3.1 Preparation and characterization of BSA-CuAuNCs
In the work, BSA-CuAuNCs with double-peak emission were synthesized by one-pot hydrothermal method using BSA as the ligand, AA as the reducing agent, CuSO4 as the copper source and HAuCl4 as the gold source. The synthesis of red-emitting CuNCs had been described in detail by our team and applied to the detection of acrylamide in toast (Cheng et al. 2024). However, the red-emitting CuNCs were unstable under acidic environment. As shown in Fig. S1A (Supplementary Materials), the fluorescence intensity of red-emitting CuNCs at 655 nm was reduced to about 51% of the original at pH 4 within 30 min. Therefore, it is significant to improve the stability of BSA-CuAuNCs under acidic conditions. Alloy nanoclusters with stable properties had been reported, which provided a reference for improving the fluorescence stability of BSA-CuAuNCs under acidic conditions (Liu et al. 2018a; Long et al. 2018). As shown in Fig. S1B, the fluorescence intensity of BSA-CuAuNCs did not change significantly at pH 4 within 30 min, indicating good resistance to acids. However, it is noteworthy that the doping of gold did not lead to obvious enhancement of fluorescence intensity, probably because the doping did not significantly change the triplet state emission of the nanoclusters or the electron transfer from the ligand to the metal core (An et al. 2020). Even in the case that the copper-gold ratio was 92:8, there was not any red emission, and it might be attributed to the fact that gold doping disrupted the structure of nanoclusters (Kang et al. 2020). So, 98:2 was chosen as the optimum copper-gold ratio.
To verify the optical properties of BSA-CuAuNCs, we measured UV-Vis and fluorescence spectra. The absorption spectra were shown in Fig. 1A, the characteristic absorption peaks at 270–300 nm for BSA and BSA-CuAuNCs were inconsistent, which might be influenced by the synthesis of nanoclusters. There were no characteristic absorption peaks in the range of 400–700 nm for BSA-CuAuNCs, indicating that no copper nanoparticles or gold nanoparticles were formed in the system (Shojaeifard et al. 2019). It was found that the absorption at wavelengths of 300–480 nm was significantly stronger than that of BSA, and there was a shoulder at 310 nm. The fluorescence spectra were shown in Fig. 1B, and the optimum excitation wavelength of BSA-CuAuNCs was 310 nm, and the optimum emission wavelengths were 395 nm and 655 nm (strong), which were attributed to BSA and the triplet state emission of BSA-CuAuNCs, respectively (Zhao and Li 2020). The optimal emission wavelength of BSA-CuAuNCs in the near-infrared region was red-shifted from 650 nm to 655 nm with gold doping (Fig. 1A), due to a larger particle size of BSA-CuAuNCs (Cheng et al. 2024; Wang et al. 2016). BSA-CuAuNCs solution was brownish-yellow under daylight and bright red under 310 nm UV light (Fig. 1B inset).
The morphology and elemental composition of BSA-CuAuNCs were explored by TEM, XPS spectroscopy and FTIR spectroscopy. The TEM images were used to observe the morphology and particle size of BSA-CuAuNCs. As shown in Fig. 2A, TEM images presented an elliptical shape with the distribution of particle sizes ranging from 1.6–2.4 nm, and in total of 200 particles were calculated to have an average particle size of 2.20 nm (Fig. 2B). XPS spectrum was applied to probe the elemental composition and elemental chemical states. As shown in Fig. 2C, BSA-CuAuNCs contained S (168.65 eV), C (285.39 eV), N (399.89 eV), O (531.34eV), Cu (932.03 eV) and Au (84.18 eV). The characteristic peaks of 932.03 eV and 951.98 eV corresponded to the characteristics of Cu2p3/2 and Cu2p1/2, which indicated the presence of Cu (0) and Cu (Ⅰ), and the absence of characteristic peaks of 942 eV showed that no presence of Cu (Ⅱ) (Fig. 2D). As shown in Fig. 2E, BSA-CuAuNCs had characteristic peaks at 84.18 eV and 87.65 eV, corresponding to Au4f 7/2 and Au4f 5/2, indicating the presence of Au (0) (Chen et al. 2020). The FTIR spectrum was used to probe the surface structure of BSA-CuAuNCs. As shown in Fig. 2F, BSA endowed BSA-CuAuNCs with the same functional groups as it, e.g. the peak 1399 cm− 1 corresponding to -OH in -COOH, the peak of 1533 cm− 1 corresponding to N-H, the peak of 1647 cm− 1 corresponding to amide bond, and the peak of 3301 cm− 1 corresponding to -OH, and the presence of these functional groups endowed BSA-CuAuNCs with good aqueous solubility (Bi et al. 2023).
The fluorescence stability of BSA-CuAuNCs affected their application in practical detection. The stability of BSA-CuAuNCs was investigated based on fluorescence emission at 655 nm, including the effects of pH, UV radiation time and ionic strength. As shown in Fig. S2A (Supplementary Materials), the fluorescence intensity of BSA-CuAuNCs remained stable from pH 4 to 9, demonstrating good resistance to acids and bases. When BSA-CuAuNCs were irradiated with the UV lamp (365 nm) for 120 min, the fluorescence intensity was basically unchanged (Fig. S2B), indicating their outstanding photostability. There was no significant difference in fluorescence intensity when the concentration of NaCl reached 3M (Fig. S2C), suggesting excellent ionic strength resistance of BSA-CuAuNCs.
3.2 Nitrite determination
We found that nitrite quenched the emission at 655 nm of BSA-CuAuNCs and had nearly no effect on the peak at 395 nm. Based on this, a ratiometric fluorescence sensor was constructed by the relationship between fluorescence ratios of BSA-CuAuNCs (F655/F395) and nitrite concentrations. In order to obtain the best analytical performance, experimental conditions (e.g. the concentration of BSA-CuAuNCs, pH value, and reaction time) were optimized. As shown in Fig. S3A (Supplementary Materials), with the increase of BSA-CuAuNCs concentration, F655/F395 showed a tendency that it decreased and then increased and reached the minimum at the concentration of 0.6 mg/mL, and so 0.6 mg/mL was chosen as the optimal concentration. pH value was a critical factor that influenced the detection system, and it was found that under neutral or weakly alkaline conditions, the fluorescence emission was not quenched by nitrite. As pH value decreased, the F655/F395 decreased rapidly, and no change was observed after pH value reached 4. Therefore, pH 4 was chosen as the optimal pH value for the sensor (Fig. S3B). As the reaction proceeded, F655/F395 decreased gradually. The F655/F395 maintained constant after 6 min and 6 min was selected as the optimal reaction time for the sensor (Fig. S3C).
Under the optimal experimental conditions, the fluorescence spectra and F655/F395 were measured when different concentrations of nitrite were added and the fluorescence spectra were shown in Fig. 3A. With the increase of nitrite concentration from 0 to 60 µM, F655/F395 decreased gradually and showed a good linear relationship (Fig. 3B) with a linear equation of F655/F395=-0.05523C + 5.3356, R2 of 0.99806. The LOD was 0.12 µM, which met the requirements for nitrite detection in food.
To verify the selectivity and anti-interference ability of the sensor, common anions and cations (NO3−, Cl−, SO42−, HSO4−, SO32−, HSO3−, Ac−, CO32−, HCO3−, Na+, K+, NH4+, Ca2+, Mg2+, Al3+, Fe3+) were tested separately under the same condition. As shown in Fig. 3C, the concentration of other ions (120 µM) was twice the concentration of nitrite (60 µM), and the F655/F395 obviously decreased only when nitrite was added. There was not significant difference with the addition of nitrite alone when other ions were present as interferences (120 µM) (Fig. 3D). The results exhibited that good selectivity and anti-interference ability of the proposed method. Table 1 summarized reported methods for nitrite determination and the constructed sensor was compared with previous works. It could be concluded that the designed method had the advantages of high sensitivity, outstanding specificity, facial operation and rapid detection.
Table 1
Comparison of different methods for nitrite detection
Detection method | Linear range | LOD | Recovery (%) | References |
HPLC | 0.1–100 µg/mL | 0.05 mg/L | 98.9-105.7 | (Chou et al. 2020) |
Capillary electrophoretic | 1.38-23.00 mg/L | 0.11 mg/L | 93.30-109.11 | (Tembo and Aygun 2021) |
Electrochemical method | 2-1000 µM | 2 µM | 98.8-102.5 | (Yimo et al. 2023) |
Ion chromatography | 0.1–20 mg/L | 0.6 mg/kg | 83.7-107.6 | (Lim et al. 2022) |
CDs/OPD | 10–500 µM | 0.45 µM | 96.84-104.48 | (Heng et al. 2023) |
Methanobactin-AuNPs | 0–8.0 µM, 8.0–50.0 µM | 16.21 nM | 95.6-102.9 | (Chen et al. 2023b) |
AuNCs@EW/H+ | 0–24 nM, 0.4–64 µM | 2.82 nM | - | (Li et al. 2020) |
CuNCs | 10–225 µM | 4.3µM | 97.3-106.6 | (Zhou et al. 2017) |
BSA-CuAuNCs | 1–60 µM | 0.12 µM | 95.03–107.20 | This work |
3.3 Possible quenching mechanisms
The main types of fluorescence quenching were categorized as static quenching and dynamic quenching (Chen et al. 2023a). To investigate the mechanism of quenching BSA-CuAuNCs by nitrite, firstly, XPS spectra of the system with and without nitrite were measured. It was possible that the oxidation of the core of BSA-CuAuNCs led to fluorescence quenching due to oxidation capacity of nitrite (Liu et al. 2018b). However, Cu2p and Au4f spectrum exhibited no obvious change, and the Cu2p spectra showed no a characteristic peak at 942 eV, indicating the absence of Cu (Ⅱ), which demonstrated that the fluorescence quenching was not due to the oxidation of the metal core (Fig. 4A and B). Subsequently, we determined the absorption spectra in the presence and absence of nitrite. As shown in Fig. 4C, the absorption peaks at 275 nm and 355 nm appeared after the addition of nitrite respectively, indicating possible interactions occurring in the system (Gao et al. 2014). Finally, TEM images of the system with the addition of nitrite were obtained and we found that BSA-CuAuNCs underwent aggregation (Fig. 4D), which was consistent with Liu et al. (Liu et al. 2013). In summary, the quenching of BSA-CuAuNCs by nitrite could be attributed to the aggregation of BSA, which belongs to static quenching.
3.4 Detection of real samples
In order to verify the potential use of the established sensor, nitrite in kimchi samples was detected. The results were shown in Table 2. The spiked recoveries of the sensor were 95.03%-107.20%, with RSDs of 1.52%-4.26%. Satisfactory results were obtained, indicating that this sensor could be applied for the nitrite detection in real samples.
Table 2
Analytical results of nitrite detection in kimchi based on BSA-CuAuNCs system.
Samples | Spiked (µM) | Found (µM) | Recovery (%) | RSD (%, n = 3) |
kimchi | 0 | 15.62 | - | 2.19 |
10 | 26.34 | 107.20 | 4.26 |
20 | 36.58 | 104.80 | 1.82 |
40 | 53.63 | 95.03 | 3.32 |