Ph 101 Frequency (Hz) Undamaged 1/4 loss for k1 10−1 10−2 102 10−3 10 102 101 −0.5 −1 −1.5 −2 −2.5 −3 −3.5 100 −0 Phase (rad) −6 −8 101 Frequency (Hz) (f) 100 Undamaged 1/4 loss for k1 10−5 Phase (rad) 100 −1 −2 −3 −4 −5 −6 −7 100 101 Frequency (Hz) 101 Frequency (Hz) 102 Undama 1/4 loss 101 Frequency (Hz) 100 Undam 1/4 loss 10−1 10−2 10−3 10 102 Undamaged 1/4 loss for k1 101 Frequency (Hz) 101 Magnitude −4 (e) Magnitude Undamaged 1/4 loss for k1 Phase (rad) Phase (rad) −2 Undama 1/4 loss 100 Frequency (Hz) −10 100 101 Frequency (Hz) 10−5 100 −5 −6 −7 100 (d) 101 (c) 100 Magnitude 102 Magnitude Pha −2 −2.5 −3 −3.5 10 −0.5 −1 −1.5 −2 −2.5 −3 −3.5 100 101 Frequency (Hz) Undam 1/4 loss 101 Frequency (Hz) Figure 29 DTF response: (a) x1 /xg ; (b) x2 /xg ; (c) x3 /xg ; (d) x2 /x1 ; (e) x3 /x1 ; (f ) x3 /x2 ă ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ −4 100 101 Frequency (Hz) 102 (d) 10−5 101 Frequency (Hz) 100 −8 102 10−5 −1 −2 −3 −4 −5 −6 −7 100 Frequency (Hz) (f) Undam 1/4 loss 101 101 100 102 Undam 1/4 loss 10−1 10−2 10−3 10 102 Undamaged 1/4 loss for k2 101 −0.5 −1 −1.5 −2 −2.5 −3 −3.5 100 101 Frequency (Hz) Frequency (Hz) Undamaged 1/4 loss for k2 101 Frequency (Hz) 10−2 Magnitude Magnitude 101 Frequency (Hz) 100 100 Phase (rad) Phase (rad) −6 Phase (rad) Phase (rad) −4 −10 100 (e) Undamaged 1/4 loss for k2 Undama 1/4 loss 10−1 10−3 10 102 −2 101 101 Magnitude Undamaged 1/4 loss for k2 100 −5 −6 −7 100 Frequency (Hz) 100 (c) Magnitude Ph Ph −3 −0.5 −1 −1.5 −2 −2.5 −3 −3.5 100 101 Frequency (Hz) Undam 1/4 loss 101 Frequency (Hz) Figure 30 DTF response: (a) x1 /xg ; (b) x2 /xg ; (c) x3 /xg ; (d) x2 /x1 ; (e) x3 /x1 ; (f ) x3 /x2 ă ă ă ă ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ −4 100 101 Frequency (Hz) Magnitude Ph −6 −8 100 Undamaged 1/4 loss for k3 10−5 101 Frequency (Hz) 10−1 10−2 10−3 10 102 −4 Undamaged 1/4 loss for k3 −6 −8 101 Frequency (Hz) (f) 100 10−5 100 −1 −2 −3 −4 −5 −6 −7 100 −3 −4 100 102 Undamaged 1/4 loss for k3 101 Frequency (Hz) 100 10−1 10−2 10−3 10 102 −0.5 −1 −1.5 −2 −2.5 −3 −3.5 100 Undamaged 1/4 loss for k3 101 Frequency (Hz) 101 102 101 Frequency (Hz) Undama 1/4 loss −2 Magnitude Magnitude Phase (rad) −1 Phase (rad) Phase (rad) −2 101 Frequency (Hz) −10 100 Phase (rad) Undama 1/4 loss 100 (e) 101 Frequency (Hz) (d) 101 (c) 100 100 −4 102 Magnitude Ph 1/4 loss for k3 −2 Undamaged 1/4 loss for k3 101 Frequency (Hz) Undama 1/4 loss f 101 Frequency (Hz) Figure 31 DTF response: (a) x1 /xg ; (b) x2 /xg ; (c) x3 /xg ; (d) x2 /x1 ; (e) x3 /x1 ; (f ) x3 /x2 ă ă ă ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ ¨ soning, as earlier, a stiffness loss for k2 is concluded Case Study: One-Quarter Stiffness Loss for k3 From Fig 31, no identical DTF before and after damage shows up This shows that G3r did not perform correctly The wave reflection at the right end of the structure degrades all DTFs We can conclude that damage must have happened at element 3, and the same reasoning leads to a stiffness loss for k3 1997 HIGHWAYS KAMBIZ DIANATKHAH Lennox Industries Carrollton, TX INTRODUCTION SUMMARY AND CONCLUSIONS This article has introduced a wave propagation approach for computing the DTF responses of nonuniform structures By using DTF responses, boundary effects are ignored in favor of the incident path that the energy takes to travel through a structure It has been shown that the DTF responses, associated with an individual element, are sensitive to physical parameter changes which are directly in the load path from the input force to the measured sensor response The DTF provides direct information regarding the source, location, and amount of damage The number of automobiles has increased drama a result of population and job growth during the eral decades During the same period, the comm tance has also increased (1) This in turn has re congestion in many suburban areas This increas fic flow translates into higher cost for accident a rise in fuel consumption, and air pollution Th ment of Transportation has estimated that the v traffic will increase by 50% in next 25 years (2) of time and productivity and health issues caus creased carbon monoxide and dioxide are the pre factors that call for building smart highway syst ACKNOWLEDGMENTS SMART MATERIALS This work was supported by the National Science Foundation under grant CMS9625004, and Dr.’s S.C Liu and William Anderson served as contract monitors Smart material technology is progressively beco of the most important new research areas for e scientist, and designers Increased use of smart will undoubtedly influence our daily lives fundam the near future Presently, the emergence of smar als and smart structures has resulted in new app that change the way we think about materials, se tuators, and data processing Smart materials ar as materials whose properties alter predictably in to external stimuli Smart materials can be div several categories: BIBLIOGRAPHY S.W Doebling, C.R Farrar, and M.B Prime, and D.W Shevitz, Los Alamos National Laboratory Report No LA-13070-MS, Los Alamos, NM, 1995 C.R Farrar and D.A Jauregui, Smart Mater Struc 7, (5): 704– 719 (1998) C.R Farrar and D.A Jauregui, Smart Mater Struct 7, (5): 720–731 (1998) F.K Chang, Proc 2nd Int Workshop Struct Health Monitoring, Stanford University, Stanford, CA, Sept 8–10, 1999 J.F Doyle,, Exp Mech 35 : 272–280 (1995) K.A Lakshmanan and D.J Pines, J Intelligent Mater Syst Struct 9: 146–155 (1998) A.H von Flowtow and B Schafer, AIAA J Guidance, 19: 673– 680 (1986) Shape-memory alloys: Polymers or alloys member their original shape under an load and temperature through phase tra tion Typical alloys are Ti–Ni (Nitinol) and (K-alloy) Piezoelectric materials where strain result applied load or voltage (electric field), for polyvinylidine fluoride (PVDF) polymer cars that have built-in computer-aided navigation equipment, and a central control unit within each highway system to assist in traffic management OBJECTIVES OF SMART HIGHWAYS The objectives of building smart highways are safety, low maintenance, and conveyance By building smart highways, more vehicles can be on the road thereby reducing congestion and eliminating the need for building additional lanes For safety, computers installed in an automobile will perform all driving tasks, and this will enhance safety because most traffic fatalities are due to human error For general safety purposes, the following are required for a modern highway: (1) piezo MTLS that connect to electric heaters and therefore, not ice up, (2) “glow in the dark” surface material, and (3) standards for traffic flow Maintenance is a major issue in establishing engineering design parameters for estimating the life of the road versus replacement cost The convenience features of smart highways should include optical sensors on-site, cars that have Global Positioning System (GPS)/road maps and bar code road signs, and autopilot features Environmental issues also play a critical role in designing a smart highway system Finally, development of road surfaces that can break down pollutants such as nitrogen oxide gas, will be a major focus of future research for metropolitan areas such as Los Angeles and Denver that have high pollution Smart Structures Smart structures are nonbiological physical structures that have the following attributes: (1) a definite purpose and (2) means and an imperative to achieve that purpose The functional aspects of a smart highway are to integrate the normal design features and provide means of controlling traffic to optimize the traffic flow and human safety Smart highway structures are designed for normal and abnormal events Normal design conditions are deadweight, thermal expansion and cyclic traffic loads (3) A smart structure is a subset of many intelligent structures that is complex and made of innovative materials, control laws, and communications Smart structures have sensors and or actuators to help them function Smart structures generally should be light, take advantage of new computer technology, integrated sensors, actuators, and contain some sense of intelligence to attain structural performance capabilities frastructures and marine structures The most significant areas of concern for sm tures are performance, cost, sensing technology ring integration, structural stress and the need fo technology Smart structures provide useful tan efits, and adaptability is achieved by knowing its, constraints, and compatibility with existin methodologies, and the ability to learn The dra building smart structures lie in the capital inte ture of the projects, lack of understanding wha collect and how to interpret the data, lack of an i team approach, reluctance to change, lack of co and dealing with regulations Major areas of conc sign, construction, and maintenance are monit evaluating structures of bridges, soil, and conc result of stress or natural disasters such as ear that will also be a major factor in constructing ligent highway system Monitoring a bridge by deformation sensors or Doppler vibrometers to d bonded composites is currently being studied t catastrophic failures(1,4) Monitoring the structural integrity of high bridges for safety and ease of repair presents a ing field of study to find new ways to support structure of these systems Smart materials are to play an important role in civil engineering d dams, bridges, highways, and buildings Senso ded throughout a concrete and composite stru sense when any structural area is about to deg notify maintenance personnel to prepare for rep placement Smart materials will be used to im liability, longevity, performance, and reduce th operating smart highways and structures (5) plication of sensors and actuators, diagnostic m structural integrity/repair, damage detection, a hybrid vibrational control will be the major areas sion in building intelligent highways and structu smart materials A smart structure using smar ogy may include the use of fiber-reinforced (FR and optical-fiber sensors Glass or carbon fibers in matrix (6) are used instead of steel to increase th of concrete Steel tends to corrode in salt, water rosive deicing compounds, but reinforced concr susceptible Shape-memory alloys that are important el intelligent (smart) materials can be used to bu sensing structures, for example, a damping dev of shape-memory alloys can absorb seismic en in the strength of steel and concrete and also to determine whether serious damage has occurred (8) Materials technology, specifically, the use of composites and shape-memory alloys in new structures and highways will have a great impact on human society, including the creation of new industries, extension of the women frontier to space, high-speed transportation, and earthquakeresistant and disaster-preventing construction A major area of focus in building a smart highway structure is the road pavement smoothness that creates better driving conditions and increases the life of the road (9) Equipment for road smoothness in all states will be required to set a minimum standard The quality of cement and concrete also plays a significant role in developing a high quality road Paving material quality plays an important role in durability and safety of roads Use of recycled materials and polymer-modified binders have been considered for the durability of paving systems in some California highways (10), compared with traditional asphalt pavement The major obstacles to using recycled or polymer material in pavement materials are extreme loads introduced by heavy trucks that may impact the integrity of the road and the driving performance of the truck However, these new materials cost far less than ordinary asphalt and may also assist in design, construction, and easier maintenance of the roads Another concern that recycled materials and polymers must address is the material’s performance in variable weather conditions and major fluctuations in temperature Highway fatalities have declined about 20% within the past decade from 47,000 to 41,000 annually as a result of safety improvement (9) Road condition plays a critical role in highway safety Liability issues and cost-effectiveness will be significant factors in the development of modern highways in upcoming years A successful, low-cost system for modern automobiles that can reduce fatalities will be a key initial step to globalization of this system Sensors Sensors must have properties that enable them to detect small changes in a structure (8), that is, changes in strain and capability for a measurable output signal The response time of a sensor is a critical issue in monitoring crack growth within a structure; although it will not be as critical in observing stiffness changes from fatigue polymer As mentioned previously, carbon fibers are use forcement in smart structures for strain sensin can replace the need for strain gauges and optica However, the important factor in considering carb is their electromechanical properties, namely, the resistivity of the fibers under load in composite, and concrete structures The electrical resistivit modulus of elasticity are affected by tensive and sive forces (9) Fiber-optic sensors can provide information strain fluctuations as a result of stress and ea ing of a flaw in a joint or within the concrete F sensors, as related to smart structures, can reduc of failure in an aging infrastructure Although fi sensors are not smart and lack actuating capab are the predominant technology that is discuss lation to smart structures Structures instrumen fiber-optic sensors can respond to or warn of im failure and indicate the health of a structure a age Applications include the instrumentation o highways, dams, storage tanks, oil tankers, and b Systems that can measure strain or vibration h tested in the United States and Germany The complexity of such optical systems and the limi fit will make this a slow market to develop Lar and longer term demonstrations will be require acceptance from the engineering community Vibrational measurements could provide inf on any earthquake activity within a region or rioration of a structure Electromagnetic sensors advantage of steel’s magnetic permeability as a fu its internal stress also present tools for monitori cables and prestressed concrete structures (12 method, the internal stresses of highly elastic measured by determining its permeability, whic measured indirectly by its inductance (12) St sors will be a key tool for monitoring crack initia and a good indicator of structural failure Typica in addition to the those mentioned before, inclu gauge sensors, displacement transducers, accele anemometers, electrical time domain reflectom stress/strain sensing), and temperature sensors Many highway systems enforce weight restri large truck to reduce road damage The use of w motion sensors such as piezoelectric polyvinylidin (PVDF) polymer can reduce the damage caused trucks This type of polymer embedded in ela ards, and by posting the best alternate routes Today’s highway systems are long, steep grades and sharp curves that present problems especially for high traffic volume and bad weather conditions More than half of all traffic accidents occur during foggy, rainy, icy, and snowy conditions More than two-third of truck accidents occur on curves or slopes One way to reduce traffic accidents is to use advanced communication systems via satellites and place warning signs The following must be considered in building a smart highway: Traffic control centers assisted by a computer networking system where controllers adjust traffic flow In this center, video-image signals, which are sent by cameras and video cameras, mounted on poles and building, are converted to digital maps Construction of many ramp meters where traffic lights are installed in critical entrance ramps to control the flow of merging traffic Placement and design of narrow poles that support signs and light on fixed objects such as a bridge (14) for safety enhancement Installation of hundreds of sensors in the pavement to count cars as they pass and to estimate and transfer this information to the control center Broadcast of critical traffic information to alert drivers to slow down ahead and advise an alternate route Automatic toll collection where sensors read optical cards on dashboards Sufficient safety enhancement In addition to building and monitoring highways systems, control centers that are assisted by computernetworking system are also required to manage traffic and construct intelligent transportation systems Central units are a way to communicate to drivers and law enforcement officers to reduce routine accidents by improving visibility at night or in bad weather by early warning to drivers (15) An intelligent highway system that has an electronic communication system should be capable of the following tasks: (1) automatically regulate the flow of traffic, (2) provide drivers with up-to-the-minute information, (3) perform most driving tasks, (4) ease carpooling, and (5) manage and guide commercial fleets (14) ger traffic, manage traffic flow, and reduce dela transit systems can be organized for expensive s taxis and to assemble carpools and vanpools for eration (16), for smart goods movement syste sist companies to transport goods more cheaply less energy and resource use, using streamlined spections and better routing through traffic an road–rail transport systems Computer proces ogy can also be used to improve manufacturer in and communications to reduce the need for long shipping and provide faster delivery systems to e purchases from home or local stores To minimize fuel consumption and improve ciency, formulas must be developed based on fac as the lane miles of roadway, vehicle miles tra level of mail routes, the population and the s state in square miles and transmitted to a central system ADVANCED AUTOMOBILES In conjunction with smart highways, smart veh provide complete control of nearly all driving The major tasks of driving consist of navigation steering, throttle control, and avoiding accident hicle will automatically control traffic light man Most automated driving tasks have already be mented in pilot vehicles by major auto manufact Today’s automobiles carry more advanced sem tor technologies than they did in the early 198 now, chip technology has been used to enhanc formance of engines and to control airbags and brake systems However, navigation systems su global satellite (GPS) systems will dominate the eration of telecommunication advances in conges areas Modern car manufacturers have develope tion systems to pinpoint a driver’s location and developing systems that activate warnings for av jects in the blind spots Collision avoidance unit under development to steer, brake, or accelerate automatically (15) A unique feature of the mod mobile is the card key for opening car doors and formation identification that will enable identify speed and taking care of tolls The microchipcard that is slightly than a normal credit card ca within three feet of the vehicle A Siemens Sma need to build smarter and more sophisticated automobiles Smart sensors and devices can be used to control traction, steering, and suspension and monitor tire pressure and sense and orient a car automatically to road conditions Sensors that can control the speed, vibration, and temperature of vehicles could be used in conjunction with road sensors to optimize most critical functions of an automobile Sensors could also be used in the rear and front of a vehicle to warn drivers that they are getting too close to another vehicle or are being approached too closely by another automobile in addition to lane changing Optical sensors, based on the misbonding effect and speckle phenomenon technology, can be used to identify a vehicle type and its speed and also to monitor traffic flow and count vehicles on the road These sensors can be placed inside the asphalt layer of the road surface (17) The major area of focus for automobiles besides safety is the use of sensors for automation and using exotic materials such as composites to substitute for steel to improve fuel efficiency The use of new materials such as composites provides a multitude of potentials and degrees of freedom for materials design that involve increased strength, creating new functions and expanding to multiple functionalities Smart materials consist of composites that indicate exactly the direction of the future development of materials engineering and represent a change from “supporting” to “working” to build up a new materials application system that integrates structures, functions, and information Smart sonic traffic sensors placed on acoustic sensors is another alternative (18) to magnetic-loop sensors to detect vehicles from the sounds that they make More sophisticated cruise control could be tied in with sensors to reduce or increase speed instantaneously to avoid accidents General Motors and Ford have tested computerized navigation systems to pinpoint a driver’s location and to warn drivers of potential obstacles by using detection systems to steer clear of objects in blind spots and avoid collisions Modem automobiles for smart highways also being considered where by drivers can take their hands off the wheel and eyes off the road enabling advanced cruise controls take charge of the driving (19) Smart highways and uniform speed are the major requirements for this futuristic idea before such cars can be used In addition to cruise control features, this type of vehicle is equipped with radar fields Sensors emit a beep if the car is about to hit something This type of technology is currently available the traditional driving mode where the driver i mand In the second mode, the driver uses th ger seat, and the car uses sensors and HR6 tech take complete care of all driving tasks This typ cle is equipped with the latest communication to the driver can monitor the traffic and weather co access the Internet, and check e-mail In the a mode, the vehicle body turns into an aerodyna Mitsubishi has also developed cars that have mul sors to detect steep roads, curves, and hazardous s the vehicle adjusts after detecting any upcoming sign Ford also uses a new technology of light bea where the size and the shape of the beam are c with the speed and the type of road, and radar is in the vehicle This technology provides ideal s trol for safety and fuel efficiency Jaguar progress in night vision technology where infrared technol the driver the ability to monitor any object that observed during darkness Mercedes-Benz in co with Boeing is also developing a limousine equip the latest electronic features that can drive on its is also equipped with GPS technology The future car for the twenty-first-century like a rolling recreation room and a source of e ment as manufacturers progress in developing n nology This in turn will reduce traffic commuting home and office and also will require far less from drivers Future automobiles will be design resent true mobility rather than a transportation new generation of cars will possess more revo and innovative electronic features to ease driv and access communication networks for weather, worldwide web, and satellite or cellular networ far, the United States has lagged behind oth trialized nations; in 1998, of all vehicles equip navigation systems, less than 5% were purchas United States and more than 90% were sold in Eu Japan, where there is higher demand for comm technology Recent developments in automobile manufactu sist of using a laptop computer to interface with a to warn and control the devices in a car The ne computers offer ample processing power and d and can operate on 12 Vdc power On new futur the laptop computer is likely to become standard cost of remaining associated hardware is expect significantly in the near future vide latitude and longitude, speed, and direction of travel GPS is beneficial to improve safety for trucks by installing receivers and sensors in the trailer section The load of the truck then can be monitored in addition to determining tax and fuel rates (9) The driver can also receive real-time traffic and navigational information The development of a central GPS unit for nationwide use is critical for managing multiple functions for entire smart highways within all states This system could be used on land as well as in the air and on the sea Development of common equipment standards, technical feasibility and accessibility, and organizational structures will be the key issues for coordinating this system Global positioning data can presently be provided from a network of Department of Defense (DOD) satellites Planes, boats, vehicles, and mapping and survey teams can determine their positions on earth by using equipment that receives and interprets signals from these satellites For smart highway applications, the satellites provide a signal that is accurate to about 100 meters without the use of GPS Coordination between the federal government and local states may be needed to enhance joint development or sharing of Differential Global Positioning Systems equipment, facilities, and information for future use The limitation of existing GPS technology lies in highly populated areas that have large buildings and trees GPS is not functional inside a tunnel or any enclosed area (16) Present GPS technology relies on a satellite signal whose signal is received and translated by a receiver (9) The system works perfectly in an uninhabited area where it may not be as useful The price of a GPS system has fallen dramatically in recent months provided that the automobile is equipped with a portable computer Earthmate sells for less than $180 and is a high-performance, easy-to-use receiver that links to the satellite navigation technology of the Global Positioning System (GPS) UPDATE ON SMART HIGHWAY PROJECTS UNDER CONSTRUCTION Major smart highway development has been underway in the states of California and New Jersey Thus far, major problems consist of major delays in completing construction and some minor accidents due to the extreme weight of signs that require support New Jersey’s Route 80 from the George Washington Bridge to its connection with 287 in Morris County and Routes 95, 23, 46, 4, 17, 202, 287, construction begins Chrysler Corporation has vehicles, particularly large trucks for smart highw are presently being tested without any drivers the company is not betting that any major smar projects will be started soon Chrysler believes t are many old cars on the road that may interfer general concept of fully automated highways Thi a case for a two-tier highway system, one for m hicles and one for cars that are not equipped w computers The associated costs and capital for new highways must be considered relative to pote enues Another dilemma concerns turning over t of human lives on such highways to a major corp the government (20) SMART HIGHWAYS IN JAPAN Japan has been far ahead of most industrialize in developing and using smart materials; there beneficial to review the recent progress of smar highways that is a model for the rest of the worl Traffic fatalities in Japan are approximately year (21) at an annual cost of $120 billion Popul sity is also 12 times higher than that in the Unit Therefore, the benefits of constructing an ITS sy have a tremendous impact on productivity The nual budget for an intelligent transportation sys timated at 700 million, proportionally higher th the United States (21) Japan has more than 380 toll roads and development is underway to autom collection system fully Although traffic control systems have been Japan for a number of years to ease traffic, the jective in automating a traffic system in Japan b smart highway system is for safety enhancemen duction of traffic fatalities Another objective is t communication between vehicles, particularly co vehicles and public transit, by using a central tr agement system Today, Japan has more than 11 smart highways, which consist of 2,077 vehicle to monitor the number of vehicles and speed Th roads also have graphic displays and television Presently, Japanese auto manufacturers offer 40 models of navigation systems; approximate sale million units per year In Japan, ITS development began in the 1970s by construction of a road system, the E ρVc Austenite to R-phase transf R-phase to martensite phase transformation Figure 11 Mechanism of the all-round shape-memory effect [after Kainuma et al (20)] spinning produces an intrinsic reversible effect under certain conditions (22) Another example is the fabrication process of Ni–Ti shape-memory films developed by Lehnert et al (23) The method consists of successively depositing pure layers of Ni and Ti ranging from 10 to 20 nm thick The heat process allows interdiffusion of Ni and Ti and subsequent crystallization This process leads to an intrinsic two-way effect resulting in a bending motion Outlook for Reversible SMA Actuators SMAs actuators can be divided into two categories, depending on the method used to produce the reversible effect (Fig 12) The first, called monolithic, refers to methods for which no external elements are used to produce the reversible effect, and the second, called multiparts or mechanism, refers to a reversible effect obtained by adding a second element The subclass “smart design” denotes the use of design strategies to obtain a reversible motion within a single piece of material These design methods are presented later Thermal Response of SMA Actuators Time response is the well-known limitation of shapememory alloys Because the actuation method is thermal, dT = Qh − (Qconduction + Qconvection + Qradiatio dt where ρ is the material density, V the volume, c t heat, T the temperature, and Qh denotes the he Considering the usual working temperature of SM tors and an application at ground level, the loss tion can be neglected (Qradiation = 0) Convectio ply expressed by hA(T − T∞ ), and the conduction should be evaluated depending on the applicatio Then, the transient response can be expressed a  T + Q [1 − e−τ t ] ∞ hA T(t) = ,  T∞ + (T0 − T∞ )e−τ t (heating) (cooling) where T∞ is the temperature of the surrounding T0 the initial temperature, A the area of the mate volume, h the convection coefficient, and Q des uniform heat medium The time constant τ is s ρVc/ hA, where ρ is the density and c the spe For electrical heating, which is the most comm provide heat to an SMA actuator, Q is equal to I I is the current through the material and R is the resistance Using Eq (11), the heating and cooling time c pressed by t=   ρVc     hA ln   ρVc      hA ln T − T∞ T0 − T∞ (T − T∞ ) − (T0 − T∞ ) − (cooling) Q hA Q hA (heating) However, this simple approach does not consid lease and absorption of internal latent heat d phase transformation Therefore, using Eq (12) response is always underestimated To include of latent heat, a method used is based on a tem dependent specific heat For instance, Brailov (24) considered a polynomial approximation of th heat during the transformation, where the pol coefficients are determined by using differential calorimeter (DSC) measurements The delay intr the phase transformation is often negligible in h is also interesting to note that the latent heat d Composites materials Mecanism Actuator with bias spring Antagonistic design “Dead weight” actuator Antagonistic design Local transformation Local annealing / Laser annealing Local hardening Local hardening Irradiation (neutrons, ions, electrons) Smart design Thermo-mechanical treatme Two-way shape memory effect obtained by precipitate (All-round effect) Two-way shape memory effect obtained by training processes Other processes Fabrication processes Multi-layers thin films Melt-spining Figure 12 SMA reversible actuators: an outlook on the known methods for producing reversible motion Table Thermal Properties of SMA vs Copper at 300 K Ti–Ni Density kg·m−3 6450 (12.52 slugs/ft3 ) Copper (pure) 8933 kg·m−3 (17.34 slugs/ft3 ) A: 18 W/m K (10.4 BTU/ft h◦ F) 401 W/mK (231.67 BTU /ft h◦ F) Thermal conductivity M:8.6 W/m K (5.0 BTU/ft h◦ F) Specific Heat (c p ) 836 J/kg K (0.20 BTU/lb◦ F) the applied stress and the material The higher t the smaller the latent heat Of course, this stres fects the fatigue performance of the actuator 385 J/kg K (0.09 BTU/lb ◦ F) Fatigue An actuator working at a strain level of 8% does not have the same lifetime as an actuator at a strain level of 2% Therefore, fatigue spec contribute to establishing the design paramete actuator Many parameters make fatigue analysis com loy composition, loading speed, heat treatment, a cold-work, etc are some of the relevant parame need to be considered A method based on factoria be scaled down; it has to be redesigned.” Let us on the use of shape-memory alloys in microengin Stalmans suggested in (2), three different types of fatigue have to be considered: Microrobotics and Microdevices r failure by fracture due to stress or strain cycling at We usually speak of microdevices when the res the motion and the dimensions of the parts ar than the precision and dimension usually a in a workshop In a resolution-of-motion versu components representation, microrobotics is ty cated in a region defined by resolutions rang 10 microns (about × 10−4 inch) to nanometer ( 10−8 inch) and dimensions ranging from 10 mm inch) to micron (about × 10−5 inch) These b are rather a trend than a definition Figure 13 tically illustrates this idea and some well-know nisms are presented Micro-devices, which integrate other funct as controlling integrated circuits, are usually c croelectromechanical systems” (MEMS) The M also often associated with silicon-based technol constant temperature r changes in physical, mechanical and functional properties, for instance, the two-way, shape-memory effect r degradation of the shape-memory effect due to stress, strain, or temperature cycling in the transformation region Considering all of these aspects, it is difficult to give some general rules However, some recommendations are made about the maximum strain versus the number of cycles (Table 4) For binary Ni–Ti, long cycle lifetimes higher than 1,000,000 have been obtained using the R-phase transformation, which, it is known, is very stable (1,2) Dimensions of mechanisms High precision mechanisms Large scale manipulator Commonsize robots Precision robotics Mechanic m Atomic force microscope Linear positioning device Watches Mobile micro-robots Electrostatic motors mm Micro-mechanisms (MEMS) µm Molecular robotics Nano-mechanisms Figure 13 Micromechanisms: a definition as a function of dimensions and resolution of motion nm µm mm 10−2 Figure 14 An illustration of the sca on cube: ratio between electrostatic gravitational force as a function of cu sion processing However, according to the definition, MEMS should not be restricted to these special processes Therefore, as a clever definition, the term “microdevices” is used throughout this article, rather than MEMS, to qualify small mechanisms, integrated or not Scaling Effect We have all tried to manipulate small objects in our childhood and found that they stick very well to our fingers! This means that the relevant force is not gravity but rather adhesive forces For instance, electrostatic or surface tension forces cannot be neglected when considering manipulation on a millimeter scale This idea is briefly illustrated in Fig 14 where the ratio between electrostatic force and gravitational force for a small cube is plotted as a function of the cube’s dimensions These scaling effects drastically modify our perception of the microworld and a lot of things have to be redesigned to adapt to “this new world.” New actuators and new mechanical components have to be created SMA is certainly one of the most interesting candidates in this fascinating field of research as actuators and also as high strain flexible structures, as shown later SMA as a Smart Material for Microrobotics The shape-memory effect in combination with an appropriate design or treatment can be used for microactuators as well as for microfastening devices For instance, superelasticity can provide an efficient method for enhancing the stroke of flexural hinges In the previous section, SMA materials were compared to other candidates for small applications In addition to some of the advantages already described, several interesting properties, relevant to microengineering applications, can be mentioned: r As mentioned before, SMAs have the highest power/weight ratio among all known actuators in microengineering Applications that require force production can be very compact r SMAs offer solid-state actuation and thus d duce any dust They are suitable for clean r ditions r The resistivity of the material (similar to steel) is low, which allows efficient Joule hea voltage and simple electronics can be used SMA actuators r Ni–Ti is compatible with MEMS processes However, a few drawbacks need to be conside fully, such as their thermal activation, which l bandwidth For fast applications, an SMA canno with electrically or magnetically field driven actua actuator efficiency is also very poor (typically a which often excludes SMA microactuators from tions whose power consumption is very low For the power consumption of a watch motor is typic picowatts! Micromachining and Fabrication of SMA Microdev Silicon-Based MEMS: Fabrication Processes (28 technologies and related processes were maj through in microelectronics as well as in senso tors, and microsystems Silicon based processe a unique method for large-scale production an turization in the development of microactuato production methods are massively parallel and al processing However, these fabrication processe few limitations: r Silicon microstructuring technologies are pl r The technological investment is very high r These methods are usually confined to stru limited aspect ratios Ni–Ti and Ni–Ti–X (where X = Cu, Hf, Pd) allo deposited on various substrates such as silicon, si ide (SiO2 ) or titanium However, the Si substra ally avoided because of the possible formation o (SiNi) during crystallization of Ni–Ti thin films scription of this process as well as the properties of thin film produced can be found in (29) Other processes such as deposition by flash evaporation (30), laser ablation (31), and multilayer processing (23) can also be used to produce thin films Ni–Ti thin films are usually etched by using HF/ HNO3 / H2 O solutions Table shows experimental data obtained by Makino et al (32) According to their results, a solution of HF:HNO3 :H2 O = 1:1:4 allows selectively etching of a Ni–Ti film deposited on a Si substrate Different fabrication methods can be used that depend on the type of actuator When the actuator is moving freely, an intermediate layer, the so-called sacrificial layer, is required Figure 15 presents a method of producing a freely moving micromechanical element represented by a spring (33) In the first step, a polyimide layer is deposited by spin coating and is subsequently cured The second step is Ni–Ti thin film deposition on the polyimide layer In the next step, a photoresist thin film is spin coated over the Ni–Ti thin film Then, the photoresist is baked, patterned, and developed to expose specific portions of the Ni–Ti layer selectively Then the Ni–Ti film is etched Finally, the structure is freed through reactive ion etching of the sacrificial polyimide layer In another process proposed by Buchaillot et al (34), the Ni–Ti film is deposited just before removing the sacrificial layer, as is illustrated in Fig 16 (29) A sacrificial Cr layer is deposited on a SiO2 substrate A polyimide layer is spin-coated on the Cr layer Then, a second Cr layer is deposited on the polyimide The upper Cr layer is patterned by photolithography and wet etched The polyimide layer is vertically etched by oxygen plasma using the previously patterned Cr layer as a mask Then, the Ni–Ti layer whose thickness is less than that of the polyimide layer is deposited The final step consists of removing the polyimide layer by wet etching (“liftoff”) and releasing the mobile part by wet etching the Cr layer Because the Ni–Ti layer is never etched, the aspect ratio of the structure is theoretically not limited Several applications such as microvalves use membrane actuators or biomorphs of Ni–Ti on SiO2 This structure can be fabricated by using the process proposed by Wolf et al (35) (Fig 17) A Si wafer that has an oxide layer on both sides (SiO2 ) is coated by a photoresist and softbacked The back side photoresist is patterned and results in direct exposure of the Si in defined parts that creates windowlike structures Then, the Si is etched and almost all of the is depositing the Ni–Ti layer Finally, the rest of t is removed When bimorphlike structures are de front SiO2 layer is not removed, and Ni–Ti is directly on it Laser Machining A laser can be used for m cutting SMA elements Nd–Yag lasers are usu The sample is fixed on a two-axis linear stage fixed focusing objective A rotational stage is s added for tube cutting The laser is focused on th and an additional gas flow is usually used to dra molten material The cutting precision depends ject size and the material It is possible to cut she thicknesses range from a few millimeters down t (a) Ni-Ti Polyimide (b) Photoresist (c) Wet etching (HF/HNO3) (d) Plasma etching Figure 15 Thin-film microactuator fabrication proc position of the layers, (b) photoresist deposition and (c) wet etching of SMA layer, and (d) plasma etching imide layer [after Walker et al (33)] SiO2 Cr Crystallisation Polyimide Ti-Ni Figure 16 “Liftoff ” fabrication process: (1) deposition of the layers (Cr, Polyimide), (2) Cr upper layer patterning and polyimide etching, (3) Ti–Ni sputtering (4) liftoff by removing polyimide using KOH, and (5,6) CR wet etching [after Buchaillot et al (34)] (2 × 10−4 in) The spot size varies according to the sample’s thickness Results for Ni–Ti have shown that the minimum spot size ranges from 0.02 mm (8 × 10−4 in) for 0.01-mm sheet thickness to 0.08 mm (4 × 10−3 in) for 1-mm sheet thickness (0.04 in) Figure 18 shows part of a microgripper mac a Nd–Yag slab laser working in the fundamen (TEMoo) An additional treatment has been us crease the edge surface roughness For small to production volume and prototyping, laser cutt (a) SiO2 Si SiO2 (b) SiO2 Si SiO2 (c') (c) TiNi SiO2 Si Si SiO2 (d) (d') TiNi SiO2 Si Si SiO2 Figure 17 Microfabrication of Ni–Ti diaphragms with and without a SiO2 layer (a) Coating SiO2 layers and patterning of the bottom one, (b) silicon etching, (c,c’) Ni–Ti deposition on Si, (c) (the SiO2 has been etched) or directly on the SiO2 layer, and (d,d’) removing the remaining silicon (35) Figure 18 A Ni–Ti structure that was laser-cut by a Nd: Yag slab laser Acc V Spot Magn 10.0 kV 4.0 222x efficient method for Ni–Ti However, due to laser heating, a heat-affected zone exists around the cutting edge The depth of this heat-affected zone is typically of the order of 10 microns (4×10−4 inch) Electrodischarge Machining Electrodischarge machining (EDM) consists of using sparks created between an electrode and a sample to machine the material There are two different processes: diesinking EDM and wire-cut EDM In die-sinking EDM, the required shape is formed negatively in the metal by using a three-dimensional electrode Various shapes, indentations, and cavities can be created In the wire-cut process, the required shape is cut by guiding the wire along the given stretch Figure 19 A Ni–Ti micro-surgical tool: the groove was machined by electrodischarge machining The groove is 40 microns wide (0.0015 inch) by about 200 microns deep (0.0078 inch) (36) Acc V Spot Magn 5.00 kV 2.0 73x 100 µm EPFL - IGA - ISR - IOA Figure 19 shows a microsurgical tool in which about 40 microns (1.57×10−3 inch) wide and abou crons (7.9×10−3 inch) deep has been machined b micro-EDM (36) The groove is millimeters ( long EDM is the most efficient method for creat tures of high aspect ratios that have low surface r EDM is usually used for micromachining of ela tures of high aspect ratio Recent Developments in Microrobotics and Microd The State of the Art In this section, some recent applications are p This discussion is not extensive but gives som Det SE WD Exp 22.4 EPFL - ISR - H.Langen (0.01 electrode trends in the field of microengineering Electronic connectors and other small fasteners are not mentioned in this section The reader interested may consult the applications section of this encyclopedia or (1,2) for a more detailed description Active Catheter and Snakelike Robotic Systems One of the first microrobotic applications was a prototype of an active miniature endoscope for a gastrointestinal intervention system (18) This endoscope was designed to pass smoothly through the sigmoid colon, which has a very small radius of curvature The outside diameter was about 13 mm (0.51 in) As one can easily imagine, the heat dissipation of active surgery instruments is a critical issue To address this problem, authors have proposed a cooling water tube going through the structure As a result of this pioneering research, many active endoscopes for gastrointestinal surgery have been developed worldwide, for example, by Reynaerts et al (37) All of these projects are minirobotic rather than microrobotic Outside diameters range from mm (0.31 in) to 15 mm (0.59 in) In the 1990s, smaller active endoscopes and active catheters were proposed in Japan Fukuda et al (38) designed an active catheter whose diameter ranged from about 1.33 mm (0.052 in) to mm (0.079 in) The catheter has also been tested in vivo A very impressive five-degrees-of-freedom tube-type micromanipulator was recently introduced by Olympus Optical Company in Japan (39) This manipulator, shown in Fig 20, is dedicated to inspection and maintenance in narrow spaces and for medical applications The diameter is one mm (0.04 in), which makes this snakelike robot one of the thinnest in the world The active parts consist of SMA strips working in bending modes The strips have a twoway reversible effect and are heated by thermal conduction A “multifunction integrated film” (MIF) is attached to the SMA strip This circuit combines a heating function and a strain sensor by using a strain gauge measurement principle The integrated circuit is constructed by successive sputtering of Ti and Pt on a polyimide layer The authors developed a position controller based on strain sensing feedback, which gives positional accuracy of less than ±0.25 mm (0.04 in) In another version, tactile sensors have been added that give reflex functions to the catheter, and match; (b) exploded view of the inside when the tube touches something, it automatica in the opposite direction Mini-to Microgrippers At the end of the 198 (40) proposed a miniature gripper suitable for cl conditions The device consists of two bronze fin have two SMA coil springs attached to each finge springs are mounted opposite one another in an a tic or “push-pull” actuating arrangement Thes were made from a thin film sputter-deposited o substrate The alloy showed an R-phase transf around 320 K The electrical resistivity change as an internal variable for monitoring the transfo In this design, Ikuta proposed a controller sch combined a resistance and positional feedback w positional feedback is given by a photosensor In another realization, Hesselbach et al Germany developed a compliant microgripper In one version (Fig 21), the compliant struc flexural hinges is machined by electricodischarge ing (EDM) from a superelastic material Thus, th uses SMA materials for two purposes: for actuatio the guiding mechanism The compliant mechanis four-bar linkage that has a transmission ratio tween the input and the output of the mechanis Gripping jaws mm M Flexible hinges α Figure 21 The microgripper developed by Hesselbach (a) Overall view of the microgripper (5 mm = 0.2 in); and mechanical equivalent of the flexible structure A microgripper (43) for micro endoscopic (i.e scope whose diameter is about mm (0.04 inch)) is presented in more detail in the next paragrap Pushing pad Figure 22 The Lawrence Livermore National Laboratory’s microgripper (200 microns = 0.008 inch) (42) input is one jaw and the output is the other jaw, a movement of the input jaw induces exactly the opposite movement of the output jaw Hence, the jaws center every grasped object, regardless of the different stiffness of the flexural hinges or disturbance forces An SMA actuator that has a strip shape deforms the compliant structure The reversible motion is obtained by using the elasticity of the compliant structure, which acts as a bias spring A resistance feedback is used to control the position of the finger The authors have reported a lifetime of more than 450,000 gripping cycles and a close/open time of 0.5 s In microrobotics, Lee et al (42) realized one of the smallest SMA microgrippers (Fig 22) This gripper’s dimensions are × 0.2 × 0.38 mm3 (0.04 × 0.008 × 0.015 in) The design principle is a kind of “bimorph” of SMA/Si materials The Si layer acts as a bias spring The mechanism consists of two identical jaws actuated by Ni–Ti–Cu The thin films are deposited on both external sides of the gripping jaws The jaws are made of silicon and were shaped by a Preloading spring Linear Actuator and Other Small Actuators G Hayward (44) proposed a high-speed linear actu sisting of several thin Ni–Ti fibers woven in a c tating helical pattern around supporting disks The volume required by the actuator is a 17-mm cylinder × 30 mm (0.67 in × 1.18 inch) long that the category of “minirobotics” rather than “micro Preloaded springs separate the disks, which keep under tension when unheated When heated, shrink and pull the disks, together The weave the fibers accomplishes displacement amplificat tary actuator has been developed using two of th actuators mounted opposite each other In this tion, a time response of less than 100 ms was However, due to the large amount of fibers in this woven design requires a high level of elec rent (typically 4–8 A) Using Si-based processes, Buchaillot et al veloped an XY linear stage made of Ni–Ti t This linear stage consists of four leaf spring ing in the same plane as the substrate Each has two parallel leafs springs and is initially during mounting to induce a martensitic reor Supporting disks SMA fiber Notches 25 mm (0.98 inch) Side view 17 mm (0.67 inch) Top view Figure 23 The “woven structure” developed by Grant and Hayward (44) Control pressure Fluid out Fluid in C Open position Figure 24 The microvalve developed by Johnson et al (47,48) Polyimide membrane PMMA Cov NiTi microd Actuators are mechanically connected to each other and are designed to avoid any coupling between degrees of freedom Therefore, the basic actuating principle is an antagonistic design, where each actuator deforms its counterpart The whole structure is a few millimeters square (0.04 inch2 ) Many other reversible actuators have been proposed in the literature Among these designs, Kuribayashi et al (45) in 1993 proposed a millimeter-sized robotic arm The allround effect (see earlier) was used to produce a reversible effect in a Ni–Ti alloy beam Three of these microcantilevers were combined to realize a SCARA-type microrobot that has three degrees of freedom Later, Kuribayashi and Fujii developed a microcantilever made of a SMA thin film prestressed by a polyimide layer (46) The polyimide layer provides reverse motion when cooling Fluidic Applications The pioneers of this field of research were Busch and Johnson (47) In 1989, they patented a micro valve, that uses a SMA thin film that has been processed by MEMS technologies The valve shown in Fig 24, consists of a silicon orifice die, an actuator die that has a poppet controlled by a Ni–Ti shape-memory alloy microribbon, and a bias spring An electrical current heats the actuator (the Ni–Ti microribbon) When no electric current passes through the actuator, the bias spring pushes the poppet against the orifice and closes the valve When an electric current is applied, the Ni–Ti actuator contracts and lifts the poppet from the orifice, opening the valve The whole device is less than one-half square centimeter (0.2 in2 ) The displacement of the poppet is more than 100 microns (0.004 inch) and produces a force of onehalf newton In a rough approximation, flow through the valve is proportional to the current applied to the actuator The current and an appropriate feedback may be used to control the flow (48) In another realization, Kohl et al (49) designed microvalves by integrating SMA bending actuators fabricated from laser-cut, cold-rolled Ni–Ti sheets (Fig 25) Polyimide sp Valve s Pins PMMA Substrate B A Figure 25 The microvalve developed at the Forshun in Karlsruhe (49) The valve consists mainly of a PMMA housing an integrated valve seat, a polyimide membran imide spacer, and a SMA microdevice The SMA tuator is used for deflection control of the membra opens or closes an opening in the valve seat The a made of “stress-optimized” cantilever elements w tuation direction is perpendicular to the plane o strate A CONCEPT OF SMART SMA MICRODEVICES Basic Idea Active mechanical systems generally consist o generator, a coupling device, a transmission syste ing system, and an output element The basic idea grate all of these elements and functions within o piece of material to form “a smart SMA microdev concept is well adapted to microrobotics because lithic device does not require any assembly and wear, friction, and dust Nevertheless, to be rev monolithic design requires that one region of the act as a biasing spring Thus, several methods h developed to address this issue (50,51) Martensitic shape Fixation zone and vision patterns Shock absorber Hinge Gripping jaws Figure 26 The EPFL’s microgripper for submillimeter lens manipulation (43,50) The Two-Way Shape-Memory Effect (TWSME) Applied to Small Devices As presented before, thermomechanical treatment of a material can introduce internal stress This stress will generate preferentially oriented martensitic variants when cooling Using an appropriate design, the TWSME can also be used to create reversible motion within a device that has a more complex shape than wires, strips, or springs The TWSME microgripper shown in Fig 26 was designed and successfully introduced in a production line of microendoscopes (43) The design consists of crablike tweezers about mm2 (0.04 inch2 ), and the gripper has a moving arm that clamps object against a fixed part Additional fixtures have been added for visual measurements during calibration and a force limitation in case of mechanical shock to the fixed part The material is a Ni–Ti–Cu sheet 0.15 mm (0.006 in) thick The training process is explained in Fig 27: The bulk material is annealed at 515◦ C for 15 The microgripper shape is then laser-cut from the sheet At this point, the gripper does not have any reversible motion: the austenitic shape is the cut shape The microgripper is fixed on a heat source (thick resistive film, SMD-resistor, thermoelectric devices) Austenitic shape Figure 27 The training process applied to the m (thermal cycles under constraint) Then, the training process is applied to t Different training methods have been repor literature (1,2) In this special case, the be for achieving gripper specifications is to a mal cycling under constraint To perform a shaft is used to deform the gripper up t The maximum motion range is reached af cles After training cycles, the austenitic sh exactly the same as the cut shape Residu deformation resulting from the training pro served and has to be considered while desi gripper During cooling, the gripper opens When he gripper recovers its parent shape (high-tem shape) and closes its jaw For a hinge th 70 µm (2.7 × 10−3 in), a range of motion o (6 × 10−3 inch) was obtained, and a gras of 16 mN was measured If this force is with the weight of the gripper itself, the force/weight ratio is 1000, and if only the a of the gripper is considered in the gripper w timate, this ratio increases to 30,000! How force when cooling is about four times low ximately 4mN) This is a limitation of the the force induced when cooling is lower th covery force generated upon heating Nev a gripper needs forces when closing to gra ject but does not need force when opening experiments consisting of grasping and rel cles showed that motion loss saturates afte cycles After 200,000 cycles, the motion lo and the loss after 1,000,000 cycles is estima same level Local Heating of the Material (“Martensite” Spring-Biased Design) To create a reversible effect, active and passive needed simultaneously within the material On consists of controlling the heating process thr Acc.V Magn 10.0 kV 22x EPFL Figure 28 Optical switch us cal heating design principle ing part, (2) bias spring, (3) sh (4,5) fixture and connection p mm = 0.04 in) mm IGA-ISR-IOA PROJET SMA structure By carefully designing the electrical path, for example, part of the material can be in the austenitic state, and the remainder is in the martensitic state The elasticity of the martensitic region can be used as a bias spring The optical switch shown in Fig 28 illustrates this idea The device was machined from a 20 µm thick (7.9 × 10−4 inch) Ni–Ti–Cu foil The motion can be compared to that of a semaphore During assembly, the structure is prestressed by pulling the fixed part (#5, see Fig 28) Then, parts #4 and #5 are fixed and separated from each other (dashed line) When a current passes between the two pads (#4 and #5), the part that has the thinnest section (#1) is heated up to the transformation temperature, recovers its original shape, and pulls down the shutter (#3) When cooling down, the bias spring—the part that has the largest section (#2)— pulls back on the structure causing reverse motion In this design, the shutter range of motion is 190 àm (7.5 ì 103 in) The time response when heating is less than 50 ms, and the power consumption is typically mW Antagonistic or “Push-Pull” Design The working principle is the same as described before, except that the two SMA actuators are part of the same piece of material By designing an appropriate electrical path, one zone of the material is heated at a time Depending on which zone is heated, the device will move in one direction or the opposite The main difference from the previous case is that the reverse motion is due to alternative heating of actuating zones An illustration of this principle is shown in Fig 29 This device is a linear actuator that has one degree of freedom During assembly, the two leaf springs (#2, #2∗ ) are prestressed and fixed on the contact pads (#3, #3∗ ) When a current passes through one of the contacts (#3, #3*) and the main body (#5), the corresponding spring is heated up to the transformation temperature and pulls th part (#4) in its direction When the springs are opposite each other, two-way motion is obtaine tain precise motion in only one direction, a guidin (#1) has been added to compensate for machinin sembly errors The schematic of the mechanical e system is shown in Fig 29 It consists of two se link structures One four-link structure guides th in one direction, but this guiding is not purely re rather it is motion that has a high radius of c Adding a second four-link structure in series com for undesired motion in the perpendicular direct linear actuator illustrates very well how smart lithic design can be: this design is a complete lin that includes actuators and guiding elements w single piece of material Leaf spring actuator Mobile part Leaf spring IA δ Guiding system δ Figure 29 A one-degree-of-freedom linear stage The m equivalent of guidance is explained 100 0.00 0.02 0.04 0.06 Déformation (%) 0.08 Figure 30 Mechanical behavior before and after annealing a cold-worked sheet having a superelastic behavior at room temperature Local Annealing Basic Principle During the fabrication of SMA devices, annealing is required to obtain a martensitic transformation within the material In a sputter-deposited thin film, the material is amorphous after deposition Annealing is required to crystallize the material In a cold-rolled sheet, the forming process introduces a large number of dislocations which prevent martensitic transformation In that case, annealing induces recovery and recrystallization of the material Typical annealing temperatures range from 400 to 900◦ C The basic idea of local annealing is to limit annealing to selected zones of the material Martensitic transformation occurs only in annealed zones In other words, some parts are active, and others are passive Figure 30 shows a tensile test on a cold-worked sheet of Ni–Ti before and after annealing (the material is in the austenitic state) The superelastic characteristic can be seen only after annealing In the framework of a monolithic design, the annealed zone can be used for actuating purposes, and the nonannealed zone can be used as a biasing element for a reversible effect Two methods can be used for annealing small devices locally (51) The first is direct Joule heating of the material, and the second uses a laser to heat the structure Local Annealing by Joule Heating Direct Joule heating actuates most SMA actuators and devices It can also be used to heat up the material to the annealing temperature Kuribayashi et al (52) used annealing by Joule heating a microactuator consisting of a cantilever working in the bending mode An example of a structure locally “Joule annealed” is shown in Fig 31 The device is a linear actuator that has four leaf springs (51) These are the fabrication steps: The structure is cut out by laser from an “as-received” unannealed Ni–Ti–Cu sheet An electrical current is passed through springs causing the temperature to rise to th ing temperature (about 500◦ C) The right s main at room temperature An infrared cam sures the temperature during annealing Once annealing is over, the springs are pr along the axis and are fixed Because of this local annealing, one part of the m passive, and the other is active The passive pa as a spring, and the active part is the actuator chanical behavior of this structure is exactly th the spring-biased multipart mechanism presen previous section “Joule annealing” can be a low-cost and efficie of producing smart SMA microdevices However, a tations exist: r This method can be applied only to devices tain loops for an electrical path r The cross section along the electrical path carefully designed with respect to temperat bution during heating The thinnest sectio the hottest and will be annealed first r Because the power dissipated in the str directly proportional to the resistance an the cross section, the local temperature an nealing conditions will depend strongly on m tolerances Laser Annealing of SMA (the LASMA Process beam is used for local heating The laser is f the point where annealing is desired In contras annealing, laser annealing of SMA (LASMA) c plied to all kinds of designs (51,55) This method successfully used on various types and shapes of crodevices It can easily be applied to thin films, c sheets, cold-drawn wires, or other materials that work-hardened Figure 32 shows some microstructures in annealed thin film The material’s microstructur served at several distances from where the las cused Near a distance of about 800 àm (3.15 ì the transition from crystallized region to the a zone can be observed The circularly shaped bu grains growing in the amorphous matrix B2 structure (austenite) Amorphous structure Figure 32 Microstructures along the irradiated zone—observation by transmission electron microscopy Two Examples of Laser-Annealed Devices (51,55) The first example is a microcantilever moving up and down, which is shown in Fig 33 The structure is machined from a sputtered-deposited Ni–Ti thin film The LASMA process is used to crystallize the material locally We have observed that after annealing, the locally affected zone of the material is expanded This means that a prestrain can be introduced during the process Therefore, an active structure can be created without The inside part is not annealed and is the funct bias spring The prestrain is realized by deforming per’s unannealed zone well above its plastic lim heating the whole structure, only the annealed r try to recover its original shape Upon cooling, element, the unannealed portion, will push back ture and open the gripper Using the LASMA pr active element and a bias spring have been intr one single piece of cold-worked material.Due to th erties, the LASMA process is a key technology for ing smaller highly integrated microdevices Flexible Structures Flexural designs are a well-established technolog sion engineering (53,54) Their monolithic struct many advantages such as freedom from wear, n bly, and smooth displacements that are contam free However, one of their disadvantages is tha are restricted to small displacements for a given stiffness in the drive direction Moreover, like eve mechanism, the force increases linearly with the ment Considering these two limitations, SMA fle improve these aspects significantly The large def available enhances the output stroke for a reason ume Moreover, due to their “plateau,” the force rate and if the structure is prestrained, a low stif ear guiding system can be imagined (55) Smart Monolithic SMA Microdevices: An Outlook for Design Methods 04 inc h) Annealed zone Table is a summary of some methods for sma (50) The local hardening mentioned in this tab exact reverse effect of the LASMA process Instea annealing, the basic idea is to harden or amor material locally In the design context, one could the LASMA process introduces an active region passive structure, whereas local hardening intr passive element within an active structure (0.079 inch) FUTURE TRENDS mm m m (0 Bias spring Figure 33 Locally annealed microcantilever: realization and design As previously mentioned, scaling down things new design approaches This rule applies to shape alloy microdevices Successful breakthroughs h Figure 34 Locally annealed microgripper (a) Design principle; (b) the microgripper and a cylindrical lens on a match head made in this field of research, but there is still a further need to understand the behavior of small devices on the microscale It is well known that the mechanical behavior of a thin film of a few microns is not the same as the behavior of a sheet that is a few hundred microns thick As W Nix mentioned (56), thin film materials are, for instance, much stronger than their bulk counterparts This may be due to the fine grain sizes commonly found in thin films, but single crystal thin films are also much stronger than bulk materials Many issues remain unresolved or little addressed for SMA thin films: r How thin can a functional SMA thin film be? In other words, where is the size limit of the shape-memory effect? r If a micron-size structure is built, how smooth will the motion be? Will we observe a discontinuous effect? r How does the scaling law apply to fatigue properties? Magnetostriction is the spontaneous deformation of a material caused by a change in its state of magnetization Some materials have both a martensitic transformation and a Curie temperature (ferromagnetic martensites) Depending on which phase has larger saturation magnetization, a field applied either above or martensitic transformation temperature will transformation between austenite and martens temperature is below the martensitic transform martensitic variants can be redistributed by a field (57) These materials might be a solution the main drawback of nonmagnetic SMAs, their time Considerable effort is still needed to intro material in applications The design of a controller for a SMA actuato discussed in this article Nevertheless, a lot of been done within the robotics community Mos have proposed PID controllers or close cousins ( control schemes use electrical resistance as th variable in the feedback loop, as proposed by Hi (18) Other methods have been proposed by G who studied control using a Preisach represent by Grant and Hayward (44) who designed a varia ture control that consists of switching between a put signals in response to an input signal Ho of the proposed methods have been applied only SMA actuator design consisting of a wire, strip, working against a bias spring For instance, the further room for controlling very complex design antagonistic designs Table A Summary of Design Methods for Smart Monolithic Designsa Design Methods Material State Prestressed Advantages Limitations TWSME Annealed No Actuation methods, low volume required Low force in one direction, training process Local heating Annealed Yes No local treatment, simple method Low stiffness of martensite, special heating path “Push-pull” Annealed Yes High range of motion Volume required Local hardeningb Annealed No Different mechanical hardening Miniaturization LASMA Amorph Work-Hardened No Yes Highly integrated design, different annealing conditions Managing the annealing process a b Ref 50 Local hardening is the exact reverse effect of LASMA applied to a cold-rolled sheet ...Ph 10 1 Frequency (Hz) Undamaged 1/ 4 loss for k1 10 ? ?1 10−2 10 2 10 −3 10 10 2 10 1 −0.5 ? ?1 ? ?1. 5 −2 −2.5 −3 −3.5 10 0 −0 Phase (rad) −6 ? ?8 10 1 Frequency (Hz) (f) 10 0 Undamaged 1/ 4 loss for k1 10 −5... Phase (rad) 10 0 ? ?1 −2 −3 −4 −5 −6 −7 10 0 10 1 Frequency (Hz) 10 1 Frequency (Hz) 10 2 Undama 1/ 4 loss 10 1 Frequency (Hz) 10 0 Undam 1/ 4 loss 10 ? ?1 10−2 10 −3 10 10 2 Undamaged 1/ 4 loss for k1 10 1 Frequency... k3 −6 ? ?8 10 1 Frequency (Hz) (f) 10 0 10 −5 10 0 ? ?1 −2 −3 −4 −5 −6 −7 10 0 −3 −4 10 0 10 2 Undamaged 1/ 4 loss for k3 10 1 Frequency (Hz) 10 0 10 ? ?1 10−2 10 −3 10 10 2 −0.5 ? ?1 ? ?1. 5 −2 −2.5 −3 −3.5 10 0 Undamaged