Surface anisotropy improves artificial muscle response
Actuators make mechanical devices move. With all the variety of operating principles—electromagnetic, piezoelectric, hydraulic, and so on—current actuators have one thing in common: they are all packed in a hard casing. That may be an advantage in factory automation, but it is definitely a disadvantage when manipulating soft living tissue, which can be easily damaged.
The answer to such actuator-tissue interface problems may be electroactive polymers, called artificial muscles. Flexibility is one of their most attractive features. It also makes them suitable for mimicking the soft natural motion of living organisms, like the undulatory swimming robot created by our group.1,2 In addition, these materials hold promise for new types of active compliant mechanisms, such as a bistable structure that can switch its state without the need of external actuators.3
Compared with other types of electroactive polymers, artificial muscles made of ionic polymer-metal composites (IPMCs) have the benefit of being lightweight and powered by low voltage. IPMC actuators consist of an ion-conductive polymer membrane swelled with water and coated with thin metal electrodes on both sides. When voltage is applied to the electrodes, the actuator bends. Increasing the voltage increases the bending amplitude, but only until the onset of water electrolysis. Higher bending amplitude is always desirable, and one method of achieving it is to use very thin IPMC films, which tend to bend more.
If the film is too thin, however, it will not be sufficiently rigid when loaded in the plane perpendicular to the plane of bending, and will buckle and twist even under its own weight (see Figure 1). Rigidity in this direction is critically important for the payload capacity of our undulatory swimming robot (see Figure 2). We describe a way to maintain the rigidity and increase the bending amplitude of the actuator by creating a ‘preferred’ bending direction through anisotropic surface modification of the IPMC.4
We applied two different modifications to obtain an IPMC actuator with preferred bending direction. The first approach was to artificially introduce mechanical cracks in the metal electrode surface, similar to the cracks naturally formed by ‘working’ the IPMC. These naturally formed cracks run perpendicular to the bending plane. The second approach was to create directional roughness on the polymer surface prior to plating.
The IPMCs were manufactured by chemical plating.5 In this technique the ionic polymer membrane (Nafion in this study) is first roughened by sandblasting and cleaning and then impregnated with metal ions in a metal salt solution. A reducing agent is introduced, and the metal ions precipitate on the surface of the membrane to form the electrodes. This cycle is repeated five times to grow thicker electrodes. Owing to the rigidity requirement, we used the thickest commercially available Nafion membrane, with thickness of 250µm.
The cracks on the electrode surface were formed by tightly bending and rolling the IPMC after each plating cycle of the sandblasted Nafion membrane. The directional roughness was made by sandpaper on a non-sandblasted Nafion surface. We prepared four test specimens with alignment of cracks or roughness microgrooves as shown in Figure 3, and one with the basic sandblasted surface.
We compared the bending response of these specimens under water, because in air they tend to dry out and repeatability suffers. Underwater measurements were made with a non-contact laser displacement meter through the transparent container walls, and we had to adjust for refraction at the boundary. The results show that higher bending amplitude is facilitated by roughness microgrooves and artificially made cracks running in the direction in which cracks form naturally, i.e., across the length of the IPMC. The most effective surface treatment was roughening, which more than doubled the bending amplitude (see Figure 4). Even when the microgrooves and cracks ran perpendicular to the direction of natural cracks, the surface treatment improved the bending response of IPMCs compared with the standard process.
These results show that anisotropic surface modification is an effective method of creating a preferred bending direction and enhancing the bending response of IPMC actuators when driving voltage cannot be increased. The improvement in the bending response allows the use of thicker films to increase the load-carrying capacity of the IPMC actuators when loaded in the plane perpendicular to the plane of bending. Particularly, for our undulatory swimming robot, this would mean increased payload and faster locomotion. At present, we are still searching for an explanation for the bending amplitude increase, as it may be caused by a change in either the mechanical or electrical properties, or both. Preliminary results indicate that reduction of IPMC stiffness plays an important role in the direction of bending.
The authors would like to thank K. Asaka of AIST for helpful discussions,especially regarding the manufacturing process of gold-plated IPMCs.
Dr. Boyko Stoimenov is a research scientist at the Bio-Mimetic Control Research Center at RIKEN Institute, Japan. He works on artificial muscles and biomimetic robots, with a special focus on contact interfaces. He obtained his PhD from Tohoku University, Japan, and his MEng from the Higher School of Transport, Bulgaria. He has authored over 15 refereed papers in the fields of tribology and artificial muscles.
Dr. Toshiharu Mukai has been head of the Biologically Integrative Sensors Lab at RIKEN, Japan, since 2001. He received his BEng, MEng, and DrEng degrees from the University of Tokyo in 1990, 1992, and 1995, respectively. He was a research scientist at RIKEN, from 1995 to 2000. From 2000 to 2001 he was a CNRS postdoctoral fellow in France. His current research interests include sensor information processing, robotics, neural interfaces, and artificial muscles.
Dr. Jonathan Rossiter is currently a Royal Society Research Fellow at the University of Bristol. He has authored over 35 refereed papers on artificial intelligence, robotics, and sensors. He is also the author of three patents and of Fril++, the uncertain object–oriented logic programming language. From 2003 to 2005 he was a JSPS visiting researcher at RIKEN, Japan.