Tetraethoxysilane (TEOS), cetyltrimethylammonium bromide (CTAB) (3-glycidyloxypropyl) trimethoxysilane (APTES), potassium ferricyanide (K₃[Fe(CN)₆]), potassium ferrocyanide (K₄[Fe(CN)₆]), potassium hydrogen phthalate (KHP), norepinephrine (NE), bovine serum albumin (BSA), ascorbic acid (AA), uric acid (UA), folic acid (FA) were obtained from Aladdin (Shanghai, China). L-Cystine and sodium phosphate dibasic dodecahydrate (Na₂HPO₄.12H₂O) were purchased from Macklin (Shanghai, China). Hydrochloric acid was obtained from Hangzhou Shuanglin Chemical reagent (Hangzhou, China). Ethanol, calcium chloride (CaCl₂), potassium chloride (KCl), sodium chloride (NaCl), Zinc chloride(ZnCl₂), oleic acid (OA) were provided from Hangzhou Gaojing Fine Chemical Industry (Hangzhou, China). Whole human blood (healthy man) was provided by Hangzhou Occupational Disease Prevention and Control Institute (Hangzhou, China) for real sample analysis. Phosphate buffer solution (PBS) was prepared using Na₂HPO₄ and NaH₂PO₄ in a certain proportion. All chemicals and reagents are used directly without further purification. Ultrapure water (18.2 MΩ cm) was used throughout the work.

Electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV) and differential pulse voltammetry (DPV) measurements were performed on an Autolab PGSTAT302N electrochemical workstation (Metrohm, Switzerland) at room temperature. A three-electrode system was used for the above electrochemical investigation with Ag/AgCl as the reference electrode (RE), platinum electrode (1 cm × 1 cm) as the counter electrode (CE), bare or modified GCE as the working electrode (WE). For DPV measurements, the step, modulation amplitude, modulation time and interval time were 0.005 V, 0.05 V, 0.05 s, and 0.2 s, respectively. The morphology of nanochannel modified electrode were characterized by scanning electron microscope (SEM equipped with X-ray energy dispersive spectroscopy-EDX, S-4800, Hitachi, Japan). Au nanomaterials obtained after chemical etching NH₂-VMSF by NaOH (1M) was also characterized. The images were obtained at 5 kV.

Firstly, the supporting GCE (d = 3 mm) was successively polished with 0.5 μm, 0.3 μm, and 0.05 μm alumina power. Then the electrode was thoroughly cleaned by sonication in ethanol and ultrapure water, respectively. The cleaned GCE were then treated by electrochemical polarization. Specifically, anodization was performed by applying a constant potential (+1.8 V) on GCE for 300 s. Then, cathodic polarization was performed using cyclic voltammetry scan ranged from −1.3 V to 1.25 V for 3 cycles in PBS (0.1 M, pH 5). The pre-activated electrode was denoted as p-GCE.

VMSF was grown on p-GCE by electrochemically assisted self-assembly (EASA) method (Huang et al., 2022; Deng et al., 2023). Firstly, the precursor solution was prepared. After 20 mL ethanol solution and 20 mL sodium nitrate (NaNO₃) solution was mixed, CTAB (1.585 g) and APTES (0.318 mL) was added and the pH was adjusted to 3.0 with HCl. Then, TEOS (2.732 mL) was added and the mixture was stirred at room temperature for 2.5 h. When the three-electrode system was immersed in the precursor solution, a constant current density (−0.74 mA/cm2) was applied on p-GCE for 10 s. Then, the modified GCE was quickly taken out and thoroughly washed in ultrapure water followed with aging at 80°C overnight. The as-prepared electrodes contained CTAB micelle (SM) was named as SM@NH2-VMSF/p-GCE. Then, SM@NH2-VMSF/p-GCE was put into 0.1M HCl-ethanol solution and stirred for 5 min to remove SM to obtain electrode with open nanochannel array (NH2-VMSF/p-GCE). To further electrodeposit AuNPs on nanochannel array, NH2-VMSF/p-GCE was put into chloroauric acid (HAuCl4, 0.5%) solution and deposited at 0.5 V for different time. The obtained electrode was denoted as AuNPs@NH₂-VMSF/p-GCE.

NE was electrochemical detected in PBS (0.1 M, pH = 5) electrolyte. CV or DPV method was used to record the electrochemical signals of different concentrations of NE. To investigate the application of the sensor in real sample analysis, detection of NE in human whole blood was investigated using standard addition method. Briefly, human whole blood with artificially added NE was diluted by a factor of 50 with the electrolyte. Then, NE was detected using AuNPs@NH2-VMSF/p-GCE.

Figure 1 illustrates the modification of VMSF on the supporting electrode and the following electrodeposition of Au nanoparticles (AuNPs) in nanochannel array. As shown, glassy carbon electrode (GCE) is selected as the underlying electrode. As known, GCE is the most widely used electrode in electrochemical sensors. As a carbonaceous electrode, GCE has unique characteristics of high chemical stability, wide potential window and good biocompatibility. However, VMSF cannot be directly and stably combined with GCE. As illustrated in Figure 1, pre-activated GCE (p-GCE) is firstly obtained through simple electrochemical polarization and stable modification of VMSF on p-GCE can be achieved without introducing an adhesive layer. This is due to the introduction of rich oxygen-containing groups on the electrode surface by electrochemical polarization. As known, electrochemical polarization is a simple and green method to prepare highly active carbon-based electrodes with no need of complex chemical reagents and tedious operations. Generally, GCE is successively anodized through electrochemical oxidation at high potential and cathodized by electrochemical reduction at low potential. Amongst, the anodizing process at high potential produces rich defects and oxygen containing functional groups through etching GCE surface. The subsequent electrochemical reduction will partially reduce the oxygen containing groups and restore the sp² structure of the electrode surface, leading to the restore of the conductivity of the electrode. VMSF was then grown on p-GCE by EASA method. This method can be used to grow VMSF simply and quickly (usually 5–30 s). In the growth process, the negative charge of p-GCE resulting from the ionization of oxygen groups is beneficial to the electrostatic adsorption of cationic surfactant (CTAB) micelle (SM). At the same time, the -OH groups on p-GCE can form Si-O covalent bonds with VMSF through co-condensation reaction with Si-OH groups, thus enhancing the adhesion of the film. Due to the use of functional siloxane precursor containing amino groups, the obtained VMSF is rich in amino groups and denoted as NH₂-VMSF. Finally, NH₂-VMSF/p-GCE with open nano channels was obtained by removing CTAB micelles in HCl ethanol solution.

The conventional standard electrochemical probe Fe(CN)₆ ^3/4- was used to characterize the electrochemical characteristics of each electrode in the NH₂-VMSF modification process. Figure 2A exhibited the cyclic voltammograms (CVs) obtained on p-GCE, SM@NH₂-VMSF/p-GCE and NH₂-VMSF/p-GCE. As a highly active carbon electrode, p-GEC displayed remarkable redox peak of Fe(CN)₆ ^3/4-. When the nanochannel is closed by micelles, Fe(CN)₆ ^3/4- cannot reach the surface of the underlying electrode. Thus, no obvious Faraday electrochemical signal was observed on SM@NH₂-VMSF/p-GCE in addition to the charging current. As NH₂-VMSF/p-GCE was rich in amino group, which had electrostatic attraction towards Fe(CN)₆ ^3/4-, the peak current of the probe on NH₂-VMSF/p-GCE was higher than that measured on p-GCE. This phenomenon indicates that NH₂-VMSF/ITO can promote the electrochemical response of the anion electrochemical probe.

As shown in Figure 1, gold nanoparticles (AuNPs) are in-situ deposited into the nano channel array by electrodeposition. It is well known that electrodeposition is a green and controllable method to prepare AuNPs with high performance. The amino groups in the nanochannels can be used as anchoring sites to increase the stability of the deposited AuNPs. The chemical composition of the modified film on electrode was characterized by X-ray photoelectron spectroscopy (XPS). As shown in Supplementary Figure S1 (Supporting information, SI), the XPS survey spectrum of NH₂-VMSF/p-GCE contained distinct peaks corresponding to Si, C, O, and N elements resulting from the silica structure and the amine groups of VMSF. In comparison, new signal of Au_4f appears after the successful deposition of AuNPs on NH₂-VMSF/p-GCE. Figure 2B was the representative CVs of NH₂-VMSF/p-GCE and AuNPs@NH₂-VMSF/p-GCE recorded in H₂SO₄ (0.1 M). As seen, no obvious redox peaks were observed on NH₂-VMSF/p-GCE. On the contrary, peak appeared corresponding to the electrochemical oxidation of gold at the forward sweep and the reductive peak at ∼0.86 V was also observed due to the reduction of gold oxide at the reverse sweep, proving the successful deposition of AuNPs (Hou et al., 2021). Compared with NH₂-VMSF/p-GCE, AuNPs@NH₂-VMSF/p-GCE exhibited slightly decrease peak current of Fe(CN)₆ ^3/4- (Figure 2A), which may be attributed to the fact that AuNPs use amino as the anchoring target, reducing the electrostatic effect of nano channels on Fe(CN)₆ ^3/4-. Supplementary Figure S2 was the top-view SEM images of NH₂-VMSF/p-GCE and AuNPs@NH₂-VMSF/p-GCE at different magnification. As shown, the electrode surface before and after the deposition of AuNPs was smooth, indicating that the deposition of AuNPs occurred in the nanochannels. In addition, Au nanomaterials with an average diameter of about 230 nm was observed when NH₂-VMSF was chemically etched away from AuNPs@NH₂-VMSF/p-GCE using NaOH (Supplementary Figure S3). The corresponding element mapping image indicated the characteristic Au signal. The large size of the Au nanomaterials mainly resulted from the aggregation of AuNPs without the spatial confinement of NH₂-VMSF.

Figure 3 shows the comparison of CV and DPV of NE detected by GCE, NH₂-VMSF/p-GCE and AuNPs@NH₂-VMSF/p-GCE electrodes. Compared with NH₂-VMSF/p-GCE, AuNPs@NH₂-VMSF exhibited a cyclic voltammetric curve with an increased charging current, indicating the increased electroactive area after deposition of AuNPs (Figure 3A). (Ashraf et al., 2022) As seen, the redox peak of NE on NH₂-VMSF/p-GCE was slightly lower than that obtained on p-GCE resulting from the decreased electrode area after growth of nanochannel film. On the contrary, AuNPs@NH₂-VMSF/p-GCE showed the highest Faraday current after AuNPs were deposited into the nanochannel array (Figures 3A, B). Thus, the introduction of nanochannels and AuNPs can improve the electrochemical signal of NE. The phenomenon might be ascribed to the adsorption capacity and strong attraction of AuNPs towards NE. Compared with p-GCE, the electrochemical oxidation peak potential of NE on AuNPs@NH₂-VMSF/p-GCE had a slight negative shift, indicating electrocatalytic ability of AuNPs.

To achieve the highest sensitivity for NE detection, the detection conditions were optimized. Figure 4A showed the effect of electrodeposition time for the formation of AuNPs on the electrochemical signal of NE. As seen, the peak current of NE decreased with the increase of electrodeposition time. Although AuNPs can improve the electrochemical signal of NE, the size of gold nanoparticles formed by too long deposition time was larger, which may hinder the diffusion of NE to the underlying electrode. Thus, the electrodeposition time was selected as 1 s in the subsequent experiment. The influence of pH on NE detection was also investigated.

Figure 4B showed the DPV curves of NE at different pH. As seen, the oxidation peak of NE moved towards the negative potential with the increase of pH (from 4 to 8), indicating that protons participate in the electrochemical oxidation and reduction process of NE. As shown in inset of Figure 4B, the oxidation peak potential (E) of NE had a good linear relationship with pH (E = −0.058 pH + 0.660, R ² = 0.997), and the slopes were close to the slope of Nernst equation of 59 mV/pH, indicating that the number of protons involved in the electrochemical oxidation and reduction process of NE was the same as the number of electrons. Comparing the DPV oxidation peak current measured at different pH values, NE had the largest signal when pH was 5. Thus, pH 5 was selected as the optimal pH for NE detection. Figure 4C showed the CVs of NE under different scan rate (v). It can be seen that the peak current increased with the increase of scan rate, but the peak potential was almost unchanged, indicating high electron transfer rate. The oxidation peak current (I _a) and reduction peak current (I _c) were linear with the scan rate (I _a = 58.1 v—0.954, R ² = 0.996; I _c = −38.2 v + 0.386, R ² = 0.995), indicating that an adsorption-controlled electrochemical process of NE. Figure 4D showed the oxidation peak current of NE at different enrichment times. When the enrichment time increased, the peak current increased, and then reached a stable platform. Thus, 10 s enrichment was used in further experiments.

Electrochemical sensing is attractive because of advantages of low cost, fast response, easy operation, and potential for miniaturization and integration (Gao et al., 2012; Samdani et al., 2017). Under the optimized conditions, electrochemical detection of NE was investigated using DPV. Figure 5A displays the DPV curves obtained on AuNPs@NH₂-VMSF/p-GCE in presence of different concentrations of NE. As shown, the peak current increased gradually with the increase of NE concentration. The peak current (I) was linearly correlated to the concentration of NE (C) in the range from 50 nM to 2 μM (I = 1.44C + 0.027, R ² = 0.994) and from 2 μM to 50 μM (I = 0.76C + 1.28, R ² = 0.998) (Figure 5B). The limit of detection (LOD) was calculated to be 10 nM using a signal-to-noise ratio of 3 (S/N = 3). The high sensitivity is described to the signal amplification resulting from the gold nanoparticles. Supplementary Table S1 (SI) demonstrates the comparison of determination of NE using different electrodes. As shown, the LOD is lower than that obtained from molybdenum trioxide nanowires modified GCE (MoO₃ NWs/GCE), (Samdani et al., 2017), silica/titania material containing gold nanoparticles stabilized by chitosan modified carbon paste electrode (SiTi/AuNP/CPE) (Morawski et al., 2021), Sulphur nanodots decorated graphene oxide modified GCE (nS@GO/GCE) (Jaiswal and Tiwari, 2021), molecularly imprinted polymer (MIP) functionalized single-walled carbon nanotubes modified GCE (MIP-modified SWNTs/GCE) (Wang et al., 2017), 6-amino-4-(3,4-dihydroxyphenyl)-3-methyl-1,4-dihydropyrano[2,3-c]pyrazole-5-carbonitrile (ADPC) assisted Fe₂O₃@CeO₂ coreshell nanoparticles (CNs) modified CPE (ADPC/Fe₂O₃@CeO₂/GCE), (Mazloum‐Ardakani et al., 2020), MIP-coated palladium nanoparticles modified GCE (MIP-coated PdNPs/GCE) (Chen et al., 2015).

The selectivity for the detection of NE is investigated by measuring the DPV peak current in the absence (I₀) and presence (I) of the possible interferences. Figure 6A displayed the current ratio (I/I₀) obtained from AuNPs@NH₂-VMSF/p-GCE for detection of NE (5.0 μM) in the absence and presence of 10-fold of K⁺, Ca²⁺, Na⁺, Zn²⁺, AA, UA, FA, OA, Glu, Lcy or 10 mg/mL of other added interfering species. When one of the above substances with a concentration of 10 times of NE was added, the current ratio on AuNPs@NH₂-VMSF/p-GCE was close to 1.0, indicating that the peak current was basically unchanged. This shows that the sensor has high selectivity in detecting NE. To evaluate the inter-electrode reproductivity, five electrodes were fabricated in parallel. When NE (5.0 μM) was detected, a relative standard deviation (RSD) of the current value was 2.7%, indicating high reproductivity. The long-time stability of the fabricated sensors was investigated when the electroces are stored at 4°C in a refrigerator. After 7-day storage, the peak current for determination of NE (5.0 μM) remained 91% of that obtained on the first day, indicating high stability.

The fabricated AuNPs@NH₂-VMSF/p-GCE sensor can be easily regenerated by stirring in HCl-ethanol solution for 5 min to remove the residual NE. Thus, the reuse performance for the detection of NE was also investigated. Firstly, the peak current of NE (10 μM) on AuNPs@NH₂-VMSF/p-GCE was recorded, and then the sensor was regenerated. The electrochemical signal on the regenerated electrode was then measured in the supporting electrolyte. Then, the regenerated sensor was employed for the detection of NE again. Figure 6B showed the measured signal in NE using the original or regenerated elecrode in NE or the suppoting electrolyte. As shown, the peak current of NE on the original or regenerated AuNPs@NH₂-VMSF/p-GCE sensor was quite similar. At the same time, the peak current obtained on the regenerated electrode in the supporting electrolyte was remarkably small, indicating that there was almost no NE residue on the electrode after regeneration. Thus, the AuNPs@NH₂-VMSF/p-GCE sensor can be reused for the electrochemical detection of NE.

As known, the whole blood has complex matrix including blood cells (e.g., red cells, white cells and platelets), various electrolytes, fibrinogen, coagulation factors, proteins, etc. Commonly, direct electrochemical analysis in whole blood is challenging because of the easy fouling of the electrode. As the ultrasmall VMSF nanochannel has size exclusion effect, VMSF based electrode can exclude large sized substances such as red blood cells, showing anti-fouling effect. Therefore, VMSF modified electrode has great potential in direct electroanalysis of complex samples. In order to investigate the detection ability of the constructed AuNPs@NH₂-VMSF/p-GCE sensor in complex samples, the standard addition method was used to investigate the NE detection performance in whole blood. Supplementary Figure S4 (SI) displayed the DPV curves obtained when different concentrations of NE was added in the diluted whole blood (by a factor of 50). Each concentration of NE was detected 3 times. As shown in Table 1, the recovery ranges from 99.9% to 102.2%, indicating high reliability in real sample analysis. In addition, the RSD for three measurement is less than 2.9%, suggesting good reproducibility during repeated detection.