© 1999 American Chemical Society.
Volume 29, No. 12, 38-46.
Smart systems: Microphones, fish farming, and beyond
Smart materials, acting as both sensors and actuators, can mimic biological behavior.
The term “smart systems” appears with increasing frequency in scientific and engineering publications as well as in the popular press. Several smart systems have been commercialized and are widely used. Many more systems are in various stages of development. Aerospace engineers are interested in smart air foils to control drag and turbulence. Diabetics need medical systems to sense sugar levels and deliver insulin. Architects are designing smart buildings with self-adjusting windows that control how much sunlight and heat flows in and out. Tennis players will want smart racquets to make overhead smashes and delicate drop shots. Smart motion detectors will monitor authorized and unauthorized entries into buildings. Smart toilets are capable of analyzing urine to identify health problems. Smart irrigation systems will be needed to optimize the world's food supply. Smart transducers can “talk” to fish, a leading source of protein in many parts of the world, to herd them like cattle.
How do smart systems accomplish all this? They are made from “smart” materials, which have the ability to perform sensing and actuating functions and therefore are capable of imitating living systems. Four of the most widely used classes of smart materials are piezoelectrics, electrostrictors, magnetostrictors, and shape-memory alloys—all of which are ferroic. Ferroic materials have active domain walls, that is, crystallographic boundaries that can be moved by applying an external force or field. The resulting changes in the shapes of these materials are large enough to make them useful as actuators. Smart materials undergo two or more phase transformations, which can be used to tune their performance. Smart materials also can be manufactured by using multifunctional composites. These composites are made by combining two materials in a connectivity pattern that optimizes the functionality of the composite.
A sensor receives a stimulus and responds with a signal; an actuator produces a useful motion or action. By definition, smart materials are both sensors and actuators, because they perform both functions. They may or may not have control systems. “Passively smart” materials function like a vertebrate's spinal cord, producing involuntary reflex responses, without thought or signal processing. “Actively smart” systems analyze the sensed signal—perhaps for its frequency components—and then decide what kind of response to make.
The integration of sensors, actuators, and control systems is an ongoing process in the automotive, medical, military, and consumer electronic markets as systems become more miniaturized and more complex. By incorporating sensors, actuators, and evolvable chips into these systems, we can begin to emulate biological behavior. Small electronic and optical subsystems are capable of evolving into more complex systems that have the ability to sense and respond to changes in their surroundings.
Meanwhile, biochemists are making major advances in understanding how the human brain functions and what it means to be alive. In the coming century, there will be a confluence of organic and inorganic life into some kind of a composite life form—perhaps an immortal life form—with a common consciousness that transcends individual beings. The Internet and the World Wide Web seem to be leading us in that direction. The evolution of machine technology into global intelligence is the theme of a major book that was published in 1997 (1).
Sensors + actuators = smart materials
A piezoelectric ceramic videocassette tape head positioner illustrates the basic idea of a smart material (Figure 1). Piezoelectric materials have the unusual ability to generate a voltage when a mechanical force is applied or to generate mechanical force when a voltage is applied. Bimorph piezoelectric ceramics are usually made from lead zirconate titanate (PZT). Bimorph cells are constructed from pairs of piezoelectric plates. When a voltage is applied, one plate expands, the other plate contracts, and the cell bends in proportion to the voltage. The videocassette tape head positioner has large actuator electrodes that move the material and smaller electrodes that sense the position and orientation of the tape head. This combination of sensing and actuating mimics two of the functions of a living system: awareness of the surroundings and a useful response, which usually is in the form of motion.
The piezoelectric Pachinko game illustrates the principle of a smart material. Pachinko parlors with hundreds of vertical pinball machines are popular in Japan. The machine, constructed by engineers at Nippon Denso (Kiriya, Japan), is made from multilayer PZT stacks that act as both sensors and actuators (Figure 2). When a ball falls on the stack, the force of impact generates a piezoelectric voltage. Acting through a feedback system, the voltage pulse triggers a response from the actuator stack. The stack expands rapidly, throwing the ball out of the hole. The ball moves up a spiral ramp during a sequence of such events. Eventually, it falls into a hole and begins the spiral climb again.
Very smart materials can learn
During the past 20 years, in addition to smart materials, we have been developing a family of “very smart materials”—we are not yet willing to call them intelligent—that have a learning or tuning function that makes it possible for them to become smarter. These materials take advantage of nonlinear properties such as electrostriction or higher order elastic constants. By applying a bias field, we can tune the level of sensitivity, that is, the degree of “smartness”.
Looking ahead to the use of thin-film microelectromechanical systems (MEMSs) (2), more intelligent materials are coming on the scene that integrate the control system with the sensors and actuators, all in one common piece of material. MEMS-based devices that have found applications in everyday use include inkjet cartridges in printer heads, accelerometers as airbag sensors in automobiles, and tire-pressure sensors. Chemical sensor arrays and high-resolution displays are two emerging MEMS products. Figure 3 shows a digital mirror display (DMD) array developed by Texas Instruments that was fabricated by micromachining. It consists of 800 x 600 tiltable mirrors that function as pixels; the associated circuitry is batch processed on a silicon chip. The mirrors are electrostatically actuated. The DMD array is a promising technology for lightweight displays that offer high definition, high contrast, and high brightness.
We believe that when sensors, actuators, and control systems can be combined like the eyes, arms, and brain of the human body, the system will deserve the name “intelligent”. Finally—although we have no idea how to do this yet—we will make “wise” materials that make moral decisions. Perhaps they will evolve in some way like living systems (see box, Forecasts for the future).
One application: smart shock absorbers
Current research is focused on vibration suppression in automobiles using smart shock absorbers. Inside the smart shock absorber developed by Toyota (Figure 4) is a multilayer piezoelectric ceramic that has about five layers for sensing road vibrations. The multilayer stacks positioned near each wheel of the automobile have about 100 layers, all part of the same PZT ceramic, which act as the actuator. After analyzing the vibration signals, a voltage is fed back to the actuator stack, which responds by pushing on the hydraulic system of the automobile to enlarge the motion. In this way, signal processors in the automobile can analyze acceleration signals from road bumps and respond with a motion that cancels the vibration. Such active systems are used to alleviate excessive vibrations in helicopter blades and in the twin tails of F-18 fighter jets.
Traditionally, vibration problems in engineering structures were solved by using mechanical dampers or by isolating vibration with foam rubber pads, but the piezo-dampers developed by Active Control eXperts (ACX; Cambridge, MA) are readily tuned to the frequencies that most require attenuation. Passive piezo-dampers that use either a resistance–capacitance shunt circuit or a resistance–inductance shunt circuit are used as electronic shock absorbers in skis and baseball bats. A piezoelectric ceramic converts the unwanted mechanical vibrations in the skis or bats to voltage, which is dissipated as heat in the shunt circuit
Underlying structure–property relationships
Four major families of ceramic and metallic actuators are under development: piezoelectrics, electrostrictors, magnetostrictors, and shape-memory alloys. All of these materials undergo at least two phase transformations with coupled thermodynamic order parameters. These transformations lead to complex domain wall behaviors, which strongly contribute to the sensing and actuating capabilities. Domain wall motion is initiated by electric fields (ferroelectrics), magnetic fields (ferromagnetics), mechanical stress (ferroelastics), or temperature changes as they transform from ferroic to nonferroic states.
In the following sections, we review the atomistic origins of piezoelectricity, electrostrictive, magnetostrictive, and thermostrictive properties.
Piezoelectrics: Smart ceramics. Most piezoelectric transducer formulations are based on PZT, Pb(Zr,Ti)O3, one of several ferroelectric substances that crystallize with the perovskite structure, described as follows: Lead atoms are located at the corners of a cubic unit cell and oxygen atoms at the centers of the cube faces. Lead and oxygen ions have radii of ~1.4 Å, and together make up a face-centered cubic array (unit cell) ~4 Å on one side. Octahedrally coordinated titanium or zirconium ions are located at the center of the unit cell. Ferroelectric materials exhibit spontaneous electric polarization; that is, the electric dipole moments within a domain align in a common direction, even in the absence of an external electrical field.
On cooling from a high temperature, the crystal structure of PZT undergoes a displacive phase transformation with atomic displacements of ~0.1 Å. For titanium-rich compositions, the point symmetry changes from cubic m3m to tetragonal 4mm on cooling through the Curie temperature near 350 °C. (Note: The Curie temperature marks the transition from the high-temperature paraelectric phase, in which the electrical dipoles only align when a field is applied, to the polar low-temperature phase, in which the dipoles align spontaneously.) The tetragonal state with its spontaneous polarization along the [001] crystallographic direction (i.e., perpendicular to the square face) persists down to temperatures approaching 0 K. Structural changes are illustrated in Figure 5.
To use these piezoelectric ceramics with their large polarizations, compositions near a second phase transition are chosen. At the Curie point, PZT converts from a paraelectric state with the ideal cubic perovskite structure to a ferroelectric phase located near a morphotropic phase boundary between the tetragonal and rhombohedral states. The morphotropic phase boundary delineates two solid phases that remain in a near-equilibrium state over a very wide temperature range. Very large piezoelectric coupling between electric and mechanical variables is obtained near this phase boundary. Much of the current research in this field involves looking for other morphotropic phase boundaries to further enhance the electromechanical coupling factors.
Two piezoelectric effects are used in PZT transducers: direct and converse. The direct effect relates polarization to stress and is used in sensors. The converse effect relates strain to electric field and is used in actuators.
Piezoelectric ceramics are prepared by “poling” them; that is, subjecting them to a large dc electrical field to align their domains. For a poled ceramic with symmetry
m (an infinite number of mirror planes parallel to the poling direction), the appropriate tensor coefficients are d31, d33, and d15. These piezoelectric coefficients make intrinsic and extrinsic contributions. Under mechanical stress parallel to the dipole moment, the spontaneous polarization (Ps) is enhanced along the poling direction (X3); and when stress is applied perpendicular to that dipole moment, electric charges develop along X3. These are the d33 and d31 effects, respectively. When the dipole is tilted by shear stress, charges appear on the side faces (the d15 coefficient).
There are extrinsic contributions to the piezoelectric coefficient as well, and these can be extremely large, often involving the domain wall motions.
Ferroelectric ceramics such as PZT do not become piezoelectric until electrically poled. Poling is carried out under intense electric fields at elevated temperatures below the ferroelectric Curie point at which the domains are easily aligned. Titanium-rich compositions in the PZT system favor a tetragonal modification, with sizable elongation along [001] and a large spontaneous polarization in the same direction. Six equivalent polar axes in the tetragonal phase correspond to the [100] direction (the unit cell edge that forms one side of a square) and directions of the cubic paraelectric state. A rhombohedral ferroelectric state is favored for zirconium-rich compositions. The distortion and polarization are along <111> (body diagonal) directions, giving rise to eight possible domain states.
The compositions that pole best lie near the morphotropic boundary between the rhombohedral and tetragonal ferroelectric phases. For these compositions, 14 possible poling directions exist over a wide temperature range, explaining in part why the ceramic piezoelectric coefficients are largest near the morphotropic boundary. Phase changes between the rhombohedral and tetragonal phases also occur during the poling process.
Electrostrictors: Very smart ceramics. Piezoelectricity is a third-rank tensor that relates strain and electric field. Electrostriction is a fourth-rank tensor that relates strain to the square of the electric field. Above the Curie temperature, the perovskite structure is cubic (centrosymmetric, or nonpolar), and the electrostriction effect is more important than the piezoelectric effect because third-rank tensors disappear in centrosymmetric media. It leads to what we call very smart ceramics.
In a smart ceramic, the direct piezoelectric effect is used for sensing, followed by feedback through the converse piezoelectric effect. In a very smart material, we monitor the change in capacitance of the material, then feed back with both dc and ac fields: first to tune the magnitude of the electromechanical coupling coefficient, then to drive it. For higher order coupling coefficients such as those that describe electrostriction, three coupled effects arise (rather than two): change in the dielectric constant with stress, field dependence of the piezoelectric voltage coefficient, and electrically driven mechanical strain. The electrostrictive ceramic becomes a tunable transducer.
We began work on several of these electrostrictive materials almost 20 years ago for active optic systems. Much of this work was done with the Itek Corp. in Lexington, MA. During the Cold War, many satellites flying over the Soviet Union used active optic systems to eliminate the effects of atmospheric turbulence. Electrostrictive materials have an advantage over piezoelectrics in adjusting the position of optical components, because much less hysteresis is associated with the motion.
Work on active optic systems has continued over the years. Similar multilayer actuators were used to correct the positioning of the optical elements in the Hubble telescope. Supermarket scanners use actuators and flexible mirrors to interrogate bar codes optically.
Relaxor ferroelectrics. Disordered perovskites contain regions where “active” ions (those that promote ferroelectricity) are in close proximity. Ordered perovskites usually have low dielectric constants because active and inactive ions are evenly dispersed and the linkage between “active” ions is severed. In partially disordered structures such as the relaxor ferroelectrics, the dielectric constant can be extremely large, making disordered materials useful as capacitor dielectrics and as electrostrictive actuators. The most widely used compositions are modifications of lead magnesium niobate (PMN), Pb3MgNb2O9.
Relaxor ferroelectrics, often perovskite materials, are characterized by temperature-sensitive microdomains that result from the many different “active” ion linkages in the disordered octahedral framework. Each NbO6 octahedron may be bonded to zero to six other NbO6 octahedra (with the remaining connections involving MgO6 octahedra). Connections between these octahedra are assumed to be essential to ferroelectricity and high anisotropy coefficients (K values). As the temperature decreases from the high-temperature paraelectric state, ferroelectric microdomains gradually coalesce to macrodomains, giving rise to a diffuse phase transformation. These polarization fluctuations are also dependent on bias field and the frequency used to measure the dielectric or piezoelectric constant. The dielectric constant drops off rapidly with increasing frequency (hence the name “relaxor”) because it takes time for the polarization fluctuations to respond. The dc bias fields favor coalescence, giving the same effect as lowering the temperature.
Relaxor behavior is very common among lead-based perovskites, suggesting that the “lone pair” electrons of Pb2+ play a role in the microdomain process, possibly by adjusting their orientations. Electrostriction is described by a 6 x 6 matrix that relates strain to the square of the electric polarization. This kind of matrix is familiar to most materials scientists because electrostriction is a fourth-rank tensor almost identical to elasticity in form. For a cubic crystal, we deal with the same coefficients, 11, 12, and 44 that would normally be used to describe the elastic properties of a cubic crystal. In this case, strain is induced electrically rather than mechanically.
Whereas piezoelectricity is observed in polar materials, electrostrictive transducers use cubic (nonpolar) materials, whose compositions are near a phase instability with microscopic regions that fluctuate in polarization. On average, atoms are located in the ideal cubic sites but continually shift off these positions. The underlying origin of these effects is a partial ordering of the PMN perovskite structure in which the niobium and the magnesium atoms of PMN alternate in position but over only a few unit cells (usually 30–50 Å). Within these ordered islands, an external field acts upon fluctuating dipoles to make large electrostrictive motions.
Magnetostrictive actuators. PZT and PMN ceramics are outstanding ferroelectric actuators. However, equally interesting developments are taking place in the field of ferroelastic and ferromagnetic materials. All these ferroic materials have a domain structure in which the walls can be moved with electric fields, magnetic fields, or mechanical stresses.
Magnetostrictive alloys (e.g., Terfenol-D, Tb1–xDyxFe2) function well as sensors and actuators. (The name Terfenol refers to terbium and iron, and the Naval Ordnance Lab, NOL, where it was developed. The “D” refers to the dysprosium-containing variety.) Magnetostrictive materials exhibit strain (i.e., a change in dimension) in proportion to the direction and extent of magnetization. High-power actuators can deliver forces >50 MPa with strains up to 0.6%, whereas magnetostrictive sensor materials can provide hundreds of times the sensitivity of semiconductor strain gages. Magnetoelastic materials (in which elastic strain alters the magnetization) also have tunable elastic moduli that can be controlled by external magnetic fields.
Many magnetomechanical transducers and actuators have been designed and manufactured with Terfenol-D. The high energy density, ruggedness, and reliability of these actuators make them attractive for vibration suppression and high-power sonar. Thin films of magnetostrictive rare earth-iron alloys can be sputtered onto silicon and patterned by etching or sputtering through masks. Micropump and microvalve membranes and cantilevers appear to be promising MEMS components.
The rare earth atoms in Terfenol have large orbital moments that interact with fields to give large magnetostrictive strains. The rotation of magnetization is largely responsible for the shape change. The field-induced strain in Terfenol-D is ~100 times larger than strains in iron and nickel.
The iron in Terfenol produces the high Curie temperature. The rare earth terbium and dysprosium atoms produce the large magnetostriction. In combination, these three elements produce the useful alloys. The phase diagram of Terfenol is the magnetic equivalent to the morphotropic boundary of PZT. A portion of the magnetic phase diagram of Terfenol is plotted in Figure 6. Terfenol is cubic and paramagnetic at high temperature, and it undergoes a magnetic phase transformation to a rhombohedral structure with magnetic spins parallel to the <111> family of crystallographic directions. Near room temperature, it is poised on an instability with the spins ready to reorient into the tetragonal directions, the former <100> directions of the cube. There is a complex domain structure both above and below the transition, and like PZT, Terfenol is poised on a rhombohedral–tetragonal phase boundary.
The figure of merit for magnetostrictive actuators is proportional to the saturation strain coefficient. But in addition to a large shape change, the strain must be easy to move. The figure of merit for magnetostrictive actuators is the ratio of the saturation strain (
) to the anisotropy coefficient (K; an indicator of the ease of rotation of the magnetization). In addition to producing a large shape change, the strain must be easy to move. TbFe2 has a very large and a large positive K, which reduces its figure of merit. DyFe2 has a K of opposite sign. K greater than zero indicates the spins preferentially align along <111>, and K less than zero means <100> is preferred. By tuning the composition of the ternary alloy to near the point at which K goes to zero, one can make an easily movable strain in this magnetostrictive alloy, maximizing the figure of merit.
Thus, it is important to alloy TbFe2 with DyFe2. A large shape change is very useful in actuators and transducers, but controlling the shape change with small applied fields is also important. In contrast, a large shape change frozen in position is of no practical value. To lower the driving field, a second phase change is positioned near room temperature by adjusting the alloy composition. It can be done by lowering the Curie temperature, but this method demagnetizes the actuator and greatly reduces the magnetostriction coefficient. It is preferable to choose a composition near the rhombohedral–tetragonal phase boundary where the easy axis, the preferred direction of the magnetic dipoles, switches from <111> to <100>. Compositions near Dy0.7Tb0.3Fe2 have large magnetostrictive coefficients with easily controlled shape changes. Below room temperature, the magnetic symmetry changes from rhombohedral to tetragonal with a significant decrease in the magnetostrictive shape change.
Shape-memory metals. The final example of an actuator material is the shape-memory alloys, which are thermally driven in contrast to the magnetic drive of mag
The shape-memory alloys undergo martensite-type phase transformations similar to those observed in steel processing. Two characteristics of martensitic phase changes are the absence of long-range diffusion and the appearance of a shape change. Ferroelastic phase transformations are distortive and diffusionless and have much in common with martensitic transformations. Ferroelastic crystals exhibit mechanical hysteresis between the stress and strain caused by stress-induced movement of domain walls. Martensites also are internally twinned, but mech
Typically, these materials are partially ordered as they undergo transition from a body-centered cubic structure to a partially ordered CsCl structure (Figure 7). The shape-memory effect takes place at a martensitic transformation from the CsCl-like structure into a distorted multidomain martensite phase. Under stress, the martensite deforms easily; when reheated, it returns to the original morphology of the high-temperature structure.
Single-phase smart materials
Most of the best actuators are primary ferroics. These ferroelastic, ferromagnetic, or ferroelectric solids are poised on an instability, often with two or more phase changes involved. PZT is cubic at high temperatures and is poised on a tetragonal–rhombohedral phase boundary. Partially ordered PMN is poised on a cubic-rhombohedral transition. The shape-memory alloys also are partially ordered and are poised on a martensitic phase transformation. Magnetostrictive Tb–Dy–Fe alloys are cubic at high temperatures and are operated at a rhombohedral–tetragonal spin reorientation.
In addition to the four materials that we discuss here, other kinds of actuators are under development. Field-induced phase transitions in modified lead zirconate ceramics involve transitions among paraelectric, antiferroelectric, and ferroelectric phases. Two phase transitions also are involved in photostrictive materials (in which electrical and mechanical characteristics change under illumination) and in chemostrictive materials such as human muscle (which consists of partially hydrated polymeric systems in which phase transitions take place in the polymer and in the surrounding sheath of water molecules during actuation).
Functional composites
Composite materials, another approach to actuation and sensing, are closer to our own interests. So far, we have described materials that are single-phased—at least at high temperatures. Another approach to making smart materials is to bring together two or more materials, each of which has an associated phase transition. For example, in our transducer program, we combine polymeric materials (which have phase transitions in which the elastic properties undergo large changes) and ferroelectric materials (in which the dielectric properties have an associated instability). Because the two materials have different kinds of instability, we can build up structures that are especially good for sensing and actuating (4).
In working with this family of functional composites, we do not have to optimize all of the tensor coefficients—only those that appear in the figure of merit. Thus, we have built different connectivity patterns (Figure 8) into these materials by using electrically soft materials (ferroelectrics) that have high dielectric constants and electrically hard materials (polymers) that have very low dielectric constants. Although the polymers are electrically hard, they are mechanically soft; compliance coefficients are several orders of magnitude larger than those of ceramics. Using the connectivity patterns in Figure 8, we build up parallel and series connections that optimize particular combinations of tensor coefficients.
Composite materials are used in many structural applications, but their use in the electronics industry has been limited. As the advantages and disadvantages of composite sensors and actuators become better understood, we expect this picture to change. Composite electromechanical transducers offer many advantages over single-phase transducers (see box, Composite electromechanical vs single-phase transducers).
Better microphones can be made by reducing acoustic impedance. Acoustic impedance can be reduced by partially replacing ceramics with a soft polymer that better couples the transducer vibrations to water and to human tissue (i.e., eardrums). Inserting electrodes inside the transducers lowers the drive voltage, improves sensitivity to hydrostatic waves, and enlarges displacements. Composite transducers illustrate a very general approach that applies to not only piezoelectric materials but also many other functional composites.
Composite piezoelectric transducers (Figure 9) have been manufactured and tested in our laboratory during the past two decades. These functional composites incorporate several underlying ideas, including
- connectivity patterns that lead to field and force concentration;
- use of periodicity and scale in resonant structures;
- symmetry of a composite structure and its influence on physical properties;
- polychromatic percolation and coupled conduction paths;
- varistor action and other interfacial effects;
- sum, combination, and product properties;
- coupled phase transformation phenomena; and
- the important role that porosity and inner spaces composites play in many functional composite materials.
Up to now, we have discussed primarily piezoelectric transducers with electromechanical sensing and actuating functions. However, the idea of smart materials is much more general (Figure 10). There are many types of sensors, actuators, and feedback circuits.
Many such sensors and actuators can be fabricated in the form of a multilayer ceramic package. Multilayer packages originally consisted of low-permittivity dielectric layers, interconnected through metallized structures called “via holes”, with metal circuitry printed on each layer. Buried capacitors and resistors have been added to the three-dimensional packages. Smart sensors, adaptive actuators, and display panels—with thermistors and varistors to guard against current and voltage overloads—continue to be developed.
Putting instability to work
Two transformations are involved in most of these smart materials. Because these materials often are primary ferroics (ferroelastic, ferroelectric, or ferromagnetic), they have domain wall motions that assist in the sensing and actuating processes. The ferroics are operated near an instability to make these domain walls—and their associated dipoles and strains—movable. We identify three kinds of commonly used actuator materials (and others appear to be possible):
- In the first kind, as in PZT or Terfenol, Curie temperature is high, and the actuator is operated near an orientational change of the electric or the magnetic dipole moment.
- The second kind involves a partially ordered phase, as in electrostrictive PMN or the shape-memory alloys. These materials are operated near a diffuse phase transition with two coexisting phases: the high-temperature or austenite-like phase and the low-temperature or martensitic-type phase.
- The third kind involves composite materials with coupled phase transformations. For example, piezoelectric ceramic fibers embedded in elastomers can be used to make biomedical transducers. The fibers produce sound waves, and the elastomer adjusts the acoustic impedance. This reduces sound reflection at the interface between the transducer and the biological tissue, allowing more efficient penetration of the acoustic waves into the tissue.
The underlying reasons for the material choices are fairly obvious. Why use primary ferroics? Because small external fields or forces, whether they are strains, electric dipoles, or magnetic dipoles, can be used to produce large responses. Why use a cubic prototype phase? Because the symmetry of cubic phases produces many equivalent orientation states, making it unnecessary to grow single crystals and allowing us to use polycrystalline materials instead. Why is partial ordering advantageous? Because it provides many nucleation sites for generating a diffuse phase transformation. Why operate materials near a morphotropic transition? Because the instability ensures persistent disequilibrium over a wide range of temperatures.
By imitating sensor, actuator, and analyzer mechanisms already used by living organisms, it may be possible to build systems and devices that mimic the functions of life itself. If motion sensors, audio receivers, and microphones are artificial eyes, ears, and voices, can artificial minds be far behind?
References
(1) Dyson, G. Darwin Among the Machines; Perseus Books: Reading, MA, 1997.
(2) National Research Council. Microelectromechanical Systems; National Academy Press: Washington, DC, 1997.
(3) Uchino, K. Piezoelectric Actuators and Ultrasonic Motors; Kluwer Academic Publishers: Boston, MA, 1997.
(4) Newnham, R. E. Mater. Res. Soc. Bull. 1997, 22, 20-34.
(5) Newnham, R. E. Bull. Am. Ceram. Soc. 1996, 75 (10), 51-59.
