Journal Club Theme of September 2013: Stretchable Ionics

Zhigang Suo's picture

In a paper just published in Science, we describe a class of devices fabricated using stretchable, transparent, ionic conductors.  These devices are highly deformable and fully transparent.  They can operate at frequencies above 10 kHz and voltages above 10 kV. We demonstrate a fully transparent loudspeaker that plays music. See a YouTube video.  The Supplementary Materials contain experimental methods, theory, and more movies. The Science magazine conducted a podcast interview, which covered some of the same ground in this post.

Motivation.  The past century has seen the rise of electronics—engineered devices in which electrons carry electrical charge.  The successful rise of electronics, however, does not diminish another success with a much longer history—ionics, Nature’s solution to charge transport, based on ions and water.  We live in two parallel worlds: the world of electronics and the world of ionics.

The two worlds communicate with each other.  We watch TVs and listen to radios.  We drive cars, guided by GPS.  More intimate communication between these parallel worlds—the engineered (using electrons) and the natural (using ions)—is creating a hybrid field:  bioelectronics. Examples of applications include electrode arrays, where the electronics of medical instruments meet the ionics of tissues and cells, and brain-machine interfaces, through which cortical ionic impulses control prosthetic arms.  One person's brain can control the other person's fingers, through electronics and the Internet.  The thought of a human can move the tail of a rat.

The emergence of bioelectronics highlights a fundamental challenge:  Electronic devices are mostly made of hard materials, and human bodies are mostly made of soft tissues.  To enable electronics to meet skin, heart and brain, we need stretchable conductors. Stretchable conductors are also needed in non-biomedical applications, such as in soft robotics and tunable optics.  See a perspective written by John Rogers in the same issue of Science. 

Limitations of existing stretchable conductors.  Existing stretchable conductors are mostly electronic conductors, including carbon grease, micro-cracked gold films, serpentine-shaped metallic wires, carbon nanotubes, graphene sheets, and silver nanowires. Attributes other than conductivity and stretchability are also important in specific applications.  Conductors may need to operate at high frequencies and high voltages, remain conductive whilst undergoing areal expansions of 1000% or more, be biocompatible, and be transparent. 

While electronic conductors struggle to meet these demands, ionic conductors meet most of them readily.  Many ionic conductors, such as hydrogels and ionogels, take a solid form, and are stretchable and transparent. Many hydrogels are biocompatible and conformal to tissues and cells down to the molecular scale. 

What are stretchable, transparent, ionic conductorsWhen salt dissolves in water and forms ions, the ions are  electric charges mobile in water.  Thus, saltwater is a transparent, ionic conductor.

But saltwater is a liquid, and our devices need a solid electrical conductor.  We make a hydrogel by combining saltwater with a polymer network.  Saltwater enables electrical conduction, and the polymer network provides the solid form.  The hydrogel is like jello, but is far more stretchable.  The hydrogel can be stretched more than five times its length. 

We have used hydrogels in much of this work, but we have also used nonaqueous ionic conductors.

How to use ionic conductors to make high-voltage devices.  The interface between an electronic conductor and an ionic conductor will undergo electrochemical reactions when the voltage across the interface exceeds a value on the order of 1 V.  How can we  use the ionic conductor to make a high-voltage device without causing electrochemical reactions?  Our device consists of two capacitors in series: one capacitor is the interface between the electrode (electronic conductor, copper) and the electrolyte (ionic conductor, hydrogel), and the other capacitor is the dielectric.  The two capacitors have vastly different capacitances.  When a high voltage is applied cross the device, the voltage drop across the electrode/electrolyte interface is tiny, so that electrochemical reaction will not occur.  Our device achieve electromechanical transduction without electrochemical reaction.

How to use ionic conductors to make high-frequency devices.  The conductivity of the ionic conductor is about 6 orders of magnitude smaller than the conductivity of copper. How can we  use the ionic conductor to make a high-frequency device, such as a loudspeaker?  In our device, the ionic conductor and dielectric form a layered structure.  The ionic conductor provides the resistance R, and the dielectric provides capacitance C.  Their product RC gives the time delay of the device.  Even though the resistance R of the ionic conductor is large, the capacitance C of the dielectric is very, very small.  Consequently, the RC time delay is very short.  Our actuators are not limited by the electrical resistance, but by mechanical inertia.

How does voltage cause deformation?  A dielectric elastomer is an electrical insulator, and is a highly stretchable elastomer.  When an external circuit applies a voltage to a membrane of the dielectric elastomer across its thickness, the membrane behaves like a capacitor.  Positive charge accumulates on one face of the membrane, and negative charge on the other face.  The positive and negative charges attract each other, causing the membrane to reduce its thickness and expands its area.  In the last decade, intense effort has been devoted to the development of dielectric elastomer transducers.  See an animation of dielectric elastomer actuator at Wikipedia.

Artificial neuromuscular system.  A dielectric elastomer functions as an artificial muscle, capable of large deformation in response to electrical stimulation.  An ionic conductor functions as an artificial neuron, capable of bringing in the electrical stimulation.  That is, the ionic conductor innervates the dielectric elastomer.  The pair of materials together mimic the functions, but not the anatomy, of a neuromuscular system. 

Artificial neurosensory systemThis paper focuses on the use of ionic conductors in devices operating at high frequency and under high voltage, but the layered electrolytic and dielectric elastomer also works for applications that require low voltage or low frequency.  When stretched mechanically, the layered material increases area and reduces thickness, so that its capacitance increases.  This characteristic will enable transparent sensors operating at low voltage, capable of measuring strains over a large range, and conformal to soft tissues.  Such devices mumic functions of neurosensory systems.     

Mechanics of artificial neurons Ionic conductors mimic neurons, capable of sending electrical signals between brain, muscle and sensory organ.  The brain, muscle and sensory organ can each be real (e.g., in bioelectronics), or artificial (e.g., in soft robotics). 

The Supplementary Materials of the paper describes the mechanics of artificial neurons.  As mentioned above, the frequency of actuation is not limited by the RC delay, but by mechanical inertia.  The actuation strain decreases as the frequency increases.  The RC delay does limit the length of ionic interconnect.  The amplitude of electrical signal decays over distance.  It is intriguing to compare an ionic interconnect to a long neuron. 

We also show that viscoelasticity causes creeping movement of the artificial muscle in response to a suddenly applied voltage.  The elasticity of the ionic conductors can also be important.  We make the ionic conductors of low elastic modulus (~kPa) and small thickness (~0.1 mm), so that the ionic conductors do not constrain the deformation of the dielectric elastomers.  We show that, at low frequencies, the actuation strain is limited by electromechanical instability.

Potential applications.  Here are some obvious targets:

  • Stretchable, ionic conductors can be used to make artificial neuromuscular and neurosensory systems for soft robots
  • Stretchable, biocompatible, ionic conductors can be used to make biomedical devices.
  • Stretchable, transparent, ionic conductors can be used to make tunable optical devices.
  • Transparent loudspeakers might be attached to windows to achieve active noise cancellation.

Every device designer can ask this question:  Can I replace the electronic conductor in an existing device with an ionic conductor?  The device may lose some performance, but may gain other attributes, such as stretchability, transparency, and biocompatibility.  What can I do with these attributes?  Can I start from scratch, and design a stretchable ionic device that does not even have an electronic counterpart?

Our theoretical estimate has shown that ionic conductors can operate at much higher frequencies than 10k Hz.  One should be able to make devices much faster than loudspeakers.  What high-frequency, ionic devices make sense?

Hydrogels dry as water evaporates.  We use hydrogels as stretchable, transparent, ionic conductors because they are easy to make and inexpensive.  Hydrogels are the ionic conductors of choice to demonstrate concepts, and to fabricate devices that require biocompatibility.  We also note that the diversity of ionic conductors creates a large pool of candidates, some of which avoid this problem. For example, ionic liquids and gels swollen with ionic liquids are nonvolatile ionic conductor. We show that ionic liquids can indeed be used as conductors for dielectric elastomers.

Challenges and opportunities.  The development of ionic conductors for stretchable devices raises many questions in mechanics and materials science.  Here are a few examples:

  • Will ionic conductors have long lifetimes?
  • Will ionic conductors be compatible with electronic conductors and dielectrics?
  • How do we ensure adhesion when the devices are stretched repeatedly?

There is only one kind of electron, but there are infinite many kinds of ions. This diversity will enable ionic conductors to be designed for many applications. Life uses primarily ions—rather than electrons—to carry electrical charge.  In creating biomedical and engineering devices, it is well to consider the opportunity:  the hard and the soft do not necessarily have to meet through electronic conductors; they may as well meet through ionic conductors.

We love to hear your thoughts.  Many of you have been working on soft materials, stretchable electronics, sensors and actuators.  Some of you are experts on large deformation of soft materials, or large displacement of flexible structures.  We would love to hear from you about opportunities and challenges.  Also, we would love to learn about your work in related areas.  Please leave your comment below.  As always, your comment can be on anything related to the topic, and will be especially valuable if it connects to your own work, or the work you know well.  We love to discuss with you.

Links cited in this post

Pradeep Sharma's picture


Dear Zhigang,

Thanks for a very readable and interesting perspective...the questions you raised, e.g. Reliability of ionic conductors, adhesion etc are intriguing avenues for future mechanics research...A couple of questions arose as I was reading through your post related to the basic physics of these materials:

(I) What physical parameter dictates the voltage beyond which electrochemical reactions will take place at the electronic-ionic conductor interface?

(2) I would think that with the large deformations involved, the dielectric constant (and hence capacitance) itself would change with deformation. Is there some evidence of that in your work or is it even an issue?

(3) In typical Si-based semi-conductor electronics, there is a strict temperature limitation; e.g. 200 C junction temperature for most devices. Presumably, given the materials involved, the operating temperature limits will be much lower for ionic conductor based electronics?

Regards, Pradeep

Zhigang Suo's picture

Response to questions related to stretchable ionics

Dear Pradeep, Thank you very much for your comment.  Here are quick responses to your questions.  Hopefully other people will also jump in.

  1. Voltage across the electrode/electrolyte interface beyond which electrochemical reaction will occur.  Really good question.  We have not done systematic experiments to proble the limiting voltage.  We tried to search for answer in the literature, but did not find good review of this issue.  Our experimental setup does provide a means to explore this question.  We briefly toyed with the idea of looking at the copper electrode in the microscope after some time of operation, and observing the evolution of microstructures.  But we did not pursue the idea.  Let's hope that someone can give us input on this important question.
  2. Does dielectric constant depend on large deformation?  Another really good question.  The short answer is no.  A dielectric elastomer is a polymer network of very long chains.  In such a network, deformation and polarization are two nearly independent processes.  Deformation involves the stretching of the polymer chains, and polarization involves rotation of monomers.  Because the polymer chains are very long, each chain contains many monomers, and the crosslinks almost do not constrain the rotation of individual monomers.  The dielectric behavior of an elastomer is nearly the same as a polymer melt (a liquid). This molecular interpretation is supported by experimental observations:  the measured dielectric constaint changes a few percent after an areal expansion of 25 times.  In theoretical calculations, it is common to assume that the dielectric constant of an elastomer is independent of deformation, a model known as ideal dielectric elastomer.  For further detail, see my early review of the theory of dielectric elastomers.
  3. Temperature limitation.  Very limited data are available in the literature.  In our own work, not reported yet, we had some reasons to go about 100 C, using carbon grease as electrodes.  The lifetime is short.  But we have not sorted out the fundamental limits to the temperature of operation. 

Konstantin Volokh's picture

Re: Response to questions related to stretchable ionics

Hi guys,
Interesting. It is especially exciting that the ionic conductors mimic living tissues. Here are my five cents concerning the second question: Does dielectric constant depend on large deformation? I considered the influence of large deformations on the dielectric parameter in this paper. The influence is negligible. Thus, it is possible to assume the dielectric parameter constant.

Cai Shengqiang's picture

Thanks for the very inspiring paper

Dear Zhigang,

Thanks for sharing the very inspiring paper and promote further discussions here. 

I have one quick question regarding the measurement of the resistivity of the saltwater gel.  I suppose that the resistivity results shown in the paper is obtained by applying static voltage and measuring the current. Am I right?  Any measurements of the dependence of the resistivity (or impedance) of the gel on the frequency of applied voltage?   

Another desirable merit of the gel conductor appearing in my mind is probably about the adjustability of their mechanical properties. For example, the stiffness of the gel conductor can be easily changed or even patterned in the fabrication process. Different applications may need conductors of different stiffness. 

Cheers and congratulations to all the authors of the excellent paper.  shengqiang 

Jeong-Yun Sun's picture

Thank you for the questions.

Hi Shengqiang,

Thank you for your interest.

Yes, we used static voltages to measure the resistivity of hydrogel. But we measure the currents after saturation.

We haven't varied frequencies. Therefore, we don't know the frequency effect yet.

Follwings are the details about the measurements.

The resistivity of the hydrogels was measured as the hydrogels were pulled by a uniaxial force.  The resistance was measured by using four-point probes. To minimize the effect of ions built up on the surface of the probes, the resistance was measured with three relatively large voltages (20 ~ 50 V; electrochemical potentials are much smaller) and the corresponding currents after saturation. Hydrogels containing 1.37, 2.74 and 5.48 M of NaCl were used. When the hydrogels were not stretched, the measured molar conductivity was 120.19 Scm2/mol, which was close to a reported value of 118.5 Scm2/mol for aqueous solutions [J. O’M. Bockris, A. K. N. Reddy, Modern Electrochemistry, vol. 1 (Plenum Press, New York, ed. 2, 2002), p. 434.]

Please let me know if you have further questions!

Thank you.


Adrian S. J. Koh's picture

With this paper, a path to mass commercialization is clearer

Zhigang, once again congratulations!

Ionic conductors potentially solves many problems, and may show the way to eventual mass commercialization of electromechanical transducers and other physically-similar applications.  I have the following questions:

1.   As gel-like ionic polymers have rather low fracture resistance (as compared with the resident dielectric), how durable will such active dielectrics be?

2.   I note the RC time-scale is small.  Does current leakage play a major role in the transduction process?

3.   I have to ask on its function as a generator: As mentioned, the capacitance should be small to ensure a small RC time.  This also implies that the charges held by the dielectric due to pre-charge may be very small.  This minute amount of charges may be dissipated in such a short time that no change in electrical state may be gained.  My question is, what would your view be on its use as a generator?

Thank you.



Christoph Keplinger's picture

Exciting times for dielectric elastomer generators!

Dear Adrian,

thank you for your commendation of our work and for these interesting questions!

Here are some answers:

ad 1) One nice fact about stretchable, transparent, ionic conductors is that they comprise a large class of materials. With respect to the use and fracture resistance of hydrogels, please have a look at the following paper:
Jeong-Yun Sun, Xuanhe Zhao, Widusha R.K. Illeperuma, Kyu Hwan Oh, David J. Mooney, Joost J. Vlassak, Zhigang Suo. Highly stretchable and tough hydrogels . Nature 489, 133-136 (2012).
The hydrogel presented in this paper features the same excellent fracture energy as natural rubber.

ad 2) The RC time-scale of our design (Fig. 1 from the Science paper) is small compared to a circuit that does not contain the dielectric: For a circuit with only electrode-electrolyte-electrode, the two interfaces between the electronic and the ionic conductors will behave as electric double-layer capacitors ( and with very high capacitances, thereby causing a very long RC time. With the introduction of the dielectric in series, the capacitance of the series connection is dominated by the comparably much smaller capacitance of the dielectric. This allows for a small RC time.

The capacitance arising from the dielectric layer is the same when using ionic conductors and electronic conductors. Thus this does not influence the leakage properties.
There could be an influence on current leakage when switching to ionic conductors though: the nature of charge carriers changes and thereby the interaction with the dielectric. We did some preliminary testing to better understand this difference. Please have a look at Figure S7 from the supplementary materials section. The experiment presented there showed that there is little difference between the leakage timescales of electronic (carbon grease) and ionic (hydrogel and ionic liquid) conductors, at least within the parameter space (maximum electric field, humidity, temperature, types of dielectric and ionic conductor) tested there.

ad 3) As stated in the "ad 2)" section, the capacitance of a dielectric elastomer generator using electronic conductors is the same as with ionic conductors. For highly stretched states of the generator, the ionic conductors might even be advantageous: they do not show percolation effects, thus potentially allowing for even better performance. Moreover, the resistance in highly stretched states is comparably low.

Future research on the use of ionic conductors in dielectric elastomer generators could yield very interesting new areas of application: fully transparent generators?; maybe bio-compatible generators?

Exciting times!


Xuanhe Zhao's picture

Transparent Conformal Electrodes based on Ionic Solutions

Dear Zhigang,

Thank you for posting this very inspiring entry! Congratulations to Christoph, Jeong-Yun, and all co-authors of this milestone paper!

Ionic solutions are abundant in nature and human body with examples ranging from seawater to body fluids. These ionic solutions may not be "good conductors" according to many criteria such as conductivity and stability. However, they do find important niche applications in many fields where high transparency and high deformability are required. I definitely want one of the clear loudspeakers with gels electrodes if commercialized.

My group also used electrodes based on ionic solutions in many of our studies. Instead of making them into gels, we put them into containers or simply used ambient ionic solutions such as seawater or body fluids as grounded electrodes. Once a high electric voltage is applied, the transparent conformal ionic-solution electrodes indeed enabled observation of interesting phenomena and important applications such as dynamic patterning and antifouling coatings .

I also echoed the challenges (or opportunities) you proposed for this emerging field. For example, it is known that water in polymer dielectrics under high voltages can cause the water-tree and electro-cavitation instabilities, which are detrimental. How to guarantee the reliability and stability of the system in high-voltage applications? Will gel electrodes fatigue under cyclic deformation? Furthermore, we found graphene and metal electrodes can be made superhydrophobic. Is it possible to integrate the merits of different types of electrodes into one system? All these may be interesting questions and therefore opportunities for the mechanics communicity.


Philipp Rothemund's picture

Re: Transparent Conformal Electrodes based on Ionic Solutions

Dear Xuanhe,

thank you for your comments.

You are right. Ionic solutions have higher resistivies than electronic conductors and often low stabilities. However, the sheet resistivites of ionic conductors can be orders of magnitude lower than the sheet resistivities of electronic conductors at high stretch, as is shown in figure 4 of the paper.  We also used an ionic liquid as ionic conductor in the paper (see figure S6 of the supporting material). Ionic liquids are known to have high thermal stabilities and to be non-volatile. They can be made into a gel (K. H. Lee et al., "Cut and stick" rubbery ion gels as high capacitance gate dielectrics. Adv. Mater. 24, 4457-4462 (2012)). The pool of ionic liquids is almost infinite because of the high combinability of the ions and offers the chance of finding a good ionic conductor for a wide range of applications.

You also pose the question of reliability and stability of systems which use ionic conductors. This is indeed an important factor, which will determine the practical use of ionic conductors, also with commercialization in mind. The use of ionic conductors in high voltage applications involves new, not yet investigated physics. Of importance is not only, whether systems with ionic conductors intrinsically have different lifetimes than systems with electronic conductors, but also the influence of the choice of ionic conductor on the performance and reliability. We are already intensely working on this questions and will present results soon.

The stretchable ionic conductors give the opportunity to design soft machines with new characteristics and functionalities. By combining ionic and electronic conductors in a single device, one may be able to use the advantages of both worlds-the worlds of electronic and ionic conduction.


Zhigang Suo's picture

hydrogels are much less resistive than carbon grease

Carbon grease has been widely used as the conductor for dielectric elastomer devices.  According to the supplier, the resistivity of carbon grease is about 1 ohm-meter.  According to our measurement, the resistivity of hydrogel is about 0.02 ohm-meter (Fig 4A of the paper).  Thus, the hydrogel is much less resistive than carbon grease.

Of course, resistivities of Cu, Ag, Au  are much lower, about 6 orders of magnitude lower than that of hydrogel.  But it is a challenge to make metals into stretchable, transparent conductors.  See a review article by Lipomi and Bao.   

Zhigang Suo's picture

Game changer: an issue becomes a feature

Dear Xuanhe:  You have identified an issue:  hydrogels are not as stable as metals.  In Yonggang's comment, he has changed the game.  He has turned this issue into a feature.  Hydrogels can be made biodegradable, so that a device can disappear after serving its useful function.  You must have read their paper on transient electronics.

Xuanhe Zhao's picture

Application-Oriented Design of Materials

Dear Zhigang,

I really appreciate this point raised by you and Yonggang. We learned various design creteria for materials and structures in traditional applications; conductivity, stability and price are among those for conductors used in traditional electronics. However, the emerging of new applications and technologies indeed changed the game. For example, the dynamic antifouling coatings we developed simply use ambient seawater and body fluids as grounded electrodes, which are even less conductive and stable than ionic gels, but they did satify the requirements of our application at almost zero cost. There is no doubt one still needs to consider a set of criteria in designing materials and structures, but the design criteria actually evolve with applications and technologies.


Zhigang Suo's picture

Soft materials deformed by air and liquid

Dear Xuanhe:  Your work of using saltwater as conductor to actuate soft materials is very exciting.  The work also points to a large family of devices that mix fluids with soft materials.  Elastomers are combined with either pneumatics or hydraulics. Deformation can be very large, and instability is commonplace.  They are used to build soft robots.  Here are some recent examples:

A common theme is to use large deformation (and instability) to do useful things.  This theme has been discussed in iMech jClub hosted by you  and by Douglas Holmes.

The use of large deformation and instability to achieve functions has once again brought mechanics to the forefront of engineering.  We need to create new designs that connect structures to functions.  We need to develop material models and computational tools to enable the nascent field of soft machines.

Zhigang Suo's picture

Transparent loudspeaker proves a point. But what is the point?

The transparent loudspeaker appeals to people. Sandrine Ceurstemont at New Scientist called it ionic music. Jesse Emspak titled his report at Fox News as "transparent artificial muscle plays music".  Reports at Gizmodo and Engadget are followed with comments from readers.  The comments are drawn to the transparent loudspeaker.  It is water-clear and lightweight.  It can even be made foldable or deformable. 

Will the transparent loudspeaker produce really high-quality sound?  Will it last long?   Maybe, as we improve the choice of materials.  But this paper does not answer these questions.

Caroline Perry in her news release noted that the transparent artificial muscle plays Grieg to prove a point.

Here is the point: stretchable ionic conductors, together with stretchable dielectrics, can make high-frequency, high-voltage, stretchable devices. In addition to conductivity and stretchability, many ionic conductors are transparent and biocompatible.

Our own interest in this work goes beyond the specific materials used (hydrogel, VHB, and copper), or specific devices demonstrated (transparent actuator and transparent loudspeaker). We hope that the work will help to start a change of perspective. In some devices, electronic conductors can be replaced with ionic conductors, and these ionic conductors bring new characteristics.

If ionic conduction is good enough for life for all this time, it must be good enough for many engineered and bioengineered devices.  Let's liberate ionic conduction from slow devices such as batteries.

Christoph Keplinger's picture

Transparent, Stretchy Speaker Shows 'Soft Machine' Potential

Inspired by all the news coverage on our transparent loudspeaker (a multisource video news service that analyzes world news and produces 2-to 3-minute, streaming video clips) has produced a well made video with live commentary. Have a look here:

Christoph Keplinger's picture

News on the transparent loudspeaker

Another recent news video on the transparent loudspeaker:

I also enjoyed reading the story from Meghan Rosen in ScienceNews, containing several quotes from reseachers also working on soft machines:

Nanshu Lu's picture

Re: Journal Club Theme of September 2013: Stretchable Ionics

Dear Zhigang, 

Thank you very much for sharing with us your "behind-the-sceen" insights here. It is indeed very exciting to watch the rise of, if I may take the liberty to name, "eletrionics", inspired by this landmark work. Congratulations to the whole crew who made it happen!

Compared to stretchable electronics including stretchable optics and sensors, there are much fewer studies on soft and stretchable actuators and generators (see a summary of them in "Flexible and stretchable electronics paving the way for soft robotics") mainly due to the lack of mechanically-matched highly-deformable conductor-actuator pairs. Dielectric elastomer paired with ionic conductors offers an ideal solution and oppurtunity in both soft acturators and soft generators.

The three questions I have are: 

1. In addition to artificial muscle, soft and elastic actuators can find significant applications in artificial hearts and heart valves. In these cases, would the cycliability and high voltages pose any challenges? 

2. Again for implantable applications, what are the possible effects of ionic fluid in the environment on the ionic conductos? If encapsulations are required, is there a material solution for that yet?

3. What is the adhesion condition at the ionic conductor-dielectric elastomer interface? I understand the hydrogel is very soft (~kPa) so the driving force for delamination is probably very small, but am still curious if any debonding/sliding has been observed?

Congratulations again and cannot wait to see followups.



Zhigang Suo's picture

Reponse to Nanshu's comments

Dear Nanshu, thank you very much for your comments.  Here we enjoyed reading your article "Flexible and stretchable electronics paving the way for soft robotics".  I also read the State-of-the-field discussion of soft robotics with great interest.  One of your points comes across clearly: Stretchable electronics have so far been effective in making sensors, but soft robots will need actuators.  Congratulations on becoming an associate editor of the new journal Soft Robotics.   

Incidentally, there is a book titled Iontronics.  The book focuses on mobile ions in organic electronic materials.  We have called our project Stretchable Ionics.

Now responses to your questions.

1.  Will cycles and high voltages pose any challenge?  Like most applications, fatigue can be an issue.  However devices operating over millions of cycles have been demonstrated.  As the field moves forward, issues of reliability will receive more attention.  Right now I don't see fatigue as being a show stopper.  High voltages are a serious issue.  At the moment, the electric field needed for actuation is 10^8 V/m.  Even when you make individual layer as thin as 10 um, the voltage will be at about 1000 V.  For sensing, however, very low voltage is needed, as discussed in my response to Yonggang.  There are more to the issue of high voltage.  I'll try to get Christoph and Tiefeng to respond.  They both have given the matter serious attention. 

2.  What are possible effects of ionic fluid in the environment on the ionic conductors?  I assume some form of encapsulation will be needed in many applications.  Hydrogels have been used inside bodies in tissue engineering and drug delivery.  We should be able to learn from past experience.

3. What is the adhesion condition at the ionic conductor-dielectric elastomer interface?  Jeong Yun discovered that the adhesion between the hydrogel and the dielectric was improved if, before attachment, he dried the surface of the hydrogel with nitrogen gas.  The adhesion was good enough for our demonstration, but delamination deserves serious attention in future research.

Christoph Keplinger's picture

High voltage devices can be safe and small!

Let me further comment on the issue of high voltage:

It is a common believe, that high voltages are something dangerous and something that you want to avoid. In reality there is much more to say about that. The dangers associated with high voltages are largely porportional to the capacitance of your circuit and consequently the maximum current and power exposure in the case of short circuits.
As it is the case with dielectric elastomer transducers, they work with high voltages, but very low currents/capacitances. Typically, the capacitance of a dielectric elastomer actuator is in the range from 1 to 1000 picofarads. Even in the case of touching such a device, that would be safe. Everyone of us has experienced electrostatic discharges originating from sources with small capacitances from objects in our surroundings, such as for example a pullover.

Moreover, high voltage can come in a small package: As we do not necessarily need high electrical currents to operate dielectric elastomer actuators, we can use amplifiers with lower power ratings. As an example, please have a look at the following coin-size amplifier from EMCO:

This little device can be powered from a watch or cell phone battery and produce voltages up to 6000V!


Yonggang Huang's picture

Dear Zhigang, I read,

Dear Zhigang,

I read, with great interests, your paper on this very inspiring and intriguing topic.  This is really a breakthrough, and has a lot of potentials for new electronics and ionic systems.  The success of
ionic type conductors in dielectric actuators will undoubtedly inspire many
interesting follow-up works in the directions of soft robotics.

Here are some questions:

Is it possible to use the applied strain to tune
the resistance of the ionic type conductors in a relatively wide range (e.g., changing several times), in a reversible manner?  This is difficult to accomplish by using
traditional metal conductors.

Is the hydrogel bio-dissolvable (under the
chemical environment of human body)?  Is
it possible to actively design the lifetime of the ionic conductors, such that
they can disappear after desired working time?

Can you envision any applications of the ionic
type conductors where only small voltage (e.g. <20 V) is needed, which may
be more suited for bio-integrated usages?  

Thanks again for posting this inspiring work.

Zhigang Suo's picture

Reply to Yonggang's Comments

Dear Yonggang: Thank you so much for your kind words. They meant a lot to me and my students. Your work with John Rogers has shaped the field of stretchable electronics. In our paper, we have tried to link the two exciting fields: stretchable electronics and dielectric elastomer transducers. The two fields have identified a common need: highly stretchable electrical conductors. Our paper describes a solution: ionic conductors. This solution will create new questions, as well as new opportunities. Comments from you and others have identified some of both.

Here are quick responses to your questions.

1. Can strain change the resistance of an ionic conductor by a large amount in a reversible manner? Yes. By a purely geometric argument, one can show that the resistance of a conductor is quadratic in stretch (p.6 of Supplementary Information ). This theoretical predicton matches well with our experimental data (Fig. 4A of the paper). In the experiment, we use uixial force to stretch a hydrogel elastically beyond 6 times its original length. The resistance increases beyond 36 times.

2. Is the hydrogel biodegradable?
Our experiment uses covalently crosslinked polyacylamide hydrogel, which is not biodegradable. However, biodegradable hydrogels have been studied for some years. One should be able to find a hydrogel both biodegradable and electrically conductive.

3. Can you envision any applications of ionic conductors where only small voltage (e.g. <20 V) is needed
? Yes. Our paper ends with the following sentences: "This paper focuses on the use of ionic conductors in devices operating at high speed and under high voltage, but the layered electrolytic and dielectric elastomer also works for applications that require low voltage or low frequency. When stretched mechanically, the layered material increases area and reduces thickness, so that its capacitance increases. This characteristic will enable transparent sensors operating at low voltage, capable of measuring strains over a large range, and conformal to soft tissues."

Matt Pharr's picture

Re: Reply to Yonggan's Comment

Dear Zhigang,

I agree that you can increase the capacitance by increasing the area and decreasing the thickness. This will also increase the response time by the same factor, meaning the response time will inrease by orders of magnitude. Is this an issue?  What applications do you have in mind for low voltages?

KeithFoo's picture

Response time

Hi Matt,

Yes, you are right. When stretched mechanically, the capacitance and the response time (RC time) will both increase. By considering the geometry, we can show that the capacitance scales to the 4th power of the stretch, while the RC time scales to the 6th power of the stretch  (pg 11 of Supplementary info ). For representative values considered here, this RC time is still faster than milliseconds at a stretch of 6. Moreover, given that the RC time can always be manipulated by the actuator design in terms of size, the RC time is not a fundamental limitation for the response speed, especially for small devices.


Zhigang Suo's picture

ionic skin

Siegfried Bauer has just published in Nature Materials a perspective on the state-of-the-art of the electronic skins.  See also an article by Takao Someya entitled Bionic Skin for a Cyborg You.  

In recent years, a great many stretchable sensor arrays have been demonstrated, with ever increasing sophistication.  These sensors need be soft and stretchble, so that they do not constrain soft tissues and they do not break.  These sensor arrays use electronic conductors.  A major challenge is to make the electronic conductors stretchable.  I have reviewed mechanics of stretchable electronics and soft machines.

In the concluding paragraph of our Science paper, we described the prospect of using ionic conductors to develop sensors.  Ionic conductors can readily achieve stretchability, biocompatibility and transparency.  They better mimic nerves.  The development of sensitive skins is rapid.  We should soon see ionic skins.

I look forward to reading a post about iSkin on iMechanica.

Jizhou Song's picture

Re: Journal Club Theme of September 2013: Stretchable Ionics

Dear Zhigang,

Congratulations and thank you for sharing your inspiring insight behind stretchable and transparent ionic conductors. Stretchable electronics based on electronic conductors require a sophiscated mechanical design (e.g., serpentine-shape). Stretchable ionic conductors will offer possibilities of a simple design for stretchable electronics with a very high areal coverage.

When reading through the paper, a few questions on the resistivity comparison came up.

The resistivity of ionic conductors was compared to that of single-wall carbon nanotubes under stretch up to 800%. CNTs usually fracture at a strain of a few tens percent. Did you observe CNT fracture in experiments? Electronic properties of CNTs (conductor or semiconductor) depend on the radius and chirality. How did you choose CNTs for the resistivity comparison? 


Jeong-Yun Sun's picture

Thank you for your questions.

Dear Jizhou,

Thank you for your question.

The resistivity data for CNT are not measured by ourselves.
We used Prof. Pei's data and cited his paper [L. Hu, W. Yuan, P. Brochu, G. Gruner, Q. Pei, Highly stretchable, conductive, and transparent nanotube thin films. Appl. Phys. Lett. 94, 161108 (2009).]
Unfortunately, I couldn't get the information about the radius and chirality of CNT from their paper.
You can contact Prof. Pei for the details about CNT what they used.

Thank you.


Hanqing Jiang's picture

Very inspiring work! Thanks, Zhigang!

Dear Zhigang:

Thank you for posting this inspiring work! I have read your paper with great interests. It is a seminal work of making stretchable ionic conductors. Congradulations to you and your team!

I have some questions to discuss:

(1) Is the ionic elastomer thermally stable? Any disassociation?

(2) What new problems you envision when the stretchable ionic conductors are integrated with miniature devices? Any compatibility problems?

(3) It is a follow-up question on low operation voltage as Yonggang asked and you replied. It helps if you can provide some representative values. 

Thank you and congradulations again! 

Christoph Keplinger's picture

Thank you!

Hi Hanqing,

thank you for your interest in our work!

Here are some answers:

(1) Most of our target applications will operate at room temperature or at body temperature. For this range of temperatures most ionic elastomers will be stable. We have already mentioned, that ionic conductors represent a very large class of materials. Hydrogels will not be the best choice for applications that need to operate at elevated temperature. It will indeed be very interesting to find ionic conductors that tolerate harsher environments.

(2) I envision problems, when length scales become so small, that the resistance of circuits exceeds a tolerable range. Ionic conductors, at least in the unstretched state, have much higher resistivities compared to metals. Therefore, ionic conductors are more limited with respect to miniaturization of conductive channels. Nevertheless, ionic conductors are intrinsically soft and stretchable, thus it will be easier to interface with biological systems.

(3) For low voltage applications we have to differentiate between actuation and sensing:
Sensing will be possible with voltages much lower than 1 V.
Electrostatic actuation is governed by the electric field. Thus the required voltages will depend on the thickness of the dielectric. It should be possible to bring them below 1kV.



Jinxiong Zhou's picture

Wrinkles of hydrogel electrodes?

Dear Zhigang,

Thank you for posting this inspiring topic. We read and learned a lot from reading of your paper. Congratulations to you and your group on this beautiful work!

Circular actuator is one of the commonly used and simple demonstrations of dielectric elastomers, where fixed displacement constraint is enforced. If the applied voltage exceeds a critical value, the elastomer loses its tension state and wrinkle will occur. This phenomenon is observed by many researchers and the wrinkling process is reversible. For a dielectric elastomer with elastomeric electrodes , wrinkling is also observed.

My question is did your students observe wrinkle of hydrogel electrode in your experiment? Will the wrinkle relate to the failure of dielectric elastomer? Another question is will the wrinkling of hydrogel electrodes affect the adhesion between the hydrogel and dielectric elastomer?



Philipp Rothemund's picture

RE: Wrinkles of hydrogel electrodes?

Dear Jinxiong,

Yes, at high enough voltages we observed wrinkles of the active region during our experiments. The stiffness of the layers of hydrogel was much lower than the stiffnes of the dielectric layer. Therefore, they did not significantly constrain the elastomer.  After the application of a stepvoltage wrinkles appeared locally in the active region. Due to viscous creep of the elastomer the wrinkles gradually spread over the active region until breakdown occured. This behavior is very similar to when carbon grease is used as electrode material. See for example C. Keplinger et al. Appl. Phys. Lett. 92, 1929032 (

With the debonding of the ionic conductors from the dielectric layer you bring up a very important issue for the functional capabilities of our design of stretchable, ionic conductors and is determined by the shear strength of the interface between the different components. In our experiments with dielectric elastomer actuators, wrinkling did not lead to delamination, but we observed delamination at very high stretches and high frequencies. In our paper we used VHB 4910 as dielectric, which is an adhesive tape. This allowed us to simply stick the hydrogels-after drying their surfaces-onto the VHB. To find a good combination among the countless possible combinations of different ionic conductors and dielectric materials or even to physically bond the layers can be a possible goal of future work


Taher A Saif's picture

Journal Club Theme of September 2013: Stretchable Ionics

VERY cool paper, Zhigang. It opens up many possibilties and provides food for thought.

1) As the gel undergoes large deformation over time, one woud expect
significant ionic fluid transport through the pores of the gel, thus
dissipating energy giving (a) high internal friction and low Q, and (b)
change of some mechanical property of the gel, e.g., increased
anisotropy, with time. If the deformation is very fast, flow is expected
to be limited, and hence high Q. Thus, it appears that the damping
coefficient of the system may depend on the frequency of oscillation. 
Did you find this trend in the expt.?


2) The concept of mimicking neuro-muscular junction or the neuron itself
is intriguing. I am wandering whether memory can also be mimicked,
i.e., does the transduction depend on the prior history of defrmation or
voltage application? If so, is the retention of "memory" depends on a
minimum number of voltage aplications and the value of the voltage?    

Zhigang Suo's picture

Response to Taher on stretchable ionics

Dear Taher:  Thank you very much for your comments.  The solvent migrates in the polymer network, but maybe too slowly to cause the effect that you are thinking about.  For hydrogel, we have measured the effective diffusivity of water migrating through the polymer network in several previous papers (e.g., this paper).  A representative value is D = 10^-10 m^2/s.  The thickness of the hydrogel used in our devices is H = 100 um.  Thus, the relaxation time is on the order of H^2/D = 100s.  This relxation time is much longer than the time scale used in cyclic experiment. The frequencies used in our experiments range from 1 Hz to 20 kHz.  In our experiments, we did not observe the effect you mentioned.  But we also did not look for such effect.  Does the above estimate address your question?  Should we use another length scale in the above estimate? 

Can we make artificial memory using ionic conductors?  This question has come up several times in our internal discussions, but we have not given the matter any serious thought.  I feel that we can, given that an artificial memory only need to mimic the function, but not the anatomy, of the actual memory.  For example, a crude memory can be a structure with bitable states.  

Taher A Saif's picture

Yes, Zhigang. It answers my

Yes, Zhigang. It answers my questions. I guess the pore size is too small in the gel that makes the diffusivity so small.


As for the memory, I was more thinking along the line of analogue memory, in contrast to digital, as it happens in the natural neural system due to synaptic plasticity. It might be worthwile to explore whether "a property" of your gel system changes due to repeatitive application of "a stimulation", and remains changed over a time scale as it happens in neurons. This would mean, gel remembered its past history.

Jian ZHU's picture

Transparent artificial muscles

Dear Zhigang,

Congratulations and thanks for sharing this fabulous work.

Carbon greases are very popularly used as the electrodes in
the dielectric elastomer actuators. However, they are really messy and black. Biological
muscles may be transparent, for example, muscles in aquatic chromatophores. It
is interesting and significant to make artificial muscles closer to natural

Besides applications to loudspeakers, ionic conductors may
also have extensive applications in optics, say, soft lens or artificial eyes,
due to the transparency.

Zhigang will give a plenary talk in the 23rd computation
mechanics workshop (
It is great to meet Zhigang and learn about this inspiring work soon.


Zhigang Suo's picture

Re: Transparent artificial muscles

Dear Jian:  Thank you very much for your kind note, and for being a gracious host.  I had a good time visting Singapore.  It was wonderful to see you and Adrian setting up your own labs doing interesting experiments and making intriguing discoveries with dielectric elastomers.  The applications that you are developing are timely.

Through collaborations, you have also brought other strong groups into the field.  Dielectric elastomers are positioned to be a platform for many technologies, just as piezoelectric ceramics have been.  To develop this new platform, people with diverse backgrounds can contribute in different, and often unexpected, ways.  

Best wishes for your projects and collaborations!

Yuhang Hu's picture

Dear Zhigang, Thank you

Dear Zhigang,

Thank you for sharing and congratulations to all the people contributing to this inspiring work.

Here are two throughts.

1. How about the homogeneity of the hydrogel? What is the smallest length scale that we can use to pattern it as an electrode?

2. As you said before, hydrogel is a good platform to combine mechanics with chemistry. Have you thought about incorporating some functional molecules into the material making it multi-functional? For instance, incorporating "Spiropyran " into the hydrogel, you may be able to make a device which can change color as it is stretched, and you may also change the surface wettability of the material. 


Christoph Keplinger's picture

Thank you, Yuhang!

Dear Yuhang,

great to see you joining this discussion! I hope your growing family is doing well.

Some thoughts on your questions:

1. A homogenous thickness will be important for applications that require large areas of hydrogels with homogeneous mechanical properties. Our ability to achieve homogeneous properties will go hand in hand with the employed manufacturing techniques. Hydrogels and related ionic conductors are very flexible with respect to manufacturing: let us just think about 3d printing, spin coating or related techniques.
On the other hand, if we use inhomogeneous properties such as thickness as a feature, this will be of great advantage to construct tunable optical devices, such as lenses with variable focal length and aperture.
The smallest workable lenght scales for ionic conducturs will be prescribed by the maximum allowable resistance of the stretchable circuit. In any case, I think, that 3d printing of hydrogels should get us into feature sizes with this limiting lenght scale.

2. Great ideas! We would love to further explore hydrogels with tunable properties, such as color or wettability. Indeed, the rich palette of physical, mechanical and chemical properties available form ionic conductors is one of the things that got us excited about stretchable ionics in the first place!


Zhigang Suo's picture

sticky leader of stretchable ionics

The paper originally submitted to Science contained two acronyms and a short phrase:

  • STIC (Stretchable, transparent, ionic conductor)
  • LEADER (Layered electrolytic and dielectric elastomer)
  • Stretchable ionics

"stretchable ionics" was the working title of early drafts of the paper.

The Editor indicated that the two acronyms were distracting. The phrase "stretchable ionics" made into the published paper, but the sticky leader was gone.

In addition to being distracting, the acronyms also understate the capabilities of ionic conductors.  The acronym STIC misses many other attributes of ionic conductors.  A fuller list might be

  • solid (retains shape and can be patterned, printed, multilayered)
  • stretchable
  • transparent
  • biocompatible
  • biodegradable
  • lasting (think about supercapacitors)
  • diverse (there are infinite many species of ions)
  • high speed (you step on a nail, and you lift your foot rapidly)

The acronym LEADER places undue emphasis on a particular configuration.  One can of course think of dielectric-electrolyte-electrode hybrids in many configurations other than the layered geometry.

The acronyms do serve some functions in their restricted ways.  STIC lists the key attributes demonstrated in this paper.  LEADER is a configuration that has interesting properties. 

Very interesting

Very interesting work!

The instability problem must be very meaningful.

The deformation looks like inhomogeneous, whether it is easy to control the deformation.

Zhigang Suo's picture

Commercializing dielectric elastomer transducers

In June 2011 I wrote a post on Pulse, a commercial product based on dielectric elastomers.  The product came out on Amazon, but has now been discontinued.  A principal developer of this product was Artificial Muscle Inc., which was founded in 2004.

A different product has come out from a startup company StretchSense.  Yesterday Christoph and I met Ben O'Brien, the CEO of StretchSense.  Ben met us before when he was a brilliant graduate student of Iain Anderson.  StretchSense produces highly stretchable strain sensors.  These sensors are needed for soft robotics, stretchable electronics, and bioelectronics.  The function is achieved with difficulty with other technologies, but is achieved readily using dielectric elastomers.  Be sure to watch this YouTube video produced by the company.

Another recent startup is Compliant Transducer Systems led by Gabor Kovacs.  Here is a YouTube video of their stacked actuator.

It will be useful to keep track of these commercial developments.  Please leave comments when you find more.

Xuanhe Zhao's picture

The applications of

The applications of stretchable ionics may go beyond electrodes for dielectric elastomers. For example, SPS Strategic Polymers is developing new technologies and products of PVDF-based dielectric polymers pioneered by Prof. Q.M. Zhang's group.While the PVDF-based polymers are mostly in semicrystalline state, the structure of PVDF actuator is the same as dielectric-elastomer actuator with a polymer film sandwiched between two electrodes. There is a video on the company's website that demonstrates a prototype product.

Zhigang Suo's picture

Pairing ionic conductors with dielectrics of all kinds

Dear Xuanhe:  Thank you so much for pointing to Qiming's company and to PVDF.  You make an excellent point: ionic conductors can be paired with dielectrics of all kinds.  Ionic conductors have attributes such as stretchability, transparency, and biocompatibility, but they will also bring challenges whenever a  material is newly introduced in a new application, such as compatibility and reliability.  One has to see if the new attributes are worth the trouble.

In writing the paper, we found prior work that used PVDF, a piezoelectric polymer, to make thin-film loudspeakers (Ref 21 and 22 cited in our paper).  We wrote the following. "Pairing the ionic conductors with transparent piezoelectric polymers will take advantage of the high transparency of the ionic conductors; piezoelectric polymers can be operated with lower electric fields compared with dielectric elastomers and have a linear relation between electric field and strain (21, 22)."  Not all attributes of ionic conductors are significant in a given application.  In this case, large stretchability is unimportant.

Several students here have been looking into a suitable choice of PVDF.  Your comment is very timely.  Thank you.

But more generally, one should think about pairing all kinds ionic conductors with all kinds of dielectrics for all kinds of devices.  This space is very large, and finding some particularly worthwhile combinations is really an interesting task.'s picture

A few days ago I met

A few days ago I met Micheal Wang, who is a mechanical engineer working on hard robots for many years. He showed great interets at soft robots based on DE technique.

 He gave us a talk, showing the traditional way of studying hard robots, as well as applications and commercialization. As he pointed out, besides of the beautiful idea of conceptual design, the key to push soft robots to application should be the sophiscated but very effective way of modelling on functionality and controlling, as what they did on hard robots. This idea inspires me and we are considering the further collaboration.

Zhigang Suo's picture

More on commercialization of dielectric elastomers

Iain Anderson e-mailed me and pointed out that Artificial Muscle Inc. are busy with the ViviTouch, a heptic technology based on dielectric elastomers.  Watch a video of the ViviTouch technology. 

Zhigang Suo's picture

Further more on commercialization of dielectric elastomers

Gabor Kovacs e-mailed me saying that the startup company Optotune is producing and selling successfully a product based on dielectric elastomers:  transmissive laser speckle reducer.

Benjamin OBrien's picture

RE: Commercializing dielectric elastomer transducers

Hi Zhigang, thanks for the intro! I think that the more
companies that can start up the better. This will provide more funding for
researchers, drive up the profile of the technology, and bring economies of
scale. It is an exciting time!

Zhigang Suo's picture

why sensors based on elastomers are important

Christoph points me to this interview with Ben O'Brien, the CEO of StretchSense.  Ben talks about why sensors based on elastomers are important.

Zhigang Suo's picture

Reuters video on stretchable ionics

This video on our work was produced by Ben Gruber, the reporter, of Thomson Reuters.