Introduction

Sodium dichloroisocyanurate (DCCNa), as 1-sodium-3,5-dichloro-1,3,5-triazine-2,4,6-trione, is widely used in the treatment of industrial circulating water, the cleaning and sterilization of the food, and disinfection of the agricultural industry because of its oxidative properties [1, 2]. As a germicidal and sterilizing agent, it has strong disinfection capabilities, a wide range of applications, and low toxicity [3]. Although the toxicity of DCCNa is relatively small, excessive inhalation contact has toxic effect on the human body's lungs [4], and it will produce a large amount of water-insoluble reaction residues cyanuric acid, which may cause secondary harm in the process of disinfection. For example, cyanuric acid can cause kidney damage, disrupt amino acid metabolism, and so on [5]. Furthermore, it has been listed as a drinking water pollutant candidate by the United States Environmental Protection Agency [6]. As a result, the safety performance of DCCNa cannot be ignored, and it's important to establish a quick and easy method to monitor DCCNa.

Rosmarinic acid (RA) is an organic compound widely present in plants, and it belongs to phenolic compounds, which is synthesized by caffeic acid and 3,4-dihydroxyphenyl lactic acid [7]. RA has a wide range of applications in food, medicine, cosmetics, and other fields. For example, in the realm of the food industry, RA is frequently employed as a natural antioxidant [8]. In the realm of pharmaceutical research, RA has garnered considerable interest due to its array of biological activities, including antioxidative, anti-inflammatory, antibacterial, anticoagulant and antitumor properties [9,10,11]. Similarly, due to its possession of these characteristics, it can be utilized in the production of skincare cosmetics [12]. In short, RA has many benefits and potential applications that deserve further study by researchers. In an effort to better understand and utilize it efficiently, it is significantly important to establish a simple and sensitive RA detection platform.

So far, there are fewer detection methods for DCCNa, and it is mainly the determination of cyanuric acid as well as hypochlorous acid of its decomposition products. For example, liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GS-MS) were used for the detection of cyanuric acid [13, 14]. And flow injection analysis (FIA) and chemiluminescence analysis were used for the detection of hypochlorous acid [15, 16]. But there are many techniques for detecting RA concentrations, such as high-performance liquid chromatography (HPLC) [17], high-performance thin-layer chromatography (HPTLC) [18], differential pulse voltammetry (DPV) [19], ultraviolet-visible spectroscopy [12], and so on. However, these methods may have some problems, such as requiring complex sample preparation, specialized operators to operate instruments, time-consuming, or low sensitivity. Compared with these methods, the fluorescence method is relatively simple and highly sensitive [20]. Therefore, it is necessary to establish a new and simple method to detect the concentrations of DCCNa and RA.

Over the past few decades, various research teams have developed different materials as fluorescence sensors, such as organic dyes [21], semiconductor quantum dots (QDs) [22,23,24], and metal nanoclusters (MNCs) [25]. Among them, MNCs are a class of new nanomaterials, consisting of several to several hundred metal atoms, and their particle sizes are usually less than 3 nm [25]. This new type of fluorescent nanomaterial due to its small size, low toxicity, good biocompatibility, and excellent optical performance features was widely used in the fields of chemical sensing [26], bioimaging [27], catalysis [28], and other fields. Copper is rich and inexpensive compared to precious metals (Au/Ag), so CuNCs are attracted much attention. Up until now, many ways were employed to synthesize CuNCs. For example, Kang et al. [29] used N-acetyl-L-cysteine (NAC) as a reducing agent and stabilizer to effectively synthesize CuNCs through the sonochemical method, and it could detect Ag+. Yu et al. [30] synthesized OVA-CuNCs using egg albumin ovalbumin (OVA) and 2-mercaptobenzimidazole as raw materials through one-pot method and it was used to detect RA sensitively. Zou et al. [20] synthesized Trp-CuNCs using D-tryptophan as a ligand through a rapid microwave-assisted method, and it was used for the detection of tetracycline. It’s well known that the polymers contain a large number of functional groups, which can be complexed with metals. Up to now, CuNCs were prepared by PVP as a ligand, but it has not been used for the detection of DCCNa/RA.

Hence, PVP-coated copper nanoclusters (PVP-CuNCs) were prepared by one-step method (Fig. 1), and the synthesis and detection conditions were optimized. Based on the quenching effect of DCCNa on PVP-CuNCs, an efficient method for detecting DCCNa/RA was established and used in real products.

Fig. 1
figure 1

The illustration of the synthesis of PVP-CuNCs and detection of DCCNa/RA by PVP-CuNCs

Full size image

Experimental Sections

Materials and Instruments

Fourier transform infrared (FTIR) spectra were recorded by Nicolet-6700 FT-IR spectrometer (Thermo Fisher Scientific, USA) based on potassium bromide sheets and were measured in the range of 400–4000 cm-1. The remaining materials and instruments were placed in the Supporting Information section.

Synthesis of PVP-CuNCs

The PVP-CuNCs were prepared by slightly modified previous literature reports [31]. In short, in the first place, 0.2 g of PVP (Mr = 58000) and 8.0 mL of ultrapure water were mixed fully at 25℃, and then CuSO4·5H2O (1.0 mL, 10 mM) and AA (1.0 mL, 100 mM) were added into the mixture. Later, pH of the above mixture was adjusted to 3.5 by HCl or NaOH (1.0 M) and stirred for 3 hours at 75 ℃. After that, the solution was cooled to room temperature and purified for 24 hours with a dialysis membrane (MWCO of 8000 Da). Finally, it was stored in a refrigerator at 4 ℃ for subsequent experiments.

Quantum Yield Calculation

The calculation method of the quantum yield was displayed in the Supporting Information.

Fluorescence Detection Process of DCCNa and RA

For determining DCCNa, a series of DCCNa solutions (0.06 mM) in different volumes (0, 1.5, 3.0, 6.0, 9.0, 12.0, 15.0, 20.0, 25.0, 30.0, 35.0, 45.0, 55.0, 65.0, 75.0, 85.0, 95.0, 105.0, 120.0, 135.0 and 150.0 μL) were added to 50 μL of PVP-CuNCs solution in 10 mL plastic centrifuge tubes. Then the BR buffer solution (pH = 4) was added to set the volume to 3 mL. Afterward, the mixture was fully reacted at room temperature for 1.0 min, and corresponding fluorescence spectra were recorded at an excitation wavelength of 345 nm, whose excitation/emission slit was kept at 5/5 nm.

For the analysis of RA, in the first, a series of different volumes (3.0, 6.0, 9.0, 12.0, 15.0, 20.0, 25.0, 30.0, 40.0, 50.0, 60.0, 75.0, 90.0, 105.0 and 120.0 μL) of RA solution (0.03 mM) were added to 75 μL of DCCNa solutions in 10 mL plastic centrifuge tubes. After 2-minute reaction, the 50 μL of PVP-CuNCs solution was added and fixed in 3 mL with BR buffer solution (pH = 4). The remaining steps are consistent with the previous steps. Finally, the selectivity and interference experiments of DCCNa/RA were discussed. All experiments were implemented three times.

Pretreatment and Detection of Actual Samples

All samples were purchased from the local supermarket (Nanchong, China) and they were preprocessed according to previous literatures [32, 33]. For the detection of actual DCCNa samples, a certain amount of pre-treated solution and DCCNa standard solution (0, 0.06, 0.12 and 0.18 μM) and 50 μL of CuNCs were extracted for analysis and detection according to the detection method described in the above section. For the detection of actual RA samples, a certain amount of pretreated solution and 75 μL of DCCNa solution (0.06 mM), 50 μL of CuNCs and RA standard solution (0, 0.06, 0.09 and 0.12 μM) were extracted for analysis and detection accordance according to the detection method described in the above section.

Results and Discussion

Optimization of Synthesis Conditions

In order to obtain good photoluminescence of PVP-CuNCs, its synthesis conditions were optimized. First, the influence of pH was discussed. As shown in Fig. S1A, the fluorescence intensities of PVP-CuNCs at 424 nm increased with the increase of pH (2.5–3.5), and then decreased with the increase of pH (3.5–4.5). It may be attributed to the fact that AA has a stronger reducing capacity in a larger pH environment, reducing Cu2+ to large-size non-luminous nanoparticles, while in a lower pH environment, the reducing capacity is not strong [34]. Secondly, the dosage of PVP was optimized, and the fluorescence intensity of PVP-CuNCs at 424 nm was strongest when the concentration of PVP was set at 0.3448 mM (Fig. S1B). Finally, the impact of PVP with different average molecular weights was also discussed under the aforementioned optimal conditions. As shown in Fig. 2A, its fluorescence intensity at 424 nm was the highest when the average molecular weight of PVP was 58000, which might be the higher relative molecular weight of PVP, the larger corresponding hydrated particle size of CuNCs (Fig. 2B–F), resulting in larger particle sizes in DLS measurement and lower fluorescence intensity of PVP-CuNCs at 424 nm. However, the synthesized CuNCs with lower relative molecular weight of PVP also exhibited relatively low fluorescence intensity, possibly because smaller molecular weight might reduce the interaction between PVP and Cu, thus inhibiting the charge transfer of ligands to metals. Also, the solubility of PVP with higher molecular weight in water decreased, causing that the ligands couldn’t be effectively attached to metal surfaces [35, 36]. In conclusion, the optimal pH, type and dosage of PVP for the synthesis of PVP-CuNCs were set at 3.5, Mr = 58000 and 0.3448 mM, respectively.

Fig. 2
figure 2

The influence of PVP with different average molecular weights (A) for synthesis; Particle size statistical distribution using dynamic light scattering (DLS) for the PVP-CuNCs (Mrpvp = 8000 (B), 24000 (C), 58000 (D), 220000 (E) and 1300000 (F)) systems

Full size image

Characterization of PVP-CuNCs

To characterize the PVP-CuNCs, TEM, XPS, FT-IR, UV-vis absorption, and fluorescence spectra were recorded. First of all, the TEM and XPS were measured. As shown in Fig. 3A, PVP-CuNCs were spherical and had good dispersibility in water. And its average diameters were estimated to be 2.21 ± 0.06 nm by the randomly selected 100 nanoparticles (Insert of Fig. 3A). At the same time, the lattice fringe space was measured 0.206 nm (Fig. 3B), which indicated the Cu (111) plane of the nanoclusters [37]. The elemental compositions and states of CuNCs were performed by XPS. The XPS survey spectrum of CuNCs (Fig. 3C) clearly showed that the copper nanoclusters were composed of C, N, O, and Cu. The high-resolution survey scan of Cu 2p was shown in Fig. 3D, which exhibited two distinct fitting peaks at 932.85 and 952.6 eV, representing 2p3/2 and 2p1/2 of Cu, respectively [38]. It indicated that CuNCs were composed of Cu (0) and Cu (I). In addition, since the characteristic peak of 942 eV didn’t exist, it further indicated that copper nanoclusters didn’t contain Cu (II) [39]. In the high-resolution survey scan of C 1s (Fig. S2A), 284.79, 285.71 and 287.37 eV were the locations of the feature fitting peaks of C-C, C-N and C=O bondings, respectively [40]. In the high-resolution survey scan of N 1s (Fig. S2B), 399.49 eV was the location of the feature-fitting peak of the C-N bonding [41]. In the high-resolution survey scan of O 1s (Fig. S2C), 530.88 and 523.08 eV were the locations of the feature fitting peaks of the C=O bonding and O-H bonding, respectively [42]. Furthermore, elemental analysis of CuNCs was performed (Fig. 3E), it could be clearly seen that elements such as Cu, C, N and O were evenly distributed in the material and corresponding EDX composition was given in the inset of Fig. S3. Additionally, as shown in Fig. S4, the FT-IR spectra showed that the absorption peaks of stretching vibration of -CH bonding was 2952.5 cm-1 and deformation vibration of -CH bonding were 1425.2 cm-1 and 1373.1 cm-1, and the absorption peaks of C-N and C=O bonding were 1286.3 cm-1 and 1672.0 cm-1, respectively, which were consistent with literature reports [43]. That also showed that PVP and PVP-CuNCs had similar absorption peaks, and PVP was successfully attached to the surface of CuNCs. Meanwhile, the UV-vis absorption spectra of PVP and PVP-CuNCs, showed that only the new absorption peak appeared at about 287 nm in the PVP-CuNCs, while there was no surface plasma resonance peak at 500-600 nm (Fig. S5), which indicated that there weren’t copper nanoparticles [44]. Then the fluorescence spectra showed the maximum excitation and emission wavelengths of the PVP-CuNCs were 345 nm and 424 nm, respectively (Fig. 3F). And the PVP-CuNCs solution was faint yellow in sunlight (left cuvette of the insert of Fig. 3F) and blue in a 365 nm UV lamp (right cuvette of the insert of Fig. 3F). In summary, copper nanoclusters were successfully synthesized. Finally, the quantum yield of the PVP-CuNCs was calculated to be 14.97% using quinine sulfate as the reference standard (Фs = 54%).

Fig. 3
figure 3

TEM images of CuNCs (A) (Insert: corresponding the size distribution) and corresponding HRTEM (B); XPS survey spectrum of CuNCs (C) and high-resolution survey scan of Cu 2p (D); Mapping image of CuNCs (E); Excitation and emission spectra of CuNCs (F, Insert: photographs of CuNCs in sunlight (left cuvette) and 365 nm UV lamp (right cuvette))

Full size image

Stability of PVP-CuNCs

The stability of CuNCs was studied by measuring its fluorescence intensity at 424 nm under different conditions. As shown in Fig. S6A, within the 4–10 range of pH, the fluorescence intensities of CuNCs varied slightly, showing good stability. From Fig. S6B, it could be seen that CuNCs still had a stable fluorescence intensity after 9 days of storage, indicating that it exhibited good storage stability. In addition, the salt resistance of CuNCs as shown in Fig. S6C, in the salt concentration range of 0-200 mM, the fluorescence intensities of CuNCs changed little, indicating its salt stability was better.

Detection of DCCNa

In order to obtain the best detection conditions, the effects of pH, reaction time and temperature were discussed. The optimal reaction conditions were obtained as follows: pH value of BR buffer, reaction time, and reaction temperature were 3.5, 1.0 min, and 25℃, respectively (Fig. S7A–C). Thus, DCCNa was detected under the optimal conditions. As shown in Fig. 4A, when different concentrations of DCCNa were added to the PVP-CuNCs system, the fluorescence intensity (424 nm) of PVP-CuNCs gradually decreased with the increase of DCCNa concentration. In the range of 0.030-2.400 μM, the linear fitting curve was obtained according to the relationship between (F0-F)/F0 and DCCNa concentration (Fig. 4B). The linear fitting equation was (F0-F)/F0 = 0.322 × CDCCNa + 0.011 (R2 = 0.998, F0 and F represent the FL intensities of PVP-CuNCs without and with the addition of DCCNa), and the corresponding limit of detection (LOD) was calculated to be 10.766 nM (3σ/K).

Fig. 4
figure 4

Fluorescence spectra of PVP-CuNCs after the addition of DCCNa (A) and RA (C); Linear relationship between the values of (F0-F)/F0 and CDCCNa (B) and (F-F0)/F0 and CRA (D)

Full size image

Detection of RA

To obtain the best detection conditions, the reaction time of DCCNa and RA was discussed. As shown in Fig. S7D, the fluorescence intensity did not change significantly after 2 min of reaction. As shown in Fig. 4C, when different concentrations of RA were added to the PVP-CuNCs-DCCNa detection system, the fluorescence intensity gradually at 424 nm increased with the increase of RA concentration. In the range of 0.030-0.900 μM, a linear fitting curve was obtained according to the relationship between (F-F0)/F0 value and RA concentration (Fig. 4D). The linear fitting equation was (F-F0)/F0 = 0.871 × CRA + 0.006 (R2 = 0.998, F0 and F represent the FL intensities of PVP-CuNCs without and with the addition of RA), and corresponding LOD was calculated to be 8.985 nM (3σ/K). In addition, we compared the LOD and detection range of other methods. As shown in Table S1, the built method had a lower LOD.

Selectivity and Interference Testing of DCCNa

To verify the selectivity of CuNCs for DCCNa, several representative substances were studied, such as K+, Ca2+, Na+, Mg2+, Cu2+, Al3+, Mn2+, Zn2+, Fe3+, S\({\text{O}}_{4}^{2-}\), C\({\text{O}}_{3}^{2-}\), HCO3-, I\({\text{O}}_{4}^{-}\), Cr2\({\text{O}}_{7}^{2-}\), glucose, sucrose, UA, Nam, AA, Gly, L-Val, L-Cys, L-Phe, L-His and L-Leu. It could be seen from Fig. 5A that only DCCNa caused the fluorescence of CuNCs to decrease significantly, indicating that it had good selectivity. Some substances such as kaolin, talc, and some of the above-mentioned substances may be present in the real sample as potential interfering substances. When DCCNa and interfering substances were present at the same time, anti-interference tests were conducted. As could be seen from Fig. 5B, the changes of fluorescence intensities at 424 nm were not obvious, indicating that it had good anti-interference ability.

Fig. 5
figure 5

The selectivity of PVP-CuNCs (A) and PVP-CuNCs-DCCNa (C) to different materials; The fluorescence intensity ratios of PVP-CuNCs-DCCNa (B) and PVP-CuNCs-DCCNa-RA (D) systems in the different coexistent materials

Full size image

Selectivity and Interference Testing of RA

To verify the selectivity of the PVP-CuNCs-DCCNA system for RA, some representative substances were studied, such as K+, Ca2+, Na+, Mg2+, Cu2+, Al3+, Mn2+, Zn2+, Fe3+, S\({\text{O}}_{4}^{2-}\), C\({\text{O}}_{3}^{2-}\), HCO3-, glucose, sucrose, SB, AP, PS, CA, ME, SY, citral and tartrazine. As could be seen from Fig. 5C, only RA caused the fluorescence of the system to increase significantly, indicating that it had good selectivity. Some of the above-mentioned substances may be present in the real sample as potential interfering substances. In the case of RA and interfering substances being present at the same time, anti-interference tests were conducted. As could be seen from Fig. 5D, the changes of fluorescence intensities at 424 nm were not obvious, indicating that it had good anti-interference ability.

The Fluorescence Quenching Mechanism of PVP-CuNCs by DCCNa

As described in Fig. 6A, little overlap between the UV-vis absorption spectrum of DCCNa and fluorescence excitation/emission spectrum of PVP-CuNCs was observed, hinting that the quenching mechanism of CuNCs by DCCNa was neither due to the inner filter effect (IFE) nor Förster resonance energy transfer (FRET) [45, 46]. Additionally, the Zeta potential of CuNCs increased to -2.68 mV after the addition of DCCNa, indicating the oxidation of copper on the surface of the CuNCs to Cu2+, resulting in a decrease in negative charge (Fig. 6B). The oxidation reaction of CuNCs in the presence of DCCNa may be related to Eqs. (1) and (2). At the same time, after the addition of DCCNa, no new absorption peak appeared in the UV-vis absorption spectrum of CuNCs, and the UV-vis absorption peak at 287 nm gradually decreased with increasing concentrations of DCCNa (Fig. 6C), which reconfirmed that the copper on the surface of the CuNCs was oxidized. In conclusion, the quenching mechanism of PVP-CuNCs by DCCNa was oxidative quenching, i.e., copper on the surface of CuNCs was oxidized to Cu2+.

Fig. 6
figure 6

The UV-vis absorption spectrum of DCCNa and the fluorescence excitation and emission spectra of CuNCs (A); The zeta potential of PVP-CuNCs without and with the addition of DCCNa (B); The UV-vis absorption spectra before and after adding DCCNa of CuNCs (C); The UV-vis absorption spectra of RA, DCCNa and RA + DCCNa (D)

Full size image
(1)
$$\mathrm{Cu} + \mathrm{HClO} + \mathrm{H}^+ \to \mathrm{Cu}^{2+} + \mathrm{Cl}^- + \mathrm{H}_2\mathrm{O}$$
(2)

The Mechanism of Fluorescence Increase by RA

From Fig. 6D, there was a shift in the ultraviolet absorption peak of RA when it was mixed with DCCNa. This is because RA possesses antioxidant properties and can react with DCCNa, and the phenomenon can be explained by Eq. (3) [47]. Due to the reaction between RA and DCCNa, which leaded to a decrease in the concentration of HClO, thereby causing an increase in fluorescence intensity.

(3)

Analysis of Actual Samples

To verify the reliability of the method, intra-day and inter-day detection results of DCCNa/RA were analyzed (Table S2). The RSD values were less than 2.22%/2.69%, indicating that the method had certain reliability. In order to further verify the application of this method in real samples, three real samples were selected, and corresponding results were analyzed by standard addition experiment. As shown in Table 1, the recoveries of DCCNa/RA were 95.8-104.9%/96.6-105.6%, and the RSD were less than 1.89%/2.54%. In addition, other standard methods (like titration and HPLC) were used to detect the actual samples (Table 2), and the results between the two were not significantly different. In summary, the method had a certain detection ability.

Table 1 The assays of DCCNa and RA by the CuNCs-PVP and CuNCs-PVP-DCCNa probe, respectively (n = 3)
Full size table
Table 2 Comparison of the detection result of DCCNa and RA by the work and standard methods (n = 3)
Full size table

Conclusions

In this work, blue-emitting CuNCs were synthesized for the detection of DCCNa/RA by one-step method. For the DCCNa/RA assay, corresponding linear ranges were 0.030–2.400/0.030-0.900 μM as well as LOD were 10.766/8.985 nM. The method could be used for the assay of DCCNa/RA in real samples with high recovery. At the same time, the reaction mechanisms of PVP-CuNCs-DCCNa and PVP-CuNCs-DCCNa-RA systems were also discussed. That is, hypochlorous acid produced by DCCNa dissolved in water has strong oxidation, which will oxidize Cu on the surface of CuNCs to Cu2+, causing the fluorescence quenching of PVP-CuNCs. While RA has antioxidant properties and can react with hypochlorous acid to restore fluorescence by reducing the generation of hypochlorous acid.