Optimization of Core–Shell Ternary Electrodes for High‐Performance Ionic Actuator in Soft Gripper

Abstract

Ionic actuators based on composite electrodes consisting of nanomaterials and conducting polymer typically offer the advantages of low‐voltage operation and high stability, however, electrode preparation using conventional mixing suffers from issues of ineffective dispersion of nanomaterials, greatly diminishing their synergistic effects. Here, the ternary electrode system based on SWCNTs/PEDOT: PSS/ionic liquid using the two‐step dispersion process is optimized, achieving a uniformly coated core–shell structure with high conductivity (≈392.4 S cm⁻¹). The ions migration process is analyzed according to the core–shell model, further optimization of the ternary electrode and device structure enables the actuator to realize the peak‐to‐peak strain per volt reaching 1.3% V⁻¹ and normalized blocking force of 0.15 MPa V⁻¹ (≈89.2 times its own weight), with stable performance maintained over 1 million cycles. Therefore, the actuator can be utilized for the assembly of multi‐clawed grippers to grasp precision components or larger objects. Multiple connected actuators fulfill a complex deformation, indicating promising applications in smart grippers, bioinspired robotics, and human–machine interaction.

A core–shell ternary electrode (Opt‐SWCNTs/PEDOT:PSS/IL) is developed via a simple two‐step dispersion and vacuum filtration process, with component ratios optimized to achieve excellent mechanical toughness and electrical conductivity based on simulation results. The resulting actuators demonstrate high strain and blocking force, enabling precise gripping and complex deformation, showing great potential for soft robotics and next‐generation electrochemical actuators.

1. Introduction

Soft actuators have garnered widespread attention as an integral component of next‐generation intelligent robots owing to their commendable dexterity, biocompatibility and adaptability to complex environments.^{[ 1 ]} The actuation methods employ diverse physical mechanisms to induce deformation in soft materials, including pneumatic inflation,^{[ 2 ]} electric‐field,^{[ 3 ]} thermal,^{[ 4 ]} optical,^{[ 5 ]} humidity actuators^{[ 6 ]} and others.^{[ 7 ]} Among these, electrochemical ionic actuators have been extensively developed due to their substantial actuation strain, fast response, lower power consumption and operational stability in the air, which could be hypothesized to result from the migration of cations or anions with different molecular sizes towards the electrodes, causing electrodes expansion and achieve the bending actuation.^{[ 8 ]} Aqueous electrolyte membrane has been used to prepare the ionic actuator at the early stage, but its performance declines rapidly in an open‐air environment due to water evaporation and electrolysis.^{[ 9 ]} Hence, non‐aqueous actuators were prepared by using ionic liquids, which show high conductivity, wide potential window and high stability, effectively meeting the practical applications of large deformation.^{[ 10 ]}

The electrode, a carrier for reversible ion insertion and deintercalation, constitutes the foundation for fabricating ionic actuators, and its performance could be improved by enhancing the ion capacity and migration rate. Common electrode materials could be classified into three categories: firstly, 1D/2D nanomaterials, including carbon materials (such as graphite carbon,^{[ 11 ]} carbon nanotube,^{[ 12 ]} graphene^{[ 13 ]} and graphdiyne^{[ 14 ]}), MXene,^{[ 15 ]} metal^{[ 16 ]} or metal oxide,^{[ 17 ]} etc. These electrode interfaces will form the electrostatic double layer, which plays a crucial role in the preparation of high‐performance actuators. Chen et al. combined carbon nanotube electrodes with Nafion/ionic liquid membranes to prepare high‐stability actuators capable of generating newton‐level output blocking force.^{[ 18 ]} Additionally, the utilization of conducting polymers, such as polypyrrole (PPy),^{[ 19 ]} polyaniline (PANI),^{[ 20 ]} and poly(3,4‐ethylenedioxythiophene)/poly(styrene sulfonate) (PEDOT:PSS),^{[ 21 ]} has been investigated. The electrical field induces changes in the oxidation state of the polymer and subsequently alters its geometric dimensions for actuation.^{[ 22 ]} On this basis, composite electrodes formed by mixing the above two materials were produced, which aimed to integrate the advantages of both.^{[ 23} , ^{24 ]} For instance, a PEDOT:PSS/Ti₃C₂T _x electrode with a layered structure demonstrates a high bending strain of ≈1.37% without any back‐relaxation.^{[ 25 ]} Ordered and active nanochannels (vertically aligned CNT) are formed by a facile electrochemical process, which provides a rapid pathway for ion transportation, leading to a wide frequency response and fast actuation speed.^{[ 26 ]} However, these fabrication processes are intricate and yield electrodes with uncontrollable morphologies. It is evident that the simple blending methods cannot fully harness the advantages of both materials, and novel strategies for the preparation of electrodes with electrochemical stability is necessary.

In this work, we employed a two‐step dispersion process, using Triton X‐100 and dimethylsulfoxide (DMSO) for the pretreatment of SWCNTs and PEDOT:PSS, respectively, and subsequently mixed them with ionic liquid to achieve the optimized core–shell ternary electrode (Opt‐SWCNTs/PEDOT: PSS/IL). Large‐area flexible films could be produced via simple vacuum filtration, offering advantages such as simple fabrication, low cost, and easily large‐scale preparation. The optimal ratios of the components were investigated, leading to the high‐performance actuators with responsiveness under low voltage (<1 V) and high frequency (10 Hz), with stability maintained even after 1 million cycles. Moreover, by adjusting parameters including ionic liquid concentrations, electrolyte‐to‐electrode thickness ratio, and device width, the actuator maintains the peak‐to‐peak strain of 1.3% V⁻¹ while its normalized blocking force is 0.15 MPa V⁻¹ (approximately 89.2 times its own weight at 2 V). Consequently, these actuators were employed to fabricate multi‐clawed grippers to grasp accurately small components or heavier objects, while serially connecting multiple actuators enables more complex deformation. We anticipate the developed devices will have broad applications in the next‐generation electrochemical actuators, smart grippers, and soft robots.

2. Results

2.1. Preparation of Core–Shell Ternary Electrode System

Single‐walled carbon nanotubes (SWCNTs) have received considerable attention in the electrochemical actuators, due to their excellent mechanical strength, high electrical conductivity, and good chemical stability. The aromatic thiophene rings in PEDOT: PSS could form strong π–π stacking interactions with the surface of the carbon nanotubes, making it effective for dispersing the SWCNTs and preventing its aggregation. However, the dispersion of SWCNTs by a certain amount of PEDOT: PSS is limited. To overcome this limitation, we proposed a two‐step dispersion process. SWCNTs are first dispersed using the Triton X‐100, which has long molecular chains that can wrap around the SWCNTs to ensure their separation. Then the dispersion is thoroughly mixed with DMSO‐modified PEDOT:PSS under ultrasonication to yield a uniformly stable dispersion of SWCNTs/PEDOT:PSS. Subsequently, a certain amount of ionic liquid (EMIM:BF₄) is added to the prepared solution for electrodes fabrication (Figure S1, Supporting Information). An optimized ternary electrode (Opt‐SWCNTs/PEDOT:PSS/IL) is fabricated using the vacuum filtration process (Figure  1a). The SEM image in Figure 1b shows the random distribution of fibers in the optimized electrode, with the inserted cross‐sectional image highlighting its layered stacking structure. Compared to traditional film‐casting method which requires stringent control on solvent evaporation to prevent film cracking (Figure S2, Supporting Information), vacuum filtration enables the preparation of dense, layered and flexible free‐standing films, thereby greatly enhancing the actuator performance. Detailed preparation procedures can be found in Section A of Supporting Information. Moreover, Figure 1c illustrates the mechanical properties of electrode, demonstrating its sufficient toughness and suitability for electrochemical actuators.

Figure 1.

Fabrication and structure of the ternary electrode system. a) Schematic illustration of multilayer fiber electrode preparation using vacuum filtration. b) SEM image of the optimized ternary electrode (Opt‐SWCNT/PEDOT: PSS/IL), and the embedded cross‐sectional image illustrates the layered fibrous structure. c) Tensile properties of the aforementioned electrode exhibit its excellent toughness. Inset: optical photograph of the electrode. d) TEM images of SWCNTs and PEDOT: PSS after sequential optimization through the two‐step dispersion process reveal a distinct core–shell structure. e) Raman spectroscopy of the sequential optimization of the electrode. For the optimized ternary electrode, the surface modification of PEDOT: PSS and IL coating on the SWCNTs shows a rightward shift in peak positions, indicating the increased sensitivity to charge transfer.

Figure 1d shows the TEM images of the sequential optimization process of the ternary electrode, the fibers exhibit the smaller diameters, with the surfactant and conductive polymer further enhancing its dispersion. A distinct core–shell structure is formed by the coating of PEDOT:PSS and ionic liquid on the SWCNTs surface. If only PEDOT: PSS is used to disperse the SWCNTs, the dispersion will be poor (More details could be found in Figure S6, Supporting Information). The corresponding Raman spectroscopy of the sequential optimization of the electrode is presented in Figure 1e. When PEDOT: PSS and ionic liquid are mixed with SWCNTs, the G band intensity significantly decreases, indicating the effective coating of the carbon nanotubes. The G band normally has two types of peaks (G⁺ and G⁻), with the G⁺ peak associated with carbon atom vibration along the nanotube axis. Coating PEDOT: PSS and ionic liquid on its surface shifts this peak to the higher frequencies, enhancing the sensitivity of carbon nanotubes to charge transfer and rendering it suitable for the fabrication of electrochemical actuators.

2.2. Analysis of Ion Migration

Figure  2a illustrates the electrical conductivity variation of the electrodes under different fabrication processes. The electrode based on the SWCNTs mixed with DMSO‐modified PEDOT: PSS (without Triton X‐100) exhibits the highest conductivity, approximately 392.4 S cm⁻¹. Using Triton X‐100 reduces the conductivity; however, it could be restored to 356.7 S cm⁻¹ after cleaning with acetone/IPA solution. Further tests on the changes in conductivity and thickness with different PEDOT:PSS and ionic liquid are depicted in Figure 2b,c. Increasing PEDOT:PSS content notably thicken the electrode, while conductivity initially experienced a slight increment before declining. Its morphology changes from fibrous to granular when the content of PEDOT: PSS increases, accompanied by a significant enhancement in Young's modulus and maximum tensile strain, which can be found in Figure S4 (Supporting Information). On the other hand, the influence of the ionic liquid on the electrode properties is minor, with conductivity and thickness remaining constant across different concentrations. EDS data indicate that the B and F elements are uniformly distributed within the electrode, suggesting the successful incorporation of the ionic liquid (Figure S5, Supporting Information). The specific capacitance of electrodes is analyzed in non‐aqueous 1.0 m EMIM: BF₄/acetonitrile electrolyte solutions at a scan rate of 0.01 V s⁻¹ (Figure 2d), which increases successively with the process optimization. For the optimized ternary electrodes prepared by using DMSO, Triton X‐100 and ionic liquid, the specific capacitance value could reach 40.54 F g⁻¹, which is slightly better than the flexible electrodes prepared by conventional casting method.^{[ 24 ]}

Figure 2.

Analysis of ion migration in ternary electrode. a) Sequential optimization of conductivity variation in ternary electrode. Mixing with different b) PEDOT:PSS and c) ionic liquids to investigate variations in electrode conductivity and thickness. d) Specific capacitance of different electrodes in non‐aqueous 1.0 m EMIM:BF₄/acetonitrile electrolyte solutions at a scan rate of 0.01 V s⁻¹. e) Probable mechanism of ion migration in electrodes with various ratios of PEDOT:PSS to SWCNTs. Small amounts of PEDOT:PSS show poorer toughness and susceptibility to breakage, while excessive PEDOT:PSS may impede ion migration into the interior. Moderate content of PEDOT:PSS will fully exploit the synergistic effects, enhancing the performance of actuator. f) Simulation analysis of ion migration. Simplifying the electrode to the core–shell structure with internal SWCNTs and external PEDOT:PSS reveals that thinner film enables rapid ion infiltration, but thicker electrodes predominantly accumulate ions at the boundaries. Ions could traverse the outer layer and migrate to the inner for the moderately mixed PEDOT:PSS, effectively leveraging the electrode structure. g) XPS N 1s spectra of ternary electrode under different voltages, the two groups of N 1s peak (‐1 V) may be originated from the migration of interfacial cations near the SWCNTs and overlaid bulk cations stabilized by the surrounding IL molecules, further substantiating the migration process of ions.

During ions migration to the ternary electrode, simultaneous electrostatic double‐layer (DL) and faradaic capacitors (FC) occur, fully leveraging their synergistic effects is crucial for achieving the high‐performance actuator. Hence, a probable mechanism of ion migration in electrodes with various ratios of PEDOT: PSS to SWCNTs is shown in Figure 2e. In case of low PEDOT: PSS content mixed with SWCNTs, inadequate binding of SWCNTs may lead to insufficient polymeric chain interactions and poor electrode cohesion, rendering the electrode prone to tearing. In contrast, excessive PEDOT: PSS results in a significant increase in electrode thickness and reduced conductivity. As ions migrate to the electrode, the predominance of faradaic capacitance hinders rapid transference to the interior. To further validate the above hypothesis, ion migration simulations were conducted on electrodes with different compositions by the finite element analysis method (COMSOL Multiphysics), and the electrode was simplified to consist of internal SWCNTs encased by external PEDOT:PSS shells. The results reveal that ions rapidly permeate the entire thinner electrode, while they merely diffuse to the boundaries of the thicker electrodes. For the moderate content of PEDOT:PSS, ions migrate into the inner SWCNTs after passing through the outer layer, fully harnessing the electrochemical activity of the electrode structure. These simulation results are consistent with the theoretical analysis.

Building on these findings, XPS analysis was employed to investigate ion migration characteristics under different voltages using traditional three‐electrode setup, with voltages applied for approximately 300 s. Given the presence of ionic liquid (EMIM: BF₄), the N 1s was selected to observe the ions migration, and according to the literature reports, PEDOT: PSS exhibits the following reaction with EMIM: BF₄:^{[ 27 ]}

$$\begin{array}{c}\mathrm{PEDO}{\mathrm{T}}^{n+}:\mathrm{PS}{\mathrm{S}}^{n-}+m\mathrm{EMI}{\mathrm{M}}^{+}+m^{e-}\leftrightarrow \mathrm{PEDO}{\mathrm{T}}^{\left(n-m\right)+}:\hfill \\ \mathrm{PS}{\mathrm{S}}^{\left(n-m\right)-}+m\mathrm{PEDO}{\mathrm{T}}^0+m\left(\mathrm{EMI}{\mathrm{M}}^{+}:\mathrm{PS}{\mathrm{S}}^{-}\right)\hfill \end{array}$$ (1)

$$\begin{array}{c}\mathrm{PEDO}{\mathrm{T}}^{n+}:\mathrm{PS}{\mathrm{S}}^{n-}+m{\mathrm{BF}}_4^{-}\leftrightarrow \mathrm{PEDO}{\mathrm{T}}^{\left(n-m\right)+}:\hfill \\ \mathrm{PS}{\mathrm{S}}^{\left(n-m\right)-}+m\mathrm{PS}{\mathrm{S}}^0+m\left(\mathrm{PEDO}{\mathrm{T}}^{+}:{\mathrm{BF}}_4^{-}\right)+m{e}^{-}\hfill \end{array}$$ (2)

With +1 V applied to the working electrode, the electrode absorbs the anions (BF₄ ⁻) and repels cations (EMIM⁺), so the N 1s peak position of imidazolium does not change significantly. However, the interaction between EMIM⁺ and PSS⁻ occurs when applied voltage at ‐1 V, there are two deconvoluted peaks of N 1s peaks. This can be attributed to the migration of interfacial cations near the SWCNTs, and overlaid bulk cations stabilized by the surrounding IL molecules, further supporting the ions migration under voltage. The peak shifts of N, F, and S elements under various voltages were analyzed in Figure S7 (Supporting Information), confirming the ion insertion into electrode. Additionally, impedance analysis in a three‐electrode system also revealed that the optimized ternary electrode exhibits impedance comparable to pure SWCNTs, demonstrating its high ion accommodation and transport capability (Figure S8, Supporting Information).

2.3. Characterization of Actuator Performance

To validate the performance of the ternary electrode, it was hot‐pressed with Nafion/IL electrolyte to fabricate the actuator (Figure  3a), and the detailed preparation process is referred to Figure S9 (Supporting Information). Ions in the actuator migrate to the opposite electrodes under voltage, resulting in different electrode expansion due to ion size discrepancies, thereby inducing the bending of actuator. The ionic liquid (EMIM:BF₄) employed in this work is characterized by substantial cation‐anion size disparities and rapid migration rates, making it well‐suited to prepare high‐performance actuators. Furthermore, the bending deformation of the actuator could be simplified to a cantilever beam model, whereby the strain could be deduced from their bending curvature. Generally, smaller curvature radius results in larger deformation and greater actuator displacement, and the bending process was simulated via the finite element analysis method. Drawing upon the preceding ion migration simulations, a consideration of ion size was integrated to analyze the actuator bending variations. The results indicate that ions progressively diffuse across the entire electrode with the passage of time or increasing voltage, triggering the actuator bending, details can be found in Figure S10 and Movie S1 (Supporting Information). Figure 3b presents the strain variations of actuators fabricated with different electrodes, which increase steadily through optimization of the fabrication process. The ternary electrode system achieves a strain of 2.45%, representing an approximately 23.6% improvement compared to the SWCNTs/PEDOT: PSS electrode, and additional data analysis is found in Figures S11 and S12 (Supporting Information).

Figure 3.

Actuator performance of ternary electrode. a) Schematic diagram of the actuator structure and its bending model under voltage, with the upper right showing the bent actuator within 25 s at 2 V voltage; the lower right presents the formula for calculating the bending strain based on the cantilever beam model. b) Bending strain of the actuator using different electrodes, which gradually increases with continuous optimization of the fabrication process, reaching a maximum of approximately 2.45%. c) Voltage and d) frequency response of actuators based on the ternary electrode with different PEDOT:PSS and SWCNTs ratios. e) Bending response of actuators with square wave input potentials varying from 0.05 to 10 Hz at ±2 V. f) Voltage and g) frequency response of actuators based on the ternary electrode with different ionic liquid content (the ratio of PEDOT:PSS to SWCNTs is 0.8). h) Stability measurement of the actuator, and the cycle period exceed 1 million cycles. Inset: peak‐to‐peak displacement of actuator before and after cyclic endurance.

As previously discussed, the SWCNTs to PEDOT: PSS ratio significantly influences the performance of the ternary electrode actuator, Therefore, electrodes with varying ratios of SWCNT to PEDOT: PSS (0.2, 0.4, 0.6, 0.8, 1.6 and 3.2) were fabricated (with fixed 5 wt% DMSO), however, the first two ratios resulted in the difficulties in forming a freestanding film due to the lower polymer content, thus the voltage (Figure 3c) and frequency (Figure 3d) response tests were conducted on the latter four ratio groups. The results show that the actuator initially exhibits a linear response to voltage, gradually reaching saturation, likely influenced by ion concentration and its electrochemical window. Voltages exceeding 2 V may induce the decomposition of the ionic liquids, leading to performance tapering off. Meanwhile, the strain gradually decreases with the higher frequency, and the corresponding decline in peak‐to‐peak displacement from 0.05 to 10 Hz is shown in Figure 3e, which could be attributed to the insufficient duration of ion migration to the electrode at higher frequencies. However, both datasets suggest that the actuator achieves maximum strain at SWCNTs to PEDOT:PSS ratio of 0.8, suggesting optimal synergistic effects between the two components. On the basis of the above results, additional ionic liquid was introduced at different ratios relative to SWCNTs (0.2, 0.5, 1.0, 2.0 and 5.0), and their voltage and frequency response are shown in Figure 3f,g. It can be seen that ionic liquid minimally impacts actuator performance, showing enhancement with the addition of small quantities, but excessive ionic liquid leads to a slight decline, possibly due to induced carbon nanotube aggregation. Hence, the optimal ratio of SWCNTs, PEDOT: PSS and ionic liquid in the ternary electrode is estimated to be 1: 0.8: 1, serving as the basis for the subsequent experiments. Detailed analysis could refer to Figures S13 and S14 (Supporting Information), and the voltage and frequency response are shown in Movies S2 and S3 (Supporting Information). Figure 3h shows that the ternary electrode system actuators possess excellent stability, maintaining stable performance after one million cycles. It still maintains stable performance under extreme conditions, including high temperature (≈100 °C), high humidity (90%), and fatigue bending tests (≈10000 cycles), demonstrating its reliability for practical applications (Figure S15, Supporting Information).

2.4. Optimization of Device Structure

Figure  4a shows the design concept of a smart actuator, which typically requires the fulfillment of two essential criteria: substantial strain and blocking force, which often presents a trade‐off relationship. This work aims to seek the optimal balance between these two parameters from the perspective of ionic liquid concentration, device thickness, and width. A higher concentration of ionic liquid implies more ion migration under voltage, leading to greater deformation. Additionally, considering the cantilever beam model, increasing device thickness and width effectively enhances the blocking force, while longer devices suffer from significant energy dissipating along the free length, leading to a decline in blocking force (Figure 4b). To facilitate comparative data analysis, a stimulus voltage of 2 V and a frequency of 0.05 Hz were applied on the actuator. Four parameters were chosen to assess the actuator performance: displacement, thickness, strain and blocking force, and its performance with different ionic liquid concentration (10%, 40%, 70%, 100%) is illustrated in Figure 4c. Mixing the ionic liquid with Nafion forms a two‐phase interaction structure, where higher concentrations increase the conductivity of the phase and ion channels, thereby affecting properties such as thickness, conductivity and Young's modulus (Details can be found in Section E of Supporting Information). Generally, under the same concentration, thinner electrolytes exhibit lower ion transport hindrance, yielding actuators with greater deformation. However, thicker electrolytes tend to possess higher Young's modulus, leading to reduced deformation but a larger blocking force. Therefore, with increasing concentrations of ionic liquid, the displacement of actuators increases, while its thickness gradually declines. This phenomenon could be attributed to the actuator with a higher concentration of electrolyte being softer and more prone to flattening during high‐temperature hot pressing, thereby resulting in the rise in strain and decreased blocking force.

Figure 4.

Analysis of device structure optimization. a) Schematic diagram of the design concept for a smart actuator, aiming to enhance the strain and blocking force by adjusting ion concentration, thickness and area. b) Enhancing actuator blocking force via cantilever beam model analysis, from a geometric perspective, involves increasing device thickness and width while reducing the free length. Four parameters are used to assess the actuator performance, including displacement, thickness, strain and blocking force. Performance variations with different ionic liquid concentrations in c) the electrolyte, ratios of electrolyte thickness to d,e) electrode thickness, f) width, and g) actuator stacking. h) Comparison of strain and normalized blocking force of the developed actuator with literature data.

Further tests were carried out to investigate the impact of the thickness on the actuator performance, which could be divided into the electrolyte (t _electrolyte) and electrode thickness (t _electrode). To maintain the high strain of the actuator, we used Nafion with high ionic liquid concentration as the electrolyte. Initially, the electrode thickness was altered (Figure 4d), revealing that both the overall device thickness and displacement decline with the decreasing electrode thickness, leading to a notable diminish in blocking force. Although there is a slight increase in strain, the blocking force is reduced, which suggests the thinner electrode thickness will reduce the space available for ion accommodation, rapidly deteriorating device performance. On this basis, sufficiently thick electrodes were chosen, and the electrolyte thickness was gradually increased to enhance its performance, as illustrated in Figure 4e. The findings reveal that augmenting the electrolyte thickness notably raises the overall device thickness, accompanied by the decrease in displacement. According to the theoretical calculations, thicker devices were anticipated to generate a continuous rise in blocking force. However, experimental observations showed that these samples are prone to tearing during hot press. Moreover, a large amount of ionic liquid was squeezed out, which could weaken the interfacial adhesion between electrode and electrolyte, causing decreased device performance. Further detailed analysis can be found in Section F of Supporting Information. The relationship between deformation and blocking force did not exhibit a monotonic trend. For example, at a strain of 1.6%, the blocking force was approximately 3 mN, whereas it would decrease to about 2.2 mN at a strain of 2.59%, and the blocking force is approximately 89.2 times its own weight. Hence, there is a mutual constraint between these two factors, suggesting the optimal ratio (t _electrolyte /t _electrode) should be controlled with the range of 5:1 to 10:1.

The width of the device does not significantly affect its displacement or strain; but substantially enhances the blocking force, reaching 6 mN with a width of 5 mm, as shown in Figure 4f. In addition, attaching multiple devices together (simply with adhesive tape) could improve its performance (Figure 4g). Although stacking a device with fewer layers generates a larger blocking force, the improvement was far less compared to the constraining effect arising from the interaction among the different actuators when increasing the number of layers (greater than three layers), leading to a decline in overall performance. The peak‐to‐peak strain and normalized blocking force of the actuator in this work were evaluated in comparison to those reported in the literature (Figure 4h). A trade‐off between large strain and blocking force observed in our devices. Thinner devices are more prone to deformation, with a peak‐to‐peak strain per volt reaches 1.3% V⁻¹ and a normalized blocking force of 0.15 MPa V⁻¹. The blocking force is increased to 0.46 MPa V⁻¹ for thicker devices, but its strain decreases to 0.7% V⁻¹. Nevertheless, the performance of the actuator with the optimized ternary electrode outperforms that of current carbon‐based electrode ionic actuators. In addition, its simplified fabrication process and lower cost further highlight its promising application potential.

2.5. Applications of the Multifunctional Actuators

The potential application of ionic actuators in grasping precision components was investigated by designing a two‐claw gripper for handling microlens, including cuboidal, cylindrical, dodecahedral, and polyhedral shapes, as shown in Figure  5a. The results demonstrate that the gripper can effectively perform gripping, lifting, holding, and stable releasing processes even with a small contact area, suggesting its ability to manipulate small or fragile objects with confined spaces (Movie S4, Supporting Information). The blocking force of the ionic actuator increases linearly with voltage, as shown in Figure S26 (Supporting Information), thus an appropriate driving voltage could be selected to avoid damaging fragile objects. Furthermore, we investigated the collaborative operation of multiple actuators, and Figure 5b illustrates a schematic of two actuators connected in series. They exhibit synchronized motion when subjected to the in‐phase voltage, resulting in large deformation. In contrast, applying reverse voltage induces counter‐directional motion, leading to the S‐curve deformation (Movie S5, Supporting Information). This demonstrates the synergistic capability of the fabricated soft actuators, enabling complex bending deformation for potential applications in simulating animal locomotion, such as fish swimming or animals crawling on land. Ultimately, a four‐claw gripper was fabricated and integrated with a robotic arm to tackle more complex scenario, namely the Tower of Hanoi puzzle, where at least 7 grasping actions are required to complete one transfer (Figure 5c). This demo not only demands actuators to possess sufficient deformation but also requires ample blocking force, and our actuators effectively accomplish this task (Movie S6, Supporting Information). Considering these features, the high‐performance ionic actuator based on an optimized ternary electrode has presented clear evidence for broad potential in smart gripper or soft robotic arms.

Figure 5.

Multiple applications of the actuator. a) Two‐claw gripper for grasping the microlens (including cuboids, cylinders, dodecahedrons, and polyhedral), indicating its potential application in microfabrication. b) Demo of the multi‐segment actuators moving simultaneously, enabling larger deformation or complex bending by applying different phase voltages. c) Integrating the robotic arm with a four‐claw gripper for demonstrating the Tower of Hanoi game.

3. Conclusion

In summary, we optimized a ternary electrode system (Opt‐SWCNTs/PEDOT:PSS/IL) by incorporating DMSO and Triton X‐100 based on the two‐step dispersion process. Compared to the simple mixing SWCNTs and PEDOT: PSS, the optimized fibers exhibit the distinct core–shell structure, with SWCNTs serving as the core and PEDOT:PSS/IL forming the outer layer. Additionally, vacuum filtration enables the production of electrodes with superior density, conductivity and toughness, and offers advantages such as low cost, simple operation, and high efficiency. Meanwhile, the ion migration process was elucidated through simulation and data analysis using a core–shell model. The optimal ratio between SWCNTs, PEDOT:PSS, and ionic liquid was further investigated to maximize the synergistic effect, providing a robust foundation for developing high‐performance actuators. Optimizing design parameters such as ionic liquid concentration, electrolyte to electrode thickness ratio, and device width has resulted in a high strain of the actuator while maintaining the large blocking force. Therefore, these actuators are fabricated as multi‐clawed grippers to grasp different objects, as well as being connected in series to achieve more complex deformations. However, the hot‐pressing method can squeeze out the ionic liquid for thicker devices, resulting in poor adhesion at the electrode‐electrolyte interface and reducing the device performance. Improving the interface adhesion between the two components and fabricating thicker devices with large deformation and high blocking force would further expand their application in smart grippers, soft robotics and human–machine interaction.

4. Experimental Section

Materials

20% Nafion dispersion solution and high purity single‐walled carbon nanotube (XFS‐02, >90%, 5–30 µm in length, outside diameter in 1–2 nm, SSA >380 m² g⁻¹) were purchased from the Scientific Hub services, and Jiangsu XFNANO materials. Poly(3,4‐ethylenedioxythiophene)‐poly(styrenesulfonate) (Sigma‐Aldrich, PEDOT: PSS, 1.3 wt% dispersion in H₂O, conductive grade), 1‐ethyl‐3‐methylimidazolium tetrafluoroborate (Sigma‐Aldrich, EMIM: BF₄, 98%), dimethylsulfoxide (Sigma‐Aldrich, DMSO, anhydrous, >99.9%), N,N‐dimethylacetamide (Sigma‐Aldrich, DMAc, anhydrous, >99.9%), acetone (Aik Moh, Tech grade), and isopropyl alcohol (Aik Moh, IPA, Tech grade) were used as received without further purification. Durapore PVDF membrane filters (0.22 µm pore size) were obtained from Merck Millipore, Sweden. Double‐deionized water was utilized throughout the study.

Fabrication of Ternary Electrode

Dissolve 10 mg of SWCNTs and 50 mg of Triton X‐100 in 100 mL of deionized water and disperse it by using a probe sonicator in an ice water bath. Subsequently, 5 wt% DMSO‐modified PEDOT: PSS solution was introduced into the above mixture and subjected to further ultrasonic treatment to ensure uniform blending. The dispersion was centrifuged (4000 rpm, 30 min) and 80% of the supernatant was collected, a varying amount of ionic liquid was incorporated and the ultrasonication was performed again to achieve uniform mixing. A vacuum filtration apparatus was used to filter the resulting solution, followed by sequential washing with IPA/acetone (volume ratio 1:1) and ample deionized water. Finally, the sample was dried in an oven (80 °C, 4 h) and peeled off to achieve a high conductivity flexible electrode.

Fabrication of Nafion/IL Electrolyte

A small amount of Nafion aqueous solution was thoroughly dried in an oven (80 °C, > 12 h) to obtain solid Nafion, then placed in a vacuum oven to remove the residual moisture. Similarly, the ionic liquid to be used is also placed in the vacuum oven to remove moisture. DMAC was added to dissolve Nafion using magnetic stirring in an oil bath (75 °C, >24 h). Different amounts of ionic liquid were added to the solution and stirred thoroughly to form a homogeneous solution (>12 h). The resulting solution was poured into a petri dish and cured in an oven (150 °C, 2 h) to form a film. Afterwards, the film was transferred to a vacuum oven for annealing treatment to obtain Nafion/IL electrolyte.

Fabrication of Actuator

The flexible composite electrode and Nafion/IL electrolyte can be stacked together to form a sandwich structure and placed into a compression mold. The sample was hot pressed (120 °C, 90 psi, 20 min) to ensure intimate bonding. After cooling to room temperature, the assembled actuator is removed and shaped for testing. For the detailed principles and methods refer to the Supporting Information.

Simulation of Ion Migration Characteristics

The primary and secondary current distribution and transport of diluted species (tds) modules were used to investigate ion migration in the electrode. The optimized ternary electrode system was simplified into a core–shell 2D structure, with carbon nanotubes as the core and PEDOT: PSS/IL as the shell. This ternary electrode was combined with Nafion/IL electrolyte to form a sandwich structure. To simplify the computational workload, it was assumed that ions were primarily distributed at the electrode‐electrolyte interface. After setting the ions concentration and transport properties, a ground is applied to one side of the electrode and different potentials to the other. The entire model is meshed with extra fine triangular elements, and changes in ion migration characteristics could be observed by varying the thickness of the outer shell. To further analyze ion‐induced expansion or bending of the actuator, the solid mechanics module was added, utilizing the hygroscopic swelling module for multiphysics coupling. By applying a fixed constraint to one end of the model, the bending process of the ionic actuator under different voltages and durations could be evaluated (COMSOL Multiphysics simulation service provided by ZhaoRan Technology Co. Ltd.).

Characterization and Measurements

The morphologies of the samples were recorded by field‐emission scanning electron microscopy (JEOL 7600). X‐ray powder diffraction analysis was performed by Bruker D8 Discover with nickel‐filtered Cu Kα radiation. FTIR spectra of ternary electrodes with different ratios were characterized by ATR‐FTIR spectroscopy (Frontier, PerkinElmer) in the range of 4000 – 600 cm⁻¹. Stress‐strain behavior was measured by the uniaxial tensile test (Criterion Model 42, MTS). The morphology of the two‐phase interaction structure in Nafion/IL film was investigated by SAXS (Nanoinxider, Xenocs) with an exposure time of 20 min. The sheet resistance of the electrode was measured by an HPS2524 analyzer, and its conductivity could be calculated by the following equation:

$$\sigma \left(\mathrm{S}/\mathrm{cm}\right)=\frac{1}{R\left(\mathrm{\Omega }/\mathrm{sq}\right)·t\left(\mathrm{cm}\right)}$$ (3)

here, R and t are the sheet resistance and thickness of the electrode, respectively.

All the electrochemical measurements were measured by the electrochemical workstation (PGSTAT302N) from Metrohm Autolab, and the specific capacitance could be obtained by the following equation:

$$C(F/g)=\frac{S}{2·\nu ·\mathrm{\Delta }V·m}$$ (4)

here, S, v, and ΔV are the area, scan rate, and potential window of CV curve, and m represents the mass of the electrode.

To analyze ion migration in the electrode, the three‐electrode system was used to apply voltage on the electrode. Pt, thicker silver wire, and Opt‐SWCNTs/PEDOT: PSS/IL serve as the counter electrode, reference electrode, and working electrode, respectively. Using the non‐aqueous 1.0 m EMIM: BF₄/acetonitrile solution as the electrolyte, different voltage was applied for approximately 300 s, and the ion migration within the electrode was characterized by X‐ray photoelectron spectroscopy (XPS, Physical Electronics Inc. Quantera II). A razor blade was used to cut the actuator (length × width: 7 × 2 mm² rectangular shape) and secured with a clamp. The signal generator (JDS6600) was used to apply voltage to the actuator, and the laser positioning system (Panasonic HG‐C 1050) was employed to measure the displacement variation (δ). The distance between the measuring point to the fixed end was approximately 3.5 mm (L), and the strain difference was estimated according to the following equation:^{[ 28 ]}

$$\epsilon =\frac{2t\delta }{L^2+{\delta }^2}$$ (5)

here, t is the thickness of the actuator thickness.

A force gauge (FUTEK, USB220) was utilized to measure its tip blocking force. Considering the force is proportional to the dimension of the actuator, it was converted into a normalized blocking force using the following equation:

$$F_{\mathrm{Nor}.}=\frac{FL}{w{t}^2}$$ (6)

here, L, w, and t are the length, width, and thickness of the actuator, respectively.

Conflict of Interest

The authors declare no conflict of interest.

Supporting information

Supporting Information

Supplemental Movie 1

Supplemental Movie 2

Supplemental Movie 3

Supplemental Movie 4

Supplemental Movie 5

Supplemental Movie 6

Zhang Y., Gupta A., Lu Q., Lv J., Chen S., Hu T., Yu J., Mandler D., Lee P. S., Optimization of Core–Shell Ternary Electrodes for High‐Performance Ionic Actuator in Soft Gripper. Adv. Mater. 2025, 37, 2503297. 10.1002/adma.202503297

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.