Introduction

Heavy metals are natural components in the Earth’s crust, released into the environment as a result of human activities, and resistant to natural degradation processes. The consequences of even minute quantities of heavy metals (1 ppm) on human well-being cannot be ignored [1,2,3,4]. Among these heavy metals, copper ions (Cu2+) exhibit both beneficial and detrimental impacts on human physiology. On one hand, they contribute significantly to vital metabolic functions; on the other hand, an excessive influx of Cu2+ can lead to various diseases, including anemia and Alzheimer’s disease [5, 6]. Notably, the typical concentration of naturally occurring Cu2+ in freshwater ranges from 0.20 to 30 ng/mL [7, 8]. Different countries establish distinct regulations governing the permissible Cu2+ levels in water, highlighting the importance placed on maintaining appropriate Cu2+ concentrations. Consequently, the recognition and precise detection of Cu2+ in water sources have become increasingly important. Current methodologies for Cu2+ detection encompass techniques such as atomic absorption spectroscopy (AAS) [9], atomic emission spectrometry (AES) [10], and inductively coupled plasma mass spectrometry (ICP-MS) [11]. However, these conventional methods entail intricate sample preparation and instrument operation, rendering them unsuitable for rapid sample analyses.

In recent years, SERS technology has garnered considerable attention for its ability to detect Cu2+ due to its high sensitivity, low detection limit, and reduced time constraints. A notable advancement was made by Liu et al., who introduced a novel and stable SERS sensor for Cu2+ detection [12]. Their approach entails utilizing Ag NPs coated with inositol hexaphosphate (IP6) as the SERS substrate. The substrate was further modified with L-cysteine, and Cu2+ is quantitatively identified using R6G as the SERS probe. Similarly, Guo et al. reported a method for Cu2+ detection using Ag-Au satellite structures. Here, Ag NPs were functionalized with 4-mercaptobenzoic acid (4-MBA), while Au NPs were modified with 4-mercapto pyridine (4-Mpy) [5]. After the addition of Cu2+, the carboxyl groups in 4-MBA and the N atoms in pyridine formed coordination bonds with Cu2+, inducing the aggregation of both Ag and Au NPs. This aggregation resulted in enhanced SERS signal peaks characteristic of 4-MBA and 4-Mpy. Despite these advancements in utilizing SERS for Cu2+ detection, they still face problems such as complex probe molecules and substrates. Therefore, there is an urgent need for a simpler approach to Cu2+ detection using simple SERS substrates and probe molecules. Although these methods have achieved the detection of Cu2+ through SERS technology, they still face problems such as complex probe molecules and complex substrates.

In this work, we introduced a SERS sensor characterized by its simple sample preparation, high sensitivity, and rapid Cu2+ detection capabilities. The SERS substrate consisted of 4-MBA functionalized onto a core-shell structure of Au@Ag NPs. By introducing an iodine ion (I−) solution, a dynamic system was established capable of Cu2+ detection. Upon the addition of Cu2+, the reaction between I− and Cu2+ initiated the etching of the Ag shell layer. This etching led to a reduction in the hot spot region of the system, subsequently attenuating the enhancement effect on Raman probe molecules. “Hot spots” for SERS were defined as the area where great EM enhancement existed [13, 14]. Consequently, the SERS signal intensity associated with 4-MBA at 1586 cm−1 decreased. A robust linear correlation was observed between the change in SERS signal intensity at 1586 cm−1 and the logarithm of Cu2+ concentration within the range of 10−9 to 10−4 M. Notably, the LOD reached a low value of 1.108 × 10−9 M. The applicability of this methodology extended successfully to the analysis of Cu2+ within tap water samples, thereby introducing a new approach to detecting trace amounts of Cu2+ in real-world environments. The schematic representation of this quantitative detection approach is illustrated in Fig. 1.

Fig. 1
figure 1

Schematic illustration of Cu2+ sensor

Full size image

Experimental Section

Materials

All reagents and solvents used were of analytical grade, and the chemicals were used with no further purification. The materials encompassed 4-MBA, sodium citrate, ascorbic acid (AA), KI, HAuCl4·3H2O, AgNO3, Cu(NO3)2, HgCl2, Co(NO3)2, Mn(NO3)2, Ni(NO3)2, Zn(NO3)2, Fe(NO3)2, Fe(NO3)3, and Cd(NO3)2, and ethanol was purchased from Energy Chemical Co., Ltd. (China). Deionized water was utilized in the preparation of the experimental solutions.

Instrumentation

The characterization of Ag NPs and Au@Ag NPs involved the utilization of a JEM-2100 ultra-high-resolution transmission electron microscope (JEOL, Japan) to determine their geometry and size. Additionally, the size of Ag NPs and Au@Ag NPs was determined using a nanoparticle zeta potential analyzer (DLS). Ultraviolet-visible (UV-vis) absorption spectra were recorded using a PerkinElmer Lambda 35 spectrophotometer (326 nm; Norwalk, CT, USA). Elemental composition was established using K-Alpha X-ray photoelectron spectroscopy (XPS), while ion content was quantified using an inductively coupled plasma mass spectrometer (ICP-MS). Material composition was determined via X-ray diffractometer (XRD) analysis. SERS measurements were conducted using a Renishaw inVia reflex confocal microscope (Renishaw, UK). These measurements were facilitated by a He-Ne laser (532 nm, exposure time of 10 s). The laser spot diameter was 1 μm, and a laser power of 50 mW was utilized throughout the experimentation.

Synthesis and Characterization of Au@Ag NPs

The synthesis procedure for Au@Ag NPs is based on existing literature with slight modifications [15]. First, 60 mL of deionized water was combined with 950 μL of a 5 g/L chloroauric acid solution in a reaction vessel; the mixture was subsequently magnetically stirred. The temperature was then increased to 120 ℃ and maintained for 1 min. Following this, 600 μL of an aqueous 1% trisodium citrate solution was rapidly injected into the solution while stirring vigorously; the mixture was left to boil for 5 min. During this process, the solution changed from colorless to purple-red, signifying the successful formation of Au NPs. The resulting colloidal solution was allowed to cool to room temperature and was subsequently stored at 4 ℃.

In a separate container, 3 mL of the Au NP colloidal solution was sonicated for 10 min. Subsequently, 270 μL of a 10 mM AA solution was added, and the mixture was stirred for 15 min. While maintaining stirring, 150 μL of a 10 mM AgNO3 solution was added dropwise to the mixture, and the mixture was stirred for 20 min. During this reaction, the solution changes from purple-red to orange, indicating the successful formation of Au@Ag NPs. The resulting colloidal solution was stored at 4 ℃ for future use.

Preparation of Cu2+ Standard Solution and Samples

First, Cu(NO3)2 was dissolved in deionized water to yield a 1 mM solution. Subsequently, a range of Cu2+ solutions with varying concentrations were prepared by diluting the 1 mM Cu(NO3)2 solution accordingly. For real water sample analysis, tap water from Shenyang, Liaoning Province, China, was collected. The collected water samples were then filtered through a 0.45-μM filter membrane to remove insoluble particles and impurities.

Cu2+ Detection

First, 200 μL of 4-MBA solution, 200 μL of Au@Ag NPs, 200 μL of KI, and 200 μL of either Cu2+ solution or various concentrations of spiked water solution were sequentially transferred into a centrifuge tube. After complete sonication and thorough mixing at ambient temperature, SERS detection was carried out. Using a He-Ne laser (532 nm) with a laser power set at 50 mW, the change in SERS intensity at 1586 cm−1 corresponding to 4-MBA was selected as the quantification benchmark. Each experimental iteration was conducted in triplicate to ensure robustness and consistency.

Results and Discussion

Characterization of Au@Ag NP SERS Substrate

To gain deeper insights into the dimensions and structure of the substrate and to validate its successful synthesis, we conducted comprehensive characterizations involving TEM, DLS, and UV-vis analyses, as shown in Fig. 2. Initially, we characterized the Au NPs during the substrate synthesis process. As illustrated in Fig. 2A, the TEM image of the Au NPs revealed a spherical morphology, depicting a uniform dispersion and relatively consistent size. From the DLS data in Fig. 2B, it was evident that the size of these Au NPs was approximately 15 nm. Furthermore, Fig. 2C displays the UV absorption spectrum of the Au NPs, wherein the absorption peak was situated at 523 nm, effectively confirming the successful synthesis of a spherical Au core. Subsequently, our focus shifted to the core-shell configuration of the Au@Ag NPs, characterized via TEM, DLS, and UV-vis analyses as shown in Fig. 2D–F. The TEM images substantiated the successful encapsulation of the Ag shell layer around the exterior of the Au NPs, resulting in the formation of Au@Ag NPs. These possess a uniform size, further validated by the DLS measurements indicating an average particle size of approximately 34 nm. Notably, the UV absorption peaks were located at 394 nm and 491 nm, corresponding to the plasmon resonance of the Ag shell and Au core [16]. This multifaceted characterization robustly confirmed the successful synthesis of Au@Ag NPs, thus establishing their viability as detection substrates for subsequent analyses.

Fig. 2
figure 2

TEM, DLS, and UV-vis images of Au NPs and Au@Ag NPs

Full size image

To validate the composition of the synthesized substrate as a valuable metal substrate encompassing both Au and Ag, XPS analysis was conducted on both the Au NPs and Au@Ag NPs, as depicted in Fig. 3. As shown in Fig. 3A, the XPS spectrum indicated that the substrate contains gold and silver components. The binding energy of 83.9 eV corresponds to the 4f orbital of Au, specifically Au 4f5/2, while 87.7 eV corresponds to Au 4f7/2 in Fig. 3B. The binding energies of the Ag 3d orbital were recorded at 368.2 eV (Ag 3d5/2) and 374.2 eV (Ag 3d3/2) in Fig. 3C. These distinctive binding energies derived from the XPS analysis affirm the presence of both Au and Ag components within the synthesized substrate.

Fig. 3
figure 3

XPS images of Au@Ag NPs

Full size image

Detection and Characterization Mechanism for Cu2+

The presence of sulfhydryl groups within 4-MBA molecules enabled a strong S-Ag bond formation, facilitating the adsorption of 4-MBA onto the surface of Au@Ag NPs. This interaction generated robust SERS signal peaks within the hot spot regions of Au@Ag NPs. Specifically, the SERS signal peaks of 4-MBA were mainly located at 997, 1012, 1077, 1138, 1181, 1586, and 1706 cm−1 [17, 18]. Among these, the signal peak at 1586 cm−1 exhibited both strength and distinctiveness, making it an ideal candidate for quantitative analysis. Therefore, the peak at 1586 cm−1 was selected as the characteristic peak for quantitative analysis. The Cu2+ detection mechanism relied on a series of chemical reactions. Since (Cu2+/Cu) [φθ (Cu2+/ Cu) = 0.3402 V], (Ag+/Ag) [φθ (Ag+/Ag) = 0.7996 V]. By considering the relationship between electrode potential and the direction of the reaction, it became evident that direct interaction between Cu2+ and Au@Ag NPs is unlikely. However, upon the introduction of I− into the system, Cu2+ underwent a reaction generating CuI precipitates and I2. Subsequently, I2 quantitatively reacted with Ag to produce AgI, inducing a series of reactions: 2Cu2+  + 4I−  = 2CuI + I2, I2 + 2Ag = 2AgI. The occurrence of these reactions led to the etching of the Ag shell layer. As the Ag shell thins, the SERS enhancement effect diminishes, resulting in a reduction of the SERS signal attributed to 4-MBA molecules at 1586 cm−1. Consequently, a linear correlation was established between the concentration of Cu2+ and the SERS signal intensity of 4-MBA molecules.

To validate the accuracy of our explanation regarding the Cu2+ detection mechanism, we initially conducted a comprehensive analysis of the overall changes occurring during the detection process, utilizing UV-vis and TEM techniques. In Fig. 4A, the UV-vis spectrum displayed noteworthy alterations, upon introduction of I− and Cu2+ into the system, resulting in absorption peaks at 426 nm and 558 nm, respectively, where were related to the SPR absorption of AgI and Au, respectively [19, 20]. Figure 4B displays a TEM image of the etched system, clearly illustrating a significant thinning of the Ag shell layer when compared to the previous image. This tangible transformation provided visual evidence of the etching of the Ag shell layer, thereby confirming the occurrence of the reaction. Furthermore, to further prove the occurrence of the etching reaction during the detection process, we delved into the valence changes of elements within the process and analyzed the resultant reaction products. The change in element valence during the reaction was analyzed via XPS characterization. In Fig. 4C, the Ag within the Au@Ag NPs was identified as Ag0, with its binding energy located at 368.2 eV and 374.2 eV. However, after the addition of I− and Cu2+, discernible shifts in binding energy were observed, accompanied by the emergence of new spectral peaks at 366.3 eV and 372.6 eV, which were attributed to generated Ag+ ions. The detection system was further characterized by XRD, as shown in Fig. 4D. It could be observed that the diffraction pattern reveals distinct peaks at 38.2°, 44.5°, 64.6°, and 77.7°, which correspond to the characteristic crystal planes of AgI (110) and (220) and Ag and Au (111) and (200), respectively. This outcome validated the formation of AgI and, by extension, the occurrence of the anticipated etching reactions.

Fig. 4
figure 4

UV-vis, TEM, XPS, and XRD images of the Au@Ag NP etching reaction

Full size image

To validate the feasibility of the SERS detection principle, a comprehensive analysis was conducted using SERS technology. Figure 5 shows the SERS signals of the Au@Ag NPs-4-MBA-I−-Cu2+ system. From the figure, curve a was the Raman spectrum of Au@Ag NPs, where it was evident that Au@Ag NPs did not exhibit any discernible characteristic peaks. This observation effectively confirmed that the substrate does not induce signal interference during Cu2+ detection. Curve c was the Raman spectrum of the Au@Ag NPs-4-MBA system, indicating an easily analyzed SERS spectrum of 4-MBA in the 800–1800 cm−1 range. Notably, a robust SERS characteristic peak at 1586 cm−1 emerged, which could be attributed to the stretching vibration mode of the benzene ring within 4-MBA. This particular peak was designated as the representative SERS characteristic peak. Curve b was the SERS spectrum of the Au@Ag NPs-I− system; it proved that the Au@Ag NPs-I− system did not coexist. Curve c was the SERS spectrum of the Au@Ag NPs-4-MBA-I− system following the introduction of 10−3 M Cu2+. Notably, a significant decrease was observed in the characteristic peak at 1586 cm−1 compared to curve d, which was due to the etching of Au@Ag NPs by the reaction products of I− and Cu2+. The resultant reaction led to a decrease in the thickness of the Ag shell and a decrease in the “hot spot” region, which weakened the enhancement effect on 4-MBA molecules and led to a sharp decrease in SERS intensity at 1586 cm−1.

Fig. 5
figure 5

SERS spectra of (a) Au@Ag NPs, (b) Au@Ag NPs + 10−3 M I−, (c) Au@Ag NPs + 10−3 M 4-MBA + 10−3 M I−  + 10−3 M Cu2+, and (d) Au@Ag NPs + 10−3 M 4-MBA

Full size image

Optimization of the Analytical Conditions

To enhance the sensitivity of Cu2+ detection, various factors influencing the testing process were optimized. These optimizations involved the concentrations of 4-MBA, Au@Ag NPs, and I− solution, as well as reaction time.

The concentration of probe molecules significantly impacted the intensity of SERS signals. Elevated concentrations could lead to multi-layer adsorption on the substrate surface, while excessively low concentrations resulted in single-layer unsaturated adsorption. In contrast, an optimal concentration yielded monolayer saturated adsorption, which aligned with the EM enhancement mechanism for maximum SERS signal strength. Figure S1A illustrates that the highest SERS intensity was observed at a concentration of 5 × 10−3 M, indicating of the monolayer saturated adsorption of 4-MBA onto the Au@Ag NP surface. Therefore, 5 × 10−3 M was selected as the optimal concentration for 4-MBA.

The concentration of the substrate affected the size of the “hot spot” region, thereby affecting the robustness of the SERS signal emitted by the Raman probe molecule. If the substrate concentration was too low, the probe molecules would not have enough adsorption sites, resulting in a weak SERS signal. Conversely, an excessively high substrate concentration could induce substrate aggregation, leading to a reduced area accessible for probe molecule adsorption and a reduced SERS signal strength. The concentrations of the previously synthesized Au@Ag NPs were denoted as c. Based on this, Au@Ag NP substrates were derived at concentrations of 2c, c/2, c/3, and c/4 using centrifugation and dilution techniques. These five substrate concentrations were optimized, and the results are shown in Fig. S1B. The SERS signal at 1586 cm−1 for 4-MBA exhibited the highest intensity at a concentration of c. Therefore, a substrate concentration of c was selected for subsequent detection.

The detection of Cu2+ was carried out via an etching reaction, which was a quantitative process. KI operated as a pivotal reactant, significantly influencing the progression of the etching reaction and subsequently affecting the Cu2+ detection results. From Fig. S1C, it could be observed that there were no significant changes in the SERS signal peak intensity at 1586 cm−1 under varying KI concentrations, indicating that the concentration of KI had no significant effect on the reaction dynamics. This could be attributed to the substantial disparity between the concentration of Cu2+ to be detected and that of KI. Therefore, any KI concentration could be utilized to react with the target Cu2+. For this study, a KI concentration of 10−3 M was selected.

The reaction time affected the completeness of a reaction. After the reaction was complete, there would be no significant change in the SERS signal. From Fig. S1D, after 10 min of reaction, the SERS signal tended to stabilize, indicating that the reaction was complete. Therefore, 10 min was selected as the optimal reaction time.

Selectivity for the Au@Ag NP-4MBA-KI Detection System

To validate the selectivity and anti-interference capability of the SERS sensor towards Cu2+, we examined the effects of various metal cations, including Co2+, Fe3+, Fe2+, Mn2+, Ni2+, Zn2+, Hg2+, and Cd2+, as shown in Fig. 6. Notably, the introduction of Cu2+ initiated the etching of the Ag shell layer, leading to a reduction in the signal enhancement capability of the SERS substrate, which resulted in a significant decrease in the SERS intensity at 1586 cm−1. Interestingly, the presence of other metal cations did not yield a significant impact on the generation of SERS signals from 4-MBA molecules, indicating that the SERS sensor had strong anti-interference capabilities and selectively detected Cu2+.

Fig. 6
figure 6

Selectivity and anti-interference capability of the sensor for Cu2+ detection

Full size image

Quantitative SERS Detection of Cu2+

To evaluate the sensitivity and potential applicability of the proposed method for quantitative analysis, we conducted measurements using various concentrations of Cu2+ standard samples. Under the optimal conditions, our method displayed a linear correlation between Cu2+ concentrations and SERS intensity within the range of 10−9 to 10−4 M with a LOD of 1.108 × 10−9 M. This correlation was well-described by the equation Y = 1047.72072X + 15,026.12795 (R2 = 0.999), as depicted in Fig. 7, which illustrated the connection between SERS signal intensity and Cu2+ concentration. To provide context, the experimental results of other methods for measuring Cu2+ are listed in Table S1, indicating the simplicity of our method.

Fig. 7
figure 7

A SERS spectra of Au@Ag NP-4-MBA-I− system with different concentrations of Cu2+ (from a to g are 10−4, 10−5, 10−6, 10−7, 10−8, 10−9, and 0 M) and B plot of corresponding intensity

Full size image

Reproducibility of Cu2+ Detection

To evaluate the reproducibility, a total of 20 data points were collected through random sampling under the established optimal experimental conditions. The outcome of this assessment is shown in Fig. 8. It was worth noting that the SERS intensity at 1586 cm−1 is relatively uniform across the data points, and the relative standard deviation of the 20 data sets remains below 5%, which served to validate the reproducibility of the proposed SERS sensor methodology.

Fig. 8
figure 8

A SERS intensity at 1586 cm−1 measured from 20 random positions and B series of SERS spectra of a 10−5 M Cu2+ solution from 20 random positions

Full size image

Analysis of Cu2+ in Tap Water Samples

To demonstrate the practicality of the developed SERS method, we utilized it to quantify Cu2+ concentrations present in tap water under the optimized conditions using the standard addition method. Different concentrations of Cu2+ were tested three times, and the results are shown in Fig. S2. Furthermore, we conducted a comprehensive comparative assessment involving the developed SERS method and the conventional ICP-MS method; the pertinent data are shown in Table 1. Remarkably, the Cu2+ concentrations detected by the SERS sensor are consistent with those measured using the ICP-MS method. The recovery rates for Cu2+ ranged from 93.68 to 100.91%. This highlighted the reliability and practical applicability of our method for the accurate detection of Cu2+ levels in tap water.

Table 1 Cu2+ detection in tap water samples via the SERS and ICP-MS methods
Full size table

Conclusion

In summary, we introduced a novel SERS sensor based on an etching reaction mechanism for the detection of Cu2+. The SERS detection substrate, Au@Ag NPs, coupled with the SERS probe molecule 4-MBA, played a pivotal role in this method. After adding I− to the system, a specific reaction between I− and Cu2+ was initiated, generating iodine species capable of etching the Ag shell. The extent of Ag shell etching varied with the quantity of iodine produced, subsequently affecting the enhancement effect of 4-MBA molecules varies and influencing the intensity of SERS signals generated at 1586 cm−1. Based on the experimental results, a robust linear correlation between the logarithm of Cu2+ concentration and I0-I was evident within the concentration range of 10−9 to 10−8 M, with a low LOD of 1.108 × 10−9 M. Moreover, the practical applicability of this method extended to the detection of Cu2+ in real water samples, effectively presenting a new method for Cu2+ detection. Whether the substrate and Raman probe molecules fail after the test solution has been left for some time, whether the test data are still accurate, whether the test method is time-sensitive, and the effect of different thicknesses of the silver shell layer and different particle sizes of the gold nuclei on the reaction need to be investigated subsequently.