Sol-Gel Synthesis of Ruthenium Oxide Nanowires To Enhance Methanol Oxidation in Supported Platinum Nanoparticle Catalysts.

A template-directed, sol-gel synthesis is utilized to produce crystalline RuO2 nanowires. Crystalline nanowires with a diameter of 128 ± 15 nm were synthesized after treating the nanowires at 600 °C in air. Analysis of these nanowires by X-ray powder diffraction revealed the major crystalline phase to be tetragonal RuO2 with a small quantity of metallic ruthenium present. Further analysis of the nanowire structures by high-resolution transmission electron microscopy reveals that they are polycrystalline and are composed of interconnected, highly crystalline, nanoparticles having an average size of ∼25 nm. Uniform 3 nm Pt nanoparticles were dispersed on the surface of RuO2 nanowires using an ambient, solution-based technique yielding a hybrid catalyst for methanol oxidation. Linear sweep voltammograms (LSVs) and chronoamperometry performed in the presence of methanol in an acidic electrolyte revealed a significant enhancement in the onset potential, mass activity, and long-term stability compared with analogous Pt nanoparticles supported on commercially available Vulcan XC-72R carbon nanoparticles. Formic acid oxidation LSVs and CO stripping voltammetry revealed that the RuO2-supported Pt nanoparticles exhibit significantly higher CO tolerance, which leads to higher catalytic stability over a period of several hours. X-ray photoelectron spectroscopy results suggest that crystalline RuO2 leads to less-significant oxidation of the Pt surface relative to more widely studied hydrous RuO2 supports, thereby increasing catalytic performance.

Introduction

The electrochemical oxidation of small organic molecules (SOMs) is a key electrochemical process and has broad technological applications in fuel cells, sensors, and catalysis.[1−5] Recently, most attention has been focused on the oxidation of alcohols such as methanol and ethanol, utilizing an ever-broadening array of Pt-based electrocatalysts.[4,5] Considerable increases in catalytic activity and stability have been achieved by tuning the size, composition, and morphology of Pt catalysts. While the catalyst is responsible for much of chemistry, the catalyst support can also play a crucial role in the mechanism of SOM oxidation.[6] Traditionally, platinum catalysts are supported on nanoparticulate carbon such as Vulcan XC-72R.[6−8] However, carbon supports are susceptible to oxidation and degradation during the SOM oxidation, which can lead to loss of Pt utilization and poor long-term durability of the catalyst.[9]

In addition to challenges with the stability of the catalyst support, the effective oxidation of SOMs on Pt requires high overpotentials because of catalyst poisoning effects.[10−12] For example, methanol oxidation (eq ) on Pt follows an indirect pathway that leads to the preferential formation of CO as a partially oxidized intermediate.[10] At low overpotentials, adsorbed methanol is rapidly converted to CO via a multi-step dehydrogenation process that occurs at Pt–Pt pair sites. Because CO oxidation requires significant overpotential, the kinetics of methanol oxidation is hindered at low overpotentials by the high coverage of CO. This effect is commonly referred to as “CO poisoning” and it is a key challenge with oxidizing a broad range of SOMs on Pt. One approach to overcome the challenge of CO poisoning is to alloy Pt with oxophilic metals such as Ru to promote a bifunctional mechanism.[13,14] In these alloys, the Ru sites adsorb oxide species at lower potentials than Pt and facilitate the oxidation of CO through a process referred to as CO spillover. Recent studies have also shown that the formation of hydrous ruthenium oxide from the oxidation of Pt–Ru alloy nanoparticles (NPs) may be an active phase that facilitates CO oxidation.[15−17]

In light of the challenges surrounding catalytic activity and stability, there is a growing interest in metal oxides as supports for Pt-based SOM catalysts.[18,19] Metal oxides such as TiO2 and RuO2 are stable over a wide pH range and are less susceptible to oxidation and degradation than traditional carbon-based supports.[20−32] More importantly, metal oxide supports have been shown to facilitate SOM oxidation on Pt NPs via several mechanisms. For example, oxides improve the three-dimensional dispersion of Pt NPs, leading to better Pt utilization.[20,32] Analogous to the bifunctional mechanism of Ru dopants, the surfaces of metal oxides can provide oxide species or facilitate the transport of hydroxide species to support CO spillover at the interface between the Pt particle and the oxide surface.[23,25,30] In addition, several reports have shown that oxide supports lead to significant changes of the electronic structure of Pt via the strong metal/support interaction (SMSI) effect.[6,22,27,32] The structural interaction between the oxide support and Pt catalysts leads to significant variations in the d-band vacancy of the Pt NP and can promote either reversible or irreversible oxidation of Pt, depending on the strength of the interaction. A higher degree of oxophilicity of the Pt catalyst is expected to further enhance CO tolerance.

Among the wide range of metal oxides, ruthenium oxide has a comparably high conductivity and its surface is hydrated in acidic media, leading to the presence of beneficial hydroxyl species.[18] Past work has focused primarily on hydrous ruthenium oxide, which is partially crystalline and has varying stoichiometry, depending upon the reaction conditions. Hydrous RuO2 leads to enhanced SOM oxidation performance of Pt but is less conductive than crystalline RuO2 and the strong SMSI effect leads to partial irreversible oxidation of the Pt surface.[6,22,27,32] Relatively few studies have investigated the support effect in well-defined, crystalline, RuO2 nanostructures.[18]

In this work, we synthesize crystalline, tetragonal RuO2 NWs with a uniform diameter of ∼130 nm using a template-directed sol–gel synthesis method. A solution-based method was employed to deposit a uniform dispersion of Pt NP on the surface of the RuO2 NWs. The methanol oxidation reaction (MOR) performance of the hybrid Pt NP/RuO2 NWs is compared with that of Pt NPs supported on commercially available Vulcan XC-72R carbon NPs (Pt NP/C). The RuO2 support leads to measurable enhancements in the onset potential, kinetics, Pt utilization, and long-term stability of MOR relative to the commercial Pt NP/C. A study of the key mechanistic intermediates including CO and formic acid reveals that the crystalline RuO2 supports lead to significant improvements in the CO tolerance of the Pt NPs supported on RuO2. In addition, X-ray photoelectron spectroscopy (XPS) results suggest that crystalline RuO2 leads to less-significant oxidation of Pt NPs relative to the results observed for hydrate RuO2 supports in the prior literature.

Results and Discussion

The synthesis of the RuO2 NWs was accomplished via a template-assisted, sol–gel technique based on previously developed methods for other metal oxides.[33,34] A novel two-step approach was developed to prepare uniform nanowires without the presence of excess bulk material that is typically formed on the surface of the template. Details regarding the development of the two-step process can be found in the Supporting Information section. In the first step, a polycarbonate filter membrane with a nominal pore diameter of 200 nm was impregnated with a RuCl3 sol by vacuum filtration. Once the ethanol solution was loaded into the template, the surface was polished to remove ruthenium residues, while leaving the precursor in the pores undisturbed. The gelation process was initiated by exposing the saturated template to propylene oxide vapors in a glass reactor. The color of the template immediately began to change from a brownish-red color to jet-black once exposed to the vapors of propylene oxide. This color change indicated that the gelation reaction between RuCl3 contained within the pores and the propylene oxide had occured. The reaction proceeded for several minutes to allow for the penetration of vapors into the template pores.

To convert the sol–gel into crystalline RuO2, the template was annealed at 600 °C for 30 min. This heating step also resulted in the vaporization of polycarbonate, thus alleviating the need for further processing. Scanning electron microscopy of the resulting powder (Figure ) revealed that the material consisted almost entirely of well-dispersed, one-dimensional, NW structures, with only a small fraction of the material consisting of NPs. The diameter of NWs was determined to be 128 ± 15 nm with lengths of up to 3 μm. The measured diameter is ∼70 nm smaller than the nominal diameter of the 200 nm pore dimensions. The smaller size of the wires relative to the template pore diameter can be explained by the removal of the organic species and densification of the structures during crystallization, which has been observed previously.[34]

Figure 1

(A) SEM image showing the 1D structure of the RuO2 nanomaterials with an average diameter of 128 ± 15 nm. (B) SEM image showing the RuO2 nanomaterials have various lengths up to ∼3 μm.

Powder XRD was performed to characterize the crystallinity and structure of the RuO2 nanowires. The diffraction pattern (Figure ) shows that the sample is composed primarily of tetragonal RuO2, with unit cell parameters of a = 4.49 Å and c = 3.11 Å. In prior reports, characterization of hydrous RuO2, RuO2·xH2O, by XRD yielded broad, undefined peaks consistent with the amorphous nature of the hydrated material.[35,36] Typically, temperatures of 400 °C or higher are necessary to fully crystallize RuO2 producing well-defined diffraction peaks. In this case, a heat treatment at 600 °C was employed to crystallize the RuO2 NWs, leading to the well-defined peaks observed in the XRD pattern. In addition to crystalline RuO2, there is also a small impurity of hexagonal Ru, which indicates that a small fraction of RuCl3 at the core of the wires was not exposed to the propylene oxide vapors. The presence of Ru may have key benefits in terms of catalysis because it can react with the Pt precursor during the Pt NP deposition step, leading to the formation of Pt–Ru alloy NPs.

Figure 2

Powder XRD pattern of the RuO2 nanomaterial showing that the major phase of the material is RuO2 (red lines, COD 9007541) with a small Ru impurity (blue lines, COD 9008513).

High-resolution transmission electron microscopy (HRTEM) was performed to examine the microstructure of the RuO2 NWs. The nanowire morphology is evident in Figure A and it was confirmed that the 1D structures were solid NWs and not hollow nanotubes. The size of the nanowires measured from the TEM images is 139 ± 19 nm, which is in agreement with the measurements from the scanning electron microscopy (SEM). It is evident from the TEM images that the nanowires are composed of interconnected NPs with a diameter of ∼25 nm. The polycrystalline nature of the NWs is consistent with the sol–gel synthesis because the gelling process forms a xerogel within the pores of the template. A similar motif was observed in Cr2O3 nanowires produced by a similar template-based sol–gel technique.[34] It is important to note that the nanowires produced have no surface coatings as a result of surfactants or residual template materials. This allows for direct contact between the catalyst NPs and the RuO2 surface, which is beneficial for their use as a catalyst support.

Figure 3

(A) TEM image of RuO2 nanomaterial, revealing the filled, polycrystalline nanowire structure. (B) TEM image of the Pt NPs dispersed onto the surface of the RuO2 nanowires. (C) HRTEM image of the RuO2 nanowires revealing the {110} plane of the polycrystalline grains.

The Pt NPs were loaded onto the surface of the RuO2 NWs via a solution-based reduction process using sodium borohydride as a reducing agent. This resulted in the production of well-dispersed Pt NPs on the surface of the NWs. HRTEM images after Pt NP deposition (Figure B) confirm the high degree of dispersion and reveal that the Pt NPs have an average size of 3.3 ± 0.7 nm (see Figure S2 for a histogram of particle size distribution). It can also be seen from high resolution images that the NPs dispersed on the surface of the nanowires are in intimate contact with RuO2. Additionally, the {010} planes of the RuO2 substrate are clearly visible, confirming that the RuO2 NWs comprise individual, highly crystalline NPs (Figure C).

XPS was performed (see the Supporting Information for additional details) to examine the electronic properties of the Pt catalyst and the support materials.[37,38] Survey spectra (Figure S3) collected from the Pt NP/RuO2 NW and Pt NP/C samples confirm the presence of metallic Pt on both substrates with a nominal loading of 20%. Fits of the high-resolution scans of the Ru peaks shown in Figure A reveal that the Ru 3d5/2, 3d3/2, and 3p3/2 peaks are located at 280.9, 285.1, and 463 eV, respectively, and can be assigned to the +4 oxidation state of ruthenium.[39] The sharp cutoff of the Ru 3d5/2 peak suggests that the metallic ruthenium observed in the XRD is relegated to the core of the material because it is not detectable by XPS. To examine the effect of the substrate on the electronic properties of the Pt catalyst, high-resolution scans were obtained of the Pt 4f peaks (Figure B). In both cases, the Pt 4f7/2 and Pt 4f5/2 peaks were located at 71.3 and 74.7 eV, respectively, consistent with metallic platinum.[40] The absence of a measureable shift in the Pt 4f peaks to higher binding energies suggests that the crystalline RuO2 support does not contribute to significant oxidation of the Pt surface. This result contrasts with those of hydrous ruthenium oxide supports, which were found to result in a shift in the peaks and measurable oxidation of the Pt surface.[22,26]

Figure 4

High-resolution XPS scans of the 3d and 3p (inset) Ru peaks (A) in the Pt NP/RuO2 NW catalyst. High-resolution XPS scans of the Pt 4f peaks (B) of the Pt NPs supported on the RuO2 NW and carbon supports.

Following the above analyses, the electrochemical properties of the supported Pt particles were examined in 0.1 M HClO4 by cyclic voltammetry (CV). Initially, the samples were cycled from 0 to 1.3 V to achieve a steady electrochemical state. No significant changes were observed in the hydrogen adsorption or oxide reduction regions of the CV for the Pt NPs supported on either the RuO2 NWs or the carbon NPs. The stability of the Pt NP/RuO2 NW CV is consistent with the high degree of crystallinity of the RuO2 support, which is less susceptible to structural reconfiguration under electrochemical conditions than partially crystalline, hydrous RuO2 supports.[20] In prior reports, the use of hydrous RuO2 supports typically leads to significant capacitive effects because of their amorphous structure.[6,25−27] In this case, the double layer capacitance is only slightly higher in the RuO2 NW support, when compared to the carbon support. This is attributed to the combined effects of the high crystallinity of RuO2 NW supports and the high conductivity of tetragonal RuO2.

The CVs of the Pt NP/RuO2 NW and Pt NP/C catalysts shown in Figure display the characteristic hydrogen adsorption/desorption region below 0.4 V. The reversible peak at 0.1 V is consistent with surface defect sites associated with spherical platinum nanostructures and suggests that both Pt catalysts have similar surface structures.[5,41] The electrochemical surface area (ESA) was determined from the integrated hydrogen adsorption/desorption charge, and both catalysts were found to have nearly identical ESAs (Table ) of approximately 50 m2·g–1. The similarities in the surface structure and surface area of the platinum highlights the support material as a key factor in determining catalytic performance.

Figure 5

Cyclic voltammograms of Pt NP/C (A) and Pt NP/RuO2 NWs (B) catalysts obtained in 0.1 M HClO4 at a scan rate of 20 mV·s–1.

Table 1

Electrochemical Data of Pt NPs as a Function of Support Material

catalyst	ESA, m·g–1	onset potentiala, mV	specific activityb, mA·cm–2	mass activityb, A·mg–1	mass activity at 2.5 hc, A·mg–1
Pt NP/C	53.8	681	0.26	0.14	0.11
Pt NP/RuO2 NW	49.6	652	0.43	0.23	0.15

Potential measured at J = 0.1 A·mg–1.

Current density measured at 0.7 V in LSVs.

Measured via chronoamperometry at 0.7 V.

Both catalysts also display reversible oxidation of the Pt surface at potentials above 0.6 V. The position of the surface oxide reduction peak of the Pt NP/RuO2 NW is located at 0.757 V, which is shifted by 41 mV to lower potentials when compared with the Pt NP/C. A similar shift is also observed in the onset for surface oxidation in the anodic sweep. These results collectively suggest that the interaction between the RuO2 support and Pt NPs leads to a stronger interaction with oxygen adsorbates. The formation of surface oxide species at lower potentials is beneficial in the oxidation of methanol because oxide species promote the oxidation of the adsorbed CO intermediate at lower potentials.

The linear sweep voltammograms (LSVs) depicted in Figure A reveal the catalytic performance of the supported Pt NPs toward MOR. Comparison of the LSVs indicates that RuO2 NW support contributes to a significant enhancement in catalytic activity over the entire onset region (Figure A inset) relative to the carbon support. As a control, LSVs were collected from RuO2 NW and carbon supports and no appreciable faradaic current was observed. Thus, methanol oxidation results from the presence of Pt. To quantify the effect of the support on catalytic performance of the Pt NPs, the onset potentials measured at a current density of 0.1 A·mg–1 and mass and specific activities at 0.7 V are shown in Table for the RuO2 NW and Pt NP/C. The onset for MOR is shifted by 30 mV to lower potentials in the Pt NP/RuO2 NW catalyst, suggesting a lower overpotential for methanol activation and oxidation.

Figure 6

Characterization of the electrocatalytic performance for methanol oxidation of supported Pt NPs. LSVs (A) normalized to Pt mass collected in 0.1 methanol at a scan rate of 20 mV·s–1. The inset highlights the onset region for methanol oxidation. Chronoamperograms (B) collected at 0.7 V of Pt mass activity as a function of time.

The mass activity of the Pt NP/RuO2 NWs was determined to be 0.23 A·mg–1, representing a 1.6-fold enhancement in activity compared with the Pt NP/C. A similar enhancement of 1.7-fold was also noted in the specific activity of the RuO2 NW catalysts (Figure S5). Collectively, these results suggest that the RuO2 support leads to significant improvements of the MOR activity at low overpotentials. Although it is difficult to compare results with those obtained in prior reports under different conditions, the Pt NP/RuO2 NW catalysts maintain comparable or better results. For example, Ahn and co-workers observed activities of 0.05 A·mg–1 for the Pt NP supported on crystalline Ru/RuO2 nanofiber supports at approximately 0.7 V with a methanol concentration of 2 M in sulfuric acid and with a scan rate of 50 mV·s–1.[30] In terms of hydrous ruthenium oxide supports, Fujishima and co-workers observed specific activities of 0.1 mA·cm–2 for hybrid Pt/hydrous ruthenium oxide nanostructures on conductive diamond at approximately 0.7 V in 0.1 M HClO4 containing 3.2 M methanol.[22]

To evaluate the long-term stability of the Pt catalysts and supports, chronoamperometry was performed at 0.7 V. The chronoamperograms in Figure B show that both catalysts undergo a brief induction period of 1–2 min where catalytic activity increases from the values measured in the LSV. This can be attributed to an activation of the Pt surface sites as a result of the surface oxidation that occurs at these potentials. Subsequently, the activity of both catalysts decreases, approaching steady-state values of 1.5 and 1.1 A·mg–1 for the Pt NP/RuO2 NW and Pt NP/C catalysts, respectively. Over the course of the test, the Pt NP/RuO2 NWs maintain a higher mass activity, confirming that the enhanced performance in the LSV is retained after several hours of operation. Moreover, the Pt NP/RuO2 NW catalysts evolved CO2 gas, which was visible in the first 10 min of the stability test. The significant bubble formation resulted in the saw-tooth pattern in the chronoamperogram. By contrast, there was far less-visible gas formation from the Pt NP/C electrode. This result suggests that the RuO2 NW support may lead to a higher faradaic efficiency in the oxidation of methanol to CO2, although additional experiments beyond the scope of this report would be necessary to confirm the quantity of gas produced by each catalyst.

We investigated the performance of the catalysts as a function of the support material toward the oxidation of formic acid (eq ) and carbon monoxide. These species are two key intermediates in the mechanism of methanol oxidation.[10,42] LSVs of formic acid oxidation are shown in Figure A. The Pt NP/RuO2 NW catalysts maintain higher overall specific activity over the entire formic acid oxidation window, which is consistent with the enhanced MOR performance. The formic acid oxidation LSVs for both catalysts is characterized by a peak in the region of 0.8–1.1 V with a corresponding shoulder at lower potentials. In terms of the mechanism, the peak-shoulder feature in the LSVs of both catalysts has been widely shown to be associated with catalysts that undergo an indirect mechanism for the oxidation of methanol.[43,44]

Figure 7

LSVs (A) collected in 0.1 M formic acid and CO-stripping voltammograms (B) collected in 0.1 M HClO4 at scan rates of 20 mV·s–1. Prior to CO stripping, electrodes were immersed in a CO-saturated 0.1 M HClO4 solution for a period of 45 min.

The indirect mechanism is expected for pure platinum and platinum-enriched alloy catalysts because of the high density of Pt active sites on the surface. In these catalysts, the dehydrogenation of methanol occurs at Pt–Pt pair sites, resulting in the rapid formation of an adsorbed CO intermediate at low overpotentials. The complete oxidation of CO to CO2 is delayed to higher potentials because CO oxidation requires the presence of adsorbed oxygen species, which do not form on Pt at potentials below 0.5–0.6 V. Thus, the high coverage of CO inhibits MOR activity and requires a high overpotential to achieve effective MOR kinetics.

Because CO is the primary, rate-limiting intermediate of the indirect pathway, the kinetics of methanol oxidation is highly dependent upon catalyst’s ability to oxidize CO. To determine the relative onset for CO activation, CO-stripping voltammograms (Figure B) were obtained after allowing the catalysts to fully adsorb CO from a saturated solution. The CO-stripping peak in the Pt NP/RuO2 NWs occurs at 734 mV, which is more than 75 mV lower than the onset for CO stripping in Pt NP/C. The onset for CO oxidation for both catalysts is closely aligned with the onset in the MOR LSVs, suggesting that CO oxidation limits the kinetics at low overpotentials. Thus, the enhanced MOR activity in the RuO2 NW supports can be attributed to improved CO oxidation performance induced by the RuO2 support. In prior reports, enhanced performance in RuO2-based supports was attributed to a variety of effects including improved catalyst dispersion,[16,24] improved transport of hydroxyl and proton species at the catalytic interface,[25] improved CO tolerance due to a bifunctional effect,[22,23,29,30] as well as the beneficial changes in the electronic structure of platinum from the SMSI effect.[25,32]

A common explanation for the enhanced performance in metal oxide-supported Pt is derived from the bifunctional mechanism of methanol oxidation in Pt1–Ru alloys. In PtRu alloys, the Ru surface sites are oxidized at much lower potentials forming Ru–OH surface species, which facilitate the oxidation of the CO intermediate formed on the Pt sites. Like metallic Ru, metal oxides also facilitate the formation of surface hydroxyl species at low overpotentials. However, unlike in uniform PtRu alloys, effective CO oxidation via the bifunctional mechanism can only take place at the Pt–RuO2 interface. The diffusion of CO from isolated Pt sites to the Pt–RuO2 interface should also be relatively slow because of the strong adsorption of CO on Pt. Moreover, RuO2 catalysts form relatively thick oxide layers on the surface in the potential window for MOR on Pt, which may also slow the transport of hydroxyl groups to Pt catalysts.[18] Thus, it is likely that the enhanced activity cannot solely be attributed to the bifunctional effect but is a combination of a bifunctional effect and electronic effect of the support on the Pt catalyst.

In terms of the electronic effect, several reports have shown that oxide supports lead to significant changes of the electronic structure of Pt via the SMSI effect.[20,25−27,32] For example, X-ray absorption spectra of Pt NPs supported on mixed Ti–Ru metal oxides showed significant variations in the d-band vacancy of the supported Pt NPs relative to pure Pt.[27,32] In the case of the widely used hydrous ruthenium oxide supports, XPS measurements have shown that the SMSI effect is strong and leads to partial, irreversible oxidation of the Pt particle surface.[22,26] Although this increases the available oxide species for the bifunctional oxidation of CO, the oxidation of the Pt particles leads to a reduction of the available Pt active sites, potentially lowering the mass activity of the catalyst.

Interestingly, we did not observe significant irreversible oxidation of the Pt NPs supported on the crystalline RuO2 support via XPS, which is consistent with recent studies.[30] From our CV results, it appears that the crystalline RuO2 has a less significant effect on Pt, leading to a small but measurable increase in the oxophilicity of the Pt surface sites indicated by 41 mV shift in the oxide peak. Similar shifts in the oxide region were observed previously.[6] The stronger interaction with oxide species and the lower onset for surface oxidation can facilitate CO oxidation on the Pt particle itself as well as at the Pt–RuO2 interface. Thus, the advantages of crystalline RuO2 may extend beyond improved stability and conductivity and may also contribute to improvements in the mechanism of MOR relative to hydrous ruthenium oxide.

Conclusions

The oxidation of SOMs has broad applications in fuel cells, catalysis, and sensors. However, the practical development of catalysts for these reactions is hindered by the poisoning effects of partially oxidized carbon species. Traditional strategies for improving SOM oxidation on carbon-supported Pt-based catalysts have focused largely on tuning the properties of the Pt catalyst itself. However, tuning the physicochemical properties of the support material represents a second and equally important pathway for reducing the effects of poisoning in SOM oxidation catalysts. In this report, we explore crystalline RuO2 as a support material for platinum catalysts considering its high conductivity and beneficial surface properties.

We employ a template-assisted, sol–gel method to produce crystalline RuO2 nanowires and subsequently decorate their surfaces with small Pt NPs (∼3 nm). Electrochemical investigations of the RuO2-supported Pt NPs revealed that the crystalline Pt NPs supported on RuO2 NWs required a lower overpotential to oxidize methanol and had better long-term MOR activity than Pt NPs supported on traditional carbon supports. XPS results suggest that crystalline RuO2 leads to less significant oxidation of the Pt surface relative to more widely studied hydrous RuO2 supports, thereby increasing catalytic performance. From a broader perspective, our results provide further evidence that support materials can not only contribute to new active sites for SOM oxidation but also influence the nature of the catalyst itself through strong interactions between the catalyst and its support. Thus, rational approaches to the catalyst design should include tuning the structural and electronic properties of both the catalyst and the support material.

Experimental Section

The RuO2 nanowires were synthesized using a modified sol–gel method presented by Walker et al.[33] Briefly, 0.42 g of RuCl3·xH2O was dissolved in 3.5 mL of 200 proof ethanol. The solution was stirred using a magnetic stir bar for 1–2 h to ensure complete dissolution of the RuCl3. Once a homogenous solution was obtained, the solution was filtered through a 200 nm polycarbonate template (Whatman, Nuclepore track etch) using vacuum filtration and a glass frit support. Approximately 50 drops were distributed over the template with applied vacuum to load the pores of the template with RuCl3. Following this loading procedure, the template was polished using an Arkansas whetstone to remove the excess Ru precursor on the surface of the template.

The loaded template was subsequently placed in a glass reactor containing 2 mL of propylene oxide and was heated to 65 °C. Care was taken to keep the template sequestered from the liquid propylene oxide to prevent dissolution of the template. The Ru precursor in the template reacted with the propylene oxide vapors for a period of 5 min. The formation of the hydrous RuO2 gel was indicated by change in the color of the template from a red-brown color to black. After treatment with propylene oxide, the template was placed into a porcelain crucible and heated to 600 °C in a muffle furnace for 30 min. Following this heating procedure, the crucible was removed from the furnace and allowed to cool to room temperature. The samples were then collected for further analysis.

To deposit small platinum NPs onto the RuO2 support, as-synthesized RuO2 NWs (∼2 mg) were dispersed into 1.25 mL of deionized water by sonication for 5 min. The solution containing the RuO2 NWs was combined with a 1 mL aliquot of a solution containing dihydrogen hexachloroplatinic acid (H2PtCl6, Alfa Aesar, 99+%). The concentration of the H2PtCl6 solution was prepared to yield 20% by mass Pt in the resulting Pt/RuO2 NWs. The solution containing both RuO2 and H2PtCl6 was stirred for 5 min before the addition of 1 mL of an aqueous NaBH4 solution (Alfa Aesar, 98%, 3 mg/mL). The deposition proceeded for a period of 30 min before the product was collected by centrifugation and washed with water three times. The same procedure was performed to deposit Pt NPs on commercial Vulcan XC-72R carbon NPs. Details of characterization and electrochemical methods can be found in the Supporting Information section.