All chemicals used were of analytical grade or of the highest purity available. Chloroauric acid (HAuCl₄) was obtained from Sigma–Aldrich (USA) and used as received. BSA was obtained from Genview. The used metal salts Pb(NO₃)₂, Ni(NO₃)₂·6H₂O, Cd(NO₃)·4H₂O, FeCl₂·4H₂O, Mg(NO₃)₂·6H₂O, CaCl₂·2H₂O, Co(Ac)₂·6H₂O, Zn(Ac)₂·2H₂O, BaCl₂·2H₂O, CuSO₄·5H₂O, Hg(NO₃)₂·2H₂O, Cr(NO₃)₃·9H₂O, Mn(Ac)₂·4H₂O and NaOH were purchased from Beijing Chemical Reagent Company (Beijing, China). The stock solution of Hg²⁺ was 5 mM. All glassware was thoroughly cleaned with freshly prepared 3:1 HCl/HNO₃ (aqua regia) and rinsed thoroughly with Mill-Q (18 MΩ cm⁻¹ resistance) water prior to use.

The fluorescence spectra were measured using a Cary Eclipse Fluorescence spectrophotometer (USA Varian). Both the excitation and emission slit widths were set to 20 nm, and the measurement was done at 20°C. The excited wavelength was 470 nm.

BSA-GNPs were fabricated by a previously reported method 23. In a typical synthesis, aqueous HAuCl₄ solution (5 mL, 10 mM, 37°C) was added to BSA solution (5 mL, 50 mg/mL, 37°C) under vigorous stirring. Two minutes later, NaOH solution (0.5 mL, 1 M) was introduced, and the mixture was incubated at 37°C for 12 h. The color of the solution changed from light yellow to light brown, and then to deep brown. The deep brown solution of BSA-GNPs emits an intense red fluorescence (Fig. 1b (a), under UV light 365 nm). The changes of fluorescence intensity of BSA-GNPs at various reaction times were showed in Fig. S1. The fluorescent BSA-GNPs show emission peak at 640 nm, excitated by 470 nm (Fig. 1b (a)).

The as-prepared BSA-GNPs were diluted to 40 times, the resulting BSA-GNPs solution was used as the detection probe. The samples were added into the BSA-GNPs probe with a volume ratio (V/V) 4:1, which was the mixture of the 400 μL Hg²⁺ solution and 100 μL BSA-GNPs probe. In the experiments of sensitivity of the detection, the solution of Hg²⁺ was prepared from 0.1 nM to 4 μM. To confirm the practical application of the probes, a water sample from a lake was filtered through a 0.2-μm membrane. A 100 μL of 100 μM Hg²⁺ solution was added into 900 μL lake water, the resulting solution was 10 μM, which was further diluted a series of different concentrations of samples with lake water.

The overall detection strategy is shown in Fig. 1a. BSA is a large globular protein with a good essential amino acid profile and bear fluorescence emission groups (tryptophan, tyrosine and phenylalanine). So, there are lots of amino acid residues on the surfaces of BSA-GNPs, which can be proved by the FTIR spectra (Fig. S2) 24.

To evaluate the attachment of BSA onto the surface of Au, the FTIR spectra of BSA and BSA-GNPs were shown in Fig. S2. Compared to the spectrum of pure BSA, the characteristic peaks of BSA-GNPs such as at 1,654, 2,958 and 3,421 cm⁻¹ can be clearly seen, which indicating the modification of BSA on the Au nanoparticles. Compared with the spectrum of pure BSA (Fig. S2a), the characteristic peaks of BSA-GNPs such as at 1,654, 1,546 cm⁻¹ can be clearly seen (Fig. S2b), which are attributed to the amide I and amide II vibrations of BSA on the Au nanoparticles 2425. The FTIR spectra indicated that the basic structure of BSA still maintained after modified the gold nanoparticles. In addition, the residues of amino acid can readily coordinate with heavy metal ions 2627, when Hg²⁺ coordinated with the BSA-GNPs, the fluorescence emission of BSA-GNPs may be quenched.

The fluorescence emission of the BSA-GNPs was shown in Fig. 1b; the maximum emission and excitation wavelengths were at 640 and 470 nm, respectively. The fluorescence emission was readily quenched in the presence of Hg²⁺. As illustrated in Fig. 1b, the BSA-GNPs solution showed strong emission at 640 nm (curve a), when added the solution of 4 μM Hg²⁺ (curve b), the fluorescence emission was quenched immediately. The inset is the corresponding photograph, the BSA-GNPs solution showed strong red emission under UV light and the red emission of BSA-GNP-probes was quenched by the addition of Hg²⁺. The corresponding TEM images of original and the aggregated BSA-GNPs were shown in Fig. S3. After the addition of Hg²⁺, the dispersed BSA-GNPs with diameter in ca. 2 nm (Fig. S3a) aggregated seriously (Fig. S3b). Therefore, Hg²⁺ detection could be easily realized via monitoring the fluorescence quenching of the BSA-GNPs under the UV light.

To evaluate the detectable minimum concentration of Hg²⁺ in aqueous solution by fluorescence quenching, the various concentrations of 0.1 nM–4 μM Hg²⁺ were added into BSA-GNPs solution, respectively. As shown in Fig. 2, all the maximum fluorescence emissions of the BSA-GNPs were at about 640 nm, and the fluorescence intensity of the BSA-GNPs showed a gradual decrease with the concentration of Hg²⁺ increasing from 0.1 nM to 4 μM. Notely, the fluorescence was quenched completely at about 4 μM. However, the red emission of BSA-GNPs can not be observed after added into 2.5 μM Hg²⁺ under the UV light. The inset is a linear between the fluorescence intensity with the different concentrations of Hg²⁺. There is a well linear relationship over the concentration range from 1 nM to 4 μM with the R² = 0.98876, indicating the ultrasensitive detection of Hg²⁺. The detection limit of the BSA-GNPs probe for Hg²⁺ detection was 0.1 nM, which was much lower than the EPA standard for the maximum allowable level 2 ppb (10 nM) in drinking water. To confirm the rapid response of BSA-GNPs to Hg²⁺, we measured the response time of fluorescence quenching by monitoring the fluorescence changes for 5 min after the addition of Hg²⁺ into the BSA-GNPs (Fig. S4). As a result, the fluorescence quenching by Hg²⁺ was completed almost within 1 min, allowing the rapid detection of Hg²⁺. The experimental results indicate the feasibility of detecting Hg²⁺ accurately using the BSA-GNPs probes.

We investigated the selectivity of our new approach for Hg²⁺ over other metal ions (500 μM Cd²⁺, Cr³⁺, Ba²⁺, Zn²⁺, Ca²⁺, Mn²⁺, Mg²⁺, Fe²⁺,Co²⁺ and 50 μM Ni²⁺, Cu²⁺, Pb²⁺) under the same conditions. As indicated in Fig. 3a, the solution of BSA-GNPs after added other ions has fluorescence emissions at about 640–650 nm. Importantly, the fluorescence of the BSA-GNPs was quenched completely in the presence of Hg²⁺, while almost 100% fluorescence quenching was not observed in the presence of other metal ions. All of the fluorescence intensities of BSA-GNPs after the addition of other ions were much higher than that of Hg²⁺, although the concentration of other ions was 10–100 times more higher than that of Hg²⁺ solution, indicating the BSA-GNPs probe was highly selective for Hg²⁺ over the other metal ions. To further illustrate the selectivity of this BSA-GNPs-based detection system, we used the enhanced ratio to express the intensity ratio, that is, the value of intensity of ion/intensity of control (I/I₀). In the Fig. 3b, the ratio of intensity of fluorescence spectra I/I₀ (640 nm) can be observed clearly. For example, the I/I₀ value of Cd²⁺ (500 μM) was at least 10 times higher than that of Hg²⁺ (4 μM), suggesting the detection limit for Hg²⁺ was almost 100 times higher than Cd²⁺. And the solutions of 50 μM Ni²⁺, Cu²⁺, Pb²⁺ were over 10 times higher than Hg²⁺. By comparison, the decrease in fluorescence intensity was more remarkable in presence of Hg²⁺. The addition of other ions with high concentrations to the BSA-GNPs has some effect on the intensity of fluorescent spectra, but did not occur fluorescence quenching completely. Additionally, we also investigated the red emission of BSA-GNPs added with all the ions under UV light. The BSA-GNP probes were added 50 μM of Cd²⁺, Cr³⁺, Ba²⁺, Zn²⁺, Ca²⁺, Mn²⁺, Mg²⁺, Fe²⁺, Co²⁺, Ni²⁺, Cu²⁺, Pb²⁺ and 4 μM Hg²⁺, respectively. As shown in Fig. 3c, the red emission of BSA-GNPs was quenched completely in presence of Hg²⁺; however, the red emissions of BSA-GNPs upon addition of other ions were observed clearly. The experimental results indicate that only the addition of Hg²⁺ caused fluorescence quenching completely, revealing the exceptional specificity of this present method.

To further evaluate the selectivity, the Hg²⁺ detection in the presence of other metal ions was investigated. The fluorescence responses of BSA-GNPs to different concentrations of Hg²⁺ in mixture solutions were obtained in Fig. S5. The results indicated that the fluorescent intensity of BSA-GNPs still showed a gradual decrease with the concentration of Hg²⁺ increasing coexisting with other metal ions in solution 1. Compared with mixture solution 1 (Fig. S5a), the fluorescent intensity of BSA-GNPs decreased a lot coexisting with Cu²⁺, Pb²⁺ and Ni²⁺ (Fig. S5b), but it did not affect the detection of Hg²⁺. Though other metal ions coexisted, the experimental results show that Hg²⁺ can be detected against other heavy metals.

To demonstrate the potential practical application of the assay to measure the Hg²⁺ content in lake water, a water sample from a freshwater lake on our campus was collected and filtered through a 0.2-μm membrane and then analyzed by ICPMS (Table S1). We did not detect the presence of Hg²⁺ ions in the lake water samples, which was in good agreement with ICP-MS data. To determine the concentration of Hg²⁺, we applied a standard addition method 89132829. The intensity of fluorescence spectra of the samples was decreased with the increase of Hg²⁺ concentrations from 0.1 nM to 4 μM. From Fig. 4, it is observed that although the sample may contain some unknown contamination that did not interfere with the BSA-GNPs-based sensing system detecting aqueous Hg²⁺, these experimental results indicate BSA-GNPs-based sensing system can be utilized for analyzing the practical samples.