Introduction

The design and construction of robots for surgeon’s assistance is becoming a critical issue for clinical advances. Reducing operator fatigue, improving accuracy and increasing repeatability consume most of the efforts in the area of computer-integrated surgery and robotic assistants [1, 2].

The simultaneous use of magnetic resonance images for the visual control of the surgeon area imposes limitations to the use of ferromagnetic materials [3]. Polymers are becoming the most suitable materials for developing two basic robotic components: actuators [413] and mechanical sensors [1423]. Indeed artificial muscles based on conducting polymers work as haptic motors [2427]: one physically uniform device fulfils both actuating and sensing requirements. From an engineering point of view, they are very robust motors due to its Faradaic nature: the position of the motor is under linear control of the consumed charge, the rate of the movement is under linear control of the applied current, and different devices with different geometry or including different masses of conducting polymers produce the same angular displacement under flow of the same charge per unit of conducting polymer mass and per unit of time [2831].

Different families of materials constitute the named conducting polymers [3237]. One of the remaining problems to get confidential products from artificial muscles based on conducting polymers is that some of those conducting polymer families use to develop active creeping effects. After submitting the device to a potential cycle, or to a charge cycle, the device doesn’t recover its initial position originating a continuous displacement of the movement range on consecutive actuation cycles [3843]. Creeping effects require quite complex self-compensation control processes to get confident products [4450]. A more precise knowledge of the creeping effect origin may allow its elimination or an easier theoretical compensation. Not many efforts have been done in this direction. Recently, we have discovered from the voltammetric and coulo-voltammetric responses of self-supported electrodes of polypyrrole (PPy) blends with organic macro-anions (MA) in aqueous solutions cycled up to −2.5 V that most of the involved charge is consumed by the reversible film oxidation/reduction with exchange of cations (C+) and water (S) with the electrolyte following the general reaction (1) [28]:

$$ {\left[\left({\mathrm{Pol}}^0\right){\left({\mathrm{MA}}^{\hbox{-}}\right)}_n{\left({\mathrm{C}}^{+}\right)}_n{\left(\mathrm{S}\right)}_m\right]}_{\mathrm{gel}}\rightleftarrows {\left[\left({\mathrm{Pol}}^{n+}\right){\left({\mathrm{MA}}^{\hbox{-}}\right)}_n\right]}_{\mathrm{s}}+n\left({\mathrm{C}}^{+}\right)+m\left(\mathrm{S}\right)+n{\left({\mathrm{e}}^{\hbox{-}}\right)}_{\mathrm{metal}} $$
(1)

In addition, a minor fraction of the consumed charge gives irreversible reactions at high cathodic potentials, overlapping reaction 1 backwards. Coulo-voltammetric responses allow a quantitative determination of this irreversible charge from self-supported electrodes of polypyrrole blends with dodecyl-bencylsulphonic acid (DBSH) [51] or para-benzolsulfonic acid (PBSH) [52] or from polypyrrole electrodes coating metals [53]. This irreversible open coulo-voltammetric fraction is not present in responses from PPy electrodes exchanging small anions [51, 54]. Irreversible charges were attributed to the hydrogen evolution from the organic acid constituent. Here, we will investigate the evolution under potential cycling of the angular movement described by a bilayer PPy-DBS bending artificial muscles with the consumed charge from the coulo-dynamic (charge-angle) responses quantifying, simultaneously, charges consumed per cycle by reversible and irreversible reactions trying to find some correlation between the irreversible charge and creeping effects.

Experimental

The electrogeneration of the polypyrrole films, the construction of the PPy/tape bilayer muscles, the experimental electrochemical procedures, and the working conditions have been described in previous papers [51, 55]. The PPy/tape bilayer is a full polymeric device without any metal content. The metal clamp allowing the PPy electrical contact with the working electrode (WE) plug from the potentiostat/galvanostat is located outside the electrolyte. In order to avoid uncertainties related to the possible presence of the irreversible hydrogen evolution by water electrolysis [53] at the polymer/metal clamp interface, a transversal lacquer strip is painted around the bilayer below the clamp contact to prevent the electrolyte ascension by capillarity. The bilayer muscle below the lacquer strip was immersed in the electrolyte.

All electrochemical studies were performed using an Autolab PGSTAT-100 potentiostat/galvanostat controlled by a personal computer using a (GPES®: General Purpose Electrochemical System) electrochemical software.

The electrochemical measurements were carried out in LiClO4 aqueous solutions. Parallel angular displacements of the bending muscles were recorded with a vision system using EVI-D31 SONY® digital cameras controlled by a Matrox® card and a control system programmed in C++ for image processing in Matlab®. The experimental setup is shown by Fig. 1.

Fig. 1
figure 1

Scheme of the electrochemical cell and configuration of electrodes (WE working electrode, CE counterelectrode, RE reference electrode) used to follow the electrochemical behaviour of the bilayer actuator (tape/PPy-DBS). A vision system constituted by two perpendicular video-cameras was used to record and follow the angular movement in aqueous electrolyte solutions

Results and discussion

The bilayer muscle was submitted in 0.1 M LiClO4 aqueous solution to consecutive potential cycles between −2.50 and 1 V, vs. Ag/AgCl (3 M KCl), at a sweep rate, ν, of 6 mV s−1. The counter electrode was a stainless steel plate (4 cm2). During the initial 10 consecutive potential sweeps, the system shows rising voltammetric responses corresponding to the adaptation, after electrosynthesis, of the internal polymer film structure to the new oxidation/reduction conditions: any structural memory from the electrogeneration is erased by cycling. After those initial cycles, stationary voltammetric responses to consecutive potential cycles are obtained. Any new experimental change (i.e., potential limits, electrolyte concentration) gives stationary responses after only two consecutive cycles. This bending bilayer tape/PPy-DBS muscle (a full polymeric system, without any metal in-between both films) translates very small PPy volume variations into large macroscopic bending angles of up to 145°. Figure 2a depicts the stationary third voltammetric response from the bilayer. Ten characteristic reference points (1 to 10) were indicated on the voltammetric response (Fig. 2a).

Fig. 2
figure 2

a Stationary voltammetric response from a tape/PPy-DBS bilayer muscle submitted to consecutive potential sweeps between −2.50 and 1.0 V at 6 mV s−1 in 0.1 M LiClO4 aqueous solution. b Coulo-voltammetric response obtained by integration of the voltammogram, and parallel evolution of the angle described (dots) by the muscle: dynamo-voltammetric response. c Pictures (c1 to c10) show the bending position of the muscle corresponding to the points 1 to 10 from Fig. 2a, b, d. d Evolution of the angular displacement of the muscle with the consumed charges, coulo-dynamic response, along the potential cycle

By voltammetric integration, the evolution of the charge (coulo-voltammetric, charge/potential or Q/E, responses [56]) consumed during the potential sweep by the PPy-DBS reactions is attained (Fig. 2b). Points 1 to 10 correlate those from the voltammogram (Fig. 2a). Negative charge increments indicate reduction reaction and positive increments refer to oxidation reactions.

In parallel to the voltammetric control (Fig 2a), the bending movement of the bilayer muscles was video recorded. Pictures c1 to c10 from Fig. 2 show the video frames corresponding to the bended muscle at the points 1 to 10, respectively, from Fig. 2a, b. The angle described by the film bottom was measured from the video frame corresponding to each muscle potential as indicated in Fig. 1 [54]. Figure 2b, dotted line, shows the dynamo-voltammetric (angle-potential) muscular response. Figure 2d shows the experimental coulo-dynamic (charge-angle) muscular response. Positive angular displacements mean clockwise bending movement of the muscle bottom and negative, anticlockwise displacements. Anticlockwise bending movement is observed during the PPy-DBS oxidation from point 2 to point 6 and clockwise bending movement during the PPy-DBS reduction, from point 1 to point 2 and from point 6 to point 10. Taking into account that the relative position of the muscle layers in the pictures is as follows: tape (left side)/PPy-DBS (right side) bending movements indicate that the PPy-DBS film shrinks by expulsion of cations and water (picture c2 to c6) during the positive increment of the consumed charge from point 2 to point 6 (Fig. 2b), that means during the PPy-DBS film oxidation (reaction 1 forwards). The film swells by entrance of cations and water driven by the PPy-DBS reduction, reaction 1 backwards, with negative increment of the charge (Fig. 2b) producing a clockwise movement (pictures c1 to c2 and c6 to c10).

Slopes from the coulo-voltammetric responses quantify the PPy-DBS layer reaction rate (reaction 1 forwards and backwards) at every experimental potential:

$$ \frac{\varDelta Q\left(\mathrm{C}\right)}{\varDelta E\left(\mathrm{V}\right)}\frac{\nu \left({\mathrm{V}\cdotp \mathrm{s}}^{-1}\right)}{F\left({\mathrm{C}\cdotp \mathrm{mol}}^{-1}\right)}\frac{1}{m\left(\mathrm{g}\right)}=\frac{\nu \cdotp \varDelta Q}{F\cdotp \varDelta E\cdotp m}\left(\mathrm{mol}\cdotp {\mathrm{s}}^{-1}\cdotp {\mathrm{g}}^{-1}\right)=r\left(\mathrm{mol}\cdotp {\mathrm{s}}^{-1}\cdotp {\mathrm{g}}^{-1}\right) $$
(2)

where Δ indicates variation, F is the Faraday constant (96 500 C mol−1), and ν (mV s−1) is the experimental potential sweep rate.

Any QE slope variation identifies either, a different reaction driven structural processes in the PPy/DBS film, its potential ranges and the involved charges [51, 53]. Thus, Fig. 1b corroborates that PPy/DBS film reduction-swelling (reaction 1 backwards) goes on between the initial potential (point 1) and the QE minimum, point 2. There, the film oxidation-shrinking (from point 2 to 4) and its subsequent oxidation-conformational packing (from point 4 to 6) with expulsion of cations and water (reaction 1, forwards) takes place. The separation point between oxidation-shrinking and oxidation-compaction regions is the closing potential.

After the QE maximum, the film reduction starts with a very slow reduction-relaxation process (points 6–7), followed by the fast reduction-swelling process (point 7 to 8) and then a slow reduction process (points 8 to 10) attributed to the formation of vesicles (water and ions surrounded by a DBS layer) [33, 5759] in the film, similar to those observed in colloidal dispersions of the surfactant [60].

The reduction charge (Fig. 1b full line) always drives the volume increment of the PPy-DBS film due to the entrance of cations and water giving a parallel angular displacement (Fig. 1b, dotted line).

The slow reduction process at cathodic potential (from 9 to 10 and from 1 to 2) overlaps the parallel presence of an irreversible process, revealed by the open part of the QV response (Fig. 1b) and attributed to the DBSH reduction with slow hydrogen evolution [51, 53]. The charge difference between the initial and final points of the QE response at the cathodic potential limit (the open QE part) is the irreversible charge consumed by the irreversible and slow hydrogen evolution [51, 53]. The charge difference between the QE closed loop maximum and its minimum is the reversible film oxidation/reduction (reaction 1) charge and, considering the different slopes, the charge involved by each of the reaction driven structural processes.

For those bilayer muscles constituted by CPs which coulovoltammetric responses doesn’t show any irreversible reaction in a similar potential ranges a linear coulo-dynamic (Q/α) response is attained, proving the Faradaic origin of the movement in the full potential range [2931, 54]. Here despite the presence of some reduction irreversible charge the coulodynamic response (Q/α), Figure 2d, corroborates the faradic (reaction 1) nature of the motor movement in the full potential range:

$$ \alpha ={\alpha}_0+kQ $$
(3)

where α (degrees) is the described angle, α 0 is the initial position of the muscle, Q (C) is the consumed charge and k (degrees C−1) is the coulodynamic constant (the empirical slope from Fig. 2d) of the system conducting polymer-tape-electrolyte. Through reaction 1 the charge controls the number (n) of monovalent cations moving into the film during reduction or expulsed during oxidation: n = Q/e, where e represents the electron charge. Every new cation moving into the film membrane drives the entrance of a defined number of water molecules for osmotic balance [6165]. The result is that the charge controls the PPy-DBS film volume variation, the concomitant variation of the generated stress gradient across the bilayer muscle, and the correlated amplitude of the described angle.

The most significant aspect from Fig. 2b and 2d is the presence of both, a cathodic irreversible reduction charge of 25 mC, versus a total reversible charge of 252 mC, and an irreversible clockwise (related to the muscle position at the beginning of the potential sweep: point 1 to point 10 distance) bending shift of 14° (creeping angle) versus a total described angle of 145°. As a consequence of the creeping effect after four consecutive cycles, each producing a creeping shift of 14°, the reduced muscle’s bottom at the end of the cathodic sweep moved outside the solution. Under similar conditions, bilayer muscles constituted by conducting polymers without irreversible reaction doesn’t present, according with the literature [30, 31, 54], creeping effects. Those results point to the irreversible charge as origin (or one of the possible components) of the creeping effect in PPy-DBS films. This charge originates the irreversible hydrogen evolution from DBSH component and a parallel irreversible film swelling per potential cycle, as deduced from the creeping angle.

Once identified, the possible origin of the creeping effect different experimental approaches will be investigated during subsequent works in order to compensate physical (creeping) and chemical (local concentrations) effects of the creeping driven irreversible reaction.

Conclusion

Coulo-voltammetric (charge-potential) responses from −2.5 to 1 V from a full organic tape/PPy-DBS bilayer muscle in aqueous solutions indicates the presence of parallel slow irreversible reactions at high cathodic potential overlapping the reversible reduction of the PPy layer. The charge spent by the irreversible process is around 10 % of the coulo-voltammetric charge. A constant angle of 145° is described (go and back) per cycle but shifting 14° clockwise per cycle (creeping effect). The reversible, per cycle, film oxidation/reduction charge measured form the coulo-voltammetric-closed loop is 252 mC per cycle. The irreversible charge per cycle obtained from the QE open part is 25 mC per cycle. The clockwise creeping displacement per cycle points to the presence of an irreversible PPy-DBS swelling linked to the irreversible hydrogen evolution from the HDBS component. The different slopes from the QE responses corroborate the presence of the reaction driven structural changes: oxidation-shrinking, oxidation-compaction, reduction-relaxation, reduction-swelling, and reduction-vesicle formation. The coulo-dynamic (Q/α) response of the muscle corroborates that, despite the presence of the irreversible hydrogen evolution, the muscle still behaves as a Faradaic polymeric motor: the described angle follows a linear dependence of the consumed charge up to −2.5 V.