Ethylenediamine (EDA, CAS no. 107-15-3), CuCl₂ 2H₂O (CAS no. 10125-13-0), (NH₂)₂CS (CAS no. 62-56-6), and Na₂PdCl₄ (CAS no. 13820-53-6) were purchased from Macklin Biochemical Co., Ltd. (Shanghai, China). Bovine serum albumin (BSA) (CAS no. 9048-46-8; storage, 2–8C), CEA (L2C01001, 2023-12; 1 mg/ml, −20C), and CEA antibody (L1C00202, 2023-12; 1.9 mg/ml, −20C) were from Shanghai Lingchao Biological New Material Technology Co., Ltd. (Shanghai, China). Phosphate buffer solutions (PBS) were prepared with Na₂HPO₄ (CAS no. 7558-79-4) and KH₂PO₄ (CAS no. 7778-77-0). Human serum was obtained from Zibo Central Hospital. Ultrapure water (18.25 Ω) was made in the laboratory throughout the experiments. Hydrogen peroxide (H₂O₂, 30 wt%) was purchased from Shuangshuang Chemical Co., Ltd. (Yantai, China).

The electrochemical measurements were performed on the electrochemical workstation (China). Conventional three-electrode systems used in electrochemical measurement include glass carbon electrode (GCE), saturated calomel electrode, and platinum wire electrode. The Tecnai G2 F20 transmission electron microscope (Hillsboro, OR, USA) was used for transmission electron microscopy (TEM) image acquisition. The JEOL JSM-6700F microscope (Tokyo, Japan) was used to record the X-ray energy (EDX) spectrum. SEM images were taken using the FEI Quanta FEG250 Field Emission Environmental Scanning Electron Microscope (Hillsboro, OR, USA).

The material preparation of Cu₂S was consistent with that reported in the literature (Zhang et al., 2018), and the method of hydrothermal synthesis was adopted. Firstly, 1 mmol CuCl₂ 2H₂O was dissolved in 30 ml EDA and 3 mmol thiourea was added. Then, the mixture was stirred with a magnetic mixer for 2 h at room temperature. After stirring, the mixed solution was transferred into a 50-ml polytetrafluoroethylene (PTFE) lined autoclave and reacted at 80°C for 8 h. Finally, the centrifugal Cu₂S was washed with anhydrous ethanol and secondary deionized water and then dried in a freeze dryer for the next step.

Polyvinylpyrrolidone (PVP, 50 mg) and synthesized Cu₂S (7.2 mg) were added into 8 ml secondary deionized water. Then, 5 ml of 10 mmol/L Na₂PdCl₄ solution was added to the above solution. Subsequently, the mixture was stirred with a magnetic mixer for 20 min and washed with ethanol and secondary deionized water three times. Finally, the obtained black powdery Cu₂S/Pd was dried in vacuum.

The GCE (0.4 µm in diameter) was first polished with 0.05 µm aluminum oxide powder and then thoroughly rinsed with ultrapure deionized water to acquire a fresh and transparent surface. The oxidation peak of bare GCE is less than 100 mV with the reduction peak. The polished GCE was covered with deionized water to prevent oxidation, and then the electrode was blown dry with ear washers. Afterwards, the GCE electrode was modified with Cu₂S/Pd (6 μl, 2 mg/ml) and the three-electrode system was assembled. The electrode was placed into 10 ml PBS, and 40 μl H₂O₂ (5 M) was injected into PBS under the operation of chronoamperometry. Finally, the residual H₂O₂ on the surface of the working electrode was washed off with PBS and the modified electrode dried.

To illustrate the whole process of the experiment more clearly, Figure 1 shows the layered self-assembly process of the immune sensor. The CEA antibody (anti-CEA, 6 μl, 2 mg/ml) was added to the surface of Cu₂S/Pd/CuO/GCE and incubated at 4C. After washing with PBS (pH 6.81), bovine serum albumin (BSA) solution (1 wt%, 3 μl) covered the anti-CEA/Cu₂S/Pd/CuO/GCE to block nonspecific active sites between the substrate nanocomposites and CEA. After 60 min of incubation, the BSA/anti-CEA/Cu₂S/Pd/CuO/GCE was washed with PBS (pH 6.81) and was added different concentration gradients of CEA (from 6 μl, 100 fg/ml to 100 ng/ml) for 60 min to optimize the reaction conditions between CEA and anti-CEA. Finally, the prepared working electrodes were washed with PBS (pH 6.81) and stored at 4°C.

The electrochemical measurement was carried out using electrochemical workstation CHI760E. The immunosensor uses −0.4 V as the scanning potential to measure the current curve of the ampere. After the background current remained stable, H₂O₂ (5 M, 10 μl) was emptied into PBS (10 ml, pH 6.81), stirred with the magnetic stirrer, and the changes of the current response were recorded. The cyclic voltammetry (CV) test was carried out in K₃ [Fe(CN)₆ ³⁻] solution (5 mM). Under the open circuit voltage of 0.196 V, electrochemical impedance spectroscopy (EIS) analysis was performed for the potassium ferricyanide solution, and a Nernst plot was drawn to record each fixed step. All the electrochemical measurement processes were conducted at room temperature.

Figure 2 shows the morphology of the prepared material. It can be found that the obtained Cu₂S is snowflake-shaped and has a flat symmetrical structure with six orientated petals radially extending from the central button; it has a diameter of 4–6 um based on the TEM image (Figure 2A). When Pd NPs wrapped around Cu₂S, they caused the Cu₂S/Pd to have an unclear boundary line and increased the specific surface area, as shown in Figures 2B–D. In Figures 2E, F, the successful preparation of Cu₂S/Pd/CuO nanocomposites was confirmed by TEM. The mapping spectra of Cu, S, Pd, and O (Figures 3B–E, respectively) clearly indicated that the distribution of the four elements was relatively uniform, specifying that the material is well constructed. At the same time, the EDX spectra (Figure 3F) confirmed that the composite material contained the Cu, S, Pd, and O elements.

EIS can be used to compare the electrical conductivity of different materials (Strong et al., 2021). In this work, EIS was used to monitor the change of electron transfer resistance (R _et). The semicircle part represents the electron transfer limitation. The larger the semicircle diameter, the greater the R _et. Cu₂S (curve a) has poor electrical conductivity, which was obviously enhanced after palladium atoms were loaded and can be used as a substrate material. The electrical conductivity of Cu₂S/Pd/CuO (curve c) was basically the same as that of Cu₂S/Pd (curve b), as shown in Figure 4A and Supplementary Table S1, indicating that Cu₂S/Pd/CuO has good electrical conductivity.

Immunosensors need to have good conductivity and catalysis, which are important for the collective effect of the various materials (Zheng et al., 2021). The sensitivity of the unlabeled immunosensor mainly depends on the reducibility of the constructed material to H₂O₂ (Yan et al., 2018). Chronoamperometry (i–t) can be used to compare the catalytic activities of different modified materials (Lee et al., 2018). As shown in Figure 4B, naked GCE has no catalytic activity for H₂O₂ (curve a). When Cu₂S was loaded onto the electrode, the catalytic signal increased slightly (curve b); the signal of Cu₂S/Pd (curve d) was greater than that of Cu₂S (curve b). This is due to the better catalytic performance of Pd NPs and the large specific surface area of snowflake-like materials, which can support more Pd NPs. Cu₂S generated Cu₂S/CuO (curve c) in the presence of hydrogen peroxide, and its catalytic activity was enhanced. It was further found in this experiment that the catalytic activity of Cu₂S/Pd/CuO (curve e) was higher than that of Cu₂S/Pd (curve d), and the current response was more stable. Therefore, the Cu₂S/Pd/CuO nanocomposite material was used as a signal-amplifying platform to construct a highly sensitive and unlabeled electrochemical immunosensor. After the modification of Cu₂S/Pd/CuO on the GCE, the peak current after multiple scanning, calculated by measuring the CV response, was basically unchanged, indicating that the prepared nanocomposite has good stability (Figure 4C). The successful construction of the sensor was verified by EIS detection of layers of modification (Figure 4D and Supplementary Table S2). Compared with the bare GCE (curve a), the resistance of the Cu₂S/Pd/CuO (curve b) electrode was higher. When anti-CEA (curve c), BSA (curve d), and CEA (curve e) were successively modified to the working electrode surface, the semicircle diameter of the electrical impedance spectra increased continuously, which was attributed to the partial inhibition of electron transfer by proteins (Supplementary Table S2), indicating that the electrochemical immunosensor was successfully modified.

In order to obtain the best measurement results for the tumor markers, experimental conditions such as the substrate concentration of Cu₂S/Pd/CuO and the pH of PBS need to be optimized. The immunosensor was constructed based on a CEA concentration of 1 ng/ml in this study.

The pH value of PBS is important for the catalytic properties of the immunosensor because a strongly acidic or alkaline environment may inactivate antigens and antibodies, thus affecting the specificity of protein binding. When the pH of PBS changed from 5.91 to 6.81, the current signal began to increase to a peak. However, when the pH of PBS exceeded 6.81, the current response was reduced. Therefore, the maximum current signal was obtained at a pH of 6.81, which maintained good biological activity. Therefore, PBS with pH of 6.81 was selected as the best electrolyte for electrochemical measurements (Figure 5A). The concentration of Cu₂S/Pd/CuO is one of the most important parameters affecting the performance of the electrochemical immunosensor. The concentration of Cu₂S/Pd/CuO will have an impact on the electron transfer and the loading capacity of the anti-CEA. In order to obtain the best performance of the immunosensor, working electrodes with different concentrations of Cu₂S/Pd/CuO were used, and 10 μl (5.0 mM) H₂O₂ was injected into 10 ml PBS at pH 6.81. As shown in Figure 5B, the current response increased significantly with the increase of Cu₂S/Pd/CuO concentration from 0.5 to 2.0 mg/ml, and then decreased gradually with the increase of Cu₂S/Pd/CuO concentration from 2.0 to 3.0 mg/ml. Therefore, the optimal concentration for the immunosensor construction in this study was 2.0 mg/ml.

In this experiment, an electrochemical immunosensor with good conductivity and catalytic activity was prepared with the Cu₂S/Pd/CuO composite material. A series of CEA concentrations were measured by chronoamperometry. With the increase of CEA concentration, the current signal of the immune sensor decreased, indicating that the antigen impeded the electron transport (Figure 6A). In the range from 100 fg/ml to 100 ng/ml, the linear regression equation of the CEA concentration to the value and current response was Y = −14.34lgC + 105.24, and the correlation coefficient was 0.9997 (Figure 6B). The limit of detection (LOD) was 33.11 fg/ml (S/N = 3).

Compared with other reported CEA results of the immunosensor (Supplementary Table S3), the immunosensor in this study showed a comparable linear range (from 100 fg/ml to 100 ng/ml) and an improved LOD (33.11 fg/ml). The reasons for the low detection limit of the prepared immunosensor were as follows: firstly, Cu₂S/Pd/CuO has good biocompatibility, which ensures antigen activity and can effectively fix antibodies. Secondly, Cu₂S/Pd/CuO can provide a wider detection range as a multi-signal amplification platform. Finally, Cu₂S/Pd/CuO has good electron transfer performance, which benefits improving the sensitivity of the electrochemical immunosensor.

After the electrochemical immunosensor was prepared, its performance needs to be verified, such as its repeatability, selectivity, and stability. These are important parameters for the evaluation of clinical application methods. Firstly, five immunosensors were prepared with the same concentration of CEA by chronoamperometry to detect the current signal; the repeatability of the electrochemical immunosensor was then studied. The relative standard deviation (RSD) of the electrode was calculated as 1.15%, indicating that the immune sensor has good repeatability (Figure 7A). Secondly, the selectivity of the immunosensor was studied using troponin I (CTnI), immunoglobulin G (IgG), and neuron-specific enolase (NSE) as the interfering substances. As shown in Figure 7B, the RSD was less than 1.79%, indicating that the selectivity of the immunosensor is relatively reliable. Finally, the stability of the immunosensor was assessed by measuring the current response of the five working electrodes and storing them in a 4C refrigerator when not tested once a day for five consecutive days. After 15 days, the current signal obtained decreased from 100% to 92.1% (Figure 7C), showing good stability. The experimental results showed that the immunosensor has good reproducibility, selectivity, and stability and can be used for highly sensitive and quantitative detection of CEA.

In order to evaluate the application potential and reliability of the immunosensor designed in this study, a standard recovery experiment and human serum sample test were conducted. The results showed that the recoveries were 95.5%–108%, as shown in Table 1. Compared with that of commercial electrochemiluminescence immunoassay (ECLIA), the relative error was 3.70%–5.89% (Supplementary Table S4), which proves the feasibility of the designed immunosensor. These results indicate that the constructed immunosensor has potential application value in clinical serum CEA detection.