Study of Porosity-Dependent Oxygen Reduction at Porous Platinum Tips Using Scanning Electrochemical Microscopy

Platinum wire (≥ 99.9%, 76 μm diameter), Kimble microcapillary pipettes, sodium chloride (≥ 99.5%), Triton X-100, hexaammineruthenium(III) chloride ([Ru(NH₃)₆]Cl₃, 98%), potassium sulfate (powder, 99+%, A.C.S. reagent), and sulfuric acid (95.0 – 98.0%) were all from Sigma-Aldrich (St. Louis, MO). Pt wire (10-μm diameter) was from Nilaco (The Nilaco Corporation, Tokyo Ginza). Carbon paste from BASi was used for electrochemical connection between an end of platinum wire and a copper wire (Hook-up wire, 28 AWG, Solid PTFE, Spool Green) from AlphaWire. Dihydrogen hexachloroplatinate(IV) hydrate (H₂PtCl₆·xH₂O, Premion, 99.999%) was purchased from Alfa Aesar (Ward Hill, MA). Lead(II) acetate trihydrate (≥ 99.5%) was supplied by Fluka (Steinheim, Germany). All solutions were prepared with 18 MΩ•cm deionized water.

Electrochemical deposition was performed using an electrochemical analyzer (CHI 730D) and a refrigerating /heating bath circulator (RW-0525G, JEIO TECH). For all deposition processes, a Pt wire and a saturated calomel electrode (SCE) were used as the counter electrode and the reference electrode, respectively. Electrochemical measurements using SECM were performed in a four-electrode setup using an electrochemical analyzer (CHI 920C SECM) with a gold wire and Hg/Hg₂SO₄ electrode as the counter electrode and the reference electrode, respectively. Commercial SECM cell was utilized. Each porous Pt structure electroplated on a Pt foil (3 × 3 mm², 0.025 mm thick) were characterized by field emission scanning electron microscopy (FE-SEM, JEOL JSM-6700F) in order to figure out their structures and pore sizes. For understanding the distances among Pt nanoparticles, porous Pt structures deposited on a TEM gold grid (300 mesh, TED PELLA, INC.) were used to obtain FE-TEM images.

Tip and substrate electrodes were prepared by sealing Pt microwires (10 μm or 76 μm in diameter) under vacuum in glass capillaries (Kimble) followed by vertical polishing to expose the Pt microdisks at the end plane, as previously reported.³³ The prepared 10-μm diameter Pt disk ultramicroelectrode (UME) with RG = 6–8; and 76-μm diameter Pt disk electrode were utilized as the tip and substrate electrodes, respectively. RG is the ratio of the overall tip electrode radius including a glass sheath to the Pt disk radius. The Pt tip electrode was immersed in 1.2 M CaCl₂ dissolved in acetone/water mixture (1:2 in volumetric ratio) and then the Pt disk was etched by applying 5 V ac voltage (60 Hz) with respect to 0.5 mm diameter Pt wire counter electrode for 4 s using variable autotransformer (Staco Energy Products Co., Daton, OH).

Pt layers with three different degrees of porosity (npPt-I, npPt-II, or npPt-III) were electrochemically deposited on the Pt surfaces within the recessed pore of the etched Pt tip electrodes with amperometry or cyclic voltammetry (CV) using Pt wire counter electrode and saturated calomel reference electrode (SCE). First, npPt-I was electroplated in a deposition solution (denoted as Soln A) containing H₂PtCl₆, 0.3 M NaCl aqueous solution, and Triton X-100 (5 : 45 : 50 in wt%) by applying −0.21 V for 1300 (±100) s at 40°C.³⁴ For npPt-III, Soln B was used as the deposition solution composed of 1 g of Soln A and 220 μL of 10 mM Pb(CH₃CO₂)₂·3H₂O aqueous solution. In fact, npPt-III was electrodeposited by CV from +0.6 to −0.25 V for 2 (±1) cycles with scan rate of 20 mV s⁻¹ at 40°C. For the deposition of npPt-II, the deposition solution (denoted as Soln C) was slightly varied from Soln B in terms of H₂PtCl₆ : 0.3 M NaCl solution : Triton X-100 ratio (i.e., 1.5 : 45 : 53.5 in wt% was used instead of 5 : 45 : 50). npPt-II was electroplated by CV (from +0.6 to −0.25 V for 7 (±1) cycles with scan rate of 20 mV s⁻¹ at 40°C.

In all deposition processes, the deposition solutions were heated to 60°C in order to activate components inside of the mixture followed by being cooled down to 40°C before every usage; and the deposition time or the potential cycle numbers were adjusted in order to fill the etched/recessed end pores of the Pt tip electrodes with porous Pt structures completely, which resulted in an inlaid porous disk tip geometry. To extract and remove Triton X-100 from porous structures, the electrodes were stirred in deionized water that was changed with fresh water every 3 h for 4 times.

Prior to SECM measurements, all electrodes were electrochemically cleaned by potential cycling between +0.7 V and −0.7 V (vs. Hg/Hg₂SO₄) in 1 M H₂SO₄ aqueous solution. Real surface areas (RSAs) of Pt disk, npPt-I, npPt-II, and npPt-III were measured in 1 M H₂SO₄ solution with CV at a scan rate of 2 mV s⁻¹. The integrated area of a hydrogen desorption region in a given CV result was divided by a conversion factor of Pt (210 μC cm⁻²).³⁵ For SECM measurements, a Pt disk substrate electrode (76 μm in diameter) was surrounded by heat shrinkable tube (Hanmicable, Busan, Korea) in order to be fixed at the bottom of a SECM cell. The tip electrode was located at a 10-μm distance above the center of the substrate electrode in N₂-saturated 10 mM [Ru(NH₃)₆]Cl₃ dissolved in 0.1 M K₂SO₄ aqueous solution by recording a SECM approach curve. In this process, the tip and substrate potentials were held at −0.9 V and −0.3 V, respectively. After that, collection efficiency test and ORR at this distance were carried out in TG-SC mode. Collection efficiency was obtained before and after measuring ORR by CV in order to confirm that H₂O₂ was successfully detected during ORR. The tip potential was scanned from −0.3 V to −0.9 V at a scan rate 10 mV s⁻¹ while a substrate potential was held at −0.3 V to re-oxidize [Ru(NH₃)₆]²⁺ generated at the tip. ORR was performed in O₂-saturated 0.1 M H₂SO₄ solution by linear sweep voltammetry (LSV) at a scan rate 2 mV s⁻¹ while the substrate potential was constantly held at +0.56 V in order to detect H₂O₂. The substrate applied potential was chosen as a peak potential of LSV for H₂O₂ oxidation in 2 mM H₂O₂ dissolved in 0.1 M H₂SO₄ at the 76 μm Pt substrate electrode at a scan rate of 2 mV s⁻¹.

Each porous Pt film was electrodeposited within a recessed Pt platform (depth of 4.8 ± 0.3 μm, n = 12) formed by electrochemical etching of Pt disk (10 μm in diameter) surrounded with glass insulator sheath. Therefore, the prepared electrodes could have the same projected geometric surface area (GSA) with 10-μm diameter and were used as SECM tip electrodes to study ORR in TG-SC mode.

The TEM images of as-prepared nanoporous Pt structures (npPt-I, npPt-II, and npPt-III) showed the morphological similarity at nanoscopic level that all these npPt films composed of tiny nanoparticles (2–4 nm in diameter) with ca. 2-nm interstices (Figures 1d-1f). On the other hand, npPt-II and npPt-III exhibited to be rough indicating a certain degrees of porosity even at microscopic level while npPt-I was smooth and flat according to the corresponding SEM images (Figures 1a-1c). Particularly, in npPt-II and npPt-III, Pt nanoparticles gathered to make the large clusters among which inter-cluster spaces formed large pores within the npPt film layers. The sizes of these relatively large pores present in npPt-II and npPt-III were estimated based on the SEM image analysis. The pore size distribution (Figure 2) indicated that npPt-III possessed larger pores (67.2 ± 47.7 nm with the most frequent value of 32.79 nm) than npPt-II (13.2 ± 6.8 nm with the most frequent value of 8.3 nm). This implies that various Pt films with different porosities can be synthesized readily by electrodeposition with a simple modification of electroplating solution composition ratios without the need of different types of templates. The regulated different porosities of the Pt films are seemingly attributed to the altered deposition speed in each deposition solution. In fact, for npPt-III deposition, a tiny amount of lead acetate (common seeding material for Pt plating) was added to the npPt-I plating solution. Thus, the Pt deposition was possibly promoted due to the presence of this seeding material, producing the deposited layer with the largest microscale pores. Meanwhile, the deposition solution for npPt-II also contained lead acetate but its Pt precursor concentration was lower than the other two (i.e., 1.5 wt% for npPt-II vs. 5.0 wt% for npPt-I and npPt-III). The intermediate deposition rate is expected at npPt-II because the seeding lead acetate accelerates but the lower Pt precursor concentration impedes the deposition. Accordingly, the intermediate porosity in between npPt-I and npPt-III was obtained at npPt-II, possessing smaller microscale pores than npPt-III in addition to nanoscale pores.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

EDS analysis for npPt-II and npPt-III at 20 different sites (Table I) shows that Pb in these porous Pt films was found at only negligible level (≤ 0.1 atomic %) because the concentration of lead acetate in the deposition solutions was very low.

Figure 3a presents CV results in 1.0 M H₂SO₄ solution obtained with the various tip electrodes: Pt disk, npPt-I, npPt-II, and npPt-III. The RSAs of the electrodes were determined from the integrated charge values passed for the hydrogen desorption (appeared between −0.3 V and −0.65 V) on Pt surface using a conversion factor of 210 μC cm⁻².³⁵ It was reported that RSAs determined using hydrogen adsorption/desorption were decreased as the potential scan rate increased due to incomplete hydrogen adsorption.³⁶ Thus, a slow scan rate (2 mV s⁻¹) was used to determine RSAs to ensure the complete hydrogen adsorption/desorption on the entire porous Pt surface of the Pt film layers in current study. The estimated RSAs were the largest for npPt-I (71000 ± 6000 μm,² n = 4), followed by npPt-II (45000 ± 12000 μm,² n = 4), and then by npPt-III (30000 ± 11000 μm,² n = 4). Particularly, the differences for npPt-I vs. npPt-II and npPt-I vs. npPt-III were statistically significant (paired t-test, p < 0.05) as shown in Figure 3b. Since the porous films filled the recessed spaces of a similar volume, the smaller interstices among Pt nanoparticles of npPt, the larger RSA it has. In fact, the determined roughness factor (RF = RSA/GSA) showed the same tendency of the RSA order: RFs are 904 ± 76 for npPt-I > 573 ± 152 for npPt-II > 382 ± 140 for npPt-III, n = 4 samples). The RSA of a planar Pt disk electrode was 275 ± 18 μm² (n = 3) which shows a roughness factor 3.5. This value is slightly larger than a typical roughness factor of polished Pt electrodes, 2–3,³⁷ however, the roughness increases with decreasing radius of electrodes.³⁸

Zoom In Zoom Out Reset image size

Collection efficiency of a designed TG-SC mode of SECM experiment was determined using a reversible redox couple, Ru(NH₃)₆^2+/3+, at 10 μm of tip to substrate separation that was also used for the SECM study of ORR (vide infra). The applied potential to a tip electrode was scanned from −0.3 to −0.9 V (vs. Hg/Hg₂SO₄) in an aqueous solution containing 10 mM [Ru(NH₃)₆]Cl₃ and 0.1 M K₂SO₄, while the substrate potential was fixed at −0.3 V. Nearly 100% of collection efficiency (99.6 ± 0.5%, n = 9) was obtained for three different tip electrodes (npPt-I, II, and III) by comparing the diffusion-controlled steady-state current of Ru(NH₃)₆³⁺ reduction at the tip and that of the produced Ru(NH₃)₆²⁺ re-oxidation at the substrate. This confirms that 76-μm diameter substrate is large enough to collect electrochemical product generated from 10-μm diameter tip electrode at their 10-μm separation distance. Collection efficiency experimental results obtained for Pt disk, npPt-I, npPt-II, and npPt-III electrodes are shown in Figure 4, where almost identical CV curves for Ru(NH₃)₆^3+/2+ redox reactions were also confirmed for these four different tip electrodes, indicating their GSAs are very similar.

Zoom In Zoom Out Reset image size

Voltammetric curves of ORR at four different tip electrodes (bare Pt disk, npPt-I, npPt-II, and npPt-III) were obtained while the substrate electrode potential was fixed to +0.56 V vs. Hg/Hg₂SO₄ to collect any possible production of H₂O₂ from the tip electrodes. The potential of +0.56 V applied to the substrate was sufficiently positive to oxidize H₂O₂ to O₂, which was confirmed by LSV curve obtained in a solution containing 2 mM H₂O₂ and 0.1 M H₂SO₄. (Figure 5). Most SECM studies on ORR to detect H₂O₂ have used SG-TC mode rather than TG-SC mode to circumvent possible large capacitive current of large substrate electrode. This possible problem was excluded using relatively small 76 μm substrate electrode as seen in Figure 4.

Zoom In Zoom Out Reset image size

Figure 6 shows the experimental results of ORR obtained with SECM in TG-SC mode. Unlike the collection efficiency experiments carried out with Ru(NH₃)₆^2+/3+, ORR current responses varied at different porous Pt tip structures. This implies ORR is a sluggish reaction and a different portion of porous Pt films is utilized at different porous structures deposited within recessed pores of tips. In Figure 6a, the apparent ORR onset potentials were observed to be dependent on the types of tip electrodes. In fact, npPt-I showed the most positive onset potential followed by npPt-II and then by npPt-III while bare Pt disk exhibited the most negative onset potential with a considerable difference from the other three npPts. This observed tendency of the ORR onset potential is possibly attributed to confinement effect.³ In the kinetic-controlled region including onset potential, the residence time of a reactant, O₂, at/near the electrode surface is a critical factor governing the ORR efficiency.³⁹ Therefore, nanoscale pores (∼2-nm interstices) of npPt-I provides more chances for capturing O₂ near the catalyst surface, resulting in a more positive ORR onset potential. The co-presence of nanoscale- and microscale-pores in npPt-II and npPt-III caused the less positive onset potential by less effective confinement effect. Predicted negligible confinement effect of bare-Pt disk leaded to the highly negative ORR onset potential.

Zoom In Zoom Out Reset image size

As the applied potentials became more negative from the onset potentials, ORR currents increased and eventually reached to the mass-transfer controlled steady-state levels for all the four different tip electrodes. The magnitudes of ORR limiting currents had the order of npPt-III > npPt-I ∼npPt-II ≫ bare Pt. The greatest ORR current of npPt-III could be explained with two factors: (1) utilization of greater electrode area or deeper section of the recessed tip and (2) higher n value. To understand the effect of electrode area, ORR curves were normalized to RSAs of the corresponding tip electrodes (Figure 6b). The obtained tip current density (J_tip,RSA, ORR current per unit RSA) decreased in the order of npPt-III > npPt-II > npPt-I for the porous Pt series. This implies that the greatest portion of the overall Pt RSA can be utilized at the Pt film possessing both micro- and nano-porosity, that is at npPt-III. In contrast, relatively smaller portion of the overall RSA of npPt-I may participate in electron transfer resulting lower current density with respect to the overall RSA. In particular, npPt-III with larger microscale pores uses its surface more efficiently for ORR, allowing facilitated O₂ influx within the film via the microscale-pores (vide infra, Figure 8).

Zoom In Zoom Out Reset image size

The surface area utilized during ORR (i.e. the used depth of the recessed tip) seems to influence the n value (Figure 7). The effect of Pt film porosity on the n values could be interpreted by inspecting the substrate current responding to H₂O₂ re-oxidation along with the corresponding tip current of Figure 6a. In fact, the n values of the various tip electrodes were calculated using Eq. 3.²³

where i_tip is a tip current, i_subs is a substrate current, and CE is a collection efficiency which is 100% in current TG-SC mode.

Zoom In Zoom Out Reset image size

The calculated n values as a function of potential is shown in Figure 7a. A bare Pt micro-disk has significantly low n value, down to ∼3.2 at −0.4 V. This low n value could be ascribed to the fast diffusion rate of generated intermediate H₂O₂ at a microelectrode as previously reported.⁴⁰ The n values at the nanoporous Pt series were nearly 4. A closer investigation of the data reveals that n values of npPt-I and npPt-II decrease slightly as the potential becomes more negative while npPt-III maintains high n value, ∼4, quite constantly for the examined potential range. This indicates the most efficient ORR at npPt-III with the smallest generation of 2 e-transfer product, H₂O₂. The slightly decreased n values of npPt-I and npPt-II in a more negative potential region are possibly ascribed to the following reasons: The influx of ORR reactant, O₂, is limited within a rather shallow depth of the Pt films due to the presence of only small pores, and therefore, the 2 e-transfer product can be swept away from the electrode surface prior to further reduction because of too high mass transport rate in this negative potential range (i.e., too much overpotential paid). Further study is, however, required for a better understanding of this observation.

This n value can be evaluated in a different way by counting percentage of generated H₂O₂ during ORR using the Equation 4¹⁹ as shown in Figure 7b.

In a good agreement with the results of n value, ORR at bare Pt disk UME generates the largest amount of H₂O₂. This amount decreases in the order of npPt-I > npPt-II > npPt-III, suggesting the highest ORR activity of npPt-III.

Figure 8 illustrates the predicted mass transport behaviors of O₂ and H₂O₂ for npPt-I and npPt-III. Due to the presence of only small pores, the influx of O₂ is restricted through the Pt film and only the shallow Pt layer is utilized for ORR,⁴¹ while the generated H₂O₂ can easily escape from the npPt-I structure. In case of npPt-III, however, larger microscale pores facilitate the influx of O₂,⁵ and the Pt films down to much deeper layers can participate in ORR. In the meantime, the generated H₂O₂ has a greater chance to be further reduced to water in npPt-III structure, showing higher ORR limiting current, lower H₂O₂ generation, and greater n value close to 4.⁴²

ORR activity of nanoporous Pt series is summarized as follows. In the kinetic controlled region, the confinement effect is a critical factor for ORR and therefore, npPt-I having only nanoscale pores with the largest RSA shows the most positive apparent onset potential. On the other hand, in the mass-transport controlled region, the actual electrode surface area utilized for ORR (i.e., travel depth of O₂ through catalyst film) is important, and therefore npPt-III facilitating O₂ influx exhibits the highest ORR limiting current with the smallest generation of 2 e-transfer product.