RGO decorated N-doped NiCo₂O₄ hollow microspheres onto activated carbon cloth for high-performance non-enzymatic electrochemical glucose detection

Abstract

This paper reports the first effective fabrication of a high-performance non-enzymatic glucose sensor based on activated carbon cloth (ACC) coated with reduced graphene oxide (RGO) decorated N-doped urchin-like nickel cobaltite (NiCo₂O₄) hollow microspheres. Hierarchically mesoporous N-doped NiCo₂O₄ hollow microspheres were synthesized using a facile solvothermal method, followed by thermal treatment in a nitrogen (N₂) atmosphere. Subsequently, they were hydrothermally decorated with RGO nanoflakes. The resulting composite was dip-coated onto ACC, and its electrochemical and glucose sensing performances were investigated using electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV), and chronoamperometric measurements in a three-electrode system. The composite electrode sensor demonstrates admirable sensitivity (6122 μM mM⁻¹ cm⁻²) with an ultralow detection limit (5 nM, S/N = 3), and it performs well within a substantial linear range (0.5–1.450 mM). Additionally, it exhibits good long-term response stability and outstanding anti-interference performance. These outstanding results can be attributed to the synergistic effects of the highly electrically conductive ACC with multiple channels, the enhanced catalytic activity of highly porous N-doped NiCo₂O₄ hollow microspheres, and the large electroactive sites provided by its well-developed hierarchical nanostructure and RGO nanoflakes. The findings highlight the enormous potential of the ACC/N-doped NiCo₂O₄@RGO electrode for non-enzymatic glucose sensing.

1. Introduction

In the microenvironment of the human body, glucose plays a vital role as a nutrient directly involved in metabolic processes. However, elevated glucose levels can be detrimental to health, leading to conditions such as diabetes mellitus. If left uncontrolled, diabetes mellitus can pose significant risks to physical well-being and even result in fatality [1]. Consequently, accurately identifying the concentration of glucose in the blood is crucial for the diagnosis, treatment, and management of diabetes mellitus. Although traditional enzymatic glucose sensors exhibit high sensitivity and selectivity, their costliness and susceptibility to pH, temperature, and humidity constraints limit their further development [2]. As a result, non-enzymatic electrochemical glucose biosensors have garnered considerable attention in current research efforts [[3], [4], [5], [6], [7], [8], [9]].

With the advancements in nanotechnology, several non-enzymatic glucose electrodes based on transition metal oxides and other composite materials have been developed [[10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21]]. Among them, spinel NiCo₂O₄ has garnered significant interest due to its high natural abundance, low cost, non-toxicity, and environmental friendliness [22]. Additionally, the interphases of nickel (Ni) and cobalt (Co) in the spinel structure exhibit strong electrical interactions, thereby enhancing the electrocatalytic activity and chemical stability of the electrode material. However, NiCo₂O₄ also has certain limitations, such as low electrical conductivity and poor chemical/mechanical stability [23]. Therefore, there is an opportunity to enhance the glucose-sensing characteristics of NiCo₂O₄, and various approaches have been explored thus far. One such approach involves synthesizing NiCo₂O₄ in various morphologies to improve its electrochemical performance, as the surface morphology significantly affects ion transport kinetics and electrolyte permeation. In particular, hierarchical hollow nanostructures of NiCo₂O₄ have been found to significantly enhance performance in electrochemical applications. This is attributed to their large specific surface area, short diffusion routes for ions or electrons, and efficient mass transport channels [24].

Furthermore, the combination of porous carbon materials like RGO with NiCo₂O₄ has been found to enhance the electrode material's conductivity, active sites, specific surface area, and therefore its glucose sensing properties [25,26]. Additionally, doping metal oxides with heteroatom elements such as N, P, and S can increase their electrical conductivity while enhancing the intrinsic activity at each site [27,28]. Notably, the lone-pair electrons around N atoms can improve the interaction with the metal atoms, and the resulting M-O-N bonds exhibit increased electronegativity and elongated bond lengths. This reduces the attraction of Co²⁺/Co³⁺/Ni²⁺/Ni³⁺ to the 3d electrons and decreases the electron transport energy, thereby improving the reaction kinetics [29,30]. Interestingly, in the metallic oxide lattice, doping N atoms partially replace intrinsic O atoms, leading to the introduction of a significant number of oxygen defects. These defects can effectively alter the electronic structure and chemical properties of crystalline materials, as well as their electrical conductivity, by causing slight atom-to-atom disruptions and lattice deformation [31]. The synergistic effects of large multivalent metal cations and oxygen vacancies have been proven to considerably enhance the electronic conductivity and electrochemical performance of electrodes [23]. Therefore, it is crucial to consider a plausible design that incorporates doping, defects, and surface-structural engineering to improve the electrochemical characteristics and overall performance of glucose-sensing electrodes. Consequently, research focusing on high-performance glucose sensors based on activated carbon cloth coated with N-doped NiCo₂O₄ urchin-like hollow microspheres decorated with RGO nanoflakes appears necessary and has been scarcely mentioned in previous publications.

In this study, we have successfully fabricated a high-performance glucose sensor based on N-doped NiCo₂O₄@RGO composite-coated activated carbon cloth (ACC). The glucose sensor exhibits superior electrochemical performance attributed to several key factors. Firstly, the highly mesoporous structure of the composite provides a large surface area, allowing for enhanced glucose sensing capabilities. Secondly, the N-doping in NiCo₂O₄ increases the number of catalytic active sites, leading to improved electrochemical activity. Moreover, the N-doping also contributes to the modification of the electronic structure and enhancement of electrical conductivity in NiCo₂O₄. These combined effects contribute to the glucose sensor's excellent performance, including high glucose sensitivity with a very low detection limit, remarkable stability, wide detection range, and rapid response time. Furthermore, the glucose sensor exhibits exceptional selectivity for glucose compared to the majority of the primary interfering species commonly found in human serum. This selectivity ensures accurate and reliable glucose detection in complex biological samples. Overall, our findings demonstrate the significant potential of the N-doped NiCo₂O₄@RGO composite-coated ACC glucose sensor, highlighting its excellent electrochemical performance and promising application in glucose sensing.

2. Experimental details

2.1. Synthesis of N-doped -NiCo₂O₄@RGO urchin-like hollow microspheres

We employed a simple solvothermal technique to prepare NiCo₂O₄ hollow microsphere precursors. In this process, urea (10.8 g), cobalt (1.74 g), and nickel (0.87 g) nitrates were added to a solution consisting of 45 ml of isopropyl alcohol (IPA) and 9 ml of deionized water. The mixture was continuously stirred until the complete dissolution of the compounds. Subsequently, the resulting solution was transferred to a stainless steel autoclave and subjected to constant stirring for 1 h. The autoclave was then placed in an electric oven and maintained at 120 °C for 12 h during the solvothermal reaction. After completion, the autoclave was gradually cooled to room temperature. To purify the precipitates, they were retrieved from the Teflon container, transferred to conical tubes, and underwent five cycles of centrifugation using deionized water and ethyl alcohol. Subsequently, the cleaned powder slurries were dried at 80 °C for 12 h and then calcined at 350 °C for 2 h to obtain black urchin-like NiCo₂O₄ hollow microspheres (referred to as HN).

For the preparation of N-doped NiCo₂O₄ hollow microspheres, namely N₅₀–HN, N₁₀₀–HN, and N₁₅₀–HN, a similar procedure was followed. However, the NiCo₂O₄ powder samples were calcined in a vacuum tube furnace with a continuous flow of pure nitrogen gas at flow rates of 50, 100, and 150 standard cubic centimeters per minute (sccm), respectively. The calcination was carried out at a temperature of 300 °C for 2 h [32]. The obtained microspheres were then immersed in a homogeneous aqueous solution containing urea and graphene oxide (GO) in a ratio of 15:1. This mixture was placed in an autoclave vessel, and a hydrothermal reaction was conducted at 120 °C for 8 h to synthesize RGO-decorated HN, N₅₀–HN, N₁₀₀–HN, and N₁₅₀–HN microspheres, denoted as HN@R, N₅₀–HN@R, N₁₀₀–HN@R, and N₁₅₀–HN@R, respectively. Subsequently, the microspheres underwent a second drying process at 80 °C for 12 h.

2.2. Deposition of N-doped-NiCo₂O₄@RGO composites on ACC

To prepare the ACC/HN@R and ACC/N₁₀₀–HN@R electrodes, the HN@R and N₁₀₀–HN@R microspheres, along with a small amount of carbon black, were dip-coated onto an ACC substrate measuring 3 × 1 cm². In a typical procedure, 200 μL of a mixture containing DI water and IPA solution in a volume ratio of 3:1 was combined with 1 mg of the pseudocapacitive materials, 0.1 mg of carbon black (CB), and 10 μL of Nafion solution. The resulting solution was subjected to ultrasonication to produce a homogeneous black ink. The ACC substrate was then immersed in the ink for 60 s to allow for a substantial deposition of the materials. Afterward, it was dried at 60 °C for 10 min. This dip-dry process was repeated 5 times to achieve the desired coating thickness. Fig. S1 provides a schematic illustration of the synthesis process for the ACC/N₁₀₀–HN@R electrode used for glucose detection.

3. Results and discussion

3.1. Structural, surface-morphological, and porosity analysis

Fig. 1(a) illustrates the XRD patterns of the HN@R and N₁₀₀–HN@R microspheres. Comparing these patterns to previous research and the JCPDS 01-073-1702 database, it can be observed that both types of microspheres exhibit distinctive X-ray diffraction peaks that closely match the characteristic planes of cubic spinel NiCo₂O₄. It should be noted that typical diffraction peaks for RGO sheets at 25° were not observed, which could be attributed to the highly disordered structure of RGO or its low diffraction intensity [33]. The presence of sharp and distinctive key peaks in the XRD patterns indicates the well-crystallized nature of NiCo₂O₄, which is achieved through the controlled solvothermal reaction and thermal annealing processes. Furthermore, upon N-doping, some of the major XRD peaks for the HN@R microspheres exhibit a slight shift towards lower 2θ values. This shift can be attributed to the alteration in crystallographic characteristics resulting from the difference in ionic radii between N and O. According to Bragg's Law equation, the substitutional replacement of O by N leads to a change in the interplanar spacing (d-value) and a reduction in the diffraction angle of the (3 1 1) plane. Consequently, a slight lattice distortion is formed in the NiCo₂O₄ structure due to N-doping. Although the primary diffraction peak intensities are slightly reduced by N-doping, the crystallinity is only mildly affected, indicating that the overall crystalline structure remains relatively intact even after N-doping-induced lattice distortion and reorientation effects.

Fig. 1.

XRD patterns of HN@R and N₁₀₀–HN@R urchin-like microspheres.

The average crystallite size (D) of the HN@R and N₁₀₀–HN@R microspheres was calculated using Scherrer's formula [34] and found to be 11 nm and 8 nm, respectively. This calculation provides an estimate of the size of the crystalline domains in the microspheres. In addition to crystallite size, the quantity of electroactive sites on the electrode material plays a crucial role in improving glucose sensing characteristics. To gain insights into the number of defects present in the microspheres, the dislocation density (δ) was assessed using the approach proposed by Williamson and Smallman [35]. The observed δ values for the HN@R and N₁₀₀–HN@R microspheres were 0.0077 nm and 0.0173 nm, respectively. The decrease in crystallinity, as indicated by the reduction in crystallite size, and the increase in dislocation density confirm the impact of N-doping on the microspheres. The larger defects present in the N₁₀₀–HN@R microspheres may contribute to their greater catalytic activity by providing more active sites for glucose sensing. Furthermore, the smaller crystallite size in the N₁₀₀–HN@R microspheres leads to tighter packing, increased point contacts, and improved inter-particle conductivity. These factors are expected to enhance the electrochemical processes and overall performance of the N₁₀₀–HN@R microspheres as a glucose-sensing material.

Fig. 2(a–f) presents FESEM images of the HN@R and N₁₀₀–HN@R urchin-like microspheres at various magnifications. The microspheres, with a typical diameter of approximately 5–6 μm, exhibit a hollow structure and demonstrate remarkable uniformity, as shown in Fig. 2(a, d). The formation of hollow microspheres typically involves a process known as inside-out Ostwald ripening, where the outer crystalline shells mature on solid particles and undergo continuous dissolution and re-crystallization during assembly [36,37]. Fig. 2(b-c, e-f) provides enlarged views of these microspheres, revealing their urchin-like hierarchical structures with numerous surface-grown nanograsses. The nanograsses on the N₁₀₀–HN@R microspheres have a diameter of around 25 nm and a length of approximately 340 nm. In comparison to the N₁₀₀–HN@R microspheres, the interior walls of the HN@R microspheres or the roots of the nanograss appear thicker and more agglomerated. The surface nanograss structures of the N₁₀₀–HN@R microspheres, on the other hand, exhibit a more homogeneous and well-grown appearance. Consequently, the N₁₀₀–HN@R microspheres provide a significantly larger surface area with adequate porosity for electrochemical activity. Furthermore, the effective incorporation of nitrogen (N) into the spinel NiCo₂O₄ structure and the uniform distribution of all constituent elements, including Ni, Co, O, C, and N, were confirmed through the acquisition of the EDS spectrum and EDS mapping images of the N₁₀₀–HN@R microspheres, as depicted in Fig. 3(a–h). The EDS spectrum displays the relative amounts of the various elements present in the sample. Specifically, the observed weight percentages of Ni, Co, O, N, and C elements were 12.16 wt%, 24.58 wt%, 33.16 wt%, 0.56 wt%, and 29.54 wt%, respectively. Additionally, to compare the N-doping concentrations and uniform distribution of constituent elements in the N₅₀–HN@R and N₁₅₀–HN@R microspheres, the EDS spectra and EDS mapping images of these samples were also obtained, as shown in Figs. S2 and S3, respectively. These analyses provide further evidence of the successful incorporation of N and the uniform distribution of elements in the synthesized microspheres.

Fig. 2.

FESEM images of (a–c) HN@R and (d–f) N₁₀₀–HN@R urchin-like microspheres at various magnifications.

Fig. 3.

(a) EDS spectrum (b) FE-SEM image, (c) EDS layer image, and the EDS mapping of (d) O (e) Co, (f) Ni, (g) C elements and (h) N N₁₀₀–HN@R urchin-like microspheres.

In Fig. 4(a) and (b), the differential pore size distribution curves of HN@R and N₁₀₀–HN@R microspheres are presented. The N₁₀₀–HN@R microspheres exhibit significant micro- and mesoporosity, which can be attributed to their hollow structure and the presence of numerous fine nano grass-like structures on the surface, along with the uniform decoration of RGO nanoflakes. As shown in the curves, the N₁₀₀–HN@R microspheres demonstrate enhanced surface area and porosity compared to the HN@R microspheres. The surface area of N₁₀₀–HN@R microspheres is approximately 51 m²/g, with a total porosity of 82%, whereas the HN@R microspheres exhibit a surface area of about 44 m²/g. These porosimetry results are consistent with the FE-SEM observations and confirm the porous nature and improved surface area of the synthesized N₁₀₀–HN@R microspheres. The presence of a wide surface area and high porosity implies the availability of more electroactive sites, which can contribute to enhanced electrochemical activity [24]. Consequently, these properties are expected to improve the glucose-sensing capabilities of the N₁₀₀–HN@R microspheres.

Fig. 4.

Pore size distribution curves of (a) HN@R and (b) N₁₀₀–HN@R urchin-like microspheres.

Fig. 5(a) presents the XPS survey spectrum of the N₁₀₀–HN@R microspheres, providing information about the elemental composition and oxidation states on the surface. The presence of Ni, Co, O, N, and C elements in the synthesized microspheres is confirmed by the photoemission peaks' binding energies. The high-resolution XPS spectra in Fig. 5(b) show the Co 2p peaks. The peaks at 780.70 and 794.50 eV correspond to the Co³⁺ oxidation state, while the peaks at 779.42 and 794.50 eV indicate the Co²⁺ oxidation state. Additionally, the satellite peaks observed at binding energies of 796.07 and 802.80 eV are associated with Co species. The Ni 2p XPS spectra in Fig. 5(c) exhibit two peaks at binding energies of 855.28 and 872.58 eV, representing Ni 2p_1/2 and Ni 2p_3/2 of Ni²⁺ oxidation state, respectively. In Fig. 5(d), the O 1s XPS spectra are shown. The deconvolution of the spectrum reveals two oxygen peaks: O I at 529.38 eV, corresponding to metal-oxygen bonds, and O II at 531 eV, indicating oxygen vacancies [38]. It has been believed that the state of nitrogen is either a nitrogen anion [39] or atomic nitrogen atoms [40,41] based on XPS investigations of N₁₀₀–HN@R [Fig. 5(e)], which demonstrate that the N 1s core level binding energy is located at ∼395 eV. Nitrogen seems to be in the state M − N (metal-nitrogen) in substituted systems, acting as an anion's substitute. However, the valence state of this substituted species is not conclusively known. While in other accounts the valence state is not specifically stated, only the presence of an M − N bond is confirmed, it has been assumed in some cases to be a N-anion. Furthermore, additional N 1s peaks at binding energies of 400 and 403 eV are observed, which can be attributed to adsorbed organic molecules or chemisorbed molecular dinitrogen. The signal around 406 eV is associated with oxidized nitrogen [40]. The signal around 406 eV is associated with oxidized nitrogen. The deconvolution of the C 1s peaks in Fig. 5(f) reveals two main peaks at binding energies of 287.28 eV and 284.78 eV, corresponding to the C–O and C C/C–C bonds of reduced graphene oxide (RGO), respectively.

Fig. 5.

(a) Survey XPS spectra, and High-resolution XPS spectra corresponding to (b) Co 2p, (c) Ni 2p, (d) O 1s, (e) N 1s, and (f) C 1s C 1s core levels of N₁₀₀–HN@R microspheres.

3.2. Electrochemical properties and electrocatalytic activity of ACC/N₁₀₀–HN@R electrode

The EIS measurement was first performed to validate the impedance alterations during the electrode modification operations, and it was afterward utilized to examine the glucose sensing data. In Fig. 6(a), EIS measurements of ACC/HN, ACC/HN@R, and ACC/N₁₀₀–HN@R electrodes were performed in a 2 M KOH electrolyte over a frequency range of 1–100 kHz at 0.01 V. The Nyquist plots exhibit a small semicircle in the high-frequency region, representing the charge transfer resistance (R_ct) associated with faradaic processes, and a linear part in the low-frequency region, indicating capacitive behavior. The inset of Fig. 6(a) highlights that the ACC/N₁₀₀–HN@R electrode has the smallest semicircular diameter among the three electrodes, indicating a decrease in electron transfer resistance due to N-doping of NiCo₂O₄ and the presence of RGO nanoflakes on the microspheres. The initial equivalent resistance is also reduced by both N-doping and RGO decoration. Furthermore, the ACC/N₁₀₀–HN@R electrode exhibits a significantly shorter linear part with a steeper slope compared to the other electrodes, indicating improved electrolyte ion transport and enhanced capacitive behavior. This suggests that the synergistic effect of RGO nanoflakes and N-doping in NiCo₂O₄ microspheres contributes to the enhanced conductivity of the ACC/N₁₀₀–HN@R electrode. For comparison, Nyquist plots of ACC/N₅₀–HN@R and ACC/N₁₅₀–HN@R electrodes are also provided (Fig. S4), demonstrating that these electrodes exhibit lower electrochemical performance compared to the ACC/N₁₀₀–HN@R electrode. To gain a deeper understanding of the capacitive contributions and charge transfer mechanisms of the glucose sensor electrode, an equivalent circuit is fitted to the Nyquist plot of the ideal ACC/N₁₀₀–HN@R electrode, as shown in Fig. 6(b). The equivalent circuit includes an ohmic resistance (R_s) representing the combined ionic resistance of the electrolyte, interface resistance between the material and substrate, and intrinsic resistance of the current collector. The charge transfer in the surface layer of the electrode materials is characterized by the parallel connection of the R₁ and C₁ components, while the Tanh impedance, a combination of Y₀ and B, represents finite-length diffusion and blocking at the surface. The inset table of Fig. 6(b) presents the calculated parameters corresponding to the EIS Nyquist plot, and the small amount of error indicates that the fitted model accurately represents the experimental values.

Fig. 6.

(a) Electrochemical impedance spectroscopy (EIS) study (Nyquist plot) of ACC/HN, ACC/HN@R, and ACC/N₁₀₀–HN@R electrodes at 100 kHz–1 Hz under 0.01 V; (b) The fitted EIS plot for the ACC/N₁₀₀–HN@R electrode with corresponding impedance equivalent circuit. The inset table provides the values for each component of the equivalent circuit.

In addition to EIS, CV analysis was performed to investigate the electrochemical characteristics of the ACC/HN@R and ACC/N₁₀₀–HN@R electrodes in a 2 M KOH solution. Fig. 7(a) and (b) display the CV curves of the ACC/HN@R and ACC/N₁₀₀–HN@R electrodes, respectively, at various scan speeds within the potential range of 0 to +0.6 V. As the scan rate was increased while maintaining the non-rectangular shape of the CV curves, the integrated area under the curve and the current density also increased. The fact that the shape of the CV curves remains unchanged across all scan rates indicates excellent electrical conductivity and respectable rate capability for both electrodes. It is important to note that the current responses obtained for the ACC/N₁₀₀–HN@R electrode are significantly higher than those obtained for the ACC/HN@R electrode at all scan rates. Additionally, the current responses of the ACC/N₁₀₀–HN@R electrodes are superior to those of the ACC/HN, ACC/N₅₀–HN@R, and ACC/N₁₅₀–HN@R electrodes [Fig. S5 (a-c)]. This can be attributed to the enhanced porosity, nanostructure, and surface area of the ACC/N₁₀₀–HN@R electrode, which provides a larger number of redox-active sites, thereby improving electrochemical processes.

Fig. 7.

Cyclic voltammetry curves of (a) ACC/HN@R, and (b) ACC/N₁₀₀–HN@R electrodes at various scan rates under 0 to +0.6 V potential window; Cyclic voltammetry results of glucose oxidation on (c) ACC/HN@R, and (d) ACC/N₁₀₀–HN@R electrodes with a glucose concentration range of 0–3 mM in 2 M KOH at 49 mV/s; the anode peak current and the glucose concentration relationship for (e) ACC/HN@R, (f) ACC/N₁₀₀–HN@R electrodes as obtained from cyclic voltammetry curves shown in Fig. (c) and (d), respectively.

Both electrodes' CV curves exhibit a non-rectangular shape with two redox peaks, which characterize their pseudocapacitive behavior resulting from electrochemical reactions. These redox peaks are primarily attributed to the Faradaic reactions occurring within the electrode materials and the redox reactions taking place in the KOH electrolyte. The presence of oxygen functional groups in abundance on the RGO nanoflakes leads to strong oxidation reactions, displaying high redox reactivity in the positive potential range [42]. The corresponding electrochemical reactions are represented by Equation (1) and Equation (2) as follows:

NiCo₂O₄ + OH⁻ + H₂O ↔ NiOOH + 2CoOOH + e⁻ (1)

CoOOH + OH⁻ ↔ CoO₂ + H₂O + e⁻ (2)

The CV measurements were carried out at a scan rate of 49 mV/s in various concentrations of glucose (0–3 mM) added to a 0.1 M KOH electrolyte to investigate the electrocatalytic activities and mechanisms of glucose sensing on ACC/HN@R and ACC/N₁₀₀–HN@R electrodes [Fig. 7(c, d)]. Redox peaks corresponding to the Ni²⁺/Ni³⁺ and Co²⁺/Co³⁺ redox couples were observed for both electrodes, confirming the successful modification and stable deposition using the dip coating method. Noticeable increases in current response values for the anodic and cathodic peaks were observed with incremental glucose additions. The ACC/N₁₀₀–HN@R electrode exhibited significantly higher anodic current responses compared to the ACC/HN@R electrodes, indicating excellent electrochemical activity of the ACC/N₁₀₀–HN@R electrode material towards glucose. Moreover, the ACC/N₁₀₀–HN@R electrodes outperformed the ACC/HN, ACC/N₅₀–HN@R, and ACC/N₁₅₀–HN@R electrodes in terms of anodic current responses to glucose [Fig. S5 (d-f)]. The synergistic activity of Ni²⁺, Co²⁺, and Co³⁺ ions in the NiCo₂O₄ spinel structure, along with the redox couples present in NiCo₂O₄, serve as important active sites for glucose oxidation in alkaline media such as KOH aqueous electrolyte. The increased oxidation current of glucose on ACC/N₁₀₀–HN@R electrodes confirms the large electrochemically active area of N₁₀₀–HN@R microspheres, providing more metal ion active sites for the glucose oxidation reaction. Thus, the glucose oxidation process can be attributed to the inherent property of the spinel NiCo₂O₄. The electrochemical catalytic activity of NiCo₂O₄ for glucose oxidation [[43], [44], [45]] can be described as follows:

Co (IV) + Ni (III) + glucose ↔ Co (III) + Ni (II) + gluconolactone (3)

The species of NiCo₂O₄ on the electrode surface transform NiOOH and CoOOH through an oxidation reaction during the anodic scan [equation (1)]. Subsequently, the CoOOH species is converted into CoO₂ on the electrode equation (2)]. Upon the addition of glucose to the 0.1 M KOH solution, the NiOOH and CoOOH species catalyze the oxidation of glucose, leading to the production of gluconolactone [equation (3)], which increases the oxidation current. Additionally, anodic peak potential shifts towards a positive direction with an increase in the glucose concentration, indicating electrochemical polarization between the adsorbed glucose species and the oxidized metal ion intermediates on the electrode's active sites. This phenomenon is beneficial for enhancing the oxidation of high concentrations of glucose analyte. Furthermore, Fig. 7(e, f) demonstrates the relationship between the anode peak currents and glucose concentrations, with correlation coefficients (R²) of 0.9923 and 0.9384 for ACC/HN@R and ACC/N₁₀₀–HN@R electrodes, respectively. These fitting curves exhibit a linear relationship between the response current and glucose concentration within the range of 0–3 mM. Conversely, the linear relationship between the anode peak currents and glucose concentrations for the ACC/HN, ACC/N₅₀–HN@R, and ACC/N₁₅₀–HN@R electrodes is relatively inadequate, as depicted in Fig. S6.

It is widely recognized that the initial potential set during amperometric measurements significantly affects the response currents [46]. Therefore, the performance of the ACC/HN@R and ACC/N₁₀₀–HN@R electrodes was assessed using chronoamperometric measurements following the repeated addition of glucose solution (28 μM) to the 0.1 M KOH solution. Various potentials were investigated, and the characteristic current vs. time plot was recorded while the electrolyte solution was continuously stirred at approximately 200 rpm to ensure rapid and uniform dispersion of the injected glucose solution. As observed in Fig. 8(a–h), following each injection of glucose solution within the range of 28–168 M, the oxidation currents for both electrodes exhibited an increase, resulting in staircase-like curves. An average response time of approximately 10 s was recorded. Furthermore, in comparison to the ACC/HN@R electrode, the oxidation current for the ACC/N₁₀₀–HN@R electrode exhibited a significant increase with the increase in glucose concentration. As shown in Fig. 8(a, e), the response currents for both electrodes were relatively low at 0.45 V. However, at a potential of 0.60 V [Fig. 8(d, h)], the noise amplitudes were considerably large, disrupting the oxidative current response and severely limiting glucose sensing capabilities [47]. Moreover, it was observed that under inappropriate potential settings, the glucose concentration had a significant impact on noise amplitudes. With successive injections of 28 μM glucose solutions, the current responses were found to be higher with minimal noise amplitudes at the initial potentials of 0.50 V and 0.55 V for both electrodes [Fig. 8(b, c, f, g)]. Fig. 8(g) demonstrates that a potential of 0.55 V results in a well-defined staircase curve with a maximum oxidation current response and excellent stability for the ACC/N₁₀₀–HN@R electrode. The chronoamperometric response of the ACC/N₁₀₀–HN@R electrode was also superior to that of the ACC/HN, ACC/N₅₀–HN@R, and ACC/N₁₅₀–HN@R electrodes [Fig. S7 (a-c)]. Therefore, the appropriate potential for chronoamperometric detection of glucose using an ACC/N₁₀₀–HN@R electrode sensor was determined to be 0.55 V. Consequently, 0.55 V was selected as the ideal starting potential for all future chronoamperometric measurements.

Fig. 8.

The amperometric response of the electrodes (a–d) ACC/HN@R and (e–h) ACC/N₁₀₀–HN@R to the successive injections of 28 μM glucose into 0.1 M KOH solution at various potentials in the range of 0.40–0.60 V. (Rotation speed = 200 rpm).

Fig. 9(a, b) shows the equivalent linear fit curves of the steady-state current response (A) vs. glucose concentration (mM) for the two electrodes. The observed curves produced for both electrodes at various applied potentials exhibit varied slopes (S), indicating differing sensitivity to glucose, within the examined glucose concentration range of 28–168 M. The slope values at 0.50 and 0.55 V are substantial. However, with a starting potential of 0.55 V, the ACC/N₁₀₀–HN@R electrode sensor has a maximum sensitivity of 6122 μM mM⁻¹ cm⁻² (R² = 0.9985). According to the calculation, glucose has a LOD of 5 nM (S/N = 3). In comparison to the ACC/N₁₀₀–HN@R electrode sensor, the sensitivities calculated for the ACC/HN, ACC/N₅₀–HN@R, and ACC/N₁₅₀–HN@R electrodes from their linear fit curves are found to be lower [Fig. S6(d-f)]. Similar to this, chronoamperometric responses in the form of staircase-like curves were acquired corresponding to the successive additions of low and high-concentration glucose solutions into the 0.1 M KOH solution as shown in Fig. 10 (a, b) to demonstrate the broad range of glucose detection by ACC/N₁₀₀–HN@R electrode sensor. The resulting calibration curves for additions of low and high concentrations of glucose are shown in Fig. 10 (c, d). As can be observed, when glucose is successively injected with a step of 1 μM concentration within the range of 1–20 μM, the sensor reacts with excellent linearity and a correlation value of 0.9988. Whereas following the addition of glucose at a concentration step of 50 μM, it exhibits two distinct linear sections in the ranges of 50–450 μM and 500–1450 μM, respectively. The detection limit is therefore estimated to be 0.29 nM (S/N = 3) for the glucose concentration range of 1–20 μM, and the sensitivity of the glucose sensor is 8871 μM mM⁻¹ cm⁻² at 0.55 V. The detection limit (82 nM) and sensitivity (1612 μM mM⁻¹ cm⁻²) were observed to be reasonable for a larger glucose concentration range i.e. 500-1450 μM.

Fig. 9.

Linear fit curve of glucose concentration (mM) vs. steady-state current response (μA) for (a) ACC/HN@R and ACC/N₁₀₀–HN@R electrodes; E_app = 0.40–0.60 V; Rotation speed = 200 rpm. Both plots indicate that 0.55 V is the optimum starting potential with which the electrode sensors have maximum sensitivity.

Fig. 10.

(a, b) The amperometric response and respective (c, d) linear fit curves of concentration of glucose (mM) vs. steady-state current response (μA) of ACC/N₁₀₀–HN@R electrode sensor at E_app = 0.55 V for (a) low and (b) high concentration glucose injections into 0.1 M KOH solution (Rotation speed = 200 rpm).

The catalytic performance of the non-enzymatic glucose sensor was compared to previously published NiCo₂O₄-based glucose sensors, as shown in Table 1 [25,26,46,[48], [49], [50], [51], [52]]. The ACC/N₁₀₀–HN@R electrode sensor exhibited exceptional glucose-sensing ability, characterized by its remarkable sensitivity, low detection limit, and wide linear range. The sensor demonstrated a 10-s reaction time for glucose detection, which is comparable to some of the sensors mentioned in the literature. The improved intrinsic activity observed in NiCo₂O₄ may be attributed to enhancements in its electronic structure, evident through lattice distortion and increased electronic conductivity, which likely contribute to the enhanced sensing performance. Furthermore, the incorporation of RGO nanoflakes onto NiCo₂O₄ microspheres improved the material's electrical conductivity and provided additional active sites for catalytic redox processes. The highly porous hierarchical nanostructure of the sensor material, with its large effective surface area, facilitated short diffusion paths for ions or electrons and provided an abundance of electroactive sites for redox reactions, contributing to the excellent electrocatalytic sensing performance of the glucose sensor. The highly conductive ACC substrate not only served as a flexible support but also offered numerous redox-active sites and diverse channels for improved electrolyte ion transport, further enhancing the glucose sensing capability.

Table 1.

Performance comparison of ACC/N₁₀₀–HN@R electrode glucose sensor with previously reported NiCo₂O₄ materials-based glucose sensors.

Sensor electrode	Sensitivity (μM∙mM−1∙cm−2)	Limit of detection (μM)	Linear range	Reference
GO-templated NiCo2O4 nanosheets	792.72	0.28	0.00015–8.86 mM	[25]
NiCo2O4/RGO hollow nanospheres	2082.57	0.7	0.04–1.28 mM	[26]
NiCo2O4 hollow nanorods	1685.1	0.16	0.0003–1.0 mM	[46]
NiCoO2/C/GCE	549.3	0.5	0.02–2.41 mM	[48]
Urchin-like NiCo2O4	72.4	0.37	0.00037–2 mM	[49]
NiCo2O4 hollow nanospheres	1917	0.6	0.1–0.3 mM 0.3–2.24 mM	[50]
Graphene wrapped Co3O4/NiCo2O4 hollow nanocages	304	0.384	0.1–3.52 mM	[51]
NiCo2O4 nanowrinkles-RGO	38.75	38.75	0.005–8.56	[52]
ACC/NiCo2O4@RGO urchin-like microspheres	8871 6122 3796 1612	0.00029 0.005 0.062 0.0682	0.001–0.02 mM 0.028–0.168 mM 0.05–0.450 mM 0.5–1.450 mM	This work

One of the most critical analytical parameters for an amperometric biosensor is its ability to selectively detect and distinguish the target analyte from interfering species [37]. Human serum contains various interfering species such as urea, dopamine, ascorbic acid, NaCl, sucrose, and uric acid, which can interfere with the direct electrochemical oxidation of glucose due to their similar electrochemical processes. To assess the selectivity of the synthesized ACC/N₁₀₀–HN@R electrode sensor towards glucose, solutions of these interfering species were sequentially injected into the 0.1 M KOH electrolyte solution at two different concentrations: 7 and 28 M. Chronoamperometric measurements were then performed at an initial potential of 0.55 V, and the results of the interference test are presented in Fig. 11 (a, b). These results demonstrate that none of the other analyte species significantly affected the detection of glucose. Therefore, the mesoporous ACC/N₁₀₀–HN@R electrode sensor exhibits exclusive glucose detection in 0.1 M KOH solution, with maximum sensor response, even in the presence of interfering species commonly found in human blood.

Fig. 11.

Interference test of the ACC/N₁₀₀–HN@R electrode sensor in 0.1 M KOH with different concentrations of glucose and other interfering species as indicated.

Fig. 12 illustrates the amperometric response of the ACC/N₁₀₀–HN@R electrode sensor to injections of 28 M glucose in 0.1 M KOH solution for 18 days, aiming to evaluate its performance stability. The inset pictures depict the glucose-sensing responses on the first and last days. The results indicate that there is no significant decrease in response after the initial eight days, and even on the final day, the sensor still retains approximately 85% of its initial response. This demonstrates the acceptable electrochemical stability of the sensor, which can be attributed to the spinel structure of the NiCo₂O₄ and the mesoporous hierarchical nanostructure of the sensor material.

Fig. 12.

Glucose sensing response stability of ACC/N₁₀₀–HN@R electrode sensor at E_app = 0.55 V for 18 days. Inset images show the first day and final day response against the 28 μM glucose injection into 0.1 M KOH solution.

We conducted XRD analysis on the optimized ACC/N₁₀₀–HN@R electrode before and after multiple electrochemical tests (more than eight times). Fig. S8 presents the obtained results. The XRD pattern of the ACC/N₁₀₀–HN@R electrode shows distinct peaks corresponding to NiCo₂O₄, RGO, Nafion, and carbon cloth. However, after performing repeated electrochemical measurements on the electrode, the peak associated with the carbon cloth becomes more prominent, while the peaks associated with NiCo₂O₄, RGO, and Nafion appear to weaken. This indicates that prolonged exposure to the KOH electrolyte during the electrochemical experiments has led to some degradation of the electrode materials, particularly N₁₀₀–HN@R. To enhance the stability of the electrode, it is suggested to increase the number of dip-coating cycles and/or optimize the Nafion binder content in the initial ink used for the dip-coating process. These modifications may improve the electrode's durability and performance over extended periods of use.

4. Conclusions

In conclusion, we successfully developed a non-enzymatic glucose sensor by coating activated carbon cloth with hierarchically mesoporous N-doped NiCo₂O₄@RGO hollow microspheres using solvothermal, hydrothermal, and dip-coating methods. The introduction of N-doping in NiCo₂O₄ resulted in defect formation, reduced crystallite size, increased conductivity, and the creation of numerous active sites. Further N-doping facilitated the formation of well-developed and uniform hollow microspheres, leading to a significantly increased surface area with sufficient porosity for enhanced electrochemical activity. The decoration of RGO on N-doped NiCo₂O₄ hollow microspheres further improved conductivity and added more electroactive sites. The fabricated ACC/N₁₀₀–HN@R electrode exhibited excellent performance as a glucose sensor, with a maximum sensitivity of 8871 μM mM⁻¹ cm⁻² in the glucose concentration range of 0.001–0.02 mM, a low limit of detection (LOD) of 0.29 nM, and fast response time of approximately 10 s. The sensor also demonstrated remarkable sensitivities and LODs across various high glucose concentration ranges, indicating its capability for reliable glucose detection over a wide linear range. Moreover, the stability tests revealed the promising performance of the electrode sensor over 18 days, and the selectivity tests demonstrated its ability to selectively detect glucose without significant interference from other analyte species commonly found in human serum. Overall, our findings highlight the potential of the composite material consisting of urchin-like N-doped NiCo₂O₄ decorated with RGO nanoflakes on activated carbon cloth as a practical solution for the development of non-enzymatic glucose sensors.

Author contribution statement

Prashant Shivaji Shewale: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Wrote the paper.

Kwang-Seok Yun: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools, or data; Wrote the paper.

Data availability statement

The data that has been used is confidential.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Biographies

Prashant Shivaji Shewale graduated from Shivaji University, Kolhapur, India, with a B·Sc. in Physics in 2006, an M.S. in Physics in 2008, and a Ph.D. in Physics in 2013. In Busan, South Korea's Dong-Eui University, he worked as a post-doctoral researcher from 2013 to 2018. He began working as a research professor in the Department of electronic engineering at Sogang University in Seoul, Korea, in January 2019. He has so far published 25 articles in international journals as well as one book. His current research focuses on energy storage/conversion devices, electrochemical sensors, UV photodetectors, low-temperature nanocomposite-based gas sensors, MEMS, and integrated systems.

Kwang-Seok Yun graduated with a B.S. in Electronics Engineering from Kyungpook National University in 1996, an M.S. in Electrical Engineering from Korea Advanced Institute of Science and Technology (KAIST) in 1997, and a Ph.D. in Electrical Engineering from the same institution in 2002. From 2005 to 2007, he worked as a postdoctoral researcher at the University of California, Los Angeles. In March 2007, he began teaching in the Department of Electronic Engineering at Sogang University in Seoul, Korea, where he is presently a Professor. Micro transducers, sensors, actuators, MEMS, as well as signal/energy storage/conversion devices, circuits, and systems, are among his current research interests.

Appendix A. Supplementary data

The following is the Supplementary data to this article: