Liquid-free ionic conductors, known as ionoelastomers, are of great interest because of their potential for developing reliable and resilient ionic devices with elastic mechanical properties. This study presents an intrinsically stretchable diode consisting of a highly stretchable ionoelastomer bilayer vertically stacked with liquid metal electrodes. The bilayer contains two types of ionoelastomer networks: one containing fixed anions with mobile cations and one containing fixed cations with mobile anions. Both ionoelastomers use 2-hydroxyethyl acrylate to provide high stretchability. The junction between the two ionoelastomers creates a diode with excellent non-Faradaic ionic current rectification. The voltage-dependent modulation of the ionic double layer at the interface between the two ionoelastomers provides the rectification. The elastic diode works under cyclic loading to a uniaxial stretch ratio of 6 (strain of 500%), allowing the development of a highly stretchable ionic OR logic gate.
I. INTRODUCTION
Soft deformable ionotronics—that is, conductive devices based on the manipulation of ions, rather than electrons—is of great interest since living organisms use ionic signaling. This feature makes ionotronic devices potentially suitable for various applications in human–machine interfaces, soft robotics, and bionic prosthetics.1–6 Most of these applications require high stretchability, ionic conductivity, and optical transparency, which are difficult to attain with conventional electronic counterparts. Substantial efforts have been made to develop a high-performance ionic junction of two types of ionic conductors for ionic current rectification, equivalent to an electronic p- and n-type junction. An ionic conductor pair with ionic current rectification is one of the most crucial building blocks in ionotronics because of its versatile applications in ionic diodes, transistors, and integrated logic circuits.7–17 Examples of successful rectification and switching of ionic currents include aqueous polyelectrolytes,7 hydrogels,8 bipolar membranes,9 charged nano-channels,10 and metal-electrolyte redox reactions11,12 with various applications in biosensors, soft power generators, and ionic circuits.13–15 More recently, stretchable solid-state ionic diodes and transistors based on polyelectrolyte gels were demonstrated.16,17
Critical issues should be addressed to ensure excellent stability with the reliable stretchability and ionic conductivity of a soft ionic conductor: (i) control of the solvent involved in ionic conductors in ambient conditions and (ii) elaborate design of contact with a metallic electrode for ionic-to-electric current conversion during device operation as well as circuitry interconnection.18–24 Although ionic conductors based on hydrogels and ionogels are promising, the use of such liquid electrolyte-type conductors containing large amounts of solvent, ionic salts, and/or ionic liquids is often limited because of the leakage and volatility of the liquid components, resulting in rapid degradation of the ionotronic device with time.25–27 In addition, contact between a hydrogel conductor and a metal electrode often involves a Faradaic electrochemical redox reaction, which can dissolve the metal electrode and result in undesired gas generation and changes in chemical composition, making the operation of the device unstable and unreliable.
Recently, elastomeric liquid-free ionic conductors, known as ionoelastomers, have attracted significant attention because of their nonvolatile nature involving no leaky liquid components and thus excellent environmental stability.28–42 Most ionoelastomers contain nonvolatile ionic liquids with pairs of mobile ions in polymer networks.28–36 They have been successfully employed as mechanically robust ionic conductors for ionotronic sensors (see Table S1). A new class of ionoelastomers have been developed in which elastomeric polyions are paired with mobile counterions, making them suitable as ionic conductors.37–42 Only a few works demonstrate ionic diodes with such ionoelastomers.40 The intrinsic stretchability of an ionoelastomer was obtained of approximately 150%, giving rise to a deformable diode with its stretchability of approximately 50%. Furthermore, a logic gate based on the combination of the stretchable ionoelastomer diodes is yet to be demonstrated.
Here, we present an intrinsically stretchable ionoelastomer junction logic gate synchronously deformable with liquid metal. A bi-layered ionoelastomer junction is developed consisting of an elastomeric polyanion with mobile cations and a polycation with mobile anions. 2-hydroxyethyl acrylate (HEA) is added to both ionoelastomers to substantially enhance the stretchability of the bi-layered ionic conductor and to promote the synchronous deformation of the ionoelastomer and liquid metal, a eutectic gallium indium (EGaIn) alloy consisting of 75% gallium and 25% indium. An ionoelastomer junction diode of a top EGaIn/bi-layered ionic conductor and bottom EGaIn exhibits a wide electrochemical window of approximately ±3 V, giving rise to excellent non-Faradaic ionic current rectification, with a rectification ratio of approximately 22 when applying a positive–negative voltage sweep. The rectification is attributed to the voltage-dependent modulation of an ionic double layer at the junction between the ionoelastomers, reminiscent of a depletion layer at a semiconducting p–n junction. The ionic diode is successfully performed upon reversible deformation of the device up to a uniaxial stretch ratio of 6 (strain of 500%), without detrimental de-lamination of EGaIn and the bi-layered ionic conductors. Thus, we demonstrate a highly deformable ionic OR logic gate with its stretchability of 300%, combined with stencil-printed EGaIn electrodes.
II. RESULTS AND DISCUSSION
A. Synthesis and characterization of highly stretchable PIE and NIE
To fabricate a stretchable ionoelastomer diode that is synchronously deformable with liquid-metallic EGaIn (SIDS-LM), a polyanion with mobile cations, denoted as p-type ionoelastomer (PIE), was stacked on a polycation with mobile anions, denoted as n-type ionoelastomer (NIE), followed by stencil-printing EGaIn on both sides of the heterojunction (PIE-NIE) to serve as the top and bottom electrodes, as shown in Fig. 1(a). In a stretchable and liquid-free PIE/NIE heterojunction prepared by physically attaching two preformed ionoelastomers (Fig. S1), the mobile counterions (both cations and anions) nearby the heterojunction interface diffuse into the opposite domain driven by entropy. The excess fixed polyanions and polycations on the backbones of the PIE and NIE sides, respectively, remain, as shown in Fig. 1(b). Because of the restricted long-range motion of the fixed ions in the crosslinked polymer networks, an interfacial electric field from NIE to PIE develops, giving rise to a drift current of the mobile ions that exactly compensates for their diffusion current at equilibrium, as shown in Fig. 1(b) (Fig. S2).
Imidazolium-based polymeric ionic liquid monomers 1-ethyl-3-methyl imidazolium (3-sulfopropyl) acrylate [EMIM][SPA] and 1-[2-acryloyloxyethyl]-3-butylimidazolium bis(trifluoro-methane) sulfonamide [AEBI][TFSI] were synthesized for PIE and NIE, respectively (Fig. S3). Poly (ethylene glycol) diacrylate (PEGDA) (Mw: 6000 g mol−1) was also synthesized according to a previously reported procedure.43 1H NMR spectra verified the successful synthesis of [EMIM][SPA], [AEBI][TFSI], and PEGDA (Fig. S4). The polymeric ionic liquid monomers were polymerized with appropriate amounts of PEGDA and 2-hydroxyethyl acrylate (2-HEA), followed by irradiation with ultraviolet (UV) light (wavelength of 375 nm) for 2 min under N2 atmosphere, giving rise to highly stretchable, liquid-free PIE and NIE ionoelastomers. The detailed synthetic procedures for the ionoelastomers are described in the experimental sections. In the co-polymerized chain network [Fig. 1(c)], the flexible crosslinker PEGDA plasticizes the polymeric ionic liquid segment, decreasing both the crystallinity and the glass transition temperature (Tg). The addition of PEGDA also improved the ionic conductivity by facilitating polymer segmental motion.44,45
Dynamic mechanical analysis (DMA) allowed us to examine the temperature-dependent dynamic storage (G′) and loss (G″) moduli of PIE and NIE, as shown in Fig. 1(d). From the local maxima of tan δ, the Tg values of approximately −11.47 and −10.84 °C were obtained for PIE and NIE, respectively. The networked structures of liquid-free PIE and NIE were also examined using thermogravimetric analysis (TGA), and the results in Fig. 1(e) show that both polymers are nearly liquid-free with a small amount of residual water (∼3 wt. %) in PIE owing to its hygroscopic property, whereas NIE is hydrophobic because of the fluorinated TFSI anion. The successful synthesis of PIE and NIE was also confirmed by differential scanning calorimetry (DSC), Fourier-transform infrared (FT-IR) spectroscopy, and x-ray diffraction (XRD) (Fig. S5). The decreased crystallinity arising from PEGDA was observed in the XRD patterns, in which the broad peaks at 2θ = ∼20° for PIE and NIE were substantially reduced compared to those without PEGDA. The surface morphologies and elemental distribution of the ionoelastomers were explored using a scanning electron microscope (SEM) equipped with an energy-dispersive x-ray spectrometer (EDX) (Fig. S6).
The mechanical properties of both the PIE and NIE ionoelastomers were investigated in terms of their tensile stress–strain behavior, and the results are shown in Fig. 1(f). The ionoelastomers of PIE and NIE are highly stretchable, with strains at break of approximately 570% and 650%, respectively. The stretchability of the ionoelastomers is one of the largest among liquid-free polymeric ionic conductors and was largely attributed to the presence of 2-HEA in the ionoelastomers. By adding 2-HEA into PEGDA as a polymer backbone [Fig. 1(c)], the maximum stretchability of PIE and NIE increased from approximately 3.23 to 5.70, and from 3.39 to 6.50, respectively (Fig. S7). The abundant hydroxyl groups in the polymer backbone network give rise to the high stretchability in the ionoelastomer.23 Notably, the presence of 2-HEA in the ionoelastomers also rendered them synchronously deformable with the EGaIn electrodes upon reversible stretching, as shown later. The Young's moduli of PIE and NIE obtained from the slopes of the stress–strain curves were approximately 210 and 170 kPa, respectively, consistent with those obtained from DMA [Fig. 1(d)].
To ensure that our mechanically robust ionoelastomers are sufficiently conductive and suitable for an ionic junction diode, alternating current impedance measurements were performed, as shown in Fig. 1(g). Nyquist plots were used to calculate the ionic conductivity (σ) values of PIE and NIE and were found to be approximately 5.35 × 10−4 and 3.85 × 10−3 mS·cm−1, respectively, using the equation σ = t/ARB, where t is the thickness of the ionoelastomers, A is the effective surface area of the electrode, and RB is the bulk resistance of the ionoelastomers. The ionic conductivity of NIE seldomly changed under ambient conditions (25 °C and 35% RH) for 72 h, whereas PIE, containing a small amount of water, increased the ionic conductivity by 14 times that of the initial value after 72 h of exposure (Fig. S8). It should be noted that the conductivities of our ionoelastomers are comparable with those of liquid-free ionoelastomers,40,41 and sufficiently high to perform ionic rectification as shown later.
B. Mechano-electrical properties of the SIDS-LM
The mechano-electrical properties of the SIDS-LM (EGaIn/PIE-NIE heterojunction/EGaIn) were examined, and the results are shown in Fig. 2. In general, the rigid metallic electrode-ionic conductor contact limits the intrinsic high stretchability of the ionic conductor-based device due to (i) the limited stretchability of the rigid electrodes and (ii) the poorly controlled metal-ionic conductor interface.28–39 To resolve these issues, the intrinsically stretchable EGaIn electrodes were directly screen-printed on the ionoelastomers with excellent wettability of EGaIn on both PIE and NIE, as shown in Fig. 2(a). The excellent wettability of EGaIn on ionoelastomers was confirmed by contact angle measurements.46 An EGaIn droplet on either PIE or NIE exhibited an initial contact angle of approximately 120° spread over time (Fig. S9). After 12 h, contact angles of 47° and 57° were observed for PIE and NIE, respectively, which suggests the development of an attractive interface between EGaIn and the ionoelastomer. The measured contact angles were used to obtain the work of adhesion between EGaIn and the ionoelastomer, based on the Young–Dupré equation; that is, Wad = γlv(1 + cos θ), where Wad is the work of adhesion, γlv is the surface tension of EGaIn (624 mN·m−1),47 and θ is the contact angle on the ionoelastomer surface.48 The Wad values of EGaIn on PIE and NIE, which were initially approximately 310 mJ·m−2, increased for both after 12 h to approximately 1100 and 980 mJ·m−2, respectively. Interestingly, Wad increased by increasing the relative number of hydroxyl groups in 2-HEA, which suggests the crucial influence of 2-HEA on the wettability of EGaIn on the ionoelastomers (Fig. S10).
When the SIDS-LM was stretched, both the top and bottom EGaIn electrodes on PIE and NIE were synchronously deformed with the ionoelastomers, as shown in the photographs in Fig. 2(a). When the stress was released, both the ionoelastomers and EGaIn electrodes synchronously recovered to their original shapes. We quantitatively compared the deformation area ratio [(final area–initial area)/initial area] of the EGaIn electrode to that of the ionoelastomer, and the corresponding results in Fig. 2(b) show the linear variation of the ratio for EGaIn with that for either PIE or NIE upon stretching up to 500% with slopes of approximately 0.91 and 0.93, respectively. These results confirm the synchronous deformation of EGaIn on the ionoelastomers. The deformation length ratio of the EGaIn electrode compared to that of the ionoelastomer was also linear, which is consistent with the results of the deformation area ratio (Fig. S11). The stress–strain curve of our SIDS-LM shown in Fig. 2(c) exhibited a maximum strain of approximately 600% at break with a Young's modulus of approximately 180 kPa. The reversible stretchability of our SIDS-LM was also confirmed by cyclic tensile tests, and no significant degradation of the mechanical properties was observed during the mechanical cycles, as shown in Figs. 2(d) and 2(e) (Fig. S12).
The synchronous deformation of EGaIn on the ionoelastomers was mainly attributed to the attraction between the abundant hydroxyl groups in the ionoelastomers and the few-nanometer-thick Ga2O3 layer on EGaIn, which spontaneously developed under ambient conditions. In our previous work, we discovered that the numerous –OH groups in 2-HEA were responsible for the surface reconciliation of EGaIn on a 2-HEA containing hydrogel upon reversible mechanical deformation.23 Similarly, because of the attraction between Ga2O3 and the –OH groups in our ionoelastomers containing 2-HEA, the EGaIn with the native oxide layer autonomously reconstructs its surface in accordance with the deforming ionoelastomer during strain, as schematically illustrated in Fig. 2(f). The presence of an oxide layer of EGaIn on the ionoelastomer surface was confirmed using depth-profile x-ray photoelectron spectroscopy (XPS) analysis (Fig. S13). The Ga 3d spectra were used to characterize the surface oxidation state of EGaIn.23 By exposing the EGaIn surface to air [Air/EGaIn (A/E) interface], a Ga2O3 layer is formed corresponding to the Ga(III) peak at a binding energy of 20.5 eV. An EGaIn-on-ionoelastomer sample was etched down to examine the EGaIn/ionoelastomer interface, and the results showed that the Ga(III) peak at 20.5 eV was evident at the interface.
The relative change in the electrical resistance of EGaIn (ΔR/R0) on either PIE or NIE was examined as a function of applied tensile strain up to 500%, and the results are shown in Fig. 2(g). Upon stretching, the EGaIn electrodes on both PIE and NIE exhibited a slight increase in resistance, although the ΔR/R0 variation as a function of strain was significantly lower than that theoretically predicted using Pouillet's law.4,49,50 An EGaIn electrode with an oxide layer exhibited large surface roughness with sufficiently increased viscosity.51–53 We believe that the strong deviation from Pouillet's law may be related to the large surface roughness of the electrode with substantially high viscosity as confirmed with 3D confocal microscope analysis (Fig. S14). It should be noted that there are some previous works dealing with a similar deviation from Pouillet's law.49,50 Considering that Pouillet's law is suitable for the electrical resistance of an elastomer, the viscoelastic properties of liquid metal may be also responsible for the deviation behavior. We should, however, admit that further study should be needed to fully understand the resistance behavior of our EGaIn electrode with the strain much lower than that predicted from the theory. Furthermore, the small variation in ΔR/R0 of the EGaIn electrodes on both PIE and NIE was consistent with the consecutive stretching at 500% and releasing cycles of 50 times [Figs. 2(h) and S15]. The resistances were slightly increased with cycles but reached an equilibrium state,54 which suggests that our SIDS-LM is suitable for a deformable ionotronic device, as shown next.
C. Non-Faradaic rectification by SIDS-LMs
The ionic properties of the various PIE and NIE junctions with EGaIn electrodes were investigated, and the results are shown in Fig. 3. First, the electrochemical windows of PIE and NIE with EGaIn electrodes were measured by cyclic voltammetry (CV). The results show that both ionoelastomers have wide electrochemical windows of approximately ±3 V (Fig. S16), allowing sufficient room for the DC operation voltage (±3 V) for non-Faradaic ionic rectification of the SIDS-LM. We examined three junctions of homo PIE/PIE and NIE/NIE, and hetero PIE/NIE by alternating current impedance measurements. From the Nyquist and Bode plots in Figs. 3(a) and 3(b), respectively, we confirmed that the impedance values of the PIE/NIE heterojunction lie between that of the two homojunctions. The frequency-dependent phase angle results in Fig. 3(c) were interpreted using the equivalent circuit model shown in the inset, with fitted parameters such as the metal-ionoelastomer contact resistance (RC), bulk resistance (RB) of the ionoelastomer, and bulk polarization capacitance (CB). A constant phase element (CPE) in the equivalent circuit was used to describe the electric double layer (EDL) for the homojunctions or the ionic double layer (IDL) for the PIE/NIE heterojunction.40 The effective equivalent EDL capacitance (CEDL) values for the PIE/PIE and NIE/NIE homojunctions were 21 and 16 μF·cm−2, respectively, when estimated from the characteristics of CPE used in the equivalent circuit model; that is, the Brug model:55 CEDL= Qα1/α RB[(1/α)-1] where Qα is a constant and α is a constant ranging from 0 to 1. In contrast, the capacitance of the PIE/NIE heterojunction was approximately 0.9 μF·cm−2, much lower than those of the homojunctions, which confirms the presence of the low capacitance planar ionic double layer (CIDL) at the heterojunction interface (Fig. S17).
The PIE/NIE heterojunction under forward bias voltage from +0.4 to +1.0 V exhibited a characteristic exponential decay of current density with time, as shown by the red curves in Fig. 3(d). However, a reverse bias voltage imposed on the junction from −0.4 to −1.0 V resulted in the current density decaying much more rapidly than that under the forward bias, as shown by the blue curves in Fig. 3(d). Notably, because our heterojunction under both the forward and reverse bias voltages was non-Faradaic at the electrode interface without a redox reaction, the ionic conductivity of the junction was lower than that of a conventional Faradaic electrolyte system. In addition, the charge transfer resistance at the electrode–electrolyte interface (RC) was seldomly considered owing to the small resistance of the EGaIn electrode. The total charges (Q) associated with both the forward and reverse bias voltages for 20 s were calculated by integrating the corresponding curves in Fig. 3(d), and the calculated Q values were plotted as a function of the voltage, as shown in Fig. 3(e). An asymmetric behavior of the Q values with the applied voltage was apparent, reminiscent of that observed in a conventional semiconducting p-n junction, which suggests that an ionically rectifying diode was successfully developed with our EGaIn/PIE-NIE heterojunction/EGaIn device. For example, the Q value with +0.4 V (5.94 ± 0.55 μC·cm−2) was approximately 20 times greater than that with −0.4 V (0.28 ± 0.03 μC·cm−2). The slope of the curve corresponded to the capacitance of the heterojunction using the relationship of C = Q/V. Two distinct slopes were obtained; that is, approximately 1.2 μC·cm−2 with the voltage range from −1.0 to +0.2 V and 26.8 μC·cm−2 from +0.2 to +1.0 V. The small capacitance was comparable to the IDL capacitance of the heterojunction obtained from the AC-impedance measurements, and the large one to the EDL capacitance.
Under reverse bias, mobile ions drifted away from the PIE/NIE junction interface, charging the IDL capacitor. In this situation, the resistor-capacitor time (RC), the time constant of an RC circuit, diminished with a small Q value owing to the low IDL capacitance. However, under forward bias, mobile ions drifted from PIE to NIE, and vice versa, toward the PIE/NIE junction interface. Once forward bias was applied above the built-in potential, the IDL collapsed, and the interface reacted as a resistor. Under this condition, the PIE/NIE circuit was effectively dominated by the high capacitance of EDLs at metal–ionoelastomer interfaces, resulting in a long RC and high Q by one or two orders of magnitude (Fig. S18). For comparison, the current density behaviors of the homo PIE/PIE and NIE/NIE junctions were examined under the same bias condition (Fig. S19), and the corresponding Q-V curves of the homojunctions were obtained, as shown in Fig. 3(e). As expected, both the PIE/PIE and NIE/NIE homojunctions exhibited linear behaviors under both forward and reverse biases with single slopes, which corresponded to the capacitances of the homojunctions. It should be noted that compared to a parallel-type ionic diode with in-plane metal electrodes, our vertically stacked diode owns its benefits of (i) high current density arising from large contact interface area and short channel length and (ii) facile and high-density integration of diodes.
The rectification ratio defined by Qforward/Qreverse was obtained as a function of the applied forward bias, as shown in Fig. 3(f). The maximum rectification ratio was approximately 22 at ±0.4 V, which is comparable to the rectification ratios in the range 2–40, which was obtained from the ratios of the forward and reverse current values in previous Faradaic ionic diode systems.7–17 The time-dependent rectification performance of our SIDS-LM was also examined under ambient conditions (Fig. S20).56
A reliable and repetitive current rectification behavior was observed in our PIE/NIE heterojunction with EGaIn under an alternating potential of ±0.60 V at 0.05 Hz, as shown in Fig. 3(g). The ionic diode characteristics of our PIE/NIE heterojunction were further confirmed by AC-impedance measurements of the junction as a function of DC bias, and the results are shown in Figs. 3(h) and 3(i). The Nyquist plots in Fig. 3(h) changed considerably upon increasing the forward bias, mainly because of the collapse of the IDL response to the applied bias, as described earlier, whereas those under reverse bias, governed mainly by IDL, rarely changed. To further analyze the IDL response under the forward bias, we introduced an equivalent circuit model under DC bias containing additional components that represented the interfacial capacitance (CPEIDL) and interfacial resistance (RI) (Fig. S21). While increasing forward bias, RI rapidly decreased because of mobile ions that accumulated at the PIE/NIE interface, followed by the destruction of the IDL. The decrease in the RI substantially decreased the impedance at a low-frequency region (<100 Hz), as evidenced by the prominent decrease in the low-frequency peak of the phase angle [Fig. 3(i)], corresponding to the IDL capacitance.
D. Ionic rectification performance of SIDS-LM under repetitive stretching
A highly stretchable ionoelastomer junction diode synchronously deformable with EGaIn electrodes was developed, and its ionic properties were characterized, as shown in Fig. 4. Our SIDS-LM was repeatedly stretched up to 500% of its initial length with a uniaxial stretch ratio (λ) of 6, as schematically illustrated in Fig. 4(a). The photographs in Fig. 4(b) show a SIDS-LM with a λ of up to 6 without significant changes in the electrical resistance [Fig. 2(g)]. Impedance spectroscopy measurements were performed with our SIDS-LM as a function of λ, and Nyquist and Bode plots were obtained of the device upon stretching, as shown in Figs. 4(c) and 4(d), respectively (Fig. S22). The bulk resistance (RB, plateau region) of the SIDS-LM, located between the frequency range from 101 and 105, decreased as the device stretched. The RB and IDL capacitance (CIDL) of our SIDS-LM were obtained by curve fitting as a function of λ, and the results are shown in Fig. 4(e). In an elastomer with a Poisson ratio of 0.5, the RB of the ionoelastomer should theoretically decrease by 1/λ upon a uniaxial stretching ratio of λ, because the surface area of the ionoelastomer increases by a factor of λ1/2 upon contacting an electrode, whereas the thickness of the ionoelastomer decreases by a factor of 1/λ1/2. However, the CIDL of an ionoelastomer should theoretically increase by λ1/2 upon a uniaxial stretching of λ because it is mainly affected by an increase in the surface area.29,40 In our ionoelastomer junction of PIE-NIE, the variation of RB with the stretching ratio agrees well with the theoretical prediction, whereas the behavior of CIDL deviated slightly from the theory, as shown in Fig. 4(e). The results confirm that our ionoelastomer junction is close to an ideal elastomer, and it deforms synchronously with the EGaIn electrodes upon stretching.
The current developed in our SIDS-LM was carefully measured with time as a function of λ in the given forward (+0.6 V) and reverse (−0.6 V) biases, and the results are shown in Figs. 4(f) and 4(g), respectively. In both the forward and reverse bias, the current of the device increased with time, mainly because of the decreased impedance of the SIDS-LM at the low-frequency range (<10−1 Hz). The same tendency of the current to increase with λ in both the forward and reverse bias in our SIDS-LM is technologically beneficial, resulting in the strain-insensitive ionic current rectification in which the rectification ratio remained almost constant with λ, as shown in Fig. 4(h). Notably, strain-insensitive current rectification is a unique feature of the vertically stacked SIDS-LM.
E. Stretchable ionic logic gate (OR) with SIDS-LMs
A highly stretchable ionic OR logic gate was developed by employing our SIDS-LM, combined with stencil-printed EGaIn interconnectors, and its performance is shown in Fig. 5. The ionic OR logic gate consisted of two vertically stacked ionic diodes of EGaIn/PIE-NIE heterojunction/EGaIn and a resistor (RNIE) interconnected with each other, as schematically shown in Fig. 5(a). The fabricated ionic OR logic gate in Fig. 5(b) shows that EGaIn electrodes with a 5 × 5 and 1.5 × 1.5 mm2 area were used for the ionic diodes and the resistor, respectively, with an EGaIn interconnector of 1 mm in width. The equivalent circuit of the ionic OR logic gate with two inputs, A and B, and an output C terminal is shown in Fig. 5(c). In our OR logic gate, an ionic diode is considered to be conductive when the voltage at either input terminal A or B is higher than the threshold voltage. Therefore, the voltage at the output terminal C should be high and assigned as “1” in the logical state. Voltage signals were programmed with the shape of square waves at given frequencies of fA and fB for inputs A and B, respectively. Four logic input stages by the combination of the two A and B inputs were achieved of (0,0), (0,1), (1,0), and (1,1), by varying the driving voltages for A and B. By optimizing the rectification performance of our SIDS-LM, we used 2 V as the input voltage. The logical “1” state was defined when the voltage in output C was higher than a high threshold voltage of 0.5 V, and the logical “0” state was defined when the voltage was lower than a low threshold voltage of 0.2 V.
First, we examined a single input logic operation with a periodic square-shaped voltage input in A and no voltage input in B, and the results are shown in Fig. 5(d). The square waves with a voltage amplitude of ±2 V at a frequency of 0.1 Hz were applied to the ionic OR logic gate. As expected, two logic input states of (1,0) and (0,0) were produced when the input voltages in A were +2 and –2 V, respectively, as shown in Fig. 5(d). The voltage measured at output C was higher than the threshold voltage of 0.5 V with an input voltage of 2 V, and lower than 0.2 V (the low threshold voltage) with an input of –2 V. By subsequently employing additional input voltage signals in input B, we were able to realize full binary logic input states of (0,0), (0,1), (1,0), and (1,1), as shown in Fig. 5(e). In addition to the square-shaped voltage wave applied in input A, with an amplitude of ±2 V at a frequency of 0.1 Hz, another square voltage wave with the same amplitude and frequency of 0.05 Hz was applied in input B. When +2 V was imposed on both inputs A and B (1,1), the highest output C voltage was measured, which was higher than 1 V, as shown in Fig. 5(e). Considering that the measured voltage was higher than 0.5 V, the logic state was assigned as “1.” With the input combination of (0,1), the output voltage measured was higher than 0.5 V, and thus, the logic state was again assigned as “1.” The other two combinations of (1,0) and (0,0) in inputs A and B successfully developed the output voltage values in C; that is, one voltage higher than 0.5 V and the other lower than 0.2 V in our logic OR gate, giving rise to two output logic states of “1” and “0,” as shown in Fig. 5(e). The results show that our ionic OR logic gate with ionoelastomer diodes is suitable for implementing the logical expression of C = A OR B.
The synchronous deformation of the EGaIn electrodes on the ionoelastomers allowed us to develop a highly stretchable ionic OR logic gate, and the results are shown in Figs. 5(f)–5(h). An ionic logic OR gate was reversibly stretched with λ up to 4 without any noticeable degradation and delamination of the EGaIn electrodes, as shown in the photographs in Fig. 5(f). The gating performance of our stretchable ionic OR logic gate was examined with the voltage inputs A and B used in Fig. 5(e) as a function of λ, as shown in Fig. 5(g). For the 4 input logic states of (0,0), (0,1), (1,0), and (1,1), the corresponding output states of “1,” “1,” “1,” and “0,” respectively, developed successfully under all λ values examined up to 4. The output C voltage (Vc) plots measured as a function of λ in Fig. 5(h) show that the output voltages in three “1” output logic states slightly increased with λ. However, the Vc for all three output “1” states was notably higher than 0.5 V, and that for the output “0” state was lower than 0.2 V, regardless of λ, which suggests that our stretchable ionic OR logic gate successfully implements the logical expression of C = A OR B under large mechanical deformation. We believe that other deformable ionic logic gates can be realized by using the ionoelastomer, which will significantly contribute to future soft deformable ionotronics (Fig. S23).
III. CONCLUSIONS
In this study, we demonstrated a highly stretchable and liquid-free ionoelastomer heterojunction diode synchronously deformable with liquid metal electrodes. Ionic heterojunctions of p- and n-type ionoelastomers with mobile cations and anions, respectively, were developed by vertically stacking the two films, which exhibited a maximum stretchability of 600% with negligible alteration in its ionic conductivity under strain. High stretchability was successfully achieved in a vertical-type ionic diode consisting of a heterojunction ionoelastomer inserted between the top and bottom stencil-printed EGaIn electrodes due to the synchronous deformation of the EGaIn electrode on the ionoelastomer. The systematic study revealed non-Faradaic ionic current rectification reliably occurred in our ionoelastomer diode with a rectification ratio greater than 22 when it was reversibly stretched with a uniaxial strain of up to 500%. Moreover, the stretchable ionoelastomer diodes synchronously deform with EGaIn electrodes and thereby allowed us to develop a highly stretchable ionic logic OR gate with two input diodes and an output. The stretchable logic gate exhibited four binary input logic states successfully decoded into the corresponding two output logic states even when it was deformed with a stretch ratio of up to 4. Developing an all-stretchable ionoelastomer diode with a metal electrode brings us one step closer to emerging ionotronics with high environmental and mechanical stability that facilitate the human–machine interface.
IV. EXPERIMENTAL METHODS
A. Materials
Liquid metal (EGaIn, 75.5 wt. % Ga and 24.5 wt. % In, Sigma-Aldrich, 495425), 1-ethyl-3-methylimidazolium chloride ([EMIM]Cl, Sigma-Aldrich, 900771), 3-sulfopropyl acrylate potassium salt (K[SPA], Sigma-Aldrich, 251631), methoxyphenol (Sigma-Aldrich, 54050), 1-butylimidazole (Sigma-Aldrich, 348414), bis(trifluoromethane)sulfonimide lithium salt (Li[TFSI], Sigma-Aldrich, 544094), 2-hydroxyethyl acrylate (2-HEA, Sigma-Aldrich, 292818), 2-hydroxy-2-methylpropiophenone (HOMPP, Sigma-Aldrich, 405655), poly(ethylene glycol) (PEG, Sigma-Aldrich, 81260), triethylamine (Sigma-Aldrich, 471283), acryloyl chloride (Sigma-Aldrich, 549797), potassium carbonate (K2CO3, Sigma-Aldrich, P5833), magnesium sulfate (MgSO4, Sigma-Aldrich, 746452), ethanol, acetonitrile, dichloromethane (DCM), and de-ionized (DI) water were purchased from Sigma-Aldrich. 2-bromoethyl acrylate (L12502–06) was purchased from Alfa Aesar. All other chemicals were purchased from Sigma-Aldrich and were used as received. VHB 4905 and 4910 were purchased from 3M and used as received.
B. Synthesis of [EMIE][SPA] for PIE
[EMIM]Cl (6.3 g), K[SPA] (10.0 g), and methoxyphenol (10 mg) (inhibitor) were mixed in acetonitrile (30 ml) and stirred overnight at room temperature. The precipitated potassium chloride was filtered, and the solvent was removed using a rotary evaporator. The crude solution was dried at 10−1 Torr for complete evaporation of acetonitrile. The residual liquid was re-dissolved in DCM, and the solution was stored at < 0 °C overnight. The precipitated salts were filtered. After complete removal of DCM, the viscous yellow oil product [EMIM][SPA] was obtained.
C. Synthesis of [AEBI][TFSI] for NIE
Under N2 atmosphere, 2-bromoethyl acrylate (5.0 g) and 1-butylimidazole (3.64 g) were mixed in acetonitrile (30 ml) and stirred overnight at 60 °C. After the reaction, acetonitrile was removed using a rotary evaporator. 1-[2-acryloyloxyethyl]-3-butylimidazolium bromide ([AEBI]Br) was extracted using DI water and washed with DCM at least three times. Then, Li[TFSI] (8.0 g) was added to the aqueous solution of [AEBI]Br and stirred overnight at room temperature. After ion exchange, water immiscible [AEBI][TFSI] was extracted with DCM and washed with DI water at least three times. The residual water in the organic layer was dried over MgSO4. Drying of the final product [AEBI][TFSI] under vacuum <10−1 Torr resulted in a slightly yellow transparent liquid.
D. Synthesis of PEGDA (Mw: 6000 g·mol−1)
PEGDA was synthesized according to a previously published procedure.43 Under an Ar atmosphere, PEG (24 g) was dissolved in DCM (72 ml) while stirring in a three-neck round-bottom flask. After 10 min, triethylamine (0.68 ml) and acryloyl chloride (0.76 ml) were added to the PEG solution with a glass syringe, and the mixture vigorously stirred overnight at room temperature under light protection. In a fume hood, the reacted mixture in the round-bottom flask was transferred to a separatory funnel and 1.5 m K2CO3 aqueous solution (0.6l) added for phase separation, after which the mixture was kept overnight under light protection. The lowest organic phase was drained into a beaker, anhydrous MgSO4 added, and the mixture stirred until well dispersed. After filtering MgSO4, the DCM solvent in the mixture was removed to obtain a viscous solution using a rotary evaporator. The mixture was then poured into diethyl ether under stirring. The precipitated PEGDA was filtered and dried overnight in a crystallization dish.
E. Preparation of stretchable ionoelastomer diode
A schematic illustration of the preparation of the ionoelastomer junction is shown in Fig. S1. We vigorously mixed the polymeric ionic liquid monomer (0.15 g [EMIM][SPA] for PIE, 0.25 g [AEBI][TFSI] for NIE) with 1 mol. % PEGDA. Next, 2-HEA (0.02 g) and 0.5 mol. % of the HOMPP photo-initiator were dissolved in the mixed liquid to form a precursor solution. The viscous solution was then injected into a VHB film spacer (thickness = 250 μm), and the solution sealed with a PET cover. Free radical polymerization was initiated under UV light (wavelength = 375 nm) under N2 atmosphere for 2 min. After the reaction, the unreacted monomers were extracted by washing with DCM. Then, PIE (or NIE) was dried at 60 °C to remove residual solvents and water before use. After drying, free-standing ionoelastomers of PIE (or NIE) were carefully peeled off from the VHB substrate, and bilayer junctions were then prepared by attaching the two surfaces of the ionoelastomers. For direct patterning of the EGaIn on the bilayer ionoelastomer junctions, the hydrophobic patterned mask was placed on the ionoelastomer and EGaIn (0.1 mg·mm−2) was dropped, and then stencil-printed.
F. Characterization of ionoelastomers
TGA of the ionoelastomers was conducted using a thermogravimetric analyzer (TA Instruments Q500). The temperature was increased from 25 to 700 °C at a rate of 10 °C·min−1 under N2 atmosphere. 1H-NMR (Bruker BioSpin, Avance lll HD 400) spectra also confirmed the structures of the ionoelastomers. DSC measurements were performed using a differential scanning calorimeter (TA Instruments Q200) in the air. The temperature was scanned from −50 to 50 °C at a heating rate of 10 °C·min−1. A peak spectrum scan was performed using an attenuated total reflectance (ATR)-FTIR spectrometer (Vertex 70, Bruker). XRD measurements were conducted using a diffractometer (SmartLab, Rigaku) with Cu Kα radiation in the 2θ range of 10°–80° to identify the crystallinity. A field-emission SEM (JEOL-7610F-Plus) equipped with energy-dispersive spectroscopy (EDS) capabilities was employed to inspect the morphologies and compositions of the ionoelastomers.
G. Mechanical properties of ionoelastomers
The rheological properties of the ionoelastomers were measured using a DMA (Q800, TA instruments) with uniaxial tensile geometry. The ionoelastomers were cut into dimensions of 40 mm in length, 6 mm in width, and 1 mm in depth. The temperature was increased from −60 to 100 °C at a rate of 5 °C·min−1 with 1 Hz oscillatory stress. The mechanical properties of the ionoelastomers were determined using a uniaxial tensile machine (Mecmeshine, 50 N load cell) with a stretching rate of 10 mm min−1 for the cycling test and 100 mm min−1 for maximum stretchability.
H. Electrochemical properties of ionoelastomers
For the electrochemical analysis, the AC-impedance spectra were determined using a Bio-logic VPM-300 electrochemical workstation in the frequency range of 1 MHz to 0.1 Hz and an amplitude of 20 mV. The positive terminal was connected to the PIE and the negative terminal to the NIE. CV curves with a constant voltage ramp and AC were measured using an electrochemical workstation (Bio-logic VPM-300).
I. Characterization of EGaIn
Depth-profile XPS (Thermo UK, K-alpha) measurements were performed to determine the oxidation of Ga. A monochromated aluminum x-ray source (Al Kα line: 1486.6 eV) was used. Scanning was performed over a 400 μm diameter. All spectra were calibrated using the C 1s peak (binding energy = 285 eV). The resistance of the EGaIn electrode on the ionoelastomer was measured using the four-point probe method (Advanced Instrument Technology, CMT-100MP).
J. Measurement method of stretchable ionic diode and ionic OR logic gate
To measure the current of the stretchable ionic diode, a parameter analyzer (4200A, Keithley) was used under various DC biases. To verify the ionic OR logic gates, four probes were prepared using a parameter analyzer. Two of them were connected with the input circuit to generate square-wave voltage signals as binary inputs for OR logic gates. The third channel was connected to the output circuit to measure the voltage signals and current at the output terminal of the logic gates. The fourth channel was grounded. The stretching test was performed using a stretching machine during the measurement process.
SUPPLEMENTARY MATERIAL
See the supplementary material for the results that support the findings of this study.
ACKNOWLEDGMENTS
This study was supported by the Creative Materials Discovery Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (No. 2018M3D1A1058536). This study was also supported by a grant from the National Research Foundation of Korea (NRF) funded by the Korean Government (MEST) (No. 2020R1A2B5B0300269711). This work was supported by the Korea Medical Device Development Fund grant funded by the Korea Government (the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, the Ministry of Health & Welfare, the Ministry of Food and Drug Safety) (Nos 9991006750 and KMDF_PR_20200901_0077), (Nos. 1711138098 and KMDF_PR_20200901_0077). This work was also supported by the Brain Korea 21 FOUR Project funded by the National Research Foundation (NRF) of Korea, Yonsei University College of Nursing (No. F21JB7504007) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (Grant No. HI19C1344). M.D.D. thanks support from the NSF ASSIST Center (No. EEC-1160483).
AUTHOR DECLARATIONS
Conflict of Interest
The authors have no conflicts to disclose.
Author Contributions
Seung Won Lee: Conceptualization (lead); Data curation (lead); Formal analysis (lead); Investigation (lead); Software (lead); Validation (lead); Visualization (lead); Writing – original draft (lead); Writing – review & editing (supporting). Taebin Kim: Data curation (supporting); Methodology (supporting); Validation (supporting); Visualization (supporting). Michael D. Dickey: Methodology (supporting); Resources (supporting); Supervision (supporting); Validation (lead); Writing – review & editing (equal). Cheolmin Park: Conceptualization (lead); Funding acquisition (lead); Methodology (supporting); Supervision (lead); Writing – review & editing (lead). Jihye Jang: Conceptualization (supporting); Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Software (lead); Validation (supporting); Visualization (supporting). Yeonji Kim: Conceptualization (supporting); Data curation (supporting); Investigation (supporting); Methodology (supporting); Validation (supporting); Visualization (supporting). Seokyeong Lee: Data curation (supporting); Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Validation (supporting); Visualization (supporting). Kyuho Lee: Data curation (supporting); Formal analysis (supporting); Methodology (supporting); Software (supporting). Hyowon Han: Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Validation (supporting); Visualization (supporting). Hyeokjung Lee: Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Validation (supporting); Visualization (supporting). Jin Woo Oh: Formal analysis (supporting); Funding acquisition (supporting); Software (supporting); Visualization (supporting). Hoyeon Kim: Formal analysis (supporting); Investigation (supporting); Methodology (supporting); Validation (supporting).
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request.