A mixture of 3 (9.23 g, 13.9 mmol) and Cs₂CO₃ (11.30 g, 34.7 mmol) in MeCN (700 ml) under nitrogen was stirred at reflux for 30 min and then a solution of the tetraethylene glycol ditosylate (7.85 g, 15.6 mmol) in MeCN (40 ml) was added during an hour. The mixture was refluxed for 72 h and allowed to cool to room temperature. After evaporation of the solvent in vacuo, the residue was taken up in CH₂Cl₂ (30 ml × 3) and the resultant solution was washed with 1 mol/L HCl (30 ml) and brine (30 ml × 2). The organic layer was dried over MgSO₄ and evaporated in vacuo. Recrystallization of the residue from CH₂Cl₂/MeOH gave 4 a pale-yellow solid. Yield, 75%. Mp: 230–232 °C. ¹H NMR (400 MHz, CDCl₃) δ 7.17 (s, 4H), 7.08 (d, J = 8.0 Hz, 4H), 6.87 (t, J = 8.0 Hz, 2H), 3.78 (d, J = 4.0 Hz, 8H), 3.54 (s,8H), 3.45 (t, J = 8.0 Hz, 4H), 3.25–3.20 (m, 4H), 3.15 (m, 4H), 1.31 (m, 4H), 0.78 (t, J = 7.5 Hz, 6H). ¹³C NMR (101 MHz, CDCl₃) δ 156.97, 155.14, 136.29, 133.81, 132.17, 129.70, 122.77, 115.15, 72.59, 72.09, 70.50, 70.13, 69.03, 37.89, 22.53, 10.23. HRMS-MALDI calculated for C₄₂H₄₈Br₂O₇ [M + Na] ⁺ 847.1639, found 847.1652.

Under nitrogen, a mixture of 4 (10.17 g, 12.3 mmol) and cuprous cyanide (7.68 g, 86.4 mmol) in 20 ml of 1-methyl-2-pyrrolidinone was stirred at 180°C for 4 h. Then, the reaction mixture was cooled slowly to 100°C, and a solution of 23.23 g (86.4 mmol) of FeCl₃·6H₂O in 5 ml of concentrated hydrochloride and 25 ml of water was added to the reaction mixture. The reaction mixture was further stirred at 100°C for 1 h and cooled to room temperature. The solid was filtered off and recrystallized from chloroform/hexane yielding 5.5 g of compound 5 as yellow solid. Yield 62%. ¹H NMR (400 MHz, CDCl₃) δ 7.36 (s, 4H), 7.10 (d, J = 7.4 Hz, 4H), 6.90 (s, 2H), 3.82 (d, J = 5.5 Hz, 8H), 3.55 (s, 8H), 3.46 (t, J = 7.4 Hz, 4H), 3.30 (t, J = 6.1 Hz, 4H), 3.17 (t, J = 6.1 Hz, 4H), 1.24 (m, 4H), 0.74 (t, J = 7.5 Hz, 6H).¹³C NMR (101 MHz, CDCl₃) δ 159.98, 156.80, 135.72, 133.54, 133.05, 129.94, 123.05, 119.17, 106.08, 77.36, 72.52, 72.17, 70.56, 69.85, 69.13, 37.75, 22.81, 10.05; HRMS: calcd for C₄₄H₄₈N₂O₇ [M+NH₄]⁺ 734.3800, found 734.3796.

A solution of 5.18 g (9.2 mmol) of KOH in 100 ml of water was added to the suspension of 1.32 g (1.80 mmol) of 5 in 20 ml of ethanol. The reaction mixture was heated under reflux for 24 h. After cooling, the aqueous solution hydrogen chloride (10% w/w) was added dropwise until the solution became slightly acidic. The precipitate was filtered off, washed with water, and dried to yield a yellow solid product 6 (1.33 g, 96%). Mp: 296–298°C. ¹H NMR (400 MHz, CDCl₃) δ 12.52 (s, 2H), 7.81–7.77 (m, 8H), 6.98–6.94 (m, 2H), 3.93–3.80 (m, 12H), 3.62 (s, 16H), 1.42–1.31 (m, 4H), 0.69 (t, J = 7.5 Hz, 6H).¹³C NMR (101 MHz, DMSO-d ₆ ) δ 167.03, 158.48, 156.72, 135.62, 133.71, 132.13, 130.41, 125.24, 122.05, 71.17, 70.07, 69.72, 40.15, 39.94, 39.73, 39.52, 39.31, 39.10, 38.89, 35.98, 21.74, 9.52. HRMS: calcd for C₄₄H₅₀O₁₁ [M-H]⁺ 753.3280, found 753.3285.

To a round-bottomed flask was added 6 (80 mg, 0.1 mmol), 10 ml of toluene, and 1 ml of thionyl chloride, and the mixture was refluxed for 5 h. After cooling, the solvent and the excess of thionyl chloride were removed at reduced pressure to give the benzoyl chloride, which was added to a solution of 4-(dimethylamino)benzohydrazide (39 mg, 0.22 mmol) in 10 ml of dichloromethane and 0.1 ml of pyridine. The reaction mixture was stirred for 12 h at ambient temperature and filtered. The precipitate was washed with ethanol to give the bishydrazide 7 as white solid, which was used to the next step reaction without further purification. The intermediate compound 7 was added to POCl₃ (5 ml), and the resultant solution was refluxed overnight under a nitrogen atmosphere. After the reaction mixture cooled to room temperature, it was poured into ice water and extracted with dichloromethane (3 × 10 ml). The combined organic layer was washed with water and brine, respectively. Then, the solvent was removed under reduced pressure, and the crude solid was purified by silica gel column chromatography using petroleum ether/ethyl acetate (1:1) as eluent affording the product 1 as white solids. Yield, 35%. Mp: 281–283°C. ¹H NMR (400 MHz, CDCl₃) δ 7.91 (d, J = 8.0 Hz, 4H), 7.81 (s, 4H), 7.15 (d, J = 8.0 Hz, 4H), 6.93 (s, 2H), 6.72 (d, J = 8.0 Hz, 4H), 3.93 (m, 8H), 3.59–3.53 (m, 12H), 3.31 (d, J = 5.2 Hz, 4H), 3.26 (d, J = 5.2 Hz, 4H), 3.05 (m, 12H), 1.25–1.19 (m, 4H), 0.60 (t, J = 7.4 Hz, 6H). ¹³C NMR (101 MHz, CDCl₃) δ 164.80, 163.55, 158.76, 157.10, 152.20, 135.07, 133.83, 129.79, 128.32, 128.13, 122.75, 121.14, 118.35, 111.59, 111.27, 72.43, 72.24, 70.42, 70.14, 69.22, 40.11, 38.05, 22.43, 9.95. HRMS-ESI calculated for C₆₂H₆₉N₆O₉ [M + H] ⁺ 1,041.5120, found 1,041.5126. (Supplementary Figure S5, ESI).

UV-vis spectra were recorded on a Cary 3,010 spectrophotometer, and the resolution was set at 1 nm. Steady-state emission spectra were recorded on a Varian Cary Eclipse spectrometer. For all measurements of fluorescence spectra, excitation was set at 334 nm for complexation, and the excitation and emission slit width was set to be 2.5 nm. Fluorescence titration experiments were performed with CH₂Cl₂ solutions of compound 1 and varying concentrations of metal perchlorate in CH₃CN solution. During all measurements, the temperature of the quartz sample cell and chamber was kept at 25°C.

As shown in Scheme 1, calix[4]arene 3 was reacted with tetraethylene glycol ditosylate in the presence of Cs₂CO₃ to successfully afford the calix[4]crown 3 in 75% yield. The substitution reaction of 4 with CuCN gave 5 in 62% yield, which was refluxed with KOH in ethanol and treated with hydrochloric acid solution, readily providing the carboxylic acid 6 in good yield. Then, the carboxylic acid 6 was reacted with thionyl chloride, and treated with benzyol hydrazine or 4-N,N′-dimethylaminobenzyol hydrazine to generate the intermediate bishydrazide 7, which was used in the next step without purification and refluxed with phosphorus oxychloride to afford the target products 1. Except for the calix[4]arene 3, all of the intermediate calix[4]crowns 3–6 and the chemosensor 1 are in 1,3-alternate conformation, which were well established by ¹H NMR and ¹³C NMR data (Supplementary Figures S1–S4, ESI). The 1,3-alternate conformation of 5 was further confirmed unambiguously by x-ray single crystal diffraction as shown in Figure 1. The x-ray crystallographic data are collected in Supplementary Table S1.

The selectivity of the receptor 1 toward different perchlorate salts, including Na⁺, K⁺, Mg²⁺, Ca²⁺, Co²⁺, Ni²⁺, Zn²⁺, Cd²⁺, Mn²⁺, Ag⁺, and Cu²⁺, was first investigated by UV-Vis spectroscopy. The UV-Vis absorption spectra for free 1 in CH₂Cl₂ solution showed an intense and structureless absorption band (ε = 4.94 × 10⁵ L/mol·cm) peaking at 340 nm (Figure 2), which might have resulted from the spin-allowed π-π* transitions involving the phenyloxadiazole moiety (Han et al., 2006). The addition of Cu²⁺ ions in the solution of 1 resulted in a significant decrease in the absorbance with an appreciable hypochromic shift of 20 nm. In contrast, only a slight decrease was observed upon addition of other metal ions mentioned above, which suggested that the selectivity of 1 toward Cu²⁺ is much higher than the other metal ions.

Ion recognition ability of 1 was further studied by the fluorescence spectra. As shown in Figure 3, the receptor 1 exhibited a strong emission with λ_max at 405 nm in solution of CH₂Cl₂. Upon addition of Na⁺, K⁺, and Mg²⁺, respectively, almost no changes were observed in the intensity and shape of the emission spectra of 1. It is noted that the addition of Ca²⁺ might slightly increase the intensity with a bathochromic shift of ca. 15 nm, perhaps because the complexation between the Ca²⁺ and the crown ether moiety changed the space distance of the two phenyloxadiazole units and the fluorescence changed consequently. Apparently, the fluorescence response of 1 toward transition metal ions was found to be more pronounced, and the addition of Co²⁺, Ni²⁺, Zn²⁺, and Ag⁺ could quench the emission of 1 in a different extent, accompanied by a concomitant red shift of ca. 14–17 nm. In contrast, the addition of Cu²⁺ significantly quenched the fluorescence of 1 under the same conditions as the aforementioned metal ions, suggesting that there is a strong interaction between 1,3,4-oxadiazole moieties of 1 and Cu²⁺ ion over the other metal ions.

The fluorescence emission properties of 1 in the presence of Cu²⁺ and a competitive metal ion were measured to investigate the selective recognition for Cu²⁺. As shown in Figure 4, no apparent changes were observed in fluorescence intensity when 10 equivalent amounts of transition metal ions (Co²⁺, Ni²⁺, Zn²⁺, Cd²⁺, Mn²⁺, and Ag⁺) were added to the solution of 1 and Cu²⁺ (10 equiv). This suggested that the recognition for Cu²⁺ was not interrupted by the competitive transition metal ions; thus, the receptor 1 might act as a selective fluorescent chemosensor for Cu²⁺. The addition of alkali metal ions (Na⁺ and K⁺) to the solution of 1 and Cu²⁺ could increase the fluorescence intensity slightly, while the alkaline earth metal ions (Mg²⁺ and Ca²⁺) could revive the emission significantly.

In order to elicit the binding property of the chemosensor 1 toward Cu²⁺ ion, fluorescence titration of 1 (1.0 × 10^–5 mol/L) with Cu²⁺ ion (0–2 equiv) was carried out (Supplementary Figure S6). According to the fluorescence titration curves of 1 with Cu²⁺ ion at room temperatures, the association constant K _a was calculated as 1.6 × 10^–4 L·mol⁻¹ (R = 0.97526) for the 1–Cu²⁺ complex by the Benesi–Hildebrand plot (Thordarson, 2011) (Figure 5). Moreover, the emission intensity of 1 is linearly proportional to the Cu²⁺ concentration in the range of 0–20 μM.

The fluorescence changes of 1 upon addition of Cu²⁺ and Mg²⁺ ions are displayed in Figure 6. The nitrogen atoms of the 1,3,4-oxadiazle units can bind with Cu²⁺ to form the complex 1·Cu²⁺, and the paramagnetic nature of Cu²⁺ ion could strongly quench the fluorescence of the 1,3,4-oxadiazole units through the electron transfer mechanism, which is consistent to the results reported in literature (Han et al., 2012). In contrast, the polyether ring (crown-5 moiety) and the oxygens from the two propoxyl groups could provide coordination sites with the alkaline earth metal ions to form the complex 1·Mg²⁺, which will change the molecular conformation as well as the space distance of the two 1,3,4-oxadiazole units. Consequently, the decomplexations between the 1,3,4-oxadiazoles and Cu²⁺ ions took place and resulted in the increase of the fluorescence. Thus, the receptor 1 might be acted as an on–off–on switchable fluorescent chemosensor triggered by the exchange of Cu²⁺ and Mg²⁺.

To gain a better understanding about the switchable fluorescence of the chemosensor 1, DFT calculations with the GAUSSIAN 09 series of programs (Frisch et al., 2013) were carried out to analyze the molecular structures of 1 and 1·Mg ²⁺ , and DFT method B3-LYP with 6-31G(d) basis set was used for geometry optimizations (A. D. Becke, 1993). As shown in Figure 7, the distance between N1 and N2 in the free receptor 1 is 9.97 Ǻ, while the corresponding distance is 11.74 Ǻ in the complex 1·Mg ²⁺ , indicating that the molecular conformation changed simultaneously due to the allosteric effect (Kumar et al., 2012; Ni et al., 2013). The conformational change as well as the increase in distance makes it difficult for the chemosensor 1 to coordinate with Cu²⁺ ion to form the stable complex, which reasonably explains the fact that the addition of Mg²⁺ ions to the solution of 1 and Cu²⁺ can trigger the revival of fluorescence.