Introduction

In recent years sensing of metal ion attracted the interest of many research groups due to its various ill effects on environment [1–3]. Of all the toxic metal ions, cadmium gets special attention due to its high toxicity [4–8]. It represents one of the highly toxic metals whose contamination occurs from variety of geological and man made sources [9–11]. Bioaccumulation of cadmium in living organism can be extremely toxic even at low concentration [12–14]. Cadmium poisoning can lead to kidney damage, high blood pressure, hypertension, bone marrow disorder and cancer [15–19]. Classically cadmium sensing include sophisticated techniques such as atomic adsorption spectroscopy [20, 21], ICP (inductively coupled plasma) atomic emission spectroscopy [22] and graphite furnace atomic absorption spectrometry (GFAAS) [23]. Although they are accurate for trace amount detection of the metal ion, they have disadvantages such as high cost and complexity of operation. Thus there is need for analytical method for the detection of cadmium.

Recently, fluorescent chemosensors have received considerable interest due to their distinct advantages such as operational simplicity and high sensitivity [24–31]. However, till date, only a few literature are available for fluorocence sensors for cadmium [32–42]. In the design and development of cadmium selective sensor, presence of zinc in the sample often creates problem. Cadmium and zinc being in the same group of the periodic table have similar properties. Therefore they very often cause similar spectral changes while coordinated with fluorescent sensor [43, 44]. Thus it is desirable to prepare fluorescence sensors which can distinguish cadmium from zinc with high selectivity and sensitivity.

Ferrocene based chemosensors found wide applicability in the field of sensor for various ions. When the signaling unit is ferrocene, redox potential is the method of choice due to its low analyte detection limit. In this regard cyclic voltammetry is often applied to measure the change of redox potentials, which can be correlated to the binding strengths of target analyte [45–49]. However, the fluorescent activity of ferrocene based molecules for the ion sensor have not been studied explicitly [50–54]. To the best of our knowledge there is not any reference of ferrocene based fluorescent sensor for cadmium till date.

Bearing that in mind we have synthesized a series of ferrocenyl Schiff bases (1) which show fluorescent property (Fig. 1). In this paper we report the selective sensing of cadmium over a wide variety of competing ion including zinc by compound 1b using fluorocence spectroscopy.

Fig. 1
figure 1

Molecular structures of 1a, 1b, 1c and 1d

Full size image

Experimental

Synthesis

The compound 1b, 1c and 1d were synthesized by usual procedure of Schiff base synthesis. To a solution of ferrocenecarboxaldehyde (1 mmol) in dry methanol, 4-aminothiophenol (1 mmol) was added and the reaction mixture was stirred for 1 h. The red precipitate was filtered and recryatalize from DCM/Hexane mixture to afford pure ferrocenyl imine as dark red solid. Compound 1a was synthesized according to literature procedure [55]. Ferrocenecarboxaldehyde (0.22 g, 1.0 mmol) and benzylamine (0.21 g, 2.0 mmol) in presence of catalytic amount of K2CO3 were dissolved in tolune (10 ml) and reflux under Dean-Stark for 24 h. Then 4Ǻ molecular sieves were added to the hot reaction mixture and stirring was continued for another 48 h at room temperature. Then the reaction mixture was filtered and the filtrate was evaporated to dryness. The solid residue was washed with hexane and dried in vacuo.

N -[(E)-ferrocenylmethylidene]aniline (1a): Dark red solid. Yield 77 %. Mp: 105–106 °C. IR (KBr, cm−1): ν 1430, 1691, 3022. 1H NMR (300 MHz, CDCl3, TMS): δ 4.29 (s, 5H), 4.53 (s, 2H), 4.86 (s, 2H), 6.65–6.92 (m, 2H), 7.15–7.35 (m, 3H), 7.37–7.55 (m, 1H) 8.38 (s, 1H). 13C NMR (75 MHz, CDCl3, TMS): δ 68.9, 69.1, 71.2, 80.1, 120.47, 125.1, 129.0, 152.6, 161.3.

4-{[( E )-ferrocenylmethylidene]amino}benzenethiol (1b): Red Solid. Yield 80 %. Mp: 160–161 °C. IR (KBr, cm−1): ν 1442, 1690, 2587, 3071. 1H NMR (300 MHz, CDCl3, TMS): δ 1.71 (s, 1H), 4.25 (s, 5H), 4.51(s, 2H), 4.80 (s, 2H), 7.10 (d, J = 9 Hz, 2H), 7.52 (d, J = 9 Hz, 2H), 8.33 (s, 1H). 13C NMR (75 MHz, CDCl3, TMS): δ 69.1, 69.3, 71.5, 80.1, 131.4, 130.1, 133.4, 152.4, 161.8. LC-MS (m/z %): 321 [M+] (18), 279 (15), 215 (32), 157 (55), 65 (100).

3-{[( E )-ferrocenylmethylidene]amino}benzenethiol (1c): Red solid. Yield: 78 %. Mp: 130–133 °C. IR (KBr, cm−1): ν 1452, 1655, 1569, 3078. 1H NMR (300 MHz, CDCl3, TMS): δ 1.83 (s, 1H), 4.16 (s, 5H), 4.51 (s, 2H), 5.01 (s, 2H), 7.30–7.54 (m, 3H), 7.80–8.05 (m, 2H). 13C NMR (75 MHz, CDCl3, TMS): δ 68.6, 70.4, 70.7, 80.1, 121.3, 122.2, 124.3, 126.0, 134.7, 153.9, 169.7. LC-MS (m/z %): 322 [M++H] (20), 320 (100), 301 (9), 287 (12).

2-{[( E )-ferrocenylmethylidene]amino}benzenethiol (1d): Red Solid. Yield: 70 %. Mp: 119–120 °C. IR (KBr, cm−1): ν 1498, 1677, 2569, 3033. 1H NMR (300 MHz, CDCl3, TMS): δ 1.76 (s, 1H), 4.22 (s, 5H), 4.50 (s, 2H), 4.78 (s, 2H), 6.40–7.51 (m, 4H), 8.26 (s, 1H). 13C NMR (75 MHz, CDCl3, TMS): δ 69.29, 69.6, 71.47, 80.0, 117.3, 118.9, 119.9, 123.8, 129.7, 137.7, 162.2. LC-MS (m/z %): 322 [M + H] (38), 321 [M+] (37), 320 (100).

Results and Discussion

The structures of the compounds were determined using Gaussian03/DFT. To optimize the structure, density functional calculations were carried out with the Gaussian03/DFT series of programs [56]. The B3LYP formulation [57, 58] of density functional theory was used employing the LANL2DZ basis set. No symmetry constraints were imposed on the molecules. No solvent modeling was employed. Eclipsed conformation of the two cyclopentadienyl ring for compound 1a, 1c and 1d was observed on optimization. However the ferrocene moiety 1b takes a staggered conformation (Fig. 2).

Fig. 2
figure 2

Computer generated optimized structures of (a) N-[(E)-ferrocenylmethylidene]aniline (1a), (b) 4-{[(E)-ferrocenylmethylidene] amino}benzenethiol (1b), (c) 3-{[(E)-ferrocenylmethylidene]amino} benzenethiol (1c), (d) 2-{[(E)-ferrocenylmethylidene]amino} benzenethiol (1d)

Full size image

In the first set of experiment we investigated the complexation of compound 1b and Cd2+ in detail. UV–vis spectra of compound 1b were recorded at different added concentration of Cd2+. In a solution of MeOH: H2O (1:1) compound 1b shows two peaks in UV–vis spectrum at 325 nm (prominent) and 240 nm. In the presence of increasing amount of Cd2+ ion the absorption at λmax = 325 nm constantly decreases whereas that at λmax = 240 nm increases. Moreover, on gradual addition of Cd2+ a new peak appears at 230 nm. The spectra show only one isobestic point at 285 nm probably due to the equilibrium between compound 1b and compound 1b + Cd2+ (Fig. 3).

Fig. 3
figure 3

Absorption spectra of compound 1b at different amount of Cd2+ ion (0–9 × 10−4 M) in MeOH and H2O mixture (1:1)

Full size image

The fluorescence sensing property of compound 1b towards Cd2+ was carried out in a mixed solution of MeOH and H2O (1:1). Free compound 1b shows weak fluorescence emission at 290 nm upon excitation at isobestic point of 280 nm. However, when Cd2+ ion was added to the above solution, we observed a significant increase in the fluorescence emission at 290 nm. To inspect the specificity towards Cd2+, next we examined the fluorescence response of compound 1b in presence of various metal ions. This can be observed by plotting ∆F/Fo verses various metal ions (where Fo is the fluorescence of 1b in absence of metal ion and ∆F = Fm-Fo (Fm = fluorescence of 1b in presence of metal ion). As seen in the Fig. 4, the ∆F/Fo value of 1b is 1.14 in presence of Cd2+, while Zn2+ shows a low ∆F/Fo value of 0.29 under the same condition. In addition, Co, Mn, Pb and Ni show very little fluorescence increase under our experimental condition. In contrast, compound 1b displays a fluorescence quenching effect with Cu, Hg, Fe and Ag.

Fig. 4
figure 4

Fluoroscence change (∆F/F0) of compound 1b at 290 nm in the presence of different metal ions in MeOH-H2O (1:1) mixture ([compound 1b] = 0.001 M, [Mn+] = 2.0 × 10−4 M, λex = 280 nm)

Full size image

To scan the signal response towards Cd2+, the fluorescence emission of compound 1b were tested in different concentration of the metal ion (Fig. 5). On stepwise addition of Cd2+ ion, the fluorescence intensity of compound 1b at 290 nm was gradually increased. The fluorescence intensity of compound 1b increased linearly with Cd2+ upto 1:1 mole ratio and then remains constant with increasing Cd2+ ion, indicating that compound 1b and Cd2+ undergoes 1:1 complex formation.

Fig. 5
figure 5

Fluorescence spectra of compound 1b in presence of different amount of Cd2+ ion (0–9.9 × 10−4 M) in a mixed solution of MeOH and H2O (1:1). Inset: Fluoroscence intensity as a function of Cd2+ concentration

Full size image

The fact that 1:1 complexation occurs between ligand and the metal is further supported by Hill plot of the fluorescence data as shown in the Fig. 6. The plot of log{Imax-I)/(I-I0)} Vs log [Cd2+] is linear with a slope of 1.1735 and intercept 4.0905 indicating binding of one Cd2+ ion per ligand with log β = 4.09.

Fig. 6
figure 6

Hill plot for the complexion of compound 1b with Cd2+ ion from fluoroscence. I0, I and Imax are fluorescence intensity of compound 1b at zero, at an intermediate and at the infinite concentration of Cd2+ ion respectively

Full size image

For a deeper insight into the relationship between selectivity and structure of 1b, compound 1a, 1c and 1d are prepared. The fluorescence activities of all these compounds in presence of different metal ion were investigated. All the compounds have different response to the metal ions.

Compound 1a and 1c do respond to the presence of different metal ion in their respective fluoroscence, but its selectivity is very low (Figs. 7 and 8). The metal sensing property of 1a (Figs. 7a and 8a) indicates that -SH is not a binding site for the metal ion, however, its presence and position greatly influence the selectivity of the ligands.

Fig. 7
figure 7

Fluoroscence spectra of (a) compound 1a and (b) compound 1c in presence of different amount of Cd2+ ion

Full size image
Fig. 8
figure 8

Fluoroscence change (∆F/F0) of (a) compound 1a and (b) compound 1c in presence of different metal ion

Full size image

To examine the binding mode of cadmium with 1d, 1H NMR spectra was measured in MeOD: D2O (1:1) solution (Figure S1). Upon interaction with Cd2+ the iminium proton experienced approximately 0.08 ppm of downfield shift to 8.51 ppm. A similar trend was also observed for the aromatic protons. This shift of the signal can be attributed to deshielding effect of the metal ion on the protons of its proximity. It is possible only when cadmium ions bind to the nitrogen atom. From this study we have confirmed that nitrogen atom of the molecule coordinate with Cd2+ ion to form 1d/Cd2+ complex.

Interestingly, the performance of compound 1d (Fig. 9), towards selective sensing of Cd2+ in presence of Zn2+ is better than that of compound 1b. However, the fluorescence of 1d is very low makes it unsuitable to use as sensor.

Fig. 9
figure 9

Fluoroscence change of compound 1d (a) In presence of different amount of Cd2+ ion; (b) in presence of different metal ion

Full size image

The fluorescence enhancement of compound 1b in presence of Cd2+ can be explained on the basis of PET process. In absence of Cd2+ ion there is a flow of electron density from nitrogen atom to the fluorophore part, which quenches the fluorescence of compound 1b. The Cd2+ is likely to bind through N atom of compound 1b. As a result the energy of the HOMO of the receptor part of compound 1b lowered and consequently the oxidation potential of receptor unit increases whish effectively stops the PET process and accordingly fluorescence intensity increases.

The driving force for the electron transfer from the receptor unit to the fluorophore unit in the PET process can be expressed by modified Weller equation as follows.

$$ \varDelta {\mathrm{G}}_{\mathrm{et}}=-{\mathrm{E}}_{\mathrm{s}}-{\mathrm{E}}_{\mathrm{red}.\mathrm{Fluo}}+{\mathrm{E}}_{\mathrm{ox}.\mathrm{receptor}} $$

Where Es, Ered.fluo and Eox.receptor represent the singlet energy, reduction potential of the fluorophore and the oxidation potential of the receptor unit respectively. If the binding of the cation increases the oxidation potential of the receptor unit than this process will increase the ∆Get, which will decrease the PET process and consequently it allows the excited chromospheres to relax by fluorescence.

Cyclic voltammogram of compound 1b was recorded in MeOH in absence and presence of equivalent amount of Cd2+ ions using GC as working electrode and Ag-AgCl as reference (Fig. 10). Both the oxidation as well as reduction potential changes in the positive direction. The oxidation potential changes from 0.6 to 0.7 V while the reduction potential increases from 0.52 to 0.58 V. Therefore from the Weller equation it is clear that the binding of Cd2+ ion increases the ∆Get value and consequently fluorescence intensity increases.

Fig. 10
figure 10

Cyclic voltammetry of (a) 1b at different scan rate (10–500 mV/sec) (b) shift of CV peaks of 1b in presence of Cd2+

Full size image

Conclusion

In summery we have developed a structurally simple ferrocene derivative as a sensitive fluorescent probe for selective detection of Cd2+. It is excited at 280 and emit at 290 nm and displays a significant fluorescent enchantment after complexation with Cd2+. Presence of –SH group in para-position of the aromatic ring greatly enhances the fluorescence intensity and selectivity for Cd2+ over other competitive metal ion including Zn2+.