Effect of hydrophobic cations on the oxygen reduction reaction on single‒crystal platinum electrodes

Highly active catalysts for the oxygen reduction reaction are essential for the widespread and economically viable use of polymer electrolyte fuel cells. Here we report the oxygen reduction reaction activities of single‒crystal platinum electrodes in acidic solutions containing tetraalkylammonium cations with different alkyl chain lengths. The high hydrophobicity of a tetraalkylammonium cation with a longer alkyl chain enhances the oxygen reduction reaction activity. The activity on Pt(111) in the presence of tetra‒n‒hexylammonium cation is eight times as high as that without this cation, which is comparable to the activities on Pt3Co(111) and Pt3Ni(111) electrodes. Hydrophobic cations and their hydration shells destabilize the adsorbed hydroxide and adsorbed water. The hydrophobic characteristics of non‒specifically adsorbed cations can prevent the adsorption of poisoning species on the platinum electrode and form a highly efficient interface for the oxygen reduction reaction.

T he development of highly active catalysts for the oxygen reduction reaction (ORR) is a prominent issue in terms of decreasing Pt loading in cathode catalysts for polymer electrolyte fuel cells (PEFCs). Model electrocatalysts with atomically well-defined surfaces have provided detailed information regarding activation sites, optimal atomic arrangement, and the surface composition of catalysts, which have been applied to nanomaterials for practical use. In the case of Pt-based materials, the introduction of heterometals drastically improves ORR activity, depending on the surface structure [1][2][3] . PtM (M = Ni, Co, etc.) bimetal alloys perturb the density of state (DOS) of the dband and compress the lattice strain of the surface layer 4,5 . Consequently, the ORR activities on PtM(111) are increased by one order of magnitude as high as that of Pt(111) electrode 6,7 . In many heterogeneous catalysts, the catalytic activity has a volcanoshaped relationship with the atomic arrangement and alloy composition of the substrate. The variation of the d-band DOS of the substrate modifies the adsorption energy of the reaction products and the intermediate species. Therefore, the center position of the d-band DOS strongly correlates with the ORR activity. Pt 3 Ni and Pt 3 Co are located near the top in the volcanoshaped dependency between the activity and the d-band center 4 . The introduction of step and kink structures also activates the ORR; the high-index planes of Pt give the maximum activity for the ORR on the surfaces with 3−4 atomic rows of (111) terrace [8][9][10] . The catalyst development through substrate modification has reached a maturity stage, and other approaches are required for further enhancement of the ORR activity.
The activity of electrochemical reactions is affected by the solvent and electrolyte ions as well as the substrate. At the electrode/electrolyte interface, such solution species construct the electric double layer (EDL), which governs the ORR activity on Pt surfaces significantly. It is well known that strongly adsorbed anions, such as halide and sulfate ions, on the Pt surface inhibit the ORR severely [11][12][13] . The adsorbed hydroxide (OH ad ) species is formed on Pt surfaces in electrolytic solutions containing ions that weakly interact with Pt surfaces 13 . The stability of OH ad depends on the surface structure and on the electronic state of the substrate 4, [14][15][16] . Since OH ad is also an inhibitor for the ORR, the control of OH ad is a key factor in activating the ORR.
In the EDL, some hydrated cations located at the outer Helmholtz plane (OHP) are stabilized as non-specifically adsorbed species through non-covalent interactions, such as hydrogen bonds and electrostatic interactions 17 . Non-covalent interactions in the EDL shift the adsorption equilibrium and the phase transition potential of the adsorbed layer 18 . In alkaline media, the ORR activity of Pt(111) is improved exponentially as the hydration energy of the OHP cations decreases; the activity in CsOH is one order of magnitude higher than that in LiOH 19 . Cations with a high affinity toward oxygen species, such as Li + , strongly stabilize OH ad 20 , which then deactivates the ORR. Therefore, appropriate control of the structure and hydrophobicity of the interfacial cations can improve ORR activity.
In PEFCs, which use proton-exchange membranes, the control of hydrophobicity is necessary in acidic solutions. In this study, we focus on tetraalkylammonium (TAA) cations of which the hydrophobicity and interfacial structure can be controlled by the alkyl chain length. Variation of the alkyl chain length can change the hydration structure of TAA + dramatically 21,22 . We have evaluated the ORR activity on single-crystal Pt electrodes in acidic solutions containing TAA + with different alkyl chains lengths. The interfacial structures have also been determined by in situ X-ray scattering and infrared (IR) spectroscopy measurements.

Results
Electrochemical characterization. We used four types of TAA + with different alkyl chains lengths (n): tetramethylammonium (TMA + ) n = 1, tetraethylammonium (TEA + ) n = 2, tetra-nbutylammonium (TBA + ) n = 4, and tetra-n-hexylammonium (THA + ) n = 6. Figure 1a shows cyclic voltammograms (CVs) in 0.1 M HClO 4 containing 10 −5 M TBA + and 10 −6 M THA + . The solubility of TAA + in acidic solution decreases with the increase of alkyl chain length. The concentrations of TBA + and THA + used in this study are approximately equal to the saturation. The CV data show the hydrogen adsorption/desorption region between 0.05 and 0.40 V, the double layer charging/discharging region between 0.40 V and 0.60 V, and the Pt oxidation region between 0.60 V and 0.90 V. In the presence of TBA + and THA + , the onset potentials for the adsorption of hydrogen and Pt oxidation shift negatively and positively, respectively, whereas the CV data in solutions containing 10 −3 M TMA + and TEA + are identical to those without TAA + (Supplementary Figure 1a). This indicates that TAA + with longer alkyl chains affects the adsorption of hydrogen and Pt oxidation, even at low concentration.
No IR absorption band for the alkyl chain of TAA + appears, as shown in Supplementary Figure 2. Previous IR study has revealed that the specific adsorption of TAA + on Pt does not occur above 0 V 23 . Therefore, the potential shift of hydrogen adsorption is not due to site blocking by specifically adsorbed TAA + . The hydrogen adsorption step is governed by the proton-transfer process in the EDL, which depends on the conformation and hydrogen-bonding interactions of hydration water. Density functional theory (DFT) calculations suggest that hydrogen adsorption energy is affected by proton transfer in the water bilayer 24 . These results indicate that TBA + and THA + located near the electrode induce the reconstruction of interfacial water, and the inhibition of proton transfer in the EDL results in the potential shift for hydrogen adsorption. In the solutions containing 10 −3 M TMA + and TEA + , no potential shift appears even though the concentrations of TMA + and TEA + are larger by two or three orders of magnitude than those of TBA + and THA + . Since the enthalpy of hydration water with TAA + is reduced by an increase of the alkyl chain length 25 , the enhancement of hydrophobicity with longer alkyl chains strongly perturbs the interfacial water, changing the equilibrium potentials for the adsorption of hydrogen and Pt oxidation on Pt(111). Figure 1b shows linear sweep ORR voltammograms in 0.1 M HClO 4 containing TBA + and THA + (voltammograms in TMA + and TEA + containing solutions are shown in Supplementary  Figure. 1b). The specific activities (j k ) at 0.90 V in TAA + -containing solutions are shown in Fig. 1c. The ORR activity increases in the sequence THA + ≫ TBA + > TEA + > TMA + ≈ HClO 4 , and shows a correlation with the hydrophobicity of TAA + . The ORR activity in THA + is eight times greater than that in 0.1 M HClO 4 . The structural effects of the substrates on the ORR were investigated using the other typical index planes of Pt, as shown in Fig. 1d (CVs and linear sweep ORR voltammograms are shown in Supplementary Figure 3). There is no significant enhancement effect of THA + on Pt(100) and Pt(110). However, THA + enhances the activity on Pt(331) = 3(111)-(111), which gives the highest activity for the ORR on stepped surfaces, by 1.3 times. The specific activities on Pt(111) and Pt(331) in THA + -containing solution are comparable with those of Pt 3 Co and Pt 3 Ni 6,7 .
At the initial stage of the Pt oxidation, the in-plane bending mode (δ PtOH ) of OH ad is detected by IR spectroscopy. The band intensities of δ PtOH at 0.90 V correlate to the ORR activities j k at 0.90 V on single-crystal Pt electrodes, confirming that OH ad inhibits the ORR 15,16 . The stability of the hydrogen-bonding network in the coadsorbed OH ad + H 2 O layer affects the ORR activity on the (111) terrace, as described in the Discussion section. The approach of hydrophobic cation to the surface may causes the destabilization of the OH ad layer. Therefore, we determined the interfacial structures and adsorbed species on Pt(111) in THA + -containing solution in which the degree of activity enhancement was the highest in terms of surface structure and cationic species by using in situ IR spectroscopy and X-ray scattering.
Infrared spectroscopic measurement of adsorbed species. Since the absorption frequency of δ PtOH overlaps ν ClO of perchlorate ion, hydrogen fluoride (HF) is used for the IR measurements of OH ad . Figure 2a shows the IR spectra of Pt(111) and THA + modified Pt(111) in 0.1 M HClO 4 and 0.1 M HF. The potentialdependent band at 1050 cm −1 appears above 0.6 V, and is assigned to the in-plane δ PtOH 15,16,20 . Figure 2b shows the potential dependence of the band intensity for δ PtOH , and the charge density of Pt oxidation estimated from the CV data (Fig. 1a). The onset potential of OH adsorption shifts positively and the band intensity of δ PtOH at 0.90 V decreases due to the modification with THA + . This result is consistent with the charge density of Pt oxidation in the CV data. The charge density of Pt oxidation increases above 0.80 V, but the band intensity of δ PtOH decreases above 0.80 V. These facts indicate the further oxidation from PtOH to PtO above 0.80 V 14 .
The positive-and negative-absorption bands at 1650 and 1610 cm −1 are assigned to the HOH bending mode (δ HOH ) of nonadsorbed and adsorbed hydrogen bonded water, respectively 26,27 . The δ HOH of non-adsorbed water corresponds to those in the liquid phase (1645 cm −1 ) and the solid phase (1650 cm −1 ). It is known that δ HOH shifts to lower frequency by adsorption on metal surfaces. These frequencies are higher than δ HOH for an isolated water monomer (1595 cm −1 from matrix isolation measurements) 28 . Therefore, these water molecules are hydrogen bonded with neighboring oxygen species and electrolyte ions. The generation of OH ad above 0.50 V decreases the coverage of adsorbed water relative to that at the background potential of 0.3 V. The increase in the intensity of δ HOH above 0.5 V also involves the orientation change of water induced by the electrode potential, because the water dipole responds sensitively to the electric field in the EDL. The band intensities of non-adsorbed and adsorbed water are also reduced by the modification with THA + , indicating the decrease of the coverage of OH ad . No IR bands in the CH stretching and CH bending regions appear on THA + modified Pt(111) between 0.30 and 1.0 V, as shown in Supplementary Figure 2. This fact suggests that the structure of the interfacial THA + is not changed by the electrode potential.
New IR bands appear on THA + -modified Pt(111) at 1510 cm −1 , which are assigned to the HOH bending mode of adsorbed water monomers at 0.30 V. Since this band overlaps the broad positivegoing band around 1570 cm −1 , an accurate absorption frequency is unclear. The red shift from 1510 cm -1 to 1090 cm −1 in D 2 O solvent indicates that these bands are derived from water (Supplementary Figure 4). On Ni(111), Rh(111), and Ru(0001) under ultra-high vacuum condition, the δ HOH shifts from 1595 cm −1 (isolated monomer) to 1560 cm −1 (adsorbed monomer) by the adsorption via oxygen lone pair [29][30][31] . The action spectroscopic method using scanning tunneling microscopy also reveals that the signal for water monomer adsorbed on Pt(111) appear around 1550 cm −1 32 . At the solid liquid interface, since the interfacial water molecules are cross-linked via a hydrogen-bonding network, the adsorbed monomers are not a stable species on the electrode. The hydrogenbonding structure of water in the EDL is influenced by the OHP  cation and its hydration shell. The hydronium cation facilitates the construction of a tetrahedral configuration, stabilizing the ice-like structure 24,33 . Conversely, the hydration shell around hydrophobic cations such as THA + disrupts the tetrahedral hydrogen-bonding network. As a consequence, the monomeric species may be stabilized on the surface.

X-crystal truncation rod measurement of interfacial structures.
Electron density profiles along the surface normal direction were estimated from X-ray specular crystal truncation rod (CTR) scattering. Figure 3a shows the specular CTR profiles of Pt(111) in 0.1 M HClO 4 with and without 10 −6 M THA + . The CTR profiles were obtained at 0.5 and 0.9 V, which are the double layer and the OH adsorption regions, respectively. Normalized CTR profiles are shown in Supplementary Figure 5a. While the CTR profiles with and without THA + at 0.50 V are nearly identical, the CTR at 0.90 V is clearly altered by the addition of THA + . Since THA + is composed of nitrogen, carbon, and hydrogen, the electron density in the EDL is significantly lower than that of Pt. Therefore, the primary factor for this variation in the CTR is the structural change in the substrate Pt. The initial model used for structural optimization comprises three Pt layers and one oxygen layer of water molecule or OH ad layer. The vertical atomic position, occupancy factor, and Debye-Waller factors for the Pt. and oxygen layers were optimized by the least-squares method. The electron density profiles and the structural parameters of the optimized model are shown in Supplementary Figure 5b and Supplementary  . This surface relaxation is caused by charge transfer from the adsorbed water to Pt and the electric field 37 . The coverage of oxygen species (θ O ) slightly increases for the THA +containing solution. IR measurements of interfacial water have been performed by many researchers, revealing that the orientation of the adsorbed water and the hydrogen-bonding networks are altered by the interaction with the electrolyte ions 38,39 . The increase of the coverage indicates the existence of THA + at the interface. The IR spectra in Fig. 2a show that OH ad is not adsorbed at 0.50 V. However, the negative-going band at 1610 cm −1 for adsorbed water on Pt(111) without THA + is larger than that with THA + . This fact indicates that the coverage of adsorbed water θ O on Pt(111) with THA + is higher than that without THA + in 0.1 M HClO 4 . At 0.9 V, The d O-1stPt in HClO 4 solution decreases from 2.47 Å to 2.14 Å. This corresponds to the Pt-O bond length of OH ad on Pt(111) proposed by low-energy electron diffraction and in situ Xray diffraction measurements 34,40 . Potential-dependent coadsorbed structures of water and OH ad were suggested by DFT calculations and Monto Carlo simulations 41,42 . XPS measurements also indicate the coadsorption of water and OH ad 14 , hence the coverage at 0.90 V (θ O = 0.78) includes adsorbed water and OH species. The surface relaxation of the Pt layer is promoted by electron donation via an oxygen lone pair of the OH ad as well as adsorbed water 43 . Therefore, d 1stPt-2ndPt is comparable with that at 0.50 V. Conversely, in the presence of THA + , d O-1stPt is 2.38 Å, which indicates that adsorbed water is dominant instead of OH ad . It is notable that the surface relaxation is eliminated (d 1stPt-2ndPt = 2.29 Å) even though the water and OH ad are adsorbed on Pt. This result indicates that the interaction between the oxygen species and Pt is decreased by the presence of THA + .

Discussion
The enhancement of the ORR activity by the addition of TAA + is related with the hydrophobicity of TAA + . Figure 4 shows the correlation between hydration enthalpy and the ORR activity. A linear correlation is found except TMA + , and the slope for TAA + is similar to that for alkali metal cations, as reported by Markovic et al. 19 . They suggested that the non-covalent interaction between hydrated cations and OH ad affects electrocatalytic reaction on Pt 19,44,45 . In alkaline solution, the ORR activity is related to the hydration energy of alkali metal cation because cations with a strong affinity for adsorbed oxygen species block the active sites for the ORR. The order of the hydration energy of alkali metal cations (Li + ≫; Na + > K + > Cs + ) is inversely correlated to the ORR activity (Cs + > K + > Na + » Li + ). Previous IR and X-ray studies have revealed that Li + located at the OHP inhibits the surface oxidation of Pt(111) due to the stabilization effect between Li + and OH ad , whereas high-order oxidation accompanied by surface roughness proceeds in solution containing Cs + 20 . Thus, OH ad is stabilized by the interaction with species that have high-oxygen affinity. The hydration structures of interfacial cations were assumed from the OH stretching band frequency 21,22 . The hydration water around hydrophilic cations, such as H + and Li + , is coordinated with the dipole moment pointed outward and the hydrogen atoms of the hydration water, which can link with the oxygen species located in the secondary hydration sphere. Conversely, hydrophobic cations such as TAA + weaken the bonding with hydration water and strengthen intermolecular hydrogen bonding in the primary hydration shell. Consequently, the hydration shell around the hydrophobic cation is restricted to coordination with the outer oxygen species because of the complete hydrogen-bonding network within the shell. The ORR activity in TMA + is comparable with that in HClO 4 without TAA + , indicating that TMA + does not contribute the enhancement of the ORR. Since the hydration shell size of TMA + is smaller than those of the other TAA + , the complete hydrogen-bonding network may not be constructed in the shell.
The OHP cations interact with the charged species in the EDL as well as the substrate. Therefore, cations electrostatically interacting with adsorbed anions can approach the surface at the potentials more positive than the potential of zero charge (pzc) 17,20,46,47 . The coverage of the OHP cations is balanced with that of the adsorbed anions and the surface charge depending on the electrode potential. On the surface at a constant coverage of adsorbed counter anions, the coverage of the OHP cation decreases with increasing potential 17 . As described above, the IR spectra indicate that the coverage of THA + is approximately constant in the potential region examined in this study. The increase in the coverage of the negatively charged species, i.e., OH ad and O ad , is balanced by positive surface charge so that the coverage of the THA + remains constant.
While the ORR activity on Pt(111) is dramatically enhanced by the presence of THA + , there is no significant enhancement of the ORR on Pt(100), Pt(110) and Pt(331) by THA + . In acidic solution above 0.6 V, the hydration water around H + is linked with OH ad through hydrogen bond (Fig. 5a). H + -H 2 O-OH ad formation stabilizes the OH ad layer and inhibits the ORR. According to studies on the Pt(111) surface under the ultra-high vacuum condition, the coadsorption of OH ad and H 2 O forms a well-defined 3 × 3 honeycomb structure because the symmetry and OH•••O distance in the coadsorbed layer fit well with the Pt(111) lattice 14,48 . The approach of the hydration shell around THA + to the Pt surface disrupts the stabilization effect between the hydration water and the OH ad layer (Fig. 5b). The coverage of OH ad is decreased by this destabilization effect, which induces the disruption of stable hydrogen-bonding network in the coadsorbed layer. These multiple effects are enhanced by the hydrophobicity of cation and promote the access of oxygen molecules to the Pt(111) surface.
The IR spectra on Pt(100) indicate that the addition of THA + results in a slight reduction of the band intensity and band broadening of δ PtOH as shown in Supplementary Figure 6. The charge densities for surface oxidation also decrease on Pt(110) and Pt(331) above 0.7 V after the addition of THA + . For Pt(100),   (111) in HClO 4 . The activation factor for the ORR with these Pt alloys is also due to decrease in the coverage of OH ad . Therefore, a similar inhibiting effect of OH ad can be achieved by the structural improvement in EDL species. Furthermore, the supply of oxygen molecules to the electrode may be improved by weakening the hydrogen-bonding strength around the hydrophobic cations.
In summary, we evaluated the ORR activities of single-crystal Pt electrodes in HClO 4 solutions containing TAA + with various alkyl chain lengths. The ORR activity increased in the following sequence THA + n = 6 » TBA + n = 4 > TEA + n = 2 > TMA + n = 1 ≈ HClO 4 , and the ORR activity on Pt(111) in THA + was comparable with those on Pt 3 Co(111) and Pt 3 Ni(111) in HClO 4 . In situ IR and X-ray CTR measurements indicated that the coverage of OH ad on Pt(111) at 0.90 V in the presence of THA + is decreased. The weaker interaction between the hydration shell around THA + and OH ad destabilizes the OH ad + H 2 O coadsobed layer and forms a highly efficient interface for the ORR.

Methods
Electrochemical measurement. Single-crystal Pt beads for CV measurement were prepared by the Clavilier's method 50 . The crystal beads were oriented to (111), (100), (110), and (331) using Laue back reflection of X-ray, and then mechanically polished with diamond slurries 51 . A Pt(111) disk (8 mm in diameter, MaTech) was used for the X-ray CTR measurement. The samples were annealed in H 2 + O 2 flame, and then cooled to room temperature in Ar + H 2 or Ar atmosphere. The annealed surfaces were protected with ultrapure water (Milli-Q Advantage A10, Millipore). The solutions were prepared with HClO 4 (Kanto Chemical), HF (Kanto Chemical), tetraalkylammonium salts, and ultrapure water. The tetraalkylammonium salts used were tetramethylammonium perchlorate (Tokyo Chemical Industry), tetraethylammonium perchlorate (Tokyo Chemical Industry), tetra-nbutylammonium perchlorate (Sigma-Aldrich), and tetra-n-hexylammonium perchlorate (Alfa Aesar). The reference electrode used in all measurements was a reversible hydrogen electrode (RHE). The ORR voltammograms were measured in the hanging meniscus rotating disk electrode (RDE) configuration using an electrochemical analyzer (ALS 701DH, BAS) and a rotating ring-disk electrode apparatus (RRDE-3 A, BAS) 52 . The ORR activity was estimated from the kinetic current density (specific activity, j k ) at 0.90 V versus RHE according to the Koutecky-Levich equation: 1/j = 1/j k + 1/j L , where j and j L are the total current density and the limiting current density, respectively. The current densities were normalized to the geometrical surface area of the Pt surface.
Infrared spectroscopy. In situ infrared reflection absorption spectroscopy (IRAS) measurement was performed using an Fourier transform (FTIR) spectrometer (Bruker Vertex 70 v). An electrochemical IR cell with a trapezoid BaF 2 or CaF 2 window beveled at 60°was attached to the spectrometer with a narrow band mercury-cadmium-telluride (MCT) detector. A polypropylene film (Chemplex Industries) with a thickness of 6 µm was intercalated between the BaF 2 prism and the Pt(111) and Pt(100) surfaces to prevent the dissolution of the BaF 2 material 20 . A layer of water was placed between the BaF 2 prism and the polymer film to compensate for the gap of the refractive index. The interferograms were averaged over 640 scans by subtractively normalized interfacial FTIR (SNIFTIRS) with ppolarized light at a resolution of 4 cm −1 . In THA + -containing solution, the adsorption of OH ad is inhibited by the repetitive potential steps of SNIFTIRS measurement because THA + assembles on the surface at the negative potentials. Therefore, after the Pt(111) and Pt(100) electrodes were immersed in 0.1 M HClO 4 and 0.1 M HF containing THA + for 5 min, IR measurement was performed in the solution without THA + . We confirmed that Pt(111) modified with THA + shows the same voltammograms and ORR activity.
X-ray crystal truncation rod scattering. In situ X-ray CTR measurement was performed with a multi-axis diffractometer at BL13XU (SPring-8) 53 . The X-ray energy used was 20 keV. Integrated intensities were measured by rocking scans, and then corrected for Lorentz and area factors. Specular CTR was measured along the surface normal direction L in units of c* = 2π/c, |c| = 0.6797 nm. Structure refinements were conducted using the least-squares method with the ANA-ROD program 54 .

Data availability
The data that support the findings of this study are available within the paper and Supplementary