Biomimetic, low power pumps based on soft actuators

Sensors and Actuators A 135 (2007) 229–235

Biomimetic, low power pumps based on soft actuators Sonia Ram´ırez-Garc´ıa, Dermot Diamond ∗ Adaptive Sensors Group, National Centre for Sensor Research, School of Chemical Sciences, Dublin City University, Dublin 9, Ireland Received 13 February 2006; received in revised form 12 June 2006; accepted 15 June 2006 Available online 28 July 2006

Abstract A new biomimetic miniaturized pump based on soft materials is reported. The pump consists of a soft chamber made of polyurethane or polydimethyl siloxane (PDMS) tubes and a set of two soft electro-mechanical actuators in a cantilever configuration. The soft actuators were constructed using Nafion as an ionic polymer and polypyrrole as conducting material to make the electrodes by chemical in situ polymerisation. To our knowledge this is the first time this type of actuator has been reported. This choice of electrode material permits the pump to be actuated using lower potentials than when using platinum (±3 V as opposed to ±5 V when electroding using platinum). This leads to a much lower power consumption, which averaged to approximately 69 mW per stroke for an actuator that measured 3 cm length by 0.3 cm wide (as opposed to 227 mW when electroding with platinum). As the potential steps are applied to the actuators, they bend and produce a change of volume in the soft pumping chamber that generates the flow movement. The pumps developed in the present work were capable of flow rates of up to 1.6 ␮L/s, and could cope with backpressures of up to 0.025 Bar. This biomimetic pump is suitable for long term field deployable platforms, such as reagent based analytical platforms, since it is low power, corrosion resistant and also resistant to particle ingress due to its soft nature. © 2006 Elsevier B.V. All rights reserved. Keywords: Biomimetics; Soft actuators; Pumps; Polypyrrole-Nafion; Microfluidics

1. Introduction As analytical chemistry advances, more compact and miniaturized devices are being developed to achieve both lower volumes of waste and lower detection limits. Another objective of more compact and miniaturized devices is deployment in different environments while remaining inconspicuous. These devices often implement microfluidic systems. However, the pumps used with these systems are traditionally several orders of magnitude larger than the device itself and they consume large amounts of energy. New miniaturized pumps have been recently reported in the literature, which are based on different types of actuators, the most successful ones being those based on piezoelectric actuators [1,2]. However, piezoelectric actuators consume large amounts of energy, which is a great limitation for the development of autonomous field deployable sensing platforms based on microfluidic devices. Moreover, the pumps reported in this paper are based on soft actuators and have the advantage of overcom-

∗

Corresponding author. Tel.: +3531 7005404; fax: +3531 7007995. E-mail address: [email protected] (D. Diamond).

0924-4247/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.sna.2006.06.048

ing the problems generated by solid particles that would block and/or damage other rigid pumps. In the present work we report the development of a miniaturized pump based on an ionic polymer (Nafion® ) coated with polypyrrole that requires low amounts of energy to operate. To our knowledge this is the first time that this type of actuator (Ppy-Nafion-Ppy) has been reported. Traditionally Nafion has been used as a base material to construct ionic polymer–metal composites (IPMC) for various applications, e.g. as artificial muscles [3,4]. Their preparation involves the chemical reduction of platinum on both surfaces of the Nafion film which is both costly and time consuming. In this work polypyrrole was used as conducting material instead of Pt to form the electrodes on both sides of the Nafion. By doing this, the cost and time of preparation of the actuator were greatly reduced and in addition much lower power was required to actuate it. This type of actuator is normally allowed to swell in an aqueous electrolyte solution. When a potential is applied between the two platinum electrodes, the hydrated cations that are interacting with the sulfonic groups of the Nafion migrate to the anode, dragging some water molecules with them. This migration causes an increase of volume on the anode and a decrease of

230

S. Ram´ırez-Garc´ıa, D. Diamond / Sensors and Actuators A 135 (2007) 229–235

volume on the cathode, which produces the macroscopic effect of bending the material [5]. However, after a limited amount of cycles, these actuators begin to dehydrate due to evaporation and hydrolysis of water arising from the voltages required. In this case the actuators were maintained in an aqueous environment during the trials in order to avoid dehydration, as this leads to a decrease in actuation strength and a consequent reduction in observed flow rate. 2. Materials and methods The actuators were constructed using Nafion-117 films obtained by casting 150 ml of a 50% vol. mixture of dimethylformaide (DMF), purchased from Sigma and used as received, and a 5% Nafion-117 solution of saturated alcohols, also obtained from Sigma, in a 9 cm diameter circular-flat bottom mould. The electroding was performed using either platinum or polypyrrole (Ppy). When platinum was used, the procedure for electroding the Nafion film was the same reported previously [6]. The surface of Nafion-117 was roughened using sand paper and sonicated for 15 min in Milli-Q water to remove any residues. The film was then swollen by boiling it first in HCl 2 M (from Riedels de H¨aen) for 30 min, which also saturates the Nafion-117 with H+ , followed by boiling in Milli-Q water for 30 min Finally the Nafion-117 film was placed in a solution containing 2 mg of Pt(II)/ml ions overnight to allow the platinum cations to penetrate the ionic polymer. This solution was prepared using [Pt(NH3 )4 ]Cl2 in water. The total volume of solution was calculated so that there was 3 mg of Pt per cm2 of membrane area. The platinum cations trapped in the Nafion-117 were then chemically reduced to Pt0 using NaBH4 (5%, w:v solution). To ensure optimum conductivity of the Pt electrodes, a second electroding was performed by reducing more Pt(II) (also from a 2 mg of Pt/ml solution prepared using [Pt(NH3 )4 ]Cl2 salt) on the existing Pt0 electrodes using hydrazine (5%, w:v solution) and hydroxylamine hydrochloride (20%, w:v solution). The IPMCs constructed in this fashion were stored in LiCl 2 M for 3 days, to saturate the sulfonic groups of the Nafion-117 with Li+ , which was previously shown to produce stronger IPMCs [7]. When the electrodes were formed using polypyrrole (Ppy), the polypyrrole was synthesised by chemical oxidation of pyrrole on the Nafion film using a procedure reported previously [8]. A Nafion-117 film was pre-treated as above, i.e. it was first roughened using sandpaper and then sonicated for 15 min in Milli-Q water. The Nafion-117 film was then boiled in HCl 1 M for 30 min and in Milli-Q water for another 30 min to swell it. The polypyrrole was synthesised in situ by chemical oxidation and polymerisation. The Nafion-117 film was dipped in a 0.2 M pyrrole solution containing 0.005 M naphthalene-1,5-disulfonic acid disodium salt (NDSA) and a solution of 0.2 M FeCl3 was then slowly added while constantly stirring the solution. Once all the FeCl3 solution was added, the mixture was allowed to react overnight. The actuator was then washed in Milli-Q water and stored in LiCl 2 M for 3 days before use. Two pumps were developed. Pump A was made using a soft polyurethane tube (2 cm long, 1 mm i.d.) as shown in Fig. 1. The

Fig. 1. (A) Diagram of the biomimetic pump constructed using a 1 mm internal diameter polyurethane tube as pump chamber (pump A). (B) Schematics with dimensions of pump A.

configuration of this pump was very simple, consisting of a pump chamber (the soft polyurethane tube), a reservoir for the liquid pumped and a 1 mm i.d. hard polyurethane tube through which the fluid was pumped. The absence of valves or any other method of control of the direction of the flow resulted in no overall flow movement; however, this set-up permitted to evaluate the pump itself. Pump B was made using a PDMS tube (1.5 cm long, 6–7 mm i.d. and 100 ␮m thick walls) as shown in Fig. 2. The configuration of this pump consisted of a pump chamber (the PDMS tube), and conical inlet and outlet (made using 0.5–20 ␮m

Fig. 2. (A) Diagram of the biomimetic pump constructed using a PDMS tube as pump chamber (pump B). (B) Schematics with dimensions of pump B.

S. Ram´ırez-Garc´ıa, D. Diamond / Sensors and Actuators A 135 (2007) 229–235

pipette tips, type A). This system permitted the control of the direction of the flow in the absence of valves by the nozzle diffuser principle, as described previously [9,10]. The outlet was connected to 20 cm of a 1 mm i.d. hard polyurethane tube. Two strips of Ppy-Nafion-Ppy actuators were placed on each side of the pump chamber in a cantilever configuration so they moved as a pair of tweezers (see insert in Fig. 2). As ±3 V potential steps were applied they moved away/toward each other deforming the pump chamber and producing a pumping action. The actuators were constantly kept in Milli-Q water in order to avoid a decrease of mechanical efficiency by dehydration due to evaporation and hydrolysis of water. Scanning electron microscopy (SEM) and electron diffraction X-ray spectroscopy (EDX) were used to characterise the actuators. A Hitachi-2000 instrument was used at a working voltage of 20 KeV in the secondary electron mode. The actuators were actuated by applying a square wave potential (±5 V or ±3 V as specified in text, and the frequency was 0.2 Hz (5 s steps), 0.33 Hz (3 s steps), 1 Hz (1 s steps) and 2 Hz (0.5 s steps), also as specified in the text). A CH-Instruments potentiostat model 630 B was used. 3. Results and discussion 3.1. Characterisation of the actuators Two different types of actuators were considered for the fabrication of the pump: ionic polymer–metal composites (IPMC) fabricated using platinum, and Ppy-Nafion-Ppy actuators fabricated by chemical polymerisation of polypyrrole on the Nafion surface, avoiding the need of depositing a platinum layer. Polypyrrole is a conducting polymer that can be synthesised by chemical or electrochemical oxidation of the pyrrole monomers. If it is synthesised by chemical oxidation of the monomers, a platinum layer between the Nafion and the polypyrrole layer is not necessary. This makes the actuators and hence the pump structure cheaper and much easier to construct. When using Pt, great performance differences were observed between actuators batches due to uneven plating of the Nafion surfaces, whereas when Ppy was chemically synthesised on the Nafion surfaces, the performance of these actuators was very reproducible. The performance of both actuators was studied and compared. The actuators that were compared were constructed using Nafion of similar thickness and pre-treated using the same procedure but one was coated with Pt and the other with Ppy. SEM and EDX images were obtained of both actuators confirming comparable thickness (see Fig. 3A-1 and B-1). EDX images of the IPMC actuator mapping F and Pt (Fig. 3A-2 and 3A-3, respectively) permitted the determination of the electrode thickness, being of the order of 17 ␮m. The total thickness of the IPMC actuator was approximately 200 ␮m. The same study was performed with the Ppy-Nafion-Ppy actuators. From the SEM (Fig. 3B-1) and elemental distribution maps (Fig. 3B-2 and B-3) it was possible to estimate the approximate thickness of the Ppy layers (ca. 20 ␮m) and the total thickness of the actuator (ca. 190 ␮m).

231

In order to assess the actuation characteristics, the amplitude of movement was measured in degrees. This prevented any dependence of the results on the length of the actuators. These experiments were performed with actuators in a dry environment. The measurements were done by fixing the end of the actuators to the centre of a semicircle containing a radial grid used to measure the bending angle of the actuators. The electrical connexion to the potentiostat was made at the fixed end of the actuator, at the centre of the grid. Fig. 4 shows the results obtained. As it can be seen from the chronoamperograms, the movement of the Ppy-Nafion-Ppy actuators (Fig. 4B) were more reproducible than the movement of the IPMC, registering the same angle for each stroke and utilising the same charge. The chronoamperogram for the IPMC showed a slight increase in charging current during the first three to four strokes followed by a decrease due to evaporation and electrochemical breakdown of water due to the higher voltages used in this case (±5 V). This also resulted in a decrease of the angle of movement (from 40◦ to 35◦ in Fig. 4). However, IPMCs can achieve wider angles of movement than Ppy-Nafion-Ppy. For example, 50◦ was the maximum angle we obtained for IPMC actuators when 10 s pulses were applied, compared to 35◦ with a Ppy-Nafion-Ppy actuator under equivalent conditions. This observation could be explained by an increased mechanical rigidity of the actuator when Ppy was used instead of Pt. The maximum amplitude at the highest speed was achieved by applying ±3 V potential steps with the Ppy-Nafion-Ppy actuator, whereas with the IPMCs ±5 V potential steps were required. Also, the charging currents were lower for the PpyNafion-Ppy actuators (of the order of 20–30 mA) than for the IPMCs (45–50 mA) (see Fig. 4) which means that the average power consumption for the Ppy-Nafion-Ppy actuators is much lower (69 mW/stroke) than for the IPMCs (227 mW/stroke). The power consumption was determined by dividing the total charge passed trough the system on each pulse by the time length of the puls and then multiplied by the voltage. The charge passed through the system on each pulse was determined by numerically integrating the current intensity vs. time curves using the raw data. However, the strength of the Ppy-Nafion-Ppy actuators (200 mg which is equivalent to a force of 1.96 N) was lower than that of the IPMCs (500 mg which is equivalent to a force of 4.90 N). The strength of the actuators was measured using a precision balance. The actuator was placed parallel to the plate of the balance touching the plate. The balance was then set to zero and the actuator was stimulated with the appropriate voltage. The pressure exerted by the actuator on the plate of the balance was then measured and converted to a force. This probably underestimates the capabilities as it does not allow to distortion on the actuator which can compensate for the displacement. Despite their lower amplitude of movement and their lower strength, Ppy-Nafion-Ppy actuators were chosen to construct the pump because of their higher rigidity which makes them more effective when deforming a soft polymeric tube, and also because of their much lower power consumption, since one of the objectives of this pump is to integrate it into a field deployable sensing platform. Also, Ppy-Nafion-Ppy actuators were simpler and cheaper to make than the IPMCs.

232

S. Ram´ırez-Garc´ıa, D. Diamond / Sensors and Actuators A 135 (2007) 229–235

Fig. 3. (A-1) SEM image of the cross-section of the IPMC actuator, (A-2) EDX map of F of the cross-section of the IPMC actuator, and (A-3) EDX map of Pt of the cross-section of the IPMC actuator. (B-1) SEM image of the cross-section of the Ppy-Nafion-Ppy actuator, (B-2) EDX map of N of the cross-section of the Ppy-Nafion-Ppy actuator, and (B-3) EDX map of F of the cross-section of the Ppy-Nafion-Ppy actuator. The SEM and EDX images were obtained at an applied voltage of 20 KeV.

3.2. Biomimetic miniaturized pump Two different pump configurations were studied. The first pump constructed consisted of a pump chamber made of a 2 cm long polyurethane tube of 0.1 cm i.d. The type of polyurethane was very soft so it could easily be deformed by the actuators. The pump chamber was then connected to a 0.1 cm i.d. tubing using an adapter. Two Ppy-Nafion-Ppy actuator strips of 0.3 cm × 1.5 cm were placed on both sides of the tube in a tweezers configuration. As the actuators were stimulated with a 3 V potential steps they would close and open squeezing and releasing the polyurethane tube, hence producing the pumping action. This pump was called pump A (a picture and a schematic representation of this pump are illustrated in Fig. 1). Due to the lack

of valves or any other system to control the direction of the flow, there was no overall displacement of liquid. However, this configuration was suitable to characterize the design of the pump from the point of view of flow rates and the influence of the size of the pump chamber on the flow rate. The flow rate measurements were performed by partially filling the system with an aqueous dye solution and hermetically blocking the inflow. The distance traveled by the meniscus was measured and the equivalent volume was divided by the width of the potential step in seconds to obtain the flow rate (␮L/s). The second pump was a valve-less pump based on the nozzlediffuser principle [9,11], that was called pump B (see Fig. 2). The actuators used were the same as in pump A. Fig. 5 illustrates the nozzle diffuser principle applied to the pump B. This

S. Ram´ırez-Garc´ıa, D. Diamond / Sensors and Actuators A 135 (2007) 229–235

233

Fig. 5. Schematics of the valve-less pump based on the nozzle-diffuser principle.

Fig. 4. (A) Chronoamperogram obtained during actuation of the IPMC actuator when applying ±5 V square pulses. (B) Chronoamperograms obtained during actuation of the Ppy-Nafion-Ppy actuator when applying ±3 V square pulses. The line-graphs correspond to the current intensity vs. time graph, and the square dots correspond to the amplitude of movement (in degrees) vs. time.

pump showed flow mainly in one direction, with very small backflow (negligible compared to the main flow). The backflow observed could not be measured using the techniques available at the moment. Efforts are being made to develop a technique based on video processing to determine the backflow of the system. The flow rates were determined by measuring the time required for the meniscus to travel a known distance in the 0.1 cm i.d. tube and dividing the equivalent volume by this time. The flow rates registered using this configuration are shown in Table 1. The flow rates were calculated by measuring the time required for the meniscus to travel a known distance, and dividing the equivalent volume of liquid by the measured time. As shown on Table 1, as the duration of the potential steps was reduced (from 5 s to 0.5 s), the flow rate increased (from 0.4 ␮L/s to 1.3 ␮L/s). Most of the actuation happens in the first second of the potential pulse, so the extra time only delays the flow. As the potential steps decreased in duration, the tube was deformed more quickly and the dead times (when no further actuation happened) also decreased. During the characterization studies of this actuator it

was found that when the frequency of the potential was larger than 1 Hz, the amplitude of the movement started to decrease simply because of the speed of the actual movement of the actuator which is limited by the speed of migration of the hydrated ions from one side to the other of the actuator. When the actuator was integrated in the pump, the amplitude of the movement was also limited by the diameter of the pump chamber and its stiffness. However, when the potential frequency was 2 Hz (potential steps of 0.5 s), a smaller deformation of the pump chamber was already observed. Higher frequencies were therefore not tested since no appreciable deformation of the pump chamber was observed. Table 1 also displays the energy consumed at each flow rate with pump B, obtained by dividing the power required by the flow rates. The same calculation was made for a commercial peristaltic pump for microfluidic systems (BVT Technologies, model 2PP10.S, average current 100 mA and a DC potential of 3 V, power required 300 mW [12]), and for a miniature piezoelectric pump (Deak Technologies. Inc., models DTI-200-12P and DTI-200-12A, which consumes 150 mW of power [13]). It can be seen in Table 1 that the pump developed in the present work consumes less energy than a piezoelectric pump or than a commercial peristaltic pump (pump B requires 50% of the energy consumed by a piezoelectric pump and 25% of the energy required by the commercial peristaltic pump to produce a flow rate of 0.4 ␮L/s). This is a very important feature for miniature pumps for field deployable platforms, where the energy consumption is a critical feature. In addition to this feature, the pumps developed in this work do not have problems of corrosion due to rusting and since they are fully made of soft

Table 1 Correlation between flow rate and the width of the potential pulse obtained for pump B Puls width (s)

Flow rate (␮L/s)

Energy consumption (mJ/␮L) Pump B

Miniature peristaltic pump (BVT Technologies, model 2PP10.S, average current 100 mA, DC potential 3 V, power required 300 mW [12])

Miniature piezoelectric pump (Deak Technologies. Inc. models DTI-200-12P and DTI-200-12A, power required 150 m W [13])

5 3 1 0.5

0.4 0.8 1 1.3

188.50 119.50 240.25 187.92

750.00 375.00 300.00 230.77

375.00 187.50 150.00 115.38

Comparison of the energy consumption for different commercial miniature pumps and the pumps developed in this work.

234

S. Ram´ırez-Garc´ıa, D. Diamond / Sensors and Actuators A 135 (2007) 229–235

Table 2 Comparison between the flow rates obtained for pump A and pump B with closed inlet and correlation with the with of the potential pulses used

Table 3 Measurement of the flow rates for different backpressure set-ups using pump B with blocked inlet

Pump

Pulse width (s)

Flow rate (␮L/s)

Back pressure (bar)

Flow rate (␮L/s)

Pump A

5 3 1 0.5

0.16 0.26 0.79 1.3 1.4 1.6 1.6 1.6

1.6 1.6 1.6 1.3 0.4 0.1

Pump B-blocked

5 3 1 0.5

0.001 0.0035 0.005 0.01 0.02 0.025

and flexible materials, they will be able to deal with ingress solid particulates. Pump B was also tested when there was no overall flow through the system. This was accomplished by partially filling the system with an aqueous dye solution and closing the inlet to produce a watertight seal (pump B-blocked). The flow rate was measured as the displacement of fluid during each stroke divided by the width of the potential pulse applied, i.e. 5 s, 3 s, 1 s and 0.5 s (as done with pump A). The results are shown in Table 2, labelled as pump B-blocked. In this case higher flow rates are observed compared to the previous configuration of the pump due to an almost complete lack of backflow. These experiments carried out under open tube conditions permitted the comparison of this pump with pump A. As shown in Table 2, when pump A was used (a 1 mm i.d. polyurethane tube as pump chamber) the flow rates were smaller. Hence, the flow rates can be controlled not only by the width of the potential step but also by changing the size of the pump chamber. When pump A was used, the actuators were able to completely close the tube during the first second of actuation. However, even a 100% deformation of a smaller tube would translate into a much smaller volume of liquid being displaced. The flow rates of the two pumps were similar only when the strokes were very short (0.5 s).

Fig. 6. Schematics of the set up for backpressure experiments. The height of the outflow with respect to the pumping chamber was controlled using a movable platform. The backpressure was calculated using the vertical height of the outflow with respect to the pump chamber.

Pump B-blocked was also used to measure the capability of the actuators to cope with backpressure. The flow rate was calculated by measuring the distance travelled by the liquid and dividing the corresponding volume of fluid by the width of the potential step (in seconds). For these experiments the potential step chosen was 3 s (frequency of 0.33 Hz) for all the experiments. The backpressure was controlled by increasing the height of the end of the 1 mm i.d. tube relative to the level of the PDMS pump chamber, as shown in Fig. 6. Table 3 illustrates the results obtained. From this experiment we estimated that this pump was capable of producing the maximum flow rate up to backpressures of 0.005 bar (5 cm × 0.001 bar/cm). For higher pressures a decrease in flow rate was observed, although the pump was capable of producing a flow rate of 0.1 ␮L/s with a backpressure of 0.025 bar.

Fig. 7. Schematics of the final configuration of the biomimetic pump based on Ppy-Nafion-Ppy actuators.

S. Ram´ırez-Garc´ıa, D. Diamond / Sensors and Actuators A 135 (2007) 229–235

The reproducibility of the fluid movement in both pumps is limited by the reproducibility of the actuation, which is strongly dependent on the level of hydration of the actuators, similarly to what happens with living organisms. As expected, the actuators became dehydrated due to evaporation and electrochemical hydrolysis of water the strength and amplitude of movement decreased, hence decreasing the flow rate. The actuation mechanism of this type of actuators involves the migration of cations due to the potential applied between the two electrodes attached to the ionic polymer, and movement of water molecules that cause a differential swelling in both sides of the Nafion film. Therefore, actuation cannot happen in the absence of water. Therefore, to maintain the effectiveness of the pumping device, the actuators need to be maintained in an aqueous environment using a hydration jacket. The schematics of this set up are shown in Fig. 7. This approach is actively being pursued in our laboratory. At present, modified flap valves are being optimised in order to minimise their size and their dead volume to make them suitable to this system. The final size of the valves is not yet optimised, for this reason they are not given in Fig. 7. An alternative approach is to use non-volatile electrolytes such as ionic liquids [14]. However this approach led to actuators capable of producing much lower strains. Further studies using different ionic liquids are currently being carried out in our laboratory. 4. Conclusions A prototype of a biomimetic miniaturized pump has been presented. The pump was fully constructed using soft materials. The actuators chosen to construct it were formed by Nafion sandwiched between two polypyrrole layers chemically synthesised on the Nafion film. The absence of a Pt layer between the Nafion and the Ppy substantially reduced the cost of producing these actuators. Also, the procedure employed to make Ppy-Nafion-Ppy actuators was less time consuming and more reproducible than that of IPMCs. The resulting pump consumed very small amounts of energy, and was able to cope with backpressures of up to 0.025 bar. The power consumption of the pump developed in the present work is very competitive with commercially available pumps for micro-fluidic systems such as the dual channel peristaltic pump manufactured by BVT Technologies, model 2PP10.S and Miniature Piezoelectric Pumps by Deak Technologies. Inc., models DTI-200-12P and DTI-20012A. This biomimetic pump is suitable for field deployable platforms, such as reagent based analytical platforms, since it is low power, corrosion resistant since it has been entirely constructed with polymeric materials and also resistant to ingress due to its flexibility. Further improvements such as implementation of valves or other engineering problems are currently under development. Acknowledgement The authors wish to thank Science Foundation Ireland for grant support under the Adaptive Information Cluster Award (SFI 03/IN3/1361).

235

References [1] P. Woias, Micropumps—past, progress and future prospects, Sens. Actuat. B 105 (2005) 28–38. [2] N.-T. Nguyen, T.-Q. Truong, A fully polymeric micropump with piezoelectric actuator, Sens. Actuat. B 97 (2004) 137–143. [3] M. Shahinpoor, Ionic polymer–conductor composites as biomimetic sensors, robotic actuators and artificial muscles—a review, Electrochim. Acta 48 (2003) 2343–2353. [4] M. Shahinpoor, Y. Bar-Cohen, T. Xue, J.O. Simpson, J. Smith, Ionic polymer–metal composites (IPMC) as biomimetic sensors and actuators, Proc. SPIE (1998), Paper N. 3324-27. [5] M. Shahinpoor, Y. Bar-Cohen, J.O. Simpson, J. Smith, Ionic polymer–metal composites (IPMCs) as biomimetic sensors, actuators and artificial muscles—a review, Anal. Chim. Acta 289 (1998) 143–153. [6] M. Shahinpoor, K.J. Kim, Ionic polymer–metal composites: I. Fundamentals, Smart Mater. Struct. 10 (2001) 819–833. [7] K.J. Kim, M. Shahinpoor, Ionic polymer–metal composites: II. Manufacturing techniques, Smart Mater. Struct. 12 (2003) 65–79. [8] A. Malinauskas, Chemical deposition of conducting polymers, Polymer 42 (2001) 3957–3972. [9] E. Stemme, G. Stemme, Nozzle/diffuser pump, Swedish Patent Applic. N. 9 300 604-7. [10] A. Olsson, G. Stemme, E. Stemme, A valve-less planar fluid pump with two pump chambers, Sen. Actuat. A Phys. 46–47 (1995) 549–556. [11] E. Stemme, G. Stemme, A valveless diffuser/nozzle-based fluid pump, Sens. Actuat. A 39 (1993) 159–167. [12] http://www.bvt.cz/, 2006. [13] http://news.thomasnet.com/fullstory/454483/954, 2004. [14] M.D. Bennett, D.J. Leo, Ionic liquids as stable solvents for ionic polymer transducers, Sens. Actuat. A 115 (2004) 79–90.

Biographies Dr. Sonia Ram´ırez-Garc´ıa received her Honours degree in chemical science from the Universitat Aut`onoma de Barcelona, Spain, in 1997. She received her doctorate in chemical sciences in 2003. The award came from the Universitat Aut`onoma de Barcelona, but the research was done jointly with Dublin City University, Ireland, under the joint supervision of Prof. Salvador Alegret and Prof. Robert Forster. She is currently working as a postdoctoral researcher in the Adaptive Sensors Group, in the National Centre for Sensor Research, Dublin City University, Ireland, where she has also done part-time lecturing. Her research interests include the development and characterisation of conducting polymers, biomimetic systems based on actuators and the development of sensors and biosensors for environmental and biomedical applications. Dermot Diamond received his PhD from Queen’s University Belfast (Chemical Sensors, 1987), and was Vice President for Research at Dublin City University, Ireland (2002–2004). He has published over 150 peer reviewed papers in international science journals, is a named inventor in twelve patents, and is co-author and editor of two books on sensors and data processing. He joined DCU in 1987 as a member of the School of Chemical Sciences, and is a founder member of the National Centre for Sensor Research (www.ncsr.ie). In 2003, he helped to negotiate the award of D 5.6 million from Science Foundation Ireland to the ’Adaptive Information Cluster’ (AIC), a joint initiative linking the NCSR, the Centre for Digital Video Processing (DCU) and the Smart Media Institute (University College Dublin). He is currently Director of the AIC. His research interests are wide ranging, from molecular recognition, host-guest chemistry, ligand design and synthesis, electrochemical and optical chemical sensors and biosensors, lab-on-a-chip, sensor applications in environmental, clinical, food quality and process monitoring, development of fully autonomous sensing devices, wireless sensors and sensor networking. He is particularly interested in developing the potential of analytical devices and sensors as information providers for wireless networked systems, i.e. building a continuum between the digital and molecular worlds.