Effects of Hydrophobic Species on the Oxygen Reduction Reaction on the High-Index Planes of Pt3Fe

Pt3Fe(111), Pt3Fe(775) = 7(111)–(111), and Pt3Fe(544) = 9(111)–(100) electrodes show the highest activity in the low-index planes, n(111)–(111), and n(111)–(100) series of Pt3Fe, respectively. The surfaces of these electrodes were modified with hydrophobic species such as THA+, melamine, and ionic liquid ([MTBD][beti]), and the effects on the oxygen reduction reaction (ORR) were studied. All the hydrophobic species improved the ORR activity on all the electrodes examined. The ORR activity of Pt3Fe(111) in 0.1 M HClO4 containing 0.1 μM melamine was 2.1 times higher than that of Pt3Fe(111) without melamine, giving 39 times higher activity than that of bare Pt(111). The durability was improved on all the electrodes examined in melamine-containing solution.


Introduction
Polymer electrolyte fuel cells (PEFCs) have attracted attention as an energy system with low environmental pollution because they emit only water. However, due to its high overpotential of the oxygen reduction reaction (ORR), a large amount of Pt is used as an electrocatalyst of PEFCs. We must reduce the amount of Pt loading to promote the use of PEFCs. Therefore, it is necessary to develop electrocatalysts with higher activity for the ORR and durability.
One of the methods of developing electrocatalysts with higher ORR activity is the control of the surface structures of the electrodes [1][2][3][4][5][6][7][8]. Marković et al. found that the ORR activity of the low-index planes of Pt increases in the order Pt(100) < Pt(111) < Pt(110) in 0.1 M HClO 4 [1]. Systematic study using the high-index planes can determine the structure of the active site of the ORR at the atomic level. Feliu et al. investigated the ORR on the high-index planes of Pt in acidic solutions and found that the ORR activity increases with increasing step atomic density [2,3]. Hoshi et al. found that the (111) terrace edge enhances the ORR activity of the Pt electrode [5]. Theoretical calculations predicted that the enhancement of the ORR is attributed to the change in the structure of adsorbed water at the (111) terrace edge, resulting in the suppression of Pt oxide formation [9].
Stamenkovic et al. found that the ORR activity of Pt 3 Ni(111) is 10 times higher than that of Pt(111) [8]. Wakisaka et al. investigated the effect of Co ratio on the ORR activity of Pt 100-x Co x and found that Pt 100-x Co x (111) with x = 27 has 27 times higher ORR activity than pure Pt(111) [15,16]. Surface X-ray scattering (SXS) found that Co was enriched up to 98% in the second layer of Pt 75 Co 25 (111). The positively charged Co in the second layer promotes strong electron transfer to the Pt skin, resulting in the high ORR activity [17].
The ORR of PtFe alloys was studied by several groups expecting higher activity. Wang et al. found that multimetallic Au/FePt 3 nanoparticles, in which FePt thin film was deposited on Au(111) surface, had three times higher ORR activity than Pt/C. After the 6000 potential cycles, Au/FePt 3 became 7 times more active than Pt/C and gave higher durability [18]. It is also reported that the ORR activity of interconnected surface-vacancy-rich PtFe nanowires is 10 times higher than that of commercial Pt/C. The activity decreased only 8.1% after 10000 potential cycles, indicating extremely higher durability compared to Pt/C, of which activity was halved after 5000 cycles [19].
We studied the ORR activity on single-crystal electrodes of Pt 3 Fe. The ORR activity increases as Pt 3 Fe(100) < Pt 3 Fe(110) < Pt 3 Fe(111). Pt 3 Fe(111) has 20 times higher activity than Pt(111) [20]. In the n(111)-(111) series of Pt 3 Fe, the ORR activity reaches the upper limit at 4 ≤ n. In the n(111)-(100) series, the ORR activity plotted against step atom density d S gives volcano shape with the highest activity at Pt 3 Fe(544) = 9(111)-(100). The ORR activities of the surfaces with (100) terrace are much lower than those with (111) terrace, giving no structural effects [21], as are the cases of the high-index planes of Pt 3 Co and Pt 3 Ni [6,7].
Pt-OH is a blocking species of the ORR on Pt electrodes [22][23][24]. Pt-OH is stabilized by hydrogen bonding with water molecules. Density functional theory (DFT) calculation predicted the adsorption forms of Pt-OH and Pt-O on Pt(111) and Pt(332) surfaces [9]. These Pt oxides were destabilized due to the change of water structure around the step. This prediction is consistent with the experimental ORR activity order: Pt(111) < Pt(332).
Hydrophobic species can also change water structure around electrode surfaces; the ORR activity can be improved by modifying the electrode surface with hydrophobic species [25][26][27][28][29][30][31]. Miyabayashi et al. found that the ORR activity and durability of Pt nanoparticles were enhanced by the modification with octylamine (OA) and alkyl amine containing pyrene ring (PA) [25]. Saikawa et al. studied the ORR activity on n(111)-(111) series of Pt modified with OA/PA = 9/1 and found that the activity increases on the surfaces with 7 ≤ n [26]. Infrared reflection absorption spectroscopy (IRAS) indicates that ice-like water with smaller cluster size increases the ORR activity on Pt(111) modified with OA/ PA [27]. Modification of Pt and PtPdCo nanoparticles with melamine enhances the ORR activity [28]. Melamine also increases the ORR activity of n(111)-(111) series of Pt where Pt(111) gives the highest increase ratio [29]. Tetran-hexylammonium cation (THA + ) enhances the ORR activity of Pt(111) by 8 times [30]. Modification with ionic liquids (ILs) such as [MTBD][beti] improves the ORR activity of Pt nanoparticles [31]. These results suggest that further improvement of ORR activity of Pt alloys is expected by the modification with these hydrophobic species.
In this study, we aim to improve the ORR activity and durability of Pt 3 Fe single-crystal electrodes by the modification with hydrophobic species such as THA + , melamine, and ionic liquid ([MTBD][beti]) ( Fig. 1).

Experimental
A Pt 3 Fe single-crystal bead was prepared according to the method reported previously [20,21]. A Pt single-crystal bead was prepared from 1 mmφ Pt wire (99.99%, Tanaka Kikinzoku Kogyo K.K.) as reported by Clavilier et al. [32]. To prevent oxidation of Fe, 0.5 mmφ Fe wire (99.99%, Johnson Matthey) was added to the Pt single-crystal bead in an ultrapure grade atmosphere of Ar (95%) + H 2 (5%) using an induction-heating furnace. The Pt 3 Fe single crystal was oriented with the back reflection Laue method and mechanically polished with a diamond slurry. The strain on the electrode surface caused by the mechanical polishing was removed by annealing in an Ar (95%) + H 2 (5%) atmosphere at approximately 1200 °C using an induction-heating furnace.
Electrolytic solutions 0.1 M HClO 4 were prepared from ultrapure water treated with Milli-Q Reference (Millipore) and ultrapure grade chemicals (Kanto Chemical).
Pt 3 Fe single-crystal electrodes were modified with THA + (Alfa Aesar), melamine (FUJIFILM Wako Pure Chemical) and [MTBD][beti] according to the following procedure. THA + modification was done by immersing the single-crystal electrode surface in a 10 −5 M THA + solution. The excess THA + was rinsed with ultrapure water. Melamine modification was carried out using the following two methods.

Melamine-Immersed Modification
The single-crystal surface of Pt 3 Fe was immersed in a 0.7 mM melamine aqueous solution. The excess melamine was removed with ultrapure water. Electrochemical measurements were done in 0.1 M HClO 4 without melamine.
[MTBD][beti] modification was performed according to the method reported previously [31]. Cyclic voltammograms (CVs) and linear sweep voltammograms (LSVs) were measured using an electrochemical analyzer (ALS 701CH) and a rotating disc electrode apparatus (RRDE-3A, BAS) with a hanging meniscus rotating disc electrode (RDE) configuration [33]. CV and LSV were measured in 0.1 M HClO 4 at a scanning rate of 0.01 V s −1 . CVs were measured in Ar-saturated solutions. LSVs were measured in the positive direction from 0.05 to 1.0 V (vs. RHE) at a rotation rate of 1600 rpm in O 2 -saturated solution. All the potentials were referred to a reversible hydrogen electrode (RHE).
ORR activity was evaluated using the specific activity (j k ) at 0.90 V at the first scan using the Koutecky-Levich equation (1): where j and j l are the total current density at 0.90 V and the limiting current density at 1600 rpm, respectively [34,35].
(1) 1∕j = 1∕jk + 1∕j 1 A geometric surface area of a single-crystal electrode equals electrochemically active surface area, because the surface of a single crystal is atomically flat.
The accelerating durability test was conducted using a protocol shown in Fig. 3. After holding the initial potential at 0.60 V for 30 s, the potential was stepped from 1.0 (3 s) to 0.60 V (3 s). This step was cycled up to 5000.  (110) and (100) step, respectively [21,36,37]. Melamine modification decreased the charges of the adsorbed hydrogen region and the oxide formation region. After melamine modification, the peaks due to (110) step (0.12 V) and (100) step (0.27 V) disappear, suggesting that melamine is adsorbed on the steps. The same trends were observed in THA + and IL-modified CVs ( Fig. S1 and Fig. S2, respectively). The coverage of the adsorbed hydrophobic species (θ) was estimated by the following equation (2): where Q H (bare) and Q H (modified) show the charge of hydrogen desorption before and after hydrophobic species modification, respectively. Table S1 summarizes the values of Q H and θ.

Results and Discussion
LSVs of melamine modified Pt 3 Fe (111) There are two reaction paths in the ORR: four electron reaction that produces water and two-electron reaction that produces H 2 O 2 . When the two-electron reaction path is open, the limiting current density decreases at lower potential [38]. No change of the limiting current density was found at lower potentials after melamine modification (Fig. 5). This is also the case of [MTBD][beti] modification (Fig. S4). On the other hand, THA + modification decreased the limiting current density at lower potentials (Fig. S3). THA + modification promotes the two-electron reaction that causes the damage to polymer electrolyte. Slight decrease of the limiting current densities shows the formation of H 2 O 2 at lower potential in Fig. 5. However, the amount of H 2 O 2 is small and the electrode is rotated at 1600 rpm; the H 2 O 2 molecules are removed from the electrode surface at 0.90 V. H 2 O 2 will not affect the ORR activity at 0.90 V.  (331), of which ORR activity is the highest in bare Pt electrodes, are also shown for comparison [5].
All the hydrophobic species improved the ORR activity on all the electrodes. Increase ratio of the activity by melamine-added modification was higher than those by other modifications in the Pt 3 Fe single-crystal electrodes examined. The ORR activity of Pt 3 Fe(111) after melamineadded modification showed the highest activity, giving j k 2.1 and 8.6 times higher than bare Pt 3 Fe(111) and Pt(331) = 3(111)-(111), respectively. The activity of Pt 3 Fe(111) after melamine-added modification reaches 39 times as high as that of bare Pt(111) [20].
The ORR activities of melamine-added modification were higher than those of melamine-immersed one in Fig. 6. In the case of melamine-immersed modification, melamine molecules exist only on the electrodes surface. Some melamine molecules will be removed from the surface during LSV with RDE. Melamine molecules exist on the electrode surface and in the solution in the case of melamine-added modification; the removed melamine molecules will be supplied from the solution. That is why melamine-added modification gave higher activity than melamine-immersed modification. Figure 7 shows increased with the increase of the potential cycles. The charge of the adsorbed hydrogen region (0.05-0.4 V) increased with increasing the potential cycles on all the bare electrodes. If the surface is heavily roughened by the potential cycles, the charge of the electric double layer between 0.4 and 0.6 V will increase. However, the electric double-layer currents did not change during the potential cycling. These results suggest that the increase of the charge in the adsorbed hydrogen region is not mainly due to the roughening of the electrode surface, but to the leaching of Fe that changes the composition of Pt 3 Fe to pure Pt. It is noteworthy that the defect peaks due to the (110) and (100) steps did not appear and the increase of the charge in the adsorbed hydrogen region was negligible after melamine-added modification. These facts indicate that melamine molecules are adsorbed at the defects (steps) and prevent the leaching of Fe from Pt 3 Fe single-crystal electrodes.
Voltammograms in the adsorbed hydrogen region were almost the same up to 5000 cycles after the melamineadded modification. This result shows the coverage of melamine is almost constant during accelerating durability test. Figure 8 shows the LSVs of Pt 3 Fe(111), Pt 3 Fe(775) = 7(111)-(111), and Pt 3 Fe(544) = 9(111)-(100) before and after melamine-added modification during durability test. The plot of j k vs. number of potential cycles is shown in Fig. 9.
The ORR activity of bare Pt 3 Fe(111) decreased by 62% after 5000 cycles, whereas the decrease was only 29% after melamine-added modification. Using the ratio of the deactivation as a measure of the durability, the durability was improved 2.1, 1.8, and 2.0 times by the melamine-added modification on Pt 3 Fe(111), Pt 3 Fe(775) = 7(111)-(111), and Pt 3 Fe(544) = 9(111)-(100), respectively. The improvement of durability by melamine-added modification was highest for Pt 3 Fe(111). Molecular size of melamine is smaller than other hydrophobic species. Melamine is easily adsorbed at the defects of (110) and (100) steps formed during the durability test and prevents the leaching of Fe from the steps [28]. Figure 10 shows the predicted adsorption model of melamine. From the DFT calculation of the dipole moment and the IR spectra of melamine adsorbed on Pt(111) and Pt(331) in ultrahigh vacuum (UHV), it is possible that adsorbed melamine is tilted on Pt(111) and Pt(331). However, it is not clear at present why melamine adopts these adsorption forms. Detailed study is now on progress in UHV in our laboratory and will be published later.