticles. As expected, the Ru-Pt coordination was established as long as bridged-oxygen was removed.
35
Figure 17. Ru K-edge EXAFS spectra of Ru/C immersed with HClO4, DI water and KOH.
0 1 2 3 4 5
36
Figure 19. Ru K-edge EXAFS spectra of hydrogen reduced Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 (group B). followed by hydrogen reduction (group C).
37 Table 3
The fitting results from the analysis of Ru K-edge EXAFS spectra.
Path
38
Figure 21 demonstrates the Pt L3-edge XANES spectra of hexachloroplatinic acid adjusted to various pH values. The large peak at the absorption edge was attributed to 2p3/2 to 5d electronic transition of Pt which is also known as white line [10, 47]. The white line intensity is able to provide information on the oxidation state of Pt due to its correspondence to d-band vacancy. A larger of the white line intensity, refers less d-band is occupied. That is, a high intensity white line is obtained for the Pt with a high oxidation state. Pt ions in hexachloroplatinic acids with various pH values revealed high white lines intensity similar due to the +4 oxidation number in Figure 16. For the white line of metallic Pt foil, its low intensity was designated as the 0 oxidation step.
For Pt atoms that were deposited onto Ru nanoparticles via the displacement reaction, their oxidation states can be observed in Figure 22. The Pt atoms that were deposited onto Ru nanoparticles with the immersion bath of hexachloroplatinic acid adjusted to pH 1 were partially reduced. For pH 2.2, although the oxidation state was slightly higher than that of pH 1, the partial reduction was observed. However, the Pt was nearly reduced for pH 8 based on its white line as high as that of Pt ions in hexachloroplatinic acid. A reasonable assumption could be adopted as physical adsorption for Pt ions onto Ru nanoparticles for the case of pH 8. In Figure 23, the whit line intensities were generally lower than that in Figure 22. This result implied that for hydrogen
39
reduced Ru nanoparticle before immersion, group B, the Pt atoms were reduced even more compare with group A. Since the deposition through displacement reaction was expected to rely greatly on the surface, the hydrogen pretreatment that eliminated the surface oxides of Ru nanoparticles demonstrated clear evidence over reduced oxidation states of Pt atoms. In Figure 24, the intensities of white lines for group C were nearly as small as Pt foil because of the final hydrogen reduction.
Nevertheless, the oxidation states were slightly higher for pH 1 and pH 2.2 than that of pH 8. This result indicated that sub-monolayer Pt for pH 1 and pH 2.2 facilitated each Pt atom to be a surface atom. For pH 8, the formation of Pt cluster might lead to a lower oxidation state due to the existence of protected inner Pt atoms.
40
Figure 23. Pt L3-edge XANES spectra of Pt foil, hydrogen reduced Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 (group B).
41 H2PtCl6 and followed by hydrogen reduction (group C).
The Pt L3-edge Fourier transform EXAFS spectra for hexachloroplatinic acids, group A, group B and group C are shown in Figure 25 to 28, respectively. In Figure 25 to 27, the peaks at 1.7 Å and 2.0 Å (without phase correction) correspond to the Pt-O bond and Pt-Cl bond at the first shell coordination, respectively. In Figure 28, the peaks at 2.1 Å and 2.7 Å (without phase correction) correspond to the Pt-Ru bond and Pt-Pt bond at the first shell coordination, respectively. The EXAFS fitting results are summarized in table 4. Spieker et al. proposed that chloride ion ligands on the Pt complexes can be exchanged by hydroxide ligands or aqueous ligands due to the hydrolysis reaction in a thorough research of dilute hexachloroplatinic acid with various pH values [12].
Similar phenomena were observed in this work in which EXAFS fitting indicated a significant ligand change influenced by pH values based on Pt-Cl and Pt-O coordination number. The coordination numbers of Pt-O were 1.31, 1.65 and 3.58 for pH 1, pH 2.2 and pH 8 hexachloroplatinic acids, respectively. These results were coherent with Spieker et al in which ligand exchange by hydroxide or aqueous ligands with the increase of pH values. On the contrary,
42
EXAFS results determined that the coordination numbers of Pt-Cl were 4.69, 4.35 and 2.42 for pH 1, pH 2.2 and pH 8 hexachloroplatinic acids, respectively. In other words, the number of chloride ligands on the Pt complexes was reduced due to the hydrolysis reaction when pH value was increased. The variation of Pt ligand species should be responsible for causing notable influence toward the behavior of displacement reaction. For deposited Pt as group A, the sum of coordination number for Pt-O and Pt-Cl was decreased as presented in table 4, indicating the ligands were detached from the Pt complexes when Pt ions were reduced by Ru. Furthermore, under various pH values, the portion of Pt-O and Pt-Cl bonding for Pt on Ru was consistent to that in hexachloroplatinic acids. However, Pt-Ru bonding was not formed due to the bridged-oxygen structure. Similar outcome was obtained for group B. For the EXAFS fitting result in Figure 28, the coordination environment was changed dramatically due to the final hydrogen reduction which removed the bridged-oxygen. Table 4 also lists the coordination numbers of Pt-Ru which were 5.75, 4.23 and 3.21 for pH 1, pH 2.2 and pH 8 of group C, respectively. The results suggested that at low pH values, Pt ions were feasible of being reduced by Ru and replacing Ru atoms. It was obvious that the incorporation for the Pt into Ru nanoparticles occurred and contributed the Pt-Ru bonding as high as 5.75 and 4.23. At a high pH value, the Pt ions were merely physically adsorbed on the surface of Ru which its coordination number was further decreased. The coordination numbers of Pt-Pt are 1.71, 2.43 and 4.24 for pH 1, pH 2.2 and pH 8 of group C, respectively. At low pH values, the Pt atoms were assumed to form a sub-monolayer based on the relatively low Pt-Pt coordination number. On the contrary, the Pt clusters were expected to be formed for pH 8 consistent to the XRD results, giving a higher Pt-Pt coordination number.
43
44
Figure 27. Pt L3-edge EXAFS spectra of hydrogen reduced Ru/C immersed with pH 1, pH 2.2 and pH 8 H2PtCl6 (group B). followed by hydrogen reduction (group C).
45 Table 4
The fitting results from the analysis of Pt L3-edge EXAFS spectra.
Path
46 C pH 2.2
Pt-Ru 4.23 2.71 7.69 5.60
Pt-Pt 2.43 2.76 8.65 4.82
C pH 8
Pt-Ru 3.21 2.71 6.69 6.25
Pt-Pt 4.24 2.75 7.26 4.66
By summarizing the analytical results from XRD, ICP-MS, TEM, EDX and XAS spectra, the mechanism of Pt displacement reaction on Ru nanoparticles is described as follows. Pt ions were reduced by metallic Ru nanoparticles accompanied by the dissolution of Ru atoms at the surface.
The dissolution of Ru atoms could be resulted from being replaced by Pt. Meanwhile, the formation of Ru oxide layer on the surface of it nanoparticle took place during the immersion into aqueous
The dissolution of Ru atoms could be resulted from being replaced by Pt. Meanwhile, the formation of Ru oxide layer on the surface of it nanoparticle took place during the immersion into aqueous