A seawater triggered dynamic coordinate bond and its application for underwater self-healing and reclaiming of lipophilic polymer

Bulk polymer capable of repeatedly underwater self-healing and reclaiming is synthesized under the inspiration of the formation of a mussel byssus cuticle.


Characterization
Fourier transform infrared spectra were recorded using a Bruker EQUINOX55 FTIR spectrophotometer between 400 and 4000 cm -1 with a resolution of 4 cm -1 by the KBr sample holder method. Attenuated total reflectance (ATR) spectra were collected with a Bruker TENSOR 27 spectrophotometer between 1000 and 2000 cm −1 with a resolution of 4 cm -1 .

S2
Proton nuclear magnetic resonance spectrum ( 1 H NMR) was obtained by using DMF-d 7 as the solvent on a Bruker AVANCE III spectrometer (400 MHz). The spectra were internally referenced to tetramethylsilane (TMS) standard.
Water contact angle was measured with a Krüss DSA 100 apparatus at ambient temperature. All the measurements were taken after a 3 µL water droplet was placed on the sample for about 1 min.
Evaluation of self-healing performance in seawater is described as follows. The synthesized polymer was filled in a silicone mold, and dried under vacuum at 40 C for 48 h for the subsequent characterization. In the case of visual inspection, the sample (2 mm thick) was cut by a razor in artificial seawater at 25 C to produce a scratch 20~50 µm wide, and then, the damaged film was allowed to stay in the water for healing. Photos recording change of the wound were taken by a camera attached to an optical microscope. As for quantitative assessment, underwater tensile test was performed on dumbbell-shaped specimen cut from the polymer film (2 mm thick) according to ISO 527-3 at a crosshead speed of 500 mm min -1 with a LWK-5 universal tester. Firstly, the specimen was bisected in artificial seawater, and then the Rheological data were obtained from a strain-controlled ARG2 rheometer with 25 mm parallel-plate geometry (disk-shaped specimens: 10 mm in diameter and 2 mm in thickness). Frequency sweeps at 0.2 % strain were conducted at different temperatures.
Stress relaxation was measured at 25 °C in tensile mode at a constant strain of 10 % with 01dB-MetraviB DMA-25N.
Cyclic tensile tests were conducted at 25 °C with a LWK-5 universal tester.
Crosshead speed of loading is 100 mm min -1 , and that of unloading is 10 mm min -1 .
The Mössbauer measurements were performed with a Topologic 500A spectrometer and a proportional counter at room temperature. 57 Co(Rh) moving in a constant acceleration mode was used as the radioactive source. The Doppler velocity of the spectrometer was calibrated with α-Fe foils. The spectra were fitted with Lorentzian peaks using Moss Winn 3.0i, and the free recoil fraction was assumed to be the same for all iron species. Model compounds Fe[DOPA] n were prepared and characterized as follows to simulate the polymer HBPU-DMPA-[Fe(DOPA) 3 ] of this work. Dopamine hydrochloride (DOPA, 1 g) was dissolved in 10 ml distilled water, and then anhydrous ferric chloride was added at mole ratios of nDOPA : nFeCl 3 = 3:1,  As the absorbance at 280 nm on UV-Vis spectrum of aqueous solution of dopamine originates from catechol group, the peak height increases with a rise in S8 concentration of dopamine. Accordingly, a linear regression was made to correlate the dependence of the absorbance at 280 nm of aqueous solution of dopamine on concentration of dopamine ( Figure S4a). By using the relationship shown in Figure   S6a as a calibration curve, the content of dopamine in the polymer can be determined as follows. After the incorporation of dopamine into the reaction system, a certain amount of sample was taken out at set interval. Then, the sample was put into water for 24 h at room temperature, allowing completely dissolving out of the unreacted dopamine. Afterwards, the concentration of the unreacted dopamine in aqueous solution was quantified by UV-Vis spectroscopy with the aid of the plot in Figure   S4a, and plotted as a function of reaction time ( Figure S4b). When the amount of unreacted dopamine approached to a steady value, it means that the reaction is almost completed, so that the data at the reaction time of 4.5 h in Figure S4b can be used to estimate the ultimate content of dopamine in the polymer via the equation:  The data in Figure S8 and Table 1 show that under the condition of pH = 9 the  In summary, two states of iron are always observed in the anhydrous samples, while single state of iron is observed in water saturated samples. This proves that the DOPA-iron complexation becomes dynamic in the presence of water and the dynamic manner is immobilized after removing water. Because of the highly symmetric structure and resonance of bis-coordination, the samples at pH = 7 is an exception.
Even the dynamic coordination is fixed in the anhydrous sample, the dynamic feature S13 is still displayed so that similar state of iron is found in both water saturated and anhydrous samples.

Figure S9a
shows that G' is higher than G" in most cases and both G' and G" decrease with decreasing frequency. Nevertheless, the two curves intersect at certain frequency under 80 and 100 C, meaning that the rheological behavior of the specimen is changed from elastic-like behavior (G' > G") to viscous-like behavior (G' < G") at low frequency regime. This transition is indicative of dynamic coordination-dissociation of catechol-Fe 3+ crosslinks. At certain low frequencies (for a S14 given temperature), the dynamic bonds start to exert significant influence on rheological performance of the polymer because the disconnected network needs time to be reconstructed. Besides, the crossover frequency between G' and G" increases with increasing temperature. Because this frequency is related to the dissociation rate constant of the reversible network (refer to M. C. Roberts, M. C. Hanson, A. P.
This is in agreement with the general behavior of reversible reactions.
In   Figure S13 shown below), especially at 24 h (i.e. the time of self-healing), and their polarities are also similar (refer to Figure S10), the observed crack disappearance should not be due to differences in polymer swelling between the two samples. are attributed to the symmetrical stretching and antisymmetric stretching of carboxyl groups, are obviously higher on the spectrum of the cut surface than those on the spectrum of the molded surface. It means that more carboxyl groups are exposed on the cut surface. The difference might originate from the fact that when the specimen was made in silicone mold by casting, the hydrophilic carboxyl groups tend to hide beneath the specimen surface during the solvent evaporation. Accordingly, there are less carboxyl groups on the molded surface of the specimen.

Wavenumber [nm]
As-synthesized Saturated with artificial seawater for two months Figure S16. UV-vis spectrum of HBPU-DMPA-[Fe(DOPA) 3 ] that was kept being saturated with artificial seawater and exposed to air for two months at 25 C in comparison with that of the as-synthesized version.
The absorbance maximum at ~280 nm is the characteristic absorption peak of dopamine, while the peak at 400∼500 nm represents the formation of Fe  (7): 2651). The characteristic absorption peak of dopamine quinone at ∼400 nm does not appear on the two spectra, suggesting that dopamine was not oxidized during either synthesis or service in seawater.

S22
*The negative signs are only indicative of movement direction of the probes.
Although HBPU-DMPA-DOPA also exhibits similar behavior in the course of AFM pull-off force measurement due to hydrogen bonding, its peak pull-off force that represents the maximum interaction between the same material is much lower than that of HBPU-DMPA-[Fe(DOPA) 3 ] (refer to Figure 3d and Table S1). Additionally, the peak pull-off forces of HBPU-DMPA-[Fe(DOPA) 3 ] decrease with decreasing pH, while those of HBPU-DMPA-DOPA nearly do not change with pH (refer to Table   S1). The results well agree with the characteristics of catechol-Fe 3+ complexation and hydrogen bonding, respectively.