2.1. Optical response of 1
To investigate the response of 1 to CN¯ ion, a solution of NaCN was gradually added to CH3CN : H2O buffered solution (7:3, pH 7.4) of 1 and after each addition the absorption and emission spectra were recorded.
The UV–Vis spectra of 1 exhibited two absorption peaks at 290 and 321 nm. Upon incremental increases of CN¯ (0–1.5 equiv.) to a solution of 1, the absorbance peaks were gradually decreased with concomitant increase of the newly appeared peak at 410 nm (Fig. 1A). In addition, a clear isosbestic point at 355 nm was formed, indicating a distinct interaction between 1 and CN¯.
On the other hand, the excitation of compound 1 at 430 nm was led to fluorescence emission at 442 nm. Upon the addition of CN¯ (0–1.5 equiv.) to 1, the intensity of original peaks gradually increased with a virtually unchanged emission shift (Fig. 1B).
These optical CN¯ sensing were also confirmed by the obvious color changes of solutions, as shown in insets of Fig. 1.
The optical response of 1 toward CN¯ fits well with the Stern-Volmer equation to confirm their strong interaction (Figs. 2A and 2B). The binding constants (Ka) were calculated using the Benesi-Hildebrand equation and found 2.8 x 103 and 1.0 x 105 M− 1 by UV-vis and Fluorescence methods, respectively.
The binding mode of 1 with CN¯ was measured and gave a 1:1 stoichiometry by Job plot analysis (Figs. 2C and 2D). According to calculations, a high sensitivity for fluorescence (LOD = 1.2 nM) and UV–vis (LOD = 2.3 nM) methods were obtained which are much lower than the maximum allowable level of CN¯ ions in drinking water (1.9 µM) set by the WHO.6
To evaluate the interference of CN¯ with the following sodium anions (NO2−, NO3−, SCN¯, HS¯, S2−, Br¯, Cl¯, F¯, I¯, H2PO4¯, IO3−, IO4−, ClO4−, BrO3¯, ClO3−, MoO42−, SO32−, S2O32−, S2O42−, S2O52−, SO42−) and chloride cations (Li+, Na+, Ca2+, Ba2+, Sr2+, K+, Mg2+, Al3+, Cu2+, Cd2+, Co2+, Fe3+, Ni2+, Hg2+, Ag+, Mn2+, Pb2+, Zr4+), competitive experiments by 1 were investigated which showed no interference between CN¯ and other anions (Figs. 3A and 3B).
Upon addition of different sodium anions to solutions of 1, it was distinct that other anions, except CN¯, induced no visual and emission color changes (Figs. 3C and 3D). The high selectivity of CN¯ over chloride metal ions confirmed that the deprotonation process is superior to metal complexation (Fig. S4).
2.2 Optical response of 2
The binding ability of 2 toward CN¯ ions were also studied by UV–vis and fluorescence spectroscopy (Fig. 4).
When 20 equiv. of CN¯ was gradually added to 2, a new peak at 417 nm was observed with an isobestic point formation and a pale yellow color appearance (Fig. 4A).
Upon addition of CN¯ ions (0.2 µM) to a solution of 2 (0.015 µM), the fluorescence peak at 466 nm was surprisingly blue-shifted to 446 nm (Fig. 4B). This non-linear emission response 2 to CN¯, as shown in above inset of Fig. 4B, suggested the different behavior of 2 with that of 1.
The binding mode of 2 with CN¯ was measured and gave a 1:1 stoichiometry by Job plot analysis (Fig. S5). The calculated detection limit by fluorescence (2 µM) is remarkably lower than that of UV–vis (10 nM). Emission intensity ratio (I466/I446) method was found to improve the ratiometric detection of CN¯ (0.9 µM, Fig. S6). This unexpected detection limit is due to formation a new emissive product (see Sect. 3.3.3 and Fig. 10).
In a further experiment, the competitive experiments showed that there is no interference between CN¯ and other ions in a solution of 2, as shown in Figs. 5A and 5B. Moreover, the significant color change of 2 in the presence of CN¯ is observable under 365 nm UV light, as shown in Figs. 5C and 5D.
Furthermore, probe 2 shows a highly selectivity of CN¯ over metal ions due to the unique reaction of CN¯ on 2 (Fig. S7).
2.3 Practical Application
2.3.1 pH response
First, the ability of both receptors for sensing of CN¯ was evaluated in the various ranges of the buffer solutions.
As shown in Fig. 6 and Fig. S8, a pH range of 5.0–9.0 is allowed for CN¯ analysis. Accordingly, we set all measurements at buffer pH = 7.4 which is applicable for biological samples.
2.3.2 Reversibility experiment
To determine the reversibility of solution of 1 toward CN¯, 1 equiv. of HCl was added, leading to disappearance of its yellow color to colorless solution (Fig. 7A). This color change in the presence of HCl is also confirmed by the disappearance of the absorbance peak at 417 nm, suggesting the reversible deprotonation-protonation cycle in even after several cycles.
In contrast, probe 2 shows the irreversible behavior under the same experiment to suggest a nucleophilic reaction was taken place on it.
The reversibility of probe 1 was also studied on silica gel plates. As shown in Fig. 7B, a colorless test strip was prepared by immersing TLC plate into MeOH solution of 1 (10− 4 M). When this colorless paper was immersed into solution of CN¯ (10− 4 M), a yellow color appeared on it. This yellow color subsequently returned back to original color of 1 when plate immersed into solution of HCl (10− 4 M). These results showed that sensor 1 can work well in both solution and solid state as portable kits for CN¯ analysis.
2.3.3 Solution and solid kit tests
Furthermore, the performance of 1 and 2 for qualitative detection of CN¯ was evaluated. As shown in Fig. 8, the gradual color change from colorless to yellow with the addition of various concentrations of CN¯ was detectable by both probes.
The visual color changes are accompanied with gradual increasing of absorption peaks of 1 in 30% aqueous CH3CN at 417 nm. However, it was hard to trace visual color change of 2 in 30% aqueous CH3CN. In contrast, an improvement in color changes was observed for 2 when the measurements were carried out in 30% aqueous MeOH at 375 nm (Fig. 8B and 8D). However, its calculated detection limit is weak (2.0 µM).
2.4 Proof of the sensing mechanism
2.4.1 1H NMR measurements
To get insight into the binding interactions between both sensors with CN¯, 1H NMR measurements were performed in the presence of NaCN, as shown in Fig. 9.
When 1 equiv. of CN¯ was added to the solutions of 1 or 2 in DMSO-d6, the corresponding signals of the OH protons entirely disappeared together with upfield shift for aromatic protons, indicating the formation of strong hydrogen bonds between CN¯ and OH groups followed by deprotonation.
In contrast, the imine protons of 1 and 2 showed the different response in the presence of CN¯. The imine protons of 1 were slightly shielded from 8.7 to 8.5 ppm, while those of 2 at 8.9 ppm disappeared and a broad peak at 4.5 ppm, assigned to the amine protons, appeared as attributed to the nucleophilic addition of CN¯ (Fig. 9B).
2.4.2 NaOH experiment
To support this finding, the interaction between 1 or 2 with OH¯ as a strong base were investigated by fluorescence spectroscopy. As shown in Fig. 10A, the addition of OH¯ (1 equiv.) and CN¯ (2 equiv.) to 1 gave the identical emission spectrum, confirming the deprotonation process is taken place. In contrast, the fluorescence response of 2 towards OH¯ and CN¯ was different (Fig. 10B), supporting the results of NMR and the irreversibility of 2 by HCl.
2.4.3 The evidence for formation of 4 and its response to CN¯
Based on the above results, we figured out the formation of a new product 4 during the titration of 2 with CN¯.
Therefore, a mixture of 2 (1 equiv.) and CN¯ (1 equiv.) in MeOH was stirred at room temperature for 1h which gave quantitative yield of dihydroxyquinoxaline 4. The formation of 4 was well supported by 1H NMR and mass spectroscopy and was identical to those of reported data (Figs. S3 and S10).35,36 As shown in Fig. 10C and 10D, 1H NMR and fluorescence spectrum of 4 and 2 in the presence of CN¯ is identical (Scheme 2B), approving the CN¯ sensing of 2 is occurred via 4.
This significant fluorescence response of 4 to CN¯ made us interested in using it as a new CN¯ sensor. Probe 4 shows two emission peaks at 470 and 520 nm.
Upon addition of CN¯ (0–1.0 equiv.) to a solution of 4, the fluorescence peak at 520 nm was gradually reduced (Fig. 11). At the same time, the peak at 470 nm was gradually increased. This unique feature can be attributed to gradual breaking of intramolecular hydrogen bonds of 4 followed by the deprotonation process. This response is fast and show highly sensitivity to CN¯ (LOD = 6.5 nM) with significant color change from yellow to blue under 365 nm UV light (below inset of Fig. 11) with a binding constant of 5 × 103 M− 1 between 4 and CN¯.
2.4.4 The proposed sensing mechanism
First, to prove the critical role of hydroxyl groups of 1 and 2 on the sensing of CN¯, compound 3, having methoxy groups, was posed to CN¯ with no absorbance response (Fig. S11).
In the second step, we considered why CN¯ recognition by 1 and 2 is taken place in different pathways.
As clear from the experimental data, 1 can recognize CN¯ via its tetradentate site between two catechol groups, as shown in Scheme 2, which keeps CN¯ far from imine groups, prohibiting its nucleophilic attack on imines.
In contrast to 1, the experimental results show that CN¯ recognition by 2 is taken place via intramolecular aldimine condensation cyclization (IACC), as depicted in Scheme 3, leading to formation of 4. The proposed mechanism for 2, however, involves cyanide attack at an aldimine and tautomerization to form a carbanion which attacks a second aldimine and the subsequent tautomerization is followed by elimination of cyanide and formation of 4 (Scheme 3). This intramolecular aldimine condensation cyclization (IACC) process is taken place due to shorter distance between two arms of salophene 2 as compared with those of S7-9.