A highly selective and sensitive fluorescent sensor based on a 1,8-naphthalimide with a Schiff base function for Hg²⁺ in aqueous media

2.1 Materials and methods

All chemicals and solvents used in the synthesis were obtained from commercial sources and applied without further purification. Dilute hydrochloric acid or sodium hydroxide were used for adjusting pH values. Britton–Robinson buffer was prepared with 40 mm acetic acid, boric acid and phosphoric acid. Tris-HCl buffer (pH = 7.2) was prepared using double distilled water. The metal cation sources, NaNO₃, KNO₃, CaCl₂, MgSO₄, AlCl₃, Pb(NO₃)₂, Fe(NO₃)₃·9H₂O, Ni(NO₃)₂·6H₂O, Zn(NO₃)₂·6H₂O, Cu(NO₃)₂·5H₂O, Hg(NO₃)₂·H₂O, AgNO₃, Co(NO₃)₂·6H₂O, Cr(NO₃)₃·9H₂O, 50%Mn(NO₃)₂, Cd(NO₃)₂·H₂O were analytical reagent grade and were dissolved by using double distilled water.

C, H and N elemental analyses were determined using a Carlo Erba 1106 elemental analyzer. ¹H NMR spectra and ¹³C NMR spectra were obtained with a Varian VR300 MHz spectrometer with TMS as internal standard. Electrospray ionization mass spectra (ESI-MS) were obtained on a Bruker micrOTOF-Q system. The IR spectra were recorded in the 4000–400 cm⁻¹ region with a Nicolet FT-VERTEX 70 spectrometer using KBr pellets. Electronic spectra were recorded on a LabTech UV Bluestar spectrophotometer. Fluorescence spectra were recorded with an F-7000 FL spectrophotometer, which had been subjected to absorbance normalization. The pH values were measured with a Delta 320 pH meter. The melting points were measured in a X-4 microscopic melting point apparatus. TLC was performed on silica gel, Fluka F60 254, 20 × 20, 0.2 mm. All the detections were carried out at T = 25 °C.

2.2 Synthesis and characterization of HL

The detailed synthetic procedure of HL is shown in Scheme 1. The intermediate products 1–2 were prepared according to the procedure previously reported [39, 40]. The mixture of N-allyl-4-(2-aminoethyl)amino-1,8-naphthalimide (2) (1 g, 3.4 mmol), 0.77 g (5.1 mmol) 3-methoxysalicylaldehyde and 40 mL of ethanol was heated to 80 °C for 4 h. After the mixture was cooled to room temperature, the precipitate produced was filtered, washed with ethanol and dried to give 1.2 g of final product N-allyl-4-(ethylenediamine-3-methoxysalicylidene)-1,8-naphthalimide Schiff base (HL). The progress of the reaction was monitored by TLC (the TLC eluent being dichloromethane–acetone = 1:1, R_f = 0.5). Yield: 1.2 g, 83%; m. p. 175–176 °C. – IR (KBr pellet, cm⁻¹): ν = 1694(s), 1645(s), 1633(m), 1584(s), 1474(s), 1340(m), 778(m), 668(m) – UV/Vis (DMF): λ = 271, 341, 437 nm – ESI-MS (C₂₅H₂₃N₃O₄): m/z = 428.28 [M−1]⁻. – ¹H NMR (400 MHz, DMSO-d₆): δ = 8.65 (d, 1H, J = 8.4 Hz, Ar-OH), 8.53 (s, 1H, –N=CHAr), 8.42 (d, 1H, J = 7.2 Hz, Ar-H), 8.25 (d, 1H, J = 8.4 Hz, Ar-H), 7.67 (t, 2H, J = 15.6 Hz, Ar-H), 7.01 (d, 1H, J = 8 Hz, Ph-H), 6.96 (d, 2H, J = 7.6 Hz, Ph-H), 6.91 (d, 1H, J = 8.8 Hz, Ar-H), 6.77 (t, 1H, J = 15.6 Hz, Ar-NH), 5.88–5.96 (m, 1H, allyl-H), 5.06–5.11 (m, 2H, allyl-H), 4.62 (d, 2H, J = 4.4 Hz, allyl-CH₂), 3.95 (t, 2H, J = 10.8 Hz, NH-CH₂), 3.75 (s, 2H, N-CH₂), 3.77 (s, 3H, Ph-OCH₃) ppm. – ¹³C NMR (400 MHz, DMSO-d₆): δ = 167.06, 163.34, 162.69, 151.67, 150.47, 148.00, 134.10, 133.23, 130.66, 129.37, 128.46, 124.28, 123.12, 121.27, 120.14, 118.29, 117.54, 116.72, 114.77, 107.80, 104.09, 56.14, 55.71, 43.38, 41.24 ppm. – Elemental analysis for C₂₅H₂₃N₃O₄ (429.48): calcd. C 69.92, H 5.41, N 9.78; found C 69.90, H 5.43, N 9.79% (See Supplemental Figures S1–S3 in the Supporting Information available online).

Scheme 1:

The synthetic route leading to HL and the structures of the intermediates and of the product.

3.1 Photophysical characteristics of HL

The ability of HL to emit absorbed light energy is characterized quantitatively by the quantum yield of fluorescence Ф_F. The fluorescence quantum yield has been calculated according to equation (1) using N-butyl-4-n-butylamino-naphthalimide (Ф_F = 0.81 in ethanol) as a standard. Herein, Ф_ref is the emission quantum yield of the standard, A_ref and A_sample represent the absorbance of the standard and the sample at the excited wavelength, respectively, while S_ref and S_sample are the integrated emission band areas of the standard and the sample, respectively, and n_ref and n_sample are the solvent refractive indeces of the standard and sample, respectively [41].

The photophysical properties of the substituted 1,8-naphthalimides are known to depend mainly on the polarization of their chromophoric system [42]. Light absorption in this molecule generates a charge transfer interaction between the substituents at C-4 position and the imide carbonyl groups. The photophysical characteristics of HL were investigated in DMF, acetonitrile, tetrahydrofuran and dichloromethane solution. Moreover, the Stokes shift (ν_A–ν_F) is an important parameter for fluorescent compounds that indicates the difference in the properties and structure of the fluorophore between the ground state S₀, and the first exited state S₁. The Stokes shift (cm⁻¹) has been calculated by the equation (2) [43, 44]. The absorption (λ_A) and fluorescence (λ_F) maxima, the extinction coefficient (ε), the Stokes shift (ν_A–ν_F), and quantum fluorescence yield (Ф_F) of HL are presented in Table 1.

From the data in Table 1, it can be seen that the fluorescence quantum yield Ф_F of HL in polar solvents is lower than in non-polar solvents. This fact can be explained by the photoinduced electron transfer process, which is favored in polar solvents while the fluorescence emission is lower [45]. The Stokes shift values for HL under study were in the 3535–4042 cm⁻¹ region. It appears that the Stokes shift values depend on the media, and that they are larger in the case of polar solvents when hydrogen bond formation or dipole-dipole interactions are favored in comparison to non-polar media, which is in a good accordance with other investigations on 1,8-naphthalimide derivatives [46, 47]. HL shows an absorption band with maximum range of about 422–437 nm that is typical for 1,8-naphthalimides substituted in C-4 position with alkylamines. The respective fluorescence maxima appear in the range 496–523 nm. The molar extinction coefficient (ε) of HL under study in the long-wavelength band of the absorption spectra ranges from 11,338 to 24,885 cm⁻¹, meaning that ε is larger in polar solvents. In summary, the polarity of the organic solvents is of great importance for the photophysical properties of HL under study and especially for the quantum fluorescence yield and the Stokes shift.

3.2 Investigation of the pH effect on the sensing performance of HL

As the pH of a system is considered to have a significant influence on the performance of a sensor, the spectroscopic characters of HL were investigated in the pH range 1.81–11.82 (Britton–Robison/DMF buffer, 1:1, v/v). As seen from Figure 1, the fluorescence intensity of HL is relatively stable in the range of pH from 5.72 to 9.15, implying that the performance of HL is actually pH-independent in this range, meaning that the design of the sensor is reasonable. All subsequent experiments were performed in solution at pH 7.2.

Figure 1:

Fluorescence behavior of HL in solution (Britton–Robison/DMF buffer, 1:1, v/v) at different pH values (excitation at 430 nm, T = 25 °C).

3.3 Selectivity and interference investigation of HL

High selectivity for a specific analyte in the presence of competing species is another important feature of a sensor. The signaling fluorescence properties of HL (10 µm) in the presence of various metal cations (Na⁺, K⁺, Ca²⁺, Mg²⁺, Al³⁺, Pb²⁺, Fe³⁺, Ni²⁺, Zn²⁺, Cu²⁺, Hg²⁺, Ag⁺, Co²⁺, Cr³⁺, Mn²⁺ and Cd²⁺) were investigated in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2). The fluorescence changes of HL are depicted in Figure 2a. It is found that among the various transition metal ions only Hg²⁺ causes a significant fluorescence enhancement of HL (Ф_Hg+HL/Ф_HL = 2.28) with the emission maximum at 526 nm. These results indicate that HL could act as a highly selective fluorescence sensor for Hg²⁺ in the aqueous medium (pH = 7.2). To validate the higher selectivity of HL for Hg²⁺ relative to other metal ions, fluorescence competitive experiments of Hg²⁺ with other metal cations (Na⁺, K⁺, Ca²⁺, Mg²⁺, Al³⁺, Pb²⁺, Fe³⁺, Ni²⁺, Zn²⁺, Cu²⁺, Ag⁺, Co²⁺, Cr³⁺, Mn²⁺ and Cd²⁺) were investigated in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2) and their fluorescence intensities were recorded. As shown in Figure 2b, when excited at 430 nm, the fluorescence intensity of other coexisting ions shows a similar pattern to that with only Hg²⁺, indicating that the HL–Hg²⁺ system was hardly affected by these coexistent ions. According to the above results, HL is a reliable highly selective sensor for Hg²⁺ in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2) with high binding affinity, suggesting the potential for wider use in practical applications.

Figure 2:

(a) Fluorescence spectra of HL (10 μm) in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2) towards Na⁺, K⁺, Ca²⁺, Mg²⁺, Al³⁺, Pb²⁺, Fe³⁺, Ni²⁺, Zn²⁺, Cu²⁺, Hg²⁺, Ag⁺, Co²⁺, Cr³⁺, Mn²⁺, and Cd²⁺ (2 eq) with an excitation at 430 nm. (b) Competitive experiments in the compound HL + Hg²⁺ system with interfering metal ions at 526 nm.

3.4 Fluorescence titration of HL towards Hg²⁺

In order to obtain insight into the binding properties of HL (10 µm) with Hg²⁺, fluorescence titration experiments were carried out at various concentrations of Hg²⁺ in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2) at room temperature (Figure 3a). Upon addition of increasing amounts of Hg²⁺, the fluorescence intensity of HL gradually increased. When the amount of Hg²⁺ added was about 5 × 10⁻⁶ m, the fluorescence intensity approached a maximum. As more Hg²⁺ was titrated, the fluorescence intensity showed negligible changes. No shifts of the emission spectra with increasing Hg²⁺ concentration was observed, which indicated that there was no integration of the fluorophore and receptor in the excited state. The nonlinear curve fitting of the fluorescence titration gives a 2:1 stoichiometric ratio between HL and Hg²⁺ (Figure 3b). The association constant is determined by the Benesi-Hildebrand equation [48, 49]:

Figure 3:

(a) Fluorescence titration spectra of HL with Hg²⁺ (0–1.5 eq) in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2) at room temperature with an excitation at 430 nm. (b) Fitting of the fluorescence titration curve of HL in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2). (c) Benesie–Hildebrand linear analysis plots of HL at different Hg²⁺ concentration. (d) Curve of fluorescence intensity at 526 nm of HL (1 × 10⁻⁵ m) versus increasing concentrations of Hg²⁺ (0.5–4.0 μm).

Herein, F is the fluorescence intensity at 526 nm at any given Hg²⁺ concentration, F₀ is the fluorescence intensity at 526 nm in the absence of Hg²⁺, and F_m is the maximum fluorescence intensity at 526 nm in the presence of Hg²⁺ in solution, and n is the stoichiometric mole ratio, which is 2 in this case. From the fluorescence titration of Hg²⁺ ion, the association constant for L and Hg²⁺ has been calculated to be 1.56 × 10¹² m⁻¹ and evaluated graphically by plotting log[(F − F₀)/(F_m − F)] against log[Hg²⁺] (Figure 3c). Figure 3d also displayed the good linearity between the emission at 526 nm and concentrations of Hg²⁺ in the range from 0.5 to 4 μm, indicating that HL can detect quantitatively relevant concentrations of Hg²⁺.

The detection limit based on the definition by IUPAC was calculated to be 0.18 µm with the following equation [50]: C_DL = 3 σ/k, where σ is the standard deviation of the blank solution, k is the slope of the intensity versus the sample concentration. This value is far below U.S. EPA-regulated limits of 31.5 μm [51] and it confirms the ability of HL to monitor low concentrations of Hg²⁺ ions commonly encountered in both environmental and physiological systems.

3.5 Mechanism of recognition of HL towards Hg²⁺

Even though the stoichiometric ratio of 2:1 between HL and Hg²⁺ was obtained by the above-mentioned fluorescence titration experiment, additional evidence for the exact ratio needed to be provided by the Job’s method. As shown in Figure 4, the concentration of the sensor L and Hg²⁺ reached a turning point when the molar fraction of [Hg²⁺]/[L + Hg²⁺] was about 0.345 at 526 nm, indicating a high binding tendency between HL and Hg²⁺ with a 2:1 stoichiometry as expected.

Figure 4:

Job’s plot at 526 nm for determining the stoichiometry of HL and Hg²⁺ in solution (DMF/Tris-HCl buffer, 1:1, v/v, pH = 7.2).

Based on the fluorescence titration spectra and the Job’s plot, we propose a possible binding model of HL and Hg²⁺ as shown in Scheme 2 [52, 53]. The fluorescence enhancement of the HL response to Hg²⁺ could be attributed to photoinduced electron transfer (PET) [12, 54], [55], [56]. Before binding with Hg²⁺, the moderate fluorescence intensity of HL is probably due to the lack of a lone pair of electrons with suitable energy at the O-hydroxy Schiff base group, therefore causing sufficient intramolecular photoinduced electron transfer (PET) [57], [58], [59], [60], [61], [62], [63]. However, when HL was bound to Hg²⁺, the PET process was blocked and the complex was more rigid, thus a significant enhancement of fluorescence was observed.

Scheme 2:

Proposed binding model of HL with Hg²⁺ to give L₂Hg.