User login


You are here

Journal Club for February 2018: HASEL artificial muscles for high-speed, electrically powered, self-healing soft robots

Christoph Keplinger's picture

Eric Acome, Nicholas Kellaris, Timothy Morrissey, Shane K. Mitchell, Christoph Keplinger

University of Colorado Boulder



The Keplinger Research Group recently published a pair of papers in Science and Science Robotics introducing a new class of artificial muscles which we have termed Hydraulically Amplified Self-healing ELectrostatic (HASEL) actuators. HASEL actuators harness an electrohydraulic mechanism to activate all-soft-matter hydraulic architectures. They combine the versatility of soft fluidic actuators with the muscle-like and self-sensing performance of dielectric elastomer actuators, while simultaneously addressing critical challenges of both. We demonstrated the versatility and muscle-like performance of these different types of HASEL actuators through exemplary applications including a soft gripper handling delicate objects, a self-sensing robotic arm, and a transparent artificial muscle (YouTube overview of HASEL).



Traditional robots primarily consist of rigid components with limited degrees of freedom and functionality, whereas plants and animals have evolved a host of soft structures to effectively function in unpredictable environments. Over the past few decades, pioneering innovation within the fields of soft robotics and stretchable electronics has strived to bridge the mismatch in mechanics between the engineered and the biological worlds.

Today, soft robotic devices are rapidly finding a diverse range of applications spanning from industrial food handling to health care. We are encouraged to advance this field by wide interest from both academia and industry, which is signified by recent funding calls such as the NSF Emerging Frontiers in Research and Innovation (EFRI) program with “Engineering Frontiers of Soft Robotics” selected as a 2018 topic, and the rapid growth of start-up companies such as Soft Robotics Inc.

A critical component of soft robotics are soft actuators. Currently, two classes of soft muscle-mimetic actuators dominate the literature – soft fluidic actuators (predominantly based on pneumatic operation) and electrically powered dielectric elastomer actuators (DEAs) – both of which have advantages and drawbacks. Fluidic actuators readily feature diverse modes of actuation such as bending and twisting, but require complex and inefficient systems of tubes, valves, and sources of pressurized fluid. DEAs excel with muscle-like performance of actuation and are controlled and powered electrically, but they face hurdles towards large scale applications due to susceptibility to catastrophic failure from dielectric breakdown and electrical aging.


Introducing Hydraulically Amplified Self-healing ELectrostatic (HASEL) actuators

HASEL actuators harness an electrohydraulic mechanism to drive shape change of soft active structures. By directly applying electrostatic forces to an insulating hydraulic fluid, HASEL actuators combine the versatility of soft fluidic actuators with the muscle-like and self-sensing performance of dielectric elastomer actuators.

In contrast to soft fluidic actuators, where inefficiencies and losses arise from fluid transport through systems of long tubes and channels, HASEL actuators generate hydraulic pressure locally via electrostatic forces acting on liquid dielectrics distributed throughout the system. In contrast to DEAs, where dielectric breakdown through elastomeric membranes limits lifetime and reliability, we show that the use of liquid dielectrics in HASEL actuators enables self-healing attributes with immediate and full recovery of actuation performance even after tens of breakdown events.
In our recently published papers, we I) discussed fundamental physical principles of HASEL actuators, II) demonstrated the robust and muscle-like performance of HASEL using prototypical designs and geometries, and III) illustrated the wide potential and versatility of HASEL actuators by demonstrating exemplary applications such as soft grippers and self-sensing artificial muscles powering a robotic arm - all using only widely available, low-cost materials and industrially-amenable fabrication techniques for soft and flexible materials. We believe there is a wide range of potential applications for HASEL that are still untapped as a plethora of geometries, materials and more advanced fabrication strategies remain unexplored.


Three illustrative designs of HASEL

a) Donut HASEL actuators (YouTube video, Fig. 1)

  • Linearly expand upon application of voltage
  • Force and strain can be tuned using hydraulic principles
  • Harness a pull-in instability to achieve large strains
  • Reliably self-heal after dielectric breakdown

Fig. 1. Basic components and fundamental physical principles of HASEL actuators illustrated using the donut HASEL geometry. (A) A schematic of a donut HASEL actuator shown at three different applied voltages, where V1<V2<V3. (B) Typical actuation response of a HASEL actuator with geometry shown in (A), which deforms into a donut shape with application of voltage (C). (D and E) Strain and force of actuation can be tuned by modifying the area of the electrode. (F) The use of a liquid dielectric confers self-healing capabilities to HASEL actuators. (Figure reproduced from [i])


b) Planar HASEL actuators (YouTube Video, Fig. 2)

  • Apply electric field over entire region containing liquid dielectric
  • In-plane expansion, resembling common mode of actuation used for DE actuators
  • Achieve higher strain at same voltage compared to DE actuators
  • Capable of large actuation strain
  • Can be scaled up to deliver large forces
  • High specific power 
  • Self-heal after dielectric breakdown
  • Self-sense deformation

Fig. 2. Components and structure of a planar HASEL actuator which functions as a linear actuator. Linear actuation can be achieved by implementing a fixed prestretch in one planar direction and applying a load in the perpendicular planar direction. (Figure reproduced from [i])


c) Peano-HASEL actuators (YouTube Video, Fig. 3)

  • Linear contraction on activation without relying on prestretch, stacked configurations, or frames
  • Capable of high-speed continuous actuation 
  • Inexpensive and industrially amenable materials and fabrication techniques
  • Do not have to rely on highly stretchable electrodes or shell material
  • Can be designed to be highly transparent

Fig. 3. Schematic of a Peano-HASEL actuator with three units. The actuator linearly contracts as voltage increases from V1 to V3. (Figure reproduced from [ii])


Standing on the shoulders of giants: HASEL actuators build upon recent breakthroughs in mechanics and materials

Here are a few important examples:

  • Stretchable ionics [1-3], Journal Club from Zhigang Suo
  • Robust bonding of hydrogels to elastomers [4]
  • Dielectric elastomer actuators [5-7]
  • Fabrication techniques for soft fluidic actuators [12-13]
  • Harnessing instabilities in soft active structures [14-15], Journal Club from Xuanhe Zhao
  • Peano fluidic actuators [16-18]
    • Heat sealing fluidic actuators
  • Inspiration from industry
    • Transformer oil that self-heals after dielectric breakdown


Opportunities and challenges for mechanics, materials, and beyond

The introduction of HASEL raises many questions and opens opportunities for future breakthroughs in mechanics, materials, electronics, and robotics. A few examples to stimulate discussion include:

  • Improved self-healing from electrical damage:
    • Use of a liquid dielectric designed specifically for this application rather than borrowing material from other high voltage applications (high voltage transformers).
    • Electrical breakdown leads to the formation of gas bubbles through vaporization. How could we avoid, mitigate, or remove gas bubbles?
    • Peano-HASEL actuators have yet to demonstrate reliable electrical self-healing, as the thin BOPP experiences significant mechanical damage during breakdown. 
  • Self-healing from mechanical damage: HASEL devices in service may experience damage beyond electrical breakdown, namely mechanical damage such as tears, abrasions, or cuts. Mechanically self-healing dielectric [19] and ionically conductive [20] materials could expand self-healing abilities of HASEL. New self-healing materials may be especially useful in the Peano-HASEL system.
  • Improving lifetime of actuators: Some possible failure modes of HASEL actuators call for a better fundamental understanding of fracture and fatigue of soft and flexible materials, and interface mechanics of layered structures (hydrogels, elastomers, thermoplastics, liquid dielectrics).
  • Tunable liquid dielectric: The performance of HASEL actuators can be tuned through modification of the liquid dielectric. For instance, the dynamics of actuation can be changed by varying the viscosity of the liquid dielectric. Optimization of the liquid dielectric for higher permittivity and dielectric breakdown strength would also increase actuation strength.
  • Fluid mechanics: Optimization of fluid flow in these soft hydraulic structures would lead to improved efficiency and speed.
  • Nonlinear mechanics and electromechanical coupling in soft hydraulic structures: What new types of electromechanical instabilities can be explored in electroactive soft hydraulic structures?
  • Advanced fabrication methods: How can advanced fabrication methods such as additive manufacturing contribute to a next generation of HASEL actuators with significantly more complex designs? Further exploration of new designs and geometries may lead to additional actuation modes with unique and surprising applications.
  • Miniaturized high-voltage electronics: Development of compact, inexpensive, safe, and energy efficient high-voltage electronics will extend the area of application for HASEL actuators.
  • Applications in robotics:
    • How can we integrate arrays of HASEL actuators with distributed sensing and computation to develop the next-generation of soft robots?
    • Practical applications of soft robotics will require a better understanding of agile, non-equilibrium control of sensor- and actuator-rich synthetic systems.


We look forward to hearing your thoughts

The development of HASEL actuators has heavily built upon knowledge and research from the mechanics and materials community. We look forward to an exciting discussion with experts in these areas and to hearing your questions and comments.


Links cited in this post


[i] E. Acome, S. K. Mitchell, T. G. Morrissey, M. B. Emmett, C. Benjamin, M. King, M. Radakovitz, C. Keplinger, Hydraulically amplified self-healing electrostatic actuators with muscle-like performance. Science 359.6371 (2018): 61-65.

[ii] N. Kellaris, V. Gopaluni-Venkata, G.M. Smith, S.K. Mitchell, C. Keplinger, Peano-HASEL actuators: Muscle-mimetic, electrohydraulic transducers that linearly contract on activation. Science Robotics 3.14 (2018): eaar3276.

[1] C. Keplinger, J.-Y. Sun, C. C. Foo, P. Rothemund, G. M. Whitesides, Stretchable, transparent, ionic conductors. Science 341, 984–987 (2013).

[2] J.Y. Sun*, C. Keplinger*, G.M. Whitesides, Z. Suo, Ionic Skin. Advanced Materials 26 (45), 7608-7614 (2014).

[3] Y. Bai, B. Chen, F. Xiang, J. Zhou, H. Wang, Z. Suo, Transparent hydrogel with enhanced water retention capacity by introducing highly hydratable salt. Appl. Phys. Lett. 105, 151903 (2014).

[4] H. Yuk, T. Zhang, G. A. Parada, X. Liu, X. Zhao, Skin-inspired hydrogel–elastomer hybrids with robust interfaces and functional microstructures. Nat. Commun. 7, 12028 (2016).

[5] R. Pelrine, R. D. Kornbluh, Q. Pei, J. Joseph, High-speed electrically actuated elastomers with strain greater than 100%. Science 287, 836–839 (2000)

[6] S. Rosset, H. R. Shea, Small, fast, and tough: Shrinking down integrated elastomer transducers. Appl. Phys. Rev. 3, 031105 (2016)

[7] Z. Suo, Theory of dielectric elastomers. Acta Mech. Sol. Sin. 23, 549–578 (2010)

[8] G. Kovacs, L. Düring, S. Michel, G. Terrasi, Stacked dielectric elastomer actuator for tensile force transmission. Sens. Actuators A 155, 299–307 (2009)

[9] F. Carpi, C. Salaris, and D. De Rossi, Folded dielectric elastomer actuators. Smart Materials and Structures 16.2 (2007): S300

[10] J. A. Koh, C. Keplinger, R. Kaltseis, C.-C. Foo, R. Baumgartner, S. Bauer, Z. Suo, High-performance electromechanical transduction using laterally-constrained dielectric elastomers part I: Actuation processes. Journal of the Mechanics and Physics of Solids 105, 81-94 (2017).

[11] C. Keplinger, M. Kaltenbrunner, N. Arnold, S. Bauer, Capacitive extensometry for transient strain analysis of dielectric elastomer actuators. Appl. Phys. Lett. 92, 192903 (2008).

[12] F. Ilievski, A. D. Mazzeo, R. F. Shepherd, X. Chen, G. M. Whitesides, Soft robotics for chemists. Angew. Chem. Int. Ed. 123, 1930–1935 (2011).

[13] P. Polygerinos, N. Correll, S. A. Morin, B. Mosadegh, C. D. Onal, K. Petersen, M. Cianchetti, M. T. Tolley, R. F. Shepherd, Soft robotics: Review of fluid-driven intrinsically soft devices; manufacturing, sensing, control, and applications in human-robot interaction. Adv. Eng. Mater. 19, 1700016 (2017)

[14] C. Keplinger, T. Li, R. Baumgartner, Z. Suo, S. Bauer, Harnessing snap-through instability in soft dielectrics to achieve giant voltage-triggered deformation. Soft Matter 8, 285–288 (2012). doi:10.1039/C1SM06736B

[15] J. T. B. Overvelde, T. Kloek, J. J. A. D’haen, K. Bertoldi, Amplifying the response of soft actuators by harnessing snap-through instabilities. Proceedings of the National Academy of Sciences 112.35 (2015): 10863-10868

[16] R. Niiyama, D. Rus, S. Kim, Pouch motors: Printable/inflatable soft actuators for robotics, in Proceedings of the 2014 IEEE International Conference on Robotics and Automation (ICRA) (IEEE, 2014), pp. 6332–6337.

[17] S. Sanan, P. S. Lynn, S. T. Griffith, Pneumatic torsional actuators for inflatable robots. J. Mech. Robot. 6, 031003 (2014).

[18] A. J. Veale, I. A. Anderson, S. Q. Xie, The smart Peano fluidic muscle: A low profile flexible orthosis actuator that feels pain. Proc. SPIE 9435, 94351V (2015).

[19] C.H. Li, C. Wang, C. Keplinger, J.L. Zuo, L. Jin, Y. Sun, P. Zheng, Y. Cao, F. Lissel, C. Linder, X.Z. You, A highly stretchable autonomous self-healing elastomer. Nature chemistry, 8(6), (2016), p.618.

[20] Y. Cao, T. G. Morrissey, E. Acome, S. I. Allec, B. M. Wong, C. Keplinger, C. Wang, A transparent, self-healing, highly stretchable ionic conductor. Advanced Materials 29, 1605099 (2017).



Zhigang Suo's picture

Dear Christoph:  Fantastic work! Heartiest congratulations on the publication of two spatacular papers!  

Several members in the group at Harvard and in China had elements of the idea, but have not pushed so far and so powerfully as you have done.  We had a name for an ongoing outline:  GEO, standing for Gel, Elastomer, and Oil. 

Here is a paper titled organic liquid-crystal devices based on ionic conductors.  We encountered the combination of the three materials, GEO, in our experimental setup, commented on some interesting attributes more basic than the liquid crystal devices, and did a rudimentary theoretical analysis in the supplementary information (Fig. S4).

The students here at Harvard and in China are all your admirers.  Me, too.  May the field that you help initiate engage more researchers, inspire deep scientific questions, and open new applications. 

A technical question.  Have you found good reference on self-healing of transformer oil?  This is really a neat idea.

Christoph Keplinger's picture

Dear Zhigang:  Thank you for your positive comment -- this means a lot to me and my students!


Our work on HASEL artificial muscles synergizes different areas I have studied during my PhD and postdoc, and our papers benefit from all the things I have learned from my academic advisors. My ambitious aim when starting my own lab in Boulder was to invent a new class of artificial muscles that enables a next-generation of soft robots that are high-speed, versatile and electrically powered, thereby avoiding the issues that arise from the use of compressors and valves in soft pneumatic actuators:

1) I have started to appreciate the wide range of possible geometries of soft fluidic actuators when working as a postdoc with George Whitesides at Harvard. Using liquid dielectrics as a hydraulic fluid, HASEL actuators inherit the versatility of soft fluidic actuators.

2) Even before I started my PhD with Siegfried Bauer, he introduced me to the fascinating world of dielectric elastomer actuators (DEAs). I was immediately convinced that this type of actuator is extremely promising, when seeing the speed, efficiency and ability to capacitively self-sense deformation. Just like dielectric elastomer actuators, HASEL artificial muscles are driven by Maxwell stress, and thus inherit the high-performance and the capacitive self-sensing abilities of DEAs. In contrast to DEAs, HASELs use a liquid dielectric, which self-heals after dielectric breakdown, and thus avoid central issues of DEAs, such as catastrophic dielectric breakdown and electrical aging. For this exact reason, industry widely uses liquid dielectrics in HV transformers.

3) Zhigang, when working together with you as a postdoc, I have started to realize that the nonlinearities and instabilities in soft active materials are a feature, and not a problem. Donut HASELs undergo a safe electromechanical instability to reach large deformations. I am absolutely certain that the mechanics community will very soon find new types of electromechanical instabilities in various designs of HASEL actuators. Additionally, HASEL actuators benefit from the unique properties of stretchable, transparent ionic conductors, a topic we introduced together in 2013.


GEO, standing for Gel, Oil, and Elastomer is without a doubt an incredibly interesting combination of materials, that might have applications far beyond the field of soft actuators. I am looking forward to seeing what the Suo group comes up with in this area! My lab has been working on HASEL actuators since more than two years, and we had central results ready more than a year ago. It is always a gamble when aiming to publish in top journals such as Science or Nature, as these outlets require top quality stories, figures, videos and writing -- all of which takes a lot of time and increases the chances that other groups publish the idea first. I am incredibly proud of my team of first-rate students, who pushed the initial papers introducing HASELs over the finish line with creativity, dedication and hard work. 4 of them -- Eric Acome, Nicholas Kellaris, Timothy Morrissey, Shane K. Mitchell -- have helped me put together the journal club post and will be discussing with us here.

When thinking about industrial applications of HASEL artificial muscles, we realized that the fundamental principles do not depend on elastomers and conductive gels. In our Science Robotics paper on Peano-HASEL actuators, we introduce a materials system with thin polymer films and evaporated metal layers. This materials system is amendable to large scale industrial fabrication, and it does not depend on highly stretchable elastomers or conductive gels. Depending on the specific application, HASELs can be built from a wide variety of different materials.


Zhigang, thank you for your question on liquid dielectrics. While being widely used in the high voltage industry, there is a surprisingly small number of academic papers available on the topic of self-healing in liquid dielectrics. A while back I wanted to better understand the fundamentals of dielectric breakdown and looked for good literature; initially I focused on solid dielectrics, but then found work from the Research Laboratory of Electronics (RLE) at MIT that opened my eyes about the benefits of liquid dielectrics and the rich physics of electrical breakdown: "Mechanisms Behind Positive Streamers and Their Distinct Propagation Modes in Transformer Oil" by Professor Markus Zahn and colleagues. Here is a link:



HASEL actuators solve important issues, but they open up just as many new questions, challenges and opportunities. In our journal club post, we mention a few areas we look forward to working on together with the mechanics, materials, and robotics communities!


Xuanhe Zhao's picture

Hi Christoph,

Congratulations on these two stellar papers! I am glad to see the field of dielectric elastomer actuators maintains so vibrant after almost two decades, thanks to innovative ideas/works from colleagues like you, Siegfried and Zhigang. Two quick questions, in addition to the on-going discussion.

1. Technical one. What is the energy efficiency of the HASEL acutator? While fluid is involved in the actuation process, I guess the energy efficiency of the actuator may be still quite high, comparable to viscoelastic DEAs.

2. Practical one. High-voltage can be a concern, as you make the power density of the actuator higher. What will be the field of targeted applications of HASEL acutators? 

BTW, thanks for referring to our works on harnessing instabilities in soft materials and tough bonding of hydrogels. 

Christoph Keplinger's picture

Hi Xuanhe,

thank you very much for your positive comments!

I am also very excited that the field of soft active materials is so active and productive now. I think it is important to note that HASEL actuators significantly draw from the advances in soft fluidic actuators. In particular, the Peano-HASEL actuators directly build on the great papers on Peano fluidic actuators [16-18]. Also, in contrast to DEAs, HASEL actuators do not have to rely on elastomers (Peano-HASELs use a thin flexible polymer shell), and the fundamental principles of actuation do not work without the use of hydraulic liquids. Therefore, I think HASEL actuators are best described to synergize the activation mechanism of dielectric elastomers (Maxwell stress) with the versatility and ease of fabrication of soft fluidic actuators, and thus form a new direction of research for artificial muscles.

I am answering your questions below:

1) At the moment we measure energy efficiency of around 20-30% for HASEL actuators (see figure S4 of the Science paper, where we report a cycle with 21%, for an actuator lifting a weight of 100g for >>10% strain). It is important to note that this number is based on a rigorous analysis of electromechanical conversion efficiency of a full cycle in work conjugate planes of force-displacement and voltage-charge (a lot of work in the field of artificial muscles does not consider full cycles when stating efficiency, and it is thus very important to compare numbers with care). The measured efficiency of HASEL actuators is comparable to typical experimental values for DE actuators -- whereas DE actuators have potentially high efficiencies (based on estimates up to 80%) experimentally measured efficiency ranges from 10 to 30%.

A number that is even more important for untethered operation of soft robots is system efficiency (this includes everything coming from the main energy source (such as battery pack), through amplifiers and switches, and finally resulting in mechanical work). In an upcoming paper, we analyze system efficiency of HASEL actuators that are driven by miniature HV electronics. The results will be very important for some practical applications.

2) Yes, high voltage could potentially be a concern for practical applications that are not designed well. I am personally not too worried about the safety aspect; after all, we constantly carry lithium ion batteries with us that can explode -- I think the risk with well insulated and shielded HV actuators is lower than that. As far as practical applications are concerned, we have been very positively surprised by the strong reaction to our papers from a range of different industries, and from companies representing several hundred thousand employees. We are already funded by a large car manufacturer to explore active surfaces based on HASELs, which could find use both inside and outside of the car. Peano-HASELs are attractive for robotic applications, due to their ability to linearly contract upon application of voltage without relying on stacked configurations or prestretch. Additionally, HASELs will find use in various types of haptic interfaces, life-like prosthesis, valves, pumps, vibration control, positioning systems and basically everything else that requires silent, efficient, lightweight, soft, inexpensive and high-speed actuators.

Yes, we made direct use of your invention in the area of tough bonding of hydrogels -- It feels great to build upon groudbreaking work from friends!

Thank you!



Jiawei Yang's picture

Really impressive work! Christoph. I am very enjoying reading, but still have some questions.

1. When the actuator lift weights, the part composed of elastomer membrane and soft electrode is stretchable, does that compromise the actuation?

2.The voltage applied determines how heavy the weight can lift. But for lifting heavy weight, your voltage is very high, does that hinder the practical application? 

3. When the elastomer membrane contacts, does the high voltage cause electric breakdown of the elastomer?

4. What is the maximum liquid you can put in the actuator? Is there any limitation? 

Thank you, and again really appreciate your idea!

Shane.K.Mitchell's picture

Dear Jiawei,

Thank you for your interest in our work and your stimulating questions. I have answered them below:

1. When the actuator lift weights, the part composed of elastomer membrane and soft electrode is stretchable, does that compromise the actuation?

Answer: In the case of the donut actuators, a very soft membrane can hinder actuation under heavy loads, as the liquid dielectric tends to bulge the side walls of the soft shell instead of translating the force upwards to lift the weight. For this reason, we used a stiffer elastomer, namely PDMS, instead of other softer silicone based products, such as Ecoflex.

2.The voltage applied determines how heavy the weight can lift. But for lifting heavy weight, your voltage is very high, does that hinder the practical application?

Answer: It is important to note that HASEL actuators are not strictly voltage dependent. The electrostatic stress generated within the shell is proportional to the electric field squared and dielectric constant of the solid-liquid composite dielectric. In our initial work, we used relatively thick membranes (on the order of 1mm) to construct the actuators, as these membranes were easier to handle and sufficient for proof of concept work. These thick membranes resulted in the need for rather high voltages to generate the necessary electric fields to trigger actuation. The voltages used in our initial work are indeed too high for practical applications; however, some strategies for reducing the operating voltage include increasing the dielectric constant of the composite structure or decreasing the distance between the electrodes by utilizing thinner membranes for the shell. Currently, we have designs which operate at 10 times less voltage while still generating high forces, making the electronics practical and the actuators extremely useful!

3. When the elastomer membrane contacts, does the high voltage cause electric breakdown of the elastomer?

Answer: Typical operation of HASEL actuators does not cause breakdown of the composite dielectric structure. That is, we primarily do not use electric fields which are higher than the dielectric strength of the composite dielectric (though, if we did, the actuators could self-heal from the breakdown). There is indeed a large change in the electric field after a pull-in transition; however, the maximum field is still below the dielectric strength of the composite and so breakdown does not occur. Additionally, there is always a layer of solid dielectric between the two electrodes, which prevents the electrodes from shorting during operation.

4. What is the maximum liquid you can put in the actuator? Is there any limitation? 

Answer: The content of liquid dielectric is far from optimized. We tested a few different oil contents to roughly determine what worked well for our demonstrations, but did not perform a rigorous analysis to optimize the fill amount. It’s important to note that as the amount of liquid inside the elastomer shell is increased, the distance between the electrodes in the relaxed state also increases. Therefore, the voltage required to generate the electric field necessary to trigger actuation also increases drastically. If the applied voltage is held constant throughout the pull-in transition, the electric field at the end of the transition may exceed the dielectric strength of the material and result in dielectric breakdown.



canhui yang's picture

Hi, Christoph, Thanks for such a wonderful and inspiring work.

Naturally, muscle contains water and can self-heal. Previous artificial muscle (DEA) mimics functions of muscle, but suffers from electric breakdown. Whereas synthesizing self-healing dielectric elastomer could be an awkward alternative for common mechanical engineers, using liquid dielectric to engender self-healing ability for artificial muscle is simple yet effective. This is really a brilliant work.

The opportunities and challenges section is already very comprehensive, and will lead to many follow-up researches across multiple disciplines.

One particular issue about the practical usage of this device: the hydrogel electrodes are exposed. This may cause several challenges: 1. Whereas hydrogels can retain water with dissolved hygroscopic salts, the water content fluctuates with the ambient humidity. Does the fluctuation of water content (i.e. resistance) of hydrogels affect the device performance? 2. The hydrogel electrodes may contact with other materials, and short circuit might happen if the materials are conductive. 3. Using hydrogel makes the overall device transparent. But the device cannot work in water environment.  Have you found potential solution for this issue?


Thank you.

Timothy Morrissey's picture

Hello Canhui,

We are thrilled you appreciate the "simple yet effective" solutions we have presented with HASEL. We also hope our opportunities and challenges section leads to some other expert community members of iMechanica in joining us in pushing the boundaries of HASEL. 

Your questions about the hydrogel are very important to us. I answer them directly below

1. Whereas hydrogels can retain water with dissolved hygroscopic salts, the water content fluctuates with the ambient humidity. Does the fluctuation of water content (i.e. resistance) of hydrogels affect the device performance? 

The short answer, yes ambient humidity can influence the properties of the hydrogel. Ambient atmosphere is especially important in climates such as ours, in Boulder, Colorado where we have very dry air, especially in the winter months. We found that in most atmospheric conditions the hydrogel would remain hydrated for hours and there was no significant influence on the performance of the device. However, in the drier winter months, it is true our hydrogels may dry out which may increase the resistance of the electrode. However, since our HASEL devices currently use high voltage, the changes in resistance has minimal influence on actuation performance. We have not done a comprehensive study on atmospheric conditions, hydrogel resistance, and actuator performance. That may be very useful for advanced lifetime test.

Also, we are very fortunate to "stand on the shoulders of giants" and follow the hydrogel recipe developed by Y. Bai. et. al. in Professor Suo's research lab.  These PAM LiCl hydrogels have very good water retention properties. Do you know of other hydrogel compositions that we should consider for enhanced water retention? 

Lastly, in some cases, we encapsulated the hydrogel in a thin layer of ecoflex. The ecoflex was applied to the hydrogel via a spin coating process. This thin layer of ecoflex helps keep water in the hydrogel and does not significantly influence mechanical properties of the actuator. Still, ecoflex is somewhat permeable to water vapor and while improving water retention of the hydrogel electrode, it does not completely solve all hydrogel hydration issues.  The supplemental material section of the Science paper clearly discusses which systems are encapsulated and how.

2. The hydrogel electrodes may contact with other materials, and short circuit might happen if the materials are conductive. 

This is true, if the electrodes came in contact with conductive material this would cause a short circuit. In real-world application we feel this is not a common situation. Additionally, when the device is encapsulated with a thin layer of ecoflex or Kapton film, this insulating layer will make shorting the electrodes even less likely. 

Along the lines of electrodes contacting other materials, it may be possible that the HASEL device is in an environment where the electrodes are physically damaged (cut, scratched, etc.) We think it may be very interesting to incorporate self-healing conductive materials such as work we published in Advanced Materials last year (ref 20 above, 

3. Using hydrogel makes the overall device transparent. But the device cannot work in water environment.  Have you found potential solution for this issue?

Again, this is true, that this device may not work in a water environment as the hydrogels would be influenced by the surrounding water. For an application such as this we can speculate a few possible solutions. 

The first possible solution may be to look for other transparent ionically conductive electrodes such as ionogels. Again, the self-healing material we collaborated with Chao Wang on (ref 20 above) might be one possible material, plus this material has the added benefit of physical self-healing. This material used ion-dipole interaction to enable both self-healing and to keep the ionic liquid in the polymer matrix. The transparency of this material was influenced by atmospheric water vapor so this self-healing material may not be the perfect electrode for a water environment, but might point you in the right direction.

A second possible solution could again be the encapsulation route. Simply sealing the hydrogel inside a polymer that is not permeable to water might allow the device to operate in a water environment. 

Lastly, it is highly possible instead of using ionic conductors, to operate HASELs with electronic conductors. Peano-HASEL has already been demonstrated with electrical conductors. It is, of course, unlikely this device would still be transparent using metallic layers

Do you have any other ideas for this? What specific application did you have in mind, sounds interesting!

Thanks again for your comments. Please let us know what you think of our ideas to address them and reach out with any new questions.

Amit Pandey's picture


great job


Xiaoyan Li's picture

Hi Christoph, it is an amazing work!

Two years ago, I and my collaborators (Prof. Hujian Gao at Brown and Prof. Hui Wu at Tsinghua) built an electrochemical actuator based on a rechargeable battery of LiFePO4 cathode and Si anode, by taking advantage of the giant volume expansion in Si anode microparticles after full lithiation. Our electrochemical actuator of the LiFePO4||Si battery can drive a high load greater than 10 MPa with a device response time less than 1 second. The driven voltage of the device is less than 4 V, which is two order-of-magnitude lower than that of piezoelectric materials. Such actuator might be used for the robotics by the clever design. The relevant paper is titled by Cycling of a Lithium-Ion Battery with a Silicon Anode Drives Large Mechanical Actuation.

Your HASEL actuator couples the hydraulic and electrostatic force, and exhibits the muscle-lie performance. I am curious on two aspects of your HASEL actuators: (1) whether/how the environmental factors (such as temperature and humidity) affects the performance and behavior of HEASEL actuator? (2) how about the fatigue behaviors (especially reliability of actuators under high-cycle fatigue) of HASEL actuator? Thank you very much in advance!

Eric Acome's picture

Hi Xiaoyan, 

Thank you for your kind comments and thoughtful questions about our work.

Your research on electrochemical actuators is very interesting. That is a clever way to turn what is usually a problem – volume change from lithium insertion and extraction – into something beneficial! The ability to drive >10 MPa at 4 V is very impressive, especially with 1 second response time. I am curious to know how these perform after a high number of cycles.

We appreciate your questions about operation in different environments and fatigue life. Below are my answers to your questions.

1)    Whether/how environmental factors (such as temperature and humidity) affects the performance and behavior of HASEL actuators?

In our case, humidity can be an important environmental factor. We have used ionic conductors, namely PAM hydrogel swollen with LiCl, (based on reference 3 above) as our electrodes. While LiCl is hygroscopic, below a certain threshold (~10% relative humidity) water loss of the hydrogel becomes significant. Colorado is a dry climate, yet we only have issues with ionic conductors drying out during the winter months. At higher humidity levels we have not observed any change in performance of the actuators. Some HASEL actuators do not require stretchable conductors and as a result flexible electronic conductors can be used.

We have not investigated the effect of temperature on our devices, but this could be very important to consider for certain applications. We know that properties such as viscosity of the liquid dielectric will depend on temperature. The liquid dielectric we used (Envirotemp FR3) has a pour point of -20 deg C, which means below this temperature the liquid does not flow, which would be detrimental for actuation performance. Understanding the limitations of current HASEL actuator materials and determining materials for extreme conditions will be important to consider in some applications.

2)    How about the fatigue behaviors (especially reliability of actuators under high-cycle fatigue) of HASEL actuator?

We have performed some initial tests of life cycle for HASEL actuators. Donut HASEL actuators were tested for more than a million cycles under a load of 150 g. Under these conditions actuation strain was 15% and we noticed no change in performance after a million cycles. We ended the test for sake of time and not because the device failed. It would be interesting to perform longer term tests for this type of HASEL actuator to determine the maximum number for life cycle.

We also tested fatigue behavior for two other designs of HASEL actuators. A planar HASEL actuator, which is made of silicone elastomer and pre-stretched onto a rigid frame, failed after 158,000 cycles from mechanical rupture. Peano HASEL actuators, made from heat-sealed biaxially-oriented polypropylene failed after 20,000 cycles due to electrical breakdown through the heat-seal. Additional experimental testing and a better fundamental understanding of fatigue of soft and flexible materials would be useful for improving the long-term performance of HASEL actuators.

Thank you again for your comment and questions!

Eric Acome

Xiaoyan Li's picture

Hi Eric,

Thank you very much for your detailed replies.

The full lithiation in silicon leads to a huge volume expansion of 300%, which mechanically degrades, fractures and even pulverizes the anodes. After full lithiation, the lithium concertation x of LixSi reaches up to 4.4. In our work, we just partially inserted the lithium till the maximum lithium concentration is about 0.8. At that time, the volume expansion is only about 30%, which avoids the mechanical degradation due to large volume change from high lithium concertation. We tested the cyclic performance of our lithium-battery-based actuator by repeating the lithiation and de-lithiation with maximum lithium concentration of 0.8, and found that this actuator worked very well after a 100,000 cycles.

You work about HASEL actuator is very impressive and inspiring, and would have a significant impact on development of artificial muscles and soft robotics. It would be greatly expected that your novel HASEL actuator will be widely applied for the robotics in near future. Thank you very much again.

Paul Le Floch's picture

Hello Christpoh,

The work of your group is very inspiring, and open up new possibilties for soft actuators. This is a breath of fresh air for the field!

I would like to come back on the recent comments of Timothy and Eric about the durability of the device in ambient atmosphere, and in water environment. Our group recently published a paper about wearable and washable hydrogels, which could be useful to design HASEL actuators with enhanced durability: Wearable and Washable Conductors for Active Textiles

We notably show that combining the effect of an hydroscopic salt and an elastomer coating can be a solution to create hydrogel devices that doesn't dry out in ambient air. Also, I would like to draw your attention to the fact that silicone elastomers have extremely high permeabilities toward water and oxygen. Ecoflex is convenient for proof-of-concept devices, but there are other rubbers, such as butyl rubber (a crosslinked polyisobutylene), which have a permeability that is about a hundred times lower than Ecoflex. This material is used to make air-tight chamber, and innerliner of tires. It is actually known to be the elastomer with the lowest permeabilities.

In our paper, we also show that the diffusion of salt (NaCl) from a hydrogel to DI water can be slowed down quantitatively with the presence of a thin butyl rubber coating. Using this material as a coating, HASEL actuators could probably work in water environment for some time.

Thank you for your amazing work! 

Shane.K.Mitchell's picture

Hello Paul,

Thank you for your comment! Your work is fascinating and will surely push wearable ionics to the next level of feasibility. I really like the idea of coating the electrodes in a butyl rubber for protection from various environmental factors. As Tim and Eric have mentioned, the air is extremely dry here in CO, so our hydrogels don’t stay hydrated for long in ambient conditions. When our actuators are not in use, we store them in a makeshift humidity chamber made from an old fish tank sealed with a garbage bag and filled with some cups of water. Frugal science at its best.

On a slightly different but related note, we have noticed some interesting phenomena pertaining to the swelling and permeability of the silicone shells. Initially we tried to use a silicone-based transformer oil as the liquid dielectric, but noticed that the oil would swell and warp the elastomer rather quickly (over the course of a few hours). Eventually we settled on a vegetable based transformer oil (Envirotemp FR3) which works quite well. However, over time frames between weeks and months we noticed that this liquid dielectric begins to permeate through the silicone membrane, causing the actuators to appear to ‘sweat’. This sweating can affect performance if a significant amount of liquid seeps out of the actuator.

One of the most fascinating facts about HASEL actuators is that their structure lends itself to a wide range of materials. It would be very interesting to make the shell from a butyl rubber; then the entire actuator could function while submerged in water. I don’t know much about the dielectric properties of butyl rubbers, nor how they interact with liquid dielectrics; however, the liquid dielectric can be tailored to fit the requirements imposed by the shell material.

Thanks again for your comment!

Jiawei Yang's picture

Hello Christpoh and Shane,

Thank you for your previous explanation. I would like to bring up one more question about the adhesion between hydrogel electrode and the elastomeric shell. Have you measured the adhesion energy? Do you see any debonding during actuation due to the large deformation of the shell? In particular, during the repetitive actuation, this bonding interface undergoes cyclic deformation, fatigue of bonding may be potentially an issue. Do you have idea how to overcome this and ensure a long-term reliability? Thank you again!

Eric Acome's picture

Hi Jiawei

Thanks for another great question!

Creating strong adhesion between the elastomer shell and hydrogels was one of the first challenges we encountered for HASEL actuators. This was particularly important for the planar HASEL actuators (Fig. 2 in this blog post). We were able to utilize the method presented by Yuk, et al. to covalently bond the elastomer and hydrogel. This technique worked very well for the planar HASEL actuators and we didn’t have issues with hydrogel de-bonding. However, there are other instances where adhesion between materials of HASEL actuators could be improved.

First, after bonding hydrogels to the elastomer shell, we encapsulate the hydrogel with a thin layer of Ecoflex (silicone elastomer). This was achieved by spin-coating uncured Ecoflex over the hydrogel electrodes. This layer was not very robust and in some cases it would peel away to expose the hydrogel. One way to improve this would be to incorporate methods like the one mentioned above by Paul Le Floch (Wearable and washable conductors for active textiles)

In the case of the donut (Fig. 1 of this post) and Peano HASEL (Fig. 3) actuators, we simply rely on the ‘stickiness’ of the hydrogels to keep the electrodes on the elastomer or polymer shell. This is sufficient because the electrodes only need to be flexible. However, strong adhesion between the electrode and shell materials would be beneficial for improving fabrication and long term reliability.

Quick methods for bonding hydrogels to surfaces of elastomers and polymers would be very useful for constructing HASEL actuators. The method recently presented by Wirthl et al. for instant bonding of hydrogels could reduce fabrication time. I am less aware of work focusing on bonding hydrogels and polymers. Since the substrate (polymer) is flexible yet not stretchable, there may not be much need for research on cyclic deformation of these material systems. Regardless, a method for obtaining strong adhesion between hydrogels and polymers would be useful for Peano HASEL actuators (Fig. 3) which are made from biaxially oriented polypropylene.

In the future we are excited to see developments in adhesion between different soft and flexible materials. For example, a recent paper by Taylor et al. investigates a simple method for bonding elastomers and thermoplastics. Research in this area broadens the range of useful materials and fabrication methods for HASEL actuators and other soft robotic technologies.

Are there other methods for adhering different materials which you think would be useful to look into? Maybe different materials systems besides hydrogel conductors and elastomers would be more convenient and have other advantages for HASEL actuators?


Eric Acome

Jiawei Yang's picture

Thank you, Eric! I realize the significance of adhesion in the actuators, and the adhesion will be vital for either effectve actuation or long time use. There will be many oppotunities to develop diverse adhesion methods for different materials in different applications.

Adhesion of hydrogel and other materials is relatively new, and the methods are pretty much you mentioned. one more paper is by Cha, et al.

For other material systems, there are many commercial products, and also a large body of literature available. Please see some provided in this list.

Looking forward to seeing the next generation of HASEL.




Ruobing Bai's picture

Dear Christoph and all authors,

Thank you for sharing this facile and strong work. As creative and productive as you have always been. In addition, it is really enjoyable to read your little summary of progresses in soft materials, and future challenges.

My colleages and I have recently been quite interested in the mechanical properties and behaviors of soft materials under prolonged loads, i.e. fatigue. Here are several examples of different fatigue behaviors we encountered.

1. Under a static load, a pre-cut hydrogel can sustain the load for a long time, but then suddenly fracture completely. [1]

2. under cyclic loads, a hydrogel is very susceptible to pre-existing flaws, and can fracture gradually by a mechanical load much smaller compared to the critical load to cause catastrophic fracture [1, 2, 3].

3. Under cyclic loads, the material property, such as stress-stretch behavior, is not affected in some soft materials (e.g. single covalently crosslinked PAAm hydrogel), but is dramatically different with loading cycles in some other soft materials (e.g. double-network hydrogels). [2]

4. Also under cyclic loads or prolonged large monotonic stretch, a plastic liquid (such as the carbon crease for DEA) on an elastomer (such as VHB) can form various types of instability patterns. [4, 5]

I can imagine all the above examples can take place somehow in specific kinds of soft robot designs. For example, fatigue under cyclic loads must be an important consideration for soft robots used as grips. However, we have seldom place our scientific studies on fatigue into practical engineering applications, like HASEL. I am wondering, what kinds of such fatigue behaviors have you encountered during your design and application of the HASEL devices? If you have, how did you resolve these issues? What are the still remaining challenges regarding fatigue in these devices, if HASEL is to be employed in broad, industrial-level applications?


Thank you again and best regards,


Nicholas Kellaris's picture

Hi Ruobing,

Thanks for the resources on fatigue behavior in these soft systems – particularly hydrogel! It would certainly be beneficial to investigate fatigue behavior in these actuators in more detail.

From what we have noticed, the sort of fatigue and crack propagation that is investigated in [1-3] does not occur in our actuators. In the highly-elastic planar HASEL actuator demonstrated in the first HASEL paper published in Science, we use the benzophenone treatment from Yuk et al. (ref 4) to bond the hydrogel to our elastomeric actuator shell. This reinforcement prevents any sort of crack formation or propagation. In the Peano-HASEL paper, we are using flexible but inextensible materials as the actuator shell, so the hydrogel does not experience any significant strain. The only mechanical failures we have observed occurred in hydrogels that had dried out and become brittle.

As you may have seen in the two HASEL papers, we conducted initial lifetime tests for the three types of actuators presented. The results underscore the fatigue present in these actuators. The donut HASEL actuators actuated for > 1 million cycles without degraded performance or failure (we stopped testing for the sake of time before actuator failure). The soft compliant material used and low material strain likely contributed to the very high lifetime. The planar HASEL actuators lasted for ~ 160,000 cycles before mechanical rupture. While these use a very elastic silicone polymer, the high loads (1 kg) used during testing exert large stresses on the material which lead to mechanical fatigue and failure. Finally, the Peano-HASEL actuators tested actuated ~ 20,000 cycles before failure. Here we use an inelastic system that depends on bending and buckling in a polypropylene shell for actuation. As you can imagine, this system is very susceptible to fatigue. In fact, during many of the lifetime tests, failure occurred through the heat seal in an area subject to repeated bending/buckling during actuation.

The work shown here was more about introducing HASELs as a new platform for soft actuator technology. Moving forward, there are many materials/geometries we could explore that would minimize the fatigue experienced. Speaking for Peano-HASELs, I think that using a more compliant material in areas that undergo significant deformation would limit this fatigue. That, and improved actuator geometries that spread deformation over a larger area rather than localized creasing and buckling.

Hopefully that gives you a little more context for the fatigue in our actuators. It’s certainly an important subject of research – I’m sure the soft robotics community will come up with some creative solutions!

Let us know if you have any more questions!

Tiefeng Li's picture

Dear Christoph:  Great work! Congratulations ! 

As many group members from Prof. Zhigang Suo, we have also work in soft robotics in Zhejiang University. We have tried quite alot in robots driven by dielectric elastomer with onboard power source (such as the Fast-moving soft electronic fish, 3, 4, Science Advances, 2017 The fully soft robotic fish can move and turns quickly. However, the electric breakdown problem remains as a great challenge on this type of robots for practical application. The idea that self healing in your paper will truly inspire the design for artificial muscles.  

technique questions:

1. the liquid is enveloped  in between of two elastomric membranes, when the high voltage is applied, will the elastomeric membranes also suffer breakdown? 

2. For our experience in dielectric drieven soft robots, the required high voltage is still quite a challenge for the life time and efficiency, for this novel desigh of actuator, do you have any idea to lower the actuating votlage in future?

thank you again for these  nice papers and ideas. recalls me the great time that we work together in Harvard,already 9 years ago,~ time flies.

Timothy Morrissey's picture

Hi Tiefeng.

Thank you very much for the congratulations. We are indeed familiar with your electronic fish work which we find very fascinating. I particularly enjoy the use of the liquid as an electrode in a DEA! To answer your two questions: 

1. the liquid is enveloped  in between of two elastomric membranes, when the high voltage is applied, will the elastomeric membranes also suffer breakdown? 

That is correct. The liquid is enveloped by an elastomeric shell which suffers damage during breakdown. The shell is shown as the white material in Fig 1A above as well as the grey material in Fig1F above. Please note in Fig 1F that after breakdown a small area of the elastomer shell is damaged. We found that the elastic behavior of the ecoflex and PDMS material allowed the damaged area to essentially reseal and keep the liquid inside the device even though there was now a small hole present. The use of elastic hydrogels on the outside also helps to ensure the liquid stays encapsulated even after breakdown.  Figure S14 of the Science paper discusses breakdown a bit more.

 Additionally, it is important we point out that the dielectric strengths of the liquid and shell material should be similar for full self-healing from electrical damage. In the case of Peano-HASEL which uses non-eleastic BOPP, we can apply high electric fields before any breakdown, since the BOPP has very high dielectric strength. After the first breakdown however, not only does the shell leak because the BOPP does not reseal, but there is a path of purely liquid dielectric between the electrodes. Since the liquid has a much lower breakdown strength than the BOPP, you can't apply the same high field as you once could before the initial breakdown. We think a use of a mechanically self-healing material such as we discussed in [19,20] and in our initial post above, might be very interesting for future work. 

2. For our experience in dielectric drieven soft robots, the required high voltage is still quite a challenge for the life time and efficiency, for this novel desigh of actuator, do you have any idea to lower the actuating votlage in future?

Yes, we completely agree the use of high voltage is still a challenge for these devices. As you are also aware, it is the electric field that dictates DEA and HASEL performance, rather than the total voltage. Using thinner materials will significantly reduce required voltage; we have used think and easy to fabricate layers for our proof of concept work in Science. We also hope to explore new materials that have different physical characteristics such as permittivity. We are also interested in exploring different classes of dielectric liquids which might help this cause. 

We are confident to have some of this work requiring much lower voltages ready for display at SPIE next month. If you are joining that conference you should certainly plan on checking out the EAP in action session.

[19] C.H. Li, C. Wang, C. Keplinger, J.L. Zuo, L. Jin, Y. Sun, P. Zheng, Y. Cao, F. Lissel, C. Linder, X.Z. You, A highly stretchable autonomous self-healing elastomer. Nature chemistry, 8(6), (2016), p.618.

[20] Y. Cao, T. G. Morrissey, E. Acome, S. I. Allec, B. M. Wong, C. Keplinger, C. Wang, A transparent, self-healing, highly stretchable ionic conductor. Advanced Materials 29, 1605099 (2017).

Dear Christoph and all authors,

Thanks for the inspireing work. Heartiest congratulations to you all.

As you mentioned in the post "Donut HASELs undergo a safe electromechanical instability to reach large deformations", my question is does the planar HASEL actuator also survive the electromechanical instability to obtain large deformation? Also, prestretch helps to eliminate electromechanical instability for dielectric elastomer actuator. I was wondering the reason you prestretch the planar HASEL actuator?

Thanks and regards

Qin Lei

Eric Acome's picture

Hi Qin Lei,

Thank you for your kind comment and questions!

Planar HASEL actuators operate in a way that is similar to dielectric elastomer actuators (DEAs). Electrodes cover the entire liquid dielectric region and when electric field is applied through the layers of dielectric, the device decreases in thickness and expands in area. In contrast to donut HASEL actuators which experience a safe electromechanical instability, planar HASEL actuators experience dielectric breakdown from similar electromechanical instabilities as DEAs. However, the liquid dielectric layer of HASEL actuators enables self-healing from dielectric breakdown.

Just as prestretch helps eliminate electromechanical instabilities in DEAs, prestretch heps prevent electromechanical instabilities in planar HASEL actuators and improves performance. A recent paper by Koh et al. presents a detailed theoretical analysis of performance for laterally constrained DEAs and we found that some of these principles also applied to planar HASEL actuators.

As Christoph alluded to above, we like to figure out how instabilities can be used advantageously. Many papers have taken this approach to achieve interesting and remarkable performance with fluidic actuators (Overvelde et al.) and DEAs (Keplinger et al.). We think that electrohydraulic coupling in HASEL actuators could enable a number of useful instabilities and nonlinear performance. The pull-in instability of donut HASEL actuators is just one simple example and we are looking forward to investigating more in the future.

We would love to hear the ideas and suggestions of the mechanics community on how we could take advantage of the structure and materials of HASEL to create some interesting soft actuators.

Thanks again for your interest!


Eric Acome

Dear Christoph:

Thank you very much for posting this fabulous and inspiring work in IMechanica, which enables us to learn more about HASEL.

Soft actuators can contribute significantly to soft robotics. It is great to learn HASEL can couple dielectric elastomers and fluidic actuators, and exhibit unique attributes (for example, self-healing). I may have the following questions, regarding the output force and the applied voltage.

1) In Fig. 1E, it seems that smaller electrodes can lead to a higher force. This force is induced by the Maxwell stress? Output forces can play an important role in robotic applications. Any suggestions for enhancing these forces (for example, use fluid of high dielectric constant)?

2) What is the density of liquid (which may affect the energy density of the actuator)?

3) Can HASEL be driven by lower voltage, with specific fluid? If the required voltage is lower than 1kV, the robot can be driven by onboard low-voltage battery, associated with a voltage amplifier of a very small volume. Consequently, the robot can achieve untethered design (which may significantly improve the robot’s movement and functionalities). 

Thanks again for the great work! Congratulations!




Shane.K.Mitchell's picture

Dear Jian,

Thank you for your kind words about our recent work. We are beyond grateful for the overwhelming support and encouragement we have been experiencing from the academic community. To answer your questions:

1.) The donut HASEL actuators seen in Fig. 1 follow Pascal’s hydraulic principles ( The Maxwell stress induces a hydraulic pressure within the shell, which is independent of the electrode size. This pressure acts on the shell, with the output force dictated by the surface area of the shell which is in contact with the load. Thus, more surface area in contact with the load during actuation means more force. We demonstrated that an actuator with a smaller electrode yields a larger force than an identical actuator with a larger electrode. However, the smaller electrodes move less fluid during actuation, and therefore this actuator achieved less strain. The actuator with the larger electrode is the exact opposite (lower force but greater strain).  By simply varying the size of the electrode, one can tune the performance of the actuator to suit a specific application.

2.) The density of the liquid dielectric we used in our papers is 0.96 g/cm3.

3.) HASELs are driven by Maxwell stress which is proportional to the electric field squared and the dielectric constant of the composite dielectric. Therefore, one could utilize a composite dielectric structure with very high permittivity to achieve high Maxwell stresses at lower voltages. An easier approach to lower the operating voltages would be to use thin shell membranes with high dielectric strength, since the Maxwell stress has a squared dependence on the field. Peano HASEL actuators do exactly that, and utilize a thin but flexible polymer shell.

Lowering the operating voltage below 1kV is certainly a goal of ours. However, I would argue that with the material system utilized in the Peano HASEL paper, untethered operation is already possible. There are a few companies (EMCO and Pico Electronics) which specialize in miniature DC-DC high voltage amplifiers which can be powered by a cell phone battery and output 10kV. We currently have multiple untethered demonstrations which we will be showcasing at SPIE in Denver CO in the beginning of March.


With that being said, lowering the voltages even more is a necessity to further decrease the complexity of the electronics. This decrease in complexity will lead to even smaller electronic packages that can switch at very high speeds and are extremely lightweight.


KevinTian's picture

Greetings Christoph (and Co.),

Great work! This certainly advances many possibilities with soft-robotics, so I'm excited to see where this work leads.  A couple of notes/questions I had:

1. Regarding the air-bubbles: since you're using PDMS, I'm surprised that the gas doesn't simply permeate through the elastomer

2. Regarding Water-retention: Paul already discussed this above, but I wanted to add that LiCl itself is not the most hygroscopic salt out there, so if water-retention is all that you're gunning for you can opt to try other salts with even lower equilibrium RH (see Greenspan).  

3. A few folks have asked about breakdown, but I'm curious (especially since it's a square-wave/high-voltage signal) whether you've noticed any chemical reactions related to the hydrogel electrodes?

4. The fabrication process you've described seems closest to a hot-embossing of polymer films (as least for the peano-HASEL), which certainly would make it very amenable to scaled up fabrication.  I was curious about which advanced fabrication methods you thought were most applicable/promising to your system, as the need for a cavity to fill with the liquid dielectric does complicate the process.  Implementing 



KevinTian's picture

I cannot seem to edit my above posting, so I apologize as it was somewhat incomplete.  On the matter of the air-bubbles, having re-read the paper it's a bit more apparent why the issue arises in your design, so do disregard that comment.  Though having an elastomeric sealing composed of just PDMS should allow for diffusion of gas through the membrane (although as noted previous, it also allows for quite a few others to diffuse through it too).  

On the matter of fabrication I had intended to mention that in extrusion printing, I could foresee the usage of something like Jennifer Lewis' fugitive inks to print cavities in an elastomeric material that could then be filled post-print.  The capability to freely selectively bond materials (beyond using a filler material to avoid it, like the injection of air as you have actively done so) doesn't seem present in the field just yet.



vidyacharan's picture

Hello Kevin,


Thank you for your insightful comments and suggestions.


The motivation behind this work was to present a new idea that can push the limits to soft actuator technology and open up more possibilities of what we can do. Hence, we resorted to using readily available materials to demonstrate the concepts.


  1. Yes, the PDMS should be permeable to gases. However, it is a relatively slow process. Given enough time, the gas generated during a breakdown should be able to permeate outside.
    New materials that allow fast permeation of gases will improve the performance and reliability following dielectric breakdowns, since the idea is to have the actuator up and running immediately following any breakdown event, without temporarily compromising on the dielectric strength until the gas diffuses out of the system.

  2. That is a great suggestion on hygroscopic salts! For this paper, we used materials that we had prior knowledge about and immediate access to. Hence we used LiCl as it was shown to be more hygroscopic and a good fit for use in conductive hydrogels, as opposed to NaCl or other common salts.

    For future work, we will take cues from your suggested reference and experiment with other salts that can help to prevent drying of hydrogels in low-humidity environments.

  3. In the Peano-HASEL design, we use strips of copper tape as an intermediary to connect the power source to the hydrogel electrodes. At the interface of the copper and hydrogel, we did notice some chemical reaction forming a tiny bit of salt deposit. (green colored deposits, mostly Copper Chloride formed by Cu reacting with the LiCl). However, it did not affect the performance or conductivity of the electrodes. Using a less reactive material system might be able to prevent this from happening.

The Peano-HASEL actuators benefit from the use of an inextensible material as the shell, giving us a wide material selection, most of which are heat-sealable thermoplastics. This makes it simple to heat-seal the pouches, similar to hot embossing, followed by filling with a liquid dielectric. This should be a very simple process to achieve in an industry. That was the motivation behind using heat-sealable BOPP sheets and demonstrating the ease of fabrication.

The planar HASEL actuators on the other hand need an elastomeric system. They could be fabricated from thermoplastic elastomers such as Thermoplastic Polyurethanes. This would potentially make the fabrication process as simple as of Peano-HASELs. Developing thermoplastic elastomers with performance comparable to silicone rubbers would be great for making HASELs.

3D printing is a wonderful method to fabricate complex geometries of HASELs. Like you have noted, we need to use a filler material that can be purged out to create the cavity needed for the liquid dielectric. The paper from Jennifer Lewis’ group on using fugitive inks presents a very simple and scalable method of doing this. Thank you for sharing that. Most 3D systems cannot print multiple materials, and have to use support structures made of the same material while printing with cavities, which remain inside permanently. This could compromise on performance especially in the case of HASELs which take advantage of a completely liquid dielectric. However, with multi-material printers and such interesting solids that are water or alcohol-soluble, we could have support material that can be purged out completely and have clean cavities to fill with dielectric.

Are you aware of any commercially available printable materials that are water soluble? This group for example has presented an ingenious idea using melted sugar as a soluble support material. (

In the last 1-2yrs, there has been a rise in 3D printers capable of working with silicones and gels. So we can soon expect to see developments in rapid prototyping of these actuators, especially with interesting geometries.'s picture

Dear Christoph,

Congrulations on your excellent work. As Zhigang mentioned, you are our admire to push the idea so far!

A few students of Zhigang in China are working on the combination of gel, elastomer and liquid. I'm also involved. Zhigang gave it a beautiful name of GEO. We have been playing with it for quite a while.  I have  two questions:

1. When you add oil in, the advantage you presented is the self-healing property. The idea is really neat. I'm curious that becasue oil can take a much higher value of permittivity than elastomer, so are you able to output a much higher force using oil compared with tranditional dielectric elastomers, such as VHB?

2. We have done some analysis about the pull-in in GEO, although our configuration is not quite similar with yours. So do you think pull-in is an essential feature in your design and why? We mechanics peaple are quite interested at this.



Christoph Keplinger's picture

Dear Tongqing,

thank you very much for your positive comment!

Yes, Zhigang mentioned some still unpublished efforts on GEO in an earlier comment; I am curious to see this when the papers go online. GEO is without doubt a great combination of materials that is exceptionally rich in mechanics and materials innovation. As for GEO, I would suggest that you maybe consider making your analysis even broader. We have had some very nice results with Peano-HASELs, which do not rely on elastomers and can also be made with thin metal films as electrodes -- and still, Peano-HASELs can serve as very good electrohydraulic artificial muscles. Maybe some of the mechanics concepts you are working on related to GEO could be generalized and made broader to also apply to systems, that consist of a liquid dielectric, thin polymer films (including but not limited to elastomers) and conductors (hydrogels, but also thin metal films or other types of stretchable conductors). Of course, you would have to adapt the beautiful name of GEO which you might not want to do. In some sense GEO can be viewed as a specific case of HASEL, using elastomers and hydrogels. I look forward to seeing your results and a discussion in iMechanica!

Here are answers to your questions:

1) Self-healing is only one advantage of using liquid dielectrics. I think what might be even more or equally important is the ability to have direct electrical control over soft hydraulic actuators, which are incredibly versatile and can achieve a lot of different actuation modes. We are still exploring what we can do with different types of oil with high values of permittivity or dielectric strength, both of which influence Maxwell stress. It is also a nice feature that HASEL actuators can be optimized for a specific application by making use of hydraulic amplification, as we have shown with donut HASEL actuators.

2) We have shown that donut HASEL actuators undergo a pull-in transition, which can be clearly seen from our experimental data. Whether this is a desired feature or not depends on the specific application. Some applications might benefit from bistable behavior (maybe soft mechanical switches or triggers, maybe also valves), other applications do not benefit from pull-in. The design of HASEL actuators is very versatile and we already have versions of these actuators that do not show this type of highly nonlinear response, but react with monotonic, almost linear behavior (which can be a simplifying feature for controls in robotics applications).

Happy new year of the dog!




Subscribe to Comments for "Journal Club for February 2018: HASEL artificial muscles for high-speed, electrically powered, self-healing soft robots"

Recent comments

More comments


Subscribe to Syndicate