*3.2. Mechanical Properties*

To well investigate mechanical properties of 3D printed hydrogels, tensile test and pure shear test were performed to characterize strength and fracture energy of 3D printed hydrogels. After photopolymerization and calcium chloride soaking, 3D printed tough hydrogels were successfully achieved. As shown in Figure 3a, the 3D printed hydrogels could be easily wrapped into a knot without any damage. After stretching, this knot could return to the original state. This result indicates that the 3D printed tough hydrogels own excellent mechanical properties and good elasticity. The stress-strain curves showed that the 3D printed hydrogels without calcium chloride solution soaking own a weak tensile strength, however, the tensile strength is largely improved after soaking process (as shown in Figure 3b). The improved tensile strength derives from formation of calcium ionically crosslinked alginate network during soaking process. The broken of hydrogels usually contains two sequential steps: initial fracture formation (nucleation) and subsequent fracture propagation (growth) [17]. In A2S2 gels, no nucleation is found due to the agar chain pullout mechanism [37]. With increasing strains, the agar chains pullout from the aggregated agar helical bundles progressively. However, the agar network remains integrated during this process. With addition of alginate chains, the weak entanglements between chains nearly restricts the pullout process of agar. This mechanism allows that the A2S2 gels own relatively long elongation. However, the A2C2 gels showed a distinctive breaking process with a higher strength and smaller elongation. With the addition of the Ca2+, the alginate chains are ionically crosslinked. The stress is concentrated on the ionically crosslinked alginate chains and unzip ionically crosslinked alginate chains preferentially leading to breakage of printed hydrogels at relatively low elongation. Therefore, A2C2 gels achieved the higher toughness at 603.22 ± 61.78 kJ m<sup>−</sup><sup>3</sup> while the A2S2 gels depicted a little smaller toughness at 493.27 ± 42.00 kJ m<sup>−</sup>3, although the A2C2 had a larger strength 488.75 ± 58.31 kPa and the A2S2 owned a very small strength at 142.67 ± 19.60 kPa (Table 2). In addition, the mechanical properties of printed gels with different compositions have been also summarized in Table 1 and Figure S2. With increasing amount of alginate and agar, the strength and toughness are improved, however, the elongation is nearly changed due to the existence of calcium

crosslinked alginate network. Table S2 and Figure 3c summarized the mechanical properties of the printed gels with various printing infill angles. Three different infill angels including 0◦, 45◦ and 90◦ were employed in this experiment and the design of 3D printing constructs was presented in Figure S1. The mechanical strength and elongation of these printed gels were similar. Therefore, the presence of calcium crosslinking of alginate network makes the infill have no influence on mechanical properties of printed gels.

**Figure 3.** (**a**) Photograph of 3D tough hydrogels, (1) original state, (2) under tension and (3) back to original state, (**b**) stress-strain curve of 3D printed gels with or without CaCl2 soaking, the insert is strength of 3D printed gels, means with different letters are statistically different at P < 0.05, (**c**) strength and elongation of 3D printed gels with different printing angles (A2C2), (P > 0.05), means with different letters are statistically different at P < 0.05.


**Table 2.** The mechanical properties of the printed hydrogels with different composition.

The pure shear test was used to calculate the fracture energy of 3D printed tough hydrogels. The schematic diagram of the pure-shear test was presented in Figure 4a and detail of measurement was located in experimental section. The fitted stress-strain curves of notched and un-notched hydrogels clearly were presented in Figure S3. According to Equation (2), the fracture energy of 3D printed hydrogels with different compositions were calculated. As shown in Figure 4b, with calcium ionically crosslinked alginate, the fracture energy is largely improved due to increased crosslinking points. When increasing the amount of agar chains, the fracture energy also increases, which demonstrates that higher crosslinking degree leads to higher fracture energy.

As shown in Figure 5a, the fitted stress-strain curve of notched samples with various compositions clearly presents that the strength at fracture are both improved with increasing amount of agar and with calcium chloride solution soaking. In addition, we can clearly find that the weak gel (A1C2 and A1S2) crack propagates very quickly and even immediately (Figure 5a,b). That means the notched sample is totally treated in a quick time. However, with increasing agar amount, the crosslinking degree is improved and the notched gel sample of A2S2 is slowly torn (Figure 5c). However, after calcium chloride soaking, the 3D printed tough gels (A2C2) have much slower crack propagation (Figure 5d). That means the combination of agar and calcium alginate (CA) network can greatly prevent the crack propagation.

**Figure 4.** (**a**) Schematic diagram of the pure-shear test for measuring fracture energy of hydrogels, and (**b**) Fracture energy of 3D printed hydrogels, means with different letters are statistically different at P < 0.05.

**Figure 5.** (**a**) Fitted stress-strain curve of notched samples with different compositions, and stress-strain curve of notched sample (**b**) A1S2, (**c**) A1C2, (**d**) A2S2, and (**e**) A2C2.

#### *3.3. Swelling and Cytotoxicity*

The swelling properties of printed gels are also systematically investigated. In the soaking process, the water molecules penetrate into polymeric hydrogels resulting in an expansion of polymeric networks and a low concentration region of polymeric chains, which leads to mechanical fracture [11]. Therefore, di fferent chemical structure and crosslinking density can result in di fferent swelling properties. As shown in Figure 6a, the A2S2 gels depict a much larger swelling ability, compared to A2C2 printed gels. The presence of calcium ionically crosslinked alginate not only increases the crosslinking degree, but also restricts the agar and polyacrylamide chains, causing limited swelling ability.

On the other hand, with increasing amount of agar and calcium crosslinked alginate, the swelling ability is quenched, which demonstrates that the high amount of hard network leads to an inferior swelling ability (Figure S4a and S4b). These results are consistent to the mechanical properties of printed gels. As shown in Figure S4c, the printed gels with di fferent infill angles present exactly the same swelling ability, which demonstrates that the infill angles have no influence on chemical structures and crosslinking degree.

The cytotoxicity of 3D printed hydrogels was also evaluated via U87-MG cells. Figure 6a depicted similar viable cell quantities between the control group and A2S2 hydrogels conditioned group in both 24 and 48 h. These results reveal that the 3D printed hydrogels own high biocompatibility after removing the unreacted acrylamide monomer. Compared to 24 h culture group, the number of viable cells after 48 h culture increased. This result demonstrates that the gels conditioned medium cannot affect cellular reproduction. The Figure S5 presented the live and dead cell images. Cells circled with green were viable, indicating these 3D printed hydrogels own high biocompatibility. After removal of the unreacted acrylamide monomer and other residues, the 3D printed hydrogels own low cytotoxicity with excellent mechanical strength and toughness, which can be considered as a potential candidate in wearable electronics.

**Figure 6.** (**a**) Swelling ratio of A2C2 and A2S2 gel, (**b**) Cell viability of U87-MG cells after culturing 24 h and 48 h (P > 0.05), means with di fferent letters are statistically di fferent at P < 0.05.

#### *3.4. Conductivity and Sensors*

As shown in Figure S6a, high transparency of 3D printed gel was achieved, which allowed these 3D printed hydrogels to detect resistance without a ffecting optical signals. Due to existence of large amounts of water inside hydrogels, hydrogels o ffer physical similarity to biotissues, and also own excellent capability to contain lots of ions [38]. By introducing Ca2+ and Cl- ions in the 3D printed hydrogels, conductive hydrogels could be used as ionic wires in the circuit. The Figure 7a depicted the bright light when 3D printed conductive hydrogels were connected into the circuit. The light was <sup>o</sup>ff, when conductive hydrogels was moved out of circuit. It demonstrates that 3D printed hydrogels are ionically conductive by containing Ca2+ and Cl- ions. The resistance of conductive hydrogels is measured via a multi meter. The conductivity of 3D printed hydrogels is around 13.9 mS/cm, which is similar to the conductivity of electrolyte, which demonstrates that the conductivity of conductive

hydrogels is close to the conductivity of calcium chloride solution. To well investigate conductivity of hydrogels, calcium alginate (CA)/PAAm double network (DN) hydrogels were fabricated via injection molding method according to our previous work with various concentration of ions, SA and AAm [11]. As shown in Figure S7a, with increasing concentration of CaCl2, the conductivity of DN hydrogels enhances. In addition, the conductivity of DN hydrogels with various concentration of CaCl2 were close to the that of CaCl2 solution with same concentration. These results strongly indicate that conductivity of hydrogels derives from ions. Furthermore, Figure S7b and S7c exhibited no significant change of DN hydrogels' conductivity with various concentration of SA and AAm. It demonstrates that the polymeric network in DN hydrogels has no e ffect on conductivity of ionically conductive hydrogels.

Figure 7b showed reduced brightness of the light when the 3D printed hydrogels were stretched, indicating that the stretch largely enhanced the resistance of hydrogels. This result exhibits that the conductivity of these conductive hydrogels is dependent on the strains. When the stretch was released, the brightness of light turned up again. To further investigate the resistance change of 3D printed hydrogels, the resistance change was measured with various stretches. We first assume the 3D printed hydrogels are incompressible and the conductivity is constant during stretching (Figure 7c). The resistance ratio is given by R/R0 = (L/L0) 2, where R and R0 mean the resistance of the stretched hydrogels and initial hydrogels, respectively. Figure 7d showed that the experimental data of the conductive hydrogels were close to the curve of theoretic equation. The small deviation of experimental data from theoretic curve might be caused by damage in the hydrogels. These results indicate that the measured resistance of the conductive hydrogels is reasonable. By comparison with other electronic conductors like indium tin oxide (ITO), silver nanowires (AgNWs), graphene, single-wall nanotubes (SWNTs), the conductive hydrogels, ionic conductors, owned lower conductivity than these mentioned electronic conductors. However, when high transmittance and stretchability are necessary, these conductive, transparent and stretchable hydrogels have specific advantages. At high stretch, these hydrogels had lower sheet resistance than these above mentioned electronic conductors. This result is consistent with previous literature [39].

**Figure 7.** (**a**) Photograph of printed hydrogels connected in the electric circuit, (**b**) photographs of the light changes of elongation connected in the electric circuit, (**c**) schematic diagram of hydrogels when stretching, (**d**) the normalized resistance of printed hydrogels is measured as a function of stretch, means with di fferent letters are statistically di fferent at P < 0.05, plotted against the ideal geometric behavior, and normalized resistance for ITO [40], AgNWs [41], graphene [42], SWNTs [43].

As depicted in Figure 8a, the resistant change of these conductive hydrogels showed a good linear correlation to strain in a range of 0–1.5, which indicates that the conductive hydrogels own a relatively large sensing range. The strain sensitivity of conductive hydrogels can be defined as the slope of resistance change rate ( *R*−*R0*/*R0*=Δ *R*/*R0*) versus applied strain (λ), formulized as *S*=δ(Δ *R*/*R0*)/δλ [38]. A gauge factor (3.83) was achieved via these conductive hydrogels that is superior to previously reported hydrogel-based strain sensor (0.478) [38]. This result demonstrates the conductive hydrogels exhibit a high sensing sensitivity. As shown in Figure 8b, the relative resistance change of these conductive hydrogels was exhibited during a step-by-step loading-unloading cycle at di fferent strains. This curve clearly showed the relative resistance of conductive hydrogels had a step-like trend. In addition, the relative resistance directly increased or decreased when conductive hydrogels were stretched or released to certain strain. It is meaningful that conductive hydrogels have quick response ability, because no hysteresis was found during strain change. Moreover, a good sensing stability of conductive hydrogels was also observed, since the relative resistance was kept stable during load-holding or unload-holding period at di fferent strain. As a wearable strain sensor, it is also important for the conductive hydrogels to own high stability. The Figure 8c showed the relative resistance change for 100 tensile cycles under 10% strain. The resistance of the conductive hydrogels was similar to the original level and these hydrogels showed no visible damage or delamination after 100 cycles of 10% strain. These results demonstrate that the conductive hydrogels have superior stability in sensing. The Figure 8d showed no obvious relative resistance change in temperature range of 20 to 40 ◦C, which means that these conductive hydrogels own high reliability when they are attached onto the human body.

**Figure 8.** (**a**) The dependence of sensing sensitivity of conductive hydrogels with the applied strain. The strain sensitivity (S) can be defined as the slope of resistance change rate (Δ*R*/*R0*) versus applied strain (λ), formulized as *S* = δ(Δ*R*/*R0*)/δλ, means with di fferent letters are statistically di fferent at P < 0.05, (**b**) The relative resistance changes vs time when a loading−unloading cycle of conductive hydrogels at di fferent strains, (**c**) the resistance ratio of conductive hydrogels as a function of fatigue cycle number (P > 0.05), means with di fferent letters are statistically di fferent at P < 0.05, and (**d**) the resistance ratio of conductive hydrogels as a function of temperature (P > 0.05), means with di fferent letters are statistically di fferent at P < 0.05.

In addition, we developed a wearable resistive strain sensor by these 3D printed conductive hydrogels and fixed it onto an index finger by copper tapes to monitor finger bending (Figure 9a). When the index finger bended step-by-step, the relative resistance change of this sensor rose up in a step-like trend (Figure 9b), which is similar to results in Figure 8b. The relative resistance of this wearable resistive strain sensor directly enhanced without hysteresis after the finger bending to a gesture, which exhibits fast response ability. During gesture-holding process of the index finger, the resistance of the sensor could remain at a constant, which shows the good sensing stability of this wearable resistive strain sensor. The Figure 9c also depicted a repeatable response during finger bending, which demonstrated that this wearable strain sensor has a good sensing stability in repeated usage. These results indicate that this conductive hydrogels-based strain sensor can be applied for human motion monitoring with high sensing sensitivity and stability.

**Figure 9.** (**a**) Photographs of finger bending, (**b**) resistance change when finger bending and (**c**) repeated response of the resistive strain sensor.
