Adsorption of ionomer and ionic liquid on model Pt catalysts for polymer electrolyte fuel cells

Graphical Abstract

The ionic liquid of [C4C1im][NTf2] is preferentially adsorbed on the step sites of Pt surface through the anionic moieties and mitigates the adsorption of ionomer on the Pt surface.

Abstract

The adsorption of the perfluoro-sulfonic acid polymer of Nafion and ionic liquid (IL) of 1-butyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide on the surface of Pt was investigated via voltammetric analyses, using stepped Pt single-crystal electrodes with (111) terraces and (110) steps, and surface-enhanced infrared absorption spectroscopy (SEIRAS) analyses using a Pt polycrystalline electrode. Sulfonate anion in Nafion was adsorbed on the stepped Pt single-crystal electrodes and suppressed the oxygen reduction reaction (ORR) activity by more than 50%, regardless of the terrace width. The IL molecules were preferentially adsorbed on the step sites through a simple IL coating procedure. The SEIRAS analysis indicated that the IL molecules were stable on the Pt surface throughout potential cycles, where the anionic moieties were in contact with the Pt surface and reoriented depending on the potential. The IL modification prior to Nafion coating mitigated ionomer adsorption on the Pt surface. However, the mitigation effect was not reflected in the ORR activity because water production led to IL desorption during the ORR activity measurement. Accordingly, IL modification is a promising method for improving the performance of Pt catalysts in polymer electrolyte fuel cells; however, further studies to prevent the leaching of IL are required for practical applications of this approach.

1 INTRODUCTION

Polymer electrolyte fuel cells (PEFCs) have promising applications in automobiles as well as other industries and play key roles in sustainable energy systems. Developing highly active and durable catalysts for cathodes, where the oxygen reduction reaction (ORR) occurs, is essential in the development of PEFCs. The activity and durability of ORR catalysts as well as oxygen permeability in the local area of catalysts, which strongly affects the power density of the cell, frequently have trade-off relationship with each other,^[1] and understanding the ORR mechanisms (e.g., the locations of reaction sites and the role of spectator species) at the atomic and molecular scales is essential for overcoming the trade-off issue.^[2] In the catalyst layer of PEFCs, polymer electrolyte, referred to as ionomer, is used as proton conductor and binder, and elucidating the behavior of ionomer molecules in the ORR is crucial in PEFC-related studies.^[3]

Using the model catalyst of the Pt (111) single-crystal electrode, the adsorption of the anionic moieties in the ionomer on the Pt surface was evaluated by the characteristic peaks observed in the cyclic voltammogram (CV),^[4] and the amount of anion adsorption, as well as its effect on the ORR, was quantified.^[5] The results indicated that the ORR activity was significantly suppressed by more than 60% at 0.9 V (i.e., less than 40% of the Pt surface area was available for ORR) by the coating of the Pt (111) surface with Nafion ionomer, while the coverage of sulfonate anions on the Pt surface was small (ca. 0.1 monolayer [ML]). Furthermore, the experimental method of voltammetry using Pt (111) was applied to ionomers with various molecular structures,^{[3, 5, 6]} and a strategy for designing the molecular structure of ionomer was established.

Compared to the single-crystal electrode, the real Pt nanoparticle catalyst has a complex surface morphology with defect sites such as edges and corners in addition to (111) facets with finite sizes.^[7] Therefore, correlating the results obtained from the single-crystal experiments with those based on Pt nanoparticles is imperative. The effect of Nafion coating on the ORR activity of Pt nanoparticles was studied in the configurations of the membrane electrode assembly and rotating disk electrode (RDE)^[8] combined with the CO displacement method^[9] for the quantification of the sulfonate coverage. The results indicated that the ORR activity of the Pt nanoparticles was suppressed by more than 50% through the adsorption of the sulfonate anions.

In addition, the effects of the ionomer molecular structure on the ORR activity of Pt nanoparticles and Pt (111) single-crystal electrodes were similar.^[3] Thus, the information acquired from the studies involving Pt (111) single-crystal electrodes has been successfully applied to the development of practical PEFCs with Pt nanoparticles.^{[3, 8]} However, experimental studies on the ionomer effects bridging the Pt (111) surface with the real nanocatalysts and the detailed mechanisms of ionomer adsorption on Pt nanoparticles have not been published thus far. Elucidating the effects of ionomer and its adsorption mechanisms is crucial, particularly for the application of shape-controlled Pt-based nanocatalysts, which often exhibit lower ORR activities in cells than those expected from RDE experiments.^[10] For predicting the behavior of these catalysts in PEFCs, it is necessary to clarify the dependence of the ionomer adsorption mechanism and ORR activity on the surface morphology of Pt catalysts. In this respect, experiments involving high-index planes (i.e., stepped surfaces) of single-crystal electrodes provide beneficial information because they possess various sites in a controlled manner. The adsorption of Nafion on stepped single-crystal Pt surfaces was studied by Ahmed et al.,^[11] but its effect on ORR activity was not examined.

In addition to designing the molecular structure of ionomer, controlling the interface between the catalyst surface and ionomer is a promising approach for mitigating the ionomer-induced suppression of the ORR activity. For example, employing mesoporous carbon as catalyst support, where Pt nanoparticles inside the pores are not in contact with the ionomers, has been effective for improving ORR activity.^[12] Another approach is to reduce the amount of ionomer in the vicinity of the catalysts while maintaining the proton conductivity in the catalyst layer using ionomer fibers.^[13] Recently, the mitigation of the ionomer-induced ORR suppression by modifying the catalyst surface with foreign materials such as thin carbon layer,^[14] hydrophobic cations,^[15] and ionic liquids (ILs)^[16] has been reported. The approach of catalyst surface modification is promising because the modification process can be applied as a post-treatment to the synthesized catalyst without significantly changing the catalyst synthesis conditions. In addition, it is notable that some modifiers improved the intrinsic activity and durability of the catalyst.^{[2, 17]} Among the aforementioned surface modifiers, the beneficial effects of ILs on the intrinsic activity and durability of the catalyst have been confirmed by several groups.^[18] The suppression of oxide formation has been proposed as the origin of the improvements from voltammetric behaviors; however, further analyses on the adsorption site and orientation of IL molecules are required to clarify the role of ILs.^[19]

In the present study, the adsorption of Nafion and IL on Pt catalyst and its effects on ORR activity are investigated using high-index planes of a single-crystal electrode. Nafion adsorption is analyzed by focusing on the effect of (111) terrace width on the adsorption site and the degree of ORR suppression. Furthermore, IL adsorption is studied with a stepped Pt single-crystal surface, where the adsorption sites can be identified in situ from the voltammogram.^{[2, 20]} Surface-enhanced infrared absorption spectroscopy (SEIRAS)^[21] with a Pt polycrystalline electrode is applied to study the adsorption state of the IL. These characterizations provide useful information for discerning the mechanisms of the reported ORR activity and durability enhancements by the IL modification. Subsequently, sequential coatings of IL and Nafion are applied to a stepped single-crystal electrode to test the mitigation of ionomer-induced ORR suppression by the overlayer introduced between the catalyst and ionomer.^{[14, 15]}

2 METHODS

2.1 Voltammetry with stepped single-crystal electrodes

The electrode surface was prepared by annealing a Pt single-crystal disk surfaced with a plane of (hkl) (99.99%, 0.196 cm², purchased from MaTecK) in a reductive atmosphere.^[2] The tested surfaces were (111), (554), (443), (332), (221), and (331), which comprise (111) terraces with atomic widths of n = ∞, 9, 7, 5, 3 and 2, respectively, and (110) monatomic steps. After annealing, the electrode surface was coated with a Nafion thin film with a thickness of approximately 35 nm using a method described in a previous study.^[6] Electrochemical measurements were conducted in 0.1 M HClO₄ in the hanging-meniscus RDE configuration. CVs under deaerated conditions and a linear sweep voltammogram (LSV) under O₂ saturated conditions were recorded to evaluate the ionomer adsorption and ORR activity, respectively.

To study IL modification, the Pt surface plane of (443) and the IL of 1-butyl-3-methylimidazolium bis(trifluoromethanesulfonyl)imide ([C₄C₁im][NTf₂]) depicted in Figure 1^[22] were used. After annealing the (443) single-crystal disk, the electrode surface was protected with a droplet of ultrapure water. Afterward, 20 μl of the IL solution (0.02 wt.% of the IL in a 75 vol% 2-propanol aqueous solution) was prepared from 99.5% [C₄C₁im][NTf₂] (IOLITEC GmbH) was deposited on the surface and then dried under Ar flow. The amount of the IL solution deposited was equivalent to 15 ML of the IL molecule. The IL-coated electrode was subjected to RDE tests with and without a sequential coating of the Nafion thin film.

FIGURE 1

(a) Top and side views of Pt (443) drawn using VESTA (ver. 3)^[22] and (b) the molecular structure of [C₄C₁im][NTf₂]

2.2 SEIRAS measurements with a polycrystalline electrode

The experimental setup for the SEIRAS measurements under potential control is described elsewhere.^[23] The measurements were performed with a Pt polycrystalline film on a Si prism with the Kretschmann attenuated total reflection (ATR) configuration before and after coating with the IL (Figure 2).

FIGURE 2

Schematic of the surface-enhanced infrared absorption spectroscopy (SEIRAS) measurement procedure. P. S. means potentiostat

The chemically deposited Pt film on the prism ^[24] was cleaned in 0.1 M HClO₄ with potential cycles in the range of 0.08–1.33 V versus a reversible hydrogen electrode (RHE). A SEIRA spectrum was recorded at 0.43 V to determine the background under the electrochemical conditions, and the obtained spectrum is denoted as BG2. The electrolyte was then removed, and a droplet of ultrapure water was deposited to keep the electrode wet. Subsequently, a SEIRA spectrum was recorded to determine the background for the IL coating process, and the obtained spectrum is denoted as BG1. Then 400 μl of the IL solution (0.25 wt.% of [C₄C₁im][NTf₂] in ethanol) was deposited on the Pt surface while a water film remained on it. During the evaporation of the solvents, SEIRA difference spectra with respect to BG1 were obtained without potential control to confirm the adsorption of the IL on the Pt film (Measurement 1). Then, the electrolyte of 0.1 M HClO₄ was added, and a SEIRA difference spectrum with respect to BG2 was obtained at 0.43 V (Measurement 2). Finally, to examine the effect of the potential on the adsorption state of the IL, SEIRA difference spectra, with respect to the spectrum at 0.43 V after the IL coating and electrolyte addition, were recorded while the potential was scanned from 0.43 to 0.08 V, then to 1.33 V, and then back to 0.43 V at a scan rate of 10 mV/s. We performed all the SEIRAS experiments while keeping the cell attached to the spectrometer because the actual background spectrum would be altered by a slight change in the cell-spectrometer configuration. To remove the electrolyte and coat the IL onto the Pt film without dismounting the cell from the spectrometer, we employed an electrolyte container with a detachable upper part, as shown in Figure S1.

3 RESULTS AND DISCUSSION

3.1 Ionomer adsorption

Figure 3 shows the CVs and LSVs for the single-crystal Pt surfaces with and without the Nafion coating. In the CV for the Nafion-coated Pt (111) surface, sharp redox peaks ascribed to the adsorption and desorption of sulfonate anions are observed in the double-layer region (i.e., in the range of 0.4–0.6 V).^[4] In addition, the CV exhibits the suppression of the butterfly peaks in the range of 0.6–0.8 V, indicating the inhibition of hydroxyl adsorbate (OH_ads) formation by the adsorbed sulfonates. Redox peaks in the double-layer region are also observed for the Pt (554), (443), and (332) surfaces, indicating that the sulfonate anions in Nafion are also adsorbed on the (111) terraces of these surfaces.

FIGURE 3

Cyclic voltammograms (CVs) and linear sweep voltammograms (LSVs) for (a) Pt(111), (b) Pt(554), (c) Pt(443), (d) Pt(332), (e) Pt(221), and (f) Pt(331) with and without Nafion coating

The cathodic scan for the Nafion-coated Pt (221) surface shows a small sharp peak at 0.18 V in the hydrogen adsorption region, in addition to a small broad peak at 0.5 V in the double layer region. For the Nafion-coated Pt (331) surface, sharp redox peaks are observed at 0.16 V, whereas no peaks are discernible in the double-layer region. These observations suggest that the adsorption of sulfonate anions occurs in the double-layer region for surfaces with large terrace widths (i.e., when n≧3), whereas it occurs in the hydrogen adsorption region (i.e., at potentials lower than 0.3 V) for surfaces with small terrace widths (i.e., when n ≦ 3). For the Pt (221) and Pt (331) surfaces, the currents on the higher potential side of the anion adsorption peak are suppressed and this can be ascribed to the suppression of the anodic adsorption of some species.

Jinnouchi et al.^[25] demonstrated through density functional theory (DFT) calculations that OH species are adsorbed on the step sites on Pt surfaces with narrow terraces at low potentials (ca. 0.3 V). Therefore, the species whose adsorption was suppressed on the Pt (221) and Pt (331) surfaces as shown above in the present study are probably the OH species on the step sites, and accordingly, the step sites are the adsorption sites of the sulfonate anions on the Pt surfaces with narrow terraces. The change in the site for sulfonate adsorption, which competes with OH_ads formation, from step to a terrace with increasing terrace width was also proposed by Ahmed et al.^[11] using Pt surfaces comprising (100) terraces or steps. In conclusion, sulfonates in Nafion are adsorbed on all Pt surfaces, whereas the adsorption sites can be different (terrace or step) depending on the terrace width.

The coverage of the adsorbed sulfonates was estimated from the charge of the desorption peak by assuming a one-electron transfer reaction for the process.

$$$$\begin{equation}{\rm{Coverage}}\left( {{\rm{ML}}} \right) = \frac{{{\rm{Charge}}\;{\rm{of}}\;{\rm{anion}}\;{\rm{desorption}}\;{\rm{peak}}\;{\rm{per}}\,{\rm{electrode}}\,{\rm{area}}}}{{{\rm{Number}}\;{\rm{of}}\;{\rm{suraface}}\;{\rm{Pt}}\;{\rm{atoms}}\;{\rm{per}}\;{\rm{electrode}}\;{\rm{area}}\; \times \;{\rm{elementary}}\;{\rm{charge}}}}\end{equation}$$$$(1)

The method of determining the background for calculating the charge is illustrated in Figure S2. In this study, the Pt atom at the root of the step, which is partly hindered by the step Pt atom, was not counted as a surface Pt atom. Figure 4 shows the relationship between the sulfonate coverage and step density. The coverage tends to decrease with increasing the step density, but it increases again from Pt (221) to (331), reaching the same level as Pt (111).

FIGURE 4

Relation between sulfonate coverage and step density of Pt surface

In the LSVs (Figure 3), the Nafion-coated electrodes exhibit lower currents both in the activity-dominating (i.e., above 0.8 V) and diffusion-limiting (i.e., 0.3–0.8 V) regions than the corresponding bare electrodes. As the thickness of the Nafion film (ca. 35 nm) is much smaller than that of the diffusion layer in the electrolyte with the rotation rate of 1600 rpm, ca. 15 μm,^[26] the difference in the diffusion-limiting current is probably not owing to the additional diffusion resistance by the Nafion film, but owing to the difference in the convection conditions in the experiments performed with and without the Nafion coating. It is not clear whether the Nafion film affected the convection condition; however, the ORR activity can be estimated from the LSV with the usual mass transport correction^[26]:

$$$$\begin{equation}{i_{\rm{k}}}\;\left( E \right) = \frac{{{i_{\rm{d}}}i\left( E \right)}}{{{i_{\rm{d}}} - i\left( E \right)}} \end{equation}$$$$(2)

where $${i_{\rm{k}}}( E )$$ is the ORR kinetic current (activity), i_d is the diffusion-limiting current, $$i( E )$$ is the measured current, and E is the electrode potential.

Figure 5 shows the relationship between the ORR activity at 0.9 V and step density for the bare and Nafion-coated Pt surfaces. For the bare Pt surface, the ORR activity exhibits a volcano-type trend and is maximized at Pt (221). Similar results have been reported in previous studies^[27] and this trend has been attributed to the optimized binding energy of the oxygenated species at an intermediate step density.^[28]

FIGURE 5

Relations between the oxygen reduction reaction (ORR) activity at 0.9 V and step density of Pt surfaces with and without Nafion coating

The ORR activities were suppressed by the Nafion coating for all the Pt surfaces, whereas the volcano-type trend was still observed. Figure 6a shows the ratio of the ORR activity at 0.9 V on the Nafion-coated Pt surface to that on the bare Pt surface as a function of the step density. For the Pt (111) and (554) surfaces, which have wide terrace widths, the ratios are lower than 0.4, indicating the significant suppression of the ORR activity by the Nafion coating. For the other Pt surfaces, the ratios are slightly higher (ca. 0.5); however, the suppression degrees are still significant.

FIGURE 6

Ratio of the oxygen reduction reaction (ORR) activity at 0.9 V on Nafion-coated Pt electrode to that on bare Pt electrode (a) versus step density, and (b) versus sulfonate coverage

Figure 6a also shows that the ratio does not systematically change with the step density. Figure 6b shows the relationship between the ORR activity ratio and estimated sulfonate coverage (shown in Figure 4). In contrast to the presumption that the suppression of the ORR activity becomes significant with increasing sulfonate coverage,^[5] an ambiguous correlation is observed.^[5] The suppression degree is probably determined by a combination of various factors—the activity of the bare Pt surface, sulfonate anion coverage, adsorption site of the sulfonate anions. Although further studies are necessary for understanding the complex behavior shown in Figure 6, sulfonate adsorption was found to significantly suppress the ORR activity on the Pt surface regardless of the terrace width and to be a critical issue in developing highly efficient PEFCs.

3.2 Ionic liquid adsorption

3.2.1 Voltammetric analysis with a single crystal surface

Figure 7 shows the CVs of the bare and IL-coated Pt (443) surfaces as well as the bare Pt (111) surface. The CV for the Pt (443) coated with the IL of [C₄C₁im][NTf₂] was stable during the potential cycles and the voltammogram obtained during the fifth cycle is shown. For the bare Pt (443) surface, sharp peaks due to under-potentially-deposited hydrogen (H_upd) at the (110) step site are observed at 0.13 V, in addition to the plateaus due to H_upd on the (111) terrace.^[7]

FIGURE 7

Cyclic voltammograms (CVs) of the bare and coated Pt (443) with [C4C1im][NTf2] as well as the CV of bare Pt (111)

Following the IL coating, the sharp peaks disappear, whereas the (111) plateaus remain; hence, H_upd formation on the steps is considered to be blocked by some species. This observation suggests that the IL molecules are selectively adsorbed on the step sites, and is reasonable because low-coordinated Pt atoms generally possess large binding energies.^[2] This behavior is reminiscent of previous results with gold atoms^[2] and organic molecules^[20], which improved the ORR activity of stepped Pt single-crystal electrodes by changing the catalytic properties of the remaining free Pt sites. For a real Pt nanoparticle catalyst, Zhang et al.^[18] observed the disappearance of a pre-peak during CO stripping voltammetry with IL modification and suggested the preferential adsorption of the IL on defect sites as the origin of the ORR activity enhancements.^{[2, 18, 29, 20]}

The ORR activity enhancement was, however, not observed in the present study, as shown in Figure S3. The CV after the ORR measurement (Figure S4) exhibits the re-appearance of the step-induced peaks, suggesting the desorption of the IL molecules from the step sites during the ORR measurement. Although the IL was stable during potential cycles under inert conditions (Figure 7), the water produced during the ORR might facilitate IL desorption. The differences in the experimental conditions between the present study and previous studies where the ORR activity enhancement was observed^{[18, 29]} are the catalyst type (bulk electrode vs. Pt nanoparticles supported on high-surface-area carbon) and the IL type ([C₄C₁im][NTf₂] vs. [7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene][bis(perfluoroethylsulfonyl)imide] (also called [MTBD][beti])). Further studies are necessary for clarifying the origin of the difference in the dissolubility of ILs and applying the concept of IL modification in practical PEFCs.

3.2.2 SEIRAS analyses with a polycrystalline electrode

Figure 8 shows the SEIRA spectra recorded immediately after the IL coating (Measurement 1, Section 2.2) and after adding 0.1 M HClO₄ and holding the potential at 0.43 V (Measurement 2, Section 2.2). The spectrum of the bulk IL measured using the ATR setup without the Pt film and the band assignments ^{[23, 30]} are also shown in the figure.

FIGURE 8

Surface-enhanced infrared absorption spectroscopy (SEIRAS) spectrum of Measurement 1 (the difference spectrum just after the ionic liquid (IL) coating with respect to just before the coating, i.e., with respect to BG1) and SEIRA spectrum of Measurement 2 (the difference spectrum for the IL-coated Pt in 0.1 M HClO4 with the potential hold at 0.43 V with respect to the Pt under the same electrochemical condition without the IL, i.e., with respect to BG2). The bulk IL spectrum measured using the attenuated total reflection (ATR) setup without the Pt film and the assignments for the bulk spectrum [23, 30] are also shown. The results for the ranges of (a) 3200–2800 cm^–1 and (b) 1500–1000 cm^–1 are separately shown

In the SEIRA spectrum of Measurement 1 (i.e., the difference spectrum recorded immediately after the IL coating with respect to the spectrum recorded immediately before the coating (i.e., BG1)), the bands ascribed to the anionic moieties of the IL (i.e., ν_as(SO₂), ν_s(CF₃), and ν_as(CF₃)) are discernible as those in the bulk spectrum, as shown in Figure 8b. The significant blue-shift of the ν(CF₃) band, when compared with the one in the bulk spectrum, is a typical feature of the SEIRA spectra of ILs containing the [NTf₂] anion, which can be explained by the large dispersion of the refractive index.^{[23, 30]} In contrast, the bands ascribed to the cationic moieties of the IL (i.e., ν(CH)_ring and (CH)_alkyl) in the SEIRA spectrum are not discernible when compared with those in the bulk spectrum (Figure 8a), although the cations exist in the IL thin film. Because the IR absorption is more enhanced for molecules located closer to the Pt surface, the results suggest that the IL is bound to the Pt surface through the anionic moieties before the addition of the electrolyte.

Next, the SEIRA spectrum of Measurement 2 (i.e., the difference spectrum for Pt in 0.1 M HClO₄ with the potential hold at 0.43 V after the IL coating with respect to the spectrum of Pt recorded under the same electrochemical conditions but without the IL coating (i.e., BG2)) is discussed. In Figure 8a,b large positive band is observed at 1120 cm^–1. This band is not due to the IL molecules but due to the perchlorate anions although the reason why the perchlorate anion band is enhanced by the IL coating is not clear. In the region of 1410–1200 cm^–1, three positive bands (labeled A–C in Figure 8b) are observed. The bands A and C can be ascribed to ν_as(SO₂) and ν(CF₃), respectively, because similar wavenumbers corresponding to these bands are observed in the spectrum for Measurement 1 and the bulk spectrum. Thus, the bands ascribed to the anionic moieties of the IL are observed after the electrolyte addition and potential control. The band B has no corresponding band(s) in the spectrum for Measurement 1 and the bulk spectrum and may be ascribed to impurities. In the region of 3200–2800 cm^–1 (Figure 8a), a broad negative peak due to H₂O is observed probably because the hydrophobic IL coating removes the H₂O molecules directly covering the Pt electrode. The bands for the cationic moieties are not discernible; hence, the adsorption configuration immediately after the IL coating (i.e., the IL molecule is adsorbed through the anionic moiety) is retained even after the addition of the electrolyte at a potential of 0.43 V.

Figure 9 shows the SEIRA spectra recorded during a potential cycle in the range of 0.08–1.33 V, starting from 0.43 V with a negative sweep. In this measurement, the background spectrum was recorded at the starting point of the cycle (i.e., 0.43 V), which was immediately before the potential cycle. In the region of 3200–2800 cm^–1 (Figure 9a), corresponding to the cation vibrations, neither positive nor negative bands are discernable. In contrast, in the region of 1500–1000 cm^–1 (Figure 9b), corresponding to the anion vibrations, characteristic spectral changes are observed. In the wavenumber range of 1410 –1330 cm^–1 (labeled region A), corresponding to ν_as(SO₂), the spectrum grows positively at the low-wavenumber side (i.e., 1380–1330 cm^–1) and negatively at the high-wavenumber side (i.e., 1410–1390 cm^–1) at high potentials. Although the wavenumbers of the observed bands do not completely agree with the reported values (1350 and 1362 cm^–1),^[23] the positive and negative bands may be assigned to ν_as(SO₂) out-of-phase (o.p.) and ν_as(SO₂) in-phase (i.p.), respectively. In the wavenumber range of 1290–1200 cm^–1 (labeled region C), the spectrum grows negatively at the low-wavenumber side (i.e., 1250–1200 cm^–1), which corresponds to ν_as(CF₃), and positively at the high-wavenumber side (1290–1250 cm^–1), which corresponds to ν_s(CF₃), at high potentials. Such opposite behaviors of the anion vibrations cannot be explained by simple adsorption/desorption processes but can be explained by reorientation on the basis of the surface selection rule of SEIRAS, in which only the vibrational modes with oscillating dipoles perpendicular to the surface are observable.^[21] At the starting point of the cycle (i.e., 0.43 V), the orientation of the anions should be random because the relative intensity of each vibrational mode is similar to that observed for the bulk IL, as shown in Figure 8. To simulate the potential-dependent change, the direction of the oscillating dipole of each vibrational mode was obtained from DFT calculations using Gaussian09. We found that “parallel orientation”, in which the two CF₃ groups pointing toward the bulk and the two SO₂ groups pointing toward the Pt surface (see Figure S5), shows the oscillating dipoles of ν_as(SO₂)_i.p. and ν_as(CF₃) nearly parallel to the surface and those of ν_as(SO₂)_o.p. and ν_s(CF₃) possessing considerable perpendicular components. According to the SEIRAS selection rule, the two former vibrations are not observable, while the two latter vibrations are observable in the SEIRA spectrum; therefore, the above spectral changes (i.e., the negative growth of the ν_as(SO₂)_i.p. and ν_as(CF₃) bands and positive growth of the ν_as(SO₂)_o.p. and ν_s(CF₃) bands) can be ascribed to the change in orientation from random to “parallel” at high potentials. This molecular behavior is not surprising because the negatively charged part (SNS) is expected to approach the Pt surface with increasing potential, and these orientations have also been observed in SEIRAS experiments on the [C₄C₁im][NTf₂]/Au interface^[30] and X-ray photoelectron spectroscopy experiments with a 1,3-dimethylimidazolium bis(trifluoromethyl)imide ([C₁C₁im][NTf₂]) thin film on Au.^[31] These spectral features in regions A and C disappear at the end of the potential cycle and this reversible behavior indicates that the anionic moieties are not desorbed but only reoriented during the potential cycle.

FIGURE 9

Surface-enhanced infrared absorption (SEIRA) spectra during a potential cycle between 0.08–1.33 V starting from 0.43 V with a negative sweep where the background was taken at the starting point (0.43 V). The results for the ranges of (a) 3200–2800 cm^–1, and (b) 1500–1000 cm^–1 are separately shown

As discussed above, the anionic moieties are adsorbed during the potential cycle, while its orientation changes depending on the potential. In a previous study,^[6] the [NTf₂] anion (CF₃SO₂NSO₂CF₃^–) was found to be less adsorptive on the surface of Pt than SO₄^2–, CF₃OCF₂CF₂SO₃^–, and CF₃CF₂SO₃^–. As the latter anions are desorbed outside the intermediate potential range of 0.4– 0.9 V,^[5] no desorption of the [NTf₂] anion throughout the potential cycle in the range of 0.08–1.33 V suggests that the cationic moieties inhibit the desorption of the anionic moieties. Hence, it is suggested that the cationic moieties are present on the Pt surface throughout the potential cycle. (The absence of the cation bands in the SEIRA spectra is probably owing to its distant location compared to the anion and smaller oscillating dipoles of the C–H vibrations.) A possible mechanism is that the hydrophobic cations inhibit the hydration of the anionic moieties.

From the CV profiles of the Pt single-crystal and the SEIRAS analyses of the Pt polycrystalline electrode, the effects of the IL on the properties of Pt catalyst can be outlined as follows. The IL molecules are preferentially adsorbed on the defect sites through the anionic moieties, and the adsorbed IL can affect the stability of adsorbed species on the terrace sites and enhance the ORR activity of the catalyst. In a previous study involving IL modification,^[18] where Pt nanocatalyst supported on high-surface-area carbon was used, ORR activity enhancement was observed and ascribed to the protection of the defect sites from oxidation, on the basis of CO stripping voltammetric analysis. In our analyses, the hypothesis of spontaneous defect protection by the IL molecules was validated using a stepped single-crystal electrode. Probably, the IL molecules observed in the SEIRAS measurements were also adsorbed at the defect sites on the Pt polycrystalline electrode. In addition, the dissolution of the IL [C₄C₁im][NTf₂] during ORR was demonstrated in our study, and further studies are necessary for mitigating IL dissolution and applying this concept to practical PEFCs subjected to long-term operation.

3.3 Mitigation of Nafion adsorption by IL molecules

The IL-modified Pt (443) surface was coated with a Nafion thin film. The dissolution of the IL during ORR may be inhibited by the sequential coatings and the concept of mitigating ionomer adsorption on Pt surface by the IL overlayer can be tested. Figure 10 shows the CVs for the Nafion-coated Pt (443) surface pre-modified with the IL before and after the ORR activity measurements. The results suggest that Nafion adsorption was mitigated by the IL modification as indicated by the disappearance of the anion adsorption peak in the double-layer region. Recently, Li et al.^[16] reported that ionomer adsorption on a Pt surface was mitigated by the coexistence of the IL of [MTBD][beti] in the ionomer film. In their experiment, however, the IL was mixed in the ionomer dispersion before the ionomer coating, and the mitigation effect was explained by physical exclusion or electrostatic screening effects by the IL existing in the ionomer layer. In contrast, in the present study, the IL and Nafion coatings were sequentially performed, and therefore, the IL possibly prevents the physical contact between the Pt surface and the ionomer. Actually, as shown in Figures 7 and 10, the CV profile with both the IL and Nafion coatings before the ORR activity measurement is essentially identical to the CV profile solely with the IL coating; therefore, the IL molecules should act as a separation layer between the Pt surface and ionomer film by adsorbing on the step sites. The results also indicate that the IL was dissolved during the ORR measurements, resulting in the loss of the mitigation effect, as indicated by the re-appearance of the anion adsorption peak. Thus, the expected inhibiting effect by the Nafion thin film on the IL dissolution during ORR mentioned above was not observed in this experiment. Consequently, the ORR activity was significantly lower for the IL-Nafion-coated Pt surface than for the bare Pt surface (Figure S6), indicating the adsorption of the ionomer on the IL-lost Pt surface. A schematic of the mitigation and re-occurrence of ionomer contact is shown in Figure 11. In the study by Li et al.,^[16] ORR activity enhancement via the mitigation of Nafion adsorption by IL molecules was observed for Pt (111). It is yet to be ascertained whether the loss of the mitigation effect by the IL in the present study is owing to the difference in the type of IL or the coating procedure from the previous study. In either case, the effectiveness of using ILs for mitigating ionomer adsorption and the challenges of retaining the IL on the surface of Pt during the ORR were demonstrated in the present study.

FIGURE 10

Cyclic voltammograms (CVs) for the Nafion-coated Pt (443) pre-modified with the IL before and after the oxygen reduction reaction (ORR) activity measurement. The CV for the bare Pt (443) is also shown

FIGURE 11

Schematic of the mitigation of ionomer contact by ionic liquid (IL) and re-contact due to IL dissolution

4 CONCLUSION

The effects of the coatings of ionomer and IL on the ORR activity of Pt surfaces were investigated through voltammetric analyses using stepped Pt single-crystal electrodes with (111) terraces and (110) steps as well as SEIRAS analyses using a Pt polycrystalline electrode. The sulfonate anions in the ionomer were adsorbed on the stepped Pt single-crystal electrodes, and the adsorption site was the (111) terraces for the surfaces with wide terrace widths (n ≧ 3) and the (110) steps for the surfaces with narrow terrace widths (n ≦ 3). The sulfonate adsorptions suppressed the ORR activity on the stepped Pt single-crystal electrodes by more than 50% (i.e., less than 50% of the Pt surface area was available for the ORR). The IL molecule of [C₄C₁im][NTf₂] was preferentially adsorbed on the step sites and this observation is consistent with the findings of previous studies. The SEIRAS analysis suggested that the IL molecules were adsorbed through the anionic moieties and that the cationic moieties played a role in preventing the desorption of the IL molecules. IL pre-modification was also found to mitigate ionomer adsorption on the Pt surface; therefore, this is a promising approach for improving the ORR activity of the catalyst. In the present study, however, the mitigation effect was not reflected in the ORR activity because of the desorption of the IL during the ORR activity measurement, probably owing to the water production. Therefore, countermeasures against the leaching of the IL during the long-term operation of PEFCs are necessary.

ACKNOWLEDGMENT

We thank Prof. Katsuyoshi Ikeda (Nagoya Institute of Technology) for allowing us to use the IR spectrometer.

CONFLICT OF INTEREST

The authors declare no conflict of interest.

Open Research

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.