Encyclopedia of Smart Materials (Vols 1 and 2) - M. Schwartz (2002) Episode 2 potx

Proposed Classification System for Smart Materials and Systems Category Fundamental Material Characteristics Fundamental System Behaviors Traditional materials: Materials have given prop

Leadingedge

TrailingedgeDynamic pressure

Active aeroelastic wing: Blending ofleading and trailing edge effectiveness

Figure 22 Achievable roll performance by combining leading

and trailing edges.

for example, by an all-movable aerodynamic surface that

has adaptive rotational attachment stiffness This also

pro-vides high effectiveness at low speeds, and excessive loads

from diverging components or flutter instabilities at high

speeds can be avoided

The usable aeroelastic effectiveness for conventional

concepts is rather limited between take-off and cruise

speed Aileron reversal usually occurs between the cruise

speed and limit speed, and too high an effectiveness of

lead-ing edge surfaces must be avoided at the limit speed On

the other hand, adaptive all-movable concepts can provide

high effectiveness at all speeds and avoid excessive loads at

the high end of the speed envelope, as indicated in Figs 23

and 24 This means, for example, that a stabilizer surface

can be built smaller than would be required by “rigid”

aero-dynamic low-speed performance

NEED FOR ANALYZING AND OPTIMIZING THE DESIGN

OF ACTIVE STRUCTURAL CONCEPTS

Of course, active materials and structural components,

together with the stimulating forces, need a correct

Active aeroelastic conceptsRange of aeroelastic

1.0

Rigid aircraft

Figure 23 Aeroelastic effectiveness of conventional and

adaptive-all-movable active aeroelastic concepts.

For conventional activeaeroelastic concepts

Usable range of aeroelastic effectiveness byflight envelope

For advanced activeaeroelastic concept

Dynamic pressure Dynamic pressure

Figure 24 Usable range of aeroelastic effectiveness for

conven-tional and advanced active aeroelastic concepts.

description in theoretical structural or multidisciplinaryanalysis and optimization (MDAO) models and methods.Once this is provided, the actively deforming structureneeds another approach for static aeroelastic analysis Thedeflections of selected control surfaces of an aircraft thathas conventional control surfaces can be predescribed foraeroelastic analysis For an actively deformed structure,initial deformations without external loads first need to bedetermined, for example, by static analysis

As described before, the deformations achievable in junction with the distribution of external aerodynamicloads are essential for the effectiveness of active structuralconcepts for aircraft control This requires efficient toolsand methods for simultaneous, multidisciplinary analyti-cal design The best design involves optimizing

con-rexternal shape,

rarranging the passive structure (topology),

rsizing the passive structure,

rplacing and sizing the active elements, and

ra control concept for the active components.

The aims of this approach are the optimum result for theobjective function (minimum weight, aerodynamic perfor-mance), fulfillment of all constraints like strength, and alsooptimization of additional objectives, such as minimum en-ergy As depicted in Fig 25 for the optimization of a passivestructure that has different constraints for the requiredrolling moment effectiveness, the energy required to actu-ate the control surface can be considerably reduced, even

if the required (low) roll rate is already met

MDO does not mean combining single discipline lytic tools by formal computing processes It means first agood understanding of what is going on This is essentialfor a conventional design Only from this understandingcan the creative design of an active concept start

ana-It is also very important to choose the proper analyticmethods for individual disciplines Usually, not the high-est level of accuracy is suitable for the simulation of impor-tant effects for other disciplines This also refers to refin-ing the analytic models, where local details are usually not

Structural weight

Rollingmoment

effectiveness

Baselinestaticdesign1.0

Aileronhinge moment [kNm]

Rigid

50.0

10.0

Figure 25 Optimization of the rolling moment and hinge

mo-ment of a trailing edge aileron of a low aspect ratio fighter wing.

interesting for interactions It is more important to keep

the models as versatile as possible for changes in the

de-sign concepts and to allow the simulation as many variants

as possible This also means an efficient process for

gener-ating models, including the knowledge of the user for this

process Fully automated model generators can create

ter-rible results, if the user cannot interpret or understand the

modeling process

Any improvement in a technical system is often referred

to as an optimization In structural design today, this

ex-pression is mainly used for formal analytic and numerical

methods Some years after the introduction of finite

ele-ment methods (FEM) for analyzing aircraft structures, the

first attempts were made to use these tools in an automated

design process Although the structural weight is usually

used as the objective function for optimization, the major

advantage of these tools is the fulfillment of aeroelastic

con-straints, not the weight saving Other than static strength

requirements, which can be met by adjusting the

dimen-sions of individual finite elements, the sensitivities of the

elements to aeroelastic constraints cannot be expressed so

easily

In the world of aerodynamics, the design of the required

twist and camber distribution for a desired lift at minimum

drag is also an optimization task Assuming that minimum

drag is achieved by an elliptical lift distribution along the

wingspan, this task can be solved by a closed formal

so-lution and potential flow theory More sophisticated

nu-merical methods are required for the 2-D airfoil design or

for Euler and Navier–Stokes CFD methods, which are now

maturing for practical use in aircraft design

Formal optimization methods have been used for

con-ceptual aircraft design for many years Here, quantities

such as direct operating costs (DOC) can be expressed by

rather simple equations, and the structural weight can be

derived from empirical data Formal methods such as

op-timum control theory are also available for designing the

flight control system

So, one might think that these individual optimization

tasks could easily be combined into one global aircraft

op-timization process The reasons that this task is not so

simple is the different natures of the design variables of dividual disciplines and their cross sensitivities with otherdisciplines The expression multidisciplinary optimization(MDO) summarizes all activities in this area, which haveintensified in recent years It must be admitted that todaymost existing tools and methods in this area are still singlediscipline optimization tasks that have multidisciplinaryconstraints

in-To design and analyze active aeroelastic aircraft cepts, especially when they are based on active materials

con-or other active structural members, new quantities are quired to describe their interaction with the structure, theflight control system, and the resulting aeroelastic effects

re-SUMMARY, CONCLUSIONS, AND PREDICTIONS

In the same way as it was wrong in the past to demandthat an aircraft design to be as rigid as possible, it’s wrongnow to demand a design that is as flexible as possible

It is sometimes said that smart structural concepts cancompletely replace conventional control surfaces But thislooks very unrealistic, at least at the moment The majordifficulties for successful application are the limited defor-mation capacity of active materials, as well as their strainallowables, which are usually below those of the passivestructure However, this can be resolved by proper design

of the interface between the passive and active structures.But the essential difficulties are the stiffness and strainlimitations of the passive structure itself It cannot be ex-pected that the material of the passive structure just needs

to be replaced by more flexible materials without an sive weight penalty It is also not correct to believe that

exces-an active aeroelastic concept becomes more effective, if theflexibility of the structure is increased Aeroelastic effec-tiveness depends on proper aeroelastic design, which needscertain rigidity of a structure to produce the desired loads

A very flexible structure would also not be desirable fromthe standpoints of aerodynamic shape, stability of the flightcontrol system, and transmission of static loads

Because large control surface deflections are required

at low speeds, where aeroelastic effects on a fixed surfaceare small, it is more realistic to use conventional controlsurfaces for this part of the flight envelope and use activeaeroelastic deformations only at higher speeds This wouldstill save weight on the control surfaces and their actua-tion system due to the reduced loads and actuation powerrequirements

To produce usable deformations of the structure also atlow speeds, all-movable aerodynamic surfaces that have

a variable attachment stiffness are an interesting option.This concept relies on development efforts for active de-vices that have a wide range of adjustable stiffness.The reasons that we have not seen more progress to date

in successfully demonstrating smart structural concepts inaeronautics may be that

rspecialists in aircraft design do not know enough

about the achievements in the area of smart rials and structures, and

mate-rsmart materials and actuation system specialists,

who try to find and demonstrate applications in

aeronautics, do not know or care enough about world conditions for airplane structures.

real-What we need is more awareness on both sides, as well

as stronger efforts to learn from each other and work

together

Although there are strong doubts about useful

applica-tions of smart structures for aircraft control, it should

al-ways be remembered how often leading experts have been

wrong in the past in their predictions, in many cases even

on their own inventions Norman R Augustine quotes some

of them in his famous book “Augustine’s Laws” (46):

r“The [flying] machines will eventually be fast; they

will be used in sport but they should not be thought of

as commercial carriers.” – Octave Chanute, aviationpioneer, 1910

r“The energy produced by the breaking down of the

atom is a very poor kind of thing Anyone who expects

a source of power from the transformation of theseatoms is talking moonshine – Ernest Rutherford,physicist, ca 1910

r“Fooling around with alternative currents is just a

waste of time Nobody will use it, ever It’s too gerous it could kill a man as quick as a bolt of

dan-lightning Direct current is safe.” – Thomas Edison,inventor, ca 1880

Also quoted by Augustine (46), the eminent scientist Niels

Bohr remarked: “Prediction is very difficult, especially

about the future.”

At the moment it looks more realistic that new hybrid,

concentrated active devices, positioned between a passive

but properly aeroelastically tailored main aerodynamic

surface and the corresponding control surfaces are showing

the like Hopefully this article will inspire useful

applica-tions of smart structures and prevent some unnecessary

3 W Schwipps, Schwerer als Luft—die Fr ¨ uhzeit der

Flugtech-nik in Deutschland Bernhard & Graefe Verlag, Koblenz,

Germany, 1984.

4 M.I Woods, J.F Henderson, and G.D Lock, Aeronaut J.

(2001).

5 E Pendleton, B Sanders, P Flick, and O Sensburg Int Forum

Aeroelasticity Struct Dynamics, Madrid, Spain, 2001.

6 J.J Bertin and M.L Smith, Aerodynamics for Engineers.

Prentice-Hall, Englewood Cliffs, NJ,1979.

7 O Sensburg, H H¨onlinger, T.E Noll, L.J Huttsell, J Aircraft

19 (5): 1981.

8 C.R Larson, E Falanges, S.K Dobbs, SPIEs 5th Annu Int.

Symp Smart Struct Mater., San Diego, CA, 1998.

9 C.R Crowe and J.M Sater, AGARD Symp Future Aerosp.

Technol Serv Alliance, Paris, 1997, CP-600, Vol I.

10 A.R McGowan, J Heeg, and R.C Lake, Proc 37th AIAA/ ASME/ASCE/AHS/ASC Struct Struct Dynamics Mater Conf., Salt Lake City, UT, 1996.

11 O.L Skopik, Wie berechnet und konstruiert man selbst ein Flugzeug Vienna, Austria, 1917.

12 C.M Christensen, The Innovator’s Dilemma Harvard

Busi-ness School Press, Boston, 1997.

13 H.I Flomenhoft, The Revolution in Structural Dynamics

Dy-naflow Press, Palm Beach Gardens, FL, 1997.

14 A.R Collar, J R Aeronaut Soc 613–636, August 1946.

15 J Schweiger, J Simpson, F Weiss, E Coetzee, and Ch Boller,

SPIEs 6th Annu Int Symp Smart Structures Mater., Newport

Beach, CA, 1999.

16 T.A Weisshaar, RTO Meet Proc 36 (AC/323(AVT)TP/17) from the RTO AVT Specialists’ Meet Struct Aspects Flexible Aircraft Control, Ottawa, Canada, October 1999.

17 E.W Pendleton, D Bessette, P.B Field, G.D Miller, K.E.

Griffin, and J Aircraft 37(4), (2000).

18 J Aircraft 32(1), Special Section (1995).

19 O Sensburg, G.; Schneider, V.; Tischler, and V Venkayya, cialists’ Meet Struct Aspects Flexible Aircraft Control RTA Meet Design Issues, Ottawa, Canada; October 1999.

Spe-20 A.H.G Fokker and B Gould, Flying Dutchman The Life of Anthony Fokker Henry Holt, NY, year unknown.

21 T Noll, H H¨onlinger, O Sensburg, and K Schmidt, Active Flutter Suppression Design and Test, a Joint U:S.-F.R.G Pro- gram ICAS- CP 80-5.5 Munich, Germany, 1980.

22 K.B Lazarus, E Saarmaa, and G.S Agnes, Active Smart terial System for Buffet Load Alleviation Bellingham, WA,

Ma-1995, SPE Vol 22447/179.

23 R.M Hauch, J.H Jacobs, K Ravindra, and C Dima, Reduction

of Vertical Tail Buffet Response Using Active Control CP-95-1080, Washington, 1995.

AIAA-24 R.W Moses, 40th AIAA-SDM Con., St Louis, 1999.

25 J Simpson and J Schweiger, SPIE 5th Annu Int Symp Smart Struct and Mater., San Diego, CA, 1998.

26 T.A Weisshaar, J Aircraft 17(6), (1980).

27 J Schweiger, J Krammer, and H H¨ornlein, 6th AIAA/NASA/ ISSMO Symp Multidisciplinary Anal Optimization, Seattle,

WA, 1996, CP-4169.

28 N Khot, F Eastep, and R Kolonay, A Method for ment of the Rolling Maneuver of a Flexible Wing AIAA-CP-96- 1361.

Enhance-29 K Griffin and M Hopkins, 36th AIAA Struct Struct mics Mater Conf., New Orleans, 1995.

Dyna-30 C.L Giese, G.W Reich, M.A Hopkins, and K.E Griffin, An Investigation of the Aeroelastic Tailoring for Smart Structures Concept AIAA-96-1575-CP, 1996.

31 A.R McGowan, L.G Horta, J.S Harrison, and D.L Raney,

RTO Meet Proc 36 (AC/323(AVT)TP/17) RTO AVT ists’ Meet Struct Aspects Flexible Aircraft Control, Ottawa,

Special-Canada, October 1999.

32 S.L Padula, J.L Rogers, and D.L Raney,

Multidisci-plinary Techniques and Novel Aircraft Control Systems 8th AIAA/NASA/ISSMO Symp Multidisciplinary Anal Opti- mization, Long Beach, CA, 2000, AIAA-CP-2000-4848.

33 J Aircraft 32(1), Special Section (1995).

34 E Pendleton, K.E Griffin, M.W Kehoe, and B Perry, Conf Proc AIAA-96-1574-CP.

35 P.M Flick and M.H Love, Specialists’ Meet Struct Aspects Flexible Aircraft Control, RTA Meet Design Issues, Ottawa,

Canada, October 1999.

36 S.I Kuzmina, G.A Amiryants, F.Z Ishmuratov, V.A Mosunov,

and V.V Chedrik, Int Forum Aeroelasticity Struct Dynamics,

Madrid, Spain, 2001.

37 J Schweiger and O Sensburg, Int Forum Aeroelasticity

Struct Dynamics, Madrid, Spain, 2001.

38 R.W Moses, CP, 40th AIAA-SDM Conf., St Louis, 1999.

39 J Simpson and J Schweiger, Industrial Approach to

Piezo-electric Damping of Large Fighter Aircraft Components SPIE

6th Annu Int Symp Smart Struct Mater., San Diego, CA,

1998.

40 A.R McGowan, SPIE 5th Annu Int Symp Smart Struct.

Mater., Newport Beach, CA, 1999.

41 T.A Weisshaar, 41st Annu Isr Conf Aerosp Sci., Tel Aviv and

Haifa, February 2001.

42 R Barrett, in Smart Materials and Structures No 5, IOP, UK,

1996.

43 P Costa, P.A Moniz, and A Suleman, 42nd AIAA SDM Conf.,

Seattle, WA, 2001, AIAA-2001-1361.

44 H.P Monner, E Breitbach, Th Bein, and H Hanselka,

Aero-naut J 104(1032), (2000).

45 J Schweiger, F Weiss, T Kullrich, 8th AIAA/USAF/NASA/

ISSMO Symp Multidisciplinary Anal Optimization, Long

Beach, CA, September 2000.

46 N.R Augustine, Augustine’s Laws AIAA, Reston, VA, 1997.

An inextricable link has existed historically between

a building’s characteristics—form, appearance, and

function—and the characteristics of the different

materi-als that were available and suitable for construction As

exemplified by historical building traditions in stone and

wood, early architects sought to understand intuitively

the intrinsic physical behavior of commonly available

materials to exploit their properties in designing and

constructing buildings Conversely, later innovations in

the type and availability of materials strongly impacted

the development of new architectural forms as architects

began to respond to changing societal demands and new

building functions emerged This trend is illustrated by

the development of steel in the nineteenth century and

the related emergence of long-span and high-rise building

forms Today, architects are beginning to look forward to

using the developments in smart materials to bring new

solutions to long-standing problems and also to exploit the

potential of smart materials in developing new building

functions, forms, and responses The wide variety of smart

materials available has great potential for use within the

field, but, in this area, their applications remain only

marginally explored

MATERIAL CONSIDERATIONS IN ARCHITECTURE

Unlike materials used for specific applications or productssuch as in refractory linings or engine blocks that arefundamentally chosen on the basis of performance crite-ria and cost, the choice of materials for architectural usehas always been based on very different types of criteria.Performance and cost obviously play a role, but the finalselection is often based on appearance and aesthetics, ease

of constructability in terms of labor skill, local or regionalavailability, as well as the material used in nearby exist-ing buildings The multimodal nature of the selection pro-cess coupled with the wide-ranging array of building types,uses, and locales has resulted in a material palette that en-compasses all of the major material classes

TRADITIONAL MATERIAL CLASSIFICATIONS

IN ARCHITECTURE

The Construction Specification Institute (or CSI) devised

a classification system in 1948 that is used throughoutthe architectural design and building construction indus-tries The classification system is bipartite: the first half

is devoted to the broad classes of materials typically used

in buildings, including paint, laminate, and concrete, andthe second half categorizes standard building componentssuch as doors, windows, and insulation The emphasis inboth major groupings is on application, not on fundamentalbehavior or properties For example, in Division 6 the char-acteristics of wood are discussed in relationship to theirrelevance to the intended application: the grade of woodsuitable for load-bearing roof structures or the type of woodsuitable for finish flooring

The CSI index serves as a template for communicationamong architects, contractors, fabricators, and suppliers.After the preliminary design of a building is completedand approved, architects prepare construction documents(known as CDs) that will serve as the “instructions” forconstructing the building Accompanying each set of CDsare the “Construction Specifications”: a textual documentthat defines each building element documented in the CDsand specifies the material or component The ConstructionSpecifications serve as a binding contract that construc-tion professionals and contractors must follow Trade asso-ciations and manufacturers of building products routinelywrite their material and product specifications in CSI for-mat to streamline the specification process for architects,and many architectural firms maintain an internal set ofConstruction Specifications that is used as the baseline forall of their projects

TRADITIONAL TECHNOLOGY CLASSIFICATION

IN ARCHITECTURE

The CSI index also categorizes the technologies used inarchitectural design and construction Unlike the standardtechnology classifications used in engineering sciencesthat categorize according to process and product, theCSI specifications categorize by system As in the CSI

material classes, the focus of the technology classes is

also on application The technologies are divided into

two major groups: the first is devoted to building

op-erational systems such as HVAC, lighting, and

plumb-ing systems, and the second is devoted to buildplumb-ing

con-struction systems such as structural, drainage, and

ver-tical circulation systems The specifications for the

build-ing operational systems are almost entirely supplied by

manufacturers

PROPOSED CLASSIFICATION SYSTEM

FOR SMART MATERIALS

The introduction of smart materials into architecture poses

a challenge to the normative classification system A smart

material may be considered as a replacement for a

con-ventional material in many components and applications,

but most smart materials have inherent “active” behaviors,

and, as such, are also potentially applicable as

technolo-gies For example, electrochromic glass can be

simultane-ously a glazing material, a window, a curtain wall system, a

lighting control system, or an automated shading system

The product would then fall into many separate categories,

rendering it particularly difficult for the architect to take

into consideration the multimodal character and

perfor-mance of the material Furthermore, many smart

mate-rials are introducing unprecedented technologies into the

field of design, and are also making more commonplace

many technologies, such as sensors, which previously had

only limited application in highly specialized functions

Table 1 describes a proposed organization in which smart

materials establish a sequential relationship between

ma-terials and technologies The proposed organization also

maintains the fundamental focus on application of the

traditional classification system

Table 1 Proposed Classification System for Smart Materials and Systems

Category Fundamental Material Characteristics Fundamental System Behaviors

Traditional materials: Materials have given properties Materials have no or limited

Natural materials (stone, wood) and are “acted upon” intrinsic active response

High performance materials: Material properties are designed

Polymers, composites for specific purposes

Property-changing and energy-exchanging respond intelligently to varying responses to external stimuli and materials external conditions or stimuli can serve as sensors and actuators

Intelligent components: Behaviors are designed to Complex behaviors can be

Smart assemblies, polyvalent walls respond intelligently to varying designed to respond intelligently

external conditions or stimuli in and directly to multimodal demands discrete locations

Intelligent environments Environments have designed Intelligent environments consist

interactive behaviors and of complex assemblies that often intelligent response—materials combine traditional materials and systems “act upon” the with smart materials and

characteristics are enabled via a computational domain

TAXONOMY OF SMART MATERIALS

Four fundamental characteristics are particularly relevant

in distinguishing a smart material from the traditionalmaterials used in architecture: (1) capability of propertychange (2) capability for energy exchange, (3) discretesize/location, and (4) reversibility These characteristicscan potentially be exploited either to optimize a materialproperty to match transient input conditions better or tooptimize certain behaviors to maintain steady-state condi-tions in the environment

Smart Material Characteristics Property Change The class of smart materials that has

the greatest volume of potential applications in ture is the property-changing class These materials un-dergo a change in a property or properties—chemical,thermal, mechanical, magnetic, optical, or electrical—inresponse to a change in the conditions of the material’senvironment The conditions of the environment may beambient or may be produced via a direct energy input In-cluded in this class are all color-changing materials, such

architec-as thermochromics, electrochromics, and photochromics,

in which the intrinsic surface property of the molecularspectral absorptivity of visible electromagnetic radiation

is modified by an environmental change (incident solarradiation, surface temperature) or an energy input to thematerial (current, voltage)

Energy Exchange The next class of materials predicted

to have a large penetration into architecture is the exchanging class These materials, which can also be called

energy-“first law” materials, change an input energy into other form to produce an output energy in accordancewith the first law of thermodynamics Although the energy

an-converting efficiency of smart materials such as

photo-voltaics and thermoelectrics is typically much less than

those of conventional energy conversion technologies, the

potential utility of the energy is much greater For

exam-ple, the direct relationship between input energy and

out-put energy renders many of the energy-exchanging smart

materials, including piezoelectrics, pyroelectrics and

pho-tovoltaics, excellent environmental sensors The form of

the output energy can further add direct actuating

capa-bilities such as those currently demonstrated by

electrore-strictives, chemoluminescents and conducting polymers

Reversibility/Directionality Some of the materials in the

two previous classes also exhibit the characteristic of

ei-ther reversibility or bidirectionality Many of the electricity

converting materials can reverse their input and output

energy forms For example, some piezoelectric materials

can produce a current from an applied strain or can

de-form from an applied current Materials that have a

bi-directional property change or energy-exchange

behav-ior can often allow further exploitation of their transient

change rather than only of the input and output energies

and/or properties The energy absorption characteristics of

phase changing materials can be used either to stabilize an

environment or to release energy to the environment,

de-pending on the direction in which the phase change is

tak-ing place The bidirectional nature of shape-memory alloys

can be exploited to produce multiple or switchable outputs,

allowing the material to replace components composed of

many parts

Size/Location Regardless of the class of smart material,

one of the most fundamental characteristics that

differen-tiates smart materials from traditional materials is the

discrete size and direct action of the material The

elimi-nation or reduction in secondary transduction networks,

additional components, and, in some cases, even packaging

and power connections allows minimizing the size of the

active part of the material A component or element

com-posed of a smart material can be much smaller than a

simi-lar construction using traditional materials and also will

require less infrastructural support The resulting

compo-nent can then be deployed in the most efficacious location

The smaller size coupled with the directness of the

prop-erty change or energy exchange renders these materials

particularly effective as sensors: they are less likely to

in-terfere with the environment that they are measuring, and

they are less likely to require calibration

Relevant Properties and Behaviors

Architectural materials are generally deployed in very

large quantities, and building systems tend to be highly

integrated into the building to maintain homogeneous

in-terior conditions Materials and systems must also

with-stand very large ranges of transient exterior conditions

The combination of these two general requirements tends

to result in buildings of high thermal and mechanical

in-ertia Therefore, even though the typical building uses

several different materials for many functions, there areonly a few areas in which the characteristics of smart mate-rials can be useful The transient environmental conditionsexperienced by most buildings often results in oversizingsystems to accommodate the full range of the exterior en-vironmental swing The swings may be instantaneous, as

in the case of wind, diurnal, or seasonal These conditionsinclude those that affect both heat transfer and daylighttransmission through the building envelope (also known

as the building fac¸ade or exterior skin) as well as those thatcreate dynamic loading on the building’s structural supportsystem For the building envelope, the property-changingclass of smart materials has the most potential application,whereas the energy-exchanging class is already finding ap-plication in building structural systems

Buildings consume two-thirds of the electrical energygenerated in the United States, and the majority of thatelectrical energy is used to support the building’s ambi-ent environmental systems, primarily lighting and HVAC(heating, ventilating, and air conditioning) systems Theintent of these systems is to effect a desired state in theinterior That state may be defined by a specified illumi-nance level or by an optimum temperature and relativehumidity Because conditions are generally maintained at

a steady state, the primary need is for more efficaciouscontrol Energy-exchanging materials have potential ap-plication as discrete sources, particularly for lighting deliv-ery systems, and also as secondary energy supply sources.The most significant applications of smart materials inbuildings, however, has been and will continue to be assensors and actuators for the control systems of these am-bient environmental systems

Smart Material Mapping

The material properties and/or characteristics that aremost relevant to architectural requirements are mapped inTable 2 against examples of smart material applications

CATEGORIES OF APPLICATIONS

One of the major difficulties in incorporating smart rials into architectural design is the recognition that veryfew materials and systems are under single environmen-tal influences For example, the use of a smart material tocontrol conductive heat transfer through the building en-velope may adversely impact daylight transmission Fur-thermore, because most systems in a building are highlyintegrated, it is difficult to optimize performance withoutimpacting the other systems or disrupting control systembalancing As an example, many ambient lighting systemsinclude plenum returns through the luminaires (lightingfixtures) that make it particularly difficult to decoupleHVAC from lighting systems The following discussion es-tablishes four major categories of applications for smartmaterials and takes into account the material/behaviormapping described in Table 2 but also considers the com-plex systems that are affected The four categories—glazing materials, lighting systems, energy systems, and

mate-Table 2 Mapping of Smart Materials to Architectural Needs

Architectural Need Relevant Material Characteristic Smart Material Application Control of solar radiation Spectral absorptivity/transmission Electrochromics

transmitting through the building of envelope material Photochromics

Suspended particle panels Relative position of envelope material Louver control systems

r exterior radiation sensors

transfer through the building envelope material Phase change materials Control of interior heat generation Heat capacity of interior material Phase change materials

Relative location of heat source Fiber-optic systems

Thermoelectrics Lumen/watt energy conversion ratio Photoluminescents

Light-emitting diodes Secondary energy supply systems Conversion of ambient energy to Photovoltaics

electrical energy

Illuminance measurements Photoelectrics Occupancy sensing

Relative location of source Fiber optics

Electroluminescents

CO 2 and chemical detection Biosensors Relative location of source Thermoelectrics

Electrorheological Shape-memory alloys

monitoring/control systems—are also intended to be

con-sistent with the more normative and identifiable

classifi-cation systems of architecture

Glazing Materials

Whether serving as windows or as glass curtain walls,

glazing materials are extensively used on the building

en-velope Originally incorporated and developed during the

twentieth century for aesthetic reasons, the current use

of glazing materials also considers the delivery of daylight

into the building’s interior The majority of developments in

high-performance glazing materials have focused on

ther-mal characteristics—spectral selectivity to reduce radiant

transmission to the interior or low emissivity to reduce

ra-diant loss to the exterior Glazing introduces the

problem-atic condition in which, depending on the exterior

envi-ronmental conditions, performance criteria that have been

optimized for one set of conditions may be undesirable in

a matter of hours or even moments later The ideal

glaz-ing material would be switchable—managglaz-ing the radiant

transmission between exterior to interior to transmit

so-lar radiation when the envelope is conducting heat out

(typical winter daytime condition) and reflect solar ation when the envelope is conducting heat into the build-ing (typical summer daytime condition) Photochromics,thermochromics and thermotropics have been proposed asswitchable glazing materials, although only thermotropicsare currently being developed commercially for this appli-cation The basic operation of these materials is that ei-ther high incident solar radiation (photochromic) or highexterior temperature (thermochromic or thermotropic)produces a property change in the material that increasesits opacity, thereby reducing radiant transmission to theinterior When incident solar radiation lessens or when theexterior temperature drops, the material reverts to a moretransparent quality, allowing more solar radiation to trans-mit to the interior

radi-There are numerous circumstances, however, for whichthis type of switching is neither desirable nor useful Di-rect solar radiation into the building can create over-heated zones in particular locations, even in the dead ofwinter Winter sun altitude is also much lower, therebysignificantly increasing the potential for glare if solarradiation is not controlled During the summer, reduc-ing the radiant transmission may increase the need for

Human perceptionsand actions

External stimuli(Light level)

Direct user controle.g., switches

Liquid crystalfilm

Laminations(Film laminated betweenglass layers)

Sensor control(Light level sensor)

Interface

Building enclosure element (wall)with controllable transparency

Enablingtechnologies

Figure 1 Typical current use of a smart material in architecture Only a single behavior is

controlled.

interior lighting systems and, because all electrically

gen-erated light has a lower lumen/watt ratio than daylight,

might exacerbate the building’s internal heat gains As

a result, the majority of efforts to develop smart glazing

have focused on the electrically activated chromogenics—

electrochromics, liquid crystal panels, and suspended

particle panels (see Fig 1) By using an electrical

in-put to control transparency, these materials can be more

easily incorporated into the control schemes for energy

management systems and/or lighting control systems The

optimum balance among lighting needs, heating/cooling

re-quirements, and occupant comfort can be determined, and

the transparency can be adjusted to meet these demands

in highly transient conditions

Lighting Systems

Most high efficiency lighting systems—fluorescent, HID

(high intensity discharge)—are relatively unsuitable for

low-level lighting or task lighting Furthermore, the typical

ambient lighting system requires enormous infrastructure

for support: electronic control systems, ballasts, integrated

cooling, light diffusers/distributors (often part of the

lumi-naire or lighting fixture) The efficiency and economics of

these systems drop as the overall lighting requirements

be-come smaller or more discrete Ambient systems are also

difficult to dim and to focus, so that very low-efficiency

in-candescent/halogen systems are still widely used for task

or discrete lighting requirements The low efficiency of the

typical lighting system results in producing a substantial

amount of heat and can be responsible for as much as 30%

of a commercial building’s cooling load The development

of fiber-optic lighting systems allows decoupling the

deliv-ered light from the primary energy conversion processes for

generating light This has the dual advantage of allowing

light delivery to any location in a building, which is much

more efficacious than using ambient lighting systems to

de-liver light, as well as removing the heat source from the

oc-cupied space Current applications for fiber-optic systems

include many museums and retail display areas, where theremoval of the heat source can profoundly improve the en-vironmental conditions of the objects under display and thediscrete nature of the light allows better highlighting andfocusing

Ambient lighting systems are generally designed to vide a standard illuminance level throughout a space at aspecified height (usually three feet above the floor) Thehuman eye, however, responds to the relative luminancecontrast between surfaces in the field of vision A light-ing level of 100 footcandles may be too low for reading ifthe surrounding surfaces provide little contrast and may

pro-be too high if the surfaces provide high contrast The vision of light into smaller and more discrete sources al-lows optimizing contrast within the field of vision Fur-thermore, the design of lighting for managing contrastenables using lower levels of lighting Sources produced

di-by the various luminescents—chemo, photo, electro—arestarting to find application in architectural interiors, par-ticularly as emergency lighting systems, because they havelow and in some cases no input power requirements LED(light-emitting diode) systems are also being developed aslow energy lighting delivery systems The latest develop-ments in polymer LED technology have produced lightingfixtures that have precise color control They provide ex-cellent color rendition and also allow for color variation—features that are difficult to achieve in standard lightingsystems

Energy Systems

The majority of buildings in the United States are nected to a utility grid and as such have little need forprimary energy conversion on-site There are numerouscircumstances, however, where secondary energy conver-sion can be quite useful, including back-up power genera-tion, peak demand control, and discrete power for remoteneeds For these situations, photovoltaic energy systemsare increasingly becoming popular because they can be

con-readily deployed on roofs or integrated directly into the

building envelope to take advantage of the incident

so-lar radiation Two other developments in smart materials

hold greater promise for managing energy needs within

a building The large interior heat loads of most

build-ings coupled with a diurnal exterior temperature swing

has encouraged investigation into thermal mass systems

for maximum exploitation of a building’s thermal

iner-tia Although theoretically sound, thermal mass systems

have three major problems: (1) very slow response time,

(2) the inability to switch off the phenomenon when it

is not desirable, and (3) the large embodied energy

re-quired to provide the necessary mass of material Phase

change materials offer the advantages of thermal mass

and very few of its disadvantages The materials can be

tuned to particular temperatures and can have very rapid

responses Much less mass is required, and therefore, the

materials can be packaged and distributed throughout the

building much more efficiently and strategically By

lay-ering phase change materials and other smart

materi-als, such as electrochromics or thermotropics, there may

be a potential to add switching capability that allows

ac-tivating or deacac-tivating of the inertial behavior of the

materials

The removal of heat generated in a building is

becom-ing an increasbecom-ing concern as point loads from lightbecom-ing,

computers, and other electrical equipment escalate

Am-bient HVAC systems do not distinguish between

human-generated and equipment-human-generated cooling needs The

ability to manage and remove the heat generated by a

point load without affecting the ambient environmental

system could improve the operation of the ambient system

and significantly reduce the energy requirements

Ther-moelectrics are currently being explored for their potential

to manage point loads discretely Already serving as heat

sinks in the majority of microprocessor cooling packages,

thermoelectrics could be incorporated into integrated

cool-ing for many other types of point sources Although the

devices are not practical for cooling air directly because of

their low coefficient of performance (COP), they are ideal

for managing the conjugate heat transfer that is

charac-teristic of most nonhuman heat sources encountered in a

building

Monitoring and Control Systems

The increasing push to reduce the energy used by

build-ing HVAC systems has led to tighter buildbuild-ings to reduce

infiltration and to larger resets for the control equipment

This combination of an impermeable building envelope and

more variable interior conditions has led to an increase

in occupant complaints and indoor air quality problems

Many of the strategies intended to reduce energy can

im-pact human health adversely, and much discussion of the

appropriate compromise between the two requirements

continues One solution that holds promise is DCV, or

“de-mand controlled ventilation.” DCV adjusts interior

venti-lation depending on the presence of occupants; it reduces

ventilation when no occupants are in a room or zone and

increases ventilation as more occupants enter Because the

human need for fresh air is linked to activity, simple pancy sensors are not enough The level of carbon dioxide

occu-in a room has been proposed as a good surrogate for theamount of fresh air needed in a space, but many concernshave arisen in regard to other chemical contamination,such as finish material outgassing, that is not connected

to occupancy Chemical sensing for building monitoringhas previously been too expensive to incorporate and tooslow to be useful New developments in smart sensors forenvironmental monitoring, particularly biosensors, holdgreat promise for optimizing the controls of ambient HVACsystems

The need to control various kinds of motions and,

in particular, vibrations in a structure appears in manyforms At the level of the whole building structure, ex-citations resulting from seismic or wind forces can re-sult in damage to both primary structural systems andnonstructural elements User discomfort can also result.Many pieces of delicate equipment in buildings also need

to be protected from external vibrations by using similarstrategies Alternatively, many pieces of equipment used inbuildings can produce unwanted vibrations that can prop-agate through buildings In response to these needs, meth-ods of mitigating structural damage have been proposedthat seek to control overall structural responses via con-trollable smart damping mechanisms used throughout astructure Several smart base isolation systems for miti-gating structural damage in buildings exposed to seismicexcitations have also been proposed These dampers arebased on various electro- or magnetorheological fluids orpiezoelectric phenomena Piezoelectric sensors and actua-tors, for example, have been tested for use in vibrationalcontrol of steel frame structures for semiconductor manu-facturing facilities

Active control can be used to modify the behavior ofspecific structural elements by stiffening or strengtheningthem Structures can adaptively modify their stiffnessproperties, so that they are either stiff or flexible as needed

In one project, microstrain sensors coupled with ceramic actuators were used to control linear buckling,thereby increasing the bucking load of the column several-fold

piezo-Several new technologies provide capabilities for age detection in structures Various kinds of optical-fibersensors have been developed for monitoring damage in ma-terials as diverse as concrete and fiber-reinforced plasticcomposite laminate structures Optical fibers are usuallyembedded in the material Strain levels can be measuredvia wavelength shifts and other techniques Crack devel-opment in structures made of concrete, for example, hasbeen monitored via optical-fiber sensors, and special dis-tributed systems have been developed for use in the struc-tural health monitoring of high-performance yachts Dis-tributed fiber-optic systems have also been proposed forleak detection in site applications involving infrastructuresystems Other site-related structural applications includeusing optical-fiber sensors for ground strain measurement

dam-in seismically active areas Other applications where smartmaterials serve as sensors include the use of embeddedtemperature sensors in carbon-fiber structures

FUTURE DESIGN APPROACHES IN ARCHITECTURE

The previous sections have outlined and discussed smart

materials in conjunction with needs currently defined

in architecture In some cases, smart materials have

been proposed as replacements for conventional

materi-als, and in other cases, smart materials have been

pro-posed for improving the functionality of standard

build-ing systems All of these developments can be

posi-tioned into the third category titled “Smart materials” in

Table 1 The impact of incorporating these materials into

standard architectural practice will be significant,

partic-ularly in regard to energy use and building performance,

but far more interesting potentials derive from

reconsider-ing smart materials as fundamental conceptual elements

in design rather than only as mprovements to existing

elements

As architectural design has always traditionally

in-volved integrated systems and materials—the building

envelope construction depends on the building’s

struc-tural system, the building’s HVAC system depends on the

envelope construction—then the greatest potential may

come from using smart materials to dis-integrate

cer-tain components, behaviors, or environments within the

building A smart component would be one that

func-tions intelligently without infrastructural support and also

will not disrupt the performance of surrounding systems

An example of a smart component might be a luminaire

that can sense relative luminances within a visual field

and self-adjust its focus, dimming capacity, and position

Activated smart materials

Liquid crystal film

Thermo-electric devices

Other

Direct action smart materials

Phase change materials

Photochromic films

Other

Sensor control

Computationalcontrol

Interface

Enablingtechnologies

Smart building enclosure (Wall)Discrete and transient control of multiple behaviors

Figure 2 Control of multiple behaviors via smart building assemblies.

Several different smart materials would be involved inthe development of this component, including sensorsand actuators, electroluminescents or LEDS, and perhapseven shape-memory alloys A smart assembly would op-erate at the next level of functionality beyond the smartcomponent There are many “high-tech” assemblies cur-rently used in architecture These assemblies integrateseveral types of components and technologies to achievemultiple functions For example, many of the most ad-vanced envelope systems incorporate mechanical shad-ing systems, thermal and ventilation control systems, andmultiple layers of glass into a highly integrated assem-bly intended to preserve view without incurring energypenalties A smart assembly would be designed to man-ifest the same behaviors, but do so in the most strate-gic manner (see Fig 2) Shading could be accomplished

at the micron or molecular level by using smart als, and thermal control could take place discretely andtransiently by selective placement of phase change ma-terials and thermoelectrics The smart assembly wouldmaximize functionality and minimize the number ofcomponents

materi-Many development activities have been focused on posals for “smart rooms” (see Fig 3) Most of these pro-posals accept the building as a traditional structure andseek to insert certain technologies into a room to add in-creased functionality Ubiquitous computing, teleconfer-encing, smart boards, voice and gesture recognition sys-tems, and wireless communication systems are among themany smart technologies being developed for incorporation

pro-Surrounding environment

Human actionsand decisions

User specifiedcontrol

Smart assemblies and devicescomprising the space

Smart enclosures

- Multiple functions

Smart environmental systems

Smart user support devices

- Appliances

- Workstations

- Other

Computationalcontrol

Interface

Sensor control

- Stimuli sensorsCriteria associated

with use of environment(e.g., work performancemeasures)

Environmentaland other stimuli

Smart devices

Sensors &

controls Controlled environment

Figure 3 Smart rooms: In the current paradigm of a smart room, new smart devices are added

to increase functionalities The controlling interface is visibly and operationally present.

User-centeredenvironment

Embeddedinterfaces

Smart rooms:

Future paradigms

Human perceptions,actions and decisions

Smart assembliesand devices

Ubiquitous embedded interfaces(Transparent to user andcomputationally-driven)

Figure 4 Smart rooms—future paradigms: The interface will disappear to the user.

into buildings A more interesting and provocative question

might be, “What would a room or building of the future be

like if we could exploit smart materials and technologies to

redesign the environment?” Smarter structures and

con-struction materials might allow significant reductions in

the size of the static building components—buildings could

become thinner, lighter, and more flexible The energy

in-tensive ambient systems in buildings could be reduced or

even eliminated if we allowed full interactivity between

the occupant and the environmental behavior (see Fig 4)

Ambient lighting systems could be replaced by discrete

sources that respond to the viewer HVAC systems could

be minimized if only the zone around an occupant were

conditioned Fundamentally, actions could be discrete and

direct—the minimum necessary at the point and time for

maximum effect

BIBLIOGRAPHY

1 D.M Addington, Boundary Layer Control of Heat Transfer in

Buildings, Harvard University Dissertation, Cambridge, MA,

1997.

2 D.M Addington, Discrete Control of Interior Environments in

Buildings, Proc ASME Fluids Eng Div., 1998.

3 E Allen, Fundamentals of Building Construction: Materials and Methods J Wiley, NY, 1999.

4 R.E Christenson and B.F Spencer, Coupled Building Control

Using Smart Damping Strategies, SPIE 7th Int Symp Smart Struct Mater., 2000.

5 J Hecht, City of Light Oxford University Press, NY, 1999.

6 N.K Khartchenko, Advanced Energy Systems Taylor &

Francis, Washington, DC, 1998.

7 R.B Peterson, Micro Thermal Engines: Is there Any Room at

the Bottom, Proc ASME Heat Transfer Div., 1999.

8 K Satori, Y Ikeda, Y Kurosawa, A Hongo, and N Takeda, Development of Small-Diameter Optical Fiber Sensors for

Damage Detection in Composite Laminates, SPIE 7th Int Symp Smart Struct Mater 2000.

9 R Shekarriz and C.J McCall, State-of-the-Art in Micro- and

Meso-Scale Heat Exchangers Proc ASME Adv Energy Syst Div., 1999.

10 M.D Symans, G.J Madden, and N Wongprasertt, Analytical and Numerical Study of a Smart Sliding Base Isolation System

for Seismic Protection of Buildings, SPIE 7th Int Symp Smart Struct Mater., 2000.

Batteries are the major power sources for portable

elec-tronic devices and toys They are also used in

automo-biles for starting, lighting, and ignition (SLI batteries) At

present, the worldwide battery market exceeds $30 billion

per year Rapid technological advances and

miniaturiza-tion in electronics have created an ever-increasing demand

for compact, lightweight batteries For example, popular

portable electronic devices such as cellular phones,

lap-top computers, and camcorders require batteries of high

energy density Additionally, a need for more efficient use

of available energy resources as well as air-quality

con-trol have created enormous interest in electric vehicles

For example, the major automobile manufacturers around

the globe are engaged in developing advanced batteries

for electric vehicles in response to increased

environmen-tal regulations and legislative mandates The advanced

and high energy density batteries have become possible

due to the discovery and development of smart materials

and processes This article, after providing a brief

introduc-tion to the basic electrochemical concepts and the

princi-ples involved in batteries, presents the materials and

elec-trochemical aspects of high energy density (lithium-ion)

batteries

ELECTROCHEMICAL CONCEPTS

A battery is an electrochemical cell that converts the

chem-ical energy of a reaction directly into electrchem-ical energy This

section covers briefly the fundamental principles of

electro-chemical cells For more detailed information, readers are

referred to several excellent texts available in the

litera-ture (1– 4)

Electrochemical Cells

Figure 1 shows a schematic of an electrochemical cell that

consists of three components: an anode or negative

elec-trode, a cathode or positive elecelec-trode, and an electrolyte or

ionic conductor During the electrochemical reaction, the

anode M is oxidized and it gives up electrons to the

exter-nal circuit:

and the cathode X accepts the electrons from the external

circuit and is reduced:

The electrolyte, on the other hand, acts as a medium forcharge transfer between the anode and cathode as ionsinside the cell The overall cell reaction is given by addingthe two half-cell reactions (1) and (2):

M+ X → Mn ++ Xn −. (3)

The amount of electricity that passes through an trochemical cell is related by the Faraday law to the masses

elec-of reactants involved and products formed If a current elec-of

I amperes flows in the circuit for a time of t seconds, then the amount of charge Q transferred across any interface

in the cell is equal to It coulombs Now, in accordance with the Faraday law, the number of moles Nmof the reactants

M or X [see Eqs (1) and (2)] consumed by the passage of It

coulombs is given by

where n, NA, and e are, respectively, the number of electrons

given up or accepted by each M or X, Avogadro’s number,

and the charge on an electron The product NAe is called the Faraday constant F, which is equal to 96,487 C mol−1,and Eq (4) can be reduced to

weight in grams divided by the number of electrons n

in-volved in the reaction

Thermodynamics of Electrochemical Cells

The driving force for an electrochemical cell to deliver trical energy to an external circuit is the decrease in thestandard free energyG oof the cell reaction [Eq (3)] Thefree energyG o is related to the standard cell potential E o

elec-by

where n and F are, respectively, the number of electrons

involved in the reaction and the Faraday constant The cell

potential E ois the difference between the electrode

poten-tials of the cathode and anode The values of E ofor variouselectrochemical couples are given in terms of standard re-duction/oxidation potentials in textbooks and handbooks

68

Load

CathodeX

Separator

Figure 1 Schematic of an electrochemical cell.

(5,6) A positive value of E omeans that the cell reaction

oc-curs spontaneously The standard potential E ois the

equi-librium potential when all of the cell components are in

their standard states For example, the solution species

have unit molar activities, the gases have pressures of 1

atmosphere, and the solid phases are in their most stable

form in their standard states For conditions other than the

standard state, the cell potential E is given by the Nernst

where R is the gas constant, T is the absolute temperature,

and aM n +, aX n −, aM, and aXare the activities of the products

and reactants involved in cell reaction (3) At room

tem-perature T= 298 K, the Nernst equation can be simplified

The cell potential also depends on the temperature and

pressure The dependences are related to the

where S is the entropy change and V is the volume

change Thus, the measurement of the cell potential can

be used to determine thermodynamic quantities such as

G, S, enthalpy change H, and equilibrium constants.

Polarization Losses in Electrochemical Cells

The amount of electrical energy that an electrochemicalcell can deliver is related to the free energy change of the

cell reaction [Eq (7)] However, when a current I is passed

through the cell, part of the energy is lost as waste heatdue to polarization losses in the cell The polarization losscan be classified into three types: activation polarization,concentration polarization, and ohmic polarization Acti-vation polarization is related to the kinetics of electrodereactions Concentration polarization is related to the con-centration differences of the reactants and products at theelectrode surfaces and in the bulk as a result of mass trans-

fer Ohmic polarization, usually referred to as internal IR

drop, is related to the internal impedance of the cell, which

is a sum of the ionic resistance of the electrolyte and theelectronic resistance of the electrodes

The different polarization losses are indicated ically in Fig 2 as a function of operating current (2) The

schemat-operating (measured) cell voltage Eopis given by

where Eoc is the open-circuit voltage andη is the

overvol-tage from polarization The overvolovervol-tageη is a measure of the deviation of the cell voltage Eopfrom the equilibrium

open-circuit voltage Eoc The overvoltageη from the three

different polarizations is given by

Figure 2 Variation of cell voltage with operating current

illus-trating polarization losses: (a) ohmic polarization, (b) activation polarization, and (c) concentration polarization.

intrinsic properties of the electrodes and electrolytes as

well as the engineering design of the cell will influence

the polarization losses and hence the performance and

ef-ficiency of electrochemical cells

BATTERIES

Performance Parameters

As mentioned in the previous section, a battery is an

elec-trochemical device that converts stored chemical energy

directly into electrical energy The performance

character-istics of a battery are assessed in terms of several

param-eters discussed later (1–3) The cell voltage Eopis the

dif-ference between the electrode potentials of the cathode Ec

and anode Ea:

Although the theoretical capacity Q of a cell or half-cell

is given by Eq (6), it is often convenient to calculate the

specific capacity Qspfor purposes of comparison The

spe-cific capacity Qspis obtained by dividing the capacity Q of

the cell or half-cell by the mass m or volume V of the cell

or half-cell and is usually expressed in terms of Ah/kg or

The available energyε of a cell is given by the product of

the cell capacity Q and the average operating voltage Eop

and again is usually given in terms of either gravimetric

energy density (specific energyεsp) in Wh/kg or volumetric

energy density in Wh/L:

The power P delivered by the cell is given by the product of

the current I flowing and the associated cell voltage Eopand

is generally given in terms of gravimetric power density

(specific power Psp) in W/ kg or volumetric power density

in W/ L:

where Ispis the current density (current per unit weight or

volume)

The discharge characteristic of a battery is another

im-portant parameter, which is given in terms of a plot of cell

voltage versus capacity The discharge profile and the

fi-nal capacity obtainable depend on the current density Isp

used Figure 3 compares the discharge profiles for various

current densities A useful way of defining the influence

of current density on discharge curves is in terms of C

Figure 3 Discharge profiles at various C rates that illustrate the

influence of current density.

rates:

where Idand Qn are, respectively, discharge current and

nominal capacity For example, a C rate of τ implies that

the nominal capacity of the cell is delivered in 1/τ hoursunder the specified current density In an ideal battery, thedischarge voltage drops sharply to zero when the chemi-cal reaction reaches completion and the stored energy isfully consumed (Fig 3) The discharge curves deviate fromthe ideal curve as the discharge rate (or current density)increases due to the polarization losses discussed in theprevious section

Coulometric and energy efficiencies and cycle life of ondary (rechargeable) batteries are some additional impor-

sec-tant parameters Coulometric efficiency qcis defined as

where Ed and Ec are, respectively, the average discharge

and charge voltages A qC< 1 implies the occurrence of

un-wanted side reactions that produce heat during the ing process Intrinsic cell materials characteristics, cell en-gineering, and cell operating conditions such as current

charg-density and temperature can all influence qC A qE< qC

implies a deviation of the discharge and charge curves

from the open-circuit voltage profile Again, polarization

losses arising from materials characteristics, cell

engineer-ing, and operating conditions can influence qE

The cycle life of a battery is the number of times it can be

charged and discharged repeatedly before the cell capacity

falls below a limiting value Generally, the limiting value

is set around 70 to 80% of the nominal capacity The cycle

life depends on the reversible characteristics (structural

and chemical stability) of the electrode materials, cell

en-gineering, and operating conditions such as temperature,

current density, and depth of discharge

Design Considerations

The equilibrium cell voltage Eoc and the capacity Q of a

battery are determined by the intrinsic properties of the

electrode materials The cell voltage can be maximized by

choosing anode materials that have a smaller work

func-tionφaand cathode materials that have a larger work

func-tionφc In other words, the anode should be a good reducing

agent that has a large negative reduction potential, and the

cathode should be a good oxidizing agent that has a large

positive reduction potential A schematic energy diagram

of an open circuit is shown in Fig 4 The open-circuit

volt-age Eocof the cell is given by

Table 1 Major Primary Battery Systems

Cell Voltage Capacity

Magnesium Mg MnO 2 Mg + 2MnO 2 + H 2 O → Mn 2 O 3 + Mg(OH) 2 2.8 271

a

or by the difference between the electrode potentials ofthe cathode and anode [Eq (14)] Thermodynamic stability

considerations also require that the Fermi energies EFof

the cathode and anode lie within the band gap Eg of theelectrolyte, as shown in Fig 4, so that no unwanted reduc-tion or oxidation of the electrolyte occurs This implies alimitation of

Alkali and alkaline-earth metals that have a smaller φa

or a larger negative reduction potential are attractive odes, and higher valent transition-metal compounds thathave a largerφcor larger positive reduction potentials areattractive cathodes to maximize the cell voltage The cellcapacity, on the other hand, is determined by the atomic

an-or molecular weight of the elements an-or compounds used

as electrodes and the degree of reaction (number of trons involved) per mole of the electrode material [Eq (6)].Lightweight elements such as hydrogen, lithium, or oxy-gen and low molecular weight compounds are preferred aselectrodes to maximize cell capacity

elec-In addition to high cell voltage and capacity, severalother criteria are important in designing a battery toachieve high efficiency and minimal energy loss The elec-trolyte should have good ionic conductivity, but should be

an electronic insulator to avoid internal short-circuiting.High ionic conductivity in the electrolyte is essential to

minimize the IR drop or ohmic polarization Using a given electrolyte, the IR drop due to electrolyte resistance can

be reduced, and the rate capability can be improved by ahigher electrode interfacial area and thin separators Theelectrode should have a high electronic conductivity and

diffusion rate for the ions to minimize the IR drop The

electronic conductivity of the electrodes can be improved

by adding electrically conducting additives such as bon The electrode reaction rates at the cathode and anodeshould be high enough to minimize activation polariza-tion This is commonly achieved by using a porous elec-trode design, which can reduce the local current density byproviding high surface area Adequate flow or passage ofelectrolytes is essential to facilitate mass transfer and min-imize concentration polarization Electrode porosity andpore size, optimum separator thickness and structure, andconcentration of the reactants in the electrolytes are im-portant factors in minimizing concentration polarization

car-In addition to these points, the electrolyte should have

Table 2 Major Secondary Battery Systems

Cell Voltage Capacity

Nickel–cadmium Cd NiOOH Cd + 2NiOOH + 2H 2 O → 2Ni(OH) 2 + Cd(OH) 2 1.35 181

aBased only on active cathode and anode materials.

good chemical stability and should not undergo any

di-rect reaction with the electrodes In rechargeable

batter-ies, chemical reversibility on the electrodes is crucial to

maintaining good capacity retention Raw materials and

fabrication costs, cell safety, and environmental factors are

additional considerations

Types of Batteries

Batteries can be classified into two types: primary

(non-rechargeable) and secondary ((non-rechargeable) batteries

Electrode materials undergo irreversible chemical

reac-tions in primary batteries, but they exhibit reversible

chemical reactions in secondary batteries Some major

pri-mary and secondary battery systems are given in Tables 1

and 2 (2) The tables give the cell reactions, voltage, and

ca-pacity for each system Most of the primary and secondary

systems are based on aqueous electrolytes; the

lithium-based primary systems in Table 1 and the lithium-ion

system in Table 2 are based on nonaqueous electrolytes

The aqueous systems are limited in cell voltage (≤ 2.1 V)

due to a smaller separation Egbetween the highest

occu-pied molecular orbital (HOMO) and the lowest unoccuoccu-pied

molecular orbital (LUMO) of water (Fig 4) and the

con-sequent vulnerability of water to reduction/oxidation

reac-tions at higher cell voltages The use of nonaqueous

elec-trolytes that have a larger Eg, on the other hand, permits

higher cell voltages in lithium-based systems

SMART BATTERIES

The discovery of smart materials and the development of

new processes have revolutionized the electronics

indus-try over the years The continued reduction in the sizes

and weights of popular portable electronic devices such as

cellular phones and laptop computers has driven the

par-allel, development of smart batteries to power them In

this regard, lithium-ion batteries have become appealing

because they offer higher energy density (volumetric and

gravimetric) compared to other rechargeable systems (Fig

5) such as lead–acid, nickel–cadmium, and nickel–metal

hydride batteries (7) Lithium-ion batteries are smaller

and lighter compared to other systems Lithium-ion

bat-teries have become a commercial reality since Sony

Cor-poration introduced them in 1990 as a result of the

dis-covery of new materials over the years The history,

prin-ciples, current status, and future challenges of lithium-ion

technology are briefly discussed in the following sections

For more detailed information, the readers are referred toseveral references in the literature (1–3, 8–13)

us-and a cathode material that has a larger work functionφc,

as shown in [Eq (23)] Lithium metal—the lightest solid

in the periodic table—has a high specific capacity and ahigh standard oxidation potential (smallφa) and is an at-tractive anode for achieving high energy density Because

of this objective, batteries that consist of metallic lithium

as an anode and a lithium insertion compound LixMyXz

(M= transition metal and X = nonmetal) as a cathode came appealing during the 1970s; a lithium insertion com-pound is a host matrix into/from which the guest species

be-Li+ can be reversibly inserted/extracted This concept of

a secondary lithium battery was initially demonstratedusing a layered metal sulfide TiS2 as the cathode and anonaqueous electrolyte consisting of a lithium salt such asLiClO4dissolved in an organic solvent such as propylenecarbonate In this cell, the Li+ions produced at the anode

by oxidation of the metallic lithium during discharge grate through the electrolyte and are inserted into the vander Waals gap between the sulfide layers of TiS2, and theelectrons flow through the external circuit from the anode

mi-to the cathode mi-to give LixTiS2 During the charging process,the Li+ions are extracted from LixTiS2and the electrons

Figure 5 Comparison of the gravimetric and volumetric energy

densities of various rechargeable battery systems.

In this cell, however, the chemical reactivity of metallic

lithium with the nonaqueous electrolyte results in forming

a passivating film on the anode Although the passivating

film prevents further corrosion, it leads to nonuniform

plat-ing of lithium durplat-ing chargplat-ing, which results in total cell

failure due to dendritic short-circuiting and also in serious

safety problems due to local overheating These

difficul-ties of the metallic lithium anode forced the use of lithium

insertion compounds as both anodes and cathodes These

cells are called lithium-ion cells or rocking-chair cells

be-cause the lithium ion shuttles or rocks between the cathode

and anode hosts during the charging/discharging process

(Fig 6) This strategy, however, requires careful selection

of cathode and anode pairs to maintain high cell voltage

(>3 V) and to minimize the added weight of the insertion

compound anode

Although the concept of secondary lithium batteries was

initially demonstrated by using a sulfide cathode, it was

recognized during the 1980s that it is difficult to achieve

high cell voltage using sulfide cathodes because an

over-lap of the higher valent Mn +:d energies and the top of the

S:3p energy and the formation of S2 −

2 ions lead to an cessibility of higher oxidation states for Mn + in a sulfide

inac-LixMySz; the stabilization of the higher oxidation state is

essential to maximize the work function φc and thereby

the cell voltage Eoc [Eq (23)] On the other hand, the

lo-cation of O:2p energy much below the S:3p energy and a

larger increase of the Mn +:d energies in an oxide compared

to those in a sulfide, due to a larger Madelung energy, make

the higher valent states accessible in oxides Accordingly,

transition-metal oxide hosts were pursued as cathodes

during the 1980s (14–16)

Figure 7 compares the electrochemical potential ranges

of some lithium insertion compounds versus metallic

Figure 7 Electrochemical potential ranges of some lithium

in-sertion compounds with reference to metallic lithium.

lithium Among them, LiCoO2, LiNiO2, and LiMn2O4

oxides that have a higher electrode potential of 4 V sus metallic lithium have become attractive cathodes forlithium-ion cells Graphite and coke that have lower elec-trode potentials < 1 V versus metallic lithium and are

ver-lightweight have become attractive anodes In a ion cell made from, for example, a LiCoO2 cathode and acarbon anode (Fig 6), the lithium ions migrate from theLiCoO2cathode to the LixC6anode through the electrolyte,and the electrons flow through the external circuit fromthe cathode to the anode during the charging process Ex-actly the reverse reaction occurs during the dischargingprocess

lithium-A lithium insertion compound should have several tures to be a successful electrode (cathode or anode) inlithium-ion cells:

fea-rThe cathode should have a high lithium chemical

potential (µLi(c)), and the anode should have a lowlithium chemical potential (µLi(a)) to maximize the cellvoltage:

Eoc= µLi(c)− µLi(a)

The voltage is determined by the energies involved inboth electron transfer and Li+ transfer The energyinvolved in electron transfer is related to the workfunctions of the cathode (φc) and anode (φa) as shown

in Eq (23), whereas that involved in Li+transfer isdetermined by the crystal structure and the coordi-nation geometry of the site into/from which Li+ ionsare inserted /extracted (17) If we consider only elec-

tron transfer, then Eoccan be given by Eq (23) Thisimplies that the Mn + ion in the insertion compound

LixMyOzshould have a high oxidation state to be used

as a cathode and a low oxidation state to be used as

an anode

rThe insertion compound LixMyOzshould allow

inser-tion/extraction of a large amount of lithium x to imize the cell capacity This depends on the number

max-of available lithium sites and the accessibility max-of tiple valences for M in the insertion host

mul-rThe lithium insertion /extraction process should be

re-versible and have no or minimal changes in the hoststructure across the entire range x of lithium inser-tion /extraction to provide a good cycle life

rThe insertion compound LixMyOz should have good

electronic conductivityσeand Li+-ion conductivityσLi

to minimize polarization losses during the ing/charging process and thereby to support a highcurrent and power densities

discharg-rThe insertion compound LixMyOz should be

chem-ically stable and should not react with the trolyte across the entire range x of lithium insertion/

elec-extraction

rThe Fermi energies of the cathode and anode in the

entire range x of lithium insertion /extraction shouldlie within the band gap of the electrolyte, as shown inFig 4, to prevent any unwanted oxidation or reduction

of the electrolyte

rThe insertion compound LixMyOzshould be

inexpen-sive, environmentally benign, and lightweight

Layered Cobalt Oxide Cathodes

LiCoO2has a layer structure in which the Li+and Co3 +ions

occupy the alternate (111) planes of a rock salt structure, as

shown in Fig 8, to give a layer sequence of –O–Li–O–Co–

O– along the c axis This structure has an oxygen stacking

sequence of ABCABC along the c axis, and the Li+ and

Co3 +ions occupy the octahedral interstitial sites of the

cu-bic close-packed oxygen array Accordingly, it is designated

as an O3 layer structure The structure provides reversible

extraction /insertion of lithium ions from /into the lithium

planes Two-dimensional motion of the Li+ ions between

the strongly bonded CoO2layers provides fast lithium-ion

diffusion (highσLi), and the edge-shared CoO6octahedral

arrangement that has a direct Co–Co interaction provides

good electronic conductivityσe necessary for a high rate

A large work functionφc for the highly oxidized Co3+/4+

couple provides a high cell voltage of around 4 V, and the

discharge voltage does not change significantly as the

de-gree of lithium extraction /insertion x in Li1 −xCoO2changes

(Fig 9) These features have made LiCoO2 an attractive

cathode, and most of the commercial lithium-ion cells are

currently made from LiCoO2

However, only 50 % of the theoretical capacity of LiCoO2

that corresponds to a reversible extraction of 0.5 lithium

per Co (practical capacity of 140 Ah/kg) can be

practi-cally used The limitation in practical capacity has been

attributed in the literature (18) to an ordering of Li+

ions and consequent structural distortions around x= 0.5

in Li1 −xCoO2 However, it has been shown more recently

that the limited capacity could be due to the tendency of

Li−xCoO to lose oxygen (or react with the electrolyte) at

Li

Coc/2

O

a

Figure 8 Crystal structure of layered LiCoO2

a deep charge when (1− x) < 0.5 (19) Figure 10 shows the

variation of the oxidation state of cobalt and the oxygencontent as the lithium content (1− x) varies The data inFigure 10 were obtained by chemically extracting lithiumfrom LiCoO2using the oxidizing agent NO2PF6in a non-aqueous (acetonitrile) medium and determining the oxida-tion state of cobalt by a redox (iodometric) titration Con-stancy of the cobalt oxidation state and an oxygen contentsignificantly less than 2 at low lithium contents demon-strate the chemical instability of Li1 −xCoO2cathodes at adeep charge when (1− x) < 0.5 The tendency of Li1 −xCoO2

to lose oxygen at a deep charge is consistent with the cent X-ray absorption spectroscopic (20) and electron en-ergy loss spectroscopic (21) data The spectroscopic data in-dicate that the holes (removal of electrons) are introduced

Figure 10 Variations of the oxidation state of the

transition-metal ions and oxygen content as lithium content varies in

Li 1 −x CoO 2−δand Li 1 −x Ni 0.85Co 0.15O 2−δ.

into the O:2p band rather than the Co:3d band during the

electrochemical extraction of lithium Introduction of a

sig-nificant amount of holes into the O:2p band will lead to

evolution of oxygen from the lattice However, note that

neutral oxygen in the presence of electrolytes in

lithium-ion cells may not be evolved under conditlithium-ions of overcharge

when (1− x) < 0.5 Instead, the cathode may react with the

electrolyte due to the highly oxidized nature of the deeply

charged Li1 −xCoO2cathode

Figure 11 shows the X-ray diffraction patterns of the

Li1−xCoO2 samples that were obtained by chemically

ex-tracting lithium from Li1 −xCoO2 The samples maintain

the initial O3 layer structure (CdCl2structure) for 0.35≤

(1− x) ≤ 1 For lithium contents (1 − x) < 0.35, a

sec-ond phase begins to form as indicated by the appearance

of a shoulder on the right-hand side of the (003) reflection

centered around 2θ = 20◦ The intensity of the new

reflec-tion increases as the lithium content decreases further, and

the end member CoO2−δconsists of reflections

correspond-ing only to the new phase The X-ray diffraction pattern

of the new phase could be indexed on the basis of a

two-phase mixture consisting of a major P3 two-phase and a minor

oxygen stacking sequences of ABBCCA and ABABAB,

re-spectively The Li+ions occupy prismatic (trigonal prism)

and octahedral sites, respectively, in the P3 and O1

struc-tures The formation of the P3 and O1 phases from the

initial O3 structure is due to sliding of the oxide ions,

as shown in Fig 12 The driving force for the sliding

ap-pears to be structural instability caused by the formation

of oxygen vacancies at low lithium contents (Fig 10) The

observed transformation of the O3 phase at low lithium

content is consistent with that found in electrochemically

prepared samples (22) The tendency to lose oxygen and theassociated structural transitions limit the practical capac-ity of LiCoO2cathodes

Layered Nickel Oxide Cathodes

LiNiO2has an O3 layer structure (Fig 8) like LiCoO2, andthe Ni3+/4+ couple that has a larger φc provides a highcell voltage of around 4 V However, LiNiO2 suffers from

a few drawbacks: (1) difficulty in synthesizing LiNiO2as

a perfectly ordered phase without mixing Li+ and Ni3 +

ions in the lithium plane (23,24), (2) Jahn–Teller tion (tetragonal structural distortion) associated with alow spin Ni3 +:d7ion (25), (3) irreversible phase transitionsduring the charge/discharge process, and (4) safety con-cerns in the charged state As a result, LiNiO2 is not apromising material for commercial cells However, some

distor-of these difficulties have been overcome by partially stituting cobalt for nickel For example, the compositionLiNi0.85Co0.15O2, has been shown to exhibit attractive elec-trochemical properties (26) It has a reversible capacity ofaround 180 Ah/kg (Fig 9) and excellent cyclability Thiscapacity is 30% higher than that of LiCoO2, and it corre-sponds to 65% of the theoretical capacity The substitution

sub-of cobalt for nickel has been found to suppress the cationdisorder and Jahn–Teller distortion, as indicated by X-rayabsorption fine structure studies (25) The higher capacity

of LiNi0.85Co0.15O2has made it an attractive alternate forLiCoO2

However, the reason for the higher capacity ofLiNi0.85Co0.15O2 compared to the analogous LiCoO2 cath-odes was not clear in the literature The structural sta-bility of the LiNi0.85Co0.15O2 cathodes during long-termcycling, particularly under mild heat, also remained to

be assessed Recent experiments on Li1 −xNi0.85Co0.15O2

samples obtained by chemically extracting lithiumfrom LiNi0.85Co0.15O2 show that the higher capacity ofLiNi0.85Co0.15O2compared to that of LiCoO2is due to its re-sistance to losing oxygen at low lithium contents Figure 10compares the variations of the average oxidation state

of the transition-metal ions and the oxygen contents

as lithium content varies in Li1 −xNi0.85Co0.15O2 and

Li1 −xCoO2 The data show that the former system hibits better stability without losing much oxygen at adeep charge Figure 13 shows the X-ray diffraction pat-terns of the Li1 −xNi0.85Co0.15O2samples that were obtained

ex-by chemically extracting lithium from LiNi0.85Co0.15O2 In

this case, the initial O3 structure is maintained for a wider

lithium content 0.23≤ (1 − x) ≤ 1, and the new phase isformed at a lower lithium content (1 − x) < 0.23 More

importantly, the X-ray diffraction pattern of the end ber NiO2−δ could also be indexed on the basis of an O3

mem-structure but had smaller lattice parameters compared

to the initial O3 phase The observation of an O3

struc-ture for the chemically prepared NiO2−δagrees with thatfound for the electrochemically prepared sample (27,28).The absence of a significant amount of oxygen vacanciesappears to prevent the sliding of oxide ion layers and thestructural transformation The absence of oxygen loss and

the maintenance of the initial O3 structure to a much

lower lithium content (1− x) compared to that in Li1 −xCoO2

permit a higher capacity in the Li1 −xNi0.85Co0.15O2

system

The differences in oxygen loss behavior between the

Li1 −xCoO2and the Li1 −xNi0.85Co0.15O2systems can be

un-derstood by considering qualitative energy diagrams for

Li1 −xCoO2and Li1 −xNiO2(Fig 14) In LiCoO2that has a

Co3 +:3d6 configuration, the t2g band is completely filled,

and the eg band is empty As lithium is extracted from

LiCoO2, the Co3 + ions are oxidized to Co4 +, which is

ac-companied by removal of electrons from the t2gband

Be-cause the t2g band overlaps the top of the O:2p band,

deeper lithium extraction where (1− x) < 0.5 results in

a removal of electrons from the O:2p band as well The

removal of a significant amount of electron density from

the O:2p band will result in oxidation of the O2 −ions and

an ultimate loss of oxygen from the lattice In contrast,

the LiNiO2 system that has a Ni3 +:3d7 configuration

in-volves the removal of electrons only from the e band For

LiNi0.85Co0.15O2, the electrons will be removed from the eg

band for (1− x) > 0.15 Because the egband lies well abovethe O:2p band, this system does not lose oxygen down to

a lower lithium content The band diagrams in Fig 14 areconsistent with the recent spectroscopic evidence for theintroduction of holes into the O:2p band rather than intothe Co:3d band in LiCoO2(20,21) and into the Ni:3d band

in Li1 −xNiO2and Li1 −xNi0.85Co0.15O2(29,30)

To assess the structural stability of Li1 −xNi0.85Co0.15O2

cathodes during long-term cycling, chemically prepared

Li1 −xNi0.85Co0.15O2 samples were subjected to mild heatand examined by X-ray diffraction (31) The data show a

decrease in the c/a ratio of the unit cell parameters of,

for example, Li0.35Ni0.85Co0.15O2when heated at T > 50◦Cdue to migration of the Ni3 +ions from the nickel plane tothe lithium plane Interestingly, the cobalt oxide Li0.35CoO2

that has a similar degree of lithium extraction (charging)

shows little or no decrease in the c/a ratio when heated

under similar conditions Thus, the Li−xNi.85Co .15O

Co3 + /4 +:t2g

Figure 14 Comparison of the qualitative energy diagrams of

Li1−xCoO2 and Li1−xNiO2

cathodes experience structural instability under mild heat,

whereas the Li1 −xCoO2cathodes do not under similar

con-ditions Although the LiNi0.85Co0.15O2cathode has higher

capacity (180 Ah/kg) than the LiCoO2cathode (140 Ah/kg)

and is more resistant to losing oxygen from the lattice

compared to LiCoO2, the structural instability experienced

due to cation migration may become an issue under

cy-cling at higher temperatures (T > 50◦C) The differences

in the structural stability between the two systems can be

explained by considering the mechanism of cation

migra-tion The migration of transition-metal ions from

octahe-dral sites in the transition-metal plane to the octaheoctahe-dral

sites in the lithium plane needs to occur via the

neighbor-ing empty tetrahedral sites, as shown in Fig 15 While

the low spin Co3 +:3d6ion that has strong octahedral site

stabilization energy (32) is unable to migrate to the

neigh-boring tetrahedral site, but the low spin Ni3 +:3d7ion that

has moderate octahedral site stabilization energy is able

to move to the tetrahedral site under mild heat

Figure 15 Schematic representation of the diffusion processes

of nickel ions in Li1−xNi0.85Co0.15O2 Dotted and solid squares

re-fer to tetrahedral site and lithium-ion vacancy, respectively T1

and T2refer to tetrahedral sites at (0, 0, 0.125) and (0, 0, 0.375),

respectively.

Li (8a site)

Mn (16d site)

O (32e site)

Figure 16 Crystal structure of LiMn2 O 4 spinel.

Spinel Manganese Oxide Cathodes

Although LiCoO2and LiNi0.85Co0.15O2 are attractive didates, both Co and Ni are expensive and relativelytoxic These considerations have created much interest

can-in manganese oxides because Mn is can-inexpensive and vironmentally benign (33–35) In this regard, LiMn2O4

en-that crystallizes in a three-dimensional cubic spinel ture (Fig 16) has become appealing (16) In the LiMn2O4

struc-spinel, the Li+and the Mn3+/4+ ions occupy, respectively,the 8a tetrahedral and 16d octahedral sites of the cu-bic close-packed oxygen array A strong edge-shared oc-tahedral [Mn2]O4 array permits reversible extraction ofthe Li+ ions from the tetrahedral sites without collaps-ing the three-dimensional spinel framework An additionallithium-ion can also be inserted into the empty 16c oc-tahedral sites of the spinel framework to give the lithi-ated spinel Li2[Mn2]O4 However, electrostatic repulsionbetween the Li+ ions in the 8a tetrahedral and 16c oc-tahedral sites, which share common faces, causes a dis-placement of the tetrahedral Li+ions into the neighboringempty 16c sites to give an ordered rock salt structure thathas a cation distribution of (Li2)16c[Mn2]16dO4 Thus, theo-retically, two lithium ions per LiMn2O4formula unit could

be reversibly inserted/extracted Although the edge-sharedMnO6octahedral arrangement that has direct Mn–Mn in-teraction provides good electrical (small polaron) conduc-tivityσe, the interconnected interstitial sites in the three-dimensional spinel framework provide good lithium-ionconductivityσLi

The lithium extraction/insertion from/into the 8a hedral and 16c octahedral sites of the Li[Mn2]O4spinel oc-curs in two distinct steps (16) The former occurs at around

tetra-4 V (Fig 9) maintaining the cubic spinel symmetry; in trast, the latter occurs at around 3 V by a two-phase mecha-nism involving the cubic spinel Li[Mn2]O4and the tetrago-nal lithiated spinel Li2[Mn2]O4 Although both involve the

con-Mn3+/4+ couple, the 1 V difference between the two cesses reflects the differences in the site energies (17) asdifferentiated by Eq (23) and (25) A deep energy well forthe 8a tetrahedral Li+ ions and a high activation energyrequired for the Li+ions to move from one 8a tetrahedralsite to another via an energetically unfavorable neighbor-ing 16c site lead to a higher voltage of 4 V (33) The cubic totetragonal transition from Li[Mn ]O to Li [Mn ]O is due

pro-t2 (dxy, dyz, dzx)

(dxz, dyz)(dxy)

Figure 17 Illustration of the Jahn–Teller distortion in

man-ganese oxides: (a) Mn4+:3d3 that has cubic symmetry (no Jahn–

Teller distortion) and (b) Mn3+:3d4 that has tetragonal symmetry

(Jahn–Teller distortion).

to the Jahn–Teller distortion of the single electron in the

egorbitals of a high spin Mn3 +:3d4ion (Fig 17) A

coopera-tive distortion of the MnO6octahedra that have long Mn–O

bonds along the c axis and short Mn–O bonds along the a

and b axes results in macroscopic tetragonal symmetry for

Li2[Mn2]O4

Although, in principle, two lithium ions per LiMn2O4

formula unit could be reversibly extracted/inserted

from/into the Li[Mn2]O4 spinel framework, the cubic to

tetragonal transition is accompanied by a 16% increase

in the c /a ratio of the unit cell parameters and a 6.5%

in-crease in unit cell volume This change is too severe for

the electrodes to maintain structural integrity during the

discharge/charge cycle, and so LiMn2O4 exhibits drastic

capacity fade in the 3 V region As a result, LiMn2O4has

limited practical capacity of around 120 Ah/kg (Fig 9) that

corresponds to an extraction/insertion of 0.4 lithium per

Mn in the 4 V region Furthermore, even though it has

lim-ited capacity, LiMn2O4tends to exhibit capacity fade in the

4 V region as well, particularly at elevated temperatures

(50◦C) The capacity fade in the 4 V region has been

at-tributed to a dissolution of manganese into the electrolyte

originating from a disproportionation of Mn3 +into Mn4 +

and Mn2 +(36) and the formation of tetragonal Li2[Mn2]O4

on the surface of the particles under conditions of

nonequi-librium cycling (37)

The difficulties of lattice distortions in the LiMn2O4

spinel have motivated strategies to suppress Jahn–Teller

distortion One way to suppress Jahn–Teller distortion is to

increase the average oxidation state of manganese because

Mn4 +:3d3does not undergo Jahn–Teller distortion The

ox-idation state of manganese can be increased either by

alio-valent cationic substitutions or by increasing the oxygen

content in LiMn2O4 Using this strategy, Thackeray et al

(33,38,39) pioneered the Li–Mn–O phase diagram For

ex-ample, substituting Li for Mn in Li+xMn O increases the

oxidation state, and the end member Li4Mn5O12(x= 0.33)has an oxidation state of 4+ for Mn Similarly, the ox-idation state increases as oxygen content increases inLiMn2O4+δ, and the end member Li2Mn4O9(δ = 0.5) willhave an oxidation state of 4+ for Mn However, these de-fective spinels are difficult to synthesize by conventionalhigh-temperature procedures, and more recently solution-based syntheses have been pursued to obtain them (40,41)

It is also difficult to extract lithium from Li4Mn5O12 and

Li2Mn4O9because Mn4 +is difficult to oxidize further andtherefore they are not suitable for lithium-ion cells thathave carbon anodes Both Li4Mn5O12 and Li2Mn4O9 ex-hibit most of their capacity in the 3 V region that corre-sponds to insertion of additional lithium into the 16c sites.Nevertheless, the Jahn–Teller distortion has been shown

to be delayed until late in the discharge process in both tems (33,38,39); the cubic symmetry without Jahn–Tellerdistortion has been shown to be preserved to x = 2.5 in

sys-Li4 +xMn5O12and x= 1.7 in Li2 +xMn4O9

Other Oxide Cathodes

The difficulties of the LiMn2O4spinel also motivated theinvestigation of several nonspinel manganese oxides, par-ticularly by employing low-temperature synthesis (33–35).LiMnO2obtained by conventional synthesis does not crys-

tallize in the O3 structure of LiCoO2; it adopts an thorhombic rock salt structure in which the oxygen array

or-is dor-istorted from the ideal cubic close packing (42) ever, LiMnO2isostructural with layered LiCoO2can be ob-tained by ion exchange of NaMnO2(43) or by partial sub-stitution of Mn by Cr or Al (44) Unfortunately, both theorthorhombic LiMnO2and the layered LiMnO2(O3 struc-

How-ture) that have close-packed oxygen arrays tend to form to spinel-like phases during electrochemical cycling

trans-In this regard, Na0.5MnO2—designated as Na0.44MnO2 inthe literature—adopts a non-close-packed structure andhas drawn some attention because it does not transform tospinel-like phases (45,46) However, only a small amount

of lithium could be extracted from the ion-exchanged ple Na0.5−xLixMnO2 although additional lithium could beinserted into Na0.5−xLixMnO2 Therefore, it is not attrac-tive for lithium-ion cells that use carbon anodes Neverthe-less, it has been shown that it is a promising candidate forlithium polymer batteries that employ metallic lithium an-odes (45) Additionally, it has been shown that amorphousmanganese oxides LixNayMnOzIηsynthesized in nonaque-ous media exhibit high capacity (300 Ah/kg) and good cy-clabilty (47,48) However, the capacity occurs across thewide voltage range of 4.3 to 1.5 V that has a continuouslysloping discharge profile, which is not desirable for com-mercial cells Not much lithium could be extracted from theinitial material LixNayMnOzIη, and therefore, these amor-phous oxides also are not attractive for lithium-ion cellsfabricated using carbon anodes

sam-To improve the cyclability of the LiMn2O4spinel, tial substitution of Mn by several other transition metals

par-M= Cr, Co, Ni and Cu in LiMn2 −yMyO4has been pursued(49–52) These substitutions, however, result in the devel-opment of two plateaus that correspond to the removal oflithium from the 8a tetrahedral sites: one around 4 V that

corresponds to the oxidation of Mn3 +to Mn4 +and the other

around 5 V that corresponds to the oxidation of the other

transition-metal ions The capacity in the 4-V region

de-creases, whereas that in the 5-V region increases as y in

LiMn2 −yMyO4increases Although an increase in the cell

voltage is attractive from the view point of energy density,

the LiMn2 −yMyO4oxides are prone to suffer from oxygen

loss and safety concerns in the 5-V region

LiVO2 adopts the O3 structure of LiCoO2 Although

lithium can be readily extracted from LiVO2, the

vana-dium ions migrate to the lithium planes for (1− x) < 0.67

in Li1 −xVO2 (53) Similarly, the LiV2O4spinel is plagued

by the migration of vanadium ions during the charge/

discharge process (54) Interestingly, a number of other

vanadium oxides such as VO2(B)—a metastable form of

VO2obtained by solution-based synthesis (55)—and V6O13

(56) exhibit high capacity and good cyclability However,

these oxides that have no lithium are not attractive for

lithium-ion cells made using carbon anodes LiCrO2 also

crystallizes in the O3 structure of LiCoO2, but it is

diffi-cult to extract lithium from this material

Iron oxides offer significant advantages in both cost and

toxicity compared to other oxides Although LiFeO2

ob-tained by conventional procedures adopts a different

struc-ture, layered LiFeO2that has the O3 structure of LiCoO2

can be obtained by an ion-exchange reaction of NaFeO2

Unfortunately, the layered LiFeO2 does not exhibit good

electrochemical properties because the high spin Fe3 +:3d5

ion tends to move around between the octahedral and

tetra-hedral sites This problem could, however, be overcome by

designing complex iron oxides that consist of poly ions such

as (SO4)2 −and (PO4)3 − For example, it was shown in the

1980s that Fe2(SO4)3that has a framework structure that

consists of FeO6octahedra which share all six corners with

(SO4)2 −tetrahedra has a capacity of around 130 Ah/kg and

a flat discharge voltage of 3.6 V (57) However, the poor

electronic conductivity of the Fe–O–X–O–Fe (X= S or P)

linkages leads to poor rate capability Nevertheless,

fol-lowing this initial concept of using poly ions, LiFePO4that

crystallizes in an olivine structure has recently been shown

to be a promising material; it exhibits a flat discharge

volt-age of around 3.4 V (58) However, LiFePO4appears to

suf-fer from limited rate capability due to insuf-ferior electronic

(σe) and lithium-ion (σLi) conductivity compared to other

oxide cathodes such as LiCoO2and LiMn2O4

Carbon Anodes

Carbon has become the material of choice for anodes in

lithium-ion cells (59,60) due to its light weight and low

electrochemical potential that lies close to that of metallic

lithium (Fig 7) Although the intercalation of alkali metals

into graphitic carbon was known for some time, the

recog-nition of a practical carbon anode began in 1989 (61) A

sub-sequent announcement by Sony Corporation in 1990

com-mercializing lithium-ion cells that have carbon anodes and

LiCoO2cathodes intensified the interest in carbon Carbon

materials can be broadly classified into two categories:

soft carbon (graphitic carbon) and hard carbon (glassy

car-bon) The former has a better ordering of graphene

lay-ers compared to the latter, and they differ significantly in

1.5Irreversiblecapacity

Figure 18 Typical first discharge/charge curves of graphite The

difference between the discharge and charge curves is irreversible capacity loss.

their physical and chemical properties The hard carbonsare typically obtained by thermal decomposition of, forexample, phenolic and epoxy resins and products frompetroleum pitch They have significant amounts of openmicropores, which tend to become closed when heated atincreasingly higher temperatures Some hard carbons havebeen shown to consist of single graphene sheets

Figure 18 shows a typical first discharge/charge curve

of graphite It has a theoretical capacity of 372 Ah/kg thatcorresponds to an insertion of one lithium per six carbonatoms (x= 1 in LixC6) It shows a significant amount of ir-reversible capacity during the first discharge/charge cycledue to unwanted, irreversible side reactions with the elec-trolyte These side reactions can lead to disintegration ofthe carbon anode and a decrease in cycle efficiency Moreimportantly, using electrolytes that consist of propylenecarbonate (PC), natural graphite cannot be charged be-cause it leads to an evolution of gas at around 1 V This

is a major drawback of the graphite anode However, usingelectrolytes that consist of other solvents such as ethylenecarbonate (EC) and diethyl carbonate (DEC), the side re-actions are suppressed, and it can be cycled without muchdifficulty

The hard carbons generally have higher capacity thangraphite Several models have been proposed to accountfor the increased capacity: adsorption of lithium on bothsides of the single graphene sheets, accommodation of ex-tra lithium into nanometer size cavities, and storage of ad-ditional lithium at the edges and surfaces are some of theexplanations Generally, the capacity of carbon materialshas been found to increase as the fraction of single layermaterial increases However, the hard carbons have a slop-ing discharge profile between 0 and 1 V, unlike graphite,which has a nearly flat discharge profile between 0 and0.3 V (Fig 18) As a result, for lithium-ion cells that usethe same cathode, the hard carbons will lead to a slightlylower cell voltage compared to graphite

Structural modifications to improve the performance ofcarbon anodes are being pursued by various groups In

this regard, texture control, surface modification by mild

oxidation, and incorporation of other elements such as

B, O, Si, and P have been studied For example, carbon

fibers that have different textures and compositions such

as C1 −y−zSiyOzand BC2N have been investigated

Other Anodes

The irreversible capacity loss encountered in carbon

an-odes has motivated the search for other anode hosts In this

regard, a few materials have drawn attention Li4Ti5O12

that has a cubic spinel structure accommodates three

ad-ditional lithium ions per formula unit into the empty 16c

sites and has negligible change in volume (<0.1%) at a flat

discharge voltage of around 1.5 V (62) However, the higher

voltage (1.5 V) and lower capacity (160 Ah/kg) compared to

that of carbon make it less attractive SnO2is another

can-didate that has reversible capacity as high as 600 Ah/kg at

0 to 2 V, but it exhibits high capacity loss during the first

cy-cle (63) More recently, some intermetallic compounds such

as Cu6Sn5that have an NiAs-type structure have shown

promise (64), and further development of these systems is

in progress

However, all of the preceding candidates and carbon

an-odes do not contain lithium, and they require that the

cath-odes be in the lithiated (discharged) form In this regard,

anode hosts based on some lithiated transition-metal

ni-trides (59) and intermetallic compounds (65) are appealing

For example, Li7 −xMnN4and Li3 −xFeN2exhibit capacities

of, respectively, 210 and 150 Ah/kg and a flat discharge

voltage of around 1.2 V Li2.6−xCo0.4N and Li2.6−xCu0.4N

whose structures are similar to that of Li3N exhibit a much

higher capacity of around 470 Ah/kg at 0.2 to 1 V; the

for-mation of an amorphous phase during the initial stages

in the latter systems leads to high capacity Two lithium

per formula unit can be reversibly extracted from Li2CuSn

(65) Further development work is necessary to assess the

full potential of these lithium-containing anodes If

suc-cessful, they have the possibility of being coupled with

some of the lithium-free cathodes such as vanadium oxides

However, the resulting lithium-ion cells may have a lower

cell voltage compared to the currently available

lithium-ion cells fabricated using LiCoO2 cathodes and carbon

anodes

CONCLUDING REMARKS

After providing an introduction to the concepts of

elec-trochemical cells and batteries, lithium-ion batteries that

offer much higher energy density compared to other

rechargeable systems were discussed Lithium-ion

batter-ies have become commercial due to the development of

smart electrode (cathode and anode) host materials

State-of-the-art lithiumion cells are made by using a

transition-metal oxide cathode such as a layered lithium cobalt

ox-ide and a carbon anode However, only 50 to 65% of the

theoretical capacity of the currently available oxide hosts

can be practically used Chemical instability, structural

instability, or safety concerns in the charged state limit

the usable capacity, depending on the cathode host The

challenge is to develop smart cathode hosts in which atleast one lithium ion per transition-metal ion can be re-versibly extracted/inserted and at the same time reducethe material cost; such a cathode can nearly double theenergy density There are also possibilities for increasingthe capacity of anode hosts by focusing on amorphous ma-terials An alternate approach is to develop cells that uselithium-containing anodes and lithium-free cathodes Thisstrategy will allow using some of the already known highcapacity cathodes such as vanadium oxides that have bet-ter chemical stability and safety characteristics From theview points of safety, cycle, and shelf life, cathodes thathave a lower voltage (3 to 4 V) but have a significantly in-creased capacity may be desirable for future applications

ACKNOWLEDGMENTS

The author thanks the Welch Foundation (Grant F-1254)for financial assistance and Dr A M Kannan, SeungdonChoi, and Ramanan Chebiam for their assistance with thefigures

Tech-8 M.S Whittingham and A.J Jacobson, Intercalation Chemistry.

Academic Press, NY, 1982.

9 J.P Gabano, Lithium Batteries Academic Press, London,

14 K Mizushima, P.C Jones, P.J Wiseman, and J.B Goodenough,

Mat Res Bull 15: 783 (1980).

15 J.B Goodenough, K Mizushima, and T Takeda, Jpn J Appl.

Phys 19: 305 (1983).

16 M.M Thackeray, W.I.F David, P.G Bruce, and J.B.

Goodenough, Mat Res Bull 18: 461 (1983).

17 M.K Aydinol and G Ceder, J Electrochem Soc 144: 3832

(1997).

18 J.N Reimers and J.R Dahn, J Electrochem Soc 139: 2091

(1992).

19 A Manthiram, R.V Chebiam, and F Prado, Chem Mater 13:

2951 (2001), p 269; R.V Chebiam, A.M Kannan, F Prado, and

A Manthiram, Electrochem Commun 3: 624 (2001).

20 L.A Montoro, M Abbate, and J.M Rosolen, Electrochem Solid

State Lett 3: 410 (2000).

21 A Hightower, J Graetz, C.C Ahn, P Rez, and B Fultz, 198th

Meet Electrochem Soc Phoenix, AZ, October 22–27, 2000,

24 A Hirano, R Kanno, Y Kawamoto, Y Tekeda, K Yamamura,

M Takano, K Ohyama, M Ohashi, and Y Yamaguchi, Solid

State Ionics 78: 123 (1995).

25 I Nakai, K Takahasi, Y Shiraishi, T Nakagome, and

F Nishikawa, J Solid State Chem 140: 145 (1998).

26 W Li and J Curie, J Electrochem Soc 144: 2773 (1997).

27 X.Q Yang, X Sun, and J McBreen, Electrochem Commun 2:

100 (2000).

28 L Croguennec, C Pouillerie, and C Delmas, J Electrochem.

Soc 147: 1314 (2000).

29 A.N Mansour, X.Q Yang, X Sun, J McBreen, L Croguennec,

and C Delmas, J Electrochem Soc 147: 2104 (2000).

30 M Balasubramanian, X Sun, X.Q Yang, and J.Mcbreen,

J Electrochem Soc 147: 2903 (2000).

31 R.V Chebiam, F Prado, and A Manthiram, J Electrochem.

Soc 148: A49 (2001).

32 J E Huheey, Inorganic Chemistry: Principles of Structure and

Reactivity Harper & Row, NY, 1972, p 294.

33 M.M Thackeray, Prog Solid State Chem 25: 1 (1997).

34 A Manthiram and J Kim, Chem Mat 10: 2895 (1998).

35 A Manthiram and J Kim, Recent Res Dev Electrochem 2: 31

(1999).

36 S.J Wen, T.J Richardson, L Ma, K.A Striebel, P.N Ross, and

E.J Cairns, J Electrochem Soc 143: L136 (1996).

37 M.M Thackeray, Y Shao-Horn, A.J Kahaian, K.D Kepler,

E Skinner, J.T Vaughey, and S.A Hackney, Electrochem Solid

State Lett 1: 7 (1998).

38 M.M Thackeray, A de Kock, M.H Rossouw, D.C Liles,

D Hoge, and R Bittihn, J Electrochem Soc 139: 363 (1992).

39 R.J Gummow, A de Kock, and M.M Thackeray, Solid State

43 A.R Armstrong and P.G Bruce, Nature 381: 499 (1996).

44 I.J Davidson, R.S McMillan, and J.J Murray, J Power Sources

47 J Kim and A Manthiram, Nature 390: 265 (1997).

48 J Kim and A Manthiram, Electrochem Solid State Lett 2: 55

(1999).

49 C Sigala, D Guyomard, A Verbaere, Y Piffard, and M.

Tournoux, Solid State Ionics 81: 167 (1995).

50 H Kawai, M Nagata, H Takamoto, and A.R West,

Elec-trochem Solid State Lett 1: 212 (1998).

51 Q Zhong, A Bonakdarpour, M Zhang, Y Gao, and J.R Dahn,

J Electrochem Soc 144: 205 (1997).

52 Y Ein-Eli, W.F Howard, Jr., S H Lu, S Mukerjee, J McBreen,

J.T Vaughey, and M.M Thackeray, J Electrochem Soc 145:

1238 (1998).

53 L.A de Picciotto, M.M Thackeray, W.I.F David, P.G Bruce,

and J.B Goodenough, Mat Res Bull 19: 1497 (1984).

54 L.A de Picciotto, M.M Thackeray, Mat Res Bull 20: 1409

(1985).

55 C Tsang and A Manthiram, J Electrochem Soc.144: 520

(1997).

56 C Lampe-Onnerud, J.O Thomas, M Hardgrave, and S.

Yde-Andersen, J Electrochem Soc 142: 3648 (1995).

57 A Manthiram and J B Goodenough, J Power Sources 26: 403

O Yamamoto, eds., Wiley-VCH, Weinheim, 1998, p 98.

60 M Winter and J.O Besenhard, in Lithium Ion Batteries: damentals and Performance, M Wakihara and O Yamamoto,

Fun-eds., Wiley-VCH, Weinheim, 1998, p 127.

61 R Kanno, Y Takeda, T Ichikawa, K Nakanishi, and

O Yamomoto, J Power Sources 26: 535 (1989).

62 D.W Murphy, R.J Cava, S.M Zahurak, and A Santoro, Solid

State Ionics 9 & 10: 413 (1983).

63 W Liu, X Huang, Z Wang, H Li, and L Chen, J Electrochem.

Soc 145: 59 (1998).

64 K.D Kepler, J.T Vaughey, and M.M Thackeray, Electrochem.

Solid State Lett 2: 307 (1999).

65 J.T Vaughey, K.D Kepler, R Benedek, and M.M Thackeray,

of a material because it is implanted in the body for longperiods Among many traditional materials, including met-als, alloys, and ceramics, that are available commercially,only a limited number are currently used as prostheses

or biomaterials in medicine and dentistry The tions in the second category require excellent mechanical

applica-characteristics as well as biocompatibility The third

cate-gory is used mainly for transducers

Among smart materials, the Ti–Ni shape-memory

al-loy (SMA) has attracted the most attention for biomedical

applications in the first and second categories due to its

excellent biocompatibility and mechanical characteristics

Research on biomedical applications of the SMA started in

the 1970s with animal experiments initially, followed by

clinical tests The first example of a successful biomedical

application of the SMA was a bone plate, which was used to

repair broken bones Now, many medical and dental

appli-cations of SMAs are available, and many new appliappli-cations

are being developed On the other hand, piezoelectric

ma-terials have been widely used as transducers for medical

ultrasonic devices due to their sensor function that uses

piezoelectricity and the actuator function that uses inverse

piezoelectricity

In this article, the properties of SMAs for biomedical

applications are discussed next, followed by some clinical

examples Recent examples of biomedical applications of

SMAs are summarized there after, and finally the recent

examples of biomedical applications of piezoelectric

mate-rials are summarized

PROPERTIES OF SMAS FOR BIOMEDICAL APPLICATIONS

The properties of SMAs that are important and have led

to its wide acceptance in biomedical applications are

dis-cussed in this section Of these properties,

biocompatibili-lity, which simply means the ability of a material to be

accepted by the body, is the most important, especially for

implants The other important properties include

super-elasticity, the shape-memory effect, hysteresis, and fatigue

resistance The properties of SMAs for biomedical

applica-tions are discussed in detail in (1)

Biocompatibility

The biocompatibility of a material is its most important

property if it is used as prostheses or biomaterials in

medicine and dentistry Biocompatibility means that the

material is nontoxic during the implanted period Because

all materials generate a “foreign body reaction” when

im-planted in the body, the degree of biocompatibility is

re-lated to the extent of this reaction Due to the rigorous

demands on material properties for biocompatibilty, only

these three metallic materials were qualified for use as

implant materials: Fe–Cr–Ni, Co–Cr and Ti–Al–V before

SMA Investigations were carried out by many researchers

on the biocompatibility of Ti–Ni (2), and an extensive

re-view can be found in (1) The results of these studies show

that Ti–Ni has superior corrosion resistance due to the

for-mation of a passive titanium oxide layer (TiO2) similar to

that found on Ti alloys This oxide layer increases the

sta-bility of the surface layers by protecting the bulk material

from corrosion and creates a physical and chemical barrier

to Ni oxidation

In in vitro dissolution studies, Bishara et al (3) found

that Ti–Ni appliances release an average of 13.05µg/day

Ni in saliva, which is significantly lower than the estimated

average dietary intake of 200–300µg/day In addition, the

measured nickel blood levels of orthodontic patients whohave Ni–Ti appliances show no significant increase during

in the body and makes it possible to create a prestressafter deployment, when necessary SMA appliances arefirst in a compact state during deployment and then re-stored to their expanded shape by heating If the phasetransformation temperature of an SMA is below body tem-perature, shape recovery can easily be induced by the heat

of the body When the phase transformation temperature

is higher than the body temperature, SMA appliances areusually heated by warm salt water or a high frequencymagnetic field

In recent studies, the shape-memory effect has also beenused for actuator functions in medical applications such

as a urethral valve and artificial sphincter which are cussed later

dis-Superelasticity

SMAs exhibit superelasticity when they are in the itic phase (1,5) Figure 1 shows the typical superelasticstress–strain curve (solid line) compared with the stress–strain curve of stainless steel (dashed line) As shown inthe figure, an important feature of superelastic materials isthat they exhibit constant loading and unloading stressesacross a wide range of strain As shown in Fig 1, the

A

StrainSubthreshold

Excessiveforce zone

Figure 1 Typical stress–strain curve of superelastic materials

and stainless steel The superelastic materials exhibit constant unloading stress over a wide range of strain.

effective strain range εeff(TN) of Ti–Ni that corresponds

to an optimal force zone is much larger than theεeff(SS)

of stainless steel Hence, a superelastic device can provide

constant pressure even if the pressed part recedes by a

lim-ited amount during the installed period On the other hand,

the pressure exerted by an appliance made from stainless

steel will drop drastically if the pressed part deforms, so

that performance deteriorates The orthodontic arch wire

that is presented as an example of an application in the

next section was the first product to use this property

Another example of applying this property is superelastic

eyeglass frames (6) These eyeglass frames have become

very popular in the United States, Europe, and Japan and

are available in almost every optician’s store These frames

can be twisted a full 180◦, but more importantly the frames

press against the head with a constant and comfortable

stress Not only is “fit” less important, but small bends

and twists that may develop do not cause discomfort to the

wearer

The superelasticity of SMAs also makes it easy to deploy

SMA stents Stents made from stainless steel are expanded

against the vessel wall by plastic deformation caused by

inflating a balloon placed inside the stent Ti–Ni stents,

on the other hand, are self-expanding More details can be

found later

Hysteresis of SMA

As shown in Fig 1, superelastic SMA exhibits a hysteretic

stress–strain relationship; the stress from A to B in the

loading phase and the stress from C to D in the unloading

are different Hysteresis is usually regarded as a drawback

in traditional engineering applications, but it is useful in

biomedical applications If the SMA is set at some stress–

strain state, E, for example, upon unloading during

deploy-ment, it should provide a light, constant force against the

organ wall, even under a certain amount of further strain

release (e.g., from E to D) On the other hand, it would

gen-erate a large resistive force to crushing if it is compressed

in the opposite direction because it takes the loading path

from E to F Hence, the SMA material exhibits a biased

stiffness at point E, which is very important in designing

a SMA stent Because the stress at the loading phase from

A to B and the stress in the unloading phase from C to D

depend on the material composition of the SMA, the

desi-rable stress–strain curve can be obtained by optimizing the

material composition

Anti-Kinking Properties

As shown in Fig 1, the stress of stainless steel remains

nearly constant in the plastic region This means that a

small increase in stress in the plastic region could lead to

a drastic increase in strain or failure of a medical

appli-ance made from stainless steel (1) On the other hand, the

stiffness of superelastic Ti–Ni increases drastically after

point B at the end of the loading plateau The increase in

stiffness would prevent the local strain in the high strain

areas from further increasing and partition the strain in

the areas of lower strain Hence, strain localization is

pre-vented by creating a more uniform strain than could be

realized by using a conventional material

EXAMPLES OF BIOMEDICAL APPLICATIONS

Orthopedic Marrow Needles Figures 2 and 3 show two types of mar-

row needles that are used in the repair of a broken thighbone (4,6–8) When a stainless steel Kunster marrow nee-dle is used, blood flow inside the bone can be blocked, andrecovery can be delayed It also has the drawback of lowtorsional strength On the other hand, a Kunster marrowneedle of SMA can be inserted into the bone in its initialstraight shape and transformed to a curved shape by heat-ing, as shown in Fig 2 Hence, the SMA Kunster marrow

Thighbone

Fracture closed

SMA aftershape recovery

After heating(b)

Figure 2 Kunster marrow needle (6).

Heat

Figure 3 Marrow needles before and after heating (5,8).

needle can avoid the disadvantages of the stainless steel

needle The SMA Kunster marrow needle can also provide

a compressive force on the fracture surfaces

The marrow needle shown in Fig 3 has a

compli-cated shape for purposes of reinforcement, which makes it

difficult to insert in the broken bone Using the

shape-memory effect, insertion can be greatly improved, as shown

in the figure, without loosing the reinforcing function,

be-cause the needles can be inserted in a simpler shape and

the necessary size and shape are recovered by heating the

needle in the marrow

Currently available joint prostheses are made of bone

cement to be fixed in the bone Stress acting on the joint

prosthesis is quite intense and severe: three to six times the

body weight of the patient under nominal action, and the

stress is cycled up to 106times Conventional bone cement

causes several inconveniences: gradual loosening after

im-plantation, resultant infection, and other complications A

prosthetic joint made of Ti–Ni SMA was developed to avoid

such problems High wear resistance is also another

advan-tage of the Ti–Ni prosthetic joint

Bone Staple and Bone Plate The bone staple shown in

Fig 4 and the bone plate in Fig 5 are used to fix broken

bones (4–6) As shown in Fig 4, a bone staple made of SMA

can be inserted at low temperature in the holes opened in

the bone, and then heated by the body temperature, it

re-covers its original shape to provide a compressive force on

the surfaces of the broken bone Bone plates are attached

by screws for fixing broken bones Bone plates made of Ti–

Ni SMA are more effective in connecting broken bones than

bone plates made of conventional material because SMA

Figure 5 Bone plate used to fix broken bones (4–6).

bone plates provide compressive force on the fracture face of the broken bones as well as repair, as shown in Fig 5.Healing proceeds faster under uniform compressive force

sur-Dental Applications Orthodontic Fixtures Due to its superelasticity, Ti–Ni is

used in many applications in dentistry It is obvious thatsuperelasticity gives the orthodontists better mechanicalcharacteristics compared to conventional elastic materialssuch as stainless steel Figure 6a,b shows a clinical ex-ample of orthodontic treatment using a superelastic Ti–Niarch wire (5,9) When fixtures made of conventional elas-tic material such as stainless steel are used, the reformingforce drops, and the fixture loosens due to movement of theteeth Hence, the fixture must be replaced several timesbefore the treatment is finished An SMA fixture main-tains a constant reforming force in a wide range of teethmovement due its superelasticity, so that no replacement isrequired after the initial installation Clinical results alsoshowed faster movement of the teeth and a shorter chairtime compared with stainless steel wire

Tooth-root Prosthesis Among several methods that

re-store the masticatory function of patients missing morethan one tooth, a tooth-root prosthesis is considered themethod that creates the most natural masticatory func-tion Blade-type implants made of Ti–Ni SMA, as shown

in Fig 7, have been used in Japan (5,6) The open angle

of the blade is used to ensure tight initial fixation and toavoid accidental sinking during mastication But to makethe insertion operation easy, the tooth-root prosthesis is

Misaligned teeth before treatment(b)

Normally aligned teeth after the first stage of treatment

Figure 6 Orthodontic treatment using a superelastic arch wire:

(a) misaligned teeth before treatment; (b) normally aligned teeth

after the first stage of treatment (5,9).

implanted in the jawbone as a flat shape and then the

opened shape is changed by heating Fig 8 shows an X-ray

photograph of the implanted tooth-root prosthesis More

than 5,000 clinical examples of SMA tooth-root prostheses

have been reported

Partial Denture The key to a partial denture is the

development of an attachment for connecting the partial

Figure 7 Tooth-root prosthesis (5,6).

Figure 8 X-ray photo of implanted tooth-root prosthesis (6).

denture to the retained teeth Clasps have been tionally used for about a century as the attachment for

conven-a pconven-articonven-al denture One of the drconven-awbconven-acks of clconven-asps mconven-ade

of conventional elastic materials is loosening during use;this can be improved by replacing the elastic materials by asuperelastic Ti–Ni alloy (5,10) Another drawback of clasps

is aesthetics because they are visible in the teeth ment To solve the problem, the size of the attachmentmust be smaller than the width of the teeth so that it can

align-be emalign-bedded completely in the teeth A precision ment using a small screw has recently become available,but it has to be designed and fabricated very precisely sothat it lacks flexibility to follow the change in the settingcondition during long-term use due to the shape change ofthe jawbone Because of its flexibility, this problem can besolved by using an attachment made of SMA

attach-The SMA attachment consists of two parts: a fixed partthat is made of a conventional dental porcelain-fusible castalloy and is attached to the full cast crown on the anchorteeth and a movable part that is made of Ti–Ni SMA and

is fixed on the side of the partial denture Examples ofmovable and fixed parts are shown in Fig 9

Surgical Instruments

Since superelastic tubing became available in the early tomid-1990s, a variety of catheter products and other en-dovascular devices using Ti–Ni has appeared on the mar-ket Early applications of Ti–Ni were retrieval baskets thathave Ti–Ni kink-resistant shafts, as well as a superelasticbasket to retrieve stones from kidneys, bladders, and bileducts An interesting example is the interaortic balloonpump (IABP) used in cardiac assist procedures (Fig 10).The use of Ni–Ti allowed a reduction in the size of the de-vice compared with the polymer tube designs and increasedthe flexibility and kink resistance compared with stainlesssteel tube designs (1)

Biopsy forceps made from stainless steel are very cate instruments that can be destroyed by even very slightmishandling Ti–Ni instruments, on the other hand, canhandle serious bending without buckling, kinking, or per-manent deformation Figure 11 shows a 1.5-mm biopsy for-ceps that consists of thin wall Ti–Ni tubing and a Ti–Niactuator wire inside Together they can be bent around a

Movable part(b)

Fixed part

Figure 9 Shape-memory alloy attachment for partial denture;

(a) movable part; (b) fixed part (4,5).

radius of less than 3 cm without kinking, and still allow for

the opening and closing of the distal grasper jaws without

increased resistance The instrument continues to operate

smoothly even while bent around tortuous paths

Figure 10 The Arrow interaortic balloon pump uses a Nitinol™

tube to pressurize the balloon (1).

Figure 11 Ninitol™ tubing that has an internal actuating wire

allows this 0.8-mm diameter grasper to operate while tied in a knot (1).

Stent

The term stent is used for devices that are used to scaffold

or brace the inside circumference of tubular passages or mens, such as the esophagus biliary duct, and most impor-tantly, a host of blood vessels, including coronary, carotid,iliac, aorta, and femoral arteries (Fig 12) (1) Stenting inthe cardiovascular system is most often used as a follow-

lu-up to balloon angioplasty, a procedure in which a balloon

is placed in the diseased vessel and expanded to reopen

a clogged lumen Ballooning provides immediate ment in blood flow, but 30% of the patients have restenosedwithin a year and need further treatment The place-ment of a stent immediately after angioplasty, it has beenshown, significantly decreases the propensity for resteno-sis Stents are also used to support grafts, for example, intreating aneurysms (Fig 13)

improve-Most stents today are stainless steel and are expandedagainst a vessel wall by plastic deformation caused byinflating a balloon placed inside the stent Ti–Ni stents,

on the other hand, are self-expanding—they are shape-set

to the open configuration, compressed into a catheter, then

Figure 12 A stent that maintains vessel patency and blood flow

to the brain is portrayed in a cutaway view of the internal carotid artery (1).

Figure 13 Stentgrafts used to exclude aneurysms, to provide an

artificial replacement for injured vessels, or prevent restenosis

after angioplasty (1).

pushed out of the catheter and allowed to expand against a

vessel wall Typically, the manufactured stent’s outer

dia-meter is about 10% greater than the vessel’s diadia-meter to

ensure that the stent anchors firmly in place The

flexibi-lity of Ti–Ni is about 10–20 times greater than stainless

steel, and it can bear a reversible strain as high as 10%

Ni–Ti stents are made from knitted or welded wire,

laser-cut or photoetched sheet, and laser-laser-cut tubing The

pre-ferred devices are laser-cut tubing, thus avoiding overlaps

and welds (Fig 14)

CURRENT BIOMEDICAL APPLICATIONS OF SMA

Artificial Urethral Valve

Urinary incontinence is the involuntary discharge of urine

caused by weakness of the urinary canal sphincter muscles

due to aging and expansion of the prostate gland

How-ever, the difference in ages, the sex of patients, and the

various causes of the disorder make it difficult to treat the

disease simply by drugs or surgery In this section, an

arti-ficial urethral valve system driven by an SMA actuator is

introduced (11)

Figure 14 Stents made from laser-cut tubing (1).

Figure 15 Artificial urethral valve.

Urethral Valve The artificial urethral valve should be

compact and should have no protrusions when it is planted in the lower abdominal region In addition, itshould be attachable onto various sizes of urethrae A com-pact urethral cylindrical valve is presented in Figs 15and 16 The valve is 15 mm across and 20 mm long It iscomposed of two semicircular stainless steel shells 0.2 mmthick and a 0.2-mm circular-arc NitinolTMplate The shellsand the NitinolTM plate are fixed together by stainlesssteel clamps Further, a cylindrical sponge rubber filling isplaced inside the valve to effect uniform contact betweenthe valve and the urinary canal In the normal state, thevalve presses on the canal, so that it is choked To free thecanal, the valve is opened by actuating the SMA element.The SMA plate, which is cylindrical at body temperature,flattens as heat is increased, and the valve, which is closed

im-by the force of the bias spring in the normal state, is opened

to release the choked urethra and allow urinary flow Toheat the SMA, a NichromeTM wire, insulated by a poly-imide membrane, was placed on the surface of the nitinolplate

Transcutaneous Energy Transformer System The energy

to drive an in-dwelled valve should be supplied from side the body A transcutaneous energy transformer systemwill be effective for this purpose (12,13) The system con-sists of two induction coils that transmit electrical energywirelessly from the primary to the secondary coil In thisstudy, the primary coil was 70 mm in diameter and had

out-12 turns, and the secondary coil was 60 mm across andhad eight turns The coils are spirals formed of twistedwires that consist of 20 lengths of copper wire 0.2 mm indiameter

Experimental Setup Figure 17 is the schematic of an

an-imal experiment that used the urethrae of male dogs ofaverage weight 12 kg, whose thickness is similar to that

of human male urethrae The urinary canal uncovered bythe cut was equipped with the SMA valve and then loadedwith water at a hydrostatic pressure of 75 cmH2O, which

is comparable to human abdominal muscle pressure Thetemperature of the valve and the flow rate of the waterpassing through the canal were measured First, the valve

Sponge rubberUrethra

Flow meter

UrethraOutlet

Valve

Powersupply Dog

0 20 40 60

Time [s]

80 100 120 140

204060

Flow rateTemperature

Input voltage

2.5 V

Figure 18 Time variations of applied direct current, valve

temp-erature, and flow rate of water passing through the valve.

was heated by direct current from the transcutaneous ergy transformer system and then it was left to cool nat-urally by cutting the electric power to the heating wire

en-In the energy transformer system, the secondary coil wasplaced under the abdominal wall, and the primary coil onthe skin face-to-face with the secondary coil The capac-ity of the condenser to enhance the resonant effect of the

circuit was set at C2= 1.0 µH The frequency of the input

current, which should coincide with the resonant frequency

of the circuit, was 90 kHz Further, the peak-to-peakamplitude of the current was set at 25 V

Results and Discussion Figure 18 shows the

open-ing/closing functions of the valve when it was driven by thedirect current heating of the SMA The input current to theheating wire on the SMA, the temperature of the valve, andthe flow rate of water passing out through the urethra areplotted as functions of time The valve started to open, andthe water began to flow 20 seconds after electric currentwas supplied The current was cut off when the temper-ature of the valve reached 45◦C It is seen that the valvecompletely intercepted the water flow 30 seconds after theelectric current was cut off Thus, the valve possesses thenecessary function of an artificial urethral valve

Figure 19 illustrates the motor functions of thevalve when it was driven by the transcutaneous energy

0.02.0

Time [s]

204060

Flow rateTemperature

Energytransmitting

Figure 19 Opening/closing functions of valve driven by

transcu-taneous energy transformer system.

Heating

Cooling

SMASilicone sheets

Figure 20 Schematic of the artificial sphincter using SMA.

transformer system The distance between the primary

and secondary coils was 3–4 mm, and the induced current

to the heating wire was 7.6 V in the peak-to-peak

ampli-tude The figure shows that the valve started to open and

released the choked canal 6 seconds after the electric

cur-rent was supplied The curcur-rent was cut off when the

tem-perature of the valve reached 45◦C, which was 35 seconds

after the current supply was started It is shown that the

water flow was intercepted 25 seconds after the current

was cut off These motor functions of the valve showed

that the induction coil system worked well as a

transcu-taneous energy transformer for heating the artificial

ure-thral valve

Artificial Sphincter

Similar to the urethral valve, the development of an

arti-ficial sphincter is also required for the medical treatment

of patients who have fecal incontinence due to a colostomy,

a congenitally anorectal malformation, or surgical

opera-tions for anorectal diseases The lack of an anal

sphinc-ter is the main reason for the problem In this section, an

artificial sphincter using a SMA actuator was one of the

solutions of these problems, as introduced in (14)

Figure 21 Schema of animal experiment.

Artificial anusPressure sensor

Mechanism of Artificial Sphincter A schematic drawing

of the proposed artificial sphincter is shown in Fig 20 Theactuator consists of two SMA plates joined by two hingesand the heating coils attached to the SMA plates The SMAplates are 70 mm long The width and the thickness are 18and 0.7 mm, respectively The material used for the SMAplates, Ti51at%Ni, is known to exhibit an all-round shape-memory effect (ARSME); it reverses shape to the “mem-orized” shape in its martensitic phase Because the high-est temperature for the complete reverse transformationmight reach 55◦C, thermally insulated materials such ascork sheets and sponge rubber sheets cover the outer andinner sides of the SMA plates, respectively When elec-tric power is applied to the coils for heating, the reversetransformation occurs in the SMA plates, accompanied byshape changes from a flat shape to an arc, the restrainedshape during annealing The shape change results in agap between the two SMA plates to open the intestines.After switching off the electric power, the shape of the SMAplates recovers by natural cooling, and the intestines areclosed again

Animal Experiments Animal experiments have been

conducted to examine the fundamental functions and thebiocompatibility of the actuator The experiments were car-ried out on a pig that had the same dimensions of intestines

as those of the human body

Preliminary tests of the fundamental functions of theactuator were conducted As illustrated in Fig 21, an arti-ficial anus was made in the pigs abdomen using a rectumafter the resection of its colons Then the artificial sphincterwas installed around the rectum located between the ab-dominal wall and the peritoneum (see Fig 22) The insidepressure of the rectum generated by the artificial sphincterwas measured first A pressure sensor was inserted fromthe artificial anus into the intestine and then moved out.The pressure exhibited a rise of 50 mmHg in the regionclipped by the actuator, corresponding to the width of theSMA plates: 18 mm (see Fig 23) The pressure tests werealso carried out when the inner part of the rectum was

Figure 22 A photograph of the artificial sphincter installed in

the rectum of a pig.

subjected to pressure The inner pressure of the rectum

was generated by pouring gel into the intestines Figure 24

shows the pressure change measured during the test As

seen in Fig 24, the pressure was increased to 75 mmHg

without any leak of the gel, and then decreased by opening

the artificial sphincter using electric power Discharge of

the gel from the artificial anus was observed

A clinical test on a living pig has also been carried out

The artificial sphincter enables a controlled bowel

move-ment of the pig From the dissecting examination after six

days of experiments, neither infections nor burn scars were

observed in the body around the artificial sphincter

Fur-thermore, the artificial anus was in good condition; this

suggested that the pressure due to the artificial

sphinc-ter was tolerable Additionally, in these experiments,

50 mm Hg

Artificial sphincter Artificial anus

Figure 23 Pressure distribution in an intestine clipped by the

artificial sphincter.

20 sec

100 mm Hg

Figure 24 Pressure change in the rectum: (a) start pouring gel into the rectum; (b) switch on the

input power; (c) discharge of gel was observed.

electric power was supplied through a pair of current leadwires that penetrated the body The use of an inductivepower transmission for the power supplement of the actu-ator could lead to a complete implantation-type artificialsphincter based on further research

CURRENT BIOMEDICAL APPLICATIONS

OF PIEZOELECTRIC MATERIALS

Due to its piezoelectricity, a piezoelectric material has asensor function that can convert a mechanical signal to anelectric signal Electrical voltage generated by mechani-cal stress in piezoelectric materials decays very fast due

to charge dissipation The voltage signal takes the form

of a very brief potential wave at the onset of the appliedforce and a similar brief wave at termination It increases

as force is applied but drops to zero when the force mains constant There is no response during the stationaryplateau of the applied stimulus Voltage drops to a negativepeak as the pressure is removed and subsequently decays

re-to zero (15) The response is quite similar re-to the response

of the Pacinian corpuscle in the human skin (16),one of thesensory receptors in the dermis PVDF (polyvinylidene flu-oride) piezofilm is suitable for uses in the biomedical field,because it is very flexible and sensitive to the fast variation

of stress or strain In this section, several recent studies ofmedical applications of PVDF film are introduced

Active Palpation Sensor for Detecting Prostatic Cancer and Hypertrophy

Prostatic carcinoma and hypertrophy are examined in eral by rectal palpation where the doctor’s index finger isused as a probe together, in most cases, with ultrasonictomography The two lesions found in this study are diag-nosed by noticing their morphological features Prostatichypertrophy is a symmetrical enlargement of the prostateglands; the stiffness varies from soft to hard Prostaticcancer, on the other hand, is a hard asymmetrical uneventumor Palpation depends on the tactile perception of theforefinger, which is said to be ambiguous, subjective, andmuch affected by the physician’s experiences Hence, thedevelopment of a palpation sensor for detecting prostaticcancer and hypertrophy is important

gen-Palpation Sensor and Measurement System The geometry

of an active palpation sensor and tip probe are presented

PVDF film

PVDF film

1 mmSponge rubberVulcanized

in Figs 25 and 26 The tip probe is mounted on a linear

z-translation aluminum bar It fits into a cylindrical outer

aluminum shell 15 mm in diameter and is driven by a dc

micromotor and crank mechanism The drive mechanism

is essentially the same as that of an electric toothbrush

The probe is positioned so that its face is to the prostate

gland, and, it is oscillated at about 50 Hz at a constant

peak-to-peak amplitude of 2 mm The probe is an assembly

of layered media

The base is a thin aluminum circular plate 10 mm in

diameter, on which a cylindrical rubber sponge 8 mm in

diameter and 4 mm thick, a PVDF piezopolymer film 6 mm

across and 28µm thick as the sensory receptor, and a thin

acetate film as a protect cover for the piezopolymer film

are stacked in sequence Furthermore, a convex layer of

vulcanized rubber 3 mm across was placed on the

sur-face of the acetate film to enhance the sensitivity of the

sensor (17)

In the experiment, the sensor head is pressed

sinu-soidally against the object, and the output signal from the

piezopolymer film is collected for 100 ms, sent to a

digi-tal storage oscilloscope for every sampling time of 0.2 ms,

and further forwarded to a personal computer via a GP-IB

interface as 500 eight-bit data for processing

Signal Processing The output voltage from the

piezo-polymer film is proportional to the rate of the strain

in-duced in the film, which means that the maximum

ampli-tude of the signal from the sensor is rather superposed

by noises from the measuring system Bearing this fact in

mind, the following data analysis can be done by using the

absolute output signal of the sensor integrated across the

period of data collection:

Figure 26 Geometry of a sensory receptor.

where I is the integrated output signal and N (= 500) is

the total number of data collected in a period of 100 ms

Clinical Test Rectal palpation was done by using the

sensor to verify the discrimination function of the sor First, the doctor examined the stiffness of the prostategland of a person by using his own index finger Next, thesensor protected by a medical rubber glove was insertedinto the subject’s rectum The relative position of the sen-sor to the prostate gland was monitored by the ultrasonicdiagnostic, and the sensor was placed face to face to theprostate gland and driven to generate sensor output Theexaminees in this time were eight in total One 74-year-oldperson was suffering from a carcinoma Four persons ofaverage age 77 were patients who had prostatic hypertro-phy, and two persons (average age 52) suffered from pro-statitis The last person diagnosed had no definite lesions

sen-on his prostate gland The stiffness of the prostate glandsdiagnosed this time was “elastic soft or elastic firm,” “def-initely elastic firm,” and “hard.” The results obtained arepresented in Table 1, which shows the average, the mini-mum and maximum output, and the standard deviation

of the output It is seen clearly that the output from theprostate gland of the diagnosis “hard” is greater than theoutput of the gland of “elastic firm.”

Haptic Sensor for Monitoring Skin Conditions

Assessing the pharmaceutical action of liniments on skindisease is a matter of importance to dermatologists Thishas drawn much attention to the development of objectivetechniques for measuring the morphological features ofskin (18) Evaluating the substantiation of the cosmeticefficacy of toiletries is another of the objective measuringtechniques The features that affect the health appraisal

Figure 27 Tribosensor.

and/or the physical beauty of skin are the morphology such

as rashes, chaps, or wrinkles This noninvasive technology

has made great advances in the studies of dermatology

during the last decade, and several methods have been

de-veloped to measure the mechanical properties of the

der-mis such as measuring transepidermal water loss using

an evaporimeter (19) and image processing of a negative

replica of the dermis (20).These methods, however, fall in

the category of indirect measuring techniques for the

der-mis In this section, the development of haptic tribosensors

for monitoring skin conditions and distinguishing atophic

and normal healthy skins directly is introduced

Tribosensor A tactile sensor for measuring skin surface

conditions is presented in Figs 27 and 28 The sensor is a

layered medium, whose construction is analogous to that

of the human finger It is composed of an aluminum shell

as the phalanx, sponge rubber 3 mm thick as the digital

pulp, a PVDF piezopolymer film 28µm thick and 12 mm

across as the sensory receptor, an acetate film as a

protec-tive cover for the piezofilm, and gauze on the surface as the

fingerprint that enhances the tactile sensitivity of the

sen-sor The sensor was attached to the tip of an acrylic elastic

beam, and a strain gauge was mounted on the surface of

the beam to monitor the force applied to the skin by the

sensor

Aluminum pipeSponge rubberElectrodePVDFCellophane filmGauzePVDF

Figure 28 Schematic of PVDF piezofilm sensor.

Table 2 Segmented Frequency Range for Power Spectrum Integration

EarthArmSensor

Neck

Band-pass filter

GPIB

Computer Oscilloscope

Figure 29 Setup of measuring instruments.

Measuring Apparatus A measurement system using the

tribosensor is presented in Figure 29 The sensor wasmoved by hand over the sample skin to maintain a constantspeed and force The voltage signal from the PVDF sensoryfilm was sent to a digital storage oscilloscope as eight-bitquantitized digital signals of 4096 points and then trans-mitted to a personal computer via a GPIB board for signalprocessing The sampling frequency was held constant at

40 kHz In the process of measurement, it was necessary

to reduce the potential difference between the surface ofthe skin and the sensor to minimize the overlap of noises

on the sensor signal To this end, the subject and the sor were grounded by fitting a grounding conductor aroundthe wrist Still, some noises from the power sources, whichwere due to outside sources, were observed around 50 Hzand 100 MHz, respectively Thus, a band-pass filter thathad cutoff frequencies of 70 Hz and 20kHz was insertedafter the sensor to remove the noise effect

sen-Signal Processing and Identification of Skins Dermatitis

is a factor that affects the skin condition The neck skin ofsubject A, who was suffering from mossybacked atopic der-matitis, was compared with the healthy skins of subjectsB–E The sensor was moved by hand over a prescribed re-gion of skin to maintain a constant speed and force Hence,

a method of identifying sample skins employing signalprocessing and neural network-based training was intro-duced First, the variance was calculated as an index toextract the features of the collected data:

Table 3 Rate of Correct Answer for Discrimination of Atopic and Normal Skins

where, x( j) is the jth quantitized digital signal, ¯ x is the

average of x( j), and N is the total number of digital

signals

Next, FFT analysis was introduced to extract the

fea-tures of the collected data Careful reading of the spectra

obtained led to an understanding that there were some

dif-ferences in the distribution profiles of spectra among the

subjects In the light of this, the power spectrum was

seg-mented into several frequency ranges, and the ratio of the

power in an individual frequency range to the power of

the whole frequency range was calculated and used as the

second index that described the characters of the sample

skin:

(R s)i = (S) i /S, (S ) i=



i

i (S ) i , (3)

where P( f ) is the power spectrum density, S is the

to-tal power distributed across the frequency range

consid-ered, and (S ) i is the power at frequency level i Here, the

level i(=1–4) stands for the range of frequencies given in

Table 2

Discrimination of Skin that has Atopic Dermatitis The

neural network was trained to recognize atopic and

nor-mal healthy skins The neural network employed was

a hierarchical network; the training method used was

instructor-assisted training, in which the correct answers

were provided as an aid to the training; the training

algo-rithm employed was back-propagation Two networks were

examined One was the network that used the four power

ratios (R s)i , i = 1–4, as the cell input, and the other used

the variance V in addition to (R s)i Five sets of data were

ob-tained for each skin The neural network had an input

sen-sory layer of four or five cells, an intermediate association

layer that had the same number of cells as the input layer,

and an output response layer that had a single cell The

correct output value was set to one for the atopic skin, and

zero was set for the healthy skin Training was repeated

by using the gradient descent method until the error

func-tion became sufficiently small, up to 103 and 104 cycles

The networks were constructed separately three times for

three sets of randomly selected initial values of synaptic

weights After completing the training of 12 individual

net-works, the recognition experiments were performed on the

atopic skin and the healthy skin using the newly obtained

five sets of data for each subject The results obtained are

shown in Table 3 Here, the percentage denotes the rate

of the correct answer averaged across the three networks

of randomly selected initial synaptic weights Atopic matitis cannot be identified without the input of the vari-ance, and the recognition was perfect when the variancewas introduced as the input

3 C Trepanier et al., J Biomed Mater Res 48: 165 (1999).

4 K Tanaka, Tobuse, and S Miyazaki, Mechanical Properties of Shape Memory Alloys Yokendo, 1993.

5 K Otsuka and C.M Wayman, Shape Memory Materials.

Cambridge University Press, 1998.

6 Y Suzuki, Topics on Shape Memory Alloys Nikkan Kogyo

Shimbun, 1988.

7 Ishikawa, Kinashi, and Miwa, Collections of SMA tions Kogyo Chosakai, Tokyo, 1987.

Applica-8 H Ohnishi, Artif Organs 12: 862 (1983).

9 R Sachdeva and S Miyazaki, Proc MRS Int Meet Adv Mater.

Chonan Bio-Med Mater Eng 9: 97–112 (1999).

12 J.C Schuder, H.E Stephenson, Jr., and J.F Townsent, IRE

Internet Conv Rec Pt9–9: 119–126 (1961).

13 H Matsuki, M Shiiki, K Murakami et al., Flexible

transcu-taneous transformer for artificial heart system, IEEE Trans.

Magn 26: 1548–1550 (1990).

14 T Takagi, Y Luo, S Hara, T Yambe, S Amae, and M.W.H.

Nakamura J Soc Adv Sci (in press).

15 G Harsanyi, Sensing in Polymer Films in Sensor Applications,

Technomic, Lancaster, Basel, 1995, p 97.

16 G.M Shepherd, in Neurobiology 3e., Oxford University Press,

New York, Oxford, 1994, p 272.

17 S Chonan, Z.W Jiang, M Tanaka, T Kato, and M Kamei, Int.

J Appl Elect Mech 9: 25–38 (1998).

18 H Tagami, Fragrance J 10: 11–15 (1993).

19 T Yamamura, Measuring methods of skin surface water

con-tent, Fragrance J 10: 35–41 (1993).

20 M Takahashi, Fragrance J 10: 16–26 (1993).