**3. Results and Discussion**

Figure 3 shows the FE-SEM images of the surface and cross-section of the PC membrane etched for 30 min in a 4.0 mol/dm<sup>3</sup> aqueous NaOH solution. The membranes were found to have uniform conical pores. The surface diameter of the pores was roughly estimated to be 540 nm by taking the average of the neighboring 50 pores. On the other hand, the depth seemed scattered, which may have been because, in the sample preparation for our cross-sectional SEM observations, the membrane was not necessarily cracked exactly in the middle of the pore. Thus, we determined the maximum depth at which the resulting cut plane was assumed to pass through the close vicinity of the conical vertex. Figure 3b presents the image of the best cut plane for the four pores, where the diameter of the pore opening was almost the same as the diameter measured on the surface (Figure 3a). Thus, the tip would have been included. This resulted in a depth of 3.8 µm. *Quantum Beam Sci.* **2021**, *5*, x FOR PEER REVIEW 6 of 13

**Figure 3.** FE-SEM images of (**a**) the surface and (**b**) the cross-section of the PC membrane etched for 30 min in a 4.0 mol/dm<sup>3</sup>NaOH solution at 60 °C. **Figure 3.** FE-SEM images of (**a**) the surface and (**b**) the cross-section of the PC membrane etched for 30 min in a 4.0 mol/dm<sup>3</sup> NaOH solution at 60 ◦C.

On the other side of the membrane, no pores were observed, confirming that the bombarded ions did not penetrate through the PC film, as calculated by the SRIM code. A previously proposed method for the fabrication of conically shaped pores that do not propagate through the entire thickness, that is, non-penetrating conical pores, involves On the other side of the membrane, no pores were observed, confirming that the bombarded ions did not penetrate through the PC film, as calculated by the SRIM code. A previously proposed method for the fabrication of conically shaped pores that do not propagate through the entire thickness, that is, non-penetrating conical pores, involves

and its stopping before the etchant breaks through to the opposite side [13]. In this method, a compression-sealed two-component cell must be employed to avoid the contact of another side with the etchant during the one-side etching [16]. However, we obtained non-penetrating conical pores even without such a dedicated cell and special care for the etching. This is because the tracks in the PC film did not propagate through its entire thickness. The immersion of the irradiated films in the etchant resulted in track-etched

Non-penetrating conical pores were prepared by varying the etching time and etchant concentration. Figure 4 plots the pore diameter and depth as a function of the etching time for different concentrations of the NaOH solutions. In Figure 4a, the surface diameter is observed to be uniform with rather small error bars under all of the etching conditions. As mentioned above, the maximum depth was taken for at least 20 pores; thus, Figure 4b plots the depth with no error bars. The diameter and depth of the pores linearly increase with the etching time; therefore, the slope of the plots corresponding to the growth rate of

membranes with non-penetrating conical pores.

the chemical etching of the penetrating tracks only from one side of the irradiated film and its stopping before the etchant breaks through to the opposite side [13]. In this method, a compression-sealed two-component cell must be employed to avoid the contact of another side with the etchant during the one-side etching [16]. However, we obtained non-penetrating conical pores even without such a dedicated cell and special care for the etching. This is because the tracks in the PC film did not propagate through its entire thickness. The immersion of the irradiated films in the etchant resulted in track-etched membranes with non-penetrating conical pores.

Non-penetrating conical pores were prepared by varying the etching time and etchant concentration. Figure 4 plots the pore diameter and depth as a function of the etching time for different concentrations of the NaOH solutions. In Figure 4a, the surface diameter is observed to be uniform with rather small error bars under all of the etching conditions. As mentioned above, the maximum depth was taken for at least 20 pores; thus, Figure 4b plots the depth with no error bars. The diameter and depth of the pores linearly increase with the etching time; therefore, the slope of the plots corresponding to the growth rate of the pores was estimated by least-squares regression. At concentrations of 2.0, 4.0 and 6.0 mol/dm<sup>3</sup> , the pore diameter was enhanced at 16, 26 and 50 nm/min, while the pore depth increased more significantly at 60, 130 and 660 nm/min, respectively. Consequently, the pore diameter and depth were controlled by a combination of the etching time and the NaOH concentration. *Quantum Beam Sci.* **2021**, *5*, x FOR PEER REVIEW 7 of 13 the pores was estimated by least-squares regression. At concentrations of 2.0, 4.0 and 6.0 mol/dm<sup>3</sup> , the pore diameter was enhanced at 16, 26 and 50 nm/min, while the pore depth increased more significantly at 60, 130 and 660 nm/min, respectively. Consequently, the pore diameter and depth were controlled by a combination of the etching time and the NaOH concentration.

**Figure 4.** (**a**) Pore diameter and (**b**) depth as a function of the etching time in an aqueous NaOH solution with different concentrations at 60 °C. **Figure 4.** (**a**) Pore diameter and (**b**) depth as a function of the etching time in an aqueous NaOH solution with different concentrations at 60 ◦C.

Figure 5 depicts the aspect ratio (AR) [17] and the cone angle, α, [18] of the conical

between two generatrix lines, respectively. Based on the assumption of a perfect conical geometry, the α values were calculated by the relationship α = 2tan−1(1/AR). For example, these were estimated to be 6.8 and 8.4°, respectively, for the pores shown in Figure 3. For all of the etching times, we observed their average values change from 4.3 to 12.2 and from 13.3° to 4.7°, respectively, with the increase of the concentration of the aqueous NaOH

Figure 5 depicts the aspect ratio (AR) [17] and the cone angle, α, [18] of the conical pores, which are defined as the ratio of the depth to the surface diameter and the angle between two generatrix lines, respectively. Based on the assumption of a perfect conical geometry, the α values were calculated by the relationship α = 2tan−<sup>1</sup> (1/AR). For example, these were estimated to be 6.8 and 8.4◦ , respectively, for the pores shown in Figure 3. For all of the etching times, we observed their average values change from 4.3 to 12.2 and from 13.3◦ to 4.7◦ , respectively, with the increase of the concentration of the aqueous NaOH solution from 2.0 to 6.0 mol/dm<sup>3</sup> . This concentration dependence can be rationalized by considering that the etch rate in the depth direction was more sensitive to the NaOH concentration than that in the transverse direction, as discussed earlier. More importantly, the AR and α, as well as the diameter and depth of the conical pores, were controlled by adjusting the etchant concentration. *Quantum Beam Sci.* **2021**, *5*, x FOR PEER REVIEW 8 of 13 solution from 2.0 to 6.0 mol/dm<sup>3</sup> . This concentration dependence can be rationalized by considering that the etch rate in the depth direction was more sensitive to the NaOH concentration than that in the transverse direction, as discussed earlier. More importantly, the AR and α, as well as the diameter and depth of the conical pores, were controlled by adjusting the etchant concentration.

**Figure 5.** Dependence of (**a**) the AR and (**b**) α of the conical pores on the concentration of the aqueous NaOH solution. Both of the values were calculated using the data in Figure 4a,b, which were obtained from the experiments at all of the etching times. **Figure 5.** Dependence of (**a**) the AR and (**b**) α of the conical pores on the concentration of the aqueous NaOH solution. Both of the values were calculated using the data in Figure 4a,b, which were obtained from the experiments at all of the etching times.

Track-etched membranes with non-penetrating conical pores were used for the fabrication of the platinum cones. Figure 6a–c depicts the FE-SEM images of the platinum cone arrays obtained using the template membranes etched for 30 min in the 2.0, 4.0 and 6.0 mol/dm<sup>3</sup> NaOH solutions, respectively. The platinum cones shown in Figure 6b have base diameters and lengths of 540 nm and 3.7 μm, respectively, both of which agree well with those of the template shown in Figure 1. The same is the case for the platinum cones shown in Figure 6a. In contrast, we obtained truncated platinum cones, as shown in Figure 6c, using the conical pores with the highest aspect ratio as a template. The vapor-deposited platinum cathode could not reach the bottom of the pores, which is likely because the Track-etched membranes with non-penetrating conical pores were used for the fabrication of the platinum cones. Figure 6a–c depicts the FE-SEM images of the platinum cone arrays obtained using the template membranes etched for 30 min in the 2.0, 4.0 and 6.0 mol/dm<sup>3</sup> NaOH solutions, respectively. The platinum cones shown in Figure 6b have base diameters and lengths of 540 nm and 3.7 µm, respectively, both of which agree well with those of the template shown in Figure 1. The same is the case for the platinum cones shown in Figure 6a. In contrast, we obtained truncated platinum cones, as shown in Figure 6c, using the conical pores with the highest aspect ratio as a template. The vapordeposited platinum cathode could not reach the bottom of the pores, which is likely because the platinum atoms and ions sputtered from the target arrived inside the conical pores from random directions [19,20].

platinum atoms and ions sputtered from the target arrived inside the conical pores from

random directions [19,20].

random directions [19,20].

tained from the experiments at all of the etching times.

solution from 2.0 to 6.0 mol/dm<sup>3</sup>

justing the etchant concentration.

. This concentration dependence can be rationalized by

considering that the etch rate in the depth direction was more sensitive to the NaOH concentration than that in the transverse direction, as discussed earlier. More importantly, the AR and α, as well as the diameter and depth of the conical pores, were controlled by ad-

**Figure 5.** Dependence of (**a**) the AR and (**b**) α of the conical pores on the concentration of the aqueous NaOH solution. Both of the values were calculated using the data in Figure 4a,b, which were ob-

Track-etched membranes with non-penetrating conical pores were used for the fab-

rication of the platinum cones. Figure 6a–c depicts the FE-SEM images of the platinum cone arrays obtained using the template membranes etched for 30 min in the 2.0, 4.0 and 6.0 mol/dm<sup>3</sup> NaOH solutions, respectively. The platinum cones shown in Figure 6b have base diameters and lengths of 540 nm and 3.7 μm, respectively, both of which agree well with those of the template shown in Figure 1. The same is the case for the platinum cones shown in Figure 6a. In contrast, we obtained truncated platinum cones, as shown in Figure 6c, using the conical pores with the highest aspect ratio as a template. The vapor-deposited platinum cathode could not reach the bottom of the pores, which is likely because the platinum atoms and ions sputtered from the target arrived inside the conical pores from

**Figure 6.** FE-SEM images of the platinum cones. The cones were electrochemically deposited in the conical pores prepared in (**a**) 2.0 (**b**) 4.0, and (**c**) 6.0 mol/dm<sup>3</sup> aqueous NaOH solutions at 60 ◦C. were fabricated by a combination of ion-track etching with the electrodeposition technique.

Figure 7 depicts the EDX spectrum of the platinum cones. The platinum signals were observed at 9.42 keV (L<sup>α</sup> line), 2.04 keV (M<sup>α</sup> + M<sup>β</sup> line) and 1.59 keV (M<sup>ξ</sup> line). Meanwhile, the other emission lines were assigned to the non-metal components, likely from the electrolyte solution PRECIOUSFAB Pt3000, e.g., carbon, nitrogen, oxygen, sulfur, and chlorine. In the electroless deposition method, the other signals of the minor components, such as Sn and Ag, were observed [4]. The platinum cones without any metal contaminations were fabricated by a combination of ion-track etching with the electrodeposition technique. The platinum cones were further analyzed using TEM. Figure 8 shows a TEM image of the tip area of the cones shown in Figure 6b. The radius of curvature of the tip was 13 nm. The selected-area electron diffraction pattern in the inset of Figure 8 exhibits concentric rings composed of bright, discrete diffraction spots that were indexed to the (111), (200), (220), and (311) crystal planes of fcc platinum, indicating the polycrystalline structure of the individual cones.

**Figure 7.** EDX spectrum of the obtained platinum cones. The arrows indicate the platinum signals observed at 9.42 keV (Lα line), 2.04 keV (Mα + Mβ line) and 1.59 keV (Mξ line). **Figure 7.** EDX spectrum of the obtained platinum cones. The arrows indicate the platinum signals observed at 9.42 keV (Lα line), 2.04 keV (Mα + Mβ line) and 1.59 keV (Mξ line).

The platinum cones were further analyzed using TEM. Figure 8 shows a TEM image of the tip area of the cones shown in Figure 6b. The radius of curvature of the tip was 13 nm. The selected-area electron diffraction pattern in the inset of Figure 8 exhibits concentric

nique.

ture of the individual cones.

rings composed of bright, discrete diffraction spots that were indexed to the (111), (200), (220), and (311) crystal planes of fcc platinum, indicating the polycrystalline structure of the individual cones. **Figure 7.** EDX spectrum of the obtained platinum cones. The arrows indicate the platinum signals observed at 9.42 keV (Lα line), 2.04 keV (Mα + Mβ line) and 1.59 keV (Mξ line).

*Quantum Beam Sci.* **2021**, *5*, x FOR PEER REVIEW 9 of 13

**Figure 6.** FE-SEM images of the platinum cones. The cones were electrochemically deposited in the conical pores prepared in (**a**) 2.0 (**b**) 4.0, and (**c**) 6.0 mol/dm<sup>3</sup> aqueous NaOH solutions at 60 °C.

Figure 7 depicts the EDX spectrum of the platinum cones. The platinum signals were observed at 9.42 keV (L<sup>α</sup> line), 2.04 keV (M<sup>α</sup> + M<sup>β</sup> line) and 1.59 keV (M<sup>ξ</sup> line). Meanwhile, the other emission lines were assigned to the non-metal components, likely from the electrolyte solution PRECIOUSFAB Pt3000, e.g., carbon, nitrogen, oxygen, sulfur, and chlorine. In the electroless deposition method, the other signals of the minor components, such as Sn and Ag, were observed [4]. The platinum cones without any metal contaminations were fabricated by a combination of ion-track etching with the electrodeposition tech-

The platinum cones were further analyzed using TEM. Figure 8 shows a TEM image of the tip area of the cones shown in Figure 6b. The radius of curvature of the tip was 13 nm. The selected-area electron diffraction pattern in the inset of Figure 8 exhibits concentric rings composed of bright, discrete diffraction spots that were indexed to the (111), (200), (220), and (311) crystal planes of fcc platinum, indicating the polycrystalline struc-

**Figure 8.** TEM image of the tip of the platinum cone. The inset shows the selected-area electron diffraction pattern.

Figure 9a shows the cyclic voltammograms obtained in a 0.5 mol/dm<sup>3</sup> aqueous H2SO<sup>4</sup> solution. The base diameter, length and areal density of the platinum cones are 550 nm, 2.4 <sup>µ</sup>m and 1.0 <sup>×</sup> <sup>10</sup>8/cm<sup>2</sup> , respectively. In this areal density, nearly half of the cones could be isolated, while the remaining ones may overlap to form multiple (mainly double) cones according to the calculation using a Poisson distribution model [21]. For comparison, the platinum plate with a diameter of 1.5 cm was fabricated by electrodeposition on the PC film without any etched pores, and its circular portion (0.8 cm in diameter) was measured in the same way. Both of the samples exhibited a hydrogen adsorption and desorption region at 0.02–0.4 V vs. RHE and a double layer plateau region at 0.4–0.6 V vs. RHE with peaks for the formation and reduction of surface platinum oxide at 0.6–1.17 V vs. RHE. The ECSA for the hydrogen adsorption was 1.8 times higher for the cones than for the plate. In a previous paper [15], the surface area resulting from the double-layer capacitance was estimated for comparison with the ECSA. This was determined by the non-Faradaic double-layer charging current around 0.5 V vs. RHE. It was approximately 1.7 times higher for the platinum cones than for the plate; therefore, the increase in the ECSA could reasonably be accounted for within the allowable error.

The electrocatalytic performance for ethanol oxidation was demonstrated by CV in an aqueous solution containing 0.5 mol/dm<sup>3</sup> ethanol and 0.5 mol/dm<sup>3</sup> H2SO4. The CV curves of the platinum cones and platinum plate are shown in Figure 9b. The current in the forward scan exhibited the oxidation of ethanol, whereas in the backward scan, another oxidation current was observed, which was associated with the oxidation of intermediates of ethanol dissociative adsorption [22]. The ECSA-normalized current densities (called specific currents) at 0.7 V vs. RHE were extracted and compared with those of platinum on carbon (often referred to as Pt/C) and platinum black found in a study by Mao and coworkers [23]. The platinum cones and plate indicated approximately 0.06 and 0.02 mA/cm<sup>2</sup> , respectively, whereas both of the commercial products exhibited 0.06–0.07 mA/cm<sup>2</sup> . The

only difference between our measurements and theirs was the ethanol concentration, and the CV curve was recorded in an aqueous solution containing 2 mol/cm<sup>3</sup> ethanol and 0.5 mol/cm<sup>3</sup> H2SO4. The twofold higher ethanol concentration would lead to an increase in the current density because it likely related to a greater amount of electroactive species in the solution [24]. Therefore, our platinum cones could exhibit the best performance among these four samples. The ECSA for the hydrogen adsorption was 1.8 times higher for the cones than for the plate. In a previous paper [15], the surface area resulting from the double-layer capacitance was estimated for comparison with the ECSA. This was determined by the non-Faradaic double-layer charging current around 0.5 V vs. RHE. It was approximately 1.7 times higher for the platinum cones than for the plate; therefore, the increase in the ECSA could reasonably be accounted for within the allowable error.

**Figure 8.** TEM image of the tip of the platinum cone. The inset shows the selected-area electron

be isolated, while the remaining ones may overlap to form multiple (mainly double) cones according to the calculation using a Poisson distribution model [21]. For comparison, the platinum plate with a diameter of 1.5 cm was fabricated by electrodeposition on the PC film without any etched pores, and its circular portion (0.8 cm in diameter) was measured in the same way. Both of the samples exhibited a hydrogen adsorption and desorption region at 0.02–0.4 V vs. RHE and a double layer plateau region at 0.4–0.6 V vs. RHE with peaks for the formation and reduction of surface platinum oxide at 0.6–1.17 V vs. RHE.

Figure 9a shows the cyclic voltammograms obtained in a 0.5 mol/dm<sup>3</sup> aqueous H2SO<sup>4</sup> solution. The base diameter, length and areal density of the platinum cones are 550 nm,

, respectively. In this areal density, nearly half of the cones could

*Quantum Beam Sci.* **2021**, *5*, x FOR PEER REVIEW 10 of 13

diffraction pattern.

2.4 μm and 1.0 × 10<sup>8</sup>

/cm<sup>2</sup>

**Figure 9.** Cyclic voltammograms of the platinum cones and plate (**a**) in 0.5 mol/dm<sup>3</sup> H2SO<sup>4</sup> using a scan rate of 50 mV/s and (**b**) in 0.5 mol/dm<sup>3</sup> ethanol + 0.5 mol/dm<sup>3</sup> H2SO<sup>4</sup> using a scan rate of 20 mV/s. **Figure 9.** Cyclic voltammograms of the platinum cones and plate (**a**) in 0.5 mol/dm<sup>3</sup> H2SO<sup>4</sup> using a scan rate of 50 mV/s and (**b**) in 0.5 mol/dm<sup>3</sup> ethanol + 0.5 mol/dm<sup>3</sup> H2SO<sup>4</sup> using a scan rate of 20 mV/s.

It should be emphasized that the current density was 3.2 times higher for the platinum cones than for the platinum plate. This comparison leads to the assumption that the cone structure contributes to the improvement of the electrocatalytic activity. The high electrocatalytic activity may be exhibited by the field-induced reagent concentration [25,26]. In other words, the polar ethanol molecules are likely to approach the surface of the cone electrode because the fine tips of the cones produce high local electrical fields. Finally, we demonstrated the fabrication of platinum cones by ion-track etching and electrodeposition techniques and found them to be a promising alternative for ethanol oxidation electrocatalysts.
