No.10 - June 2006
APPLICATION OF THE MONTH: Carbon nanotube actuators
Carbon nanotube-based electrically driven actuators convert the electrical energy applied by an external source into a mechanical deformation of the nanotubes. In a similar manner, optically driven actuators convert infrared radiation at the nanotube level into mechanical deformation.
Current alternatives
A number of other materials exhibit values suited for actuation in areas ranging from robotics and micro-pumps to artificial muscles. Such conventional solutions include piezoelectric devices, shape memory alloys, electroactive polymers, or magnetostrictive materials. Their performances are expressed in terms of strain percentage, stress generated and the strain rate. Other performance issues are the motion repeatability and the actuator's endurance.
Typical piezoelectric materials include quartz (SiO2), lead zirconate titanate (PZT), lithium niobate, and some polymers such as polyvinyledene fluoride (PVDF). They offer a fast response to changes in applied voltage and a good repeatability. To summarize, piezoelectric actuators exhibit high peak stress (35 MPa), low strain (1%), very low power density (0.1 W/Kg) at very high speeds. They are quiet but require high voltages.
Shape memory alloys, such as Nitinol (NiTi) and CuZnAl, respond to heating by changing their shape. Nevertheless, in order to perform their reverse transformation, shape memory alloys require cooling and overall their performance is strongly influenced by the surrounding temperature conditions. Shape memory alloys offer very high peak stress (200 MPa), low strain (1-8%), very high power density (100kW/kg). On the other hand, their efficiency is rather poor (<5%), they are very difficult to control and much slower than other counterparts are.
Electroactive polymers (whether electronic or ionic) exhibit higher response speeds, lower densities and improved resilience when compared to shape memory alloys. Their major disadvantages are low actuation forces, mechanical energy density and lack of robustness. In terms of performance, such actuators offer high stress (5-34 MPa) and high stiffness, at low driving voltages (~2V). Despite high work density exhibited (100kJ/m3;), strain usually does not reach 2%. Electronic electroactive polymers can hold strains under DC activation, inducing relatively large actuation forces, but require high voltages (~150 MV/m). Ionic electroactive polymers usually provide bending actuation with a large displacement, but featuring low actuation force.
| |
Piezoelectric materials |
Electroactive polymers |
Shape memory alloys |
| Displacement |
< 1.7% |
> 10% |
< 8% |
| Force (MPa) |
up to 130 |
0.1 - 3 |
about 700 |
| Speed |
µsec to ms |
µs to s |
s to min |
| Fracture toughness |
resilient, plastic, elastic |
resilient, elastic |
elastic |
Combinations of rare earth elements with iron, in materials such asTbFe (Terfenol) and TbDyFe (Terfenol-D), are most commonly employed as magnetostrictive materials. In their case, the actuation mechanism is activated by an external magnetic field. The current generation of magnetostrictive materials is only capable of small displacements (0.1 to 0.5 percent), but offers a very high work output per unit mass.
As a conclusion, problems associated with the range of their operational temperature, voltage, cycle life or the work density/cycle stimulate the exploration of alternatives like carbon nanotube actuators.
Nanotube electrically-driven actuators
Although actuators made of individual carbon nanotubes have been demonstrated, usually carbon nanotube composite materials or bucky-paper are used as macroscale actuators in simple electrolytes, such as sodium chloride solution (NaCl).
The typical set-up consists of nanotube bucky-paper used as the electrodes of an electrochemical cell. Applying a voltage to the electrodes injects electronic charge into the nanotubes and results in the formation of a double layer at the electrolyte-ion interface. The system is then studied through classical electrochemistry techniques like cyclic voltametry. As the potential between the electrodes varies, the strain in the in-plane direction of the bucky-paper can be observed for both the anode and the cathode. Both electrodes bend laterally but the strain is higher for the cathode. The strain has a parabolic dependence on the potential. Unlike the case of polymer-based materials, no dopant intercalation is required for the operation of carbon nanotube actuators.
The first results using single-walled "buckypaper" for actuation were published in 1999 by Ray Baughman and co-workers and generated significant interest as the stress generated (0.75MPa) was higher than the peak capacity of human skeletal muscle (0.3MPa). The observed strain was 0.2% and the experiment could be repeated over 140,000 cycles with a 33% decrease in the actuator stroke. More recently, stress values of up to 26MPa and actuation rates of 20% s-1 have been obtained with an electrochemical set-up using carbon nanotube fibers. This remarkable result represents two orders of magnitude more than natural muscles.
The strain generated by carbon nanotube actuators remains below 2%. Although this value seems small especially compared to that of skeletal muscle (10%-20%), it compares well with piezo-electrical materials and it is obtained at voltages of only a few volts. For reference the maximum strain of ferroelectric ceramics is 0.1%.
Nanotube optically-driven actuators
Intuitively, the mechanism for optically driven actuation should be related to the IR band gap of the nanotubes and the generation of electron-hole pairs. A study of optically driven actuation effects had been performed by Zhang and Iijima in 1999 prior to the two experiments mentioned above. The authors suggested that, under IR light exposure, electron-hole pairs are generated and the different nature of the tubes would create local imbalance in the charge density. This in turn would result in electrostatic fields that should stretch the nanotube sheet. This hypothesis was supported by the measurement of a light-induced electrical current during this earlier study. The same type of current was measured by Lu and Panchapakesan and encouragingly for this hypothesis, the photo-conductivity responses were on similar time-scales with the actuator strain responses indicating the same origin.
Two experimental studies seem representative for the potential of carbon nanotubes employed as optically driven actuators.
In the first study, published in 2005, Lu and Panchapakesan reported an optical actuator made out of single-walled nanotube sheets mechanically bonded to an acrylic elastomer surface. The 30-40 µm thick nanotube sheet was obtained by vacuum filtration after dispersion into isopropyl alcohol and the thin acrylic elastomer was of similar thickness. For the measurements, a cantilever structure was fabricated by attaching the actuator perpendicularly to a 100 µm thick PVC film. The study included the extensive characterization of the deflection of the cantilever during exposure to lasers of different intensities and wavelengths.
The strain observed was between 0.01%-0.3% at light intensities of 5-120 mW/cm2 with a clear increase in the strain with the intensity of the light. Almost the same strain amplitude has been obtained when measurements were repeated over a small number of cycles. The 0.3% strain corresponded to a stress of 0.9 MPa, higher than the peak capacity of human skeletal muscle of 0.3 MPa. The strain also increased with the wavelength of the laser source, almost doubling when the energy or incident photons increased from 0.8 to 1.94 eV.
The second relevant study was published in the same year by Ahir and Terentjev from Cavendish Laboratory. This time the actuator consisted of multi-walled carbon nanotubes dispersed in a polymer (polydimethylsiloxane - PDMS). The resulting composite was clamped in a frame and was subject to stress measurements under the action of IR light from a ~675 nm source. The measurements were performed for composites with nanotube loadings ranging from 0.02% to 7% and different pre-strain values (between 2% and 40%). The stress obtained was of the order of tens of kilopascals, which was inferred by the authors to correspond to a strain of 2-4%.
Areas of application
The favorable comparison between the initial results and properties of skeletal muscles suggested a first application as artificial muscles. However, other comparisons are at least as interesting. Ion transport through nanotube actuators is much faster than through bulk polymer films. For instance, Nafion-carbon nanotube composite materials exhibit superior performances as actuators when compared to metal-doped Nafion films, due to the electrical conductivity and high aspect ratio of carbon nanotubes. As mentioned before, carbon nanotube electromechanical actuators also generate higher strains than high-modulus ferroelectrics.
The current status of the field opens doors for envisioning applications in areas including microsurgical devices, artificial limbs, artificial ocular muscles, prosthetic devices, as well as for other micromechanical and microfluidic devices. Carbon nanotube actuators could also be used as nano-tweezers for manipulating submicron clusters.
