Duty Cycle for pulse electroplating

We realized that the duty cycle, defined as Ton /( Ton + Toff), would be an appropriate indicator to solicit further information on the composition variations over T_on and T_off. As defined, a higher duty cycle is closer to the dc deposition, while a lower one represents a longer stop time between pulses. Fig. 3.5 exhibits the ratio of Pt in the PtRu nanoparticles as a function of duty cycle from the data listed in Tables 3.1 and Tables 3.3. Remarkably, the Pt ratio increased rapidly as the duty cycle was reduced. This infers that a longer T_off renders a pronounced effect of Ru loss in the PtRu nanoparticles. At higher duty cycles, we reached a plateau, with the resulting composition approaching Pt₅₂Ru₄₈. These behaviors were expected, because the galvanostatic deposition from nitroso precursors of Pt and from these samples were rather consistent, indicating that the pc deposition was insensitive to the current densities under study once the T_on and T_off were determined.

In addition, the catalyst loadings were in the range of 49.1–67.6 µg/cm², values that are in line with earlier results. Table 3.6 also lists the electrochemical characteristics for these samples.

0.0 0.1 0.2 0.3 0.4 0.5

Figure 3.5. The effect of duty cycle on the Pt atomic ratio for the PtRu nanoparticles. Data are from Tables 3.1(■), 3.3 (○), and 3.5 (×).

36

Table 3.5. Results from materials characterizations on PtRu nanoparticles with fixed values of T_on (50 ms), T_off (400 ms), and total coulombic charge (8.0 C/cm²).

a) potential at peak current density in anodic scan b) peak mass activity in anodic scan

c) potential at peak current density in cathodic scan d) peak mass activity in cathodic scan

e) ECSA from hydrogen desorption data

To account for the apparent enrichment in the Pt at shorter duty cycles, we believe a displacement reaction was taking place where the Ru in the PtRu nanoparticles was preferentially dissolved while the Pt cations in the solution were reduced during T_off. This is attributed to the observed rise in the Pt ratio when the Toff was prolonged. The nature of displacement reaction could be supported by detailed XRD analysis on the PtRu nanoparticles to determine their lattice parameter. Fig. 3.6 exhibits the XRD patterns for the PtRu nanoparticles with Toff of 100, 400, and 600 ms, respectively. The complete parameters for their synthetic conditions are shown in Table 3.1.

37

Due to the interferences from carbon cloths, the XRD profiles were expectedly rough. Nevertheless, the signals from Pt(111), Pt(200), Pt(220), and carbon were identified. Because we did not observe the characteristic Ru peaks, we rationalize that alloying of Ru into the Pt lattice was successfully achieved. Moreover, because the atomic radius of Ru is smaller than that of Pt, substitution of Ru in the Pt unit cell results in the reduction of lattice parameters. This engenders a slight shift of diffraction peaks to larger angles. To minimize undesirable noises, we selected the signal from Pt(111), as shown in Fig. 3.7, to estimate the lattice parameter with the equation below

^√

(3.1)

(3.2)

where a is the lattice parameter, max is the peak position for Pt(111), and λkα is the wavelength of X-ray. The value for a could further be used to deduce the amount of alloyed Ru, as suggested by Antolini et al., in the following relation[92] where a0 is the standard lattice parameter from a bulk Pt (JCPDS: 870646) and xRu is the atomic Ru ratio in the alloyed state. The results from these estimations are presented in Table 3.7. As listed, we noticed that when the T_off was prolonged, the ratio for the alloyed Ru also decreased. This behavior is understandable, because when the displacement reaction was occurring (at larger T_off), the Ru on the surface were dissolved preferentially, leaving those buried inside (the alloyed Ru) intact.An alternative explanation for the Pt enrichment at shorter duty cycles is a cementation process. As proposed by Spendelow and Wieckowski, spontaneous deposition of Pt from electrolyte could proceed via the oxidation of Ru substrate.[93] In such a reaction, the Ru is not dissolved but exists in an oxidized state on the surface with metallic Pt deposited from the electrolyte. This model poses a drastic contrast with that of displacement reaction, where the Ru is dissolved. Therefore, we could identify the exact mechanism in our case simply by determining the oxidative states of Ru at different duty cycles.

38

Fig. 3.8(a) presents the XPS signals from Ru (3p_3/2) on samples of T_off in 100, 400, and 600 ms, respectively. The complete synthetic parameters can be obtained in Table 3.1. The selection of Ru (3p_3/2) instead of Ru (3d_5/2) was to avoid possible interference from that of C (1 s). As shown in the diagram, all the samples revealed rather broad profiles. The signals peaks were located at 462.2, 462.0, and 461.9 eV for samples of 100, 400, and 600 ms, respectively. Earlier, the peak for metallic Ru was documented at 461.1 eV, which was close to what we measured.[94] A slight deviation is expected, because there is a minor shift of Ru signals (less than 0.1 eV) in the PtRu from that of metallic Ru.[95] Nevertheless, our broad signals seem to suggest that Ru existed in multiple oxidative states. Furthermore, the XPS signal intensity decreased considerably with longer Toff, indicating that the amount of Ru, regardless of its oxidative states, decreases when the duty is reduced. This is consistent with our argument of steady Ru loss during T_off.

39

30 40 50 60 70 80

(c) (b) Pt(200)

Int ens it y ( a. u.)

2  degree)

Carbon Pt(111)

Pt(220) (a)

Figure 3.6. The XRD patterns for the PtRu nanoparticles with fixed values of T_on (50 ms), J_a (50 mA/cm²), and coulombic charge (8.0 C/cm²), as well as Toff of (a) 100, (b) 400, and (c) 600 ms.

38 39 40 41 42

(c) (b) (a)

Int ens it y ( a.u.)

2  degree)

Figure. 3.7. XRD patterns from Fig. 3.6 with an enlarged range between 38 and 42° for lattice parameter determination.

40

Table 3.7. Lattice parameter and alloyed Ru for the PtRu nanoparticles with fixed values of Ton (50 ms), Ja (50 mA/cm²), and coulombic charge (8.0 C/cm²).

sample Pt (111) peak (deg) Lattice parameter a (Å ) X_Ru (%) Pt/Ru atomic ratio^b

TF100 40.21 3.879 33.9 ± 2 52.8/47.2

TF400 40.16 3.884 30.2 ± 2 64.1/35.9

TF600 39.97 3.902 15.9 ± 2 83.4/16.6

Pt^a 39.79 3.920

a) from JCPDS 870646

b) from ICP-Mass (as shown in Table 3.1)

In an acidic environment, we suspect the possible oxidation states of Ru are Ru⁰, RuO2, and RuO₂·nH₂O. To determine their relative ratios quantitatively, we carried out curve fitting using software (Thermo Avantage 3.20). The corresponding peaks for RuO2, and RuO2·nH2O were 462.2 and 463.8 eV, respectively.[93, 96] As shown in Fig. 3.8(b), we acquired a reasonable match for the observed XPS responses. The detailed fitting results are listed in Table 3.8. When the Toff was prolonged, we observed that the ratio for the metallic Ru increased, while the ratio for the RuO2·nH2O decreased steadily. Interestingly, the ratio for the RuO2 remained unchanged. Because the metallic Ru existed in the alloyed state buried inside the PtRu nanoparticles, the loss of Ru in the displacement reaction was likely from the RuO2·nH2O on the surface.

41

Figure 3.8. (A) XPS signals of Ru (3p_3/2) from PtRu nanoparticles with fixed values of T_on (50 ms), J_a (50 mA/cm²), and coulombic charge (8.0 C/cm²), as well as T_off in (a) 100, (b) 400, and (c) 600 ms. (B) The results of curve fitting using Ru⁰, RuO₂, and RuO₂·nH₂O.

42

Table 3.8. Results from XPS and curve fitting of PtRu nanoparticles with fixed values of T_on (50 ms), J_a (50 mA/cm²), and total coulombic charge (8.0 C/cm²).

T_off (ms)

Binding energy^a (eV)

Reference binding energy^b (eV)

Suggested species

Ratio (%)

100 461.1 461.1 Ru 36.5

462.6 462.2 RuO₂ 33.9

463.9 463.8 RuO₂·nH₂O 29.6

400 461.1 461.1 Ru 41.1

462.4 462.2 RuO₂ 34.5

463.8 463.8 RuO₂·nH₂O 24.4

600 461.2 461.1 Ru 43.8

462.4 462.2 RuO₂ 36.4

463.8 463.8 RuO₂·nH₂O 19.8

a) From XPS curve fitting b) From Ref.[93, 94, 96]

43