Physical hydrogels composed of polyampholytes demonstrate high toughness and viscoelasticity

Transcription

1 Physical hydrogels composed of polyampholytes demonstrate high toughness and viscoelasticity Tao Lin Sun 1, Takayuki Kurokawa 2, Shinya Kuroda 1, Abu Bin Ihsan 3, Taigo Akasaki 3, Koshiro Sato 3, Md. Anamul Haque 2, Tasuku Nakajima 2, Jian Ping Gong 2 1 Graduate School of Science, Hokkaido University, Sapporo , Japan 2 Faculty of Advanced Life Science, Hokkaido University, Sapporo , Japan 3 Graduate School of Life Science, Hokkaido University, Sapporo , Japan These authors contributed equally to this work. * gong@mail.sci.hokudai.ac.jp NATURE MATERIALS 1

2 Methods Synthesis of polyampholyte hydrogels Polyampholyte hydrogels are synthesized using the one-step random copolymerization of an anionic monomer and a cationic monomer. All of the chemicals were used as purchased without further purification. A mixed aqueous solution with the prescribed total ionic monomer concentration Cm (M) and molar fraction f of the anionic monomer, 0.25 mol% UV initiator, 2-oxoglutaric acid (in relative to the total monomer molar concentration), and 0.5 M NaCl was poured into in a reaction cell consisting of a pair of glass plates with a 3 mm spacing and irradiated with 365 nm UV light for 11 hours. After polymerization, the as-prepared gel was immersed in a large amount of water for 1 week to reach equilibrium and to wash away the residual chemicals. During this process, the mobile counter-ions of the ionic copolymer are removed from the gel, and the oppositely charged ions on the copolymer form stable ionic complexes either through intra- or interchain interactions. Fig. 1b and Supplementary Fig. 6 show the chemical structures of the monomers used in this work. These cationic and anionic monomers are essentially randomly copolymerized in the polyampholyte hydrogels. The random structure of P(NaSS-co-MPTC) was confirmed using a 1 H-NMR reaction kinetics study. DN and PAAm hydrogels were synthesized using the method described in reference 8. Characterization of gels Swelling measurements. The as-prepared polyampholyte gels formed in glass plates with rectangle shapes were cut into samples with fixed sizes and then immersed in water and allowed to reach the equilibrium state. The swelling volume ratio Qv was defined as the ratio of the sample volume at swelling equilibrium V to that in the as-prepared state V0, Qv = V/V0. The polymer weight fraction Cpoly (wt%) of the sample was measured by the weight change upon drying using a freeze-drying process. The swelling of samples in NaCl solution was characterized by the volume ratio of the equilibrium swollen gel sample in NaCl solution Vsalt to that in water Vwater, Qsalt,water = Vsalt/Vwater. To achieve adequate precision, three measurements were carried out on samples of different volumes taken from the same gel. 2 NATURE MATERIALS

3 Tensile and compressive test. The tensile stress-strain measurements were performed using a tensile-compressive tester (Tensilon RTC-1310A, Orientec Co.) at a deformation rate of 100 mm/min in air. The test was carried out on dumbbell-shaped samples with the standard JIS-K size (12 mm (L) 2 mm (d) 2 3 mm (w)) (Supplementary Fig. 12). For the cyclic tensile test, all of the experiments were carried out in a water bath to prevent water from evaporating from the samples. The work of extension at fracture Wb (J/m 3 ), a parameter that characterizes the work required to fracture the sample per unit volume, was calculated from the area below the tensile stress-strain curve until fracture. In the compression test, samples with cylinder shapes (~ 15 mm in diameter and mm in initial thickness) were placed on a metal plate coated with silicon oil to decrease the friction. The loading velocity was 0.5 mm/min. Tearing test. The tearing test was performed to characterize the toughness in air using a commercial test machine (Tensilon RTC-1310A, Orientec Co.). Samples of 2 3 mm (w) in thickness were cut into the standard JIS-K6252 1/2 sizes (50 mm (L) 7.5 mm (d); the length of the initial notch is 20 mm) with a gel-cutting machine (Dumb Bell Co., Ltd.) 36. The two arms of a test piece were clamped, and then the upper arm was pulled upward at constant velocity 100 mm/min while the tearing force F was recorded. The tearing energy T was calculated at a constant tearing force F using the relation T = 2F/w, where w is the thickness of the sample (Supplementary Fig. 13). Pure shear test. A pure shear test was also used to characterize the toughness, following the method established in references 18 and 37. Two different samples, notched and unnotched, were used to measure the tearing energy T. The samples were cut into a rectangular shape with a width of 20 mm and length 40 mm (a0). The sample thickness was 0.67 mm (b0). An initial notch of 20 mm in length was cut using a razor blade. The test piece was clamped on two sides, and the distance between the two clamps was fixed at 8 mm (L0). The upper clamp was pulled upward at constant velocity of 100 mm/min, while the lower clamp was fixed. The force-length curves of the samples were recorded, and the tearing energy was calculated from T = U(Lc)/(a0 b0), where U(Lc) is the work done by the applied force to the unnotched sample at the critical stretching distance Lc, and Lc is the distance NATURE MATERIALS 3

4 between the two clamps when the crack starts to propagate in the notched sample. The onset of the crack propagation was determined using the movie image recorded by a camera (Supplementary Fig. 14). We have confirmed that the tearing test and the pure shear test gave the consistent values for the same kind of samples. Impact test. The shock absorbance ratio was characterized using an impact tester (No.511-D, MYSS Tester Co.). First, plate-shaped samples with thicknesses of ~3 mm were fixed in the impact tester. Then, the hammer impacted the sample with a velocity that was determined by the impacting angle. The impacting velocity (v0) and rebounding velocity (vt) of hammer were calculated from the impact displacement and time. The shock absorbance ratio R was estimated using the relation R = 1 (vt/v0) 2. The initial impact velocity was fixed at m/s. Rheological test. Rheological tests were performed using an ARES rheometer (advanced rheometric expansion system, Rheometric Scientific Inc.). A rheological frequency sweep from 0.01 to Hz was performed with a shear strain of 0.5% in the parallel-plates geometry in a temperature range of C. The disc-shaped samples with thicknesses of ~ 3 mm and diameters of 15 mm were adhered to the plates with glue and surrounded by water. Crack tip observation. In order to observe the stress concentration during the crack growth, the crack microstructure was frozen using acetone to avoid any stress relaxation, and then the sample was observed using polarized optical microscopy (Olympus, BH-2). 4 NATURE MATERIALS

5 1) 1 H-NMR reaction kinetics study a NaSS MPTC H 2 O DMSO 5h30min 4h40min 3h40min 2h30min 0min c a d eb f ppm NATURE MATERIALS 5

6 b c [MPTC] Reaction 3 r1=5.23 r2=0.188 Reaction 2 r1=2.43 r2= Reaction 1 r1=1.45 r2=0.713 [MPTC] r1=1.48 r2= [NaSS]/[MPTC] [NaSS]/[MPTC] d 10 <N> NaSS and <N> MPTC <N> MPTC <N> NaSS Total monomer conversion, p Supplementary Figure 1 1 H-NMR reaction kinetic study for the copolymerization of NaSS and MPTC. a, Spectra evolution of reaction mixture in 1 H-NMR probe. b, Molar concentration of MPTC, [MPTC], vs the ratio [NaSS]/[MPTC] from the 1 H-NMR analysis of the 3 reaction systems with different NaSS molar fraction 0.7 (reaction 1), 0.52 (reaction 2) and 0.3 (reaction 3). c, Corrected global molar concentration of MPTC, [MPTC], vs the ratio [NaSS]/[MPTC] using the treatment proposed in reference 38. d, The instantaneous number-average sequence length of monomer NaSS <N> NaSS and MPTC <N> MPTC versus the total monomer conversion p for 0.52:0.48 composition. The dashed line is the predicted curve of sequence length change with monomer conversion. The random structure of polyampholyte hydrogel P(NaSS-co-MPTC) was confirmed using a 1 H-NMR reaction kinetics study. In order to determine the sequence length distribution of the copolymer P(NaSS-co-MPTC), the reactivity ratios r1 and r2 are needed to calculate from the 6 NATURE MATERIALS

7 monomer conversion against reaction time, where the r1 and r2 are defined as the ratio of rate constants of home-propagation reactions to cross-propagation reactions for NaSS and MPTC, respectively. The detailed method used here is described in the reference 38 and 39. The copolymerization of anionic monomer NaSS and cationic monomer MPTC was carried out in D2O and the weighted dimethyl sulfoxide (DMSO) was used as external standard substitute. Three kinds of mixed aqueous solution with total monomer concentration 0.1 mol/l, the NaSS molar fraction (0.3, 0.52 and 0.7), 0.25 mol% UV initiator, 2-oxoglutaric acid (in a concentration relative to the total monomer concentration), and 0.5 mol/l NaCl were performed in the glass tubes under the irradiation of UV light. To study the monomer conversion, we taked the samples from the reaction system and transfered to a shaded place to quench the polymer reaction each several minutes. The concentration of unreacted monomers remaining in solution was determined by the integral area ratio of 1 H-NMR signals which were detected by a 400MHz NMR system, that is, the vinylic protons of the monomers (6.8, 5.9 and 5.4ppm for NaSS, 5.7 and 5.5 ppm for MPTC) versus DMSO protons (2.8ppm) which corresponds to the symbols c, a, b, d, e and f in the spectra (Supplementary Fig. 1a). Finally, we obtained the monomer conversion spectra against reaction time. From the conversion of the monomer, we determined the reactivity ratios using the integrated form of the copolymerization equation (terminal model) for the three reactions 38, as described in Supplementary Fig. 1b. The reactivity ratios are r1 = 5.23, r2 = for reaction 3, r1 = 2.43, r2 = for reaction 2 and r1 = 1.45, r2 = for reaction 1 by fitting the terminal model, corresponding to the composition of NaSS 0.3, 0.52 and 0.7, respectively. The obtained different reaction ratios are due to the sample preparation and integration which perturb the measurement to result in the large error. To overcome this, we obtained the combined data by multiplying the inverse shift factors [ MPTC] reaction2 Q1 10 at the common ratio [ MPTC] reaction1 [ NaSS] (the final and initial [ MPTC] [ MPTC] reaction3 monomer ratio for these two reactions) with data of reaction 2 and Q2 Q [ MPTC ] reaction2 NATURE MATERIALS 7

8 at the ratio [ NaSS] with data of reaction 3, respectively, as indicated in Supplementary [ MPTC] Fig. 1c. Finally, the obtained reactivity ratio is r1 = 1.48 and r2 = Furthermore, the apparent rate constants of propagation reactions of NaSS and MPTC at the composition of fnass = 0.52 are /min and /min, respectively, according to the relationship between the monomer conversion and reaction time. Then we can predict the required time for full conversion of total monomers. The instantaneous number-average sequence length of NaSS monomer, <N> NaSS, and MPTC monomer, <N> MPTC, can be expressed with the total monomers conversion point p according to the Mayo-Lewis theory 39. The results are shown in Supplementary Fig.1d. We predicted the sequence length with the monomer conversion at high monomer conversion p (> 0.8) from the relationship between the conversion of total monomers and reaction time when the apparent rate constants of propagation reactions of monomers are known. According to these results, we can assume the mechanism of the supramolecular hydrogels network formation. The polymerization begins with the incorporation of NaSS molecules with a few MPTC molecules added due to the difference of reactivity ratios (r1 = 1.48 and r2 = 0.70). At total monomer conversion p = 0, the determination of <N>NaSS = 2.5 and <N>MPTC = 1.7 can be understood that, as an average of the incorporation to the growing polymer chains, a sequence of three NaSS molecules follows by one MPTC molecules. At p = 0.8, then <N>NaSS = 1.3 and <N>MPTC = 5.1, a sequence of one NaSS molecule would follow by 5 MPTC molecules. At higher convention (> 0.9), the polymer chains will grow with a block sequence of MPTC molecules. This polymerization process leads to the inhomogeneities of the network structure. As a result, we assume that during the dialysis process the NaSS rich segments (formed at the beginning of polymerization) and MPTC rich segments (formed at the end of polymerization) would form the strong ionic complex structure, serving as permanent cross-linking points, while other parts lead to the weak ionic complex, behaving as reversible sacrificial bonds. 8 NATURE MATERIALS

9 2) Effect of charge ratio 10 2 Swelling volume ratio Swelling volume ratio, Q v Young's modulus Fracture stress Swelling Deswelling Young's modulus, E Fracture stress, b (MPa) NaSS molar fraction in feed, f Supplementary Figure 2 NaSS molar fraction effects on the swelling volume ratio Qv, Young s modulus E, and the compressive fracture stress σb of the hydrogels P(NaSS-co-MPTC). We denote the samples using the symbol Cm-f-x and the names of the copolymers, where Cm (mol/l) is total molar concentration of monomers, f is the anion molar charge fraction, and x (mol %) is the molar ratio of the chemical cross-linker N, N -methylenebisacrylamide (MBAA) in relative to Cm in the precursor solution. Supplementary Fig. 2 shows the effect of the charge fraction (f) on the swelling volume ratio Qv, Young s modulus E, and the compressive fracture stress σb of the hydrogels P(NaSS-co-MPTC) Cm -f-4 synthesized with different NaSS molar fractions f in the feed, where the total molar concentration Cm is fixed at mol/l with a chemical cross-linker (x = 4 mol %). Here, Qv is defined as Qv = V/V0, where V and V0 are the volumes of the samples after full swelling in pure water and in the as-prepared state, respectively. Near the charge balance point (f = 0.48 ~ 0.53), the gels shrink in water (Qv < 1) relative to their as-prepared state, indicating that the Coulomb attraction prevails over the repulsion and the polymer NATURE MATERIALS 9

10 chains collapse. In the regions with sufficient charge imbalance (f < 0.48 or f > 0.53), on the other hand, the gels swell (Qv > 1), indicating that the Coulomb repulsion prevails and the polymer segments elongate. The shrinking of the gels around the charge balance point is accompanied by dramatic increases in the modulus (E) and fracture stress (σb). Thus, the optimized f value in the feed, at which Qv reaches a minimum and E and σb reach maximums, is 0.52, which is close to the charge balance point. The true charge ratio of P(NaSS-co-MPTC) Cm-f-4 (the total molar concentration is fixed at 0.875mol/L) was studied using elemental analysis (Supplementary Table1), which revealed that ftrue = 0.48 for the sample of f = 0.52 in feed, which is slightly different from the ideal stoichiometric ratio of 1:1. This indicates that the gel with complete charge balance (f = 0.5) is not in the most stable state. The attractive ion pairs cannot achieve close approach, probably because the restriction of the polymer chain conformation frustrates the electrostatic effect 23. Supplementary Table 1 The weight percentage of elements in various polyampholyte hydrogels P(NaSS-co-MPTC) f-4 through element analysis. f wt % (C) wt % (H) wt % (N) wt % (S) wt % (Cl) ftrue * f is the anion charge fraction in feed and ftrue is the true anion charge fraction in gel determined by element analysis. 10 NATURE MATERIALS

11 3) Molecular weight To discuss how the entanglements of polymer chains contribute to the toughness of the hydrogels, we have tried to estimate the molecular weight of the polymer. The P(NaSS-co-MPTC) hydrogels dissolved in 4 M saline solution at high temperature (> 50 C) after 2 days, and we can obtain homogeneous aqueous solution. It is difficult to accurately estimate the molecular weight Mw of such a kind of random polyampholyte using Gel Permeation Chromatography (GPC), so we estimated Mw, roughly, from the overlapping concentration C* of the polymer solution where the viscosity increased dramatically. The overlap concentration determined by the viscosity in 4 M saline solution was C*~ 30 g/l, corresponding to the repeat unit ~ 0.15 mol/l. C * is related to the repeat unit size a (~ 0.3 nm), the average degree of polymerization N, and the coil size of a polymer chain R as: C ~ N A N 4 R 3 3 1/2 Assuming that the polymer is in the Θ solvent, R ~ an. So we have 3 N ~( ) 3 4 NAaC 2 Where, the NA is Avogadro constant ( mol -1 ). As a result, the degree of polymerization is N ~10000, and the corresponding molecular weight is around ~ g/mol. This value falls in the common range of radical polymerization. Commonly, the entanglement concentration Ce is about several times above the overlap concentration. As shown in Fig. 2a, the gel phase appears at Cm ~ 0.7 mol/l. This value is 5 times the value of C*, consistent with the entanglement concentration Ce. This supports the argument that the gel phase starts to appear above the entanglement concentration. NATURE MATERIALS 11

12 4) Deswelling-induced complex formation as revealed by cyclic testing a b 0.6 Stress (MPa) Equilibrium gel As-prepared gel Strain (mm/mm) Supplementary Figure 3 Schematics of polyampholyte hydrogel in the as-prepared state and equilibrium state in water. a, Illustration of the dialysis process from the as-prepared state to the equilibrium state of the gel. b, Cyclic stress-strain curves of the as-prepared and water-equilibrium gel P(NaSS-co-MPTC) The equilibrium state of the gel is measured in water, while the as-prepared state of the gel is measured in air. 12 NATURE MATERIALS

13 5) Lattice model for calculating the electrostatic interaction Supplementary Figure 4 Schematic lattice model of rod-like polymer chains for calculating the electrostatic energy. a, The cationic (red) and anionic (blue) groups are alternately distributed along the polymer chains. The distance between the cationic and anionic groups is the same. The rod consists of two concentric cylinders of inner radius r and outer radius R, which correspond to the radius of bare and hydrated macroions, respectively. b, Projection of the top part of the profile in (a). We use the lattice model shown in Supplementary Fig. 4 to estimate the polymer concentration dependence of the strength of polyampholyte hydrogels. From this lattice model, the polymer volume fraction, defined as the volume of dry gel divided by the volume of wet gel, can be expressed as V V dry wet r R 2 2, (1) while the polymer concentration Cpoly is related to the polymer volume fraction, C V ~ dry poly Vwet By substituting Eq. (1) into Eq. (2), we can get C poly r ~ R 2 2 Since the size of the bare macroions r is fixed, Eq. (3) becomes R C (4) 0.5 ~ poly (2) (3) NATURE MATERIALS 13

14 According to the electrostatic interaction theory 40, the ion association energy Eion is given by the equation E ion ~ C poly 2 Ne A 4 2R r 0 (5) Here, NA is the Avogadro constant, e is the charge of an electron, R is the hydrated radius of the ion, and 0 and r are the vacuum permittivity and relative permittivity of water, respectively. Then, Eq. (5) is simply described by the scaling relation, E C ~ poly ion R From Eq. (4) and Eq. (6), the power relations between the ionic interaction and the concentration of the polymer is derived as (6) E ~ C (7) ion 1.5 poly This scaling law is in agreement with the experimentally observed relation between the tearing energy T and the true polymer concentration Cpoly, 1.8 T ~ C poly (8) This agreement quantitatively illustrates the effect of the ionic interactions on the strength and toughness of the polyampholyte hydrogels. 14 NATURE MATERIALS

15 6) Solvent-induced shape memory a b f Writing Erasing c d e 37s 6min Supplementary Figure 5 Shape memory behaviour of polyampholyte hydrogel. The as-prepared hydrogel P(NaSS-co-MPTC) with its initial straight shape (a) is deformed into a spiral shape that can be written by immersing the sample in water (b). When the sample is stretched, the spiral shape is deformed to a straight shape (c); however, it recovers its spiral shape automatically after the force is removed (d, e, b). The full recovery process takes about 20 min in water at 20 C. The spiral shape can be erased in 0.5 M NaCl solution, which causes the sample to return to its initial straight shape (f). Since the ion complexes serve as cross-linking points and lock the polymer chain conformations, we can write any desired shape to the polyampholyte hydrogels during the ion complex formation process in water and erase the shape by dissociating the ion complexes in NaCl solution. As shown in Supplementary Fig. 5, when we deform an as-prepared hydrogel from its initial straight shape (Supplementary Fig. 5a) into a spiral shape and then immerse it in water, the spiral shape is NATURE MATERIALS 15

16 memorized (Supplementary Fig. 5b, write process). After that, even if the gel is forced to deform to the straight shape, it automatically returns to the spiral shape (Supplementary Fig. 5c, d, e). Furthermore, when we immerse the spiral shape gel in 0.5M NaCl solution, the memorized spiral shape is erased and the gel recovers to its initial straight shape (Supplementary Fig. 5f, erase process). In principle, the write and erase processes can be repeated many times. A softening temperature Ts ~ 48.2 C (Supplementary Fig. 7c) is observed, and the written spiral shape is memorized either below (25 C) or above (75 C, Supplementary Movie 3) this Ts in water, which confirms that the shape memory effect is solvent-induced. This effect is different from the shape memory effects of most polymers, which are based on the glass transition temperature of the polymer ) More polyampholyte hydrogels systems and chemical structure effect MTAC 4-VPC Supplementary Figure 6 Chemical structures of cationic monomers for the polyampholyte hydrogels. Methacrylatoethyl trimethyl ammonium chloride (MTAC) and 4-vinylpyridine chloride (4-VPC). 16 NATURE MATERIALS

17 Supplementary Table 2 Structural features and mechanical properties of various polyampholyte hydrogels. Hydrogels Composition cw E σb (MPa) εb Wb tanδ R Ts Cm-f* (wt%) (MPa) (m/m) (MJ/m 3 ) (%) ( C) polyampholyte hydrogels P(NaSS-co-MPTC) (1) P(NaSS-co-DMAEA-Q) (2) P(NaSS-co-4-VPC) (3) P(NaSS-co-MTAC) Common hydrogels PAAm Single network PAMPS/PAAm double network 2-4 (4) / (5) * Cm and f represent the total ionic monomers concentration (mol/l) and the charge fraction of the anionic monomer, respectively, in the precursor solution used to synthesize the gels. The parameters cw, E, σb, εb, Wb, tanδ, and R are the water content, Young s modulus, fracture stress, fracture strain, work of extension at fracture, loss factor (10 Hz and strain 0.5%), and shock-absorbing ratio, respectively, at room temperature. Ts is the softening temperature determined by the peak of loss factor of the gels. (1)(2) The optimal compositions at the minimum swelling ratio Qv used were to obtain the strongest mechanical properties, while other samples were synthesized at the stoichiometric ratio in feed. (3) The gels were synthesized from 4-vinylpyridine (4-VP), and then the as-prepared samples were immersed in 0.5 M HCl to ionize the weak bases 4-VP to their charged forms 4-VPC, before the immersion in water. (4) The abbreviated symbol 2-4 refers to 2 M AAm and 4 mol% MBAA in the precursor solution of the gel. (5) The 1-4/ notation represents 1 M AMPS and 4 mol% MBAA for the first network and 2 M AAm and 0.01 mol% MBAA for the second network in the precursor solution. NATURE MATERIALS 17

18 a Swelling volume ratio, Q v P(NaSS-co-MPTC) P(NaSS-co-DMAEA-Q) P(AMPS-co-DMAEA-Q) Swelling Deswelling Anionic monomer fraction in feed, f b Stress (MPa) P(NaSS-co-MPTC) 53.9% water content P(NaSS-co-DMAEA-Q) 52.6% water content Strain (mm/mm) c Loss factor, Tan P(NaSS-co-MPTC) P(NaSS-co-DMAEA-Q) Temperature, T ( o C) Supplementary Figure 7 Effect of chemical structure of the ionic monomers on the behaviours of hydrogels. a, Relationship between the swelling volume ratio Qv and the anionic monomer molar fraction f of the 3 sets of gels prepared at a formulation of f-4. All the gels deswell around their charge balance points (f ~ 0.5). b, Tensile behaviours of the two physical polyampholyte hydrogels prepared at the f = 0.52 where their Qv reaches a minimum: P(NaSS-co-MPTC) , P(NaSS-co-DMAEA-Q) c, Temperature dependence of the loss factor (tanδ) of P(NaSS-co-MPTC) and P(NaSS-co-DMAEA-Q) at 10 Hz and 0.5% strain. The temperature at which tanδ reaches the maximum corresponds to the softening temperature Ts, which is indicated by the numbers in the figure. The experiments were performed in water. This one-step approach to synthesizing polyampholyte hydrogels with multiple specific mechanical properties is quite general and can be applied to various combinations of oppositely charged 18 NATURE MATERIALS

19 monomers. The Young s modulus and the viscoelastic behaviours of the synthesized gels depend strongly on the specific chemical structure, especially the hydrophobicity of the monomers. To demonstrate these effects, we used two additional monomers, cationic methyl chloride quarternised N,N-dimethylamino ethylacrylate (DMAEA-Q) and anionic 2-acrylamido-2-methylpropanesulfonic acid (AMPS), as shown in Fig. 1b. Using the ionic monomer combinations, we compare the behaviours of three series of hydrogels with hydrophobicities in the order of P(NaSS-co-MPTC) > P(NaSS-co-DMAEA-Q) > P(AMPS-co-DMAEA-Q). Their swelling behaviours of the three series of chemically crosslinked gels at various charge fractions f are shown in Supplementary Fig. 7a. The most hydrophilic P(AMPS-co-DMAEA-Q) series exhibits a sharper V-shape than the other two series; that is, a slight deviation from the charge balance point destroys the ion complex. This indicates that the hydrophobicity has a synergistic effect in stabilizing the ionic interaction 23. The linear polyampholytes were prepared at the f where their Qv reaches a minimum in Supplementary Fig. 7, f = 0.49 for P(AMPS-co-DMAEA-Q) and f = 0.52 for P(NaSS-co-MPTC) and P(NaSS-co-DMAEA-Q) ). No physical hydrogel is formed from the most hydrophilic combination of AMPS and DMAEA-Q. In contrast, tough physical gels are formed in the relatively hydrophobic combination of NaSS and DMAEA-Q, the same like the combination of MPTC and NaSS. As shown in Supplementary Fig. 7b, the most hydrophobic physical hydrogels, P(NaSS-co-MPTC) , which contains 53.9 wt% water, is very tough and shows clear yielding, a high strength (σb = 1.7 MPa), and a relatively large extensibility (εb = 730%). The less hydrophobic physical hydrogels, P(NaSS-co-DMAEA-Q) , which contains 52.6% water, is very ductile, showing a large extensibility (εb = 1550%). Thus, the more hydrophobic the gel, the stronger the ion bond is. The less hydrophobic physical hydrogels, P(NaSS-co-DMAEA-Q) , shows almost perfect self-healing that is better than that of the more hydrophobic and rigid sample, P(NaSS-co-MPTC) After healing for 24 h at room temperature, the two cut surfaces completely merged together and the healing efficiency reached as high as ~ 99%, as shown in Fig. 3f. The strength of the ion bonds also dramatically influences the viscoelasticity and, therefore, the NATURE MATERIALS 19

20 damping behaviours of the physical hydrogels. For example, at room temperature, the less hydrophobic P(NaSS-co-DMAEA-Q) with a weak ion complex shows a shock-absorbance ratio R as high as 95.5%, while the hydrophobic P(NaSS-co-MPTC) with a strong ion complex exhibits an R of 76.6% (Supplementary Table 2). The viscoelastic and shock-absorption features of the polyampholyte hydrogels are presented in the movies showing the vibration and rebound experiment, in which the elastic double-network gel is used as the reference (Supplementary Movies 4 and 5). 8) Structure analysis a 250 b 0 Dry gel 200 Intensity (a.u.) Dry gel Wet gel Heat flow (uw) -1x10 4-2x10 4 Wet gel -3x q (nm -1 ) -4x Temperature ( o C) Supplementary Figure 8 Structure analysis by WAXS and DSC. a, WAXS spectra of wet and dry P(NaSS-co-MPTC) samples. The WAXS patterns were obtained by a Rigaku X-ray crystallograph under Cu radiation (λ = nm). The measurement was carried out using an X-ray generator with a voltage of 40 kv and a current of 20 ma. The specimen-to-detector distance was 250 mm, and the exposure time was 5 min. b, DSC scanning for wet and dry P(NaSS-co-MPTC) samples performed at a heating rate of 10 C/min from -150 C to 300 C. The two peaks (around 0 C and 120 C) in the DSC curves are assigned to the melting point and boiling point of water in the polymer, respectively. Thermal analysis was performed using a differential scanning 20 NATURE MATERIALS

21 calorimeter (Model DSC22, SII Nano Technology Inc.) connected to a thermal analysis system (model, SSC 5100). We have attempted to analyse the structure of the hydrogels P(NaSS-co-MPTC) using several methods. The gel shows no X-ray diffraction peaks in the wide-angle X-ray scattering (WAXS) range, while a broad peak at q ~ 12.7nm -1, corresponding to a characteristic length of 0.49 nm, appears for the dried sample (Supplementary Fig. 8a). Furthermore, no thermal melting peaks from the ion complex structure appear in differential scanning calorimetry (DSC) analysis either for the hydrogels or for the dried sample (Supplementary Fig. 8b). These results indicate that the polyampholyte hydrogels are amorphous with no crystalline structure, which is different from the behaviour of ionomers that form crystalline domains 43. 9) Rheological results a Storage modulus, G' Loss modulus, G'' (Pa) G' G'' Tan 0.1 C C 16.1 C C 32.1 C 40.1 C 48.1 C 56.1 C 64.1 C 72.1 C 80.1 C 88.1 C 95 C Frequency, (Hz) b Loss factor, Tan ln a T E a = 71kJ/mol E a = 308kJ/mol /T ( 10-3 K -1 ) Supplementary Figure 9 Dynamic mechanical behaviours of the polyampholyte hydrogels P(NaSS-co-MPTC) a, Frequency (ω) dependence of the storage modulus G, loss modulus G, and loss factor tanδ of polyampholyte hydrogel. The measurements were performed from 0.01 to 15.8 Hz at a shear strain of 0.5% at different temperatures from 0.1 C to 95 C, and the results were obtained by performing classical time-temperature superposition shifts at a reference temperature of NATURE MATERIALS 21

22 24.1 C. b, Arrhenius plot depicting the temperature dependence of the shift factors for the sample. The apparent activation energy values were calculated from the slope of the curve. These values are smaller than the covalent bond dissociation energy Ec-c ~ 347 kj/mol, which ensures the preferential dissociation of the ion complexes under deformation. The dynamic behaviours of the polyampholyte hydrogel at different temperatures and frequencies follow the principle of time-temperature superposition well. Supplementary Fig. 9a shows the master curves of G, G, and tanδ for the polyampholyte hydrogel at a reference temperature of 24.1 C. It is noted that G is larger than G over the whole frequency range from 10-7 to 10 6 Hz, indicating that the sample, even without any chemical cross-linking, is always in the gel state with predominantly elastic properties. The longest relaxation time, estimated as the reciprocal of the frequency at which the storage modulus reaches a plateau at about Pa, is about s at 24.1 C. The apparent activation energy Ea is obtained from the Arrhenius equation, at Ae Ea / RT, where at is the shift factor, R is the ideal gas constant, and A is a constant 44. The temperature dependence of the shift factor at shows that the activation energy of the gel varies over a wide range, kj/mol, as shown in Supplementary Fig. 9b, corresponding to kbt at room temperature. All these results indicate a wide distribution of strengths of the ion associations, which corresponds to the random structure obtained from the radical polymerization of the sample. The upper range of the activation energy is less than but close to the covalent bond dissociation energy, Ec-c ~ 347 kj/mol (~ 140 kbt). This explains why a rigid and tough gel is observed even without any covalent cross-linker in this system. We also found that the activation energy Ea of the less hydrophobic system hydrogel P(NaSS-co-DMAEA-Q) has a narrower distribution of Ea ( kj/mol) than the more hydrophobic system P(NaSS-co-MPTC) NATURE MATERIALS

23 10) Biocompatibility In order to evaluate the biocompatibility and anti-biofouling properties of polyampholyte hydrogels P(NaSS-co-MPTC), Chinese hamster lung fibroblast cells (JCRB0603:V79) and RAW macrophages (Primary Cell Co. Ltd. Japan) were used to perform the cytotoxicity test and adhesion test, respectively. Cytotoxicity test of polyampholytes hydrogels using V79 cells a Colony formation rate (%) Gel Gel Gel Positive RM-A Positive RM-B Concentration (%) b Colony formation rate (%) 120 Positive RM-C Concentration (ug/ml) Supplementary Figure 10 Cytotoxicity Test of substances using V79 cells. a, Colony formation rate of polyampholyte physical hydrogels Poly(NaSS-co-MPTC) , and , positive control substance (RM-A) and positive control substance (RM-B) with different concentration extracted from these substances. b, Colony formation rate of positive control substance (RM-C) with different concentration extracted from the substance. This study was conducted to investigate the cytotoxic effects of the extracts from polyampolyte hydrogel using Chinese hamster lung fibroblast cells (JCRB0603:V79) by the colony formation method, referring to the standard method: ISO : 2009 (E) Biological evaluation of medical devices Part 5: Tests for in vitro cytotoxicity; and ISO : 2007 (E) Biological evaluation of medical devices Part 12: Sample preparation and reference materials. Physical polyampholyte hydrogels Poly(NaSS-co-MPTC) with compositions , , and were used to investigate the cytotoxic effects toward the Chinese hamster lung NATURE MATERIALS 23

24 fibroblast cells (JCRB0603: V79); V79 cells obtained from the Department of Cancer Cell Research, Institute of Medical Science, University of Tokyo on March 9, 1994 were used. High-density polyethylene film (negative RM) was used as the negative reference material (abbreviation: negative RM) that shows no cytotoxicity. Polyurethane film containing 0.1% zinc diethyldithiocarbamate was used as the positive reference material A (abbreviation: positive RM-A) that shows moderate cytotoxicity. Polyurethane film containing 0.25% zinc dibutyldithiocarbamate was used as the positive reference material B (abbreviation: positive RM-B) that shows weak cytotoxicity. Zinc dibutyldithiocarbamate directly dissolved in dimethyl sulfoxide was used as the positive reference material C (abbreviation: positive RM-C). These controls were purchased from Hatano Research Institute, Food and Drug Safety Center. MEM culture medium containing 5% fetal bovine serum (abbreviation: M05 culture medium) were prepared by mixing the components at the following ratios. Firstly, Eagle MEM culture medium (9.4 g; containing kanamycin and phenol red) was dissolved in Japanese Pharmacopoeia water with a total volume of 1 L for injection. Secondly, the sodium bicarbonate solution was added to the sterilized mixture by autoclaving to adjust the ph to 7.2 to 7.4. Thirdly, MEM nonessential amino acid solution (0.09 mmol/l), sodium pyruvate solution (0.11 g/l), L-glutamine (0.292 g/l), and fetal bovine serum (inactivated at 56ºC for 30 minutes, 5%) were added to the system to achieve the final composition. The sterilized polyampholyte hydrogels by autoclaving (120ºC, 20 minutes), approximately 2 (thickness) 15 (diameter) mm, were put in a borosilicate glass medium bottle. Then the culture medium (10 ml/g) was added making the hydrogel reach the equilibrium state after 24 hours in a shaking incubator which was set under the condition of 37ºC, amplitude 70 mm, and 100 rpm. The culture medium after the extraction from the substrates (100% extraction) were collected and diluted to the prescribed concentrations as the testing solutions. The extracts obtained from the controls A and B were also diluted by the same method. Controls C was diluted with dimethyl sulfoxide to obtain the prescribed concentrations. The resulting solutions were used to test. Each test group for substrates with different concentration was assayed in triplicate. 24 NATURE MATERIALS

25 Cells in a growth phase were treated with 0.02% EDTA-0.25% trypsin (0.5M EDTA, 2.5% trypsin,) and harvested, and then suspended in the culture medium to obtain a cell density of 50 cells/ml counted by a hemocytometer. The suspension was dispensed to a 12-well multiwell plate (FALCON) at the volume of 1 ml (50 cells)/well which was placed in the CO2 incubator at 5.0% CO2 and 37.0 ºC. 1 ml/well of the test solution was added to the plate for culturing for 6 days before discarding of the initial culturing medium. After culturing, the culture medium for the test solution in the well was discarded, and then the cells on the substrate were rinsed with Ca 2+ and Mg 2+ free Dulbecco s phosphate buffer and fixed in methanol for approximately 5 minutes. Then the cells were stained with 5% Giemsa solution for approximately 10 to 15 minutes. Each well was observed with a stereoscopic microscope (SZ61TRC-C-D, Olympus Corporation) to count the colonies consisting of 50 cells or more. Any well obviously showing a decrease in the colony size (a decrease in the number of cells per colony) was also recorded. For each test series, the colony formation rate in each well was calculated regarding the mean number of colonies in the corresponding control group (0% or 0 µg/ml: blank) as 100%. When the colony formation rate decreased to 50% or less, the IC50 (concentration that inhibited the colony formation rate to 50% of the control mean value) value was calculated using regression analysis after logarithmic conversion of concentrations based on the concentration-response relationship. The cytotoxicity of the test substrate is classified as shown in the Supplementary Table 3 with reference to the Basic Principles of Biological Safety Evaluation Required for Application for Approval to Market Medical Devices, MHLW Notification 0301 No. 20, Office of Medical Devices Evaluation, Evaluation and Licensing Division, Pharmaceutical and Food Safety Bureau, Ministry of Health, Labour and Welfare, Japan, March 1, NATURE MATERIALS 25

26 Supplementary Table 3 Classification of cytotoxicity Degree of cytotoxicity (IC 50 value) 100% or more Weaker than the positive RM-B Stronger than the positive RM-B and weaker than the positive RM-A Stronger than the positive RM-A Classification of cytotoxicity No cytotoxicity, or very weak cytotoxicity Weak cytotoxicity Moderate cytotoxicity Severe cytotoxicity The colony formation rates of the polyampholyte hydrogels Poly(NaSS-co-MPTC) , , , positive control RM-A, RM-B, and RM-C are shown in Supplementary Fig. 10a and b. The extracts from the negative RM do not inhibit the colony formation of the cells. However, the colony formation is inhibited at concentration 2% or more of the extracts from the positive RM-A, at concentration of 60% or more of the extracts from the positive RM-B, and at concentration of 3µg/ml or more of the extracts from the positive RM-C, respectively. The IC50 values calculated from the colony formation rates are the concentration of 1.42%, 50.6% and 3.05 µg/ml for the positive RM-A, positive RM-B and positive RM-C, respectively. For polyampholyte hydrogels , , , no inhibition effect on colony formation phenomenon are observed during the concentration of 0% ~ 100%, which indicate the nontoxicity of polyampholyte hydrogels towards the V79 cells (IC50~ 100% or more) according to the degree of cytotoxicity in Supplementary Table NATURE MATERIALS

27 Cytotoxicity test and adhesion test of polyampholytes hydrogels using macrophages a b c Number of cells per cm Control of live cells Control of dead cells Hydrogels of live cells Hydrogels of dead cells 2 Time (h) 72 2 hours 72 hours 70um TCPS PAAm gel d P(NaSS-co-MPTC) gel 1 P(NaSS-co-MPTC) gel 2 Number of cells per mm TCPS PAAm gel P(NaSS-co-MPTC) gel 1 P(NaSS-co-MPTC) gel Time (h) 24 70um Supplementary Figure 11 The behaviors of macrophages on the different substrates. a, Number density of live and dead macrophages in solution after the cells were cultured on the hydrogel P(NaSS-co-MPTC) and on tissue culture polystyrene (TCPS) for 2h and 72h. b, Morphology of macrophages after culturing on TCPS for 2h and 72h in the presence of polyampholyte hydrogel P(NaSS-co-MPTC) c, Morphology of macrophages adhered on TCPS, PAAm hydrogel, P(NaSS-co-MPTC) gel 1, and P(NaSS-co-MPTC) gel 2 after culturing for 24h. Red arrows indicate macrophage adhered on the surface of substrates. d, Number density of macrophages adhered on the substrates after culturing for 2h and 24h. NATURE MATERIALS 27

28 In order to evaluate the anti-biofouling properties of polyampholyte hydrogels P(NaSS-co-MPTC), RAW macrophages (Primary Cell Co. Ltd. Japan) were used to perform the cytotoxicity test, including the direct and indirect contact tests 45-46, and adhesion test since that macrophages are highly adhesive cells responsible for immune response to implant materials. For the direct test, the cells are cultured on the surface of the hydrogel while for the indirect test, cells were cultured on the tissue culture polystyrene (TCPS) in the presence of hydrogels. The sterilized hydrogel samples with the disc-shape of 15mm in diameter for the cell culture were immersed in HEPES buffer solution to reach the swollen equilibrium state for one week by continuously exchanging the solution. The morphology of cells on the substrate surfaces were observed under the phase contrast microscope (OLYMPUS CKX31, Japan) equipped with a digital camera. The macrophages were diluted in Dulbecco s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and seeded on the substrate surfaces in a 5% CO2 humidified atmosphere at 37ºC. Cytotoxicity test Direct contact test Firstly, the cells were cultured on the surface of the physical polyampholyte hydrogel P(NaSS-co-MPTC) , and the TCPS was used as a control. The non-adherent cells were washed away from the hydrogels using PBS buffer solution while the trypsin was used to remove the cells adhered to the TCPS substrate. Then the collected cell suspension was mixed with trypan blue. The live and dead cells were counted by the hemocytometer under the microscopy. The initial seeded cell density was about and cells/cm 3 for the live cells and dead cells, respectively. As shown in Supplementary Fig. 11a, the number density of live and dead macrophages increase from and (2h) cells/cm 3 to and (after 72h culturing) cells/cm 3, respectively, when the cells were cultured on the P(NaSS-co-MPTC) On the TCPS control, the value changes from and (2h) cells/cm 3 to and (after 72h culturing) cells/cm 3, respectively. The hydrogel shows higher ratio of live cells to dead cells (3.8) than that on TCPS control (2.8), which demonstrate the nontoxicity of the polyampholyte 28 NATURE MATERIALS

29 hydrogels. Polyampholyte hydrogels of other compositions also exhibit the same non-toxic tendency (data are not shown). Indirect contact test In order to further confirm the nontoxicity of polyampholyte hydrogels toward macrophages, the cells were cultured on the TCPS in the presence of sample P(NaSS-co-MPTC) The initial seeded cell density was about cells/cm 3. Supplementary Fig. 11b shows the morphology of macrophages on TCPS after culturing for 2 h and 72h in the presence of P(NaSS-co-MPTC) Comparing with the initial seeding of the cells, the cell numbers increased after 72h culturing, indicating the proliferation of the cells. This result also indicates the non-toxicity of polyampholyte hydrogel P(NaSS-co-MPTC) towards the macrophages. Adhesion test For the cell adhesion test, we used hydrogels P(NaSS-co-MPTC) with 0.1 mol% and 0.3 mol% chemical crosslinker (coded as P(NaSS-co-MPTC) gel 1 and P(NaSS-co-MPTC) gel 2, respectively) to facilitate the direct optical observation due to their transparent features. We also used PAAm hydrogel (synthesized from 1 M AAm and 4 mol% MBAA) as a hydrophilic control. About cells/cm 3 cells were seeded on each of these samples, and after 2h and 24 h, the non-adherent cells were washed away from the substrates using PBS buffer solution. Supplementary Fig. 11c shows the optical images of macrophages on these substrates. A large number of macrophages adhere on the hydrophobic surface of TCPS while a small amount of macrophages adhere on the hydrophilic surface of PAAm gel. However, almost no cell adheres on the surfaces of the two P(NaSS-co-MPTC) hydrogels. The number of cell adhered on the surfaces, determined from three images of each sample, also confirms this, as shown in Supplementary Fig. 11d. The number of adhered macrophages on the surface of TCPS and PAAm gel increase with the culture time from 101±9.5 (2h) to 182±17.0 cells/mm 2 (24h), and from 1±1.4 (2h) to 9±2.8 cells/mm 2 (24h), respectively, while, there is no cell adhesion on the P(NaSS-co-MPTC) hydrogel 1 and the number of adhered macrophage slightly decreases from 3±4.2 (2h) to 1±1.4 cells/mm 2 (24h) on NATURE MATERIALS 29

30 P(NaSS-co-MPTC) hydrogel 2. These results indicate that the polyampholyte hydrogels have excellent anti-fouling properties against microphages. The anti-biofouling behavior of the polyampholyte gels is analogous to zwitterion polymers that usually show anti-biofouling properties The cytotoxicity test and adhesion test demonstrate that the polyampholyte hydrogels have excellent biocompatibility and anti-biofouling properties. 11) Mechanical test Tensile test Supplementary Figure 12 Geometry of tensile test sample. Sample thickness w = 2-3 mm. Fracture test In order to obtain the accurate tearing energy of these polyampholyte hydrogels P(NaSS-co-MPTC) with relatively high modulus, we use two models to determine this, including tearing test and pure shear test. 30 NATURE MATERIALS

31 a b c 6 5 Force, F (N) Streching distance, L (mm) Supplementary Figure 13 Tearing test to determine the tearing energy. a, Geometry of tearing test sample. Sample thickness w = 2-3 mm. b, Experimental picture of the tearing test. c, A typical force-extension curve of tearing test for polyampholyte hydrogel P(NaSS-co-MPTC) Tearing test For the tearing test, the method to determine the tearing energy was introduced in reference 36. As shown by the constant stretching force in Supplementary Fig. 13b, a steady state of crack propagation is obtained (Supplementary Fig. 13c). The tearing energy is calculated from the constant stretching force F as T=2F/w. T is ~ 3950 J/m 2 for P(NaSS-co-MPTC) NATURE MATERIALS 31