Designing Ionic Conductive Elastomers Using Hydrophobic Networks and Hydrophilic Salt Hydrates with Improved Stability in Air

Existing soft ionic conductors fall into two distinct categories: liquid‐rich ionic conductors containing large amounts of liquid electrolytes (≈70–90 wt.% water for hydrogels and ≈20–80 wt.% ionic liquids for ionogels), and liquid‐free ionic conductors that do not contain liquid components (e.g., ionic conductive elastomers). However, they are often plagued by dehydration, leakage of toxic ionic liquids, and air aging. Here, using hydrophobic polymer networks and hydrophilic salt hydrates, ionic conductive elastomers (s‐ICEs for short) containing only a tiny amount of bound water (≈1–5 wt.% are synthesized). Thanks to the small embedded water content, the s‐ICEs are advantageous over liquid‐rich ionic conductors in terms of enhanced mechanical/electrical stabilities and safety; they also outperform previously reported liquid‐free ionic conductors by avoiding air‐aging issues. The s‐ICEs introduced here also show excellent stretchability, good elasticity, high fracture resistance, desirable optical transparency and ionic conductivity, which are comparable to those of state‐of‐the‐art liquid‐rich and liquid‐free ionic conductors. With all the above advantages, the s‐ICE represents an ideal material for practical applications of soft ionotronics in ambient conditions.

example, most hydrogels eventually suffer from dehydration even under ambient conditions, thereby reducing their elasticity and ionic conductivity and considerably limiting the durability of hydrogels as ionic conductors. [16][17][18] Ionogels, another representative of liquid-rich ionic conductors, are advantageous over hydrogels in terms of non-volatility, broad working temperature, and wide electrochemical window. [19][20][21] Still, many ionogels are plagued by the leakage of toxic ILs, [22,23] especially when subjected to mechanical loadings such as squeezing, which may potentially lead to the corrosion of metal electrodes and safety-related issues due to the toxicity of ILs. Besides, the abundance of liquid components in liquid-rich ionic conductors may compromise the mechanical properties. For instance, the adhesion of hydrogels to various materials -from metals to elastomers and gels -is intrinsically problematic since loadcarrying polymer chains are diluted and water molecules hardly transfer forces. [24] Moreover, ionogels face a trade-off between ionic conductivity and mechanical performance (i.e., modulus and strength) since the addition of ILs enhances the conductivity but typically reduces the stiffness and strength. [13,22,25] In recent years, liquid-free ionic conductors composed entirely of crosslinked polymer networks and mobile ions have emerged. The rapid advancement of the field is illustrated by the advent of ionic conductive elastomers (ICEs), [26][27][28] in which both cations and anions are mobile, and ionoelastomers, [29] in which either anions or cations are fixed to the elastomer network while the other species of ions remain mobile. Existing liquid-free ionic conductors, however, suffer from limitations inherent to the hygroscopic nature of either the electrolytes or the hydrophilicity of the polymers. Many ionoelastomers either exhibit limited stretchability [29] or are opaque, [30] such that they cannot meet the requirements of soft ionotronics such as optoelectronic devices and optical fibers, which require stretchability and optical transparency. [31][32][33] In the last few years, our group reported novel stretchable ionic conductors, including liquid-rich and liquid-free ionic conductors. We investigated the effect of combining different electrolytes (e.g., different salts, and ionic liquids) in the same copolymer network of P (MEA-co-IBA). This copolymer network is an ideal backbone for ionic conductors, since it has excellent mechanical properties at room temperature and provides some degree of tunability through the monomer composition. To put our previous work in perspective, Figure 1 shows a map of the ionic conductivity versus strength of the two different systems synthesized from the same copolymer networks but different electrolytes. The ICEs developed based on the copoly mer networks containing different lithium salts (i.e., LiTFSI and LiClO 4 ) [27,34] show a moderate ionic conductivity but a high mechanical strength (i.e., ≈7 MPa). Note that the strength and ionic conductivity could be increased simultaneously due to the interaction between the polymer networks and ions of electrolyte salt. [27] Unfortunately, such an excellent combination of properties is only stable in dry conditions. In other words, the hydrophilic nature of Li + ions in the electrolyte salt makes liquid-free ICEs moisture sensitive, [34] resulted in an evolving (decreased) stiffness and stress at break over time due to the plasticizing effect of water. [34] Yet, stability in ambient air is essential for applications in the presence of oxygen and water. As a potential solution to this stability problem we developed copolymer networks swollen with hydrophobic ionic liquids such as [C 2 mim][NTf 2 ] and [BMMIm][TFSI] resulting in ambiently stable ionogels [16] but exhibiting a maximum strength of only ≈0.7 MPa. Therefore ionic conductors combining both a high stress at break and a long-term ambient stability remain desirable in practical applications.

Results and Discussion
In this work, we report an ionic conductive elastomer (i.e., s-ICE) with hydrophobic/hydrophilic integrated design using P (MEA-co-IBA) hydrophobic polymer networks and hydrophilic salt hydrates. Unlike liquid-rich ionic conductors containing a large amount of liquid (≈70-90 wt.% water for hydrogels; ≈20-80 wt.% ILs for ionogels) and liquid-free ionic conductors having no liquid component, the s-ICEs reported here contain ≈1-5 wt.% of bound water molecules (Figure 2a), offering longterm stability in terms of mechanical and electrical properties under ambient conditions. Thanks to the ultralow water content, the s-ICEs represent a new soft ionic conductor that can be practically used in ambient conditions or open environments for long periods of time, since they overcome stability issues of existing ionic conductors -which are related to either high or zero liquid content -including but not limited to dehydration and the leakage of toxic ILs for liquid-rich ionic conductors in ambient and harsh environments, as well as air-aging issue and low conductivity for many typical liquidfree ICEs and ionoelastomers. The s-ICEs also demonstrate ultrahigh stretchability, high fracture toughness, good elastic recovery, and self-adhesiveness, which are comparable to those of state-of-the-art liquid-rich and liquid-free ionic conductors. With all the above advantages, s-ICEs containing ultralow water content are an ideal material for practical applications of soft ionotronics.

www.advelectronicmat.de
in the hydrophobic liquid monomers used in this work and the bound water makes up ≈34% of the mass of LiClO 4 ·3H 2 O, leading to an s-ICE system containing ≈1-5 wt% of water. The glass transition temperature of the s-ICEs clearly decreases with increasing concentration of hydrated lithium salt as shown by the DSC scans ( Figure S1, Supporting Information), suggesting that the salt is dissolved in the polymer networks. The composition of each sample is given in Table 1, including the weight and molar fractions of monomers, electrolyte salt, and water. Further in this paper, we examine the physical properties of the s-ICEs and investigate how the ultralow water content contributes to enhancing the stability of mechanical and electrical performances of s-ICEs relative to their liquid-rich (i.e., hydrogels) and liquid-free counterparts (i.e., ICEs and ionoelastomers).
The obtained s-ICEs exhibit high stretchability, excellent reversible elasticity, good transparency and ionic conductivity. Figure 2c shows that the s-ICE-0.5 exhibits an ultra-high stretchability with a stretch at break in uniaxial extension of λ b = 22.48, being one of the most stretchable ionic conductors reported to date among both liquid-rich and liquid-free ionic conductive materials. The s-ICEs possess much enhanced stretch capability, strength and stiffness, relative to the pure copolymer P (MEA-co-IBA) that contains no lithium salt ( Figure S2, Supporting Information) due to the formation of physical bonds such as hydrogen bonds and lithium bonds between the ions of salt hydrates and polymer networks according to our previous findings. [27,34] Moreover, the s-ICEs possess excellent elastic reversibility. For example, the s-ICE-0.5 can immediately recover its original length after being stretched to 10 times its original length and released (Figure 2d), demonstrating highly elastic behavior of the material. The s-ICEs are also optically transparent: a 1 mm-thick s-ICE-0.5 sample has an average transmittance of 85% over the entire visible range, as shown in Figure 2e; the inset shows that the whole spectrum of visible colors can be seen through the s-ICE-0.5 sample. In addition, the s-ICEs present a desirable ionic conductivity at room temperature for soft ionotronic devices. As the concentration of LiClO 4 ·3H 2 O increases from 0.25 to 1 M, the ionic conductivity at room temperature of the s-ICEs can reach 1.95 × 10 −3 S m −1 (Figure 2f), orders of magnitude higher than that of many existing liquid-free ICEs and ionoelastomers, [26,29] because the existence of water molecules -though in very limited amount (i.e., 3 water molecules per lithium ion) -can effectively facilitate the motion of lithium ions. [34] To gain comprehensive insights into the mechanical behavior of the material, we perform uniaxial tensile tests on four series  (Figure 3a), higher than that of most previously reported ionic conductors. [14,16,17,26,28,29,36] Tensile strengths and Young's moduli of all series of s-ICEs extracted from the uniaxial tensile tests are on the same level ( In our experiments, we surprisingly observe small or no differences in mechanical performance and fracture toughness between the s-ICEs containing different content of salt hydrates. The reason behind this result may be due to two factors. First, interactions between polymer chains and ions of lithium hydrates, stiffen and strengthen the material [27,34] and (i.e., Figure S2, Supporting Information) while small amounts of bound water play the role of a plasticizer. When the concentrations of salt hydrates increase, both effects (i.e., the interaction between the polymer chains and ions of the hydrated salt and plasticizing effect of the bound water) vary simultaneously with opposite influences, resulting in no significant differences in bulk mechanical and fracture behavior of the material. Second, the differences in volume fraction of polymer, which affects the mechanical properties of the materials (Table S1, Supporting Information) are small. Small differences in the volume fractions of the s-ICEs leads to ignorable differences in the mechanical performances.
To investigate the level of reversible elasticity of s-ICEs, we carry out cyclic-loading tests of s-ICEs. In the first set of experiments, samples are stretched to a maximum stretch ratio of λ max = 11 before unloading. All the samples exhibit nearly identical hysteresis loops and levels of energy dissipation, independent of the salt concentration (Figure 3d,e; Figure S4, Supporting Information). The corresponding recovery ratio (defined as 1 − ε res /ε max , where ε res is the residual strain and ε max = λ max − 1 is the maximum strain.) of all the s-ICEs exceeds 90% under a 100 mm min -1 displacement speed (the corresponding stretch rate = 0.14 s −1 ), indicative of the highly reversible elastic behavior of the material at large deformations, regardless of the salt concentration (Figure 3e). In the second experiment, we cyclically stretch the s-ICE-0.5 samples to increasing maximum stretches -from λ max = 2 to λ max = 13. The s-ICEs demonstrate an ultimate recovery ratio of ≈90%, with a corresponding residual stretch of only λ res = 2.24 after being released from λ max = 13 (Figure 3f and Figure S5, Supporting Information). The dissipated energy U d during each cycle increases almost linearly with λ max (Figure 3g). These results indicate a highly elastic behavior of s-ICEs with rapid elastic recovery and low hysteresis at high stretch ratio, a trait highly desirable for functional materials to find uses in soft ionotronics. Soft ionotronic devices are highly integrated systems and are typically comprised of various materials with distinct functionalities. In such systems, robust interfacial adhesion between different components is a fundamental requirement, which dictates the overall structural integrity and reliability of such devices. Achieving strong interfacial adhesion of hydrogels (a typical liquid-rich ionic conductor) to other functional materials, however, has been challenging due to the abundance of water molecules, [37] which changes neighbor readily and transmit forces negligibly. The adhesion energy of natural hydrogel-elastomers interface without special interfacial treatment is typically below 1 J m -2 ; [38] existing bonding methods that enable strong adhesion between hydrogels and elastomers are often restricted to specific chemistries. [39][40][41] Unlike conventional hydrogels, the s-ICEs reported here demonstrate instant tough bonding to a wide variety of materials -from hard to soft -with high interfacial toughness: the 90° peeling tests show that the adhesion energy between the s-ICEs and copolymers of P (MEA-co-IBA), glass, copper, and VHB are 752.6, 685.7,  (Figure 3h,i). Owing to the self-adhesive nature of the s-ICEs, strong interfaces -including s-ICE/copolymer, s-ICE/glass, s-ICE/copper, s-ICE/VHB -can be formed immediately by simply attaching s-ICEs to surfaces of different materials. The interface can sustain large stretches > 10 without interfacial debonding, demonstrating a desirable level of adhesion of the s-ICEs to various surfaces ( Figure S6, Supporting Information).
It has been widely recognized that liquid-rich ionic conductors such as hydrogels and ionogels suffer from key limitations inherent to liquid electrolytes, which may leak under mechanical loadings, [42] evaporate in a dry environment, [18] or easily take in water molecules from surroundings, [22] thereby changing the mechanical and electrical properties of the material. Liquid-free ionic conductors, including ICEs and ionoelastomers, contain no liquid component, avoiding the hassles of leaking liquid materials; however, many of them are prone to water absorption and hydroxylation and are subject to humidity-induced aging under ambient conditions. [27,29] To demonstrate the ambient stability of the s-ICEs, we expose s-ICE-0.5 samples to the ambient atmosphere (≈30-50% relative humidity, RH) for 72 h and measure the stress-strain curves at 24 h intervals. The s-ICEs retain their mechanical properties during the entire testing period, which is indicative of the excellent structural stability of s-ICEs under ambient conditions (Figure 4a-c and Figure S7a, Supporting Information). For comparison, we synthesize two types of liquid-free ICEs -P(MEA-co-IBA) networks containing lithium salt LiTFSI and LiClO 4 , respectively -and measure their stress-strain curves after storing them in ambient air for 24 h. The liquid-free ICEs exhibit degraded mechanical performance after 24 h of storage due to moisture absorption ( Figure S8, Supporting Information). More evidence of the airaging behavior of liquid-free ICEs containing LiTFSI or LiClO 4 salt has been provided in the literature, [34] revealing that water molecule-assisted lithium bond breakage is the underlying mechanism for the degradation of the mechanical properties of liquid-free ICEs under ambient conditions. [34] To further showcase the s-ICEs' tolerance to extremely dry conditions, we store s-ICE-0.5 samples in a humidity-free nitrogen environment (water content <<0.01 PPm, oxygen content <<0.01 PPm). It is observed that, after 72 h of storage, the s-ICE samples do maintain their mechanical properties (Figure 4a-c, Figure S7b, Supporting Information), showing that the mechanical performance of the s-ICEs is insensitive to extremely dry conditions, in conspicuous contrast to liquid-rich ionic conductors, especially hydrogels. [18] Changes in mechanical properties of liquid-rich gels and liquid-free ICEs are often related to leakage of liquids (i.e., water loss in hydrogels) or moisture absorption (i.e., water uptake of liquid-free ionic conductors). By contrast, the mechanical properties of the s-ICEs presented in this work appear relatively stable under dry and ambient conditions with varying humidity levels. To further confirm this result, we quantify the water content absorbed by the s-ICEs from environments with different humidity levels. Figure S9a (Supporting Information) displays the weight change of s-ICEs under ambient conditions, with the RH level changing from 30% to 70%. After 96 h storage, the s-ICEs gain negligible weight by water absorption -less than 0.1 wt.% for all s-ICE samples, hardly affecting the mechanical properties of s-ICEs as demonstrated in Figure 4a-c. To gain comprehensive insights into the electrical performance of the s-ICEs, we examine the ionic conductivity of the s-ICEs under various RH levels. The s-ICEs demonstrate a relatively stable ionic conductivity for 30% < RH < 50% and exhibit a slightly higher ionic conductivity when the RH reaches ≈70% ( Figure S9b, Supporting Information). This is in sharp contrast to nearly all liquid-rich hydrogels and some liquid-rich ionogels, which tend to lose ionic conductivity due to dehydration or the leakage of ILs [16,22,42] Molecular dynamics simulations are conducted to explain the water retaining capability of s-ICEs under extremely dry conditions at the atomistic scale. For simplicity, we only consider modeling the MEA chains to represent the copolymer network because MEA is the majority component. Figure 5a,b describes the simulation model. Twenty water molecules and one LiClO 4 molecule are embedded in the central region of a sphere-shaped MEA matrix, resembling a core-shell structure (illustrated in Figure 5a) where the core part is the H 2 O-LiClO 4 cluster and the shell part is the MEA matrix (Figure 5b). In the simulation, the entire core-shell structure is subjected to a vacuum environment -much harsher than that in the glove box (namely, a dry nitrogen environment). All simulations are carried out at room temperature. The simulation results show that even under such harsh conditions, the LiClO 4 can www.advelectronicmat.de effectively retain several water molecules. Figure 5c shows the time evolution of the spatial distance histogram accounting for the number of water molecules at certain distances away from the lithium ion, and the insets show the configurations of water molecules around the lithium ion at the corresponding moment. At the beginning of the simulation (i.e., at 0 ns), water molecules are dispersed around the lithium ion by the construction of the model. As the simulation time increases, the water molecules begin to diffuse outward, which can be read from the evolving histogram (Figure 5c). If a water molecule has diffused to the surface of MEA shell, it is immediately removed in order to simulate the evaporation process. From the histograms, we can see that the longer the simulation time, the lesser the total number of water molecules. However, there are still 4-5 water molecules in close vicinity to the lithium ion (i.e., at the distance of 0.25 nm) after about 185 ns (see blue line in Figure 5d), in sharp contrast with the diminishing number of water molecules at other distances. This is a key signature of the water retaining capability of the s-ICE obtained from the simulation. To further clarify such an effect, we construct a control model in which the initial number of water molecules is changed to 4, and find that these water molecules are not diffusing outward and are bound closely to the lithium ion throughout the simulation (see green line in Figure 5d). By inspecting the simulation trajectories, it is found that the water retaining capability is due to lithium bonds formed between the oxygen atoms in the water molecules and the lithium ions. Lithium bonds are strong physical interactions able to attract water molecules, which is the reason why s-ICEs can retain water in extremely dry environments.
We also evaluate the electrical properties of a s-ICE-0.5 sample subjected to uniaxial tension. When the s-ICEs are www.advelectronicmat.de stretched, the resistance increases nearly quadratically with increasing strain (Figure 6a). The result is well in line with the ideal elastomer model where the volume and resistivity of the material are taken to be fixed: When the s-ICE is stretched by a factor of λ, the cross-sectional area of the gel is reduced by a factor of 1/λ, the resistance is thus R/R 0 = λ 2 , [27] as shown in Figure 6a. The force-responsiveness of the s-ICEs makes it an ideal material for the design of a resistive sensor. Thus, we construct a simple resistive-type sensor by directly connecting metal conductive wires to the two ends of the s-ICE-0.5 strip. Cyclically stretching the sensor to different stretches of λ = 1.5, 2, 2.5, 3, 3.5 produces repeatable and straindependent resistance changes (Figure 6b). The sensing capability of the s-ICE-based sensor is further demonstrated by directly adhering the gel film to human fingers, and bending the finger to 90° produces a reliable resistance change signal (Figure 6c). In addition, the change in resistance of the s-ICEbased sensor exhibits excellent stability and repeatability during cyclic stretching for 100 consecutive cycles to λ max = 3 (Figure 6d), indicating its prominent durability. Figure 7 shows the difference between different ionic conductors based on the same polymer networks of P (MEA-co-IBA), covering the ICEs containing different lithium salts (i.,e dry LiTFSI and LiClO 4 , respectively), [27,34] ionogels based on the ionic liquid [C 2 mim][NTf 2 ], [16] and the s-ICEs using hydrated salt introduced in this work. The design principles of these materials include:1) In all cases, the monomer composition is kept the same and the absence of crosslinker means that the materials are highly stretchable; 2) The two quantities that change depending on the additive (salt, hydrated slat, IL,  www.advelectronicmat.de water) are the stress at break and conductivity. The differences between the materials based on the different additives are: 1) The ICEs based on the dry lithium salts including LiTFSI and LiClO 4 exhibit the best mechanical strength but a lower ionic conductivity; 2) the ionogels possess the highest ionic conductivity and lowest mechanical strength among all; 3) the ionic conductivity and mechanical strength of the s-ICEs developed here by using hydrated lithium salts are a compromise between ICEs and ionogels (Figure 7a). However, the stability in the ambient conditions is important when it comes to the applications of the materials in the presence of water and oxygen under varying humidity levels. As Figure 7b shows that the s-ICEs synthesized here using the salt hydrates exhibit a higher ionic conductivity than that of the ICEs based on dry lithium salts and are stable over varying humidity levels for a long period of time under ambient air, while having a markedly higher strength than that of the corresponding ionogels.

Conclusion
In conclusion, by using hydrophobic elastomer networks and salt hydrates, we report a new hydrophobic/hydrophilic integrated design of novel ionic conductors -the s-ICEs, which contain only a tiny amount of water (about 1-5 wt.%), inherently different than existing ionic conductors that are either liquid-rich or liquid-free. Molecular dynamics simulations strongly suggest that the water present in s-ICEs in humid environments does not completely evaporate in dry environments. As evident by both experiments and molecular dynamics simulations, this small intrinsic water content of the s-ICEs imparts enhanced safety and environmental stability of both electrical and mechanical properties relative to liquid-rich ionic conductors (e.g., ionogels and hydrogels) and liquid-free ionic conductors under ambient conditions. The material also demonstrates excellent stretchability, good reversible elasticity, high fracture resistance, spontaneous adhesion to various surfaces, and desirable optical transparency. We finally demonstrate the use of s-ICEs as resistive-type stretch sensors. We hope that our work may inspire a series of new soft ionotronic devices with stable performance based on s-ICEs.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.