Biomacromolecules 2000, 1, 642-647
642
Electrical Activation of Artificial Muscles Containing Polyacrylonitrile Gel Fibers H. Brett Schreyer,* Nouvelle Gebhart, Kwang J. Kim, and Mohsen Shahinpoor Artificial Muscles Research Institute, Department of Mechanical Engineering, University of New Mexico, Albuquerque, New Mexico 87131-1386 Received May 9, 2000; Revised Manuscript Received July 17, 2000
Gel fibers made from polyacrylonitrile (PAN) are known to elongate and contract when immersed in caustic and acidic solutions, respectively. The amount of contraction for these pH-activated fibers is 50% or greater, and the strength of these fibers is shown to be comparable to that of human muscle. Despite these attributes, the need of strong acids and bases for actuation has limited the use of PAN gel fibers as linear actuators or artificial muscles. Increasing the conductivity by depositing platinum on the fibers or combining the fibers with graphite fibers has allowed for electrical activation of artificial muscles containing gel fibers when placed in an electrochemical cell. The electrolysis of water in such a cell produces hydrogen ions at an artificial muscle anode, thus locally decreasing the pH and causing the muscle to contract. Reversing the electric field allows the PAN muscle to elongate. A greater than 40% contraction in artificial muscle length in less than 10 min is observed when it is placed as an electrode in a 10 mM NaCl electrolyte solution and connected to a 10 V power supply. These results indicate potential in developing electrically activated PAN muscles and linear actuators, which would be much more applicable than chemically activated muscles. Introduction Artificial muscles are materials that can be either chemically or electrically activated, thus not requiring mechanical devices (gears, motors, etc.) for activation, and they include a variety of geometries and materials. Two prevalent material classes are conductive polymers and polyelectrolyte gels. A variety of electrically activated structures with rapid response times (a few seconds or less) and high force produced per unit mass have been made from bilayers of conductive (or conjugated) polymers, such as polypyrrole, polyacetylene, and polyaniline, and nonconductive polymers and/or metals.1-7 These materials have great potential for use as micromanipulators or microactuators due to their excellent bending and folding characteristics, but they have limited axial displacement and few large structures (with dimensions in the centimeter to meter range) have been made. Polyelectrolyte gels are known for their large volume change, and common polymers of these gels include polyacrylamide (PAM), poly(vinyl alcohol)-poly(acrylic acid) (PVA-PAA), and poly(2-acrylamido-2-methylpropane) sulfonic acid (PAMPS).8-21 Under an electric field, these gels are able to swell and contract, inducing large changes in gel volume within a few seconds or less. Such changes in volume can then be converted to mechanical work. One pitfall of polyelectrolyte gels is that either they are mechanically weak or they are unable to hold a significant load. Artificial muscles made from chemically activated polyacrylonitrile * To whom correspondence should be addressed. Current address: Department of Chemical and Nuclear Engineering, University of New Mexico, Albuquerque, NM 87131-1341. E-mail:
[email protected].
(PAN) gel fibers are mechanically stronger and have a much greater change in length than most other polyelectrolyte gels. For these reasons, PAN gel fibers have a greater potential for application as linear actuators and artificial muscles than other gels. The starting materials to prepare chemically activated artificial muscles are fibers of the inelastic polymer PAN. This simple polymer has a hydrocarbon backbone with a nitrile group attached at every other carbon (-CH2CH(CN)-)n. The method to prepare chemically activated PAN gel fibers was elucidated by S. Umemoto and colleagues.22 The method involves two steps: the first step is to anneal the fibers at temperatures between its glass transition temperature (about 110 °C) and thermal degradation temperature (about 250 °C). This cross-links polymer chains within PAN. The second step converts remaining nitrile groups on the cross-linked polymer to carboxylic acids by saponification with a strong base. The structure (such as degree of cross-linking and cyclization) and properties (such as mechanical strength and modulus of elasticity) of PAN gel fibers are determined primarily by the annealing process.22,23 There is still uncertainty about the structure, but some aspects of the structure are clear, such as type of bond formation. Structural analysis has determined that CdN and CdC bonds are present in annealed PAN, likely the products of cyclization and dehydrogenation reactions.24-26 In general, annealed PAN likely consists of cyclic structures and cross-links, but details are in dispute.24-35 Several researchers suggest a structure including aromatic nitrogen rings with intermolecular nitrile
10.1021/bm005557l CCC: $19.00 © 2000 American Chemical Society Published on Web 09/26/2000
Electrical Activation of PAN Artificial Muscles
Biomacromolecules, Vol. 1, No. 4, 2000 643
Materials and Methods
Figure 1. Structure of PAN gel fibers. Ring structures and nitrile cross-links are formed during heat treatment of PAN and carboxylic acid groups are added during saponification.
cross-links,22,24,25 where the degree of cyclization and crosslinking is a function of annealing time and temperature.22,25 The uncertainty may stem from variations in degree of cyclization and cross-linking based on differences in annealing time and temperature among researchers. Any remaining nitrile groups on annealed PAN are likely converted to carboxylic acid groups during the saponification step. This converts the heat-treated PAN fibers to elastic fibers. Umemoto and co-workers have confirmed the presence of nitrogen-containing ring structures and carboxylic acid groups on PAN gel fibers.22 Due to the presence of such structures and groups on PAN gel fibers, the tertiary structure of the polymer is sensitive to charge and pH. The resulting PAN gel fibers contract when exposed to acidic aqueous medium and elongates in a strong alkaline medium.22,36 A possible structure of PAN gel fibers based on current knowledge as described above is given in Figure 1. Early research on polyelectrolyte gels produced a poly(acrylic acid) gel that contracted up to 20% with addition of acid.37 The ability of PAN gel fibers to contract to over half its initial length makes it an appealing material for use as linear actuators and artificial muscles, but the use of acids and bases to contract and elongate PAN have limited its application. An attractive alternative is electrical activation. Several workers have made attempts toward a fast acting, electrically controllable linear actuator/artificial muscle.16,38,39 For example, a poly(vinyl alcohol)-poly(acrylic acid) copolymer with platinum deposited on its surface was activated in an electrochemical cell. The polymer decreased in length by about 5%, and both contraction and elongation took about 12 min.38 During electrolysis of water, hydrogen ions are generated at the anode while hydroxyl ions are formed at the cathode in an electrochemical cell. Electrochemical reactions can then potentially be used to control the length of PAN artificial muscles. This may be achieved by either locating a PAN muscle near an electrode where the ions are generated, or, if the conductivity of PAN gel fibers can be increased, the PAN muscle can serve as the electrode itself. The study reported herein takes the second approach, where conductivity of PAN gel fibers is increased by (1) depositing platinum on the fibers and (2) combining them with graphite fibers. In addition to reporting on results of electrical activation of PAN gel artificial muscles, mechanical properties of PAN gel fibers are also examined.
Fibers of polyacrylonitrile (PAN) were obtained from Mitsubishi Rayon Co., Ltd. (Japan). Each fiber is composed of roughly 2000 intertwined PAN fibrils. Fibers were annealed at 240 °C in an oven (Imperial Laboratory, Melrose Park, IL) for 2 h in an air environment. Annealed fibers placed in a boiling solution of 1 N sodium hydroxide (NaOH, Aldrich, Milwaukee, WI) for 30 min carried out the saponification step. Artificial muscles made of PAN gel fibers were prepared by bundling annealed fibers together and then boiling in NaOH. PAN gel fibers were contracted using 2 N hydrochloric acid (HCl, Aldrich) and elongated using 2 N NaOH. ASTM standard method E111-82 was used to determine the modulus of elasticity (E) for PAN gel fibers.40 Several bundles of PAN gel fibers were tested, all identical in length but varying in number of fibers per bundle. The load was applied uniaxially and added incrementally until failure. Stress was calculated based on the cross-sectional area of annealed PAN fibers. The diameter of a single PAN fibril was measured to be about 7.0 µm from scanning electron microscopy (SEM, JEOL 5800LV at 20 keV) analysis, and with 2000 fibrils per fiber, the cross-sectional area of a single fiber is estimated to be 7.70 × 10-8 m2. Conductivity of PAN gel fibers was increased by chemical deposition of platinum. Raw PAN fibers were first cleaned by soaking in hot (about 80-90 °C) 2 N HCl for 30-40 min, followed by a rinse with deionized water (dH2O). The fibers were allowed to dry and then immersed in 0.1% (w/ v) tetraamineplatinum chloride (Pt(NH3)4Cl2, Aldrich) in dH2O. The PAN fibers were then transferred to a reducing solution containing 300 mL of dH2O, 0.5 mL of concentrated ammonium hydroxide (NH4OH, JT Baker, Phillipsburg, NJ), and 4 mL of 5% (w/v) sodium borohydride (NaBH4, Aldrich). The solution was slowly heated to about 50-60 °C with agitation (100 rpm) along with periodic additions of 4 mL of 5% NaBH4 until all Pt salt is reduced to Pt metal. The fibers were rinsed with dH2O, followed by 2 N HCl, and again with dH2O. This process was repeated several times to seed the PAN fibers with Pt. To coat fibers with Pt, fibers were placed in a solution containing 0.05% (w/v) Pt(NH3)4Cl2, 0.05% (w/v) NH4OH, 0.1% (v/v) hydrazine monohydrate (H2NNH2‚H2O, Aldrich), and 0.05% (w/v) hydroxylamine hydrochloride (H2NOH‚HCl, Mallinckrodt Inc., St. Louis, MO). The solution was slowly heated to 60 °C over about 60 min with agitation (100 rpm) along with periodic additions of H2NNH2‚H2O and H2NOH‚HCl. This step deposited a uniform layer of Pt on PAN fibrils. The PAN-platinum (PAN-Pt) fibers were then rinsed with 2 N HCl and dH2O. The fibers were then annealed and boiled in NaOH (as described above) to prepare PAN-Pt gel fibers. Several PAN-Pt fibers were bundled together to produce a PAN-Pt gel artificial muscle. In addition to PAN-Pt muscles, PAN fibers were combined with graphite fibers to produce conductive PANgraphite (PAN-Gr) artificial muscles. Graphite fibers were prepared and donated by Sandia National Laboratories (Albuquerque, NM). Fibers of PAN were converted to graphite fibers by carbonization in an inert atmosphere above
644
Biomacromolecules, Vol. 1, No. 4, 2000
Schreyer et al.
Figure 2. Experimental setup for electrical activation of PAN-Pt (or PAN-Gr) artificial muscles: (A) power supply; (B) PAN muscle/ electrode; (C) 3.0 g weight; (D) platinum electrode; (E) 10 mM NaCl electrolyte solution.
temperatures of 1000 °C. The resulting graphite fibers had fibrils with diameters about the same size as PAN fibrils, roughly 6.4 µm as determined by SEM analysis. Graphite fibers and annealed PAN fibers were intertwined and bundled prior to saponification with NaOH. To minimize the effect of inelastic graphite fibers on expansion/contraction of a PAN-Gr muscle, the graphite fibers were cut to lengths matching that of the elongated length of PAN gel fibers prior to intertwining with annealed PAN and the following saponification step. The graphite fibers were not elastic, but they were flexible and able to bend easily when PAN gel fibers in a muscle contracted. Variations in the mass ratio (dry basis) of graphite to PAN were tested for optimal electrical activation, e.g., reproducibility and rate of contraction. A PAN-Pt or PAN-Gr artificial muscle was placed in an electrochemical cell as an electrode for electrochemical activation. The other electrode in the cell was a strip of Pt metal (1.55 × 1.70 cm) and the electrolyte was 10 mM sodium chloride (JT Baker). A BK Precision (Chicago, IL) DC power supply (model 1630) served as the source of electricity. A weight (3.0 g) was attached to the lower end of the muscle to keep it taught and vertical in the electrolyte solution. To induce contraction, the muscle was configured as the anode in the cell and for elongation by having it serve as the cathode. Electrical activation was based on the electrolysis of water. The half reaction at the anode was (E° ) reference potential) 2H2O f O2 + 4H+ + 4e- E° ) -1.23 V
(1)
The relevant half reaction at the cathode was 2H2O + 2e- f H2 + 2OH- E° ) -0.83 V
(2)
PAN artificial muscles were contracted or elongated in the cell by electrochemical generation of H+ or OH-, respectively, in the near vicinity of the PAN gel fibers. Figure 2 is a diagram of the experimental apparatus. Results Length of PAN gel fibers was strongly dependent on the pH of the aqueous solution bathing the fibers. An example of this chemical activation of a PAN artificial muscle is given in Figure 3. Increasing the pH of the aqueous solution by addition of a strong alkaline solution brought about a rapid increase in length while addition of a strong acid induced
Figure 3. Length of PAN gel fiber artificial muscle (50 fibers) as a function of pH. Onset of contraction occurred at pHs below about 4.5 while elongation initiated at pHs above 6.0.
Figure 4. Stress vs strain diagram for contracted and elongated PAN gel fiber artificial muscles. The muscles contained 25 fibers. Initial lengths of contracted and elongated muscles were 5.7 and 10.8 cm, respectively.
rapid contraction. Results given in Figure 3 were very reproducible with length increasing to twice its initial value or more upon elongation. Repeated pH activation (over 30 cycles of contraction/elongation) produces little degradation in performance as monitored by rate and magnitude of length change, and PAN gel fibers are quite stable when PAN is stored in its contracted state. A hysteresis was noted in that elongation did not begin until the pH was raised above 6 while contraction did not start until pH was decreased below 4.5. Rate of contraction is dependent on strength of acid solution. Acid solutions with concentrations of 5 N produced contraction rates of about 1.4 cm/s, while for a 1 N solution, contraction occurred at about 1.1 cm/s. When PAN gel fibers were removed from an aqueous environment and allowed to dry, they became rigid with a length close to that of contracted fibers. The mechanical properties of PAN gel fibers, especially in the contracted state, were analogous to properties of rubber, i.e., the modulus of elasticity increased as strain increased. Sample stress vs strain curves for contracted and elongated PAN gel muscles are given in Figure 4. In general, contracted PAN muscles were much more elastic and withstood a load about three times that of elongated PAN before breaking. Table 1 summarizes the results of mechanical testing. Results of electrical contraction and elongation of a 5.0 cm long PAN-Pt artificial muscle containing 20 fibers is
Biomacromolecules, Vol. 1, No. 4, 2000 645
Electrical Activation of PAN Artificial Muscles Table 1. Mechanical Properties of Contracted and Elongated PAN Gel Fibers, Where Reported Values Are Mean Values from Multiple Tests PAN gel fibers
modulus of elasticity, E (MPa)
% elongation upon break
contracted elongated
1.0 3.9
240 42
Figure 6. Electrical activation of PAN-Gr gel fiber artificial muscle in electrochemical cell. Mass ratio of PAN to graphite in the muscle was 1:2. Muscle contracted while serving as cell anode and elongated while serving as cathode. Initial length of muscle ) 5.0 cm, cell voltage ) 10 V, and cell current ) 280 mA.
Figure 5. Electrical activation of PAN-Pt gel fiber artificial muscle (20 fibers) in electrochemical cell. Muscle contracted while serving as cell anode and elongated while serving as cathode. Polarity of electrodes reversed at t ) 10 min. Initial length of muscle ) 5.0 cm, cell voltage ) 20 V, and cell current ) 120 mA.
shown in Figure 5. With 20 V applied to the cell, a PANPt muscle at the anode contracted from a length of 5.0 cm to 2.8 cm, or a 44% decrease in length, in less than 10 min. For a 20-fiber muscle pulling on a 3.0 g weight, the load was about 0.019 MPa. After 10 min, the polarity of the electrodes were switched to elongate the muscle (PAN-Pt muscle ) cathode). After electrode polarity was switched, the muscle continued to contract, but within 1 min of the switch, the muscle began to elongate. The muscle expanded from a length of 2.7 cm to 4.1 cm in 11 min, not returning to the starting length of 5 cm. Approximate mean rate of contraction was calculated to be 0.19 cm/min, while approximate mean rate of elongation was 0.11 cm/min. Resistivity of the PAN-Pt gel fiber muscle was about 1.2 × 10-4 Ω‚m, but increased to about 24 Ω‚m after activation. This increase in resistivity was attributed to Pt separating from PAN gel fibers during electrical activation. Most Pt separated from the PAN within the first cycle of activation, and within two or three cycles, the PAN-Pt muscle would fail to contract or elongate. Although these muscles no longer responded to electrical activation, they still responded to chemical/pH activation. Several muscles with various PAN to graphite mass ratios were tested. Each demonstrated high reproducibility, and a muscle with 1:2 PAN to graphite mass ratio had the fastest rate of contraction. Electrical activation of this muscle is shown in Figure 6. Here 10 V was applied to the cell and the muscle contracted 1.35 cm (27% decrease in length) within 10 min. Two cycles of contraction/elongation are shown in Figure 6, and the pattern was reproducible over several cycles. The mean rate of contraction was 0.31 cm/ min while the mean rate of elongation was 0.17 cm/min. The PAN-Gr muscle contained seven fibers of PAN that served as the contraction/elongation mechanism, while the
graphite fibers served as conducting medium for the muscle. The graphite fibers were thin and flexible, but did not change length. With a 3.0 g weight attached to the muscle, the load was determined to be about 0.55 MPa. The load applied to the muscle was calculated based on the cross-sectional area of PAN fibers (as opposed to the whole PAN-Gr muscle) since during contraction PAN fibers became taut while graphite fibers slackened, and therefore, most or all of the load was carried by PAN fibers in the muscle. Resistivity of the whole PAN-Gr muscle was measured to be about 7.8 × 10-4 Ω‚m and did not change with electrical activation. Discussion Artificial muscles made of PAN gel fibers contracted to half their initial length by chemical activation with a strong acid. This phenomenon is not new, but the mechanism of activation has yet to be fully determined. The mechanism is likely similar to that of other polyelectrolyte gels where change in size is due to switching between a hydrophobic and a hydrophilic structure. PAN gel fibers are unique in that most of the size change occurs in the axial direction. For PAN gel fibers exposed to low pH (i.e., contracted state), carboxylic acid groups would likely be protonated, rendering the polymer hydrophobic, thus expelling water and ions and possibly allowing hydrogen bonds between acid groups to form (see structure in Figure 1). When exposed to a high pH environment, carboxylic acid groups on the polymer would likely be in their salt form, encouraging a hydrophilic structure and allowing for swelling of the fibers, thus expanding the muscle. This explanation indicates that mass transfer of water and/or ions in and out of the polymer may determine the rate of contraction and elongation. The hysteresis as seen in Figure 3 indicates that a critical ion concentration is needed before the polymer changes length (H+ for contraction, OH- and Na+ for elongation). This suggests a couple of possibilities: (1) ions in solution must reach a critical concentration before displacing corresponding ions on the polymer, and/or (2) a critical fraction of groups on the polymer must be converted to the corresponding form (i.e., acid vs salt form) before the whole gel fiber changes
646
Biomacromolecules, Vol. 1, No. 4, 2000
length. Determining structures of contracted and elongated fibers would contribute significantly to understanding the mechanism of activation and the hysteresis. Mechanical testing of PAN gel fibers demonstrated that the contracted form is significantly stronger than the elongated form; i.e., the stress necessary to break contracted muscles was about three times that of elongated muscles. This suggests that the structure of the contracted polymer contains stronger primary, secondary, and/or tertiary structures than the elongated polymer. Further, the contracted form is much more elastic with behavior similar to that of rubber. The stress vs strain diagram for contracted PAN (Figure 4) was similar to that of rubber, and its modulus of elasticity (about 1 MPa) was in the range of that for synthetic rubber, which can vary from 1 to 75 MPa.41 For a comparison to human tissue, the abdominal aorta in the heart has a modulus of elasticity of roughly 0.1-0.2 MPa.42 Electrical activation of PAN artificial muscles was achieved through electrolysis of water. The conductivity of PAN gel fibers was increased by two means: depositing a coat of platinum on the fibers and intertwining PAN fibers with graphite fibers. This allowed for a PAN artificial muscle to serve as an electrode in an electrochemical cell. While serving as the anode hydrogen ions were generated near and/ or within the artificial muscle, reducing the pH in the vicinity of the muscle and causing the PAN gel fibers in the muscle to contract. Hydroxyl ions were produced near the muscle that led to swelling and expansion by reversing polarity of the cell’s electrodes. Electrical activation of PAN-Gr muscles was more reproducible than PAN-Pt muscles. PAN-Pt muscles prior to activation had low resistivities, about 1.2 × 10-4 Ω‚m, but after activation the resistivity increased to 24 Ω‚m or greater. This large decrease in conductivity was attributed to separation of the Pt metal from the PAN gel fibers during activation, thus explaining the poor reproducibility of PANPt muscles. As an alternative, PAN fibers were intertwined with conductive graphite fibers to produce a conductive PAN muscle with a resistivity of about 7.8 × 10-4 Ω‚m. This yielded a muscle that was electrically activated in the cell with consistent performance. Having graphite fibers in the muscle has the added benefit of increasing the strength of the muscle. A PAN-Gr muscle would likely fail under a higher stress due to the addition of graphite fibers than a pure PAN muscle. Muscles could then be designed around the mechanical properties of graphite as opposed to elongated PAN. As seen in Figures 5 and 6, the PAN muscle continued to contract for 1-2 min after switching from anode to cathode before the muscle began to elongate. Although the reason for this phenomenon was not examined, a possible explanation was that, initially after the change in polarity, hydrogen ions in solution rapidly migrated to the negatively charged cathode/muscle, significantly dropping the pH in the vicinity of the electrode before hydroxyl ion production overcame this effect. As with chemically activated PAN gel muscles that can contract to over half its initial length (Figure 3), both PANPt and PAN-Gr artificial muscles were able to contract to
Schreyer et al.
about 50% of their initial lengths (Figures 5 and 6). Although the addition of platinum or graphite to PAN gel muscles had little effect on magnitude of contraction, electrical contraction was noticeably slower than chemical contraction. The rates of contraction for both PAN-Gr and PAN-Pt artificial muscles in the electrochemical cells were about the same, 0.0052 and 0.0032 cm/s, respectively. This was much slower than chemical contraction with a strong acid, 1.4 cm/s. When compared to contraction rates of 90-180 cm/s for unloaded anterior tibial muscles of frogs (Rana pipiens), the artificial muscle rate of contraction was considerably slower.43 Mass transfer of water and ions in and out of the polymer likely limited rate of contraction for the PAN artificial muscles, which is often the case for polyelectrolyte gels.10-13 Decreasing the size of PAN artificial muscles would likely reduce mass transfer resistance and increase contraction rate. As a comparison, the lengths of tibial frog muscles used in ref 43 were between 5 and 10 mm, about 10-20 times shorter than PAN gel fibers used in this work. Another possible method to increase rate of contraction would be through electrochemical pH control using ion exchange membranes. Here an electrochemical cell and electrodes would be separated by an ion exchange membrane into half-cells, with the hopeful effect being to increase the rate of pH change near the artificial muscles. Other electrically activated artificial muscle materials have much quicker response times, such as polyelectrolyte gels that can change volumes of 100% or more in a few seconds and conductive polymers that can produce rapid bending and folding microstructures. Yet PAN gel fibers have two properties that distinguish PAN from other muscle materials: its ability to produce such a large change in length (few other materials are known to do so), and the fact that it is relatively a much stronger material. This warrants further research on developing a rapid, electrically activated PAN linear actuator or artificial muscle, with possible large and small scale applications where linear movement is required. Decreasing the size of PAN gel fibers and using ion exchange membranes to decrease rates of contraction, along with determining contracted and elongated gel fiber structures, are currently being investigated by the authors. Conclusions PAN gel fibers contract to half their initial length with chemical (i.e., pH) activation. Contraction of these gel fibers was likely due to hydrophobic properties of contracted PAN and the resulting expulsion of water from the polymer matrix. Artificial muscles made from PAN gel fibers were shown to have a modulus of elasticity similar and even greater than that of human tissue. The elastic properties and the unique ability of PAN gel fibers to significantly change length indicate potential use of this material for linear actuators and artificial muscles. To further improve this potential, electrical activation of artificial muscles made from PAN gel fibers was examined. By combining PAN gel fibers with either platinum or graphite fibers, the artificial muscles were able to serve as an electrode in an electrochemical cell. Through the electrolysis of water, the artificial muscles contracted over
Electrical Activation of PAN Artificial Muscles
45% of their initial length, thus indicating the feasibility of electrical activation. Although the rate of contraction was significantly slower than that of muscle tissue, decreasing water and ion mass transfer limitations would likely greatly enhance rate of contraction. References and Notes (1) Otero, T. F.; Angulo, E.; Rodriquez, J.; Santamaria, C. J. Electroanal. Chem. 1992, 341, 369-375. (2) Pei, Q.; Inganas, O. AdV. Mater. 1992, 4, 277-278. (3) Kaneto, K.; Kaneko, M.; Min, Y.; MacDiarmid, A. G. Synth. Met. 1995, 71, 2211-2212. (4) Smela, E.; Inganas, O.; Lundstrom, I. Science 1995, 268, 17351738. (5) Santa, A. D.; DeRossi, D.; Mazzoldi, A. Synth. Met. 1997, 90, 93100. (6) Smela, E.; Gadegaard, N. AdV. Mater. 1999, 11, 953-957. (7) Smela, E.; Kallenbach, M.; Holdenried, J. J. Microelectromech. Syst. 1999, 8, 373-383. (8) Osada, Y.; Hasebe, M. Chem. Lett. 1985, 1285-1288. (9) DeRossi, D. E.; Chiarelli, P.; Buzzigoli, G.; Domenici, C.; Lazzeri, L. Trans. Am. Soc. Artif. Intern. Organs 1986, 32, 157-163. (10) Grimshaw, P. E.; Nussbaum, J.; Grodzinsky, A.; Yarmush, M. J. Chem. Phys. 1990, 93, 6: 4462-4472. (11) Shiga, T.; Kurauchi, T. J. Appl. Polym. Sci. 1990, 39, 2305-2320. (12) Doi, M.; Matsumoto, M.; Hirose, Y. Macromolecules 1992, 25, 5504-5511. (13) Osada, Y.; Okuzaki, H.; Hori, H. Nature 1992, 355, 242-244. (14) Shahinpoor, M. Smart Mater. Struct. Int. J. 1992, 1, 91-94. (15) Herbert, I. R.; Tipping, A.; Bashir, Z. J. Polym. Sci., Part B: Polym. Phys. 1993, 31, 1459-1470. (16) Shahinpoor, M. Nonhomogeneous large deformation theory of ionic polymeric gels in electric and pH fields. Proceedings of the 1993 SPIE Conference on Smart Structures and Materials; Albuquerque, NM, Feb. 1993; SPIE: Bellingham, WA, 1993; No. 1916, pp 4050. (17) Shiga, T.; Hirose, Y.; Okada, A.; Kurauchi, T. J. Appl. Polym. Sci. 1993, 47, 113-119. (18) Hirai, T.; Nemoto, H.; Hirai, M.; Hayashi, S. J. Appl. Polym. Sci. 1994, 53, 79-84. (19) Hirai, M.; Hirai, T.; Sukumoda, A.; Nemoto, H.; Amemiya, Y.; Kobayashi, K.; Ueki, T. J. Chem. Soc., Faraday Trans. 1995, 91, 473-477. (20) Bohon, K.; Krause, S. J. Polym. Sci., Part B: Polym. Phys. 1998, 36, 1091-1094. (21) Rosen, O.; Piculell, L.; Hourdet, D. Langmuir 1998, 14, 777-782.
Biomacromolecules, Vol. 1, No. 4, 2000 647 (22) Umemoto, S.; Okui, N.; Sakai, T. In Polymer Gels; DeRossi, D., et al., Eds.; Plenum Press: New York, 1991; pp 257-270. (23) Vasiliu-Oprea, C.; Marcu, I.; Ciovica, S.; Dan, F. Polym.-Plast. Technol. Eng. 1999, 38, 609-620. (24) Xue, T.; McKinney, M.; Wilkie, C. Polym. Degrad. Stab. 1997, 58, 193-202. (25) Surianarayanan, M.; Vijayaraghavan, R.; Raghavan, K. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, 2503-2512. (26) Kikuma, J.; Warwick, T.; Zhang, H.; Tonner, B. J. Electron Spectrosc. Relat. Phenom. 1998, 94, 271-278. (27) Ogawa, H.; Saito, K. Carbon 1995, 33, 783-788. (28) Hu, X.; Johnson, D.; Tomka, J. J. Text. Inst. 1995, 86, 322-331. (29) Chatterjee, N.; Basu, S.; Palit, S.; Maiti, M. J. Polym. Sci., Part B: Polym. Phys. 1995, 33, 1705-1712. (30) Hu, X. J. Appl. Polym. Sci. 1996, 62, 1925-1932. (31) Rizzo, P.; Guerra, G.; Auriemma, F. Macromolecules 1996, 29, 1830-1832. (32) Bashir, Z. Acta Polym. 1996, 47, 125-129. (33) Pospisil, J.; Samoc, M.; Zieba, J. Eur. Polym. J. 1998, 34, 899904. (34) Sawai, D.; Yamane, A.; Kameda, T.; Kanamoto, T.; Ito, M.; Yamazaki, H.; Hisatani, K. Macromolecules 1999, 32, 5622-5630. (35) Dalton, S.; Heatley, F.; Budd, P. Polymer 1999, 40, 5531-5543. (36) Shahinpoor, M.; Salehpoor, K.; Mojarrad, M. Some experimental results on the dynamic performance of PAN muscles. Proceedings of the 1997 SPIE Conference on Smart Materials Technologies; San Diego, CA, Mar. 1997; SPIE: Bellingham, WA, 1997; No. 3040, pp 169-173. (37) Kuhn, W.; Hargitay, B.; Katchalsky, A.; Eisenberg, H. Nature 1950, 165, 514-516. (38) Hamlen, R.; Kent, C.; Shafer, S. Nature 1965, 206, 1149-1150. (39) Shahinpoor, M.; Adolf, D.; Segalman, D.; Witkowski, W. Electrically controlled polymeric gel actuators. U.S. Patent 5,250,167, Oct. 1993. (40) American Society for Testing and Materials. 1997 Annual Book of ASTM Standards; ASTM: West Conshohocken, PA, 1997; Vol. 03.01, E111-82, pp 221-226. (41) Van Vlack, L. Elements of Materials Science and Engineering, 5th ed.; Addison-Wesley: Reading, MA, 1987; pp 612-613. (42) Berne, R.; Levy, M. CardioVascular Physiology, 7th ed.; Mosby: St. Louis, MO, 1997. (43) Julian, F.; Sollins, M. Regulation of Force and Speed of Shortening in Muscle Contraction. The Mechanism of Muscle Contraction; 37th Cold Spring Harbor Symposium on Quantitative Biology, Cold Spring Harbor, NY, 1973; Cold Spring Harbor Laboratory: Cold Spring Harbor, 1973; pp 635-646.
BM005557L