Ionic electroactive polymer metal composites: Fabricating, modeling, and applications of postsilicon smart devices



Smart systems adapt to the surrounding environments in a number of ways. They are capable to scavenge energy from available sources, sense and elaborate external stimuli and adequately react. Electro Active Polymers are playing a main role in the realization of smart systems for applications if fields such as bio inspired and autonomous robotics, medicine, and aerospace. This paper focus on the possibility to use Ionic Polymer Metal Composites as a class of materials relevant to the realization of post silicon smart systems. The three main aspects of this new technology, i.e., fabrication methods, modeling, and applications are described with emphasis to most recent results. Attention is given to main challenges and shortcomings to be solved for technology, modelling, and control of IPMC based devices that need to be solved before this new technology can be fully exploited in real world applications. © 2013 Wiley Periodicals, Inc. J. Polym. Sci., Part B: Polym. Phys. 2013


An ever growing number of applications that once seemed impossible are becoming available at an impressive rate, because of the continuous advances in materials, hardware, and software technologies. It is possible to envisage a trend toward the development of smart systems, capable to solve even more complex problems with little or no human intervention in strategic fields such as bioinspired robotics, aerospace, and medicine, just to mention a few.1 New smart systems will be required to embed a number of different functions including, electric power generation and storage,2 signal sensing and processing, and actuating capabilities.3

This target will be partly achieved in the framework of the Moore Law's, that is, a continuous growth of electronics performance, in terms of speed increment and size reduction. Nevertheless there is a general consensus on the needing to develop “more than Moore” technologies, based on novel materials to obtain a significant diversification.

Polymeric materials have already shown to have the required functionalities and therefore will play a main role in the development of smart systems.4–6 Studies have been reported, for example, on the capability of this class of materials to scavenge energy from surrounding environment,7–10 to realize all organic electronic devices,11 and to obtain reversible energy transduction.4, 5 Though such capabilities can be obtained by using more traditional inorganic materials, these suffer for a number of shortcomings that do not plague polymers: for example, power scavenging can be obtained, just to give an example, by using piezoelectric ceramics but they are very rigid and fragile. In the same way signal processing can be obtained by using very fast silicon based electronics that unfortunately need to be mounted on rigid cards, while a significant interest exists for flexible electronics.12

This article will focus on Ionic Polymer Metal Composites (IPMCs), a class of composite materials that transform electrical stimuli into a mechanical reaction and vice-versa and that belongs to the wider class of Electroactive Polymers (EAPs). Nevertheless polymers exist that react to non-electric stimuli, for example they can be excited by using thermal, chemical, or magnetic stimuli, and/or react by producing non electrical outputs.

Though a lot of scientific articles have been produced on IPMCs since 1990s, including survey articles focused on topics relevant to this novel technology, the huge number of contributions published each year justifies the opportunity to reconsider the whole scientific production. If scientific data bases are investigated on IPMCs it emerges an ever growing interest paid to technologies and applications. This article proposes a survey on IPMCs focusing on a novel point of view: the capability of IPMCs to be used as components relevant to the realization of smart devices is investigated along with a careful analysis of more recent scientific contributions.

IPMC are quite new materials and of course many efforts are still required before they can become a mature technology, suitable to realize ready to market products. To this aim attention is given also to main challenges and shortcomings to be solved for technology, modeling, and control of IPMC based devices, before they can be fully exploited in real world applications.

The following of the article is organized as it follows: in Electroactive Polymers and Ionic Polymer Metal Composites section, an introduction to EAPs and IPMCs is given. In Ionic Polymer Metal Composites Fabrication section, the fabrication methods used to realize and to improve IPMCs are described. In IPMC Models section, a taxonomy of models available for the description of their behavior will be drawn with a focus on efforts to include such models in the framework of controlled systems. In Applications section a number of applications will be described. Conclusions section will deal with final conclusions and discussion on challenges and possible evolutions for this technology.


Polymers are experiencing an endless fortune because of a number of interesting properties that spans from low cost of base materials and easy of production, that allow to produce cheap devices, to impressing flexibility and light weight, electrical and thermal insulation capabilities, good resistance to aggressive chemicals, with numberless possibilities of applications. Also, the increased knowledge about the dependence of polymer chemical/physical properties on their molecular structure has given the possibility to tailor the properties of new synthesized materials, with a burst to envisaged applications.

In last decades electrical stimuli-responsive polymers have been discovered and their use as electromechanical transducers has been shown.6 Moreover the transduction capabilities of polymers can be reversible13 so that they can sense external stimuli and react to them, for example by changing their shape or size: they are not anymore passive tools or simple coating elements but cooperate to attain men's objectives.

The interest toward EAPs has risen at an impressive rate in the last two decades since they are capable to realize artificial muscles, that is, actuators that produce very large mechanical reaction to electrical stimuli while being light, soft, and very quiet. As an example, in Figure 1 two snapshots of the deformation produced when an EAP is connected to a sinusoidal input voltage with a 4 V peak amplitude, and 0.25 Hz frequency, are reported.

Figure 1.

Two snapshots of the deformation produced by an EAP using a sinusoidal input voltage with a 4 V peak amplitude, and 0.25 Hz frequency.

The history of EAPs is much more ancient. As a matter of fact the very first experiments on EAPs are due to W.C Roentgen, who in 1880 observed large mechanical reactions produced on natural rubber because of electrical stimuli. In that same year Jacques and Pierre Curie discovered the piezoelectric effect for quartz crystals.14 This coincidence is partly obscured by the fact that piezoelectric non polymeric materials were since then much more investigated than EAPs: the first piezoelectric device was proposed during the World War I, when the underwater sonar was proposed for submarine detection by P. Langevin.

EAPs are classified, according to their operational mechanism, into two main classes, namely electronic, or dry, EAPs and ionic, or wet, EAPs. Transduction capabilities in the first class are due to Coulomb force, with no mass transport, and this results in high actuation force and fast response time (typically in the order of milliseconds), though they require very high activation voltages and produce quite small deformations. Most investigated electronic EAPs are either ferroelectric polymers, such as polyvinylidene fluoride (PVDF), or dielectric elastomers (D-EAPs). A very nice review on dielectric elastomers can be found in Ref. 15.

In ionic EAPs the electromechanical transduction property is obtained by a net migration of ionic charges, under the effect of an applied stimulus. The charge migration eventually occurs in cooperation with some carrier or solvent. Traditionally such solvent was pure water and this deserved to this class of materials the name wet EAPs. Such characteristic was exploited in a number of underwater applications where the hydration of the devices is fully guaranteed. Applications in dry environments have anyway been proposed and the possibility to use solvents other than water has been successfully investigated.

Ionic EAPs are of great interest because they can produce very large deformations when few volts are applied. As a matter of fact, they are also slower (typically fraction of a second), and have a smaller energy density with respect to dry EAPs.

Among wet EAPs the most widely investigated classes are Carbon Nanotubes (CNTs), Conductive Polymers (CPs), and IPMCs. CNTs can be classified into single-wall CNTs (SWNTs) or multi-wall CNTs (MWNTs) according to the number of graphene layers. In Figure 2(a) schematic views of 0D to 3D structures that can be obtained by using a 2D graphene layer are reported. In Figure 2(b) a schematic view of a MWNT is shown.

Figure 2.

A graphene layer and different structures that can be originated (a) Reprinted by permission from MacMillan publishers Ltd: Nature Materials,16 copyright 2007. A schematic of a MWNT (b); available at:

CPs are a class of organic materials with an unusual combination of plastic characteristics and tailor-made semiconducting or conducting electrical behavior.

IPMCs are composite materials made of a polymer (also called ionomer or ionic polymer) since it contains ions that are weakly linked to the polymer chain. The polymer is metallized via a chemical process, on both sides, with a noble metal (usually platinum or gold), to realize the electrodes.

The following of this article will focus on IPMCs and their evolution as suitable candidates to realize electromechanical reversible transducers that can be used both as motion actuators and sensors.

If an electric field is applied across their thickness, they undergo a bending deformation or develop a blocked force, if a mechanical constrain limits the deformation. It is not unusual to observe IPMCs in beam configuration whose free end is tilted by 180° with respect to the IPMC base, when a voltage of few volts is applied (generally voltage values spanning from 1.0 V to 2.0 V are considered though experiments with voltage values up to about 5.0 V have been described). For the case of the blocked force values ranging from about 1.0 mN/V to 10 mN per volt are reported,5 where the dispersion of observed values partly depends on differences among investigated devices and partly on the undesired lack of reproducibility of IPMCs.

On the contrary, if an IPMC is dynamically deformed an mechanoelectric transduction is observed and an electrical signal is generated at the electrodes because of the ionic migration, due to imposed stresses.5–13 The sensing capability is strongly frequency dependent and the sensor output signal can be either an open circuit voltage (peak values of few millivolts have been reported for the case of an impulsive mechanical displacement5) or a short circuit current (a value of few milliamperes per meter have been reported for the case of a sinusoidal input17), depending on the used conditioning circuit.

A schematic view of the transduction phenomenon is shown in Figure 3.

Figure 3.

The transduction phenomenon for an IPMC sample.

In the literature two different approaches have been proposed so far to explain the transduction capabilities of IPMCs. In more details, the IPMC transduction behavior has been explained both as a consequence of the solvent (water) molecules migration, dragged by mobile ions, or by forces that act on the polymer backbone, because of the charge unbalance.

The nature of the used cations has a big influence on the performance of the IPMC actuator. In fact, it has been demonstrated that cations with larger hydration numbers can more effectively drag water molecules and then produce larger actuations, while the size of counterions can affect the IPMC actuation speed. Finally it has been reported that the counterion type can affect the back-relaxation phenomenon18 observed of IPMC based actuators: when a step input voltage is applied to an IPMC, this undergoes a quite quick deformation toward a steady state position, then a slower back relaxation to the starting configuration is observed. Studies are reported in the literature showing the main role of the counterion on the entity of back-relaxation or even on its presence.19 More details on the IPMC transduction mechanisms will be given in IPMC Models section.


There are a number of different types of ionic polymers available but the typical IPMC, used in the most reported investigations, is composed of a perfluorinated ionomer membrane, usually Nafion® or Flemion®, which is surface-composited by platinum via chemical process.

In Figure 4(a) a schematic view of an IPMC is reported, while a corresponding microscopic view is seen in Figure 4(b). It is possible to observe that the chemical realization of the electrodes produces a dendritic structure of the electrode-ionomer interface. IPMCs are generally soaked by using a suitable solvent, such as water.

Figure 4.

The typical structure of an IPMC. A sheet of Nafion® and platinum made electrodes realize an IPMC (a). A microscope view of an IPMC, where the irregular electrode-ionomer interface is seen (b).

The most part of research on IPMCs has been carried out on Nafion®117 based membranes. Nafion® is available in the form of unreinforced films of different thickness, reinforced films, and even a liquid solution. Nafion®117 is an unreinforced film whose typical thickness is 178 μm. Since metallic electrodes are tens of micrometers thick, IPMCs typical thickness is about 200 μm.

Developed in the 1960s by DuPont and commercialized in the 1970s, Nafion®117 is a Teflon-based polymer with short side-chains, terminated by ionic groups (typically sulfonate or carboxylate (SO3 or COO for cation exchange). These acidic side groups dissolve when the material is hydrated, causing the membrane to swell significantly (absorbing as much as 38% of its dry weight in water).20

This dissolution frees the cations, associated with each pendant acidic group, that can move within the polymer matrix, while the anions are covalently fixed to the fluorocarbon backbone. Due to its Teflon-like structure, the perfluorinated sulfonic acid polymers exhibit excellent chemical and thermal stability.

In order to impose or collect an electric signal across the thickness of the IPMC, both faces of the polymeric membrane must be plated with conductive electrodes. The most used methods to form metallic electrodes onto ionic polymer membranes are based on an ion exchange process and subsequent chemical reduction.5, 21–34 Electrodes are generally realized by using noble metals, such as platinum or gold, though other metals, such as silver, copper or palladium, have been proposed.

The manufacturing technique requires four basic steps:

  • 1A surface treatment (or roughening) of the polymer surface;
  • 2The adsorption (or ion exchange) of metal ions in the polymer;
  • 3The reduction (or primary plating) of the ions of the metal to the metallic state;
  • 4Developing (or secondary plating) of the electrodes.

Actually, the first step is a pretreatment of the ionomer surface in order to increase the surface area density, where the metal salt penetration and reduction occur, and to maximize the interfacial area between the polymer and metallic layer. It is realized by sandblasting or sandpapering the polymer surface, followed by an ultrasonic cleaning (to remove any residue on the surface) and chemical cleaning in a boiling acid solution (HCl or HNO3, at low concentration) for 30 min. The acid wash is intended also to ensure that the polymer is fully saturated with protons. The membrane is then rinsed with bi-distilled water to eliminate any residual acid and to swell the film.

The adsorption (or ion-exchange step), is performed to obtain the absorption of metal ions on the ionomer film surface. It is realized by soaking the ionic polymer in a salt solution (e.g. Pt(NH3)4HCl)) to allow the diffusion of the metal cations, via the ion exchange (expected metal adsorption on ionic polymer is about 3 mg/cm2). The films are left in the solution at room temperature for more than 10 h.

The third step is the reduction (or primary plating): during this step the reduction of the cations absorbed into the ionic polymer to the metallic state in the form of nanoparticles is obtained. The reduction can be realized using different reducing agents depending of the used cations (generally NaBH4 or LiBH4 are used), at a suitable temperature. During this phase metal layers deposit near the surface of the membrane and Pt+ counterions in the ionomer film are replaced by Na+ or Li+, for the reducing agents mentioned above. As mentioned in a previous section, the counterion type is not of secondary interest, neither its choice is limited to sodium or lithium.19, 33

The last step is the developing or secondary plating: metal is grown on top of initial metal surface in order to reduce the electrodes resistance. Additional amount of metal is plated on the deposited metal layers.

Optimization of IPMCs Manufacture

The fabrication of IPMCs can be optimized by changing some parameters in the manufacturing procedure. A list of such parameters includes bath temperature, nature of metal to be used for the electrodes, concentration of the metal containing salt, reducing agents, number of sequential cycling of adsorption/reduction, dispersing agents and co-reduction of different metals. In the following the influence of some of these parameters will be analyzed.

Physically Loaded and Interlocked (PLI) Electrode System

The standard IPMC manufacturing procedure has relatively high cost due to the use of noble metals and associated complex chemical processes, so a careful control of laboratory procedures is necessary. To reduce manufacture cost, some authors35, 36 reported a novel method of physically loaded and interlocked (PLI) electrode realization. This technique consists in physical loading by hot pressing of a conductive primary powder (Ag was used in the referred case) into the ionic polymer network forming a dispersed particulate layer. This primary layer functions as a major conductive medium. Subsequently, this primary layer of dispersed particles of conductive material is interlocked within the polymer network with smaller secondary particles (e.g. platinum or palladium) via chemical process. As a result both primary and smaller-interlocking particles are secured within the polymer network and the intrinsic contact resistances between large primary particles is reduced. Furthermore, an electroplating process can be applied to integrate the entire primary and secondary conductive phases, serving as another effective interlocking electrode. A schematic view of the resulting electrode is seen in Figure 5.

Figure 5.

A schematic process illustration of a PLI-IPMC.36 Reprinted from Sensors and Actuators: A Physical, 96, Shainpoor, M; Kim K.J., Novel ionic polymer-metal composites with physically loaded particulate electrodes as biomimetic sensors, actuators and artificial muscles, 122–135, Copyright 2002, with permission from Elsevier.

Sequential Cycling of Adsorption/Reduction

It is known that a large conductor–polymer interface area influences the capacitance of the electric double layer: adding more particles in the electrode, the capacitance and then the performances of the device increase.30, 34, 37–43 In an IPMC, the amount of particles and the interfacial area increase with the thickness of the electrode, so it is necessary to increase the number of Pt electroless plating steps during the manufacture procedure to obtain a device with better performances. This can be realized by adsorption/reduction steps performed sequentially, up to five to seven times.

Reduction of Surface-Electrode Resistance

It is widely reported that particles of platinum, used for electrodes, are in a dense form with a corresponding significant level of surface-electrode resistance. Shahinpoor and Kim25 proposed to electrochemically deposit a thin silver or copper layer on top of the platinum electrode to reduce the surface-electrode resistance.

The IPMC manufacture consists of an initial compositing process and surface electroding followed by the straightforward electrochemical deposition of silver (or copper) on top of the IPMC. It requires a rectifier and silver (or copper) solution. The rectifier controls the DC voltages and currents within appropriate ranges. A schematic view of the resulting platinum/silver electrode is seen in Figure 6.

Figure 6.

A schematic diagram illustrating the silver deposition process on an IPMC.25 Reprinted from Shahinpoor, M.; Kim, K.J., The effect of surface-electrode resistance on the performance of ionic polymer-metal composite (IPMC) artificial muscles, IOP Smart Mater. Struct., 2000, 9, 543–551.

Silver and copper solutions were prepared by the dissolution of appropriate concentrations of the salts in water. Adequate actions are required in order to obtain a uniform and optimized thickness (approximately 1–2 μm) of the silver (or copper) layer. In the referred case this was obtained by controlling AgNO3 and CuSO4 solution concentration in water, deposition-time, and solution temperatures.

The silver surface is much brighter and smoother than that of the platinum based and shows a typical silver-like color. The copper layer became bluish due to copper oxidation (as expected) but the bonding between the platinum surface and silver (or copper) was favorable in both cases. Since the electrochemical process produced a continuous metal phase, this thin layer could reduce the surface resistance of typical IPMCs by a factor of approximately 10 (surface resistance values less than 1 Ω/cm were reported).


It is known that non-noble metals are unstable and prone to oxidation when used as electrodes for IPMCs44 and this is considered as a main drawback, since IPMCs are privileged candidates to realize systems to be used in wet environments or even in water.45 Nevertheless, considering the high cost of noble metals some authors46 developed a co-reduction process for plating ionic polymer materials that gives noble and non-noble metal alloy electrodes, with a resulting reduction in the consumption of expensive precious metals such as platinum and gold.

Though the concept of co-reduction is not a new one,47–49 Bennett and Leo46 developed a new process to reduce simultaneously a noble (Pt) and a non-noble metal (Cu) on the Nafion surface and to form an alloy sublayer consisting of two metal phases that appear to be homogeneously mixed.

The aim of the study was to minimize the amount of platinum present in the electrode while eliminating the problem of electrode oxidation. To have less platinum in the electrodes is advantageous for a number of reasons. Platinum is a very rare and expensive metal and moreover it has a density that is 2.36 times greater than the density of copper. This means that electrodes that contain large amounts of platinum will be heavier and, consequently, the overall mass of the ionic polymer device will increase while its mass energy density will decrease. Moreover, the electrical resistivity of platinum is over six times larger than that of copper, which will adversely affect the surface conductivity if large amounts of platinum are used.

The copper and platinum ions must be exchanged into the membrane simultaneously and then reduced simultaneously at the membrane surface in aqueous solution of sodium borohydride. For the copper/platinum electrodes manufacture, tetra-ammine platinum chloride ([Pt(NH3)4]Cl2) was used to source the platinum ions and cupric sulfate (CuSO4) was used to source the copper ions.

The proper ratio of ions to be mixed in the exchanging solution must be determined by changing the relative molarities of the ions and varying the soaking time during experimental investigations. In particular, results in Ref. 46 showed that Nafion® membrane has a much higher preference for the tetra-ammine platinum ion than for the copper ion. A plausible explanation for this phenomenon may be that the copper ion is more solvated, and thus the platinum ions should be preferred by the polymer.

Dispersing Agents

It is widely known that the actuation mechanism of IPMCs is enhanced by the presence of a suitable solvent and water is generally used to this purpose. Unfortunately, IPMCs dehydrate during their use so that, to achieve high efficiency, it is necessary to reduce or eliminate the water leakage out of the surface.

The loss of water from the polymer membrane is due to leakage out of the electrodes porous surface, electrolysis, and so forth. To avoid this phenomenon it is necessary to obtain a more uniform electrodes surface. Unfortunately, due to a coagulation process the nominal size of platinum particles near the boundary of IPMC electrodes is generally 40 nm to 60 nm and the surface of the electrodes is not uniform. A schematic illustration of platinum coagulation during the chemical reduction process is shown in Figure 7.5

Figure 7.

A schematic illustration of platinum coagulation during the chemical reduction process.5 Reprinted from Shahinpoor, M.; Kim, K. J., Ionic polymer-metal composites: I. Fundamental, IOP Smart Mater.Struct., 2001, 10, 819–833.

This phenomenon can be alleviated if dispersing agents are used during the chemical reduction process: the additives enhance the dispersion of metal particles within the ionic polymer, control their size and reduce coagulation. As a result, a better metallic particle penetration can be obtained in the polymer (the deeper the penetration, the lower the surface resistance) with smaller average particle sizes and consequently a more uniform distribution. Obviously, the more uniform the particle distribution the more difficult for water to pass through the electrode layer will be and a lower water leakage will result.

Some authors studied the influence of various types of dispersing agents, added during the chemical reduction process. In particular, polyvinylpyrrolidone (PVP) was used in Refs. 5, 21, 50, and 51, while the role of some surfactants as size controllers has been investigated in Refs. 52 and 53.

Dodecyl sodium sulfate salt (DSS) is one of the popular compounds used to this purpose. It gave good results on controlling the dimension of silver particles;54 more specifically, it was observed that silver size decreased from sub-micron size to nanometer size with increasing of DSS concentration.

Bonomo et al.41 studied the influence of the use of PVP and DSS on the performances of IPMCs, realized by suing Nafion®117 and Pt. The influence of PVP with different molecular weights (average Mw = 10,000 and 29,000) and varying the additives concentration was investigated. The additives were added either during the primary or the secondary plating. It was observed that PVP at a higher molecular weight worked better at higher concentrations, while PVP at a lower molecular weight gave better results at lower concentrations.

Manufacture of Tridimensional IPMCs (3D-IPMCs)

IPMCs generate low actuating forces compared to other technologies and this characteristic limits the possible fields of applications as actuators so that the possibility to improve developed force is of interest. It is known that the actuating force is related to the bending stiffness and the thickness of the ionic polymer membrane. To improve the performances of IPMCs in terms of the actuating force, it is necessary therefore to fabricate thicker materials, known as three-dimensional IPMCs (3D-IPMCs).


In Refs.[55]55 to[62]62 as-received ion exchange membranes were dissolved in appropriate solvents and these were successively evaporated to fabricate 3D-IPMCs.

Recast ion exchange membranes can be also realized. The preparation of solution recasted Nafion® film samples must be done with extreme care. The starting material is a liquid Nafion® solution generally containing 10% of Nafion® and 90% of solvent (approximately one-to-one mixture of alcohol and water). Unfortunately, the cast dispersion is very prone to cracking, producing micro and macro cracks upon drying after solvent evaporation. Therefore, it is required to introduce an additive (typically a low vapor pressure solvent as dimethyl formamide (DMF), dimethyl sulfoxide (DMSO), triethyl phosphate, ethylene glycol (EG)60, 61 that makes the solvent mixture acting like an azeotrope. More specifically, the use of DMF has been reported to be the most adequate choice.

The resulting membrane must be, eventually, annealed at an high temperature (70 °C) in order to develop some crystallinity. Usually, the temperature is raised to 150 °C for further curing. The annealing process can tailor mechanical and chemical stability to the solution recasted Nafion® film. In fact, without annealing, the ionomer would remain soluble and actuation would not be possible because it requires the membrane to be saturated with solvent. In referenced contributions multiple layers of liquid Nafion® were dried one on the top of the other, and samples approximately 2 mm thick were prepared successfully. Pt based electrodes were then realized by using the absorption/reduction fabrication method.

3D-IPMCs can be also obtained starting from Nafion® liquid solution by using an alternative technique.60, 61 In that case, a metallic conducting powder (e.g. silver, platinum, palladium, gold, copper, and any other conducting powders) was mixed, at first, with the electroactive polymer solution. The fine powder was uniformly dispersed within the solution. The resulting mixture was casted as thin layer and then the electroactive polymer solution underwent the drying process of solvents. The metal conducting powder was dispersed, at that stage, within the polymer. The electroactive polymer solution (without the metallic conducting powder) was then added on the top of the first layer and dried and the procedure was repeated until the desired thickness was obtained. A layer of the metal conducting powder was successively formed by the same method described above. As a final step, the ion conducting powder-coated electrode was cured under elevated temperature. In Figure 8 a process diagram for the ion conducting powder-coated electrode made by the solution recasting method is shown.60 If necessary, the surface conductivity can be enhanced by adding a thin layer of novel metal via electroplating.63

Figure 8.

A process diagram for the ion conducting powder-coated electrode made by the solution recasting method.60 Reprinted from Kim, K. J.; Shahinpoor, M., Development of Three Dimensional Polymeric Artificial Muscles, Proc. SPIE's 8th Annual International Symposium on Smart Structures and Materials, 4329, 2001, 223–232.


Recasting techniques, described so far, suffer because of low reproducibility: they require delicate tuning of process variables (temperature, nature and concentration of the solvent, etc.). Some authors proposed a new method to obtain thick IPMCs starting from Nafion® films, known as the hot pressing method:42 a hot pressing system is used to make several Nafion® films adhere together. This method allows controlling the thickness of the IPMC with a good reproducibility.

The hot-pressing method consists of two steps: the former is a melting step which makes the interfacial layers between the films improve the adhesive characteristics; the latter is a pressing step which makes the stacked films adhere to each other. The 3D-Nafion specimen can be prepared varying the number of Nafion® 117 films (three to five sheets are usual) and using different pressure and temperature values.41, 42 After cooling the stacked film and detaching from the mold, the 3D-Nafion samples need to be cleaned by successive heating in sulfuric acid (3 wt %), distilled deionized water, hydrogen peroxide (10 wt %) and distilled deionized water to remove impurities.

The 3D-IPMCs can be manufactured by immersing the 3D-Nafion® specimen in an aqueous solution of platinum ammine complex ([Pt(NH3)4]Cl2) for 1 to 2 days to absorb Pt ions in the stacked film. This step is followed by primary and secondary plating. To improve the performance of the electrode, some authors applied additional postprocess consisting in the increase of the number of the Pt electroless-plating cycles41, 42 and in the use of dispersing agents.41 In these cases, it was possible “to modulate” the properties of the 3D-IPMCs as a function of the number of Pt plating, of the nature of dispersing agent and of its concentration.

Non-Water Based IPMCs

Water has been proposed since the beginning as the most obvious solvent to be used with IPMC actuators. The main reasons for the preference of water as the solvent are the favorable interaction of the Nafion® polymer with water and the low viscosity of water. Generally, Nafion® membranes will absorb up to 38% of their dry weight in water and achieve an ionic conductivity of 83 mS/cm when fully hydrated.20

Unfortunately water-swollen membranes lose the solvent because of water volatility: when IPMCs are used in air there is a continue dehydration due to water evaporation.

IPMC actuators can lose water also because of the small electrochemical stability window of water. In fact, due to electrolysis process, the water will separate into gaseous hydrogen and oxygen when the applied voltage is greater than 1.23 V. Both water evaporation and decomposition lead to water loss and, hence, result in a deterioration of the actuation performance.5

One way around this problem would be to use coating a barrier to contain the water inside the membrane. Shahinpoor and Kim63 reported the use of some barrier coatings (e.g. parylene, silicone rubber, isotactic polypropylene, 1-dodecanethiol); however, a barrier coating will add passive stiffness to the actuator device and reduce the amount of strain that the device can generate.64 Furthermore, if the electrolysis limit of the aqueous solvent is exceeded, the formation of hydrogen and oxygen gases at the membrane surface will create blisters beneath the barrier coating and will lead to the coating delamination from the transducer. Research activities have, therefore, been carried out to identify solvent systems other than water that while enhancing the transduction can reduce or eliminate the problems associated with the use of water.

Suitable solvents might have a low vapor pressure and a large electrochemical stability window and a high ionic conductivity. Hydrophobicity is also of importance, as ionic polymers readily absorb moisture from the air, which eventually leads to the same problems associated with using water as solvent. Suitable candidates are ethylene glycol (EG) and ionic liquids.

Ethylene Glycol

Ethylene glycol or 1,2-ethanediol (C2H6O2) is an organic solvent consisting, as water, of polar molecules. It can be used over a wide range of temperatures and, commonly, also as an antifreezer. Some of the physical properties of EG are listed in Table 1. IPMCs with ethylene glycol as the solvent have shown large solvent uptake, and can be subjected to relatively high voltages without electrolysis. They can be actuated in open air for rather long time intervals, and at low temperatures. They may be good actuators when high-speed actuation is not necessary; in fact, the reaction speed is limited by the high viscosity of EG (about 16 times higher than that of water at room temperature) and by the higher molecular weight with respect to that of water.65–69

Table 1. Physical Properties of Ethylene Glycol and Glycerol
inline image

When ethylene glycol is used as the solvent, IPMCs need to be preventively dried, to remove any water within the sample. Samples are put in a drying chamber at 100 °C for 1 or 2 days. Samples are wrapped between two filter papers and put in a container and a vacuum pump is connected to the container to take out any air in the container that might carry water vapor. Samples are then soaked overnight in a beaker containing pure EG in a 60 °C water bath.

Some authors70, 71 presented the results of a series of tests on Nafion® based IPMCs using as solvent, instead of EG, glycerol (1,2,3-propanetriol, C3H8O3). This is another polar solvent with high viscosity (1000 times the viscosity of water) (Table 1). To have fully solvated samples with glycerol, dried samples are soaked for about 8 h, in a beaker containing pure glycerol and immersed in the 70 °C water bath.

Ionic Liquids

A number of researchers replaced water in IPMCs by ionic liquids.72–80 Ionic liquids are salts containing organic cations and (mostly) inorganic anions. They exist in their liquid state at room temperature. There is no formal definition of the minimum melting point for a compound to be described as an ionic liquid; many researchers use 80 °C, although some compounds with melting points as high as 100 °C are referred as ionic liquids.81, 82

Ionic liquids have an immeasurably low vapor pressure, due primarily to the bulky and cumbersome structure of the corresponding ions, which inhibits the formation of a crystalline solid: this means that they will not be lost because of evaporation. Also, they are able to dissolve a wide range of organic and inorganic compounds, including some polymers and minerals, are non-flammable and have a high thermal stability.

Electrochemical stability windows of ionic liquids is 4 V or more, and they are thermally stable to temperatures as high as 400 °C. Furthermore, ionic liquids have high ionic conductivities and can be used as electrolytes for a variety of applications, including electrochemical capacitors83 and conducting polymer actuators.84, 85

Ionic liquids as solvents for IPMC transducers have been proposed since they are very stable and therefore will eliminate the problem of solvent evaporation, thus allowing the use of IPMC transducers in a broader range of environments. Also, the ionic liquids are themselves ionically conductive and should therefore facilitate ionic motion in the Nafion® membrane, and may enhance the ion motion as compared with water. Furthermore, the electrochemical stability window of most ionic liquids is larger than that of water. It is therefore possible to apply larger actuation voltages to IPMCs, thus increasing the available energy density of these devices. A list of ionic liquids and their physical properties are reported by Aldrich86 but the most common classes of ionic liquids used in IPMCs are those based on substituted imidazolium cations.

More specifically, 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide (EMI-Im),74 (1-ethyl-3-methyl-imidazolium tetrafluoroborate (EMIBF4),69 1-ethyl-3-methylimidazolium trifluromethane sulfonate (EMI-Tf),72, 74 1-butyl-3-methylimidazolium tetrafluoroborate (BMI-BF4),73, 80 1-butyl-3-methylimidazolium trifluoromethanesulfonate (BMI-Tf),80 and 1-buthyl-3-methyl-imidazolium hexafluorophosphate (BMIPF6)73 were the most used ionic liquids as the solvents for IPMCs.

In order to utilize an ionic liquid as a solvent for an ionic polymer transducer, a Nafion® membrane is typically first plated with metal electrodes by using the traditional impregnation/reduction process. The obtained IPMC is, therefore, fully saturated with water. Following the plating process, in order to impregnate these materials with ionic liquid, the first step is to completely remove the water from the membranes. This is done by baking the membrane in a vacuum oven at a temperature of 75 to 85 °C and a pressure of about 3333 Pa (absolute) for at least 1 h. After dehydrating the polymer samples, they are placed into a mixture of methanol (30–40% by weight)/ ionic liquid72 and sonicated at 65 to 80 °C.

The diffusion of the ionic liquid into the Nafion® membrane depends on the ionic liquid used, but is typically slow and therefore high temperatures (>80 °C) and long soaking times (>4 h) are required.

Upon removal from the ionic liquid/methanol mixture the samples are blotted with a clean filter paper to remove residual solvent on the surface. Doyle et al.87 have reported that the uptake of these ionic liquids into the membrane is proportional to the temperature at which the swelling is carried out. Following the swelling of the membranes with ionic liquid, they were dried under vacuum at 110 °C for 3 h to remove the methanol.

An alternative procedure73 used a mixture water/liquid ionic, (at room temperature for 48 h). After ion exchange, the IPMCs were dried at 90 °C for 30 min and stored in vacuum chamber.

Mixed Solvent: EG/Ionic Liquid

Some authors88 focused on non-water based IPMCs that use a mixed solvent obtained by using ethylene glycol and ionic liquids. The main idea for this research was to gain synergic effects from the mixed solvents: in fact, a large amount of ethylene glycol can be absorbed by the membrane, but the electric conductivity of ethylene is very low, while hydrophobic-ionic liquids have relatively high electric conductivities, but are very poorly absorbed by the membrane. Both the ionic liquid and ethylene glycol are miscible, so, using mixed solvents, ethylene glycol can act like a link or bridge between the membrane and the ionic liquid. Therefore, more of the ionic liquid can be absorbed by the Nafion® membrane.

To realize the device an IPMC with water as the inner solvent was dried at a temperature of 80 °C and under a vacuum pressure of 105 Pa for at least 3 h. The dried IPMC was then immersed into a mixture of an ionic liquid (1-ethyl-3-methylimidazolium trifluromethane sulfonate) and ethylene glycol at a temperature of 80 °C for 12 h (overnight). Different composition of the mixture of ionic liquid and ethylene glycol were used (30–40% weight of ethylene glycol). The solvent uptake level of the IPMC was determined by measuring the weight of the IPMC before and after the immersion.

Ionic Polymer-Polymer Composites

An improvement of IPMCs has been recently introduced by some of the authors and it is described in the following since it is a relevant evolution toward the realizations of polymeric smart devices. In fact they are all-organic transducers, named ionic polymer-polymer composites (IP2Cs) with electrodes realized by organic conductors.89–93 This class of transducers exploits the properties of organic conductors that are characterized by high conductivity values. They represent a further step toward the realization of all polymeric smart devices, with potential applications in various application domains from electronics to robotics or biochemistry. In fact organic conductors are mainly processable from solutions and allow the use of low cost deposition techniques such as spin coating, dip-coating, spin-casting or printing techniques.

In Refs. 91 to 93 all-organic devices were realized starting from Nafion® 117 and different manufacture procedures have been followed “to build” the electrode based on conducting polymers: the first method consists in the drop-casting deposition technique, the second is based on the polymerization in situ.

Regarding the drop casting, commercial formulations of a conducting polymer based on poly(3,4-ethylendioxytiophene)-poly(styrenesulfonate) (PEDOT:PSS) were used (see Table 2).94 The basic procedure to fabricate the IP2Cs requires to rough the film by sandblasting, to have a better mechanical adhesion of the organic conductor on Nafion® film, and to apply over both sides of the membrane the PEDOT:PSS by the drop-casting technique. The membranes are then dried by heating them at 60 °C for 2 h.

Table 2. Adopted Commercial Conducting Materials Based on PEDOT:PSS
inline image

When ethylene glycol or ionic liquids are used as solvents, before drop casting it is necessary to dry Nafion® to remove water, to soak the membrane overnight in the chosen solvent and to heat for some hours.

To improve the Nafion®/conducting polymer adhesion and to avoid the degradation of the electrodes during the rehydration, the manufacture procedure was improved and the organic conducting polymer (PEDOT:PSS) was polymerized directly on the Nafion® surface by oxidation of 3,4-ethylendioxytiophene (EDOT) in presence of sodium polystyrene sulfonate (NaPSS) in aqueous dispersion.89 Different times of polymerization (1/4 h, ½ h, and 1 h) were used to “modulate” the performance of the devices; in fact, the thickness of the organic electrode is function of the polymerization time and, consequently, influences both device stiffness and electrode conductivity in IP2C. In particular, to manufacture the IP2C, the Nafion® film (pre-treated by successive boiling (1 h) in 1M H2SO4, bidistilled water, 3% H2O2 and again in bidistilled water) was soaked in a solution of EDOT and NaPSS. Solid Fe(NO3)3 9H2O was added and stirring was continued for different times (1/4 h, ½ h and 1 h). The formation of PEDOT resulted by the color change of the transparent Nafion® into a dark blue-black layer being deposited on both sides of the membrane. The membrane was well rinsed with bidistilled water and then was boiled in 1M H2SO4 for 1 h (in order to perform the exchanging of Fe3+ with H3O+ ions), and finally in H2O (1 h). In order to obtain IP2Cs with EG as the solvent, water was removed drying the devices at 100 °C for 24 h, soaking IP2C overnight in a beaker containing EG and then heating them to 60 °C for 1 h.

A comparison of described production strategies, along with corresponding advantages and shortcomings is shown in Table 3.

Table 3. Classification of IPMC and IP2C Manufacture Techniques
inline image

The description of production procedures of IPMCs shows a common needing to improve some aspects of IPMC transducers. The main challenging aspects that should be solved in the near future before IPMC can become a mature technology are linked to the poor repeatability of electromechanical transduction behavior. It is likely that this largely depends on the chemical/physical actions that are required to “prepare” bare Nafion® to the realization of electrodes, that can greatly change its pristine properties, and on the chemical realization procedure of electrodes, that does not allow a tight control of their realization. A significant improvement could be obtained by using direct electrode deposition techniques, as those available for conducting polymers, since they allow a tighter control of the electrode realization.

Also, the loss of solvent has been largely investigated and partially solved by using solvents different from water. Another path that should be better investigated is represented by IPMC encapsulation without a significant reduction of transduction performance.

Though the increase of delivered force has been also considered as a drawback for IPMC actuators there are still a lot of applications (floating underwater robot, microgravity environments) where the power increase is not a fundamental issue, while the IPMC lightness and softness are considered valuable properties.


The full exploitation of a novel technology passes through the intermediate step of the comprehension of involved phenomena and of the capability to describe such phenomena by using suitable mathematical models. Such models are useful when applications are designed and become a necessary tool when complex applications are of interest, so that it is mandatory to be able to predict in advance the behavior of the device under design. Mathematical models are required that describe the behavior of the devices under investigation, with a level of accuracy imposed from the intended application.18

A quite general path in system modeling starts from models that, on the basis of experimental investigation, try to describe the observed data, to arrive at models that guarantee a larger level of generality. This evolution allows obtaining models able to predict the behavior of devices different from that under investigation or in working conditions different from those experimentally observed. Such models are a formidable tool for the user, since it is possible to predict the effects of any choice during the design step of devices and are required, for example by a number of controlling strategies that are used to modify the behavior of, otherwise, unsatisfactory performing systems.

The modeling of IPMCs has followed, in last two decades, this evolution and a significant improvement can be observed, from pioneering works to more recent models that have an ever more increasing level of complexity and accuracy. Moreover, models have been proposed that are intended to describe the IPMC transducers, their interactions with the surrounding environment and IPMC based systems. Such models are the result of efforts to develop IPMC based smart systems.

Nowadays tens of models have been proposed to describe IPMC transduction capabilities and a taxonomy is useful to help understanding the main differences among them. A way to classify models is into black box models, gray box models, and first principle or white box models. Interested readers can find a good overview on the theory of system modeling in Ref. 95. Such classification will be used in the following of the section to review the IPMC models proposed in the literature. Moreover some of the proposed models are capable to describe both the IPMC behavior working as sensors and as actuators, while others describe only one of the two possible functioning modes. Such a characteristic will be outlined in the following of the section for the described models.

Further models that have been developed to approach specific fields of IPMC applications will be described in dedicated sub-sections.

Black Box Models

Black box models are mathematical models that approximate data recorded during experiments on the device under test by using suitable identification strategies. The objective of black box models is to fit experimental data, without any attempt to describe the underlying mechanisms. Unfortunately, they are generally of little utility for devices different from that used for the model identification, or even in working conditions different from those used for the model identification.

A typical example of a black box model, proposed since the very first years, in due to Kanno et al.96–98 In Ref. 96, the combination of a number of exponential functions was used to model the tip displacement of an IPMC bender under the effect of various input voltage steps. The used approximating function limits the application of the model only to the case of step inputs. The experimental results showed, moreover, that the system was not linear.

In Refs. 97 and 98, the experimental nonlinear electromechanical reaction of IPMC actuators was fitted by a linear model consisting of an electrical lumped model, that converts the applied voltage into the absorbed current, and a linear stress generation stage that accounts for the transduction of the absorbed current into IPMC mechanical reaction. The IPMC electrical model was obtained by dividing it into 10 equal sections, in the attempt to approximate its distributed nature. Parameters in the model were identified by using the IPMC step response and the model of the electrical stage was tested against different waveforms, up to about 20 Hz. The mechanical model was identified by fitting experimental data for a step input. A view of the proposed electrical stage is reported in Figure 9,97 while a corresponding mechanical structure was proposed for the bending motion.98

Figure 9.

The lumped electrical model proposed for IPMC actuators.97 © 1995 IEEE. Reprinted, with permission, from Kanno, R.; Tadokoro, S.; Hattori, M.; Oguro, K., Modeling of ICPF (Ionic Conductiong Polymer Film) Actuator – Modeling of Electric Characteristics, Proc. of the 1995 IEEE IECON 21st International Conference on Industrial Electronics, Control, and Instrumentation, 1995, 913–918.

The structure proposed by Kanno et al. was reconsidered in successive works by other authors who changed the basic assumptions in a number of ways. In Ref. 99 the fractal nature of the IPMC electrodes was investigated and experimental data were used to identify a model of the IPMC actuator. More specifically a distributed RC line was used to model the electrical behavior of the IPMC, while the IPMC bending was identified by experimental data.

The fractal nature of IPMC actuators was further investigated in the frequency domain in Ref. 100, where a novel model for IPMC actuators, based on non-integer order models was proposed and identified.

Recently in Ref. 101 a nonlinear black-box model for IPMC actuators based on a novel Preisach type fuzzy Nonlinear Auto Regressive Exogenous (NARX) has been proposed and identified by using a modified Particle Swarm Optimization (PSO) technique.

The sensing properties of IPMCs have been known for two decades: the first article proposing an application of IPMCs as sensors dates back to 1992 thanks to Sadeghipour et al.102 In that article authors have shown that IPMC can sense pressure, though no solvent was used and the sensing element was saturated with hydrogen.

Nevertheless models described so far fail to capture the reversibility of the electromechanical coupling capability of IPMCs. This topic was addressed, even though by using still a black box approach, in Ref. 103. In the article the authors proposed a linear coupled electromechanical model suitable for IPMCs in cantilever configuration, that can be used both to model IPMC based sensors and actuators. More specifically, the IPMC transducer has been modeled by using a two-port system as represented in Figure 10.

Figure 10.

The two port model introduced by Newbury and Leo to model IPMC based transducers.103

In the model v and i are the electrical variables, that is, voltage and current at the IPMC electrodes, while f and are the mechanical ones, that is, external force and velocity at the tip of the bender. Moreover the quantities introduced are linked in the frequency domain by the equation:

equation image(1)

The model proposed in Ref. 103 was identified mainly by using the IPMC as an actuator and performing a large number of static and dynamic measurements.

The described black box models are not in any case to be considered as an exhaustive description of all the models proposed in the literature. It has been preferred to focus on models that look interesting because of their innovative contribution or of their successive exploitation.

Gray Box Models

Gray box models are obtained on the basis of some well understood theories (e.g. the beam theory and the piezoelectric theory, though adapted to the specific case) and are ruled by using parameters that can be estimated with a set of experiments. More specifically, data of relevant quantities are acquired and are used by suitable optimization algorithms that minimize some kind of cost function in order to fit model estimation to the observed data.

The evolution of the model proposed by Newbury and Leo in Ref. 103 is an example of a gray box model, largely investigated also by other research groups.

The Z impedance matrix in eq 1 has not any link with physical phenomena nor with the geometrical parameters of the IPMC transducer. This shortcoming was successively investigated by the same research group in Refs.17 and104. In the referred two articles the same model structure was investigated with a look to the competing requirements of simplicity, accuracy and generality. Moreover, elements in the impedance matrix were proposed that depend only on the characteristic of the materials, so that the model can be scaled on the basis of the size of the device under test or under design.

The transducer was represented by using an equivalent electric circuit as reported in Figure 11.

Figure 11.

A view of a pinned IPMC transducer (a) and the corresponding equivalent electrical circuit (b).104

The IPMC transducer was modeled as a pinned beam by using the Eulero-Bernoulli theory which was applied to describe its deformation, while the electromechanical transduction phenomenon was modeled by using suitable coupling terms, in a similar way as for the case of piezoelectric devices. The analogy between piezoelectricity and IPMC transduction is limited to the adopted general scheme. In fact, since IPMCs are viscoelastic materials, the Young modulus is not anymore a real constant value. On the contrary, it is modeled as a complex function of the input signal. The same holds true for the electromechanical coupling term [represented by the ideal transformer in the schematic of Fig. 11(b)] that assumes the form of a complex function of the applied signal frequency.

In the circuit reported Rdc is the system DC resistance, and Zp is the electrical impedance, essentially modeling the capability of the system to storage charges. Zm1 and Zm2 are the mechanical impedance and the inertial term respectively. In such a way a linear dynamic model of the IPMC transducer, capable to describe the system dynamics up to the first mechanical mode was obtained. The proposed model was experimentally verified in Ref. 17 both for the case of the actuator (up to about 50 Hz for the case of the free deflection, and up to about 200 Hz for both the blocked force) and for the case of the sensor (up to about 200 Hz for the open circuit voltage).

By applying suitable boundary conditions to eq 1 the circuit reported in Figure 11(b) was solved deriving equations for IPMC transducers. Two relations relevant for the actuation were obtained, e. g, by imposing = 0 (no motion is allowed and the blocked force is developed) or f = 0 (no force is applied ad the free tip deflection is observed) as a consequence of the applied voltage. These expressions are:

equation image(2)
equation image(3)


When the sensing mode is considered, if v = 0 is supposed the short circuit sensing current produced by the tip motion with a velocity can be derived:

equation image(4)

The model described so far has been further modified in Refs.105 and106 for the case of the actuator where a nonlinear dynamic model was proposed and validated. More specifically, in Ref. 105 the lumped electrical model of the IPMC actuator reported in Figure 11(b) was modified to take into account of the finite resistance of the IPMC electrodes and of the nonlinear voltage to current relationship largely reported in the literature. An example of such a voltage to current curve is reported in Figure 12(a), while in Figure 12(b) is reported the equivalent electric circuit proposed in Ref. 105. The model was investigated up to about 100.0 Hz and the influence of environmental temperature and relative humidity was addressed.106

Figure 12.

An IPMC actuator voltage to current non linear behavior (a) and the electrical model proposed in Ref.105 for the voltage to current transduction (b). Reprinted from Bonomo. C.; Fortuna, L.; Giannone, P.; Graziani, S.; Strazzeri, S., A nonlinear model for Ionic Polymer Metal Composites as actuators, IOP Smart Materials ans Structures, 2007, 16, 1–12.

Model in eq 1 has been investigated and validated, for the case of the sensor up to about 50.0 Hz, in Refs.107 and108. In Ref. 107, the authors showed the relevance of the working conditions on the Young's modulus form and on the coupling coefficient and concluded that IPMC based sensors are much less sensitive to the presence of water as the solvent, with respect to IPMC based actuators. Finally, in Ref. 108 the influence of environmental temperature and relative humidity was investigated.

Since grey box models use recorded data to estimate a number of parameters, the quality of the estimation capability greatly relies on the identification approach. This is much true when nonlinear models are of interest, as it is the case for IPMC actuators. This aspect has been investigated in Ref. 109, where a nonlinear least square method is proposed to improve the quality of the nonlinear dynamic model proposed in Ref. 105.

The grey box approach has also been applied by some of the authors also to model IP2C sensors91, 93 and actuators,110 where the new transducer models have been identified to take into account of the substitution of metallic electrodes, with polymeric ones.

As described above model proposed in Refs.[94]94 and[98]98 was a rough approximation to describe the distributed nature of IPMC actuators. Starting from that model in Ref. 111 a lossy RC distributed line was used to model IPMC transducers. The proposed model is able to faithfully reproduce the non uniform curvature observed in IPMC actuators. The model is eventually capable to predict the sensing characteristics of IPMCs.

White Box Models

White box models try to describe the behavior of IPMC transducers by the comprehension of the physical and chemical phenomena involved in the transduction process. They have opposite pros and cons with respect to black box models, since they are very hard to be derived, depending on parameters that cannot be directly observed and/or might be based on complex equations that might fail to be useful in real life applications Nevertheless, since IPMC transduction is based on nanoscale phenomena, white box models play a fundamental role in the understanding of the microscopic phenomena that cause the IPMC transduction capability so that they could allow to predict the consequences of modifications in IPMC production steps. This aspect can give a formidable contribution to the optimization of IPMC macroscale performances.

In '90s Shainpoor proposed a microscopic scale model for large electrically induced deformations in a number of ionic polymeric gels, that is, for the case of cross-linked polymer networks swollen in a liquid medium, but in the absence of any electrode.112, 113

Starting from year 2000, a number of white box models have been proposed for IPMC actuation mechanism. Broadly speaking such models can be classified in two groups depending on the phenomenon that is hypothesized to dominate the charge redistribution inside the IPMC and then to produce the electromechanical coupling.

In Ref. 18, Tadokoro et al. supposed that the IPMC actuator bending is a consequence of solvent (water) migration, dragged by mobile ions, because of an applied external electric field. Moreover the redistribution of water under the effect of osmotic pressure causes a slower back relaxation of the IPMC. The electrostatic interaction of unbalanced charges was also considered, but it was argued that the electrostatic forces are less significant than the effect produced by ions/water redistribution.

In that same year a different explanation of the IPMC actuation was proposed by Nemat-Nasser and Li,114 and Nemat-Nasser.115 Based on their model the external electric field is again the cause of the ions redistribution, but, unlike in the Tadokoro model, the IPMC deformation was believed to be caused by stresses that act on the polymer backbone, because of the charge unbalance. In other words the IPMC deformation was considered as a consequence of electrostatic forces. Nemat-Nasser and Li, eventually, tried to explain also the back relaxation phenomenon that they observed for IPMCs actuated in air.

A further contribution toward the understanding of EAP transduction capabilities was given by de Gennes et al., again in 2000,116 who addressed the problem of ionic polymer direct and inverse electroactive effect, under static fields. Here again the transduction phenomena are believed to be caused by the mobile cations (Na+ in the described case) that drift under the effect of the electric field and carry a certain number of water molecules. These molecules pile near the cathode and create an overpressure that deforms the membrane. More specifically the authors considered the case of a ionomer swollen by water.

A schematic view of the system is reported in Figure 13.

Figure 13.

The deformation model of an IPMC under the effect of an applied electric field E.116

The transduction phenomenon was modeled by using two transport equations. The charge transport occurs with a current density J normal to the membrane and the solvent transport with a flux Q. Relevant forces are due to the electric field E and to pressure p. Quantities introduced so far are linked by equations derived from thermodynamics under the form of the Onsager relations:

equation image(5)

where σ is the membrane conductance; K is the Darcy permeability; L12 = L21 = L is a cross coefficient.

Equation 5 were further investigated by the authors to derive a direct effect for the case of the actuator, in the ideal case when the electrodes are impermeable to water, and an inverse effect, for the case of the sensor, under the effect of an applied torque Γ.

As mentioned above, research activity on IPMC sensing capabilities grew at quite a slower rate than corresponding activities on IPMC sensors, this being partly because the main interest was focused on the tantalizing huge deformation capabilities of ionic EAPs, including IPMCs.

Nevertheless activity to model mechanoelectric transduction in IPMCs evolved since the pioneering articles mentioned so far. This reflects the interest in the possibility to use one technology to realize the functions that are required in a smart system. In Ref. 117, Farinholt and Leo, using the model by Nemat-Nasser and Li,114 proposed a physics-based model based on the hypothesis that a deformation imposed to an ionic polymer produces a charge accumulation at the polymer surface. They proposed a model describing how the imposed deformation produces charge accumulation and then a sensing current. Based on these assumption and on the hypothesis that charge density is coupled with the mechanical stress through a proportionality constant,114 the authors derived equations for charge density, electric field, and electric potential when short-circuit boundary conditions are considered. Such boundary conditions were justified by the use of a current to voltage conditioning circuitry.

equation image(6)
equation image(7)
equation image(8)

By considering a cantilevered beam the corresponding current is estimated to be:

equation image(9)

The interested reader can find the full list of the used symbols in the referenced article.

Plots of the normalized predicted charge density, electric field, and electric potential are reported in Figure 14,117 referring to the instant immediately after the tip deformation has been applied.

Figure 14.

Plots of the normalized predicted charge density (a), electric field (b), and electric potential (c) adapted from Ref. 117. Reprinted from Farinholt, K.; Leo, D.J., Modeling of electromechanical charge sensing in ionic polymer transducers, Mechanics of Materials, 2004, 36, 421–433.

More recently Del Bufalo et al.,118 proposed a mixture theory to model IPMC actuation. The system was modeled as a mixture of three components: a charged solid (the backbone polymer matrix), an uncharged fluid (the solvent, e.g. water), and a charged gas (the mobile ionic species). The model is based on a set of three coupled linear partial differential equations that link the IPMC reaction to the applied voltage to osmotic pressure, hydraulic pressure and electrostatic forces.

In Ref. 119, the same mixture theory is applied to the physic-based modeling of both acting and sensing of IPMCs under the effect of a linear static action.

Finally, in a recent work Aureli and Porfiri120 proposed a nonlinear model for IPMC sensors, based on the Poisson-Nerst-Plank equations. Such a model was developed to adequately model IPMC sensors when the voltage across the IPMC electrodes is larger than the thermal voltage (about 25 mV at room temperature).

It has emerged so far that, starting from the production procedure, down to the modeling of transducers, IPMC based devices are the result of complex interactive activities, where different field of knowledge contribute to the final result. The same multidisciplinary approach is required if models capable to describe the chemo-electro-mechanical processes that result in the IPMC transduction are searched for.

Multiphysics models allow obtaining white box models by describing phenomena relevant to IPMCs by using partial differential equations. The equations are then solved in the form of finite differential equations, by using suitable software tools that have become available.

In Ref. 121 the IPMC is modeled as a thin 3D body resembling the characteristics of the hydrated Nafion®, sandwiched between two metallic layers. Authors modeled the chemically induced deformations in terms of a volumetric distortion field, induced by the redistribution of ions and solvent molecules, while the electric and chemical physics were described through standard equations of the electrostatics and of mass conservation. Thermodynamic issues were finally used to derive the stress T, the electric field e, and the chemical flux j. The commercially available finite element code COMSOL was used to perform numerical experiments, for the cases of both IPMC based actuators and sensors.

The COMSOL code has been recently used to model the actuation behavior of IPMC also in Ref. 122. More specifically the effects of finite resistance value of the platinum electrodes on the deformation and on the temperature distribution were investigated. As an evolution, the authors developed a multiphysic modeling of IP2Cs. In Figure 15 a comparison between the temperature distribution of IPMCs and IP2Cs, realized by using PEDOT:PSS, are shown, when they are activated by using the same forcing signal. The plots reported in Figure 15(a,b) show the effect caused by the different conductivity of metals with respect to organic conductors in the temperature distribution.

Figure 15.

Temperature along an IPMC actuator (a),122 and the corresponding IP2C temperature values (b).

Models for IPMC in Underwater Application

When a transducer is used in any application it interacts with the environment and/or with other systems. Such an interaction can produce dramatic changes in the transducer behavior and requires the development of new models. Of course in many cases such models will be truly application dependent and some cases will be reported in the Applications section. Nevertheless there are some working conditions that are quite general to justify the inclusion of the corresponding models in this section. This is the case of models that describe the interaction of IPMCs with a fluid, since IPMCs have been largely proposed for the realization of underwater prototypes such as bioinspired jellyfishes, fishes propelled by using IPMC based fins,123–125 or to realize systems to be used in fluids, including biomedical devices and micropumps, see for example Ref. 126.

In Ref. 127, Bonomo et al. used the hydrodynamic function theory128 to model the behavior of an IPMC vibrating in a fluid. In fact, while the interaction of an IPMC with air is negligible, its interaction with denser fluids can produce significant effects. More specifically, in Ref. 127 the frequency response of a cantilevered IPMC actuator for a number of fluids relevant to envisaged applications (e.g. sea water and blood were considered) was investigated. The proposed model is based on the solution of distributed partial differential equations and is therefore capable to describe the behavior of IPMC actuators in a wide frequency range with a good approximation of resonant frequency modes. The model was tested for the cases of an IPMC actuator vibrating in air and in pure water and for frequencies up to about 40.0 Hz. In particular, for this latter case the model correctly predicted both the first and the second resonance mode of the IPMC. In Figure 16 the frequency response for an IPMC actuator immersed in air, synthetic fluid, deionized water, sea water, and human blood, respectively as predicted by the proposed model are shown.

Figure 16.

Frequency response of an IPMC actuator immersed in air, synthetic fluid, deionized water, sea water, and human blood respectively.127 Reprinted from Brunetto, P.; Fortuna L.; Graziani, S.; Strazzeri, S.; A model of ionic polymer–metal composite actuators in underwater operations, IOP Smart Materials and Structures, 2008, 17, 025029.

The concept of hydrodynamic function was also exploited in the article of Aureli et al.129 They developed a modeling framework for free-locomotion of underwater vehicles propelled by vibrating IPMCs in quiet water.

Power Harvesting Models

As mentioned in the Introduction section one of the reasons of attractiveness of EAPs, including IPMCs, is their capability to be multifunctional materials with some of the functionalities required to realize smart systems. Among these, power harvesting capabilities have attired a very quickly growing interest, since smart devices could be required to work in the absence of traditional power sources. A typical example is wireless sensor networks that could be mounted in locations where powering systems could be either unavailable or inconvenient to be mounted. Though a power harvesting device is very close, in principle, to a sensor, the focus here is on the capability of the device to efficiently collect energy from the environment, without too much attention on the information degradation, as it is the case for a sensor. Aspects such as, for example, load adaptation are relevant in power harvesting and require the development of models that deals with such topics.

The possibility of IPMCs to be used as power harvester devices has been addressed in past years, with envisaged applications both in air7, 8 and in water.130, 131 More specifically, in Ref. 8 the model proposed by Newbury and Leo,104 and adapted by Bonomo et al.107 was considered and identified. The power generation capability of IPMC transducers under the effect of base excitation vibrations in air was investigated. A power generation capability of few nanowatts was estimated when a vibrating frequency of 7.09 rad s−1 was imposed to the system.

In Ref. 130 the authors investigated the power harvesting capabilities of an IPMC submerged in a fluid, because of mechanical base excitation. The mechanical vibration of the IPMC was modeled trough the Kirchhoff-Love plate theory, while its interaction with the encompassing fluid was described by using the linearized solution of the Navier-Stokes equations, and the concept of hydrodynamic function. Also in this case developed power of the order of nanowatts was predicted.

While activity referenced so far on power harvesting by using IPMCs was based on base excitation, in Ref. 131 the possibility to harvest power from the flutter instability of an heavy flag hosting and IPMC transducer was, finally, investigated. In fact, the low bending stiffness of IPMCs allows for the use of compliant structures that can experience flutter motion at moderately low flow speed.

Though results on power harvesting capabilities of IPMCs reported so far show that the available power is very small and does not allow the powering of any real device, this field of research is quite promising because IPMC can efficiently collect energy at very low frequencies. In fact, a lot of environmental mechanical power is available at low frequencies.

It can be also argued that significant improvements of the level of collected power can be obtained. One suitable solution could be the connection of a large number of energy scavenging units that can be used thanks to the low cost of polymeric devices. Such a solution has also the beneficial effect to enlarge the usable bandwidth of the scavenging system, if the units are purposefully chosen with unmatched mechanical characteristics.

In the following Table 4 a schematic view of the described models, including their characteristic, main advantages and limitations is reported.

Table 4. Classification of the Modeling Approaches of IPMCs
inline image

Though an evolution of model performance can be observed from the analysis of reported articles, most of the authors remark that further efforts are needed to better understand, and therefore model, some aspects of IPMC behavior such as the nonlinearities, the dependence of IPMC behavior on environmental working conditions, and the time-variant behavior, mainly due to solvent loss. Nevertheless some undesired behaviors do not depend on the quality of the models. For example, as mentioned above, IPMCs show a dispersion of the electromechanical transduction capabilities linked to the poor control of the production phase that causes variable electromechanical properties, even when produced in the same batch.

Controlling IPMCs

IPMC based systems are challenging to control for the reasons mentioned above. Nevertheless, high precision control is often needed in many applications envisaged for IPMCs, such as, for example, those referring to biomedical devices or micromanipulators. As a consequence, a number of different control approaches can be found in literature, based both on black box models, grey box models and physical models of IPMC actuators.

Even if some classical approaches have been proposed, like PID or LQG controller, intelligent, adaptive or robust controller design methodologies are often applied to cope with model uncertainties, nonlinearity, time-variability and non-repeatability of the IPMC behavior.

One of the first articles in the area of IPMC control is Ref. 132 where a classical LQR controller was proposed to regulate the deformation of an IPMC actuator in order to improve the dynamic behavior. A PID controller with impedance control was proposed in Ref. 133, while in Ref. 134 an integrator anti-windup scheme was added to reduce the performance degradation due to actuator saturation.

A more advanced approach was proposed in Ref. 135 where a model reference adaptive control (MRAC) structure for tracking control, along with a pole-placement approach and a genetic optimization strategy, to tune the parameters, was introduced in order to deal with environmental variations, like humidity or ionic fluid evaporation. Another MRAC controller in which the parameters are modified on-line without seeking the bounds of uncertainties, is proposed in Ref. 136 in order to guarantee repeatable performance even with a time-varying behavior of the system. Some experimental results, over a short time period (4 min) are reported in the article.

A robust controller is designed in Ref. 137 to overcome uncertainties and non-repeatability, by using the H approach and μ-synthesis. It was based on a linear black box model and the controller performance was evaluated using only step response.

In Ref. 138 a model based frequency-weighted feed-forward controller was designed to enable fast positioning while avoiding large voltages. A feedback controller was also added to take into account unmodeled effects.

A different solution for IPMC high precision control was proposed in Ref. 139. It is based on the Quantitative Feedback Theory, in order to cope with the large parameter uncertainties, satisfying both robust tracking performance and the noise attenuation requirement. A nominal model with the corresponding uncertain bounds was determined as a preliminary step.

A modified direct self-tuning regulator consisting in a pole-placement controller with integral action, a reference model and a self-tuning strategy was designed in Ref. 140 to minimize the tracking error. The controller was designed both for IPMC actuators working in air and underwater on the basis of a black box linear model.

The elimination of the low-frequency distorsion caused by back-relaxation was dealt with in Ref. 141. More specifically, an adaptive neuro-fuzzy controller to control the deformation of IPMC for underwater manipulation was proposed, in which the membership function designed by an expert are adapted by a neural algorithm.

In Ref. 142, it was proposed a model-free approach for controlling the position of an IPMC actuated rotary linkage for micro-manipulation. The method proposed is a nonlinear tuning method which adapts the parameters of a PI controller in order to improve the steady state response. It uses a gradient descent algorithm to minimize a cost function of a controller design criterion directly from an unknown system.

More recently, in Ref. 143, a time variant black box model, identified with a real-time approach, was used to design an adaptive feed-forward controller. A noise cancellation technique was also used to alleviate the effects of plant disturbances. In Ref. 144, a controller able to work over a long period of time was proposed. In details, an iterative feedback tuning algorithm was introduced to adaptively tune the controller for a IPMC micropump for microfluidics applications during normal system operations. Also this approach is model-free and it is based on a PID controller.

Finally, a short review of control strategies proposed in literature for IPMC actuators is included in Ref. 145.


Thanks to their peculiar characteristics, IPMC devices have greatly stimulated several researchers to imagine and realize prototypes in a large variety of application fields. The number and the quality of the devices realized clearly have increased in time according to the improvements in the knowledge of the IPMC behavior and in the technology for their fabrication. The state of the art in the knowledge of their behavior and the examples of applications analyzed so far are well described in Ref. 22 which represents a good starting point to understand the evolution and the growing interest for IPMC devices.

The double capability of an IPMC membrane to work as an actuator and as a sensor has had the effect to produce two main directions in the research development. So, it is possible to find applications in which an IPMC plays the role of a sensing device and other applications where the IPMC are used as actuators. This is not a rigid partition and, even more interesting, it is possible to find applications where sensing and acting IPMCs are combined into a smart device, or even to find IPMCs, fabricated following particular dedicated procedure, in a combined role as sensor-actuator devices. In this view, each of the three cases is addressed in the following part.

IPMC as Actuator Devices

Several applications where IPMCs represent the actuation devices have been proposed. Their time evolution is characterized by the presence of sophisticated structures and the growing importance of the control systems for accurate and effective actions. Here is a description of the most promising and interesting results in the last decade.


An example of linear actuator was proposed in Ref. 146: structures composed by IPMC strips connected according to particular configurations were analyzed, mainly numerically. In more recent developments, a 2DOF actuator, obtained shaping a cylindrical IPMC, were presented in Ref. 147. Fabrication technique and experimental results showed the effectiveness of the actuator, able to bend up to 50°, in all the four directions. The same robotic structure, a biaxial bending actuator, but with a different manufacture procedure, was addressed in Ref. 148. This last realization technique allowed to get high actuating force and high flexibility. A model of its behavior and a control scheme were also developed to allow an effective implementation of the actuator.

In many applications, especially for mini and micro realizations, the transmission of power to the actuators, also in case of use of IPMC, plays a very important role that may limit their usage. A first tentative of solution to this problem was given in Ref. 149, where the results of some preliminary studies concerning the possibility of a power transmission by means of microwave transmitter and a specific realization of integrated receiver and IPMC actuator were reported. The same problem was addressed in Ref. 150 with the aim to replicate hovering flight and swimming in biological systems. A conceptual scheme of a robotic microswimmer to be powered is shown in Figure 17.

Figure 17.

A schematic of a of the robotic microswimmer.150 © 2012 IEEE. Reprinted, with permission, from Abdelnour, K.; Stinchombe, A.; Porfiri, M.; Childress, S., Wireless Powering of Ionic Polymer Metal Compostes Toward Hovering Microswimmers, IEEE/ASME Trans on Mech., 2012, 17, 924–935.

The electrical (low power, low voltage, etc.) and mechanical (weight, compliance, etc.) characteristics of IPMC actuators have stimulated their use in the realization of manipulators and grasping devices, especially for micro manipulations. In Ref. 151 a parallel architecture, with a redundant number of IPMC actuators, was presented and studied. The mechanical characteristics were discussed and its behavior under a visual feedback control action is illustrated. The mechanical structure of a 2DOF micro manipulator, together with an adaptive control system, was designed and presented in Refs.[152]152 and[153]153, actuated by two IPMC membranes. The overall characteristics evidenced by experimental results motivated the proposed usage as a manipulator for sensitive biological materials.

A planar multi DOF manipulator realization was described in Ref. 154. It was obtained from one IPMC realization by means of laser cuts which allow electrical insulation between segments. The kinematic model was derived and used for a closed loop control implementation.

Manipulators where the IPMCs are introduced as joint actuators have also been considered. For example, in Refs.[155]155 and[156]156 the guidelines for the design of a 2DOF planar manipulator with plastic links actuated by IPMC strips at the joints were described. A particular attention was given to the kinematic model computation and the control law design and implementation in order to verify its robustness with respect to all the IPMC parameters variation. An equivalent kinematic structure with a different approach to the IPMC realization problem for an enhancement of the actuation performances was described in Ref. 42 with application to the motion of artificial fingers.

The description of an innovative three-dimensional four-electrode IPMC actuator was reported in Ref. 157 along with the design, fabrication and test phases. The structure can carry an optical fiber so that a very thin laser light can be arbitrarily oriented. Its usefulness in micro endoscopic surgery was highlighted. Gripping capability have also been explored. In Ref. 158, an in depth analysis of the state of the art about this kind of application has been carried out and a series of experimental tests compared to numerical results for a prototype of two finger gripper were described proving the feasibility of such a use for micro manipulations. In fact, the bending capability of IPMC actuators can facilitate the construction of gripping devices especially for low force micro manipulation. In this view, a two finger micro gripper realization and the experimental results obtained using it in a closed loop control system were described in Ref. 159.

Taking inspiration from the behavior of biological models of locomotion, several interesting proposals have been presented in the last years with a large variety of robotic mobility architectures. Leaving a separate section to the applications in fluid environments, ground robots are here addressed. Starting from the legged locomotion, the realization and the behavior in some experimental tests of a biped robot actuated by linear IPMC actuators was described in Ref. 160. A suitable parallel and serial combinations of elementary IPMC constituted each linear actuator, to increase the force generated and the total displacement.

Increasing the number of legs, a four legged micro robot mimicking a turtle motion was presented in Ref. 161. It was fully realized by IPMC material with a suitable shape that allows each leg to be actuated independently, so getting the walking motion.

A prototype of worm-like, mimicking the undulatory locomotion of a worm, was presented in Ref. 162, where a fully IPMC based structure, composed by a chain of IPMC actuators, is presented. The actuation and then the motion was obtained by propagating a periodic signal generated by a Cellular Neural Network. In Figure 18(a,b) a schematic of the locomotion mechanism and a view of the realized prototype are shown, respectively.

Figure 18.

A schematic of the locomotion mechanism (a) and a view of the realized prototype (b) of a bioinspired worm.162 © 2006 IEEE. Reprinted, with permission, from Arena, P.; Bonomo, C.; Fortuna, L.; Frasca, M.; Graziani, S.; Design and control of an IPMC wormlike robot, IEEE Transactions on Systems, Man, and Cyb. – Part B: Cybernetics, 2006, 36, 1044–1052.

A different mechanical structure for a worm-like robot was proposed in Ref. 163, where IPMC strips actuate each of the segment that compose the worm. A sinusoidal propagating actuation produces the ondulatory behavior of the structure and, consequently, the motion. A similar but smaller worm-like robot was described in Ref. 164. The micro-IPMC structure consists of one IPMC support divided into five segments separately actuated. The motion is obtained propagating a sinusoidal wave along the robot. A feedback control system, based on image analysis, allows to drive the robot according to desired goals.

The capability of IPMC actuators to work in a fluid environment has convinced a great number of researchers to develop applications devoted to verify the effectiveness of such a choice. Studies on the physics of a IPMC based device in a fluid, with the computation of suitable mathematical models and with the measurement of the effects under different control actions, have been carried out. A particular attention has been devoted to fish-like mobile structures. Although it is possible to find some preliminary results starting from the early 1990s,165 only in the last years a large number of researchers have focused their work on this subject. Some consideration on the possibility to use IPMC actuators for robotic motion in fluid environment were presented in Ref. 45, with a preliminary study on the trust capabilities using an actuated tail. Then, in view of the way in which an IPMC actuator works for motion in a fluid, in Ref. 166 the results of an interesting study of the flow field generated by a planar strip actuated to vibrate were reported, computing the relationships between the vibration characteristics and the thrust effects produced and showing that the amplitude of the oscillation does not play a such important role.

The computation of an effective mathematical model describing the motion generated by an IPMC actuator, taking account of both form drag and viscous drag forces, was performed in Ref. 167 and validated experimentally while for a state space mathematical model of a segmented IPMC actuator, taking into account electrical, mechanical and hydrodynamics effects, one can refer to Ref. 168 for its computation and validation by means of an experimental larval zebra fish robot.

More recently, a study169 was proposed analyzing the ability of a steady turning motion of a robotic fish, with particular reference to a structure consisting of a rigid body part and a flexible caudal fin. A general mathematical framework was developed and then applied, as a case study, to an IPMC actuator-propelled robotic fish. A more experimental approach was adopted in Ref. 170, where a numerical study of the hydrodynamics of a vibrating IPMC actuator in an aqueous environment was carried out. Several characteristics of a cantilever IPMC actuator were derived, highlighting relationships between the dynamic behavior of the actuator and the hydrodynamic effects in the fluid. All the results were compared with experimental tests.

Prototype realizations of underwater systems actuated by artificial muscles based on IPMC materials can be found in Ref. 171, where an IPMC membrane actuated by a periodic input voltage constitutes the fish body while the swimming speed is regulated by changing the input frequency or in Ref. 172, with an open loop motion planning and a particular open loop oscillators based control.

In the same years, in Ref. 173 first and then in Ref. 174 a small robotic fish was proposed, propelled by means of a PVC caudal fin actuated by an IPMC membrane. Facing with this pretty sophisticated structure, where the IPMC beam is attached to a passive plastic fin and their interaction has to be considered, in Ref. 175 a model that included the interactions between the active part of the propeller, the IPMC, and the passive one, the plastic foil, was developed. The same problem was faced in Ref. 176, where a mathematical model, based on Lighthills theory of elongated-body propulsion, was developed, including both the dynamics and the hydrodynamics and defining the relationships between the periodic actuation voltage and the steady-state cruising speed, also for different caudal fin dimensions.

A different approach in the modeling of the hydrodynamic effects with added mass and damping was followed in Ref. 129, where also a modal analysis of the vibration effects of the actuator behavior was carried on. The results were verified by means of a remotely controlled prototype.

For a better understanding of all the effects in swimming behavior of IPMC actuators for underwater applications, an important role is played by contributions like177 where, through an experimental approach, the results obtained for a fish-like robot motion using different actuation periodic signals, from sine waves to square waves, and under different wave parameters (amplitude, frequency, and so on) are investigated.

A different approach was followed by researchers that preferred to focus their attention on the control system, in order to obtain better performance of underwater swimming robots actuated by IPMC membranes. A strong motivation to this idea is that such devices, mainly thanks to the actuator choice, are suitable for miniaturization, so finding a large variety of application fields with different kinds of fluid, water as well as blood, and in this cases both the control accuracy and the optimization for power consumption are considerably relevant questions. Among them, in Ref. 178 a feedback control system was proposed for maximizing the trust; an optimization approach was followed starting from the computed model and the results were verified on an experimental setup. The improvement in the control action produced by an adaptive control added to a model based classical control has been shown in Ref. 136: the simple model available and the high variability in the parameters values were effectively compensated by the adaptive contribution. In Ref. 179 an energy consumption model has been analytically computed on the basis of the IPMC actuators behavior and the hydrodynamic effects. The results were used to design an energy-saving control. In Ref. 180, starting from the observation that motion of IPMC as an actuator in a fluid environment is characterized by an oscillatory periodic behavior, a standard discrete-time repetitive controller was designed. The IPMC membrane was integrated by a strain gage sensor for the feedback measurement and the result is a feedback control for precision tracking of periodic reference trajectories.

Of course, not only fish shaped robots have been considered. A strong effort has been devoted to develop structures that can combine the reduced dimensions and low power consumption of IPMC actuators with a higher mobility. Among the first examples of legged microrobot it can be found59 where a ciliary type eight legged robot, suitable for moving in aqueous environments, was realized and analyzed. In Ref. 181 an eight legged microrobot designed to walk submerged in a fluid (water) environment and to float upward and downward was proposed. A preliminary evaluation of the capability of a leg was performed by experimental measurements and a prototype was presented and tested. The concept of legged robot was more recently used in Ref. 182 where a robot with IPMC based legs was proposed for in-pipe applications.

A tadpole shaped robot was studied in Ref. 183. The small structure was realized and tested, verifying its fully controllability acting on the periodic input voltage parameters.

Worm-like structures, fully actuated so getting an undulating motion in a fluid, were proposed in Ref. 184, with a particular attention to the characteristics of the IPMC actuation, and in Ref. 185, highlighting the importance and the effectiveness of a miniaturization procedure for several application fields like human blood flow. The control problems arising from the miniaturization constraints have been approached too. Jellyfish structure has attracted researches attention. In Ref. 186 a microrobot with such a structure was considered, using four vibrating legs as actuation appendages and focusing on its motion capabilities along the vertical direction, floating and sinking. A jellyfish gave the inspiration also for a slightly different robotic equivalent,124 always based on IPMC actuators but in a different configuration and controlling is motion through a sequence of pulse and recovery phases.

Real views of a jellyfish in swimming and of the prototype realized are shown in Figure 19(a,b), respectively.

Figure 19.

Jellyfish in swimming (a) and components of the jellyfish robot realized (b).124 Reprinted from Yeom, S.-W.; Oh, I.-K., A biomimetic jellyfish robot based on ionic polymer metal composite actuators, IOP Smart Materials and Structures, 2009, 18, 085002.

Amoeba-like structures have been tentatively proposed187, 188 too. A robotic structure moving as a ray fish was described in Ref. 189 where all the solutions adopted for minimization of the electronics are discussed and a preliminary simple control strategy was adopted to validate the motion capabilities.

In the research of smarter structures for underwater motion both technology advances and previous experiences bring to even more interesting solution. Then, very recently, in Ref. 190 a versatile micro structure actuated by 10 IPMC membranes as legs or fingers, able to walk, rotate, float and grasp, was described.

Robotic fishes can also be used for water systems monitoring. Two different remotely operated fish-like robots, using an IPMC for the actuation, were presented in Ref. 191 for measure and data collection of water temperature. Applications in sea monitoring can also be found: in Ref. 192 the supervision of seawater pollution produced by petroleum exploration platform was proposed to be performed by robotic fishes actuated by IPMC, thanks to their reduced power consumption. An optimal control was designed and proposed.

In addition to the classical application of IPMC materials as actuators, some test and experimentations have been carried on to verify the feasibility of general actuators based on IPMC elements. A specific fabrication method of a large variety of shapes for micro IPMC devices was accurately described in Ref. 193, with a particular application to the realization of micro grippers to be applied in laparoscopic surgery.

A helical shape of an IPMC actuator was addressed in Ref. 194. Experimental analysis of its behavior was described, showing its effectiveness for applications in the radius control of active stents. The possibility of controlling an IPMC using as a reference input an electromyographic signal was firstly investigated in Ref. 195 and, more recently, rediscovered in Ref. 196.

Photos of the EMG controlled hand in null and in flexion state are reported in Figure 20(a,b), respectively.195

Figure 20.

An IPMC based hand controlled by EMG signal in none state (a) and flexion state (b).195 © [2006] IEEE. Reprinted, with permission, from Lee, M.-J.; Jung S.-H.; Lee, S.; Mun M.-S.; Moon I. Control of IPMC-based artificial muscle for myoelectric hand prosthesis, Proc. First IEEE/RAS-EMBS International Conference on Biomedical Robotics and Biomechatronics, BioRob 2006, 2006, 1172–1177.

Electrodes for the acquisition were positioned at forearm and the measured signal, suitably filtered and amplified, was used to drive an IPMC beam with interesting results. The same control architecture was presented in Ref. 197 but on a more complex mechanical structure, with a measurement-actuation chain starting from the measure of EMG signals and ending with the actuation of a prosthetic arm.

The proposal of a micropump actuated by IPMC drivers was introduced in Ref. 198 with an in deep numerical analysis, through a finite element approach, of its characteristics. A different mechanical structure for the realization of a micropump based on IPMC actuation was discussed in Ref. 199, with some experimental results reported. A schematic of the production process and of the realized prototype are shown in Figure 21.

Figure 21.

A multilayered IPMC diaphragm: (a) the fabrication process of the IPMC diaphragm with photolithography and PDMS molding techniques and (b) a photograph of the fabricated IPMC diaphragm.199 Reprinted from Nguyen, T.T.: Goob,N.S.; Nguyen, V.K.; Yoo, Y.; Park, S., Design, fabrication, and experimental characterization of a flap valve IPMC micropump with a flexibly supported diaphragm, Sensors and Actuators A, 141, 2008, 640–648, with permission from Elsevier.

A bistable actuator with the advantage to require no energy for maintaining the two positions was illustrated in Ref. 200. The bending characteristic of an IPMC strip was used to let it switch between the two opposite directions. Improvement with actuator segmentation and the analysis of buckling mode can be found in Ref. 201.

The use of IPMC actuators to control an inverted pendulum was investigated in Ref. 202. A cart with the inverted pendulum over is moved by means of the deflection of an IPMC actuator. The goal was to validate the accuracy of the positioning of IPMC actuators in a closed loop control system. An improvement of such a result was presented in Ref. 203 where, for a similar mechanical structure, an LQR controller was implemented, with the additional constraint to minimize the input voltage of the membrane, and the whole system was implemented and tested.

Within the large variety of application fields, the actuation of optical devices has also been investigated. So, small motion and low energy actuation properties for IPMC devices suggested to use them for changing the focal length of lens moving a silicon plate that produces a variation of the shape of liquid lens,204 while a deformable mirror was addressed in Ref. 205, where an IPMC based actuator with a gear shape was used to produce mirror curvature.

As a final citation, it may be interesting to recall that IPMC usage has been proposed also for its physical and chemical characteristics. In Ref. 206, an IPMC was used to increase the efficiency of a electrolysis generator to produce gas for a buoyancy control of underwater vehicles.

Preliminary results on the possibility to use IPMC in more sophisticated configurations for advanced applications are available. For example, in Ref. 207, IPMC actuators arrayed in horizontal as well as vertical directions were proposed to improve the actuation performance. In this configuration, they were applied to the multifunctional control surface of a micro air vehicle (MAV).

IPMC as Sensor Device

The use of IPMC materials as sensor elements is based on the relationship between the deflection and the charge accumulation that produces a voltage difference on the electrode surfaced by the ion motion. Clearly, the effectiveness and the accuracy of these kind of sensor systems is strongly related to the knowledge of such relationship. Then the computation of a model as complete as possible is the first necessity.

Unfortunately, several parameters characterize the model of an IPMC sensor. The hydration level and the temperature are the most effective in changing the characteristic of the deflection-voltage relationship. In Ref. 208 an extensive analysis on the most recent results in literature about the computation of models able to capture as many dynamical aspects as possible was performed. In addition, a solution of temperature effects compensation was proposed, so enhancing the effectiveness of such sensor devices.

The field of applications of IPMC based sensor devices is the same as for the actuation, where the mechanical, chemical and electrical characteristics of IPMCs play a significant role in improving the performances. Then, it is possible to find biomedical applications in which the flexibility and biomimetic aspects help to get relevant and promising results.

The use of IPMCs for measuring bending changes of human tissues was proposed in Ref. 209 where an accurate model for measuring bending moment expressions for both dynamic and static electric potentials was computed from which the beam deflection curve and the pressure distribution generated on human tissue can be obtained.

Biological systems often give inspiration for construction of equivalent systems. IPMC devices are among the easiest ones to work in water or more general in fluids. This fact guided several researchers to develop bio-inspired systems with an IPMC as the sensing element. This is the case of Ref. 210 where the artificial version of a lateral line (an array of hair cell sensors), used by underwater organisms to get information on flow conditions, obstacles, and moving objects, was proposed replacing the hair cells with IPMC strips and using an artificial neural network for processing the sensor signals. The developed sensor array and the experimental set-up are shown in Figure 22.

Figure 22.

Experimental setup: the IPMC-based lateral line prototype(a); the tank for the experimental investigation of the system (b).210 © [2011] IEEE. Reprinted, with permission, from Abdulsadda, A.T.; Tan, X.; Underwater source localization using an IPMC-based artificial lateral line, Proc. 2011 IEEE International Conference on Robotics and Automation (ICRA 2011), 2011, 2719–2724.

The ability of IPMC strips to detect deformations by means of the voltage on their electrodes can be used in a quasi static way, to measure a deflection, like in Ref. 211, where a dynamic curvature sensor for curvature monitoring of deployable/inflatable dynamic space structures was proposed or in a dynamic way, to measure changes in deflection like those one has in vibrating systems. This idea was discussed in Ref. 212 where a vibration sensor has been proposed. Its model is derived and some experimental tests validated the effectiveness of the sensor.

Active cantilever beams are widely used in resonant conditions to measure a number of parameters that can influence the value of the resonant frequency and/or the value amplitude of the oscillations in resonance conditions. Piezoelectric beams are generally proposed to this aim. In Ref. 213 an IPMC based actuator was forced to work in resonant conditions and was actually characterized as a sensing system to measure tensile loads. The proposed arrangement extends the measuring frequency range to DC loads, that cannot be measured by IPMC sensors. In Figure 23 views of the developed set up and of the transduction diagram are shown.

Figure 23.

Experimental setup: The IPMC-based force resonant sensor (a); the corresponding transduction diagram.213

IPMC sensors can find interesting fields of applications also as part of a more complex sensor system. This is the case of Ref. 214 where, on the basis of the strong simplification in the mechanical realization and in the electronic interface, the use of a very small IPMC beam was proposed for in-situ measurement of gas flow properties in automotive applications.

A further example of a bioinspired system has been described in Ref. 215, where a set of IPMC sensors were used to mimic ciliate cells of the vestibular labyrinth a biological component of living beings, situated in the inner ear and devoted to perception of the angular acceleration the head is subjected to. Ferrofluids216 were used in that application to stimulate the IPMC sensors that produce a coded signal, used to detect the system motion.

In Ref. 217 a different approach to the use of IPMC/IP2C as sensor was faced. In this case, it was investigated the possibility to detect a pressure on the surface of the membrane and the relative localization of the force. Good results were obtained from a qualitative point of view for every working condition. Moreover, it was shown that also for this configuration the high sensitivity of the material to the change of parameters involved force to use it in a controlled environment.

Combined Use of Sensing and Actuation Capabilities

In several applications and prototypes the measurement and actuation actions require characteristics that are both fulfilled by IPMC devices. So, the cases in which IPMCs are used in a combined action as sensors and as actuators are not rare, especially in micro or in bioinspired realizations.

In this section such systems are described as a relevant aspect of the possibility to use one technology to realize both sensing and acting functionalities, and hence as a valuable contribution toward the realization of postsilicon smart systems.

The more intuitive combined use of IPMC as sensor and actuator consists of using two IPMC membranes, one as actuator and the second as sensor. Such a solution was proposed and investigated in Ref. 218, where a thin plastic sheet supporting the two strips is considered in the analysis. The same idea of the use of an IPMC couple sensor-actuator was also faced in Ref. 219 with particular attention to the control scheme. This same configuration was used in Ref. 220 to realize a viscometer. The actuator was used to generate a vibrating motion in the fluid and the sensor was devoted to measure the parameters of such a vibration, which are dependent on the viscosity of the material.

The very small dimensions, the very low contact force exerted make the tactile sensor proposed in Ref. 221 particularly suitable for biomedical applications. The combined action of a couple of IPMC actuator and sensor was used to design a tactile sensor in which the actuator deflection produces the pressure and then the deformation of the touched surface and the sensor measure such deformation.

The same actuator/sensor combination was proposed in Ref. 222, though the system was arranged in a different way and it was forced to work in resonant conditions. The changes in the resonance frequencies measured in correspondence of the interaction with different materials makes it possible to distinguish the material itself from its mechanical characteristics. In Figure 24 a view of the system schematic (a), of the corresponding prototype (b), and of the system behavior for different materials (c) are shown.

Figure 24.

A view of the system schematic (a), of the corresponding prototype (b), and of the system behavior for different materials (c)222. © [2010] IEEE. Reprinted, with permission, from Bonomo, C.; Brunetto, P.; Fortuna, L.; Giannone, P.; Graziani, S., A tactile sensor for biomedical applications based on IPMCs, IEEE Sensors Journal, 2008, 8, 1486–1493.

Finally a study of the interferences between the IPMC based actuator and sensor, in systems with suitably patterned electrodes, was considered in Ref. 223.

An improvement of such a contemporary use of two IPMC devices, each of them devoted to one of the two functionalities, is the realization, on the same IPMC film, of one device that exhibits both the characteristics. In such cases the sensing element is realized by using the metallic electrodes, that deform as a consequence of the actuator motion. It is worth to consider that in combined applications described so far the IPMC sensor is a generating one, while now the sensor is a resistive element. This requires different conditioning circuitry. Studies on the dependence of the electrode resistance on the IPMC deformation can be found in Refs. 224 and 225. In Ref. 226 a different pattern was proposed end experimentally investigated by using a Wheatstone bridge, as the conditioning circuitry. The same system is further modeled in Ref. 227.

A successful attempt to obtain the contemporary effects of actuation and deformation measurement for an IPMC strip was carried out in Ref. 228, where it is proven that the deflection of the actuator strip can be obtained measuring the surface voltage in points different from the input feeding ones.

A different way to evaluate the deformation in a three-dimensional micro device omnidirectional accessible elastic tweezers for microendoscopic surgery to get a force measurement at the tip was proposed in Ref. 229. It was established the quantitative relationship between the reaction force from the grip and the feedback current on the actuators. An interesting result in this context is represented by,230 where a microrobot mimicking a Venus flytrap, a carnivorous plant, was presented and where IPMC elements are used both as nanosensors, to detect the presence of an insect, and nanoactuators, to close the lobes so trapping the insect.

In Ref. 231 the motion of IPMC actuators was sensed again by using resistive elements. Those elements anyway are not realized by patterning the IPMC electrodes but strain gages are used. The system is intended to monitor underwater working IPMC actuators. In Figure 25 a view of the experimental set up is shown (adapted from Ref. 231).

Figure 25.

A view of the experimental set up including the IPMC actuator and the mounted strain gages rosette.231 © [2010] IEEE. Reprinted, with permission, from Integrated Sensing for IPMC Actuators Using Strain Gages for Underwater Applications, IEEE/ASME Tr. on Mech., 2012, 17, 345–355.

Finally, in Ref. 232 an IPMC actuator was equipped with PVDF sensing element to realize a smart acting/sensing system.

Also in this section tables are reported to have a better look to IPMC applications. More specifically in Table 5, IPMC based actuators are considered, while in Table 6 the cases of IPMC based sensors and IPMC sensor/actuator combined use are taken into account.

Table 5. Classification Applications of IPMC as Actuators
inline image
Table 6. Classification Applications of IPMCs as Sensors and IPMC Sensor/Actuator Combined Use
inline image

The huge amount of described applications is a symptom of the vivid interest on possible IPMC based applications. Also, it is possible to observe a large diversity in the type of the proposed applications.

Among application fields investigated for IPMCs a continuous interest has been devoted to the cases of robotic applications and in particular the IPMC peculiarities have been exploited to realize motion system based on tails (for underwater propulsion) and legs. The main aspects of the proposed applications in these fields are reported in Tables 7 and 8, respectively.

Table 7. Comparison of Underwater Tail Based Motion
inline image
Table 8. Comparison of Leg Based Mobile Robots
inline image

The most part of described applications are lab-scale prototypes and have been developed as a proof of concepts more than as systems ready to the market. Of course, IPMC based applications are the last step toward the full exploitation of this novel technology and suffers for drawbacks that still affect IPMC production and modeling phases.

Real life applications need of more repeatable materials and this can be obtained only if a tighter control of chemo/physical properties of obtained IPMCs will be available. Efforts are needed to better define the relationships between production steps and their consequences on produced transducers.

Though sophisticated control strategies mentioned in the previous section can greatly alleviate the effects of unmodeled uncertainty and time variance, actions are needed to improve materials stability mainly due to solvent loss.

In experiments reported so far, no relevant fatigue effects have been observed for IPMC based sensors and actuators. Nevertheless, organized experimental campaigns are required before the useful file can be defined of IPMC transducers.

Finally tough IPMC biocompatibility has been claimed for IPMCs, to the best knowledge of the Authors, no experimental studies have been conducted and are required ahead IPMC biomedical devices can be safely used.


Smart systems, capable to solve even more complex problems with little or no human intervention, are the subject of an ever growing interest from researchers and industry since they could allow dramatic increments in the performance of applications in fields such as bio inspired robotics, aerospace and medicine, just to mention a few. New smart systems will be required to embed a number of different functions including, electric power generation and storage, signal sensing, and processing and actuating capabilities.

Polymeric materials will play a main role in the development of smart systems since some of them have already been shown to have the capability to scavenge energy from surrounding environment, elaborate electrical signal, produce action, and sense signals.

Though such capabilities can be obtained by using more traditional inorganic materials, they suffer for a number of shortcomings that do not plague polymers: for example, power scavenging can be obtained, just to give an example, by using piezoelectric ceramics but they are very rigid and fragile. In the same way signal processing can be obtained by using very fast silicon based electronics that unfortunately need to be mounted on rigid cards, while a significant interest exists for flexible electronics.

In this article a class of electroactive polymers (EAPs), Ionic Polymer Metal Composites (IPMCs) that are among the most interesting EAPs and have shown the capability to harvest mechanical energy, react to electrical inputs, and produce electrical reaction when subjected to mechanical actions have been described.

More specifically in article the state of the art on IPMCs is reviewed and the three main aspects of this new technology that is, fabrication methods, modeling, and applications are presented. From reported results it emerges that, though IPMCs have been the object of interest for no more than two decades, the scientific literature is growing at an impressive rate and significant improvements have been obtained in all the fields of this new technology, from efforts in the production to the demonstration of IPMC capability to solve even more complex problems. Also, the impressive rate of growth of proposed IPMC based systems, though they are still at the stage of proof of concept lab prototypes, show that IPMCs are one of the promising technologies for the realization of smart systems.

Since a great activity exists in the field of organic electronics devoted to the realization of novel signal elaboration organic systems and promising results are already available, the realization of all the functionalities required to obtain post silicon smart systems looks not far to come and EAPs will play a significant role. Because of their peculiarities, such as the capability to work in wet environments, low activation voltage, lightness, and flexibility, IPMCs could have the privileged technology for a number of niche applications.

Such applications will require autonomous systems and IPMC energy scavenging properties could be usefully exploited. Actually, other technologies could cooperate to power the envisaged systems and, for example, polymer-based organic photovoltaic solar cells already exist that are close to the market exploitation.

Last but not the less, the obtained organic systems will be environmental safe and therefore it will be possible to use them as disposable low cost devices.

Of course the IPMC technology is still at his infancy and big multidisciplinary efforts are required to let it evolve to a mature technology.

The production of IPMCs will be required to improve the materials repeatability and time duration which are two of the main weakness of this technology. Even if the researched have been concerned with the low force value that IPMCs can deliver, niche applications have been already envisaged for present IPMCs.

The modeling of IPMCs needs to improve the prediction quality of the proposed models as it regards nonlinearities, influence of environmental conditions and time-variant behavior. To this aim, the multiphysics approach can be usefully applied to cope with the complexity of phenomena that are involved in the IPMC transduction capabilities.

Last but not the least the prototypes need to be improved and need to show their capability to solve real life problems. This will be possible only if the whole IPMC technology chain will solve aforementioned drawbacks and weaknesses.


Research activity on IP2C described in the article has been partially supported by the PRIN ′08 project ‘Innovative all-polymeric ionic transducers for post-silicon applications: realization, modeling, metrological characterization and applications’ of the Italian Ministero dell'Istruzione dell'Università e della Ricerca.

Biographical Information

original image

Viviana De Luca was born in Messina on August, first 1985. She received the “First-level laurea” Degree in Electronic Engineering on July 2008 from the University of Messina (Italy). She receved the Master Degree in Automation Engineering and Control of Complex Systems on October 2011 from the University of Catania (Italy). Currently she is PhD student in System Engineering at the Department of Electric, Electronic and Informatic Engineering at the University of Catania, where her research interests are focused on FEM multiphysics modeling of ionic-polymeric devices.

Biographical Information

original image

Paolo Di Giamberardino received the master degree (Electronic Engineering) in 1991 and the PhD (Systems Engineering) in 1995 from the University of Rome “La Sapienza”. He is with the Department of Computer, Control, and Management Engineering Antonio Ruberti at Sapienza University of Rome since 1996 as Assistant Professor. He started working on analysis and control of nonlinear discrete time and sampled dynamical systems but he expanded his research interests including robotics (motion planning, control, coordination, visual feedback, visual servoing, etc.), sensor systems and devices and, more extensively, mobile sensor networks (deployment, planning, coverage, optimizations as well as data exchange, handling, fusion, etc.).

Biographical Information

original image

Giovanna Di Pasquale was born in Siracusa in 1958. She received the M.S. degree in chemistry from the Università degli studi di Catania, Catania, Italy, in 1981. She is currently an Associate Professor of chemical fundamentals of technology from the Dipartimento di Ingegneria Industriale, Università degli studi di Catania. She has been a Scientific Coordinator of the research unit of PRIN projects and is a Scientific Coordinator of research projects funded by the University of Catania. She is the author of more than 40 publications on international refereed journals and a considerable amount of communications to national and international conferences. Her research interests include polymeric materials, specifically synthesis and characterization, photooxidation, and stabilization.

Biographical Information

original image

Salvatore Graziani received the M.S. degree in electronic engineering and the Ph.D. degree in electrical engineering from the Università degli Studi di Catania, Italy, in 1990 and 1994, respectively. Since 1990, he has been with the Dipartimento di Ingegneria Elettrica, Elettronica e Informatica, Università di Catania, where he is an Associate Professor of electric and electronic measurement and instrumentation. His primary research interests lie in the field of sensors and actuators, signal processing, multisensor data fusion, neural networks, and smart sensors. He has coauthored several scientific papers and two books.

Biographical Information

original image

Antonino Pollicino received his summa cum laude Chemistry degree in 1983 at University of Catania. He has a position of full professor of Science and Technology of Materials at the University of Catania. His scientific activity is mainly focused on polymeric materials. He spent several periods abroad (United Kingdom and Australia) to work on advanced polymeric materials. The researches carried out can be grouped in the following topics: Synthesis and study of the structure-properties relationships of high performance polymers; Study and optimization of the composition of polymer surfaces; Synthesis and characterization of nanostructured polymer systems.

Biographical Information

original image

Elena Umana received the Degree in Computer Science Engineering from the Università degli Studi di Catania in 2005. She received the Ph.D. degree in Electronic, Automatic and Complex System Control Engineering from the Università degli Studi di Catania in 2010. She has performed a post graduate stage at STMicroelectronics labs on innovative organic circuit design, characterization and modeling in 2006. She is a GMEE Member (Group of Electric and Electronic Measures). Her research activity is focused on polymeric electromechanical sensors. Her activity is developed in the lab of the Department of Electric, Electronic and Computer Science of the University of Catania. She has authored over 20 journal and conference papers.

Biographical Information

original image

Maria G. Xibilia received the Laurea degree in Electronic Engineering in 1991 and a Ph.D. in Electric Engineering from the University of Catania, Italy in 1995. After a Post Doc position at University of Catania she become Assistant Professor of Automatic Control at the University of Messina where she is, since 2005, Associate Professor of Automatic Control and Robotics. Her research activities are mainly focused on innovative topologies of neural networks and cellular neural networks, hybrid strategies for nonlinear system identification, control and optimization, soft computing, supervision and control for complex plants and virtual sensors for industrial plants.