HCG, Ab1, and Ab2 were obtained from Shanghai Lingcao Biotechnology Co., Ltd. CD, hydrazine solution, and ammonia solution were purchased from Beijing Chemical Reagent factory. Bovine serum albumin (BSA) was purchased from Sinopharm Chemical Reagent Co., Ltd. The supporting electrolyte was phosphate buffer solution (PBS) that was prepared with KH₂PO₄ (0.1 M), K₂HPO₄ (0.1 M) and KCl (0.1 M). CHI 660E Electrochemical Workstation was applied for the electrochemical measurements. GCE (with a diameter of 3 mm), platinum wire and Ag/AgCl electrode were used respectively as the working electrode, counter electrode and reference electrode.

Firstly, graphene oxide (GO) was prepared with the typical Hummers method [23]. For preparing CD/GN, the homogeneous GO dispersion of 50.0 mL (1.0 mg mL⁻¹) and 50.0 mL CD dispersion (1.0 mg mL⁻¹) were mixed in a glass flask to be stirred continually for 15 min before the addition of ammonia solution and hydrazine solution. Then, the formed mixture solution was ultrasonicated for 30 min and subsequently heated at 65°C for 8 h under stirring. Finally, the CD/GN nanohybrid was obtained after filtration and drying.

For constructing RhB/Ab2/CD/GN, 1.0 mL CD/ GN dispersion (0.25 mg) was mixed with 9.0 mL Ab2 solution (200 μg mL⁻¹) and 0.1 mL RhB solution (0.2 mg mL⁻¹), then the mixture was vibrated at 4°C for 5 h. Finally, the obtained RhB/Ab2/CD/GN was washed and stored for further use.

Firstly, GCE was modified with 8.0 μL CD/GN suspension (0.25 mg mL⁻¹) and dried. Next, 3.0 μL Ab1 (0.2 mg/mL) was incubated on the surface of CD/GN/GCE. For preventing the non-specific binding, Ab1/CD/GN/GCE was incubated in BSA solution. Next, the formed Ab1/CD/GN/GCE was interacted with HCG target and then HCG-Ab1/CD/ GN/GCE was immersed in the RhB/Ab2/CD/GN solution to construct a sandwich-type electrochemical immunosensing platform. The electrochemical impedance spectroscopy (EIS) analysis was carried out in 5.0 mM [Fe(CN)₆]³⁻/⁴⁻ solution, and the related experimental setup were as follows: initial potential, 0.25 V; low frequency, 1 Hz; high frequency, 1000000 Hz; amplitude potential, 0.005 V.

Various technologies were introduced to characterize the as-prepared CD/GN nanohybrid. Firstly, Fig. 1a–d showed the atomic force microscopy (AFM) images of GN and CD/GN, the results revealed that the average thickness of GN is ~0.5 nm. As for the CD/GN nanohybrid (Fig. 1b and d), the average thickness increases to be ~1.2 nm owing to the presence of CD molecules on the surface of GN. Then, the FT-IR spectrums of GN and CD/GN were displayed in Fig. 1e, it can be found that the pure GN exhibits typical C-C band (1132 cm⁻¹) and C=C conjugation (1565 cm⁻¹) [24–26], while the synthesized CD/GN shows several typical CD absorption spectrums: the ring vibrations at 482, 622, and 754 cm⁻¹, the C-O/C-C stretching/O-H vibrations at 1029 and 1085 cm⁻¹, the CH₂ stretching vibrations at 2915 cm⁻¹, and the O-H stretching vibrations at 3423 cm⁻¹ [26,27]. In addition, the dispersibility of CD/GN in water was also studied and the images of GN and CD/GN were shown in Fig. 1f. From the images, it can be found that the pure GN exhibits poor dispersibility, while the CD/GN exhibits superior dispersibility in water resulted from the hydrophilic exterior of CD. These characterizations indicate that the CD/ GN nanohybrid was prepared successfully and the CD molecules can improve greatly the dispersibility of GN.

Firstly, the EIS analysis was used to investigate the construction processes of the sandwich-type immunosensor, and the EIS plots of CD/GN/GCE, Ab1/ CD/GN/GCE, HCG-Ab1/CD/GN/GCE and RhB/ Ab2/CD/GN-HCG-Ab1/CD/GN/GCE in 0.1 M KCl solution containing 5.0 mM [Fe(CN)₆]³⁻/⁴⁻ were shown in Fig. 2. The results revealed that the charge-transfer resistance (R_CT) value of CD/GN/GCE is extremely small owing to the prominent conductivity of CD/GN. However, as for the other electrodes including Ab1/CD/GN/GCE, HCG-Ab1/CD/GN/ GCE and RhB/Ab2/CD/GN-HCG-Ab1/CD/GN/ GCE, the Rct values at these electrodes increased in turn, the reason is that the conductivities of biomacromolecules including Ab1, BSA, HCG and Ab₂ are not as good as that of CD/GN. This investigation revealed that the sandwich-type immunosensor layers were constructed successfully.

Next, the feasibility of the designed sandwich-type immunosensor for HCG was demonstrated by measuring the differential pulse voltammetry (DPV) responses of different electrodes. As exhibited in Fig. 3, the results showed that no oxidation peak current of RhB presented at HCG-Ab1/CD/GN/GCE in 0.1 M PBS (pH 7.0) due to the absence of RhB/Ab2/CD/ GN. However, there is an obvious peak current of RhB presented at ~0.855 V for RhB/Ab2/CD/GN-HCG-Ab1/CD/GN/GCE since the antibody-antigen recognition reaction between Ab2 and HCG (6.0 ng mL⁻¹) induced the capture of RhB/Ab2/CD to HCG-Ab1/ CD/GN/GCE. For further comparison, the sensor performances without the use of CD have been studied. As exhibited in Fig. 3, there is only an incredibly small peak current observed that can even be ignored at RhB/Ab2/CD/GN-HCG-Ab1/GN/ GCE, indicating the non-specific adsorption is extremely weak. These comparisons revealed that the as-designed immunosensing proposal is feasible for the HCG detection.

For achieving the highest sensitivity in detecting HCG, several related experimental conditions were evaluated then (HCG, 6.0 ng mL⁻¹). Fig. 4A exhibited the influence from the different pH values of PBS on the peak current of RhB/Ab2/CD/GN-HCG-Ab1/ CD/GN/GCE, it’s noted the DPV response of the immunosensor increases along with the pH increase from 5.0 and reaches a maximum at 7.0, the reason may be that the excessive acid or alkaline conditions destroy the protein structures. Fig. 4B showed the influences from the interaction time of Ab1/CD/GN with HCG and HCG with RhB/Ab2/CD/GN (HCG, 6.0 ng mL⁻¹; pH 7.0), the results revealed that the optimum time for the former is 32 min, while 40 min is for the latter, thus these times were used for the antigen--antibody interactions. In addition, the amount of Ab2 used in constructing RhB/Ab2/CD/ GN was also investigated and the results were shown in Fig. 4C. From the figure, it can be seen that the DPV current of the immunosensor increases gradually with the increase of the amount ranging from 3.0 to 9.0 μL. However, when the amount of Ab2 increases further, the DPV current showed a decrement owing to the competitive immobilization between Ab2 and RhB probe with CD, revealing the ideal balanced point between RhB and Ab2 is 9.0 μL.

Finally, under the conditions optimized above, the detection capability of the designed electrochemical immunosensor for HCG was evaluated. As displayed in Fig. 5, via plotting the calibration curve that consisted of the DPV currents of RhB upon the concentrations of HCG, the obtained results revealed that the currents increase linearly along with the increase of HCG concentrations ranging from 0.002 and 10.0 ng mL⁻¹, and the regression equation is I_p(mA) = 0.304 + 0.686C (ng mL⁻¹) (R₂ = 0.997). Herein, C is expressed the concentration of HCG. Based on S/N = 3, the calculated limit of detection (LOD) is low to 1.0 pg mL⁻¹. For understanding the detection capability of the immunosensor more direct-viewing, the comparisons between the present work and the previous literatures for the analytical parameters including linear range and LOD were performed, and the results of comparison were shown in Table 1. From the comparisons, it can be known that the detection performances of the designed electrochemical immunosensor are more superior (wider linearity and lower LOD) than those at the most of previous literatures.

The SSRR performances (selectivity, stability and reproducibility as well as real application ability) are also very important for the sensors and they were studied in the following sections with the concentration of 2.5 ng mL⁻¹ for HCG. Firstly, the selectivity of the designed immunosensor was performed that was carried out in the presence of several interfering substances (Fig. 6a). The results revealed the peak currents of RhB did not change even in the presence of cortisol (CT), alpha fetoprotein (AFE), dehydroepiandrosterone (DPD), carcinoembryonic antigen (CEA), leptin (LT), BSA, and haptoglobin (HG) (10.0 times of HCG level, that is 25.0 ng mL⁻¹), demonstrating that the designed sandwich-type immunosensor exhibited desirable selectivity for HCG detection. Next, the stability of the immunosensor was studied, and the test results showed that the signal responses of the immunosensor remain almost unchanged for four weeks through storing RhB/Ab2/ CD/GN at 4? before use (Fig. 6b), suggesting the developed immunosensor offered satisfactory stability. Then, we also investigated the reproducibility of the proposed immunosensor by recording the peak current responses of RhB at 8 constructed independently RhB/Ab2/CD/GN-HCG-Ab1/CD/GN/ GCE. As shown in Fig. 6c, the obtained relative standard deviation (RSD) value at these immunosensors is 3.65%, indicating the reproducibility of the present proposal was acceptable. Furthermore, the real application capability of the proposal was also studied by detecting HCG in real human serum via standard addition method. The serum sample was obtained from the local hospital and various concentrations of HCG were added in turn to the samples, sequentially the signal currents of the sensing platform towards the sample were recorded, and the recovery was used to evaluate the feasibility of immunosensor in real application. As displayed in Table 2, the obtained recoveries were noted in the range between 95.0 and 98.0%, revealing the designed sandwich-type immunosensor has desirable real application capability for HCG analysis.