Probing the Surface Chemistry of Nanoporous Gold via Electrochemical Characterization and Atom Probe Tomography

Surface chemistry information is crucial in understanding catalytic and sensing mechanisms. However, resolving the outermost monolayer composition of metallic nanoporous materials is challenging due to the high tortuosity of their morphology. In this study, we first elaborate on the capabilities and limitations of atom probe tomography (APT) in resolving interfaces. Subsequently, an electrochemical approach is designed to characterize the surface composition of nanoporous gold (NPG), developed from dealloying an inexpensive precursor (95 at. % Ag, 5 at. % Au), by the means of aqueous electrochemical measurements of the selective electrosorption of sulfide ions, which react strongly with Ag, but to a significantly lesser extent with Au. Accordingly, cyclic voltammetry was performed at various scan rates on NPG in alkaline aqueous solutions (0.2 M NaOH; pH 13) in the presence and absence of 1 mM Na2S. Calibrations via similar voltammetric measurements on pure polycrystalline Ag and Au surfaces allowed for a quantitative estimation for the Ag surface coverage of NPG. The sensitivity threshold for the detection of the adsorbate–Ag interaction was assessed to be approximately 2% Ag surface coverage. As curves measured on NPG only showed featureless capacitive currents, no faradaic charge density associated with sulfide electrosorption could be detected. This study opens a new avenue to gain further insight into the monolayer surface coverage of metallic nanoporous materials and assists in enhancement of the interpretation of APT reconstructions.


Introduction
Nanoporous gold (NPG) derived from selective electrolytic dissolution of Ag from Ag-Au(-Pt) alloys [1,2] has shown promising results in terms of catalysis, sensing, bioanalytical and biomedical applications [3][4][5][6][7]. Gaining insight into the surface composition of metallic nanoporous materials could not only enhance the understanding of the underlying mechanisms of nanoporosity evolution and catalysis, but would be crucial in biosensing applications, in particular, in vivo sensors. However, obtaining insight into the entire surface composition of such tortuous nanostructures has proved to be challenging despite the use of various advanced atomic-scale characterization techniques such as atom probe tomography (APT) [8,9]. While APT reconstructions describe the materials with an atomic resolution, the composition of monolayers (subnanometer bins) cannot be resolved with certainty due to APT aberrations [10][11][12]. Additionally, APT is a destructive and highly localized method of investigation, hence there is a need for a non-local and non-destructive surface analysis method to investigate the surface composition of such nanoporous materials.
This study is an attempt at shedding further light on the surface composition of NPG by augmenting APT data with aqueous electrochemical techniques. The question of how much of the surface of the NPG is still covered by Ag is raised and electrochemical characterization is employed for the investigation.
It is hypothesized that measuring electrochemical responses from selective interactions of certain species with the surface of NPG advances the analysis of the composition of its outermost monolayer. The most straightforward electrochemical approach to this problem would be a comparison of electrosorption of aqueous ions that would react readily with one element on the surface but would react to a significantly lesser extent with the other element on the surface; or ideally would not react at all. Assuming that this difference in reactivity could be captured electrochemically, a quantitative assessment regarding the surface composition of electrodes could potentially be obtained.
Electrochemical surface characterization has already been used to estimate the platinum surface coverage of NPG structures developed from electrochemical dealloying of Ag-Au-Pt alloy precursors containing 20 to 23 at. % Au and a small atomic percentage of Pt (1 to 3 at. %) [13]. The electro-adsorption of under-potentially deposited hydrogen (H UPD ) is characteristic of Pt group metals only, while over-potential deposition of hydrogen occurs on all electrode materials during the hydrogen evolution reaction [14]. Measurements of the charge density for H UPD on NPG structures (formed by dealloying Ag-Au-Pt alloys) and their comparison to the values reported in the literature for pure polycrystalline Pt surfaces [15] have allowed Pt surface coverage estimations. Such estimations have showed surprisingly good agreement with surface analysis results from X-ray photoelectron spectroscopy (XPS) [13,16]. However, whether isolated Pt atoms on the surface would interact with hydrogen, and in general the ensemble effects of Pt in H UPD on alloys are not fully elucidated yet [17][18][19][20].
In this study we leverage the selective electrosorption of hydrosulfide ions (HS − ) on Ag as measured via cyclic voltammetry to characterize the surface composition of NPG developed from electrolytic dealloying of lean noble alloys containing only 5 at. % Au (Ag 95 Au 5 ).

Sample Preparation
A 99.999% pure gold wire (purchased from Alfa Aesar, Tewksbury, MA, United States) with a diameter of 1 mm was spot welded to a copper wire. The junction was then masked using Miccroshield lacquer (purchased from Structure Probe, Inc., West Chester, PA, United States) resulting in an exposed surface area of 0.17 cm 2 . The wire was cleaned in 20% nitric acid followed by rinsing with copious amount of deionized (DI) water prior to each measurement.
A 99.999% pure silver rod (purchased from Alfa Aesar, Tewksbury, MA, United States) with a diameter of 7 mm (a geometric surface area of 0.38 cm 2 ) was mounted in epoxy and polished with abrasive paper down to 800 grit finish. Then, the surface was cleaned ultrasonically in ethanol and was finally rinsed with DI water and dried with pressurized air before each experiment.
Silver-gold alloys were fabricated using a vertical induction cold crucible semilevitation furnace custom designed by and purchased from Arcast Inc (Oxford, Maine, United States). A total of 5 g of Ag-Au binary alloy containing 5 at. % gold, balanced with silver (Ag 95 Au 5 ), which corresponds to 91.24 wt. % Ag and 8.76 wt. % Au was fabricated using 99.999% pure Au and Ag pellets purchased from Kurt J. Lesker Company (East Sussex, United Kingdom). The melting process was done twice to ensure perfect mixing. The resulting lumps of alloys were then cut into 1 mm sheets using a Buehler (Lake Bluff, IL, United States) IsoMet 1000 Precision Cutter equipped with a diamond blade. The sheets were cold rolled into 200 µm foils. Small samples ranging from 0.1 to 0.5 cm 2 were then cut out of the foils. Annealing was performed in H 2 -Ar atmosphere (2.5% H 2 , balanced with Ar) at 900 • C for 5 h. All the alloy specimens were used in the as-annealed condition without any further surface preparation. In preparation for the electrochemical measurements, the samples were spot welded to copper wires. The junction was then masked using lacquer.

Electrochemistry
All electrochemical measurements were performed using a Gamry (Warminster, PA, United States) Reference 600 digital potentiostat controlled by Gamry Instruments Framework software. The charges associated with electrosorption measured by cyclic voltammetry were calculated using Gamry Echem Analyst version 5.60.
All solutions were prepared using reagent grade de-ionized water (Type 1, 18.2 MΩ cm resistivity) and were deaerated for 30 min by high-purity nitrogen (minimum purity: 99.998%) purging via a fine porous glass frit prior to the onset of the experiments. During the electrochemical measurements, the deaeration was continued at a lower flow rate above the level of the solution to eliminate the possibility of oxygen intake. A Pt foil with a surface area of 2 cm 2 was used as a counter electrode.

Cyclic Voltammetry
Cyclic voltammetry was performed at scan rates of 20, 50 and 100 mV s −1 to investigate sulfur adsorption/desorption on the surfaces. Measurements on Ag and Au electrodes were used to calibrate the peaks as a baseline for the characterization of the surface of as-annealed Ag 95 Au 5 alloy and NPG electrodes.
A basic solution (pH 13) of 0.2 M sodium hydroxide (NaOH) with and without 1 mM sodium sulfide (Na 2 S) was used for cyclic voltammetry experiments. Both chemical compounds used in these experiments were Analar grade and purchased from Sigma Aldrich (St. Louis, MO, United States). The deaeration step preceded Na 2 S addition, which was followed by vigorous stirring by a magnetic stir bar.
The reference electrode (RE) used in 0.2 M NaOH solutions was a mercury/mercury oxide electrode with a 20% potassium hydroxide (4.24 M KOH) electrolyte (Hg/HgO; +0.09 V vs. standard hydrogen electrode (V SHE )). For the electrochemical measurements in the presence of 1 mM Na 2 S, a silver/silver sulfide (Ag/Ag 2 S) RE was fabricated by anodizing a high-purity (99.99%) Ag wire (with a diameter of 1 mm and a length of 7 cm) at +0.5 V vs. a saturated calomel electrode (SCE; +0.24 V SHE ) for 30 s in a 1 M Na 2 S solution [21]. Measurements against Hg/HgO RE for an extended period of time revealed a stable potential at −0.64 V for the Ag/Ag 2 S RE.
The upper limit of the scanned potential window was chosen at −0.07 V Ag/Ag2S to avoid bulk Ag sulfidation, which in the presence of Na 2 S would result in Ag 2 S film formation and consequent surface roughening after several redox cycles [22]. The consideration with the lower potential limit (−0.65 V Ag/Ag2S ) was to avoid large current densities for hydrogen evolution, for which the equilibrium potential is approximately −0.8 V SHE according to the Nernst equation at the conditions of the experiments.

Electrochemical Dealloying
A 250 mL electrochemical cell containing 125 mL 0.5 M HClO 4 solution, prepared from Analar grade HClO 4 (62%, Alfa Aesar, Tewksbury, MA, United States) was employed for dealloying experiments. A mercury/mercury sulfate electrode (MSE; +0.64 V SHE ) was used as the reference electrode. Dealloying was done via potentiostatic chronocoulometry at 0.3 V MSE and to a charge density of 2.5 C cm −2 . In all cases, nanoporosity was induced on all sides of the samples.
Upon dealloying, samples were removed from the cell and immediately washed with DI water, followed by immersion in a large volume of DI water for 10 min. Then, cyclic voltammetry measurements were performed immediately after rinsing with DI water and open-circuit potential (OCP) stabilization (~20 min) in the fresh electrolytes.

Electrochemical Impedance Spectroscopy (EIS)
To quantify the increase in the surface area after dealloying, and to gauge the appropriateness of the scan rates used in voltammetry of NPG electrodes, double-layer capacitance measurements were employed using EIS at OCP, AC voltage 10 mV rms, frequency range 100 kHz-0.1 Hz with 10 points/decade.

APT Limitation in Resolving Interfaces
To enhance the mechanical properties of the dealloyed materials in preparation for APT, Cu electrodeposition is used to fill the nanoporosity [8]. The differences in chemistry and structure of Cu-filled NPG samples impart surface imperfections which affect the electronic field projecting from the tip of the APT sample, which in turn leads to trajectory aberrations in the departing ions, and can cause both local magnification or de-magnification of phases, as well as overlap of atoms from different phases in the APT reconstructions. The effect of these local evaporation artifacts on the position of measured atoms is illustrated in Figure 1. As revealed by the plotted iso-concentration contours, Cu-originally absent from the NPG phase-is detected throughout almost all regions of the dataset due to the prevalence of trajectory aberrations. Although established APT data analysis algorithms, such as proximity histograms and surface excess calculations, can yield insightful information about the overall composition and shape of the ligaments [8,9,[23][24][25], the local evaporation artifacts prevent an atomic-scale look at the monolayer covering the ligament surfaces.

Electrochemical Impedance Spectroscopy (EIS)
To quantify the increase in the surface area after dealloying, and to gauge the appropriateness of the scan rates used in voltammetry of NPG electrodes, double-layer capacitance measurements were employed using EIS at OCP, AC voltage 10 mV rms, frequency range 100 kHz-0.1 Hz with 10 points/decade.

APT Limitation in Resolving Interfaces
To enhance the mechanical properties of the dealloyed materials in preparation for APT, Cu electrodeposition is used to fill the nanoporosity [8]. The differences in chemistry and structure of Cu-filled NPG samples impart surface imperfections which affect the electronic field projecting from the tip of the APT sample, which in turn leads to trajectory aberrations in the departing ions, and can cause both local magnification or de-magnification of phases, as well as overlap of atoms from different phases in the APT reconstructions. The effect of these local evaporation artifacts on the position of measured atoms is illustrated in Figure 1. As revealed by the plotted iso-concentration contours, Cu-originally absent from the NPG phase-is detected throughout almost all regions of the dataset due to the prevalence of trajectory aberrations. Although established APT data analysis algorithms, such as proximity histograms and surface excess calculations, can yield insightful information about the overall composition and shape of the ligaments [8,9,[23][24][25], the local evaporation artifacts prevent an atomic-scale look at the monolayer covering the ligament surfaces. It is noteworthy to mention that the viability of forming nanoporosity from such lean noble alloys has been debated [26,27]. For a given average ligament diameter, the theoretical minimum threshold of the Au content in binary Ag-Au precursor alloys required to render the formation of a fully Au covered nanoporous morphology after dealloying can It is noteworthy to mention that the viability of forming nanoporosity from such lean noble alloys has been debated [26,27]. For a given average ligament diameter, the theoretical minimum threshold of the Au content in binary Ag-Au precursor alloys required to render the formation of a fully Au covered nanoporous morphology after dealloying can be estimated. Modeling the ligaments as Au covered cylinders with Ag cores suggests a minimum Au content of about 5 at. % for an average ligament size of ca. 15 nm [9].

Electrochemical Characterization
Cyclic voltammograms for a pure polycrystalline Ag electrode in 0.2 M NaOH solution in the presence and absence of 1 mM Na 2 S are shown in Figure 2. The voltammetric curves obtained in solutions containing only 0.2 M NaOH ( Figure 2a) show a featureless capacitive current except for the region where the hydrogen evolution reaction (HER) is kinetically favorable. In the solutions containing 1 mM Na 2 S, however, three partially overlapping Nanomaterials 2021, 11, 1002 5 of 12 peaks and an isolated peak are present at −0.60, −0.47, −0.37, and −0.15 V Ag/Ag2S half-wave potentials (E 1/2 = (E pc + E pa )/2 where E pc and E pa represent the cathodic and anodic peak potentials, respectively), respectively, and are in excellent agreement with the values reported in the literature for sulfur adsorption/desorption potentials in similar systems [28][29][30][31][32]. Similar measurements on a pure polycrystalline Au electrode are shown in Figure 3 and with a more refined scale depicted in the insert. The behavior of Au in 0.2 M NaOH solution is purely capacitive within the potential range where the HER does not interfere. No distinct peak can be resolved in the presence of 1 mM Na2S, which attests to the weak electrochemical interactions between Au and HS − as compared to those with the surface of Ag. Although there are reports of kinetically sluggish sulfur adsorption peaks on Au in different aqueous systems than the ones used in this study [36], there are various contrasting results among different investigators [37][38][39][40][41]. The absence of any conspicuous For solutions containing Na 2 S in this investigation, the exclusive sulfur species present in the alkaline electrolyte used (0.2 M NaOH, pH 13) is hydrosulfide (HS − ) as determined from the literature pKa values for H 2 S (7.02) and HS − (17.1) [33] and potential-pH equilibrium diagrams for the S-H 2 O system at room temperature [34]. Moreover, similar voltammetric measurements using 35 S-labeled Na 2 S and radiotracer techniques [35] have revealed The appearance of more than one peak suggests that the HS − adsorption is more complex than is indicated in Equation (1), perhaps involving cation coadsorption and/or potential dependent structural rearrangement of the adlayer [29]. Generally, multiple cycles of voltammetric measurements on a number of electrodes indicated a high level of reproducibility for the observed peaks.
The well-resolved peak at −0.15 V Ag/Ag2S is attributed to the oxidation of AgSH ads into Ag 2 S ads (Equation (2)) [29,30,35]. This step does not correspond to a significant increase in the sulfur surface coverage, but instead, results from an oxidative phase transition involving a pre-existing and essentially complete AgSH ads layer [29].
Ag + AgSH ads + OH − = Ag 2 S ads + H 2 O + e − Similar measurements on a pure polycrystalline Au electrode are shown in Figure 3 and with a more refined scale depicted in the insert. The behavior of Au in 0.2 M NaOH solution is purely capacitive within the potential range where the HER does not interfere. No distinct peak can be resolved in the presence of 1 mM Na 2 S, which attests to the weak electrochemical interactions between Au and HS − as compared to those with the surface of Ag. Although there are reports of kinetically sluggish sulfur adsorption peaks on Au in different aqueous systems than the ones used in this study [36], there are various contrasting results among different investigators [37][38][39][40][41]. The absence of any conspicuous peak in the cyclic voltammetry of Au in the Na 2 S containing electrolyte used in this study indicates that HS − is suitable for the purpose of surface characterization of NPG. The current densities at potentials more negative than −0.4 VAg/Ag 2 S are considerably affected by the HER (Figures 2 and 3). Therefore, for the surface characterization discussion, only the portions of the curves that are located more positive than −0.4 VAg/Ag 2 S were considered. The peak observed at −0.37 VAg/Ag 2 S in the case of the Ag electrode was selected as the electrochemical signal that distinguishes Ag from Au since the HER contributions to the current densities around this potential are negligible on both electrodes and the behavior of Au appears to be purely capacitive. After integrating the areas below the curves corresponding to the HS − adsorption/desorption on pure Ag in 0.2 M NaOH containing 1 mM Na2S, within the potential range where the peak at −0.37 VAg/Ag 2 S is present-between −0.255 and −0.385 VAg/Ag 2 S-and averaging the values for both anodic and cathodic peaks observed at scan rates of 20, 50, and 100 mV s −1 , and subtracting the charge density corresponding to the curves measured in the absence of 1 mM Na2S, a total faradaic charge density of 30.39 ± 2.01 µC cm −2 was calculated. The plus-minus value represents the standard deviation among two sets of measurements. The current densities at potentials more negative than −0.4 V Ag/Ag2S are considerably affected by the HER (Figures 2 and 3). Therefore, for the surface characterization discussion, only the portions of the curves that are located more positive than −0.4 V Ag/Ag2S were considered. The peak observed at −0.37 V Ag/Ag2S in the case of the Ag electrode was selected as the electrochemical signal that distinguishes Ag from Au since the HER contributions to the current densities around this potential are negligible on both electrodes and the behavior of Au appears to be purely capacitive.
After integrating the areas below the curves corresponding to the HS − adsorption/desorption on pure Ag in 0.2 M NaOH containing 1 mM Na 2 S, within the potential range where the peak at −0.37 V Ag/Ag2S is present-between −0.255 and −0.385 V Ag/Ag2Sand averaging the values for both anodic and cathodic peaks observed at scan rates of 20, 50, and 100 mV s −1 , and subtracting the charge density corresponding to the curves measured in the absence of 1 mM Na 2 S, a total faradaic charge density of 30.39 ± 2.01 µC cm −2 was calculated. The plus-minus value represents the standard deviation among two sets of measurements. Figure 4 shows the cyclic voltammetry measurements conducted on the surface of asannealed (non-dealloyed) Ag 95 Au 5 alloy. Since the metallic radii of Ag and Au are virtually the same, 95% of the surface area of Ag 95 Au 5 alloy can be regarded as Ag, assuming negligible surface segregation during the annealing stage [42].  To normalize the current response of the cyclic voltammetry measurements on the NPG samples by their true surface area, electrochemical impedance spectroscopy (EIS) measurements were performed before and after dealloying in deaerated 0.5 M HClO4 solution (Figure 5a) to estimate the double-layer capacitance (Cdl) using the relationship presented in Equation (3).
Then, the ratio of the Cdl after dealloying to the Cdl before dealloying as measured at 10 Hz, which is ca. 240 for dealloying this particular Ag95Au5 sample to 2.5 C cm −2 , was used as a measure of the factor by which the surface area has been multiplied as a result of nanoporosity evolution, assuming that the surface composition change does not change the elementary capacitance.
Estimating the surface area of NPG using Cdl measurement via EIS has shown an excellent correlation with the results from the Brunauer-Emmett-Teller (BET) technique [13]. Comparisons between the EIS estimations of the Cdl of NPG and estimations made by other electrochemical techniques have indicated that the surface area estimated by EIS is 1.5-to 1.75-fold larger than estimations made by Cu UPD or Au oxidation/reduction methods [51]. However, the effects of post-porosity coarsening during measurements involving polarization of the NPG electrodes have not been considered. Due to the lower free surface energy and molar heat of sublimation of Ag compared to Au, Ag atoms in a solid solution of Ag-Au alloy are expected to preferentially diffuse and concentrate at the surface. Several studies have verified Ag surface segregation in Ag-Au systems under ultra-high vacuum (UHV) conditions [43][44][45][46] and under various oxidative atmospheres [47][48][49][50]. A recent APT investigation of surface segregation in Ag-Au alloys at elevated temperatures (323 K < T < 673 K) indicated that under an unreactive atmosphere, such as Ar (6 kPa), Ag segregation is negligible for short treatment duration (1 h) [42]. On the other hand, reductive treatments in H 2 was found to lead to Au enrichment on the surface [42]. Overall, given the annealing atmosphere of 2.5% H 2 balanced with Ar at 1 atm used for the Ag 95 Au 5 samples in our study, negligible surface segregation is a safe assumption, as also demonstrated by electrochemical characterization in the following.
The four peaks associated with the HS − adsorption/desorption on pure Ag (Figure 2b) appear dampened-each to a different extent-in Figure 4. An average value of 28.40 ± 1.08 µC cm −2 was yielded for the total faradaic charge density of the adsorption/desorption of HS − on the surface of as-annealed Ag 95 Au 5 within the potential range of −0.255 and −0.385 V Ag/Ag2S using the same method described earlier in the case of pure Ag electrode. This value is in a good agreement with what would be expected based on the calibration with pure Ag electrode (0.95 × (30.39 ± 2.01) = 28.87 ± 1.91 µC cm −2 ).
To normalize the current response of the cyclic voltammetry measurements on the NPG samples by their true surface area, electrochemical impedance spectroscopy (EIS) measurements were performed before and after dealloying in deaerated 0.5 M HClO 4 solution (Figure 5a) to estimate the double-layer capacitance (C dl ) using the relationship presented in Equation (3).
Nanomaterials 2021, 11, x FOR PEER REVIEW 9 of 13 A thorough investigation into the post-porosity coarsening of NPG developed from lean noble alloys is the subject of a future study. Preliminary observations indicated that after 1 h of exposure to 0.5 M HClO4 and simultaneous polarization at 0 VMSE, which is well below the critical potential (+0.125 VMSE; Figure 2 in [9]), the surface area reduced by a factor of ca. 0.9, as measured via EIS by the method described earlier. Since Cu UPD and Au oxidation/reduction methods require polarization of the electrodes, we have adhered to EIS at OCP to estimate the true surface areas in this study.
The dependence of the absolute value of the imaginary impedance (|Zimg|) on frequency (f) for the NPG, Ag95Au5, Ag, and Au samples in 0.2 M NaOH is displayed in Figure 5b. The plot of log(|Zimg|) vs. log(f) has a constant slope of −1 for an ideal capacitive behavior. Over at least two frequency decades, from 1 to 100 Hz (highlighted with yellow), the slope of the plots is close to −1 (red dashed lines), suggesting a nearly ideal capacitive behavior. Figure 5b indicates that the dependence of the Cdl estimations on f for the NPG sample is similar to those for flat Ag95Au5, Ag, and Au surfaces over a wide range of frequencies. Cyclic voltammetries on NPG electrodes show featureless curves ( Figure 6). The measurement on NPG in the presence of 1 mM Na2S is also shown in Figure 6b with a finer scale and the superimposition of the HS − electrosorption peak observed at −0.37 VAg/Ag 2 S on pure Ag electrode adjusted in current density to represent various Ag surface coverages. The peaks can be distinguished from the capacitive current density on Ag surface coverages as low as 2%. The ensemble effects of Ag in HS − electrosorption on binary Ag-Au systems can be examined in further work by analyzing the results from model alloys and fabricated surfaces. Then, the ratio of the C dl after dealloying to the C dl before dealloying as measured at 10 Hz, which is ca. 240 for dealloying this particular Ag 95 Au 5 sample to 2.5 C cm −2 , was used as a measure of the factor by which the surface area has been multiplied as a result of nanoporosity evolution, assuming that the surface composition change does not change the elementary capacitance.
Estimating the surface area of NPG using C dl measurement via EIS has shown an excellent correlation with the results from the Brunauer-Emmett-Teller (BET) technique [13]. Comparisons between the EIS estimations of the C dl of NPG and estimations made by other electrochemical techniques have indicated that the surface area estimated by EIS is 1.5-to 1.75-fold larger than estimations made by Cu UPD or Au oxidation/reduction methods [51]. However, the effects of post-porosity coarsening during measurements involving polarization of the NPG electrodes have not been considered.
A thorough investigation into the post-porosity coarsening of NPG developed from lean noble alloys is the subject of a future study. Preliminary observations indicated that after 1 h of exposure to 0.5 M HClO 4 and simultaneous polarization at 0 V MSE , which is well below the critical potential (+0.125 V MSE ; Figure 2 in [9]), the surface area reduced by a factor of ca. 0.9, as measured via EIS by the method described earlier. Since Cu UPD and Au oxidation/reduction methods require polarization of the electrodes, we have adhered to EIS at OCP to estimate the true surface areas in this study.
The dependence of the absolute value of the imaginary impedance (|Z img |) on frequency (f) for the NPG, Ag 95 Au 5 , Ag, and Au samples in 0.2 M NaOH is displayed in Figure 5b. The plot of log(|Z img |) vs. log(f) has a constant slope of −1 for an ideal capacitive behavior. Over at least two frequency decades, from 1 to 100 Hz (highlighted with yellow), the slope of the plots is close to −1 (red dashed lines), suggesting a nearly ideal capacitive behavior. Figure 5b indicates that the dependence of the C dl estimations on f for the NPG sample is similar to those for flat Ag 95 Au 5 , Ag, and Au surfaces over a wide range of frequencies.
Cyclic voltammetries on NPG electrodes show featureless curves ( Figure 6). The measurement on NPG in the presence of 1 mM Na 2 S is also shown in Figure 6b with a finer scale and the superimposition of the HS − electrosorption peak observed at −0.37 V Ag/Ag2S on pure Ag electrode adjusted in current density to represent various Ag surface coverages. The peaks can be distinguished from the capacitive current density on Ag surface coverages as low as 2%. The ensemble effects of Ag in HS − electrosorption on binary Ag-Au systems can be examined in further work by analyzing the results from model alloys and fabricated surfaces. On a separate note, the overpotential required for HER is considerably lower in the 0.2 M NaOH solutions containing 1 mM Na2S on the surface of pure Au, as seen in Figure  3, and on the Au-enriched surface of NPG, as seen in Figure 6. It is more difficult to discern this trend from the measurements on pure Ag or as-annealed Ag95Au5 alloy due to the interference from the current densities associated with the oxidation of AgSHads into On a separate note, the overpotential required for HER is considerably lower in the 0.2 M NaOH solutions containing 1 mM Na 2 S on the surface of pure Au, as seen in Figure 3, and on the Au-enriched surface of NPG, as seen in Figure 6. It is more difficult to discern this trend from the measurements on pure Ag or as-annealed Ag 95 Au 5 alloy due to the interference from the current densities associated with the oxidation of AgSH ads into Ag 2 S ads . There are also reports of enhanced proton and water reduction rates on heterogeneous catalysts composed of noble and/or transition metal sulfides [52][53][54][55][56].

Conclusions
New insight into the composition of the outermost monolayer of a nanoporous metallic material was gained by designing an aqueous electrochemical method for surface characterization, based on selective and reversible electrosorption of anions on the electrode.
The question of to what extent the outermost monolayer of a nanoporous structure, formed from dealloying lean noble alloys, is composed of the more noble element was raised. Limitations of APT in resolving the surface composition of the NPG were elucidated. Electrochemical methods were adopted to augment the APT data in better characterizing the surface composition of NPG developed from electrolytic dealloying of Ag 95 Au 5 alloys. Adsorption-desorption of HS − in alkaline environments (0.2 M NaOH; pH 13) was found to be sufficiently selective towards Ag as opposed to Au, hence permitting quantification of the Ag surface coverage of the electrodes.
Cyclic voltammetry measurements on a pure polycrystalline Ag electrode in 0.2 M NaOH solutions containing 1 mM Na 2 S revealed a total of four peaks below the potentials required for bulk Ag 2 S formation. Measurements on a pure polycrystalline Au electrode showed a featureless capacitive current except for the region where HER is kinetically favorable.
In measurements on an as-annealed Ag 95 Au 5 surface, four peaks appeared at similar potentials associated with HS − electrosorption on Ag electrode, although each dampened in current density to a different extent. The faradaic charge densities associated with one peak at the half-wave potential of −0.37 V Ag/Ag2S on the Ag 95 Au 5 electrode was found to be very close to 95% Ag surface coverage. This peak, which was also not considerably affected by contributions from the HER on the electrodes used in this study, was then selected for calibration purposes.
The curves measured on NPG showed a featureless capacitive current except for the cathodic currents from the HER. Preliminary sensitivity analysis indicated that the peak used in calibration can be distinguished from the capacitive current densities, even when attenuated in intensity according to Ag surface coverages as low as 2%. Future work will focus on electrochemical characterization of NPG using other adsorbates and will further detail the sensitivity threshold by testing model alloys and known surface compositions.
Additionally, strong catalysis of water reduction was observed on the surface of pure Au and on NPG where sulfur was present in the solution.
Author Contributions: All authors contributed to the conceptualization, experimental design, writing, and editing. A.F.-E. was responsible for experimentation and data analysis. B.L. was responsible for acquiring and curating APT data. R.C.N. supervised and acquired the funding. All authors have read and agreed to the published version of the manuscript.

Data Availability Statement:
The data presented in this study are available upon reasonable request from the corresponding author. The data are not publicly available since they form parts of ongoing studies.