Facile fabrication of hierarchically nanostructured gold electrode for bio-electrochemical applications

Abstract

Nanoporous gold (NPG) is one of the most extensively investigated nanomaterials owing to its tunable pore size, ease of surface modification, and range of applications from catalysis, actuation, and molecular release to the development of electrochemical sensors. In an effort to improve the usefulness of NPG, a simple and robust method for the fabrication of hierarchical and bimodal nanoporous gold electrodes (hb-NPG) containing both macro-and mesopores is reported using electrochemical alloying and dealloying processes to engineer a bicontinuous solid/void morphology. Scanning electron microscopy (color SEM) images depict the hierarchical pore structure created after the multistep synthesis with an ensemble of tiny pores below 100 nm in size located in ligaments spanning larger pores of several hundred nanometers. Smaller-sized pores are exploited for surface modification, and the network of larger pores aids in molecular transport. Cyclic voltammetry (CV) was used to compare the electrochemically active surface area of the hierarchical bimodal structure with that of the regular unimodal NPG with an emphasis on the critical role of both dealloying and annealing in creating the desired structure. The adsorption of different proteins was followed using UV–vis absorbance measurements of solution depletion revealing the high loading capacity of hb-NPG. The surface coverage of lipoic acid on the hb-NPG was analyzed using thermogravimetric analysis (TGA) and reductive desorption. The roughness factor determinations suggest that the fabricated hb-NPG electrode has tremendous potential for biosensor development by changing the scaling relations between volume and surface area which may lead to improved analytical performance. We have chosen to take advantage of the surface architectures of hb-NPG due to the presence of a large specific surface area for functionalization and rapid transport pathways for faster response. It is shown that the hb-NPG electrode has a higher sensitivity for the amperometric detection of glucose than does an NPG electrode of the same geometric surface area.

1. Introduction

Nanoporous gold (NPG) is a three-dimensional framework of bicontinuous pores and Au ligaments, prepared by dissolving the less noble component(s) from an Au alloy by using chemical or electrochemical methods. The pore size of the NPG can be tailored over the range from a few nanometers to several hundred nanometers by tuning the composition of the alloy and the dealloying conditions. Bicontinuous NPG composed of interconnected nanoscale ligaments is typically formed from alloys through the evolution of nanoporosity by percolation dealloying [1,2]. Dealloyed NPG can be formed as nanoparticles, nanowires, microwires, and films. NPG possesses advantages such as being physically robust and chemically inert, having tunable pore size, high surface-to-volume ratio, good electrical conductivity, biocompatibility, and permeability [3]. Numerous methods including chemical, electrochemical, photothermal, and electroannealing have been devised for tuning NPG morphology [4]. NPG electrodes have been used for a wide range of applications such as in the development of immunosensors [5], DNA sensors [6], enzymatic biosensors [7], enzyme-free biosensors [8], catalytic platforms [9], and in fundamental structure–property studies at the nanoscale [10]. The plasmonic properties of NPG make it a highly explored material to be used as a Surface-Enhanced Raman Spectroscopy (SERS) substrate for ultrasensitive sensing applications. Thin self-standing NPG membranes have been explored for their localized surface plasmon resonance between 500 and 900 nm [11,12]. The highly tunable morphology of NPG can be exploited to optimize the resonant frequency for each plasmonic modality. The most common method for the fabrication of NPG is the corrosion of less noble elements from their alloy with gold. In the fabrication of NPG from Ag/Au alloys, Ag atoms are dissolved under corrosive conditions such as in concentrated acid or under an applied positive potential. The dealloying process generates surface vacancies from the detachment events and then the remaining surface atoms diffuse forming a bicontinuous network of ligaments. Under free corrosion, the process takes several hours whereas with applied bias potential (typically less than 1 V above the standard potential of the less noble element) only a few minutes are enough to complete the process [13].

Electrochemical systems with microelectrodes of very high surface areas are of great interest due to their potential application in electrochemical sensor development. The introduction of nanostructures on the electrode surface can increase the effective surface area, increase the sensitivity, and enhance the catalytic activity. The electrode morphology dictates the electrochemical sensor performance as the pore size of the nanomatrix plays a critical role in the capture and transport of molecules. It has been seen that small pores help in target detection in complex samples and larger pores increase the accessibility of the target molecule thereby enhancing the detection limits of the sensor [14]. As compared to traditional planar electrodes, NPG offers 3 orders of magnitude improvement in detection limit due to increased target penetration into the porous network [15]. The commonly used methods of introducing nanostructures into the electrode include the template technique, electrodeposition, self-assembly, sputter-deposition [16], attachment of white gold leaf and a PDMS flexible support [17], and alloying of multiple metals followed by selective dealloying. The dealloying process is one of the most effective routes to fabricate NPG. The less noble metal, which is the sacrificial metal, greatly affects the morphology of the dealloyed sample because of the difference in dissolution rates. The dynamic process of nanoporous gold structure formation wherein the less noble metal atoms leave the precursor alloy by oxidation and diffusion and the remaining metal atoms undergo rearrangement and consolidation, gives rise to an interconnected network of pores and ligaments [18].

Surface-molecule interactions and systemic modification of NPG surfaces play an essential role in understanding loading capacity and release kinetics of drugs and other molecules in these nanostructured materials. Immobilization of varying functional moieties has shown different packing density and steric effects which ultimately control the loading behavior in NPG [19]. Self-assembled monolayers (SAMs) of thiolated compounds are key elements in nanoscience and nanotechnology. SAMs are molecular assemblies formed spontaneously on a surface by chemisorption and can be organized into ordered domains. The formation of the 2D assembly is driven by the presence of intermolecular interactions and the chemical bond formation of molecules with the surface [20]. SAMs on Au surfaces are widely used as building blocks for the fabrication of different types of devices used in biology, biosensing, medicine, catalysis, photonics, and electronics [21,22]. Alkanethiol SAMs can be prepared by exposure of clean gold surfaces to alkanethiols [23]. Modification of gold surfaces with SAMs represents a method to link a variety of organic and biological species to the gold surface. The homogeneity and the nature of the packing of the SAMs are important factors in their applicability as incomplete surface coverage can be detrimental for biosensing and other applications [24]. On the journey toward practical implementation of nanostructured ensembles, increasing interest has been placed on the design of multifunctional platforms. The surface chemistry, composition, and stability of such smart materials can be tuned to accomplish ultra-trace detection of target moieties, and have a promising future in various technologies ranging from medicine to energy generation [25]. These applications have been targeted extremely well with a comprehensive library of hierarchical nanostructures that provides researchers with further insights for their future work in fabricating high-performance materials.

Hierarchical porous structures are often observed in nature and have been mimicked in nanoscale configurations to modulate the physical properties of materials for better performance [26]. Porous materials with hierarchical architecture are characterized by the presence of interconnected pores on different length scales [27]. Nanoporous metals formed by dealloying usually possess unimodal pore size distributions and therefore, various strategies have been developed to create a hierarchical bimodal porous structure with two distinct ranges of pore size [28]. Functional materials with a network of nanoscale ligaments interpenetrated by two levels of pore size give rise to structural hierarchy fulfilling the two key aspects of material design strategy which are separate and mutually conflicting in nature namely, the large specific surface area for functionalization, and rapid transport pathways [29,30]. Ding and Erlebacher demonstrated the technique of dealloying/plating/re-dealloying for creating such nanostructures where Au-Ag alloy first underwent regular etching followed by heat treatment with a unimodal pore size of upper hierarchical level and residual silver demanding a second dealloying to create the lower hierarchical level. This approach remained limited to thin films [31]. This two-step dealloying strategy was in itself versatile as the annealing temperature, porosity evolution, and upper hierarchy porous architectures can be tuned to generate varying pore sizes and topology [32]. Biener et al. suggested the use of ternary alloy composed of two less noble metals removed successively via corrosion. Cu-Ag-Au was used as the starting material wherein the removal of Cu and Ag gave rise to low-density NPG samples with a bimodal structure [33]. Considering the distinct behavior and reactivity of Al, work has been done utilizing a ternary alloy of AuAgAl as the source alloy foil for fabricating hierarchical nanoporous structures with bimodal pore size [34]. A similar approach of using two-step dealloying has been used to generate hierarchical porous structures from Pd₁Ag₂Al₉₇ as a precursor alloy with controllable Pd/Ag atomic ratios [35]. The methods described using ternary alloys do not give rise to long-range ordered structures. Zhang et al. used a two-phase master alloy of Al-Au, where larger pores were created by leaching out one of the phases which gave rise to the bimodal structure but with a low degree of order [36]. A highly ordered hierarchical structure in the macroscopic dimension was reported using nested network NPG (N³PG) [37]. An attractive, simple, green, and modified sol–gel route of preparing a three-dimensional hierarchically macro/mesostructured porous gold monolith has been reported [38]. The synergistic combination of nanocasting and chemical dealloying along with colloidal bijels generated a bimodal structure with macro-and mesoporosity [39]. Rapid solidification and chemical dealloying is another simple strategy that has been used to synthesize NPG gold ribbons with bimodal channel size distributions consisting of large channels of hundreds of nanometers with highly porous channel walls of tens of nanometers [40]. To produce an ordered and anisotropic hierarchical structure with control over multiple length scales from centimeters to nanometers, a combination of “bottom-up” and “top-down” processing routes has been reported. In this, direct ink writing (DIW) was used in combination with alloying and dealloying to create a hierarchical nanoporous gold structure [41]. Hierarchical nanoporous gold film electrode with an extra high roughness factor of 1250 has been synthesized using a multi-cyclic electrochemical co-alloying/dealloying approach. The electrode generated using this process had an extra high surface area and could be used as an excellent catalyst for methanol oxidation [42]. NPG surfaces with highly tunable surface roughness are ideal candidates for surface-enhanced Raman spectroscopy (SERS) and show high sensitivity toward the detection of molecules [9].

In this work, we report the fabrication of a hierarchical bimodal nanoporous structure using two-step dealloying with different compositions of Au-Ag alloy. The advantage of using this approach is the strict control over the structure and the pore size. Here, electrochemical dealloying is used prior to chemical dealloying. This can help in the fine-tuning of the characteristic length scale of the porosity and the average distance between the ligaments. The applied potential was optimized to control the rate of dissolution and interfacial diffusion of ions. Control over the Ag dissolution rate during dealloying is effective to avoid cracking. The perspective behind using different dealloying strategies in our fabrication procedure along with determining the favorable electrochemical potential and time and temperature for annealing paves the way toward the desired final morphology by controlling the amount of residual less noble metal constituent at each stage of the process.

A bimodal pore size distribution was seen using scanning electron microscopy with enhanced electrochemical and loading properties relative to the traditional NPG. The structure was further investigated electrochemically using cyclic voltammetry, thermally using thermogravimetric analysis (TGA), and protein loading was studied using UV–vis spectroscopy in a solution depletion method. The hb-NPG electrode was then shown to give improved non-enzymatic detection of glucose when compared to a regular NPG electrode of the same geometric area.

2. Materials and methods

2.1. Reagents

Gold wire (0.2 mm diameter, 99.99 %) was obtained from Electron Microscopy Sciences (Fort Washington, PA). Fetuin from fetal calf serum (lyophilized powder), peroxidase from horseradish, sodium chloride (NaCl), and bovine serum albumin (BSA) of ≥98 % purity, all were purchased from Millipore-Sigma (St. Louis, MO). Sodium carbonate (enzyme grade, >99 %), sulfuric acid (certified ACS plus), nitric acid (trace metal grade), hydrogen peroxide (50 %), and sodium bicarbonate (certified ACS) were all from Fisher Scientific (Pittsburg, PA). Potassium dicyanoargentate (K[Ag(CN)) (99.96 %) and potassium dicyanoaurate (K[Au(CN)2]) (99.98 %), potassium hexacyanoferrate (II) trihydrate, potassium hexacyanoferrate (III), ethanol (HPLC/spectrophotometric grade), α-lipoic acid, and sodium hydroxide (99.99 %) were obtained from Sigma–Aldrich (St. Louis, MO). 6-Ferrocenyl-1-hexanethiol (FcSH) was procured from Dojindo Molecular Technologies, Inc. West Gude Dr. Suite 260, Rockville MD 20850. Milli-Q water (18.2 MΩ cm at 25 °C) was prepared using a Simplicity UV system from Millipore Corporation, Boston, MA, USA. All chemicals, reagents, and proteins were used as received.

2.2. Instrumentation

The surface morphology, thickness, and the live quantitative colored images of the fabricated electrodes were obtained using Thermo-Fisher Scientific Apreo 2C scanning electron microscopy equipped with ColorSEM technology.

Thermogravimetric analysis to estimate the SAM surface coverage was done using a Q500 thermogravimetric analyzer (TA Instruments, DE, USA). The adsorption of proteins onto NPG was studied using a Varian Cary 50 UV–Visible spectrophotometer. UV–Vis spectra and absorbance readings were acquired using the Varian Cary 50 UV–Vis spectrophotometer. A Suprasil quartz spectrophotometer cuvette with a 10 mm light path and volume capacity of 1.0 mL (model number 14-385-902C, Fischer Scientific, Pittsburgh, PA, USA) was used for all experiments.

Electrodeposition and dealloying were carried out using an EG&G Princeton Applied Research 273A digital potentiostat/galvanostat and the PowerPULSE software. Cyclic voltammetry scans and reductive desorption studies were done using a VersaSTAT 4 potentiostat/galvanostat (Princeton Applied Research, AMETEK Scientific Instruments) and the VersaStudio software.

2.3. Preparation of nanoporous gold coated gold wires

Solutions of 50 mM K[Ag (CN)₂] and K[Au (CN)₂] in 0.25 M Na₂-CO₃ were prepared. The solution used for electrodeposition was prepared by combining 3.5 mL of the solution of K[Ag(CN)₂] with 1.5 mL of the K[Au(CN)₂] solution, followed by degassing with argon for 10 min. Electrodeposition was done using a three-electrode arrangement in a glass cell containing 5 mL of solution, a platinum wire counter-electrode, and an Ag|AgCl (sat’d. KCl) reference electrode. A gold wire of length 5.0 mm and a diameter of 0.2 mm served as the working electrode. Electrodeposition was carried out for a period of 10 min at a potential of − 1.0 V (vs Ag|AgCl in KCl (sat’d.) giving rise to AuAg alloy with a composition of Au₃₀Ag₇₀ (at. %). Dealloying was carried out by immersing the electrodeposited Au + Ag alloy-coated gold wire in concentrated nitric acid for 17 h followed by rinsing with Milli-Q water and ethanol [43].

2.4. Preparation of hierarchical bimodal nanoporous gold coated gold wires

A hierarchical bimodal nanoporous gold structure with two distinct ranges of pore size was created using a sequence of electrochemical dealloying-annealing-chemical dealloying. Solutions of 50 mM K[Ag (CN)₂] and K[Au(CN)₂] in 0.25 M Na₂CO₃ were prepared. The solution used for electrodeposition was prepared by combining 4.5 mL of the solution of K[Ag(CN)₂] with 0.5 mL of the K[Au(CN)₂] solution followed by degassing with argon for 10 min. Electrodeposition was done using a three-electrode arrangement in a glass cell containing 5 mL of solution, a platinum wire counter-electrode, and an Ag|AgCl (sat’d. KCl) reference electrode. A gold wire of length 5.0 mm and a diameter of 0.2 mm served as the working electrode. Electrodeposition was carried out for a period of 10 min at a potential of − 1.0 V (vs Ag|AgCl in KCl (sat’d.) giving rise to AuAg alloy with a composition of Au₁₀Ag₉₀ (at. %). Electrochemical dealloying was carried out for a period of 10 min at a potential of 0.6 V (vs Ag|AgCl in KCl (sat’d.). Following the first dealloying, the sample was annealed at 600 °C for 3 h using a Barnstead Thermolyne 47900 digital lab furnace (model F47915). After cooling to room temperature, a second chemical dealloying was carried out by immersing the annealed gold wire in concentrated nitric acid for 17 h followed by rinsing with Milli-Q water and ethanol. Using a three-step approach of alloying, partial dealloying, annealing, and chemical dealloying, hierarchical structures made of different alloy compositions, such as Au₁₅Ag₈₅, Au₂₀Ag₈₀, Au₂₅Ag₇₅, and Au₃₀-Ag₇₀ (at.%), were created. Depending on the desired composition of the initial alloy, a different electrodeposition solution was utilized. Au₁₅Ag₈₅, Au₂₀Ag₈₀, Au₂₅Ag₇₅, and Au₃₀Ag₇₀ required the solutions to be mixed as (4.25 mL + 0.75 mL), (4 mL + 2 mL), (3.75 mL + 1.25 mL), and (3.5 mL + 1.5 mL) of the K[Ag(CN)₂ and K[Au(CN)₂ solutions, respectively.

2.5. Characterization of electrodes and roughness factor analysis

The surface morphology of the fabricated electrodes was characterized using SEM. The hierarchical nanoporous gold was vertically mounted on the SEM stage to characterize the thickness of the electrode. Elemental analysis was done using energy-dispersive X-ray spectroscopy (EDS) at the voltage of 15 kV and current of 1.6 nA. The electrochemical surface area was determined by the gold oxide stripping method by scanning from −0.2 V to 1.6 V and back to −0.2 V (vs Ag|AgCl) at 100 mV s⁻¹ in 0.5 M H₂SO₄. The charge under the oxide reduction peak was integrated to estimate the electrochemically active surface area of the NPG-covered gold wires and various compositions of hierarchical bimodal nanoporous gold-covered gold wire, using the reported conversion factor of 450 μC cm².

2.6. Preparation of self-assembled monolayers (SAM)

SAMs of lipoic acid was prepared by immersing three wires at a time into a 1 mM α-lipoic acid solution, in ethanol, of 200 μL total volume. The wires were left for approximately 17 h (overnight) and then rinsed with ethanol thrice. Lipoic acid was used in our study as it generates a significant presence of defects due to the disordered nature of its SAM. These defects are of utmost importance for our future biosensor studies providing sites for the electrooxidation of products formed via enzymatic reactions [44].

2.7. Reductive desorption of lipoic acid SAM

Reductive desorption of lipoic acid SAM was carried out in a 0.5 M NaOH solution, argon degassed for 30 min. CV scans were performed in 5 mL of solution in a three-electrode cell arrangement. The CV scan was performed between 0 and − 1.5 V (vs Ag|AgCl) at a scan rate of 20 mV s⁻¹ [44].

2.8. Study of protein loading onto fabricated electrodes using solution depletion technique.

To characterize the loading of molecules onto NPG and hb-NPG surfaces, a UV–vis spectrophotometer-based kinetic absorption study was performed. The solutions of the proteins fetuin, BSA, and HRP in PBS buffer (0.01 M, pH 7.4) were prepared separately of concentration 1 mg mL⁻¹, 0.5 mg mL⁻¹, and 1 mg mL⁻¹, respectively. 500 μL of the prepared solution was placed in the special low volume cuvette along with one fabricated electrode immersed in it. A baseline recording using only PBS buffer was done before the protein solution was added. Real-time monitoring of loading was done for 120 min at a wavelength of 280 nm.

The concentration of the protein solutions was determined using the Beer-Lambert equation A₂₈₀ = ε (280) × C × L, where A₂₈₀ is the absorbance at 280 nm using a UV-spectrophotometer, C is the concentration in M, L is the path length in cm and ε ₂₈₀ is the extinction coefficient of the protein at 280 nm in units of M⁻¹ cm⁻¹. ε ₂₈₀ values of proteins were calculated as the weighted sum of the ε ₂₈₀ value of Trp, Tyr, and Cys using the proposed equation: ε ₂₈₀ (M⁻¹ cm⁻¹) = no. of tryptophan × 5500 + no. of tyrosine × 1490 + no. of cysteine × 125. The values were as follows: ε ₂₈₀ (BSA) = 42,925 M⁻¹ cm⁻¹ [45], ε ₂₈₀ (Fet) = 19,840 M⁻¹ cm⁻¹ [46], ε ₂₈₀ (HRP) = 25,250 M⁻¹ cm⁻¹ [47]. The protein solutions were used immediately after preparation.

We must first determine the molar extinction coefficient for the protein in question. After that, the absorbance at time t = 0 min was used to calculate the starting number of molecules using the Lambert-Beer law. The number of molecules at time t is calculated using the same formula. We can calculate the number of protein molecules that have entered the NPG structure and presumably been immobilized by subtracting the number of molecules at the respective time from the initial number of molecules present at t = 0 min.

2.9. Protein desorption from the electrodes

Desorption of proteins from NPG and hb-NPG electrodes was tested by taking the electrodes that had been dipped in the protein solutions for 2 h and suspending them in PBS buffer (0.01 M, pH = 7.4) and incubating them. The PBS solution was the choice due to its common use for biological experiments in vitro [48]. Desorbed proteins were measured after 24 h and 48 h of incubation using UV absorbance. All measurements were done in triplicate [49].

2.10. Thermogravimetric analysis

NPG and hb-NPG (10:90 initial alloy) electrodes onto which α-lipoic acid (1 mM) was immobilized (for overnight) were rinsed with ethanol and water, air-dried, and placed in a platinum weighing pan, and heated inside the thermogravimetric analyzer from room temperature to 600 °C at a ramping rate of 20 °C min⁻¹. The carrier gas used was nitrogen, which was passed at a flow rate of 40 mL min⁻¹. Prior to initiating the temperature ramp, N₂ gas was allowed to flow through the sample for 5–10 min. Initial mass, mass losses, and weight change percent were obtained from the analysis. The surface coverage of the molecules on the fabricated electrode’s surface was calculated based on the net mass loss and the electrochemical surface area determined through the gold oxide stripping method [50].

2.11. Study of electron transfer in SAMs of 6-ferrocenyl-1-hexanethiol on NPG and hb-NPG

SAM formation was carried out by immersing the three electrodes, bare gold wire, NPG, and hb-NPG in a 1 mM ethanolic solution of 6-ferrocenyl-1-hexanethiol (FcC6SH) for 18 h at room temperature. After monolayer formation, the electrodes were rinsed with ethanol. Cyclic voltammograms of all the modified electrodes were recorded at a scan rate of 50 mV s⁻¹ in the presence of 0.1 M NaClO₄ (deoxygenated for at least 15 min) supporting electrolyte in a 0–0.8 V potential window [51].

2.12. Electrochemical detection of glucose and the effect of interferents

CV and chronoamperometry were performed using a bare gold electrode, NPG, and hb-NPG as working electrodes. Electrodes were washed with ethanol and water before use. After drying in air at room temperature the working electrodes were used for the detection of glucose. The CV was carried out in 10 mM glucose in 0.1 M NaOH from −0.8 to 0.8 V at a scan rate of 50 mV s⁻¹ and the current response was measured. The response was compared for all three electrodes [52].

The real-time electrocatalytic activity for glucose oxidation on a bare gold electrode, NPG, and hb-NPG at potentials of −0.39 V, −0.02 V, and 0.57 V respectively, was determined using chronoamperometry. The detection of glucose was carried out in 4 mL of 0.1 M NaOH with successive addition of 50 μL of 10 mM glucose with constant stirring of the solution at 300 rpm [52].

The non-enzymatic sensor was tested in the PBS buffer (pH 7.4) containing glucose along with interfering species which are 30 times lower than glucose concentration. Selectivity of the sensor for glucose was tested in the presence of common interfering species which were ascorbic acid (50 μL of 2 mM), lactose (50 μL of 2 mM), sodium chloride (50 μL of 2 mM), and sucrose (50 μL of 2 mM). The current response of the sensor to sequential injections of glucose with the interferents was studied in 0.1 M NaOH [53].

2.13. Data analysis

All data calculations and graphing were done using Sigma Plot 12.0. Analysis of pore sizes and ligament width along with inter-ligament distances were done using ImageJ (imagej.nih.gov/ij/). To analyze the image using ImageJ, the SEM file is converted to tiff format. A new scale is introduced to the software for the image to be analyzed using the option of set scale. To measure the interligament distance and ligament width, 30–40 different locations were measured using the freehand line. A table consisting of all the measured lengths emerges on clicking “measure” under the option “analyze”.

3. Results and discussion

Concerning our fabricated electrodes, the adjustment of pore volume and ligament size is of paramount importance. The bimodal pore structure has shown superior electrocatalytic activity and sensitivity towards glucose oxidation in alkaline solution compared to traditional NPG [54]. In this work, we start by fabricating a bimodal structure with varying ratios of Au/Ag and focusing first on the characterization based on surface morphology, surface coverage and loading capacity of the traditional NPG coated gold wire.

4. Characterization of nanoporous gold and hierarchical bimodal nanoporous gold electrode prepared by varying the ratio of the Au-Ag alloy

Structural characterization was performed using SEM and composition analysis after the first and second dealloying was performed using EDX. Electrodeposition resulted in a shell-like morphology around the gold wire. The thickness of hb-NPG around the gold wire is approximately 3 μm as calculated using the side view images from SEM and manually measuring the thickness using the scale bar.

After chemical dealloying, NPG has shown an open interconnected network of ligaments and pores [55]. SEM images of varying alloy compositions revealed that out of all the compositions, hb-NPG prepared from Au₁₀Ag₉₀ (at.%) was comprised of upper and lower hierarchy without cracks in the structure. Here, upper hierarchy denotes the presence of larger pores and lower hierarchy is used to denote the smaller pores. Fig. 1 shows the SEM images of the top and side views of the exterior of hb-NPG coated gold wire prepared from alloys of varying Au: Ag ratios. EDX analysis showed that the amount of silver left after the first dealloying was significant as seen in Fig. 2 and ranged from 80 to 85 atomic %. The colored images indicate the changes in elemental composition wherein silver (red color) and gold (yellow color) are shown at each step involved in the fabrication of hb-NPG. This is a useful feature of color SEM with the Apreo 2 instrument that different elements can be assigned different colors. The electrode of our interest for further study, prepared from Au₁₀Ag₉₀ (at.%) was further analyzed using ImageJ software for determining the structural features such as ligament width and inter-ligament distance as reported in Table 1 and Table 2. The effects of the alloy composition on microstructures of the nanoporous structure were investigated. It was seen that with an increase in the silver content the size of the ligament increases from 819 ± 71 to 1031 ± 272 nm. The structure starts to coarsen during dealloying, and factors like the alloy composition, residual silver, temperature, or electrolyte might determine how quickly coarsening occurs [56].

Fig. 1.

SEM images of a) 10:90 (Au: Ag) alloy prepared by applying − 1.0 V (vs Ag/AgCl, KCl saturated) for 10 min on gold wire, b) morphology after electrochemical dealloying, c) pore-coarsening seen after annealing, d) Larger pores seen after chemical dealloying in concentrated nitric acid, e) and f) are showing the ligament width and interligament distance in upper and lower hierarchy respectively. Thickness of the hierarchical structure deposited on gold wire is shown in (g), (h) Histograms showing the ligament width and interligament distance in upper and lower hierarchy respectively.

Fig. 2.

A) Color SEM of the a1) Au₁₀:Ag₉₀ alloy, a2) electrochemically dealloyed structure, a3) annealed structure, and a4) hierarchical nanoporous structure after the final step of chemical dealloying B) EDX spectrum showing the elemental composition of the b1) alloy, b2) electrochemically dealloyed structure, and b3) hierarchical bimodal nanoporous gold structure.

Table 1.

Distribution of ligament width and inter-ligament distance of varying alloy compositions. The evaluation was done after completion of the multistep synthesis involving electrochemical dealloying, annealing, and chemical dealloying.

Alloy composition (Au:Ag)	Mean pore size (nm)
	Ligament width (nm)	Inter-ligament distance (nm)
30:70	819 ± 71	52 ± 7
25:75	940 ± 269	50 ± 3
20:80	1031 ± 272	62 ± 11

Table 2.

Distribution of ligament width and inter-ligament distance for the hierarchical bimodal structure where alloy composition was Au₁₀:Ag₉₀. The evaluation was done after completion of the multistep synthesis involving electrochemical dealloying, annealing, and chemical dealloying.

Hierarchy level	Ligament width (nm)	Inter-ligament distance (nm)
Upper	938 ± 285	853 ± 41
Lower	51 ± 5	52 ± 15

Table 2 shows the distribution of ligament width and interligament distance for the bimodal structure made from the alloy of Au₁₀Ag₉₀. The ligament width in the upper and lower hierarchy was calculated to be 938 ± 285 nm and 51 ± 5 nm respectively. The interligament distance was comparable to their ligament width and was found to be 853 ± 41 nm and 52 ± 15 nm for upper and lower hierarchical features, respectively.

4.1. Surface area measurement and roughness factor analysis

The roughness factor of all the fabricated electrodes, which is the ratio of electrochemical surface area (ECSA) to the geometrical surface area was obtained using cyclic voltammetry using a value of 450 μC cm⁻¹ for the reduction of a single layer of gold oxide where the charge under the oxide reduction peak was integrated to estimate the ECSA [44,57].

The impact of varying alloy composition along with the essential role of the three-step fabrication process can be seen in Fig. 3. The CV curves for hb-NPG electrodes prepared from Au: Ag alloys of different compositions are shown in Fig. 3a. The increase in silver content in the initial electrodeposited alloy results in a dramatic increase in the area under the peaks for gold oxidation and reduction for the hb-NPG electrodes resulting in an increase in the roughness factor. The comparison of multistep fabricated hb-NPG (10:90) with the chemically dealloyed NPG (10:90) is also shown in Fig. 3b which is indicative of the essential role of two-step dealloying and annealing in increasing the ECSA in comparison to chemical dealloying. Fig. 3c shows the CV after electrochemical dealloying and annealing. The cyclic voltammogram shows the existence of Ag and Au on the surface of the electrode after annealing. The oxidation of silver in the alloy starts at a more positive potential of 0.52 V in comparison with pure Ag (0.4 V). On the other hand, the oxidation of Au happens at the potential of approximately 0.8 V whereas the oxidation peak for pure gold is seen at 0.96 V [58]. The observable Au, and Ag reduction peak after annealing confirmed the bimetallic composition. Prior to annealing, the gold peak is not seen which might be due to the silver covering over the gold. Compared with Au, Ag showed a higher peak intensity due to its higher atomic percentage in the alloy [59]. The peak at 0.38 V in both cases is due to the formation of Ag₂O layers, while the peak at 0.6 V was associated with the oxidation of Ag₂O to AgO [60].

Fig. 3.

Cyclic voltammograms showing a) the comparison of electrochemically active surface area for the porous structure formed from different Au: Ag atomic percentages of the alloy. The varying composition of the starting alloy was Au₁₀:Ag₉₀, Au₂₀:Ag₈₀, Au₂₅:Ag₇₅ and Au₃₀:Ag₇₀ b) Au₁₀:Ag₉₀ alloy, chemically dealloyed structure formed after immersing in nitric acid, and finally the enhancement of peak current when electrochemical and chemical dealloying is used in combination (c) the oxidation and reduction peaks after electrochemical dealloying are shown in red where the peak at 1) denotes the oxidation of Au. The oxidation and reduction peaks after annealing are shown in blue where the peak at 1) is associated with the oxidation of Ag₂O to AgO. The peak at 2) in both cases is due to the reduction to Ag₂O layers. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

The ECSA of hb-NPG was found to be 7.64 cm². Our lab has previously reported the ECSA of NPG (Au₃₀Ag70) (at.%) to be 12.5 cm² on average determined from the oxide stripping experiments for the same size gold wire substrates [43]. Table 3 summarizes the ECSA data of all the fabricated electrodes along with their surface roughness factor calculation.

Table 3.

Calculation of the roughness factor (R_f) and electrochemically active surface area (ECSA) of the final structure formed from varying Au:Ag alloy composition. Charge under the reduction peak seen in CV was evaluated for all the different compositions. The 10:90 composition formed the hb-NPG structure with the highest roughness factor.

Composition	Charge (mC) under reduction peak	Mean	Standard dev	ECSA/cm2Rf
10:90	3.51 3.44 3.37	3.44	0.07	7.64231.5
20:80	1.82 1.93 1.7	1.82	0.11	4.04122.4
25:75	1.07 0.97 0.54	1.02	0.28	2.2668.5
30:70	0.13 0.11 0.37	0.20	0.14	0.4513.6

4.2. Determination of surface coverage of lipoic acid molecules on NPG and hb-NPG using thermogravimetric analysis and reductive desorption

In the present study, we used the destructive analytical method of TGA to quantify the number of molecules loaded onto hb-NPG and NPG which in return has allowed estimation of the surface coverage of the molecules taking into account the electrochemical surface area of the fabricated electrodes. We subjected the air-dried modified NPG and hb-NPG electrodes (modified using 1 mM lipoic acid) to pyrolytic decomposition in an inert environment while scanning up to 600 °C. At this temperature, lipoic acid molecules are expected to be completely decomposed. During the TGA temperature ramp, it is possible that some coarsening of the NPG or hb-NPG may occur although the time spent in the ramp is overall about 29 min and only 10 min are spent at 400 °C and above. For as long as they remain on the gold surface during the temperature ramp, the chemisorbed lipoic acid molecules may hinder gold surface diffusion by increasing the energy barrier for lateral motion of gold atoms and hence serve to slow any coarsening [61]. A systematic study of the effects of lipoic acid modification on thermal coarsening of NPG remains to be performed. The modified electrodes were subjected to TGA and it was found that NPG experienced a weight loss of 0.05 % and hb-NPG with a weight loss of 0.1 % from the initial weight of the sample during the temperature scan. The average mass loss was found to be 3.33 μg and 6.02 μg for NPG and hb-NPG respectively. This mass loss corresponded to 0.266 μg cm⁻² of LA self-assembled onto the NPG surface equivalent to 1.3 × 10⁻⁹ mol cm⁻² (12.5 cm² electrochemical surface area) and 0.788 μg cm⁻² of LA on the hb-NPG surface corresponding to 3.8 × 10⁻⁹ mol cm⁻² (7.64 cm² electrochemical surface area). The TGA thermograms (temperature ramped at 20 °C min⁻¹) of LA loading on both the electrodes along with reductive desorption data are depicted in Fig. 4 and in Table 4.

Fig. 4.

A) TGA thermograms of lipoic acid loading on NPG and hb-NPG from room temperature to 600 °C at a ramping rate of 20 °C min⁻¹. B) Reductive desorption of lipoic acid immobilized on NPG and hb-NPG carried out in 0.5 M NaOH solution, performed between 0 and −1.5 V at a scan rate of 20 mV s⁻¹.

Table 4.

TGA and reductive desorption data showing surface coverage for Lipoic acid SAMs on NPG and hb-NPG. The data presented is compared with previously reported values.

Method	Substrate	Surface coverage for LA SAMs (mol cm−2)			Reference
Reductive desorption	NPG	2.42 × 10−10			[65]
	Flat gold surface	3.5 × 10−10			[66]
	Flat gold surface	3.00 × 10−10			[50]
	NPG	2.09 × 10−10			Present study
	hb-NPG	3.42 × 10−10			Present study
	Substrate	Wt. change (%)	Mass loss (μg)	Surface coverage LA mol/cm2
TGA analysis	NPG	0.05	3.3 ± 0.7	1.3 × 10−9	Present study
	hb-NPG	0.1	6.02 ± 0.03	3.8 × 10−9	Present study

The surface coverage was estimated using the reductive desorption method. A reduction peak associated with the Au-S bond in the SAM is observed around −0.9 V which is due to the reductive desorption of lipoic acid from the gold surface [62]) The number of lipoic acid molecules was estimated using the charge under the reductive desorption peak and the surface area of the electrodes determined from the oxide stripping experiments. The average surface coverage of lipoic acid SAMs was found to be 3.42 × 10⁻¹⁰ mol cm⁻² on hb-NPG and 2.09 × 10⁻¹⁰ mol cm⁻² on NPG. The surface coverage data from TGA suggests that the molecules might be contributing to mass loss that are from the areas inside the electrodes that cannot be reached via the electrochemical desorption technique and hence accounts for a greater number of molecules per unit area of the electrode. There may also be lipoic acid molecules physisorbed on top of those bonded to Au surfaces. Previous studies with commercially available gold plates (6.0 mm × 6.0 mm × 0.25 mm) have shown the surface coverage of 1.31 × 10¹⁸ molecules m⁻² (equivalent to 2.18 ×10⁻¹⁰ mol cm⁻²) [50]. In addition, it is theoretically estimated that a complete and ordered coverage of LA molecules on a gold surface would fall in the range of 3.45 × 10^{18 —} 4.40 × 10¹⁸ molecules m⁻² using the unit cell dimensions of LA (a = 11.744 Å, b = 9.895 Å, c = 9.246 Å). This corresponds to four LA molecules per unit cell. Some of the reported surface coverages of LA on flat gold surfaces are 3.00 × 10⁻¹⁰ mol cm⁻²(1.81 × 10¹⁸ molecules m⁻²) [63], 3.50 × 10⁻¹⁰ mol cm⁻²(2.11 × 10¹⁸ molecules m⁻²) [64], and 2.42 × 10⁻¹⁰ mol cm⁻²(1.46 × 10¹⁸ molecules m⁻²) [43,65,66].

4.3. Absorption of proteins with varying structures onto NPG and hb-NPG

The absorption of proteins from solution onto NPG electrodes has attracted tremendous amounts of attention, primarily due to the great interest in the development of biosensor devices [67–69]. Hierarchical bimodal nanoporous gold (hb-NPG) presents a new opportunity as a support for a range of applications due to the high loading capacity. The influence of surface morphology, pore size, and pore volume play a pivotal role in the study of absorption kinetics of proteins and glycoproteins [70]. Synthesis of an electrode with a bimodal pore size distribution is highly promising to achieve high capacity loading of proteins as their diffusion deeper into the material should be enabled. This suggests that the rational design of a hierarchical bimodal pore size distribution in the hb-NPG electrode and the optimization of meso- and macropores maximized the protein loading [71].

Horseradish peroxidase (HRP) is one of the most studied enzymes of the plant peroxidase superfamily owing to its use in different bioanalytical applications and diagnostic kits [72]. The HRP molecular weight is 44 kDa, about 9.4 kDa of which is from the glycan modification of the protein. The X-ray data have shown that the protein conformation is stabilized by four disulfide bonds along with the presence of two calcium-binding sites. The isoelectric point (pI) of HRP is 9 and it shows maximum activity in the pH range of 6–8 [73]. The loading experiment was performed in PBS, pH 7.4, which is lower than the pI of HRP thereby imparting a positive charge that can electrostatically bind with the negative gold electrode surface [74,75]. The real-time protein loading observed in the case of HRP can be seen in Fig. 5. HRP has 8 cysteine, 4 methionine, and 6 lysine residues on or near its surface. These sulfur-containing amino acid residues (cysteine and methionine) and –NH₂-containing amino acid residues (lysine) would facilitate the multipoint attachment of HRP on the electrode’s surface [76]. Moreover, the enhanced accessibility of the enzyme to the inner surface might be attributed to the macropores present in hb-NPG [77].

Fig. 5.

Real-time protein loading study using UV–vis spectrophotometer showing the change in absorbance and the average number of immobilized BSA, Fetuin and HRP molecules on NPG and hb-NPG electrodes. BSA, Fetuin, and HRP loading on NPG is shown in (a), (c), and (e) respectively. The loading of same proteins on hb-NPG is shown in (b), (d), and (f) respectively. Each graph depicted is an average of three measurements. Data depicted as an average (N = 3).

The second candidate selected in our loading study is a 48.4 kDa serum glycoprotein fetuin which is a key player in bone formation and human metabolic processes [78]. It is an acidic glycoprotein rich in sialic acid contributing to its net negative charge at physiological pH in the terminal residues as well as in several aspartate/glutamate-rich domains. Its isoelectric point is close to that of albumin, which is around 5 [79–81]. The aqueous solution of fetuin when exposed to the metallic surface, adsorbs to some extent due to the dynamic interplay of electrostatic, van der Waals and hydration interactions [82]. This protein is of interest for future applications of hb-NPG for development of biosensors for glycoproteins and serves as an example.

Bovine serum albumin is a larger molecular weight protein (66.4 kDa) that makes up approximately 10 % of total whey proteins in milk [83]. The structural features of BSA have opened its use in drug delivery applications [84]. The isoelectric point of BSA is 4.6 making it negatively charged at pH 7.4; however, the preferential binding of BSA to negatively charged surfaces has been attributed to the 60 lysine groups present on its surface that can have electrostatic interactions with negatively charged surfaces. The adsorption of BSA on the electrode surface can be explained using two possible mechanisms involving either “end-on” or “side-on” binding in which end-on interaction gives a higher surface coverage of BSA on the electrode surface [85].

A real-time protein loading study is graphically depicted in Fig. 5. The number of HRP molecules immobilized on the electrode during the loading experiment is found to be higher than fetuin and BSA due to the strong electrostatic interaction between protein and the metal surface. The immobilization was higher for hb-NPG than NPG for BSA and for fetuin (Table 5) and slightly lower for HRP. The dimensions of BSA are reported as 14 nm × 4 nm × 4 nm [86] indicating a surface footprint per protein molecule as small as 16 nm² or as large as 64 nm², corresponding to coverages of 6.25 × 10¹² molecules cm⁻² and 2.5 × 10¹³ molecules cm⁻², respectively and assuming complete coverage on a flat surface. Our measured estimates of 2.0 × 10¹² molecules cm⁻² and 2.85 × 10¹² molecules cm⁻² on NPG and hb-NPG are reasonable and suggest a very significant extent of protein coverage of BSA within NPG and hb-NPG. The dimensions of HRP are reported as 4.0 nm × 6.7 nm × 11.7 nm [87] indicating a surface footprint per protein molecule as small as 27 nm² or as large as 78 nm² corresponding to coverages of 3.7 × 10¹² molecules cm⁻² and 1.07 × 10¹³ molecules cm⁻², respectively and assuming complete coverage on a flat surface. Our measured estimates of 9.48 × 10¹² molecules cm⁻² and 7.0 × 10¹² molecules cm⁻² on NPG and hb-NPG are reasonable and suggest very significant extent of protein coverage of HRP within NPG and hb-NPG perhaps enhanced by electrostatic interactions. In case of Fetuin immobilization, our measured estimates of 8.63 × 10¹² molecules cm⁻² and 9.33 × 10¹² molecules cm⁻² on NPG and hb-NPG are reasonable and also suggest a very significant extent of fetuin coverage within NPG and hb-NPG.

Table 5.

Number of immobilized protein molecules on NPG and hb-NPG. Data depicted as average (3 measurements).

Proteins used in the study (molar extinction coefficient. M−1 cm−1)	No. of molecules immobilized/cm2 on NPG	No. of molecules immobilized/cm2 on hb-NPG
BSA	2.03 × 1012	2.85 × 1012
HRP	9.48 × 1012	7 × 1012
Fetuin	8.63 × 1012	9.33 × 1012

The desorption process of incorporated proteins under the same conditions revealed that the proteins did not desorb out into the buffer even after 48 h of incubation in the buffer (V = 1 mL). Proteins are well known to stick to gold surfaces via Au-NH₂ interactions with lysine residues on the protein surface or with any accessible cysteine residues in the protein structure. Therefore, it seems these proteins are strongly physisorbed onto the interior gold surfaces.

4.4. Quantitative analysis of the SAMs of ferrocene 1-hexanethiol

Cyclic voltammetric data for ferrocene moieties of FcC6SH immobilized on the Au, NPG and hb-NPG electrodes as self-assembled monolayers is depicted in Fig. 6. Voltammograms on bare gold wire and NPG electrode show sharp oxidation and reduction peaks while on hb-NPG broad peaks are seen. The shape of the voltammograms was found to be independent of the scan rate from 0 to 0.8 V. Repeated scanning does not change the features demonstrating the stability of the monolayer to electrochemical cycling. Broadening in the peak occurs due to monolayer heterogeneity and it appears that hb-NPG presents a more heterogeneous environment for SAM formation [51]. The interaction between ferrocene sites or high surface coverage leading to inhomogeneity of the sites leads to lateral repulsive interaction in the monolayer and hence a broader peak [88]. The narrow width of the redox peak suggests the formation of a densely packed monolayer, where anions are responsible for screening repulsive interactions between ferrocenium cations. The difference between anodic and cathodic peaks is quite small and is indicative of the occurrence of the redox reaction of the adsorbed species on the electrodes [89]. The cyclic voltammograms show a pair of redox peaks in the potential region between + 300 and + 400 mV, corresponding to the oxidation and reduction of the terminal ferrocene group of the monolayer. The amount of 6-ferrocenyl-1-hexanethiol was evaluated using the charge of the oxidation peak [90].

Fig. 6.

Cyclic voltammograms of SAMs of 6-ferrocenyl-1-hexanethiol with unmodified bare gold wire, chemically dealloyed electrode, and hb-NPG (solution composition reported) (NaClO₄, 0.1 M; scan rate = 50 mV s⁻¹). Increasing the concentration of the electrolyte to 0.5 M has little variation on the shape of the curve.

Surface coverage of the redox-active surface group has been calculated using the integrated charge associated with the cathodic voltammetric peak using the relation Γ = Q_cv/nFA_ECSA, where Q is the Faradaic charge, n = 1 is the number of electrons involved in the redox process for ${\text{F}}_{\text{c}}/{\text{F}}_{\text{c}}^{+}$ redox couple, F is the Faraday constant, and A_ECSA is the electrochemical surface area of the electrode. The theoretical value of maximum surface coverage of ferrocene alkanethiol on Au (111) is 4.5 × 10⁻¹⁰ mol cm⁻². High surface coverage of 5.47 × 10⁻⁹ and 30.1 × 10⁻¹⁰ mol cm⁻² has been observed for NPG and hb-NPG electrodes. This might be attributed to the lateral electron transfer or electrostatic interactions between redox groups as well as additional physisorbed molecules [91].

The values obtained for FWHM for the monolayers formed on NPG and hb-NPG are listed in Table 6 along with other voltammetric parameters. The theoretical value of FWHM for a one-electron process is 90.6 mV when the monolayer exhibits an ideal electrochemical response. The differing responses obtained from the monolayers on different electrodes suggest that repulsive interactions exist within the monolayers (FWHM > 90.6) formed on hb-NPG and attractive interactions give rise to FWHM values of less than 90.6 mV in NPG [92,93]. The values for FWHM for the bare gold wire, NPG, and hb-NPG were seen to be 9.1, 39, and 160 mV. It has been observed that the electroactive species anchored on surfaces with different microstructures can experience non-uniform local environments, the result of various forces such as electrostatic interactions, strong dipole − dipole interactions, or high mobility. The presence of such lateral interactions results in broader peaks for molecules experiencing repulsive interactions and narrower peaks for attractive interactions. The type of interaction is coming from the orientation difference of the ferrocene thiol species on the same substrate but with different morphological features [94]. The lateral interactions in the monolayer or a neighbor-attractive surface activity effect are revealed by the narrowness of the redox peak. The narrow width of the redox peak of ferrocene hexanethiol SAMs in NPG is characteristic of a densely packed layer where anions induce attractive forces between ferricenium cations. On the other hand, in hb-NPG, the redox peak of ferrocene hexanethiol SAMs is broader and less sharp. Broadened voltammetric peak indicates a kinetic inhomogeneity among the surface ferrocene population which suggests either repulsive interaction between ferrocene units or monolayer heterogeneity arising due to the presence of upper and lower hierarchical architectures on the surface of the electrode [89,95–97]. The redox potential E⁰ was determined from the average of the anodic and cathodic potentials, and the separation between the peaks is only several mV, indicating that the electron transfer occurs rapidly and reversibly [98].

Table 6.

Voltammetric parameters and surface coverage (per geometric area) values for 6-ferrocenyl-1-hexanethiol SAMs. Data depicted as average (3 measurements) ± std dev.

Electrode	Charge/μC	Surface coverage/nmol cm−2	E0=(Ec + Ea)/2 (mV)	FWHM/mV	Ic/μA	Ia/μA
Bare gold	1.1 ± 0.03	0.35	347.02	9.1	−3.46	4.6
Chemically dealloyed	16.9 ± 0.7	5.47	354.55	39.14	−25.89	37.07
hb-NPG	92.93 ± 5.5	30.1	326.13	160	27.14	29.56

4.5. Electrochemical detection of glucose

CVs for bare gold, NPG, and hb-NPG in the absence and presence of 10 mM glucose were recorded to evaluate the catalytic performance of the electrodes. The oxidation peak current was seen at −0.39 V for bare gold, at −0.02 V for NPG, and 0.57 V for hb-NPG as seen in Fig. 7. The peak current was found to be 30, 400, and 1000 μA for bare gold, NPG, and hb-NPG respectively. Compared to the bare gold and NPG, hb-NPG produced more than a two times enhancement in the current, indicating the higher sensitivity of the electrode [99]. The presence of dual-sized pores and much larger roughness in hb-NPG generates more active sites further verifying the superior catalytic performance of the electrode. The features seen in the CVs for NPG in the absence of glucose resemble those previously reported in 0.1 M NaOH [100,101] and are also similar to those reported on cooled bead gold electrodes in 0.1 M NaOH solution [102]. The formation of an adsorbed layer of hydroxide by oxidation of the water adlayer occurs followed by formation of a surface layer of Au(OH)₃ during the anodic scan that is then reduced back to Au during the cathodic scan. The CV in the absence of glucose for hb-NPG shows a similar reduction peak as NPG but has additional features in the anodic scan that could indicate the presence of different exposed crystalline faces of gold and possibly a small amount of residual silver giving rise to the very small additional cathodic peak.

Fig. 7.

CVs of (i) bare gold, (ii) NPG and (iii) hb-NPG recorded in N₂ saturated NaOH solution with 10 mM glucose.

The real-time electrocatalytic response of varying amounts of glucose on NPG and hb-NPG was evaluated using the chronoamperometric technique at an applied constant potential taken from the CV of the respective electrode. As the concentration of glucose increased from 10 mM to 50 mM, a rapid current increment was observed as seen in Fig. 8. When the glucose concentration exceeded 50 mM, the current response seen in NPG became saturated. This might be due to the adsorption of the intermediates formed during the process into the active sites on the NPG electrode which restricted the diffusion of incoming glucose molecules [52]. In contrast, in hb-NPG the current had not yet saturated at 90 mM. The current response to varying glucose concentration demonstrated a linear relationship with the correlation coefficient of r² = 0.93, indicating diffusion-controlled behavior. Previous research on the performance of glucose sensors based on gold nanostructures has shown that the high sensitivity of the nanoporous gold electrode is attributed to the electrode’s surface area that enhances the electro-oxidation of glucose. Free-standing NPG film fabricated by chemical dealloying of Au₃₅Ag₆₅ (at.%) alloy leaves gave a sensitivity of 20.1 μA mM⁻¹ cm⁻² in the concentration range of 1–18 mM [103]. Another study using a porous gold cluster film having disordered 3D hierarchical pores and network structure has shown a rapid response and high sensitivity of 10.76 μA mM⁻¹ cm⁻² towards glucose determination [104]. Urchin-like gold submicrostructures synthesized using a seed-mediated method displayed a high sensitivity of 16.8 μA mM⁻¹ cm⁻² due to their high surface-to-volume ratio [105]. Using the hydrogen bubble dynamic template synthesis followed by a galvanic replacement reaction, gold (Au) films with open interconnecting macroporous walls and nanoparticles have been effectively fabricated. The resulting porous Au sheets showed a linear range from 2 to 10 mM with a sensitivity of 11.8 μA cm⁻² mM⁻¹, and a detection limit of 5 μM towards the electrooxidation of glucose [106]. Another nonenzymatic NPG-based amperometric glucose sensor was developed using a simple, quick square-wave oxidation reduction cycle (SWORC). With a detection limit of 0.5 mol/L, the NPG electrode demonstrated under ideal conditions an excellent linear correlation between the current response signal and the glucose concentrations in the ranges of 2 mol/L to 1.375 mmol/L and 1.375 mmol/L to 15 mmol/L [107]. The NPG with a 12-nm pore size (a-NPG) was produced by electrochemical dealloying for 25 s at 303 K under anodic potential (600 mV). For the detection of glucose utilizing an a-NPG electrode, the obtained minimum sensible concentrations are 413 nM in alkaline media and 1 μM in neutral solutions [108]. With a sensitivity of 41.9 μA mM⁻¹ cm⁻² and a detection limit below 30 μM, a gold nanowire array electrode has demonstrated a linear differential-pulse voltammetric response for partial oxidation of glucose at low potentials [109]. Using a bio-inspired manufacturing technique and the natural egg-shell membrane, innovative and sensitive 3D hierarchical porous Au networks (HPANs) have been created. With two linear ranges of 1–500 μM and 4.0–12 mM, a limit of detection (LOD) of 0.2 μM, and a quick response time, this non-enzymatic glucose sensing device demonstrated exceptional performance for glucose measurement [110]. A linear amperometric response at + 0.2 V with a sensitivity of 20 mA M⁻¹ cm⁻² has been observed for the electrooxidation of glucose at nanoporous gold (NPG) surfaces created by electrodeposition of AgAu alloy layers followed by the selective dissolution of Ag [111]. A simple one-step anodic potential step approach was used to successfully build a three-dimensional NPGF. Compared to previous porous gold electrodes, the NPGF electrode demonstrated great sensitivity (232 μA mM⁻¹ cm⁻²) and a detection limit of 53.2 μA toward glucose detection [112].

Fig. 8.

Chronoamperometric curve of (i) NPG and (ii) hb-NPG towards glucose with successive addition in 0.1 M NaOH at the working potential of 0.02 V and 0.57 V, respectively. The calibration curve of the NPG and hb-NPG glucose sensor is shown below each of the figures.

The slope of the current versus concentration plot was used to calculate the sensitivity. In the present work, it was found to be 6.53 μA mM⁻¹ cm⁻² for NPG and 15.97 μA mM⁻¹ cm⁻² for hb-NPG, indicating a higher sensitivity for the hb-NPG electrode. A comparison of the sensitivity and limit of detection of the hb-NPG glucose sensor with those in some other studies using NPG-based nonenzymatic glucose sensors reported in the literature is shown in Table 7, focusing on those made of gold. The low limit of detection and high sensitivity makes hb-NPG a promising electrode for applications in electrochemical sensing. The presence of the hierarchical architecture exposing many active Au-OH sites, known to be the active site for glucose oxidation in basic solution, evidently give rise to an improved current response.

Table 7.

A comparison of the sensitivity and limit of detection of the hb-NPG glucose sensor with those in other studies involving sensors made up of nanoporous structures of gold alone reported in the literature.

Glucose sensor	Limit of Detection (μM) (S/N = 3)	Sensitivity (μA mM−1 cm−2)	Reference
Porous Au film	5	11.8	[106]
NPG electrode film	0.5	–	[107]
a-NPG	0.413 in alkaline medium and 1 in neutral solutions	–	[108]
Au nanowire array	0.003	0.04	[109]
Urchin-like Au submicrostructures	10	16.8	[105]
3D free standing NPG films	3	20.1	[103]
Hierarchically porous gold cluster film	1	10.76	[104]
3D hierarchical porous Au networks	0.2	4.362	[110]
NPG surfaces	–	20	[111]
Ultrahigh porous NPG layers	–	135	[101]
Nanoporous gold film (NPGF)	53.2	232	[112]
Hb-NPG	0.18	15.97	This study

The selective response of the electrode for glucose was seen in the presence of other competing species. The amperometric response for the selectivity of hb-NPG sensor electrodes was measured by the addition of 10 mM glucose followed by the addition of 2 mM of interferents as shown in Fig. 9. The current response to glucose was compared with some common interferents, and it was concluded that the hb-NPG-based sensor electrode was highly selective, showing negligible response for other species in the mixture under these conditions. It has been proposed that the oxidation of glucose on NPG is under kinetic control while that of the interferents is under diffusion control and that the current magnitude due to glucose oxidations increases strongly with surface area while that from interferents does not [113]. Some of the interferents may also be oxidized at higher potentials on hb-NPG than needed to detect the glucose response.

Fig. 9.

The current response of the fabricated hb-NPG glucose sensor in the presence of common interferents.

5. Conclusion

We have presented a simple procedure of producing hierarchical and bimodal nanoporous consisting of pores in two widely separated size ranges, higher active electrochemical surface area, and high roughness factor by employing a multistep process involving electrochemical dealloying-annealing-chemical dealloying. We have shown how the variation in process parameters can alter the morphology and the electrochemical performance of the electrode. It was seen that the resulting hierarchical nanoporous gold electrode possessed macro and mesopores giving rise to higher loading capacity as compared to a nanoporous gold electrode when made with Au: Ag (10:90) and annealed at 600 °C for 3 h. Furthermore, the interconnected network in the structure produces efficient mass transport, resulting in high mass activity and sensitivity compared to the other nanostructured electrodes [114].

The hierarchical electrode possessed curved interconnected ligaments with a ligament width of 938 ± 285 nm for the upper hierarchy dominated by highly active low-coordinated atoms. The enhanced ligament width of hb-NPG is an excellent platform for catalytic reactions [115]. The presence of an additional lower hierarchical ligament width of 51 ± 5 nm imparts a large specific surface area for functionalization. By transitioning from dealloying a binary alloy of Au-Ag to a dual-level hierarchy in hb-NPG, we are creating material for applications requiring fast access to larger molecules and their fast detection as seen in catalysis and sensing [37]. A significant increase in the roughness factor from 13.6 to 231.5 was seen when going from Au₃₀:Ag₇₀ to Au₁₀:Ag₉₀ alloy composition. The density of electrochemically active sites is linked with the roughness factor of the material. The density of active sites on a flat surface when multiplied by the roughness factor can help us obtain the density of electrochemically active sites [116]. The higher surface roughness of hb-NPG as compared to NPG gives rise to greater electrochemical activity. This work lays a solid foundation for the design of high-performance electrochemical biosensors. The advantages of the hierarchical design of an electrode in biosensor applications stem from high surface area and easy modification of the surface [117].

The hierarchical structure of the electrodes with a high surface area has an impact on the protein loading. In our study, we found that hb-NPG with dual hierarchical features enables higher loading capacity than the conventional NPG. The hierarchical architecture allows proteins to penetrate through their interconnected macropores and the lower hierarchical features lay a base for protein binding and communication with the substrate. Protein loading in an electrode is determined by the type of the material and its morphology. NPG and hb-NPG are both gold-based but the number of immobilized proteins is higher in hb-NPG due to the presence of macropores and mesoporosity which accelerates the binding of most proteins by physical/electrostatic interactions [118]. The effect of morphological features in the bimodal electrode has been seen to affect the shapes of CVs of ferrocene thiol SAMs. Non-ideal behavior in the form of peak broadening is observed in the case of hb-NPG. The electrochemical response of redox-active SAMs measured using CV has major a contribution coming from its electronic and supramolecular arrangement on the substrate [119].

The hierarchical electrode fabricated in our study has shown potential for monitoring human health via tracking signals from biological fluids. The hb-NPG electrode demonstrated a higher sensitivity of 15.97 μA mM⁻¹ cm⁻² for glucose detection compared to the traditional NPG (6.53 μA mM⁻¹ cm⁻²). The dual-sized porosity in hb-NPG provides a large surface area and easy penetration of the glucose molecules facilitating their oxidation [120].

Although the materials with hierarchical architecture have realized great progress, there are still some challenges. For commercialization, there is a need to develop novel preparation strategies involving fewer steps and lower costs. The electrochemical performance of cathode and anode materials can be enhanced in the future by logically bringing together the strengths of compatible materials with novel architectonics. Studies focusing on solving the issue of the large-scale production and deep investigation of the in-situ formation mechanism of the hierarchical structures will interesting to work in in the future [121].

Data availability

Data will be made available on request.