A supramolecular gel-elastomer system for soft iontronic adhesives

Electroadhesion provides a promising route to augment robotic functionalities with continuous, astrictive, and reversible adhesion force. However, the lack of suitable conductive/dielectric materials and processing capabilities have impeded the integration of electroadhesive modules into soft robots requiring both mechanical compliance and robustness. We present herein an iontronic adhesive based on a dynamically crosslinked gel-elastomer system, including an ionic organohydrogel as adhesive electrodes and a resilient polyurethane with high electrostatic energy density as dielectric layers. Through supramolecular design and synthesis, the dual-material system exhibits cohesive heterolayer bonding and autonomous self-healing from damages. Iontronic soft grippers that seamlessly integrate actuation, adhesive prehension, and exteroceptive sensation are devised via additive manufacturing. The grippers can capture soft and deformable items, bear high payload under reduced voltage input, and rapidly release foreign objects in contrast to electroadhesives. Our materials and iontronic mechanisms pave the way for future advancement in adhesive-enhanced multifunctional soft devices.

1. On Page 2, "Existing electroadhesives are commonly operated under a few kilovolts to generate sufficient adhesive force, and have limited materials selection to fabricate with". The authors may want to be careful about this argument. The electroadhesion between some materials can be achieved at potentials as low as ~1V (see the paper titled Low-Voltage Reversible Electroadhesion of Ionoelastomer Junctions, DOI: 10.1002/adma.202000600) 2. On Page 3, the schematics in Fig. 1c are nicely drawn but remain confusing. First, why does the architecture design illustrated in Fig. 1c differ from that in Fig. 4a? Which is the actual design of the soft iontronic adhesive unit used in this paper? Second, the iontronic adhesion mechanism is not explained in the manuscript. Why is the interdigitated OHGel pattern necessary for iontronic adhesion? This should be made clear in the paper. Third, the actuating mechanism of the DEA should also be clarified. Why does the DEA bend in response to voltage bias? That's not self-evident. Furthermore, for figure legend of Fig. 1c, the word "adhension" is a typo. 4. On Page 5, the author mentioned that the optimal OHGel with a polyelectrolyte-glycerol ratio of 2:1 exhibits balanced deformability, elasticity, and self-healing efficiency. However, no quantitative evidence was provided to corroborate the self-healing behavior of the OHGel. It is better to provide stress-stretch curves of OHGels before and after healing for various periods of time, and the self-healing efficiency should be discussed. 5. On page 5, "An increasing glycerol-polyelectrolyte ratio promotes water retention ( Supplementary  Fig. 3)" and thereby softens OHGel due to the enhanced degree of hydration. The reviewer thinks that water retention refers to the hydrogel's ability to retain its initial water content in open environments, whereas "water retention" in the above sentences refers to the degree of water absorption or hydration. The authors may also want to quantify the mass change (due to potential water evaporation) of the OHGel over a reasonable period of time to demonstrate the water retention capability of the material.
6. On pages 6 and 7, the authors mentioned fatigue, a word referring to the initiation and propagation of cracks in a material due to cyclic loading, which is apparently not the content discussed in this work.
7. On page 6, schematics of the self-healing mechanisms in OHGel and SHPU are provided in Fig. 2h but not explained at all. What are the primary mechanisms underpinning the self-healing behavior of SHPU and, especially, OHGel? 8. On Page 9, the authors use Timoshenko analysis to predict the bending curvature of the DEA at rest. However, the Timoshenko method is only accurate for linear-elastic material and small-strain situations. For the soft materials presented in this work, results obtained by Timoshenko analysis can only be a very rough estimation. The author should clarify this in the manuscript. 9. On Page 10, it is mentioned that the contact layer employs a hydrophobic coating on its outer surface to obviate the inherent tackiness of SHPU and enables self-cleaning by preventing the collection of contaminants. The word self-cleaning is a bit misleading. We call lotus leaf a self-cleaning material considering its unique nanoscopic architecture of the leaf surface, but we do not regard plastic bottles (e.g., bottled water) as self-cleaning just because of their non-sticky nature.
10. On Page 10, Fig. 4a might be further improved by numbering each layer, so that one can know that the top SHPU is the first layer, the biased OHGel is the second layer, the stretched SHPU is the third layer, ……, the bottom SHPU with fingertips OHGel pattern is the 6th layer. The reviewer had a hard time figuring out the meaning of Fig. 4a.
Reviewer #3 (Remarks to the Author): In this manuscript, the authors design and synthesize a supramolecular gel-elastomer system and use the inkjet printing technique to fabricate a soft gripper, which is composed of a dielectric elastomer actuator and an electrostatic adhesion segment. The soft gripper exhibits excellent adaptation to various soft and deformable objects and a large payload up to ~670. It is interesting that the use of ionic gels as the electrodes leads to fast release of the gripper from the objects. Whereas individual performance of the materials and devices have been demonstrated to different extents in literature (e.g. (1) soft grippers with intrinsic electroadhesion based on DEA, J. Shintake, 2015, Adv. Mater.; (2) integrated 3D printing of gel-elastomer system with robust heterolayer adhesion, P. Zhang, 2022, Nature Comm.; (3) device-level healing, self-healing soft electronics, J. Kang, 2019, Nature Electronics; devices with intrinsic self-healing properties of all device components, M. Khatib, 2018, Small.), this work achieves the performances in an integrated device at the same time. The authors have investigated the properties of the materials and the device in detail and the experimental results can support the main conclusions.
There are certain statements arising concerns about general concepts and must be rephased. Some detailed comments are as follows.
1. On page 2, the statement that "ion transportation in a swollen polymeric network will not be interrupted even under hyperelastic deformation" is not true. Depending on the ion transport mechanism, the deformation of polymer network will play a key part, e.g. by narrowing the ion transport path or by dragging the tethered ions along with the polymer chains, etc.
2. In fig. 1d, the chemical structures of glycerol and water do not have annotations. For example, do the red sphere represent oxygen atom? 3. Response time should be provided if "fast release from foreign surface" is mentioned repeatedly.
4. The operating voltage of the gripper herein is also on the order of 1 kV, so it is not convincing to emphasize "reduced voltage input". 5. Table 1 might include one more column to compare the payloads of different devices, now that the payload of this work is much higher than others. 6. On page 5, the statement "The incorporation of nonvolatile, hygroscopic glycerol avoided unwanted solvent loss and no sign of gel stiffening was observed over the course of this study." should be revised. Glycerol can help retain water but solvent loss is inevitably in general for gel materials, e.g. in elevated temperature or removed by the foreign bodies during contact or in aqueous environment, etc. (Kim, et

Reviewer #1
This work reports a set of materials (hydrogel electrode and TPU dielectric) and a printing fabrication process for soft electroadhesion. The results seem interesting, the figures are clear and the demos of good quality. There is a major problem with novelty and contribution though. I disagree with the authors about their analysis of the state of the art (weaknesses of current materials for electroadhesion) and major claimed contributions (ionotronic adhesion mechanism, better performing materials for electroadhesion).
I can see potential in this work given the high-quality of the results presented, but the authors should first reshape their article clarifying their novelty compared to the state of the art. The presentation of the mechanism is misleading. The comparison of materials performance for electroadhesion is not looking at the right metrics (e.g., electrostatic energy density, shear pressure, load to weight ratio). The working style should also be improved as currently it makes it difficult to understand the content. Printing of stretchable electroadhesives and self-healing seem novel and can have advantages in electroadhesion but are now only presented as marginal parts of the work. I suggest that the authors expand on these aspects of their work, clarify their analysis of the state of the art and present clear novelty and contribution.

Comment #1-1
The authors claim in the abstract an ionotronic adhesion mechanism. However, the adhesion mechanism seems to me to be no different from conventional Electroadhesion based on interdigitated electrodes insulated by a dielectric polymer. The authors replaced metals or conductive polymers with hydrogels as electrodes. Since electroadhesion is based on electrostatics, replacing metals or conductive elastomers with hydrogels as electrodes does not change the 2 mechanism for electroadhesion. Charge transportation mechanism within the electrodes changes, but electroadhesion is based on forces between static charges accumulated on the surfaces of the electrodes, which do not change between an electronic conductor and a hydrogel.

Response:
We agree with the reviewer that our iontronic adhesives share the same physical principle (electrostatic/Coulomb attraction) with electroadhesives. When ionic gel-based electrodes are biased, mobile ions in the gel would accumulate and mirror the induced charges/dipoles in the opposing surface to generate electrostatic adhesion. Herein, we recognize "electrostatic adhesion" as the underlying mechanism for our devices, and acknowledge that "iontronic adhesion" may not represent a new mechanism. Instead, we shall describe it as a new configuration of electrostatic adhesion that uses ionic gel as electrodes. The statements open-circuit condition to facilitate the adhesives' detachment from a conductive surface. In light 3 of the above statements, we believe it is proper to name our devices as iontronic adhesives and iontronic-adhesive grippers, and to describe their underpinning mechanism as electrostatic adhesion. Our previously defined "iontronic adhesion" mechanism is no longer used in the manuscript.
The reviewer also mentioned another work (Advanced Materials 32, 2000600, 2020) that realized reversible iontronic adhesion with ± 1 V. Under forwards bias, interfacial adhesion could be established between two ionoelastomers whose polymer chains are oppositely charged, through the formation of an ionic double layer. It's obvious that such adhesion force can only form between the designated active materials; replacing one of the ionoelastomer to another material would fail to render the effect. Therefore, this work reported an iontronic adhesion phenomenon at materials level; it does not offer a universal solution to generate electrostatic adhesive force on an arbitrary foreign object, and it is not related to the devising of controllable adhesive modules for soft robotics. In contrast, our work focuses on the delivery of device-level iontronic adhesion, wherein the charged ionic gel electrodes can polarize most foreign passive surfaces/objects to render normal/shear adhesion force and perform gripping/handling functionalities. Therefore, it is fair to compare our work with the precedent device-level electrostatic adhesives as summarized in Table   1, whereas the ionoelastomer work represents the progress in another subcategory of electrostatic adhesion. To clarify this point, we have modified our statement on Page 2, line 10 to "Existing device-level electroadhesives…".
The authors state (Introduction, page 2) that "Electroadhesives based on elastomeric components, such as silicone and stretchable electronic conductor (e.g. carbon-filler-percolated composite) feature improved geometrical adaptation, but at a cost of the low dielectric constant of silicones, questionable conductor-dielectric interfacial bonding, and potential mechanical/electrical failures (tearing, puncturing, loss of conductivity) under large deformation." I do not agree with this analysis for the following reasons.

Comment #1-2
First, the role of the dielectric constant in the performance of electroadhesion is questionable. It might help slightly reducing the voltage, which might or might not be critical depending on the 4 application, but alone it does not lead to higher performance. 1 -5 kV voltages can be easily provided by palm-sized battery driven power supplies, so I do not see any revolutionary advantage in reducing the voltage from 5 to 1 kV. Performances of electroadhesion (similarly to DEAs and any other electrostatic actuator/clutch) do not depend on the dielectric constant ε alone, but are rather related to the maximum electrostatic energy density of the dielectric material, which is ε E_BD^2, with E_BD being the breakdown field (see Hinchet, R., Shea, H., 2020. High Force Density Textile Electrostatic Clutch. Advanced Materials Technologies). When replacing a dielectric material for an electrostatic actuator or an electroadhesion device, the metric to be compared should be the maximum energy density ε E_BD^2, not only the dielectric constant ε. I recommend the authors to include this comparison in their article. Based on the reported data, the electroadhesion devices developed in this work use a TPU based dielectric (SHPU) with ε = 7 and operate with E=V/g = (1000 "V")/(2* 60 µ"m" ) = 8.3 "V/" µ"m" . Silicone instead has typically ε = 2.7 and can reach and exceed breakdown fields values of E=100 "V/" µ"m" (Sylgard 184, see Albuquerque, F.B., Shea, H., 2020. Smart Mater. Struct.). Therefore, from these data one cannot conclude any evident advantage in electrostatic forces when replacing Silicone with SHPU since the factor 2.6 gained on the dielectric constant would come at the price of a loss in electric field of factor 12. Beyond electrostatics, performance of electroadhesion is highly influenced by mechanical and surface properties (see Cacucciolo, V., Shea, H., Carbone, G., 2022. Peeling in electroadhesion soft grippers. Extreme Mechanics Letters). I recommend that the authors include a proper comparison of their materials and devices with the state of the art, using the most widely accepted metrics: electrostatic energy density, shear pressure, load to weight ratio.

Response:
1) The significance of voltage reduction: We employed 1 kV as the testing voltage to benchmark the adhesive patches' performance on different foreign substrates (Fig. 5 c-d). However, for many usage scenarios (like when handling metals ( Fig. 4d) and smooth/high-k dielectrics), the adhesive patches can function under several hundred volts to generate sufficient shear pressures. Under 400 V input, the iontronic adhesive produces a shear pressure of 2.26 kPa, 1.28 kPa, and 0.50 kPa on aluminum, smooth glass, and grounded glass, respectively, which corresponds to a payload of 86.1 g, 50.6 g, and 19.8 g for our four-fingered iontronic-adhesive gripper (Supplementary Fig. 26). Decreasing the operating 5 voltage from kilovolts to hundred volts is a key step to enable small-scale, untether soft robots.
While stepping up a battery output to kilovolts requires palm-sized power supplies, these is a vast selection of PCB-compatible, miniature, and ultralight (< 1g) power sources that can upscale a battery output up to 500 V.

2) Electrostatic energy density:
We agree with the reviewer that the peak performance of electrostatic adhesion is determined by the maximum energy density (κEB 2 ) of the contact dielectric layer, where both dielectric constant (κ) and dielectric breakdown field (EB) are deciding factors. However, in real world applications, electrostatic adhesives should not be operated near the marginal EB for safety and reliability reasons. The advantage of having a high dielectric constant lies in that one can achieve the same level of electrostatic energy density, or adhesion performance, using a lower voltage input in the dielectric's safe range. By replacing Sylgard 184 (κ ≈ 2.8) with SHPU (κ ≈ 6.8), a 1.56 times reduction in operating voltage can be realized while rendering the same electrostatic energy density.
Besides, 1 kV is a safe operating voltage for our devices rather than the breakdown voltage of SHPU. To calculate the maximum energy density of SHPU, we measured its dielectric strength by 6 following ASTM-D3755. As shown in Supplementary Fig. 13a, a SHPU specimen (thickness ≈ 200 μm) was mounted between two opposing brass probes (diameter = 25 mm) with the testing cup filled with silicone oil to prevent flashover and premature breakdown. A Hipot tester (Chroma 9056) was employed to supply DC potential at a ramping rate of 500 V s -1 . The breakdown voltage (UB) and SHPU film thickness (d) were recorded to calculate breakdown field (EB = UB/d). The results were analyzed by fitting into the two parameter Weibull distribution function: where P is the cumulative probability of dielectric breakdown occurring at electric field equal to or below E, β is the shape parameter that describes the scattering of data, and EB is the characteristic breakdown strength at a cumulative failure probability of 63.2%. To find EB, the function was rearranged and plotted as log(-ln(1-P)) against E ( Supplementary Fig. 13b). The characteristic dielectric strength of SHPU was determined to be 63.6 V μm -1 . This value is comparable with some commercial TPUs, such as ElastollanⓇ 1185A10 (product of BASF, EB = 88.6 V μm -1 , ACS Appl. Mater. Interfaces 2017, 9, 6, 5237-5243). Based on the same method, the dielectric strength of VHB 4905 and Sylgard 184 (without prestretch) was measured to be 37.9 V μm -1 and 27.2 V μm -1 and, respectively.

3) Proper comparison of materials and devices:
Electrostatic energy density: The dielectric constant (κ), dielectric strength (EB), and maximum electrostatic energy density of SHPU, Sylgard 184, and VHB 4905 are compared using bar charts as summarized in Supplementary Fig. 13c

Reviewer #2
This work reports an iontronic adhesion mechanism using a gel-elastomer system consisting of an ionic organohydrogel (OHGel) as adhesive electrodes and resilient polyurethane (SHPU) as dielectric layers. An ionic soft gripper that seamlessly integrates actuation, adhesive grip, and external perception is designed through additive manufacturing. The idea of this work is novel, the results are technically sound and well-delivered with ample evidence, and the manuscript is clearly written. The following comments should be addressed before the recommendation for publication can be made.

Comment #2-2
On Page 3, the schematics in Fig. 1c are nicely drawn but remain confusing. First, why does the architecture design illustrated in Fig. 1c differ from that in Fig. 4a? Which is the actual design of the soft iontronic adhesive unit used in this paper? Second, the iontronic adhesion mechanism is not explained in the manuscript. Why is the interdigitated OHGel pattern necessary for iontronic adhesion? This should be made clear in the paper. Third, the actuating mechanism of the DEA should also be clarified. Why does the DEA bend in response to voltage bias? That's not selfevident. Furthermore, for figure legend of Fig. 1c, the word "Adhension" Is A Typo.
Response: 1) We thank the reviewer for pointing out the inconsistency between Fig. 1c and Fig. 4a. Fig. 1c is a conceptual illustration for the iontronic device, whereas

Comment #2-3
Page 4. The word DEA (dielectric actuator?) should be defined at its first usage.

Response:
We have provided the full name (dielectric elastomer actuator) for DEA at its first usage (page 4, line 2).

Comment #2-4
On Page 5, the author mentioned that the optimal OHGel with a polyelectrolyte-glycerol ratio of 2:1 exhibits balanced deformability, elasticity, and self-healing efficiency. However, no quantitative evidence was provided to corroborate the self-healing behavior of the OHGel. It is better to provide stress-stretch curves of OHGels before and after healing for various periods of time, and the self-healing efficiency should be discussed.

Response:
We have supplemented the comparison between different OHGel compositions to showcase the advantages of PE10/GY5 (Supplementary Fig. 7a). The mechanical self-healing behavior of PE10/GY5 was quantitively recorded via uniaxial tensile test (Supplementary Fig. 7b). After bisecting, rejoining, and healing the sample at room temperature for 15 minutes, it could restore 96.5% stretchability and 100% ultimate tensile strength.
Supplementary Fig. 7 | a, Comparison of OHGels with different PE/GY ratio. PE10/GY3 is freestanding yet barely heals due to the limited chain motion; PE/GY7 heals rapidly, but at a cost of becoming non-free-standing; PE10/GY5 occupies both good mechanical integrity and rapid selfhealing capability. b, Mechanical self-healing behavior of the PE10/GY5 OHGel characterized by uniaxial tensile test. After bisecting, rejoining, and healing at room temperature for 15 minutes, the OHGel (PE10/GY5) could restore 96.5% stretchability and 100% ultimate tensile strength.

Comment #2-5
On page 5, "An increasing glycerol-polyelectrolyte ratio promotes water retention ( Supplementary   Fig. 3)" and thereby softens OHGel due to the enhanced degree of hydration. The reviewer thinks that water retention refers to the hydrogel's ability to retain its initial water content in open environments, whereas "water retention" in the above sentences refers to the degree of water absorption or hydration. The authors may also want to quantify the mass change (due to potential water evaporation) of the OHGel over a reasonable period of time to demonstrate the water retention capability of the material.
Response: our statement on page 5, line 14 should agree with the reviewer's point that OHGel can preserve its initial water content, rather than continuously absorbing water from air. We have modified the statement as "An increasing glycerol-polyelectrolyte ratio preserves more initial water in OHGel and thereby softens OHGel due to the enhanced degree of polyelectrolyte-solvent interaction" to address this point. After inkjet printing the pre-OHGel ink (with high water content) on to a SHPU substrate, the formulation would partially dehydrate and then gelate physically under ambient condition. After ~ 6 hours, the OHGel samples could reach dynamic equilibrium in water evaporation/reabsorption, on condition that the ambient humidity is stable (RH ≈ 60%).
Accordingly, an initial weight loss (Supplementary Fig. 4a) was observed for all OHGel samples with different PE/GY ratio. We further recorded their net weight for up to 14 days (room temperature, ambient air) and found no noticeable mass change in PE10/GY7 and PE10/GY5. PE10/GY3 further lost ~ 3.6% in net weight on day 14 due to its low glycerol content ( Supplementary Fig. 4b).

Comment #2-6
On pages 6 and 7, the authors mentioned fatigue, a word referring to the initiation and propagation of cracks in a material due to cyclic loading, which is apparently not the content discussed in this work.

Response:
We appreciate the reviewer pointing out that fatigue refers to the irreversible mechanical property change or fracture in a bulk material after being exposed to cyclic load. We agree that such phenomenon is not discussed in this work. The statement on page 6, line 4 has been modified as "…that protects OHGel electrodes from mechanical damages such as accidental perforation".
The statement on page 7, line 22 has been updated as "SHPU could recover from viscoelastic deformation as suggested by the almost identical stress-strain loops".

Comment #2-7
On page 6, schematics of the self-healing mechanisms in OHGel and SHPU are provided in Fig.   2h but not explained at all. What are the primary mechanisms underpinning the self-healing behavior of SHPU and, especially, OHGel?
Response: The self-healing mechanisms for OHGel and SHPU are briefly mentioned on page 8, line 4 and line 11, respectively. A detailed discussion on OHGel self-healing mechanism is provided below Supplementary Fig. 7.

Self-healing mechanism of SHPU:
Based on the macromolecular design of SHPU, UPy motifs carrying quadruple H-bonding sites are introduced within its soft matrix to form dynamic crosslinks. UPy undergoes rapid association and dissociation of the quadruple H-bonding units (Fig. 2d, bottom) with a UPy-UPy dimer lifetime of 1.7 s. When a bulk SHPU sample is bisected and rejoined, dynamic chain motion/diffusion in the soft domains offers the opportunity for a pair of free, non-associated UPy groups from both halves to dimerize and reform dynamic crosslink at the cut. Since the Tg of soft domain is subzero (-12 ℃), the polymer chain in soft domains exhibits high mobility at room temperature and can thus facilitate SHPU self-healing under ambient condition.

Self-healing mechanism of OHGel:
OHGel is formed by gelating P(SPMA0.5-r-MMA0.5) in a water/glycerol binary solvent. Therefore we believe that our estimation wouldn't deviate too much from the real situation.
Nonetheless, we have updated our statement on page 10, line 7 as "The DEA adopts a dielectric elastomer minimum energy structure (DEMES) whose bending curvature at rest is approximated using Timoshenko analysis".

Comment #2-9
On Page 10, it is mentioned that the contact layer employs a hydrophobic coating on its outer surface to obviate the inherent tackiness of SHPU and enables self-cleaning by preventing the collection of contaminants. The word self-cleaning is a bit misleading. We call lotus leaf a selfcleaning material considering its unique nanoscopic architecture of the leaf surface, but we do not regard plastic bottles (e.g., bottled water) as self-cleaning just because of their non-sticky nature.

Response:
The incorporation of hydrophobic coating (silica nanoparticle) on SHPU can drastically reduce its surface energy, as is evidenced by the increment in water contact angle (CA) from 86° to 149° (Supplementary Fig. 20a). The coating thereby manifests superhydrophobicity, one of the underlying mechanisms that contribute to self-cleaning effect. As demonstrated in Supplementary Fig. 20b, after directly contacting the dirt, the pristine SHPU film became contaminated whereas the one with superhydrophobic coating remained clean. In light of the above evidence, the superhydrophobic surface of coated SHPU is essentially different from a non-sticky plastic film. SHPU membranes before and after contacting the dust. The pristine SHPU collected a lot of dust due to its inherent tackiness, whereas the one with hydrophobic coating maintained clean.

Comment #2-10
On Page 10, Fig. 4a might be further improved by numbering each layer, so that one can know that the top SHPU is the first layer, the biased OHGel is the second layer, the stretched SHPU is the third layer, ……, the bottom SHPU with fingertips OHGel pattern is the 6th layer. The reviewer had a hard time figuring out the meaning of Fig. 4a.

Response:
We have numbered the layers in Fig. 4a. The Figure caption has been updated accordingly to annotate each layer with its designated number.

Reviewer #3
In this manuscript, the authors design and synthesize a supramolecular gel-elastomer system and use the inkjet printing technique to fabricate a soft gripper, which is composed of a dielectric elastomer actuator and an electrostatic adhesion segment. The soft gripper exhibits excellent adaptation to various soft and deformable objects and a large payload up to ~670. It is interesting that the use of ionic gels as the electrodes leads to fast release of the gripper from the objects. There are certain statements arising concerns about general concepts and must be rephased. Some detailed comments are as follows.

Comment #3-1
On page 2, the statement that "ion transportation in a swollen polymeric network will not be interrupted even under hyperelastic deformation" is not true. Depending on the ion transport mechanism, the deformation of polymer network will play a key part, e.g. by narrowing the ion transport path or by dragging the tethered ions along with the polymer chains, etc.

Response:
We agree with the reviewer that the deformation of ionic gels can influence the behavior of ion transportation, and hence their ionic resistivity will increase under hyperelastic stretch. Such phenomenon has been experimentally validified by Keplinger et al. (Science 341, 984-7, 2013) in the research of hydrogel-electroded DEAs, where they found the increment of ionic resistivity is related to the concentration of electrolyte (Fig. R1). For the ionic hydrogel dissolving 5.48 M NaCl, its resistivity increased slightly from 0.016 Ω·cm to 0.034 Ω·cm when the sample was elongated 6 times of its initial length (due to narrowed ion transportation pathway in the polymeric network, etc.). We have modified our statement on page 2, line 25 as "ionic gels can maintain ion transportation (with slight increment in resistivity) even under hyperelastic deformation" to clarify this point. Reproduced from (Science 341, 984-7, 2013) with permission granted.

Comment #3-2
In fig. 1d, the chemical structures of glycerol and water do not have annotations. For example, do the red sphere represent oxygen atom?

Response:
We have annotated the atoms' color in the caption of Fig. 1e.

Comment #3-3
Response time should be provided if "fast release from foreign surface" is mentioned repeatedly.
Response: For our iontronic gripper, the response (release) time from a metallic foreign surface is summarized in Fig.   4g. Under 1 kV voltage input, the adhesive patched could retract from a smooth (Sa = 0.12 μm) and a rough (Sa = 0.76 μm) metallic surface within 5.9 s and 3.1 s, respectively.
Faster release was observed when lower voltage (a few hundreds volt) was supplied. 20

Comment #3-4
The operating voltage of the gripper herein is also on the order of 1 kV, so it is not convincing to emphasize "reduced voltage input".

Response:
We used 1 kV as the standard driving voltage to benchmark the adhesive patches' performance on different foreign substrates (Fig. 5 c-d). For many usage scenarios (like when handling metals and smooth/high-k dielectrics), the gripper can function under several hundred volts (Fig. 4d) to generate sufficient shear pressures. Under 400 V voltage input, the iontronic adhesive can produce a shear pressure of 2.26 kPa, 1.28 kPa, and 0.50 kPa on aluminum, smooth glass, and grounded glass, respectively, which corresponds to a payload of 86.1 g, 50.6 g, and 19.8 g for our four-fingered iontronic-adhesive gripper (Supplementary Fig. 23).

Response:
We appreciate the reviewer addressing the difference between "water retention" and "solvent loss prevention". We agree that even for non-volatile solvents with very high boiling point (such as glycerol), they can evaporate slowly and finally cause solvent loss after a very long period of time (over the length of this study). In our iontronic devices, the OHGel electrodes are properly encapsulated by SHPU, therefore they are not likely to contact foreign bodies directly to cause solvent loss. We have revised the statement on page 5, line 13 as "hygroscopic glycerol helped to retain water" to address this point. 22

Comment #3-7
The plot of Fig. 2k is confusing. The x-axis is tensile strain and the y-axis is resistance, while the curve is R/R0 = λ2.

Response:
The expression of R/R 0 = λ 2 is theoretically correct to describe the stretch-induced resistance change in an ideally stretchable conductor.
We have rearranged it into R = R 0 (1+ε) 2 to properly mirror the variables with the axis labels, where λ = 1+ε, λ is tensile stretch (ratio between final and initial length), ε is tensile strain. The statement on page 8, line 9 was adjusted accordingly.

Comment #3-8
On page 8, the authors claim that they "deliver the first gel-elastomer system that forms strong and inherent interfacial bonding without requiring additional surface treatment or coupling agent.", which is too ambitious. See, for example, P. Zhang, 2022, Nature Comm. and Q. Ge, 2021, Sci.
Adv. The quality of the paper is significantly improved in this second version, but there are still some critical points to be clarified. In particular, the advantage of reducing the operating voltage needs to be supported by more grasping experiments below 500 V.

Response
While I believe the claim that the electroadhesion performance is improved in this work compared electroadhesives made of silicone elastomers should be removed as it disagrees with data in literature.

Comment #1-2
I agree with the authors that reducing the voltage to under 500 V is a game changer for the PCB size. However, if this is a key result for this paper, then it should be included in the main rather than in Supplementary Information. Most of the demos shown in the main of the paper now show about 1 kV of operation (Fig. 4). Please include demos that operate under 500 V in the main to support the ≤ 500 V operation as a main contribution of this work. Fig. 5e  Response: 2) We appreciate the reviewer pointing out that it's the current and not voltage that causes safety issues to human body. We have read the above paper thoroughly to obtain better understanding on this topic. Note that we are not claiming any safety-related advantages of our iontronic-adhesives in the manuscript.

Response: 1) As shown in
Also, I disagree with the authors about the fact that "The advantage of having a high dielectric constant lies in that one can achieve the same level of electrostatic energy density, or adhesion performance, using a lower voltage input in the dielectric's safe range.". Please notice that "The dielecric safe range" decreases together with the breakdown field E_BD. Assuming for example that one operates a device at 80% of breakdown field, this value is 80 -90 V per micrometer for Sylgard 184 PDMS, versus 8 -9 V per micrometer of the SHPU dielectric material presented in this work. This means that the electrostatic energy density, which is the key electrical metric for adhesion performance, is significantly lower for SHPU than for Sylgard 184 PDMS.
The breakdown field of Sylgard 184 PDMS is measured to be lower than SHPU by using the same testing protocol (Supplementary Fig. 13). Based on our data, we claim that the electrostatic energy density of SHPU is higher than that of PDMS.
Many thanks to the authors for performing breakdown strength tests using a rigorous procedure.
These results are very helpful. However, I am surprised with some of the results. repeat the experiments without the silicone oil bath.
Response 4) As suggested by the reviewer and the mentioned review paper (Hajiesmaili, 2021), breakdown strength is not an intrinsic material property, but a measured parameter that depends on sample geometries and testing conditions. Generally, the breakdown strength of an elastomer sample decreases with increasing sample thickness (EB ∝ t -0.5 ) and area under field (more defects).
The discrepancy between our Weibull distribution and the literature may be due to the different thickness, area, and testing conditions of Sylgard 184. Our Sylgard 184 specimens have a thickness of ~ 200 μm and a testing area of 4.91 cm 2 between the flat brass probes (probe diameter = 2.5 cm). As for the > 100 V/μm breakdown strength of Sylgard 184 mentioned in the review paper ( Fig. 8, Hajiesmaili, 2021), we are unable to identify the samples' geometrical information as What we learned from the review paper (Hajiesmaili, 2021) is that the reported over 100 V/μm EB of Sylgard 184 is measured with a hemispherical electrode resting on a slab of elastomer with a constant load (Fig. R1b), which corresponds to the configuration of Hertz indentation. The contact area between the hemispherical electrode tip and the elastomer has been estimated to be smaller than 0.5 mm 2 (International Journal of Smart and Nano Materials 6.4 (2015): 290-303; authored by M. Kollosche), which is ~1000 times smaller than the flat disc electrodes employed in our experiment (Fig. R1c). Given a n times larger testing area, the breakdown voltage at 63.2% cumulative probability would decrease by a factor of n -1/β (page 19, Hajiesmaili, 2021), where β is the Weibull modulus. With n = 1000 and β = 9.4 (derived from our Weibull fitting), a 52% 4 reduction in EB can be estimated. Considering other influencing factors such as different sample thickness, environmental humidity, and different setting in cutoff current that defines a breakdown event (which is 0.1 mA in our case), the lower EB of Sylgard 184 in our measurement is justifiable.
Oil bath has been commonly used in dielectric breakdown test to prevent O2 induced flashover (arc) and premature breakdown. It's considered beneficial for obtaining more accurate EB values.
The experiments were conducted immediately after placing the elastomer in the oil cup so the influence of silicone oil impregnation should be minimal.
To avoid any mislead/dispute, we have mentioned in our main manuscript (page 7, line 29) that the reported EB values are consistently measured and are valid when referring to our testing protocol. Breakdown strength may vary when different testing protocols are applied. c) Dielectric breakdown measurement setup using a pair of flat brass electrodes (used in our experiment).