2.1 Formation of complex CB@CDs
CDs were obtained by condensation ethylenediamine and citric under high temperature and pressure. 1H NMR spectra of CDs is shown in Fig S1. 7.25-7.45 ppm was the proton peak corresponding to the benzene ring of CDs (1H). 3.95 ppm was the proton peak (2H) of the methylene group in -COOCH2-,3.65 ppm was the proton peak of methylene group in -CH2-OH (3H), 3.36 ppm was the proton peak in -CONH-CH2- (4H), 2.67-2.87 ppm was the proton peak in -N-CH2× and O=C-CH2-C (5H). 2.01 ppm was the proton peak of the methyl group in -C=C-CH3 (6H). 1.20 ppm was the proton peak of the methyl group in -C-CH3 (7H). The main peak of CDs was consistent with the literature 32, so it was successfully prepared
CB[6]@CDs complex was prepared by the interaction between CB[6] and CDs through the host-guest. The binding behavior between CB[6] and CDs was represented by 1H-NMR spectrum, as shown in Fig S2. After adding CB[6], the 1H proton signal (7.25-7.45 ppm) moved to the high field by 1.6-1.7 ppm. The 7H proton signal (1.20 ppm) moved 0.45 ppm to the high field. These data indicated that some groups of CDs were inserted into the CB[6] cavity. In addition, 2H and 5H proton peaks were affected by the de-shielding effect of CB[6] port and moved to low field, and the corresponding signals 3.95ppm (2H) and 2.67-2.87 (5H) were changed. As the proportion of CB[6] increases, the integrated area of 7H gradually increases, which also indicated that the CB[6]@CDs complex increased.
2.2 Properties of hydrogel
CB[6]-MA, MH, AM and PVA monomers were subjected to free radical polymerization in water and the cavity of CB[6] was used to combine with CDs to form hydrogel in the mold. Infrared spectra were performed to characterize various valence bonds in the hydrogel, as shown in Fig. S3. It was obvious carbonyl group of CB[6] characteristic peak at 1750 cm-1 for the hydrogel without CDs (Fig. S3b), and N-H deformation vibration peak of acrylamide at 1640 cm-1. The absorbance at 1610cm-1 was derived from the N-N group of methylacrylamide hydrazide. By comparing curve b and curve c of Fig. S3, it is evident that the absorption peak of curve c at 1674 cm-1 is significantly wider, because CDs contain a large number of N-H bonds. Thus, the hydrogel with blue fluorescence was successfully prepared by the interaction between CDs and CB [6] units.
Dynamic covalent bonds, multiple responsive, are particularly sensitive to external stimuli, which constructs smart polymer material, pH-responsive dynamic covalent bonds of acylhydrazone bonds and borate bonds were introduced in the system, resulting in the transformation of the gel from the initial shape to the temporary shape. The shape memory function of hydrogel was realized by three different interaction forces: formation and destruction of dynamic covalent borate bonds and acylhydrazone bonds and reversible crystallization of polyvinyl alcohol chains. The hydrogels were prepared to size of 50 mm×10 mm ×1 mm and then bent the specimens and immerse it in borax solution (0.2 M) or glyoxal solution (0.2 M) to fix the temporary shape, respectively. The temporary shaped hydrogel was immersed in acetic acid to detect the stage of shape recovery. Temporary shapes were also fixed by freezing in the refrigerator for 2 h. However, they took a long time to thaw at room temperature, so thawing in deionized water at 60 oC to detect the stage of shape recovery, as shown in Fig 1. The temporary shapes of the hydrogels were fixed by the interactions. Eliminating these interactions would restore the shape of the hydrogel to their original state.
polyvinyl alcohol chains in hydrogel contained a large number of hydroxyl groups, which could react with borax solution to form dynamic covalent borate bonds. Therefore, the shapes of hydrogels in borax solution were fixed as "V" shape and were restored to their initial state in acetic acid (Fig 2). The reason was that dynamic covalent borate bonds breakage in acidic environment restored gels shape. To further investigate the effect of borate bonds on gels shape, we immersed gels in borax solution for different times and found that soaking time did not affect gels recovery to initial shape. In order to prove the change of valence bonds of hydrogels before and after deformation, infrared test was carried out (Fig. S4). Compared with the initial hydrogel, the peak of hydrogel@borate at 1670 cm-1 was significantly widened and the characteristic peak of B-O at 1390 cm-1 was obvious, which indicated the formation of dynamic covalent borate bonds (Fig. S4c). After the gel was immersed in acetic acid (hydrogel@acetic acid), the peak at 1670 cm-1 narrowed and the peak at 1390 cm-1 disappeared, which was consistent with the characteristic peak of the initial gel (Fig. S4d). It had been shown that the dynamic covalent borate bonds were broken. The chemical structure of hydrogel@acetic acid was consistent with the initial gel and the shape was restored.
As shown in Fig S5, linear hydrogels were immersed in glyoxal solution to fix their shape as "V". There were a large number of hydrazide groups in hydrogels, which could react with glyoxal to form dynamic covalent acylhydrazone bonds, thus fixing the shape of hydrogels. The restoration of the initial shape of fixed shape hydrogels in acetic acid was due to the fracture of the dynamic covalent acylhydrazone bonds in acidic environment. To further investigate the effect of acylhydrazone bonds on the shape of gels, these were immersed in glyoxal for different time. These were found that the time required for the hydrogel to recover to its initial shape increased when the soaking time was 4 minutes or more. This was because the number of reactions of hydrazine groups increased and the gel cross-linking density increased with the longer soaking time. As a result, it took more time to break the acylhydrazone bonds. The change of shape of hydrogels in glyoxal and acetic acid were essentially the formation and elimination of chemical bonds in hydrogels. Compared with the initial hydrogel, hydrogel@glyoxal showed a significant increase in peak intensity at 1661 cm-1 due to condensation of the aldehyde group with the hydrazide group to form C=N bonds, indicating dynamic covalent acylhydrazone bonds formation (Fig. S6b). After the gel was immersed in glacial acid (hydrogel@ glacial acetic), the peak intensity at 1661 cm-1 significantly weakened, which was consistent with the characteristic peak of the initial gel (Fig S6c). It was previously shown that the dynamic covalent acylhydrazone bonds were broken. The chemical structure of hydrogel@glacial acetic was consistent with the initial gel and the shape was restored.
The prepared hydrogel had dynamic covalent acylhydrazone bonds, borate bonds and interpenetrating network of polyvinyl alcohol chain, which made the gel mechanically superior. The hydrogels were immersed in borax, glyoxal solution, and then in acetic acid solution to judge the mechanical properties of the gel before and after deformation. As shown in Fig. S7, when hydrogels were immersed in borax solution for 0,2,3,4 and 5 minutes, temporary shape formation and soaking time did not affect the maximum tension or reduce the tensile strength. According to Fig 2, hydrogels could recover their shape within 2 minutes of immersion in acetic acid. Based on this, the above hydrogels were immersed in acetic acid solution for 2 minutes before detection. The tensile strength of the hydrogel after shape transformation was from 0.439 MPa to 0.397 MPa. The difference was 0.042 MPa (9%), which was tiny and negligible (Fig. 7Sb). No obvious adverse effect on the mechanical properties of the gel before and after deformation. In addition, hydrogels were immersed in glyoxal solution for 0,1,2,3, and 4 minutes. According to Fig. S5, these hydrogels recovered their shape within 4 minutes of immersion in acetic acid. On this basis, the above hydrogels were immersed in acetic acid solution for 2-4 minutes before detection. The variation in tensile strength of hydrogels was negligible (30%). It indicated that the mechanical properties of gels were not affected absolutely before and after deformation (Fig. S7d). The gels exhibited remarkable mechanical stability.
2.3 Recognition of metal ions by hydrogel
1H-NMR indicated that -CB[6] units were bound to CDs through host-guest interaction. In order to further clarify the role of -CB[6] group in gels, CB[6] functionalized hydrogels and hydrogels without CB[6] were prepared. The essential difference between these two hydrogels was that the binding modes of CDs in hydrogels were respectively supramolecular self-assembly binding of -CB[6] and ordinary physical adsorption. Two gels of the same quality were respectively immersed in a beaker containing 50ml deionized water. After 48 hours, the fluorescence of deionized water in the beaker were detected, as shown in Fig S8. Soaking in -CB[6] units functionalized hydrogel, the fluorescence intensity of deionized water was much lower than that of deionized water soaked in hydrogel without -CB[6]. This indicated that the self-assembly binding mode of -CB[6] and PCDs was far more compact and chemically stable than that of physical doping. hydrogel with -CB[6] could produce blue light under 365nm UV lamp irradiation. Encounter iron salt or mercury salt (0.2 mg/mL) fluorescence quenching, blue light disappeared (Fig 3). The hydrogel appeared pale yellow in natural light (Fig 3a); showed brown maple in the presence of Fe3+ ions (Fig 3b) and showed bright white in the presence of Hg2+ ions (Fig 3c), indicating the adsorption of iron and mercury ions on the hydrogel.
To further explore the effect of other metal ions (K+, Na+, Ca2+, Zn2+, Mg2+, Cu2+, Al3+, Pb2+, Ni2+, Co2+, Cd2+, Ag+) on the fluorescence characteristics of the hydrogel, the fluorescence emission spectrum of the hydrogel was tested (Fig 4c). It was found that most metal ions had little interference on the hydrogel and did not quench the fluorescence like iron and mercury ions. It was demonstrated that the hydrogel could recognize Fe3+ ions and Hg2+ ions specifically. The optimal experimental conditions were used to determine whether Fe3+ ions and Hg2+ ions could effectively quench the fluorescence of the hydrogel, and the linear response range and limit of detection (LOD) were detected. According to the Stern-Volmer equation: I0/I = 1+KSV[Q]. Where I and I0 were respectively the fluorescence intensity of hydrogel with or without Fe3+ ions. Q was the Fe3+ ion concentration, and KSV is the constant of Stern-Volmer equation33. The fluorescence intensity at 475nm decreased with the increase of Fe3+ ions concentration (0-1.25mM). There was a good linear relationship between fluorescence intensity and Fe3+ ions concentration (R2=0.957). The LOD of Fe3+ was calculated and verified to be 0.94 μm (Fig 4a). The fluorescence intensity at 475nm decreased with the increase of Hg2+ ions concentration (0-1.5mM). There was a good linear relationship between fluorescence intensity and Hg2+ ions concentration (R2=0.987). The LOD of Hg2+ ions was calculated and verified to be 0.19 μm (Fig 4b). Compared with the national standard detection, the hydrogel had strong fluorescence activity and could well identify Fe3+ ions and Hg2+ ions. The UV-vis absorption spectrum of Hg2+ or Fe3+ ions has overlap with excitation spectrum of hydrogel, as shown in Fig. 4d. The results indicate that the energy is competed with Hg2+ or Fe3+ ions, and the luminescence of hydrogel center is quenched. These results indicated that hydrogel containing cucurbit[6]uril-carbon dots had potential application value in the detection of metal ions.