An ultrasensitive high-performance baicalin sensor based on C₃N₄-SWCNTs/reduced graphene oxide/cyclodextrin metal-organic framework nanocomposite

Abstract

Baicalin (Bn) obtained from natural plants has been found to exhibit significant antiviral activity against Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). Herein, a novel ultrasensitive Bn electrochemical sensor was proposed based on graphitized carbon-nitride - single-walled carbon nanotube nanocomposites (C₃N₄-SWCNTs), reduced graphene oxide (rGO) and electrodeposited cyclodextrin-metal organic framework (CD-MOF). The sensing nanomaterials were characterized by X-ray diffraction spectroscopy, X-ray photoelectron spectroscopy, Raman spectroscopy, scanning electron microscopy, and transmission electron microscopy. Under optimal conditions, the sensor exhibited sensitive detection of Bn in a wide linear range of 1 × 10^-9–5 × 10^-7 M with an LOD of 4.6 × 10^-10 M and a sensitivity of 220 A/M, and it showed satisfactory stability and accuracy for detecting Bn in real samples (human serum and bear bile scutellaria eye drops). In addition, the electrochemical reaction sites and redox mechanism of Bn were revealed through electrochemical behavior and density functional theory. This work provided an insightful solution for detecting Bn, and extensive potential applications could be further expected.

Graphical Abstract

1. Introduction

Flavonoids, a class of polyphenolic compounds that can reduce the risk of cardiovascular disease and cancer, have aroused great interest of researchers [1], [2]. Among these, baicalin (Bn) is an important biologically active flavonoid extracted from Scutellaria baicalensis Georgi [3], which exhibits versatile physiological activities such as anti-inflammatory, antioxidant, anti-cancer, anti-microbial and anti-fungal activities [4], [5]. Therefore, as a medicine, Bn has been widely used to treat diphtheria, viral hepatitis, nephritis, upper respiratory tract infections, dysentery, scarlet fever, and prevent cerebrovascular dysfunction [6], [7]. Notably, recent researches have shown that Bn exhibits excellent antiviral activity against Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), which provides a great opportunity for the development of much-needed drugs for treating Coronavirus Disease 2019 (COVID-19) [8]. However, the amount of Bn must be strictly controlled, otherwise excessive Bn will cause some adverse physiological reactions such as muscle soreness, low-grade fever, diarrhea, and leukopenia [9]. Therefore, it is necessary and urgent to develop accurate, sensitive and innovative methods for detecting Bn. Currently, the quantitative detection of Bn mainly relies on traditional methods such as HPIC [10] and LC-MS [11], but these methods have many disadvantages including high cost, complex operation and cumbersome pre-treatment processes. Electrochemical techniques have become an attractive method due to their low cost and high sensitivity [12]. Moreover, the sensitivity of the electrochemical sensor can be greatly improved by synthesizing and using different new materials to modify the electrode surface. Therefore, a constructed ultrasensitive Bn electrochemical sensor has good application prospects.

Single-walled carbon nanotubes (SWCNTs), an excellent charge conductor and charge semiconductor, have the advantages of good mechanical strength, good chemical stability, and large specific surface area [13], [14]. Graphene has high electronic conductivity, excellent electrocatalytic activity, and a wide range of applications [15], [16]. Similarly, graphite carbon nitride (C₃N₄) is a soft material with a two-dimensional graphene-like structure, and it is considered as a promising material due to its simple synthesis, low cost, abundant nitrogen content, and many active sites [17]. Due to their great performance, these carbon nanomaterials have been attracted the attention of researchers in the field of electrochemical sensors [18], [19]. For example, Duoc et al. [20] have used double-walled carbon nanotubes and graphene hybrid films (DWCNTs-Gr), and cholesterol oxidase (ChOx) to construct sensitive arsenic (V) electrochemical sensor. Alizadeh et al. [21] constructed a citric acid sensor based on g-C₃N₄, Fe₂O₃ and BiOI. Thus, it is possible to use these carbon nanocomposites to enhance the electrochemical and electrocatalytic properties of sensors.

Cyclodextrin-metal-organic framework (CD-MOF) based on alkali metal ions and cyclodextrin (CD) has been reported as a new type of green MOF [22]. Because of the unique adsorption and encapsulation abilities of CD-MOF, it has usually been used as a carrier and adsorbent [23]. For instance, Liu et al. [24] used β-cyclodextrin MOF as precursor to synthesize porous carbon for the adsorption of herbicides. Kritskiy et al. [25] used γ-cyclodextrin-metal organic framework for encapsulation of leflunomide. However, the application of CD-MOF in the field of electrochemical sensors has been rarely reported. Therefore, it is novel to improve the sensitivity and detection capability of sensors by the powerful enrichment performance of CD-MOF.

Herein, a C₃N₄-SWCNTs/rGO/CD-MOF composite has been developed on bare glass carbon electrode (GCE) as novel high-sensitivity sensor for Bn ( Fig. 1). First, the cavity of CD-MOF can adsorb a large amount of Bn, thus increasing the concentration of Bn on the sensor surface to amplify the signal. Next, the active site on C₃N₄ can accelerate the rapid redox of Bn on the electrode surface. However, the strong bonding between SWCNTs can greatly affect the high charge mobility and stability of C₃N₄-SWCNTs. After the addition of rGO, the strong π-π interaction between rGO and SWCNTs provides good dispersion and stability of the nanocomposites. In addition, rGO can be loaded with various nanomaterials at high capacity to improve the efficiency of the sensing materials and thus the sensitivity of the sensors [26]. Thus, the sensor shows excellent detection performance for Bn with the help of synergistic effects of these materials and can be applied to the analysis in actual samples.

Fig. 1.

Illustration of the construction process of C₃N₄-SWCNTs/rGO/CD-MOF/GCE and used for detecting Bn.

2. Material and methods

2.1. Reagents and apparatus

Melamine (C₃H₆N₆) and short single-walled carbon nanotubes (SWCNTs) were purchased from Tianjin Guangfu Fine Chemical Research Institute and Chengdu Organic Chemistry Co., Ltd., Chinese Academy of Sciences. β-cyclodextrin, baicalin, and Na₂CO₃ were obtained from Aladdin. The phosphate buffered saline (PBS) was prepared with NaCl, H₃PO₄, NaH₂PO₄, and Na₂HPO₄.

CHI 660D was used to study the electrochemical behavior of the sensors. JEM-2100 transmission electron microscope, Bruker D8 Advance X- ray diffractometer, JEOL JSM-6610LV scanning electron microscope, Thermo ESCALAB 250XI X-ray photoelectron spectrometer, Thermo Nicolet 6700 FT-IR spectrometer, and Renishaw inVia Raman microscopes were used to characterize the materials.

2.2. Preparation of C₃N₄-SWCNTs

First, a 50 mL crucible containing 5 g of melamine was placed in a tube furnace. The tube furnace was heated to 550 °C at a rate of 10 °C min^-1 and held for 2 h. After cooling, pale yellow solid was collected and ground into a powder to obtain g-C₃N₄.

Next, 100 mg of SWCNTs and 50 mg of g-C₃N₄ were added to 80 mL of mixed acid (V(H₂SO₄): V(HNO₃) = 3: 1) and ultrasonically dispersed for 1 h. The dispersion was refluxed at 60 °C for 2 h and cooled to room temperature. Next, the dispersion was diluted with deionized water and filtered using a polytetrafluoroethylene (PTFE) microfiltration membrane (220 nm), and this process was repeated three times. C₃N₄-SWCNTs was collected after vacuum drying at 60 °C.

2.3. Preparation of reduced graphene oxide

Graphite oxide (GO) was synthesized by a modified Hummers method [27]. GO was dispersed in H₂O by ultrasound for 1 h. Then NaBH₄ was added into the suspension and kept stirring for 24 h at 80 °C. A black solid was obtained after filtration and washing. The reduced graphite oxide (rGO) was collected after vacuum drying at 60 °C.

2.4. Preparation of CD-MOF

CD-MOF was prepared by referring to the previous literature [24], [28]. 1.0 equiv of β-CD and 8.0 equiv of Na₂CO₃ were dissolved in H₂O and filtered. Methanol was slowly diffused into the filtrate within 14 days. The colorless crystals were collected and washed three times with methanol. CD-MOF was obtained after vacuum drying at 35 °C. In addition, γ-CD-MOF was synthesized under the same conditions as β-CD-MOF, only replacing β-CD with γ-CD.

2.5. Preparation of modified electrodes

Dispersions of C₃N₄-SWCNTs (N, N-dimethylformamide as dispersant, 1 mg mL^-1), rGO (H₂O as dispersant, 1 mg mL^-1), and CD-MOF (PBS as dispersant, 2 mg mL^-1) were prepared and sonicated for 1 h. Then, 6 μL C₃N₄-SWCNTs suspension and 4 μL rGO suspension were mixed and sonicated for 30 min. Next, the mixed suspension was dropped on a clean glassy carbon electrode (GCE, Diameter = 3 mm) and dried under an infrared lamp to obtain C₃N₄-SWCNTs/rGO/GCE. CD-MOF was electrodeposited on C₃N₄-SWCNTs/rGO/GCE by cyclic voltammetry (CV) for 25 cycles to obtain C₃N₄-SWCNTs/rGO/CD-MOF composite film (potential range: −2.0 to 2.0 V, scan rate: 100 mV s^-1). It can be found from Fig. S1 that as the number of scanning cycles increases, the redox peak and background current gradually decrease, which proves that the CD-MOF has been successfully modified on the electrode surface.

In addition, C₃N₄/GCE, SWCNTs/GCE, C₃N₄-SWCNTs/GCE, CD-MOF/GCE, C₃N₄-SWCNTs/rGO/β-CD/GCE, C₃N₄-SWCNTs/rGO/γ-CD/GCE, and C₃N₄-SWCNTs/rGO/γ-CD-MOF/GCE were prepared in similar method.

3. Results and discussion

3.1. Materials characterization

The X-ray diffraction (XRD) patterns of C₃N₄ and C₃N₄-SWCNTs are shown in Fig. S2. C₃N₄ exhibits two distinct diffraction peaks at 13.1° and 27.5°, corresponding to the stacking of in-planner repeating layers and the stacking of conjugated aromatic systems (JCPDS 87-1526) [29]. Two characteristic peaks can be observed at 43.2° and 26.0° in the XRD pattern of C₃N₄-SWCNTs, corresponding to the (100) and (002) diffraction planes of SWCNTs, which indicate that SWCNTs have been loaded on the C₃N₄ [30].

The X-ray photoelectron spectroscopy (XPS) spectra of C₃N₄, C₃N₄-SWCNTs and C₃N₄-SWCNTs/rGO has been obtained ( Fig. 2 and S3). The contents of C and O in C₃N₄-SWCNTs are significantly higher than those in C₃N₄, and Na in C₃N₄-SWCNTs/rGO is attributed to NaBH₄ in rGO, which proves that C₃N₄ has been bound to SWCNTs and rGO (Fig. 2A and S3A). The peaks of C₃N₄ at 284.8 eV and 288.4 eV correspond to the C–C/C C bond and N C–N bond [31], and the intensity of the C C/C–C bond is greatly enhanced after binding to SWCNTs and rGO (Fig. 2B and S3B). In N 1s spectra, the intensity of the N–(C)₃ bond and H–N–C bond of C₃N₄-SWCNTs/rGO is higher than that of C₃N₄ due to the covalent bonding of C₃N₄ with SWCNTs and rGO. (Fig. 2C and S3C) [32]. In the O 1s spectrum (Fig. 2D and S3D) of C₃N₄, the binding energy peak at 532.4 eV mainly comes from the adsorbed water molecules in the C₃N₄ sample, and the intensity of the peak is low [33]. The O 1s binding energy peaks of C₃N₄-SWCNTs and C₃N₄-SWCNTs/rGO at 532.1 eV and 533.5 eV are mainly attributed to the C O bond and C–O bond [34], [35]. Compared with C₃N₄, the intensity of the C O/C–O peaks of C₃N₄-SWCNTs is so high that the peaks corresponding to the adsorbed water molecules could hardly be found (Fig. 2A). The XPS results indicate C₃N₄ has been synthesized and successfully covalently bonded to SWCNTs and rGO.

Fig. 2.

(A) XPS survey spectra of C₃N₄ and C₃N₄-SWCNTs. High-resolution XPS spectra of (B) C 1s, (C) N 1s, (D) O 1s of C₃N₄ and C₃N₄-SWCNTs.

The Raman spectra of C₃N₄, C₃N₄-SWCNTs, rGO and C₃N₄-SWCNTs/rGO are shown in Fig. S4A. It can be seen that C₃N₄ has a wide band at ~1687 cm^-1, which is attributed to conjugate vibration of carbon and nitrogen in C₃N₄ [36]. The Raman spectrum of pure SWCNTs shows two typical bands, D band and G band at 1349 cm^-1 and 1539 cm^-1. Fig. S4B shows the radial breathing mode (RBM) Raman scattering of SWCNTs, which can be used to estimate diameter of nanotube. The diameter of SWCNTs can be estimated according to the following analytical expression [37]:

$$\begin{array}{c}{\omega }_{\mathrm{RBM}}=\frac{\mathrm{A}}{{\mathrm{d}}_{\mathrm{t}}}+B\end{array}$$ (1)

Where ω_RMB and d_t are the Raman frequency of RBM and the diameter of SWCNTs, respectively. A and B are coefficients related to the substrate (in this experiment, A and B are 234 nm cm^-1 and 10 cm^-1) [38]. By calculation, the diameters of SWCNTs are mainly 1.56 nm and 0.91 nm. The intensity ratio I_D/I_G of SWCNTs and C₃N₄-SWCNTs are almost identical, indicating that the bound C₃N₄ has no effect on the structure of SWCNTs [39]. Moreover, the I_D/I_G of C₃N₄-SWCNTs/rGO is significantly lower than that of rGO, which proves that C₃N₄-SWCNTs/rGO has a high graphitization degree and fewer defects.

The FT-IR spectrum of C₃N₄ shows several strong bands at 1200–1700 cm^-1, which is a typical stretching mode of C–N heterocycles (Fig. S5A) [40]. Peaks of 810 and 896 cm^-1 are attributed to characteristics of the typical breathing pattern of the s-triazine ring and the deformation mode of N–H, respectively [41]. The peak at 3000–3350 cm^-1 belongs to amino groups and physically adsorbed water molecules [42]. For the FT-IR spectrum of C₃N₄-SWCNTs, there is a broader peak at 3000–3600 cm^-1 than C₃N₄, which is due to stretching vibration of more -OH [43]. In addition, the tensile vibration of C C / C O and C–O (at 1500–1700 cm^-1 and 1005 cm^-1), CH stretching vibration (at 2953 / 2893 cm^-1), bending vibration of C–H (at 854 cm^-1) can also be observed [44]. On the other hand, a typical vibrational band of β-cyclodextrin can be found in the FT-IR spectrum (Fig. S5B). The vibration bands at 1024 (C–O / C–C stretching / O–H bending vibrations), 1646 (C O stretching vibrations), 2924 (CH₂ stretching vibrations), and 3302 cm^-1 (-OH stretching vibrations) can be observed [45]. In addition, the -OH peak intensity of CD-MOF is significantly lower than that of pure CD, which is due to the coordination effect of -OH. This result is consistent with previous reports, indicating that CD-MOF has been successfully synthesized [22], [24].

3.2. Micromorphology of materials

The surface micromorphology of materials was obtained by scanning electron microscope (SEM) and transmission electron microscope (TEM) ( Fig. 3). C₃N₄ and SWCNTs show blocky and nanofibrous tubular structures, respectively (Fig. 3a, b). The image of C₃N₄-SWCNTs displays that there are carbon nanotubes around C₃N₄, which proves that C₃N₄ and SWCNTs are well combined (Fig. 3c,d, i, j, k, l). In addition, rGO has a layered and wrinkled structure, which is its typical feature (Fig. 3e). The CD-MOF is arranged in an orderly manner and has a long block shape (Fig. 3f). Fig. 3g shows that rGO has successfully encapsulated C₃N₄-SWCNTs, and lamellar, massive, and fibrous morphologies can be observed. After electrodeposition of CD-MOF on C₃N₄-SWCNTs/rGO, many blocky particles are found on the surface, indicating successful and effective deposition (Fig. 3h).

Fig. 3.

SEM images of (a) C₃N₄, (b) SWCNTs, (c, d) C₃N₄-SWCNTs, (e) rGO, (f) CD-MOF, (g) C₃N₄-SWCNTs/rGO, (h) C₃N₄-SWCNTs/rGO/CD-MOF. TEM images of (i-l) C₃N₄-SWCNTs.

3.3. Electrochemical behavior of electrodes

Electrochemical impedance spectroscopy (EIS) and the classical electrochemical model (Rs((Rct-Zw)-CPE)) have been used to study the electron transfer resistance (Rct) of electrodes. As can be seen from Fig. 4, bare GCE (Rct = 550 Ω) and C₃N₄/GCE (Rct = 1072 Ω) have poor electrochemical performance. Conversely, the Nyquist plots of SWCNTs/GCE and C₃N₄-SWCNTs/GCE show almost no semicircle on the high-frequency region, which indicates that the electron transfer of the probe is extremely fast on these electrodes. In addition, compared with pure β-CD (Rct = 649.9 Ω), the Rct of CD-MOF (Rct = 1338 Ω) is significantly enhanced, which indicates that CD-MOF successfully formed rigid skeleton structure and reduced electron transfer performance. Due to the poor conductivity of CD-MOF, the electron transfer performance of the C₃N₄-SWCNTs/rGO/CD-MOF/GCE (Rct = 70.94 Ω) is slightly reduced compared to C₃N₄-SWCNTs/GCE. This result can also confirm that CD-MOF has been deposited on C₃N₄-SWCNTs/rGO/GCE.

Fig. 4.

The Nyquist plots of modified electrodes. Supporting electrolyte: 0.1 M KCl containing 5 × 10^-3 M K₃[Fe(CN)₆] / K₄[Fe(CN)₆]. Frequency range: 0.1 Hz to 100 KHz. Amplitude: 5 mV.

Fig. S6 shows CV curves of various electrodes in K₃Fe(CN)₆. Peak current observed on the C₃N₄-SWCNTs/GCE is higher than currents of CD-MOF/GCE and C₃N₄/GCE, indicating that C₃N₄-SWCNTs demonstrates a better electrocatalytic and sensing activity. In addition, the peak current observed on the C₃N₄-SWCNTs/rGO/CD-MOF/GCE is slightly smaller than that observed on the C₃N₄-SWCNTs/GCE and C₃N₄-SWCNTs/rGO/GCE, which means that CD-MOF modified by electrodeposition dose not significantly reduce the electrochemical performance.

Further, the electroactive surface area (Aeff) of the different modified electrodes has been calculated (the calculation procedure has been recorded in Supplementary Material) and the results show that C₃N₄-SWCNTs/rGO/CD-MOF/GCE have the largest electroactive surface area (Fig. S7).

3.4. Electrochemical behavior of Bn on different electrodes

To study electrocatalytic of different modified electrodes on Bn, CV curves and differential pulse voltammetry (DPV) curves in PBS have been recorded ( Fig. 5). The curves of GCE and C₃N₄/GCE have shown very weak peak currents (curves a, b), indicating that poor electrochemical processes have occurred on these electrode surfaces. SWCNTs/GCE has large background currents, excellent conductivity, but weak catalytic performance for Bn (curve f). Compared with curves b and f, C₃N₄-SWCNTs/GCE shows suitable background currents and good electrocatalytic performance for Bn with the help of good catalytic properties of C₃N₄ and great electrochemical properties of SWCNTs (curve c). Because of large specific surface area, good biocompatibility and stability of rGO, electrocatalytic ability of C₃N₄-SWCNTs/rGO/GCE for Bn has been further enhanced (curve d). In addition, CD-MOF can adsorb Bn on the electrode surface because of its excellent enrichment ability, which can greatly enhance the response current of Bn. Therefore, under the synergistic effect of various materials, C₃N₄-SWCNTs/rGO/CD-MOF/GCE has the largest redox current and the best catalytic performance for Bn (curve e).

Fig. 5.

(A) CVs and (C, D, E) DPVs of (a) GCE, (b) C₃N₄/GCE, (c) C₃N₄-SWCNTs/GCE, (d) C₃N₄-SWCNTs/rGO/GCE, (e) C₃N₄-SWCNTs/rGO/CD-MOF/GCE, and (f) SWCNTs/GCE in 0.1 M PBS (pH=2.5) containing 2.5 × 10^-7 M Bn. (B, F) Redox peak current of different electrodes. Scan rate of CV: 100 mV·s^-1, pulse period of DPV: 0.2 s.

To investigate the enrichment performance of different CD and CD-MOF for Bn, CV and DPV curves of Bn have been recorded on C₃N₄-SWCNTs/rGO/β-CD/GCE, C₃N₄-SWCNTs/rGO/γ-CD/GCE, C₃N₄-SWCNTs/rGO/β-CD-MOF/GCE, and C₃N₄-SWCNTs/rGO/γ-CD-MOF/GCE (Fig. S8). Compared with pure β-CD, γ-CD shows better enrichment and detection performance for Bn, which is consistent with the previously literature (curve a, b) [46]. When CD is replaced with CD-MOF (both γ-CD and β-CD), the enrichment capacity of these electrodes is further enhanced, which is attributed to the high specific surface area, large pore size, strong biodegradability and strong encapsulation capacity of CD-MOF (curve c, d) [47]. In addition, the redox currents on C₃N₄-SWCNTs/rGO/β-CD-MOF/GCE and C₃N₄-SWCNTs/rGO/γ-CD-MOF/GCE are similar, indicating that both γ-CD-MOF and β-CD-MOF have analogous adsorption properties. Due to the simple preparation and low cost of β-CD, β-CD-MOF has been chosen to construct this sensor.

3.5. Effect of scan rate on C₃N₄-SWCNTs/rGO/CD-MOF/GCE

Fig. 6A shows the CV curves of Bn on C₃N₄-SWCNTs/rGO/CD-MOF/GCE at different scan rates (v). Redox currents increase linearly with the increasing v (Fig. 6, Fig. 6B), and the calibration equations for oxidative peak (i_pa) and cathodic peak (i_pc) can be found as follows.

$$\begin{array}{c}{\mathrm{i}}_{\mathrm{pa}}\left(\mathrm{\mu A}\right)=0.06\mathrm{v}\left(\mathrm{mV}·{\mathrm{s}}^{-1}\right)+1.11\left({\mathrm{R}}^2=0.9985\right)\end{array}$$ (2)

$$\begin{array}{c}{\mathrm{i}}_{\mathrm{pc}}\left(\mathrm{\mu A}\right)=-0.06\mathrm{v}\left(\mathrm{mV}·{\mathrm{s}}^{-\mathrm{1}}\right)-3.30\left({\mathrm{R}}^2=0.9994\right)\end{array}$$ (3)

Fig. 6.

(A) CVs of C₃N₄-SWCNTs/rGO/CD-MOF/GCE at different scan rates (from a to s: 10, 20, 40, 60, 80, 100, 120, 140, 160, 180, 200, 220, 240, 260, 280, 300, 320, 340, 360 mV·s^-1) in 0.1 M PBS (pH 2.5) containing 2.5 × 10^-7 M Bn. (B) The dependence of redox peak currents on the scan rates (v). (C) The dependence of oxidation peak potential on lnv.

These characteristics indicate that the redox reaction of Bn is adsorption-controlled process on C₃N₄-SWCNTs/rGO/CD-MOF/GCE. Oxidation peak potential (E_pa) shifts negatively with the increasing v, which indicates redox process of Bn tends to be irreversible on C₃N₄-SWCNTs/rGO/CD-MOF/GCE (Fig. 6C). The relationship between scan rate and E_pa can be depicted as:

$$\begin{array}{c}{\mathrm{E}}_{\mathrm{pa}}\left(\mathrm{V}\right)=0.021\ln \mathrm{v}\left(\mathrm{mV}·{\mathrm{s}}^{-\mathrm{1}}\right)+0.37\left({\mathrm{R}}^2=0.9943\right)\end{array}$$ (4)

According to the Laviron equation [48], [49]:

$$\begin{array}{c}{\mathrm{E}}_{\mathrm{pa}}\left(\mathrm{V}\right)={\mathrm{E}}^0+\left(\frac{\mathrm{RT}}{{\mathrm{\alpha n}}_{\mathrm{\alpha }}\mathrm{F}}\right)\ln \left(\frac{{\mathrm{RTk}}^0}{\mathrm{\alpha nF}}\right)+\left(\frac{\mathrm{RT}}{\mathrm{\alpha nF}}\right)\ln \mathrm{v}\end{array}$$ (5)

where, E⁰ is the formal standard potential, R is the universal gas constant (8.314 J mol^-1 K^-1), T is the absolute temperature (298 K), α is the transfer coefficient, n is the electron number, F is the Faraday constant (96485 C mol^-1), and k⁰ is the rate constant. According to (4), (5), we have calculated that the value of αn is 1.22. The normal range of α is 0.3–0.7 [50], so the α is 0.61 and n is 2. Therefore, the redox electron transfer number of Bn is 2 on the surface of C₃N₄-SWCNTs/rGO/CD-MOF/GCE, which is consistent with the previous report [51].

3.6. Effect of pH on C₃N₄-SWCNTs/rGO/CD-MOF/GCE

As shown in Fig. 7, the redox potential shifts to the negative potential with increasing electrolyte pH. Calibration equations of E_pa, formal potential (E^θ), and reduction peak potential (E_pc) can be expressed as follows.

$$\begin{array}{c}{\mathrm{E}}_{\mathrm{pa}}\left(\mathrm{V}\right)=-0.049\mathrm{pH}+0.60\left({\mathrm{R}}^2=0.9902\right)\end{array}$$ (6)

$$\begin{array}{c}{\mathrm{E}}^{\mathrm{\theta }}\left(\mathrm{V}\right)=-0.052\mathrm{pH}+0.59\left({\mathrm{R}}^2=0.9967\right)\end{array}$$ (7)

$$\begin{array}{c}{\mathrm{E}}_{\mathrm{pc}}\left(\mathrm{V}\right)=-0.055\mathrm{pH}+0.58\left({\mathrm{R}}^2=0.9981\right)\end{array}$$ (8)

Fig. 7.

(A) CVs of C₃N₄-SWCNTs/rGO/CD-MOF/GCE at different pH (from a to j: 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0) in 0.1 M PBS containing 2.5 × 10^-7 Bn. (B) The dependence of oxidation peak potential (Epa), formal potential (E^θ), reduction peak potential (Epc) on pH. (C) The dependence of redox peak currents on pH. Scan rate of CV: 100 mV·s^-1.

According to Nernst Equation [52]:

$$\begin{array}{c}\frac{{\mathrm{d}}_{\mathrm{Ep}}}{{\mathrm{d}}_{\mathrm{pH}}}=\frac{2.303\mathrm{mRT}}{\mathrm{nF}}\end{array}$$ (9)

Where, m and n are the proton and electron number, respectively. The value of m/n is 1.1 through (7), (9), indicating that number of m and n are equal in the redox process of Bn on C₃N₄-SWCNTs/rGO/CD-MOF/GCE.

The above electrochemical studies of scan rate and pH indicate that the redox process of Bn on the sensor involves the transfer of two protons and two electrons. Gaussian 09 and Multiwfn [53] have been used to study the redox process and the reaction sites of Bn (Density functional theory study has been recorded in Supplementary Material), and the results show that the two hydroxyl groups on the benzene ring of Bn are possible sites for oxidation reactions (Fig. S9 and S10).

3.7. Optimization of conditions

3.7.1. Optimal electrolyte pH

It can be observed from Fig. 7C that the redox peak currents increase as the pH increases from 1.5 to 2.5, and the maximum value is obtained at pH 2.5. Then the redox current decreases with increasing electrolyte pH. Furthermore, analyzing the DPV curves of Bn at different pH on C₃N₄-SWCNTs/rGO/CD-MOF/GCE, the relationship between currents and pH is the same as the result of CV curves ( Fig. 8). Since the maximum redox currents can be obtained in PBS with pH = 2.5, it has been selected as the supporting electrolyte for experiments.

Fig. 8.

(A) DPVs of C₃N₄-SWCNTs/rGO/CD-MOF/GCE at different pH (from a to j: 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0) in 0.1 M PBS containing 2.5 × 10^-7 M Bn. (B) The dependence of oxidation peak currents on pH. Pulse period of DPV: 0.2s.

3.7.2. Optimal electrodeposition cycle for CD-MOF

The number of electrodeposition cycles is an important factor affecting the amount of CD-MOF modification and the enrichment performance of C₃N₄-SWCNTs/rGO/CD-MOF/GCE. Therefore, the influence of different electrodeposition cycles on the detection ability of Bn has been studied. It can be seen from Fig. S11 that the oxidation current of Bn gradually increases as the number of electrodeposition cycles increases from 0 to 25. This shows that CD-MOF has been successfully modified and significantly improved the enrichment capacity of the modified electrode. The peak current decreases significantly when the number of electrodeposition cycles is increased from 25 to 35, which is due to the weakened electrochemical and electrocatalytic properties of the modified electrodes caused by excessive CD-MOF. Therefore, 25 cycles have been chosen as the optimal modification cycles.

3.7.3. Optimal accumulation time and accumulation potential

Since CD-MOF has excellent enrichment ability, the accumulation time and accumulation potential also affect the electrocatalytic ability of C₃N₄-SWCNTs/rGO/CD-MOF/GCE for Bn. Fig. S12A shows that the oxidation current gradually increases with the accumulation time from 1 min to 26 min. However, when the accumulation time is increased from 26 min to 34 min, the current is almost unchanged, which indicates that the enrichment has reached saturation. Therefore, 26 min is selected as the optimal accumulation time. Similarly, Fig. S12B shows that the peak current increases significantly as the accumulation potential changes from − 0.2–0 V and reaches a maximum at 0 V, and gradually decreases from 0 to 0.2 V. Hence, 0 V is used as the optimal accumulation potential for the entire experiment.

3.8. The detection performance for Bn on C₃N₄-SWCNTs/rGO/CD-MOF/GCE

Under the optimal experimental conditions, the linearity and limit of detection (LOD) of Bn on C₃N₄-SWCNTs/rGO/CD-MOF/GCE has been obtained ( Fig. 9A). The current and concentration of Bn show good linear relationship between 1 × 10^-9 M to 5 × 10^-7 M, and the LOD (S/N = 3) is 4.6 × 10^-10 M (Fig. 9B, C). The linear regression equation can be described as:

$${\mathrm{i}}_{\mathrm{pa}}\left(\mathrm{\mu A}\right)=0.22{\mathrm{C}}_{\mathrm{Bn}}\left(\mathrm{nM}\right)+0.24\left(1\times {10}^{-9}\mathrm{M}\le {\mathrm{C}}_{\mathrm{Bn}}\le 3\times {10}^{-7}\mathrm{M},{\mathrm{R}}^2=0.9976\right)$$ (10)

$${\mathrm{i}}_{\mathrm{pa}}\left(\mathrm{\mu A}\right)=72.11\lg \left({\mathrm{C}}_{\mathrm{Bn}}\right)\left(\mathrm{nM}\right)-112.91\left(3\times {10}^{-7}\mathrm{M}\le {\mathrm{C}}_{\mathrm{Bn}}\le 5\times {10}^{-7}\mathrm{M},{\mathrm{R}}^2=0.9943\right)$$ (11)

Fig. 9.

(A) DPVs of Bn quantitative analysis on C₃N₄-SWCNTs/rGO/CD-MOF/GCE, concentrations of Bn (a to o- 1, 2, 5, 10, 30, 60, 100, 150, 200, 250, 300, 350, 400, 450, 500 nM). (B, C) DPVs of linear calibration between the current responses and the concentration of Bn. Pulse period of DPV: 0.2 s.

As shown in Table 1, C₃N₄-SWCNTs/rGO/CD-MOF/GCE has high sensitivity, good linear range and ultra-low LOD comparing with reported electrodes. So, the C₃N₄-SWCNTs/rGO/CD-MOF/GCE has great detection performance for Bn.

Table 1.

Comparison of various modified electrodes for detection of Bn.

Electrode	Technique	Linear range (mol L-1)	LOD (mol L-1)	Sensitivity (A/M)	Refs.
Ta2O5-Nb2O5@CTS/CPE	DPV	8 × 10-8 - 8 × 10-6	3 × 10-8	8.49	[51]
DM-β-CD-GNs/GCE	DPV	4 × 10-8 - 3 × 10-6	1 × 10-8	55.3	[54]
Co-amino-Gr/GCE	SWV	1 × 10-8 - 8 × 10-7	5 × 10-9	53.2	[55]
MoS2/GCE	DPV	1.25 × 10-7 - 1.25 × 10-5	5 × 10-8	–	[56]
SS-b-CD–Pd@RGO/GCE	DPV	2 × 10-8 - 2 × 10-5	5.2 × 10-9	1.88	[57]
boron doped diamond	SWV	1 × 10-6 - 9.5 × 10-7	2.6 × 10-7	0.05	[58]
TRGO/GCE	DPV	1 × 10-8 - 1 × 10-5	6 × 10-9	7.36	[59]
C3N4-SWCNTs/rGO/CD-MOF/GCE	DPV	1 × 10-9 - 5 × 10-7	4.6 × 10-10	220	This work

3.9. Repeatability, reproducibility, anti-interference, and stability

C₃N₄-SWCNTs/rGO/CD-MOF/GCE has been employed to detect 2.5 × 10^-7 M Bn by DPV for twelve times to evaluate its repeatability. The relative standard deviation (RSD) of currents is 1.0%, indicating C₃N₄-SWCNTs/rGO/CD-MOF/GCE has good repeatability ( Fig. 10 A). In order to evaluate reproducibility, ten independently modified electrodes have been prepared under the same conditions to detect 2.5 × 10^-7 M Bn by DPV. Fig. 10 B shows that the prepared sensors have satisfactory reproducibility with the RSD of 2.3%.

Fig. 10.

(A) DPV responses of the same modified electrode toward 2.5 × 10^-7 M Bn collected from twelve repeat measurements. (B) DPV responses of ten independent electrodes prepared under the same condition toward 2.5 × 10^-7 M Bn. (C) The DPVs of C₃N₄-SWCNTs/rGO/CD-MOF/GCE in 0.1 M PBS (pH 2.5) containing different substances. (D) Oxidation peak current of DPV on C₃N₄-SWCNTs/rGO/CD-MOF/GCE in PBS containing different substances. Pulse period of DPV: 0.2 s.

In addition, influences of some possible interferences on the oxidation current of Bn have been investigated by recording the DPV curves without or with some substances (Fig. 10 C, D). It can be found that one hundred equivalents of galactose (Gal), fructose (Fru), sucrose (Suc), glucose (Glu), citric acid (CA), ascorbic acid (AA), potassium chloride (KCl), ammonium chloride (NH4Cl), and twice equivalents of chrysin (Chr) has almost no effect on the DPV response and oxidation current of Bn, which indicates that C₃N₄-SWCNTs/rGO/CD-MOF/GCE has good anti-interference performance for Bn.

To study storage performance of C₃N₄-SWCNTs/rGO/CD-MOF/GCE, seven modified electrodes under the same conditions have been prepared. Then they are stored at room temperature (25 °C), and an electrode is taken out every five days to detect Bn and record the oxidation current ( Fig. 11A). The results show that the modified electrode still has excellent electrocatalytic performance for Bn after being stored at room temperature for one month. Similarly, in order to study the stability of C₃N₄-SWCNTs/rGO/CD-MOF/GCE, the same electrode has been used to obtain the oxidation current of Bn every two days and stored it at 4 °C (Fig. 11B). It shows oxidation current decreased by 11.9% after two weeks, indicating that the sensor has good stability.

Fig. 11.

(A) DPV responses of seven independent electrodes at different storage times. (B) DPV responses of same C₃N₄-SWCNTs/rGO/CD-MOF/GCE at different storage times.

3.10. Detection of Bn in actual samples on C₃N₄-SWCNTs/rGO/CD-MOF/GCE

3.10.1. Detection in human serum

Since Bn often has been used as medicament, to evaluate applicability of manufactured electrochemical sensor in actual samples, C₃N₄-SWCNTs/rGO/CD-MOF/GCE has been detected the Bn in human serum through a standard addition method. Human serum samples have been obtained from Xiangtan University Hospital and diluted 100 times with PBS (pH 2.5). C₃N₄-SWCNTs/rGO/CD-MOF/GCE has been obtained the oxidation current in the blank sample and the spiked sample, and the Bn concentration is calculated through the calibration equation. Table 2 shows that the recovery rate of Bn is 97.7–106.2%, which indicates that C₃N₄-SWCNTs/rGO/CD-MOF/GCE can be used to the detection of Bn in human serum.

Table 2.

Determination of Bn in human serum samples (n = 3).

	Added (nM)	Found (nM)	Recovery (%)	RSD (%)
1	0	0	–	–
2	50	53.1	106.2	1.5
3	100	99.8	99.8	2.1
4	200	199.4	99.7	1.6
5	300	293.2	97.7	2.4
6	400	406.3	101.6	1.7

3.10.2. Detection in sold medicine

Bear bile scutellaria eye drops (a kind of medicine) have been purchased from a local pharmacy and diluted 20 times with methanol. Then, 20 μL of diluent has been added to 20 mL PBS (pH 2.5). Next, DPV curves for unspiked and spiked samples have been obtained on C₃N₄-SWCNTs/rGO/CD-MOF/GCE, and the concentration of Bn has been calculated by the calibration equation. Table 3 shows that the recovery of Bn is 96.7–104.9% in the presence of actual medicine, indicating that the sensor can be used to detect actual medicines.

Table 3.

Determination of Bn in sold medicine (n = 3).

Added (nM)	Found (nM)	Recovery (%)	RSD (%)
0	68.1	–	–
50	117.3	98.4	2.5
100	170.2	102.1	1.9
150	213.2	96.7	1.4
200	277.9	104.9	2.7
300	382.5	104.8	3.2

4. Conclusion

In this work, C₃N₄-SWCNTs have been successfully synthesized by a one-pot method, and a novel ultrasensitive Bn electrochemical sensor based on C₃N₄-SWCNTs/rGO composites and cyclodextrin metal-organic backbone has been proposed. Due to the synergistic effect of these materials, the sensor exhibits excellent detection performance and ultra-high sensitivity (220 A/M) for Bn. Subsequently, the developed method has been successfully applied to the analysis of actual human serum samples and medicine samples, showing satisfactory reliability and great potential for application. This work provides a novel method for the detection of flavonoids and proposes an interesting scheme for the application of cyclodextrin metal organic frameworks in electrochemical sensors.

CRediT authorship contribution statement

Pengcheng Zhao: Conceptualization and Methodology. Linzi Huang: Validation and Investigation.Hui Wang: Methodology and Visualization Chenxi Wang: Writing – original Draft. Jia Chen: Visualization.Pingping Yang: Validation.Meijun Ni: Writing – review & editing.Chao Chen: Visualization and Supervision.Chunyan Li: Supervision.Yixi Xie: Supervision and Validation.Junjie Fei: Conceptualization, Funding acquisition and Project administration.

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

Junjie Fei is currently a professor of College of Chemistry, Xiangtan University, China. He received his Ph.D. degree in Analytical Chemistry from Wuhan University (China) in 2004. His work focuses on the design, construction and application of stimuli-responsive electrochemical sensors and novel high-sensitivity electrochemical sensors.

Yixi Xie is currently an associate professor of College of Chemistry, Xiangtan University, China. He received his Ph.D. degree in Analytical Chemistry from Central South University (China) in 2014. His work focuses on the design, construction and application of new electrochemical sensors for polyphenol detection.

Pengcheng Zhao received his PhD in Applied Chemistry from Xiangtan University (China) in 2021. His research focuses on the construction of electrochemical sensors for novel carbon nanomaterials.

Appendix A. Supplementary material

Supplementary material.

.