Encyclopedia of Smart Materials (Vols 1 and 2) - M. Schwartz (2002) Episode 4 pot

par-Background on Cement-Matrix Composites Cement-matrix composites for smart structures includethose that contain short carbon fibers for sensing strain,damage, and temperature and for

Structural materialsProperties

Functional materials

Multifunctional materials

Intelligent materialsFunctions

PropertiesFunctionsInformationSimultaneous functions

Composite

materials

Monolithicmaterials

Figure 2 Evolution of the design of materials with

macrostruc-ture and microstrucmacrostruc-ture.

to domestic appliances and the medical supply industry

The reason for this universal utilization of these new

ma-terials is their extensive range of physical properties Their

physical properties could comply precisely with the design

specification of a product In the past, metallurgists were

primarily responsible for the development of new

mate-rials Now these new materials are synthesized by eclectic

teams of specialists

The trend of employing eclectic teams of specialists was

responsible for the explosive development during the

lat-ter half of the twentieth century of a plethora of

mate-rial classifications too numerous to discuss here These

include the development of a variety of functional

mate-rials, such as gallium arsenide or magentostrictive

ma-terials where the functional properties are exploited in

practice instead of the structural properties as in

tradi-tional practice However, there is one classification worthy

of mention because it is important commercially and is

cen-tral to chronicling the evolution of a branch of materials

science

This class of modern materials is the advanced

compo-sites These engineered materials are synthesized within

two distinct phases comprising a load-bearing materialhoused in a relatively weak protective matrix The rein-forcement is typically particles, whiskers, or fibers, whilethe matrix can be polymeric, ceramic, or metallic materials

A characteristic of these composite materials is that thecombination of two or more constituent materials creates

a material with engineering properties superior to those ofthe constituents—albeit at the expense of more challeng-ing fabrication technologies

As an aside, it should be noted that the field of fibrouscomposite materials technologies is not entirely new Con-sider the recording in Exodus, chapter 5 of the Bible, of theIsraelites manufacturing bricks from a mixture of clay andstraw for the pharaoh in Egypt in 450 BC Over a thou-sand years later, the French used a combination of horsehair and plaster to create ornate ceilings in stately homes.Advanced polymeric composite materials have affordedthe engineering community the opportunity to fabri-cate products with the strongest and stiffest parts perunit weight Furthermore, by appropriately designing themacrostructure of these materials, the engineer can de-velop composites with different properties in different di-rections, or alternatively, different properties in differentdomains of a structural member Thus, not only are thegeometrical and surface attributes of the part being de-signed, but in addition the material’s macrostructure isbeing design too The infusion of these materials into nu-merous industries has been responsible for the creation ofmany generations of products in the defense, automotive,biomedical, and sporting goods industries, for example.Thus at the dusk of the twentieth century the creation

of materials with an engineered macrostructure was thestate-of-the-art In the context of advanced fibrous com-posite materials, materials science had come the full cir-cle In the beginning, naturally occurring materials were

studied by Homo habilis to select the appropriate mineral

for the tip of the spear or arrow Now, one million years

later, Homo sapiens sapiens is studying naturally

occur-ring fibrous composite materials in order to emulate theplacement of fibers for the creation of synthetic compositeparts One of the primary motivations for the growth ofthis field of scientific endeavor is that naturally occurringfibrous composites are embodied in numerous species offlora and fauna They are one of the basic building blocks

of life They are an intrinsic design of Nature’s Clearly,

if these biological systems have evolved to these maturestates through the millennia, then they are indeed worthy

of study and perhaps emulation in an engineering context.This observation provides the underpinning of the subse-quent section: biologically inspired materials

Historically, then, each era during the evolution of mankind was motivated by the insatiable demand for supe-rior weapons and more innovative products In turn, thisdemand propelled the maturing of materials science, be-cause the triumvirate of product design comprises materi-als, manufacturing and conceptual design And of coursethis quest is very evident to this day Currently the age ofsynthetic materials has been driven by the same motives.While plastics shall undoubtedly remain an increasinglydominant component of modern lifestyles, the new mil-lennium shall witness the emergence of a superior class

hu-of advanced materials These smart materials, or

intelli-gent materials, materials with diverse characteristics that

mimic flora and fauna, shall be responsible for historians

classifying it as The Age of Smart Materials, because they

will dominate the materials selection process when

tech-nologists are seeking state-of-the-art materials

BIOLOGICALLY-INSPIRED CREATIVITY IN ENGINEERING

Although human genius through various inventions makes

instruments corresponding to the same ends It will never

discover an invention more beautiful nor more ready nor

more economical than does Nature, because in her inventions

nothing is lacking and nothing is superfluous.

Leonardo da Vinci (Ms RL 19115v; K/P 114r, Royal Library, Windsor)

These profound words penned some 500 years ago by

Leonardo, that brilliant painter, sculptor, draftsman,

archi-tect and visionary engineer, are the earliest that formally

recognize the power of Nature’s creativity in the fields of

natural science such as botany, zoology, and entomology

The naturally occurring materials and systems associated

with these academic disciplines have developed their

pro-perties and characteristics over millions of years through

processes of Darwinian evolution They have been required

to survive when subjected to dynamically changing

en-vironmental conditions Indeed, these conditions demand

that only the most adaptable and fittest survive

There-fore, these biological systems are truly optimal designs

engineered by Mother Nature in response to a set of

un-written design specifications Hence they merit meticulous

scrutiny and potential emulation by humankind

Since Charles Darwin’s seminal work in 1859, entitled

“On the Origin of Species by Means of Natural Selection,”

the notion of natural selection has remained the central

theme of evolutionary biology Thus the proposition is that

all life forms evolved because organisms with traits that

promoted reproduction and survival somehow passed on

those traits to future generations Organisms without

those traits simply became extinct They failed to survive

Indeed, this powerful assertion has motivated theories in

several disciplines beyond the central theme of materials

science Hence the proposition of learning from biological

systems to advance the field of engineering has credence

Evolutionary psychology, for example, proposes that the

human mind is not a vacuous medium, but instead

com-prises specialized mental protocols that were honed by the

solving of problems faced long ago Sociobiology, on the

other hand, employs natural selection and other

biologi-cal phenomena to explain the social behavior of animals

This emulation, or mimicking, of biological systems is

of-ten called biomimetics The name is derived from the Greek

bios (life) and mimesis (imitation) It can be employed by

creative designers to develop solutions to engineering

prob-lems through the use of a direct analogy between a

natu-rally occurring system and an engineering system Thus

a meticulous and comprehensive study of a living

organ-ism can yield invaluable insight into the subtleties of its

refined design atributes developed by a lengthy process of

evolution and optimization

Already during recent centuries Nature’s creativeprowess has been recognized and exploited by many giftedindividuals Consider, for example, Sir George Cayley’swork in 1810 when he was designing a low-drag shapefor his fixed wing flying machine He exploited his know-ledge of ichthyology to propose that the geometry of thewing cross section should mimic the streamlined low-dragcross section of the trout Sir Marc Isambard Brunel pro-posed the use of caissons to facilitate the underwater con-struction of civil engineering structures by the serendipi-tous observation of a shipworm tunneling in timber George

de Mistral developed the hook-and-loop fastener, such asthose manufactured by Velcro USA, Inc, by studying howcocklebur plants tenaciously adhered to his trousers af-ter he had walked through woodland thickets Of course,there are many others too who have used a direct analogybetween a biological system and an engineering system inorder to create a new artifact for the betterment of hu-mankind

In the context of the development of advanced als from studies of naturally occurring systems with fibers,such as the stalks of celery or the skin of a banana, thefibrous composites of engineering practice are an obviousconsideration They emulate the fibrous structures of mus-cles or plants For example, the structure of the humbletree comprises flexible cellulose fibers in a rigid lignin ma-trix On the other hand, the insect cuticle comprises chitinfibers in a proteinaceous matrix

materi-On a separate level of comparison between materialsengineered by humankind and those created by MotherNature, consider the variety of materials in both clas-sifications In 1990, it was estimated that the market-place contained about 60,000 different plastics This con-trasts sharply with the observation that there are only twogroups of substances from which almost all skeletal tissues

of animals and plants are formed These groups are theamino-acid-based proteins and polysaccharides that occurtogether in different proportions in the vast majority ofbiological materials Clearly, the apparently limited chem-istry of these naturally occurring materials is compensatedfor by the tremendous diversity of their microstructure Inaddition, the creation and manufacture of these diversemicrostructures is accomplished when subjected to a smallrange of temperature, humidity, and pressure These lim-ited conditions contrast markedly with the extreme ther-modynamic environments employed to produce plastics.Finally, these processes of design and manufacture occursimultaneously They occur in unison Again, this contrastssharply with the protocols implemented in a large percent-age of industrial enterprises where design departmentsand manufacturing departments function autonomously.They do not practice simultaneous engineering

SMART MATERIALS AND STRUCTURES: CURRENT NONCOMMERCIAL TECHNOLOGIES

Biological systems are extremely complex, and in the text of biomimetics, it is therefore only inevitable that aneclectic team of specialists must be assembled to pros-ecute the research if progress is to be assured in the

con-creation of new materials Teams of trained professionals

are mandatory because biomimetics lies at the boundaries

of numerous artificial disciplines that have been created

by the academic community in order to study the

phe-nomenological problems of life Thus a team with

exper-tise in materials science, biological sciences, biotechnology,

nanotechnology, molecular electronics, artificial

intelli-gence, data-processing, microprocessors, automatic control

theory, manufacturing, applied mathematics, chemistry,

physics and the like, is required, along with the

associ-ated equipment The best configuration would be a blend

of academics, industrialists, and government employees

The creation of this type of infrastructure is often a

chal-lenging undertaking Thus there is a need for large grants

and perhaps centers of excellence

Unfortunately, in many academic institutions the

tra-ditional research philosophy is for each professor to

seek their own funding Interdisciplinary collaboration is

minimal Unfortunately, this is an outcome of the academic

system where professors are promoted because of their

individual accomplishments and natural phenomena are

broken down into distinct fields and the artificial interfaces

are well defined This infrastructure hinders potential

in-teractions between individuals in adjacent fields Thus the

conventional model is of a single investigator guiding

grad-uate students and postdoctoral fellows Therefore there

needs to be a major shift in the current research paradigm

if progress in this field is to proceed effectively

However, using the typical single-discipline

non-all-embracing philosophy, humankind has indeed created a

multitude of macroscopically smart materials that mimic

naturally occurring materials While these material

sys-tems are still quite crude relative to Nature’s materials,

they do exhibit, at the most advanced level, the behavior of

muscles (by using actuators), nerves (by using sensors),

and brains (by using control schemes, microprocessors,

fuzzy logic, artificial intelligence, etc.) at various levels of

sophistication

The diverse structural materials, functional

materi-als, and multifunctional materials can be combined in

building-block style, as shown in Fig 3 The materials that

mimic naturally occurring materials can have, at the most

advanced levels sensors, actuators, information pathways,

control schemes, and microprocessors Some people

clas-sify these materials as intelligent materials because they

can respond effectively to dynamically changing

environ-mental conditions

An intelligent material could comprise a graphite-epoxy

fibrous polymeric structural material that houses some

embedded fiber optic Mach Zender interferometric strain

sensors and conventional surface-mounted electrical

resis-tance strain gauge strain sensors The signals from these

sensors could be processed by a microprocessor that

em-ploys an appropriate control algorithm to activate an

ar-ray of actuators exploiting piezoelectric and shape-memory

phenomena to create an optimal performance under

vari-able external stimuli

It should be noted that there are naturally several

lev-els of complexity in this field At the lowest level, a

mate-rial might feature only sensors to emulate the nerves of a

Sensor

Sensor

Sensor/actuator

StructureActuator

ActuatorStructure

MicroscopicMacroscopic

Figure 3 Classification of materials design.

biological material: photonic, piezoelectric, or conventionalstrain gauges, for example, depending on the design speci-fication Alternatively, the smart material might only fea-ture actuators that are controlled manually to emulate themuscles of a biological material These actuators include

a variety of functional materials that demonstrate the verse phenomena associated with piezoelectricity, electro-rheology, magnetorheology, magnetostriction, and elec-trostriction, for example They are typically excited by anelectrical stimulus that changes the geometrical configu-ration, stiffness, or energy dissipation characteristics in acontrolled manner At the highest level, the material mightinclude the brains of a biological system These complexentities can be imperfectly represented by the exploitationand integration of diverse theories and technologies asso-ciated with neural networks, rule-based systems, a mul-titude of control schemes, signal processing, nanotechno-lgy, and microprocessors, for example Figure 4 illustratessome of the primary components of thee materials.The last decade has witnessed an explosion of articles

di-on smart materials, and the first book dedicated to thefield was published in 1992 by the author While two jour-

nals are dedicated to this field—namely the Journal of Intelligent Material Systems and Structures and the Jour- nal of Smart Materials and Structures—the eclectic na-

ture of this field, coupled with the importance of als science in the product realization process, mandatesthat most academic journals now publish regularly arti-cles on some aspect of the field The World Wide Web is anobvious source of information on these activities, but the

materi-Microprocessor-basedcomputational capabilities

Real-time control

capabilities

Smart materials

Structural materialNetwork of actuators

Network of sensors

Figure 4 Ingredients of the premier class of smart materials.

professional wishing to enter this rapidly evolving field

would be well advised to attend the SPIE’s international

symposium on smart materials and structures for an

overview of current thinking

SMART MATERIALS AND STRUCTURES: THE FUTURE

Predicting the future of a technical field with any certainty

is fraught with errors and subsequent embarrassment

Therefore I will refrain from peering to the horizon with

binoculars, preferring to view the future without an

opti-cal aid After all, there are numerous records of truly gifted

individuals failing to accomplish this task of predicting the

evolution of a scientific disciplines or a technological field

For example, the famous British scientist Lord Kelvin, who

was president of the Royal Society in 1895, stated that

“Heavier than air flying machines are impossible.” Four

years later, Charles H Duell, commissioner of the U.S

Office of Patents, commented that “Everything that can be

invented has been invented.” More recently, in 1977, Ken

Olsen, the founder of Digital Equipment Corporation, said

“There is no reason for any individuals to have a computer

in their home.” Thus the evidence is clear Nevertheless,

fu-ture objectives will be explored on the tenet that engineers

and scientists should mimic naturally occurring systems

as one method of advancing the state-of-the art

The most sophisticated class of smart materials in the

foreseeable future will be that which emulates

biologi-cal systems This class of multifunctional materials will

possess the capability to select and execute specific

func-tions intelligently in order to respond to changes in the

local environment Furthermore, these materials could

have the ability to anticipate challenges based on the

ability to recognize, analyze, and discriminate These

ca-pabilities should include diagnosis, repair,

self-multiplication, self-degradation, self-learning, and

home-ostasis

Table 1 Nature’s Rules for Sustaining Ecosystems

Use waste as a resource Diversify and cooperate to fully use the habitat Gather and use energy efficiently

Optimize rather than maximize Use material sparingly Don’t foul nests Don’t draw down resources Remain in balance with the biosphere Run on information

Shop locally

Environmental Design Issues

The maturing nexuses between professionals in the logical and engineering communities have been responsi-ble for a deeper appreciation of the awesome prowess ofNature’s creations This appreciation should be used to es-tablish new operating procedures for the design and manu-facture of the next generation of materials They canon can

bio-be distilled from studies of complex ecosystems that haveevolved through the millennia These winning strategieshave been adopted by diverse organisms; therefore, theyhave been employed by both animals and plants compris-ing diverse organs and parts This diversity neverthelessdoes not defer their harmonious functioning together tosustain life and its activities Indeed, these diverse organ-isms provide a set of rules that are worthy of analysis andpotential implementation in engineering practice by hu-

mankind After all, Homo sapiens sapiens is an organism

too They are presented in Table 1

The information in Table 1 has very broad tions in terms of environmental design protocols for theengineering profession, in addition to the creation of newclasses of smart materials

ramifica-Health Monitoring of Smart Materials

There is a cycle in all ecosystems Biological systems perience life and death Thus there is a recycling of thematerial Mature trees in a forest absorb nutrients fromthe soil, but they eventually die and collapse to the forestfloor There they crumble and decompose, providing enrich-ment to the soil This enrichment provides nourishment forthe next generation of trees and, of course, other forms ofvegetation too

ex-It appears therefore that materials that emulate thisnatural cycle could be developed Indeed, there are already

a number of biodegradable materials on the market thatdecompose after serving their useful purpose In additionthis philosophy has been responsible for new engineeringprotocols concerning the environment Thus parts are be-ing recycled, others are being re-manufactured, and there

is great concern for reducing the consumption of naturalresources

The continuous monitoring of the health of a part orproduct is one of the many new ideas set for developmentbased on the re-evaluation of practices in the context ofnaturally occurring systems The engineering approach

Birth (cure monitoring)

Life (use/performance feedback)

Death (fatigue/failure monitoring)

Figure 5 Health monitoring of Homo sapiens sapiens.

requires sensors to be embedded in a material, or attached

to the surface of a part, in order to provide information on

the behavior of the engineering artifact or its structural

integrity It is called “health monitoring” or the

cradle-through-grave” approach This approach of continually

monitoring the status of a part during manufacturing,

ser-vice, and failure mimics the spasmodic health-monitoring

activities of Homo sapiens sapiens Noninvasive techniques

are often employed to monitor the health of individuals

be-fore appropriate medical action is prescribed These

tech-niques, illustrated in Fig 5, often begin in the womb and

continue throughout life in response to various stimuli and

conditions

The analogous engineering situations employ in situ

sensors to monitor the behavior of the part during its

man-ufacture, the service life, and at failure as shown in Fig 6

Thus the sensing system would be employed initially to

monitor the state of cure of a smart part during the

fab-rication process, in an autoclave perhaps Subsequently,

the same system would monitor continuously the health

of the part during the service life For example,

dynami-cal stresses could be monitored and compared with limits

imposed by the design specification If limits are exceeded,

Death (critical monitoring)Life (health monitoring)Birth (ultrasound)

Figure 6 Health monitoring of smart structures.

then the operational envelope of the machine could be tomatically changed to restore it to the desired domain.Critical parts, subjected to complex fatigue environ-ments, could be monitored for structural integrity and im-pending failure These would include primary structures

au-of aircraft and large civil engineering structures such asdams, pipelines, and highway bridges The task of main-taining these structures by maintenance and inspectionwould be greatly enhanced by this health-monitoring ca-pability

Intelligence in Biological Materials

The intelligence to be imbued in a synthetic materialdeveloped by humankind should emulate the intelli-gent attributes found in biological systems These at-tributes do not require human involvement, and theyfunction autonomously, as evidenced by self-learning, self-degradation, and regeneration Thus the rusting of iron

in a humid environment could be considered to be a ple form of self-degradation Other functions could includethe availability to recognize and subsequently discrimi-nate, redundancy, hierarchical control schemes, and the

sim-election of an appropriate action based on sensory data.

Furthermore, a material that has been damaged and is

undergoing a process of self-repair would reduce its level

of performance in order to survive

Intelligence that should be inherent in future

genera-tions of smart materials can comprise several categories

that are derived from studies of biological systems These

are highlighted, with some of the consequences, for the

va-riety of terms listed below

Crystal Structure Changes occur in the orientation of

crystals, in the atomic configuration, and in the teratomic spacings: These changes are responsiblephase transformations, and in polymeric materials,the molecular chain can be re-configured from afolded state to an extended state

in-Molecular Structure Changes occur in the molecular

structure caused by molecular chains breaks, figured intramolecular bonds, three-dimensional in-tramolecular spacings, and antigenic or enzymaticreactions

recon-Macroscopic Structure Global changes occur in the

macroscopic structure after the diffusion and bulktransfer of fluids and ultra-fine powder

Composition The interaction between a material and

the neighboring environment at the interface results

in a change in the material composition at the surfacebecause of the attendant chemical reactions

Interfaces Interfacial changes typically pertain to grain

boundary phenomena and also reactions at the face of a material resulting from interactions withthe adjacent environment

sur-Energy The interaction between a material and the

ad-jacent environment at the interface trigger an energychange or the release of electrons or photons

Ion Transfer Ions become transformed ion materials

and in addition radicals are transferred along mer chains

poly-Charge Transfer poly-Charge transfer occurs by conductivity

in metals and charges are transferred along polymerchains in organic materials

Electronic Structure The magnetic properties of

mate-rials are changed by changes in the orientation ofelectronic spin vectors

This is the main vocabulary being used as attention

shifts to the attributes of biological materials and how they

can relate to the development of intelligence in smart

ma-terials The following paragraphs highlight some of the

primary attributes of biological systems and their

inter-pretation in the context of materials science Some are

il-lustrated in Fig 7

Mitosis Self-multiplication, self-breeding, or growth in

a biological system involves a cell creating similarcells by mitosis to replicate itself To mimic these pro-cesses, smart materials will be self-producing withthe additional implicit constraint of terminating theprocess when a prescribed state has been achieved

Intelligence from the human standpoint

Learning

Standby

HomeostasisSelf-adaptation/

surrounding-adjustment

Time-functionalresponsiveness

Otherintelligence

Prediction/

notification

Self-assembly Self-repair

multiplication

Self-Self-diagnosis(internal)Redundancy Autolysis

Intelligence at the most primitive levels in materials

(primitive functions)

Figure 7 Intelligent functions inherent in biological materials.

This might require changes in their molecular orcrystalline structure through the absorption of sub-stances from the neighboring environment or otherregions of the material

Self-repair This biological activity is closely related to

the self-multiplication function or growth It requiresthe identification of damage and the extent of thedamage before the repair process is initiated andthe material restored to its state of normality Thestate is manifest in materials that undergo changes

in crystalline structure and in the interfacial ditions at the surface of the material or at grainboundaries Future classes of smart materials couldautonomously regain their original shape throughphase transformation, even after the material hassuffered permanent deformation from surface im-pact

con-Autolysis Biological systems die and decompose when

they sustain severe injuries or damage, or they cease

to receive nutrition Smart materials with these tributes would decompose upon completion of their

at-useful life and be assimilated within the ment Smart materials with this attribute would beable to change their molecular structure or theirmacrostructure.

environ-Redundancy To enhance survivability, biological

sys-tems possess some degree of redundancy in theirstructures and functions in order to survive Smartmaterials with this attribute would feature redun-dant functions that would be dormant under nor-mal conditions However, as the material structurechanges in response to dynamic environmental con-ditions, the microstructure would be reconfigured tomaintain equilibrium through the activation of newfunctions These functions would trigger a change ofthe molecular structure of the crystalline structure

For example, a stress-induced transformation might

be triggered at the tip of a crack in order to reducethe stress concentration in that region and hinderthe crack propagation Alternatively, parts that de-mand high reliability and are being subjected to veryhigh loads could embody an innovative function thattriggers a phase transformation in order to develophigher strength properties

Learning The ability to learn is fundamental to many

aspects of the behavior of biological systems ing would be manifest in smart materials throughchanges in their physical constants, and changes inthe molecular or crystalline structure The associ-ated knowledge-base is associated with some innateattributes; others are acquired through experienceand through interactions with the local environment

Learn-These may involve inductive logic where general clusions are distilled from the observations of anevent Smart materials would benefit from this abil-ity and respond more appropriately to external stim-uli through the recollection of previous experiences

con-Some of the underpinnings would include sensoryabilities, data processing, control schemes, and ac-tuators

Autonomous Diagnosis Some biological systems contain

a self-diagnosis health-monitoring capability thatpermits the identification of problems, degradations,malfunctions, and judgmental errors This is facil-itated by a comparison between the current condi-tions and the past Classes of smart materials would

be able to monitor their own welfare so that the fects of damage on the performance of the materialcould be ascertained and corrective measures imple-mented

ef-This class of functions would be achieved bymolecular structural changes, changes in the crys-talline structure, or changes in the interfacial proper-ties at the surface or the grain boundaries Typicallychanges occur in materials that are traditionallyused in nonequilibrium states, and their functionalproperties change when they reach a state of equi-librium In addition, changes are found in materialsthat determine autonomously the appropriate time

to quickly terminate their functional behavior in sponse to the ambient environmental conditions, and

re-in materials that detect degradation before they ger a stress-induced transformation that reveals thedamaged state, perhaps using energy

trig-Prediction Some biological systems can predict the

im-mediate future and take appropriate action by ploiting sensory data and by learning from their pastexperiences Classes of smart materials designed toemulate these biological systems would need to use

ex-a combinex-ation of sensors, control ex-algorithms, ex-andknowledge This is manifest in materials that un-dergo an energy change, a change in their molecu-lar structure, or a change in the crystalline struc-ture Examples include the change in the crystallinestructure associated with a phase change, the trans-fer of electrons precipitated by the re-configuring ofthe crystalline structure and stress induced transfor-mations

Standby Smart materials should embody the state of

readiness for action displayed by biological systems.This state would probably require several subfunc-tions such as sensing, diagnosis, prediction, learn-ing, and some class of actuators This responsivenesscould be governed by the role of the material and theability to analyze the dynamically changing exter-nal environmental conditions It would be manifest

by changes in energy, molecular structure, and talline structure

crys-Information Integration Biological systems generally

evolve by storing the integrated experiences of ous generations extending over a great period of timebefore it is transferred to subsequent generations.Thus the emulation of genes or DNA would providethe basis for the creation of materials with innova-tive functions that integrate and maintain memorybanks

previ-Recognition Organisms frequently possess the ability

to recognize and discriminate when evaluating formation Smart materials with this ability wouldnot only embody analytical skills but also some mea-sure of fuzzy logic to interpret the data from sensingfunctions

in-Homeostasis Biological systems are generally

sub-jected to dynamically changing environmental ditions that cause internal systems and struc-tures to change continually Such systems surviveand maintain stable physiological states by coor-dinated responses that autonomously compensatefor these ever-changing external conditions Innova-tive functions would need to be established to ac-complish these tasks These functions would requireclosed-loop feedback systems where sensory informa-tion from both the input and the output states need

con-to be compared Furthermore, the desired conditionswould require the careful orchestration of sensors,processors and actuators

Feedback is manifest in a variety of formsincluding the transfer of charges, the change incomposition, the change in molecular structure, thetransfer of radicals and ions, and the change in thecrystalline structure Thus, for example, materials

with a transformation temperature would undergochanges when external conditions subject the mate-rial to an elevated temperature.

Adaptation Biological systems adapt to ever-changing

environmental conditions through the evolution oftheir physiological state These abilities would havegreat utility in engineering practice Examples of thisclass of materials include films of lubricants that as-sume solid or fluidic states depending on the localthermal or dynamic conditions Others include mate-rials that control their optical properties in response

to changes in external stimuli like magnetic fields,electrical fields, or heat

There are probably other biological attributes that are

worthy of emulation in the design and manufacture of

smart materials that mimic naturally occurring

materi-als, but this list provides a beginning The road ahead will

have many positive and negative undulations and many

winding turns The primary provinces will include

theo-retical studies, experimental studies, and computational

studies, followed by design and manufacture, the creation

of primitive functions, assembly and integration of these

functions, material characterization, and the development

of the supporting technologies

CONCLUDING COMMENTS

With the passing of time, the technologies associated with

materials are becoming ever more sophisticated It is

un-reasonable to contemplate the receding of this tide of

knowledge and complexity because it is motivated by both

commercial and military demands for superior materials

Composite materials, based on ideas emulating from

nat-urally occurring materials, could be considered the most

sophisticated class of materials created by humankind If

this is true, then it is only natural to invoke induction and

suggest that biological materials are the basis from which

new generations of materials will be developed

Smart structures are important because of their use in

hazard mitigation, structural vibration control, structural

health monitoring, transportation engineering, and

ther-mal control Research on smart structures has emphasized

incorporating of various devices in a structure to provide

sensing, energy dissipation, actuation, control, or other

functions Research on smart composites has emphasized

incorporating of a smart material in a matrix to enhance

smartness or durability Research on smart materials has

emphasized the study of materials (e.g., piezoelectric terials) used for making the devices However, relativelylittle attention has been given to the development of struc-tural materials (e.g., concrete and composites) that caninherently provide some of the smart functions, so thatthe need for embedded or attached devices is reduced oreliminated, thereby lowering cost, enhancing durability,increasing the smart volume, and minimizing mechanicalproperty degradation (which is usually caused by embed-ded devices)

ma-Smart structures are structures that can sense certainstimuli and respond to the stimuli appropriately, some-what like a human being Sensing is the most fundamentalaspect of a smart structure A structural composite, which

is itself a sensor, is multifunctional

This article focuses on structural composites for smartstructures It addresses cement-matrix and polymer-matrix composites The smart functions addressed includestrain sensing (for structural vibration control and trafficmonitoring), damage sensing (both mechanical and ther-mal damage related to structural health monitoring), tem-perature sensing (for thermal control, hazard mitigation,and structural performance control), thermoelectricity (forthermal control and saving energy), and vibration reduc-tion (for structural vibration control) These functionalabilities of the structural composites have been shown inthe laboratory Applications in the field are forthcoming

CEMENT-MATRIX COMPOSITES FOR SMART STRUCTURES

Cement-matrix composites include concrete (containingcoarse and fine aggregates), mortar (containing fine ag-gregate but no coarse aggregate), and cement paste (con-taining no aggregate, whether coarse or fine) Other fillers,called admixtures, can be added to the mix to improve theproperties of the composite Admixtures are discontinuous,

so that they can be included in the mix They can be ticles such as silica fume (a fine particulate) and latex (apolymer in the form of a dispersion) They can be shortfibers such as polymer, steel, glass, or carbon fibers Theycan be liquids such as aqueous methylcellulose solutions,water reducing agents, and defoamers Admixtures to ren-der the composite smart while maintaining or even im-proving the structural properties are the focus of thissection

par-Background on Cement-Matrix Composites

Cement-matrix composites for smart structures includethose that contain short carbon fibers (for sensing strain,damage, and temperature and for thermal control), shortsteel fibers (for sensing temperature and for thermal con-trol), and silica fume (for vibration reduction) This sec-tion provides background on cement-matrix compositesand emphasizes the carbon-fiber cement-matrix compositedue to its dominance among intrinsically smart cement-matrix composites

Carbon-fiber cement-matrix composites are structuralmaterials that are gaining in importance quite rapidly due

to the decrease in carbon-fiber cost (1) and the

increas-ing demand for superior structural and functional

proper-ties These composites contain short carbon fibers, typically

5 mm long; the short fibers can be used as an admixture

in concrete (whereas continuous fibers cannot be simply

added to the concrete mix), and short fibers are less

expen-sive than continuous fibers However, due to the weak bond

between carbon fiber and the cement matrix, continuous

fibers (2–4) are much more effective than short fibers in

re-inforcing concrete Surface treatment of carbon fiber [e.g.,

by heating (5) or by using ozone (6,7), silane (8), SiO2

parti-cles (9), or hot NaOH solution (10)] is useful for improving

the bond between fiber and matrix, thereby improving the

properties of the composite Surface treatment by ozone or

silane improves the bond due to the enhanced wettability

by water Admixtures such as latex (6,11), methylcellulose

(6), and silica fume (12) also improve the bond

The effect of carbon fiber addition on the properties of

concrete increases with fiber volume fraction (13), unless

the fiber volume fraction is so high that the air void

con-tent becomes excessively high (14) (The air void concon-tent

increases with fiber content and air voids tend to have a

negative effect on many properties, such as compressive

strength.) In addition, the workability of the mix decreases

with fiber content (13) Moreover, the cost increases with

fiber content Therefore, a rather low volume fraction of

fibers is desirable A fiber content as low as 0.2 vol.% is

ef-fective (15), although fiber contents exceeding 1 vol.% are

more common (16–20) The required fiber content increases

with the particle size of the aggregate, because flexural

strength decreases as particle size increases (21)

The effective use of carbon fibers in concrete requires

dispersing of the fibers in the mix Dispersion is enhanced

by using silica fume (a fine particulate) as an admixture

(14,22–24) Typical silica fume content is 15% by weight of

cement (14) The silica fume is usually used along with a

small amount (0.4% by weight of cement) of

methylcellu-lose to help the dispersion of the fibers and the workability

of the mix (14) Latex (typically 15–20% by weight of

ce-ment) is much less effective than silica fume in helping

fiber dispersion, but it enhances the workability, flexural

strength, flexural toughness, impact resistance, frost

resis-tance, and acid resistance (14,25,26) The ease of dispersion

increases with decreasing fiber length (24)

The structural properties improved by carbon fiber

addi-tion are increased tensile and flexible strengths, increased

tensile ductility and flexural toughness, enhanced

im-pact resistance, reduced drying shrinkage, and improved

freeze–thaw durability (13–15,17–25,27–38) Tensile and

flexural strengths decrease as specimen size increases, so

that the size effect becomes larger as the fiber length

in-creases (39) Low drying shrinkage is valuable for large

structures, for repair (40,41), and in joining bricks in a

brick structure (42,43) The functional properties created

by carbon fiber addition are strain sensing (7,44–58) (for

smart structures), temperature sensing (59–62), damage

sensing (44,48,63–65), thermoelectric behavior (60–62),

thermal insulation (66–68) (to save energy for buildings),

electrical conduction (69–78) (to facilitate cathodic

protec-tion of embedded steel and to provide electrical

ground-ing or connection), and radio wave reflection/absorption

(79–84) (for electromagnetic interference or EMI ing, for lateral guidance in automatic highways, and fortelevision image transmission)

shield-In relation to structural properties, carbon fibers pete with glass, polymer, and steel fibers (18,27–29,32,36–38,85) Carbon fibers (isotropic pitch-based) (1,85) areadvantageous in their superior ability to increase the ten-sile strength of concrete, even though the tensile strength,modulus, and ductility of isotropic pitch-based carbonfibers are low compared to most other fibers Carbon fibersare also advantageous in relative chemical inertness (86).PAN-based carbon fibers are also used (17,19,22,33), al-though they are more commonly used as continuous fibersthan short fibers Carbon-coated glass fibers (87,88) andsubmicron diameter carbon filaments (77–79) are even lesscommonly used, although the former are attractive for thelow cost of glass fibers and the latter are attractive fortheir high radio wave reflectivity (which results from theskin effect) C-shaped carbon fibers are more effective forstrengthening than round carbon fibers (89), but their rel-atively large diameter makes them less attractive Carbonfibers can be used in concrete together with steel fibers;the addition of short carbon fibers to steel-fiber-reinforcedmortar increases the fracture toughness of the interfacialzone between the steel fiber and the cement matrix (90).Carbon fibers can also be used in concrete together withsteel bars (91,92) or together with carbon-fiber-reinforcedpolymer rods (93)

com-Carbon fibers are exceptional in most functional erties, compared to the other fiber types Carbon fibers areelectrically conducting, in contrast to glass and polymerfibers, which are not conducting Steel fibers are conduct-ing, but their typical diameter (≥60 µm) is much largerthan the diameter of a typical carbon fiber (15µm) The

prop-combination of electrical conductivity and small diametermakes carbon fibers superior to the other fiber types in thestrain sensing and electrical conduction However, carbonfibers are inferior to steel fibers for thermoelectric compos-ites, due to the high electron concentration in steel and thelow hole concentration in carbon

Although carbon fibers are thermally conducting, theaddition of carbon fibers to concrete lowers the thermalconductivity (66), thus allowing applications for thermalinsulation This effect of carbon fiber addition is due to theincrease in air void content The electrical conductivity ofcarbon fibers is higher than that of the cement matrix byabout eight orders of magnitude, whereas the thermal con-ductivity of carbon fibers is higher than that of the cementmatrix by only one or two orders of magnitude As a result,electrical conductivity is increased upon carbon fiber addi-tion despite the increase in air void content, but thermalconductivity is decreased upon fiber addition

The use of pressure after casting (94) and extrusion (95,96) can result in composites that have superior microstruc-ture and properties Moreover, extrusion improves the sha-pability (95)

Cement-Matrix Composites for Strain Sensing

The electrical resistance of strain-sensing concrete out embedded or attached sensors) changes reversibly with

−2

−101

Figure 1 Variation of the fractional change in volume electrical

resistivity and of the strain (negative for compressive strain) with

time during dynamic compressive loading at increasing stress

am-plitudes within the elastic regime for a carbon-fiber latex cement

paste after 28 days of curing.

strain, such that the gauge factor (fractional change in

re-sistance per unit strain) is up to 700 under compression

or tension (7,44–58) The resistance (dc/ac) increases

re-versibly under tension and decreases rere-versibly upon

com-pression due to fiber pull-out upon microcrack opening

(<1 µm) and the consequent increase in fiber–matrix

con-tact resistivity The concrete contains as little as 0.2 vol.%

short carbon fibers, which are preferably those that have

been surface-treated The fibers do not need to touch one

another in the composite The treatment improves

wetta-bility with water The presence of a large aggregate

de-creases the gauge factor, but the strain-sensing ability

re-mains sufficient for practical use Strain-sensing concrete

works even when data acquisition is wireless The

applica-tions include structural vibration control and traffic

mon-itoring

Figure 1 shows the fractional change in resistivity along

the stress axis, as well as the strain during repeated

com-pressive loading at an increasing stress amplitude for

carbon-fiber latex cement paste after 28 days of curing

The strain varies linearly with the stress up to the highest

stress amplitude The strain returns to zero at the end of

each loading cycle The resistivity decreases upon loading

in every cycle (due to fiber push-in) and increases upon

un-loading in every cycle (due to fiber pull-out) The resistivity

has a net increase after the first cycle, due to damage Little

further damage occurs in subsequent cycles; the resistivity

after unloading does not increase much after the first cycle

The greater the strain amplitude, the more the

resistiv-ity decreases during loading, although the resistivresistiv-ity and

strain are not linearly related The effects in Fig 1 were

also observed in carbon-fiber silica-fume cement paste

af-ter 28 days of curing

Figures 2 and 3 show the fractional changes in the

longitudinal and transverse resistivities, respectively, for

carbon-fiber silica-fume cement paste after 28 days of

cur-ing durcur-ing repeated unaxial tensile loadcur-ing at increascur-ing

strain amplitudes The strain essentially returns to zero at

0

0.20.40.60.811.21.41.61.8

100.050.10.150.2

20 30Time (s)

5)

Figure 2 Variation of the fractional change in longitudinal

elec-trical resistivity (solid curve) and of strain with time (dashed curve) during dynamic uniaxial tensile loading at increasing stress amplitudes within the elastic regime for a carbon-fiber silica-fume cement paste.

0.010.0080.012

Figure 3 Variation of the fractional change in transverse

electri-cal resistivity (solid curve) and of strain with time (dashed curve) during dynamic uniaxial tensile loading at increasing stress am- plitudes within the elastic regime for a carbon-fiber silica-fume cement paste.

the end of each cycle, indicating elastic deformation Thelongitudinal strain is positive (i.e., elongation); the trans-verse strain is negative (i.e., shrinkage due to the Poissoneffect) Both longitudinal and transverse resistivities in-crease reversibly under uniaxial tension The reversibility

of both strain and resistivity is more complete in the gitudinal direction than in the transverse direction Thegauge factor is 89 and−59 for the longitudinal and trans-verse resistances, respectively

lon-Figures 4 and 5 show corresponding results for fume cement paste The strain is essentially totally re-versible in both the longitudinal and transverse directions,but the resistivity is only partly reversible in both direc-tions, in contrast to the reversibility of the resistivity whenfibers are present (Figs 2 and 3) As with fibers, both longi-tudinal and transverse resistivities increase under uniax-ial tension However, the gauge factor is only 7.2 and−7.1

silica-for Figs 4 and 5, respectively

Comparison of Figs 2 and 3 (with fibers) with Figs 4 and

5 (without fibers) shows that fibers greatly enhance the

0 10 200

0.0040.0080.0120.016

00.5

1.51

22.5

30Time (s)

Figure 4 Variation of the fractional change in longitudinal

elec-trical resistivity (solid curve) and of strain with time (dashed

curve) during dynamic uniaxial tensile loading at increasing

stress amplitudes within the elastic regime for a silica-fume

ce-ment paste.

magnitude and reversibility of the resistivity effect The

gauge factors are much smaller when fibers are absent

The increase in both longitudinal and transverse

resis-tivities under uniaxial tension for cement pastes, whether

with or without fibers, is attributed to defect (e.g.,

micro-crack) generation In the presence of fibers, fiber

bridg-ing across microcracks occurs, and slight fiber pull-out

occurs under tension, thus enhancing the possibility of

microcrack closing and causing more reversibility in the

resistivity change The fibers are much more electrically

conductive than the cement matrix The presence of the

fibers introduces interfaces between fibers and matrix The

degradation of the fiber-matrix interface due to fiber

pull-out or other mechanisms is an additional type of defect

generation which will increase the resistivity of the

com-posite Therefore, the presence of fibers greatly increases

the gauge factor

The transverse resistivity increases under uniaxial

ten-sion, even though the Poisson effect makes the transverse

strain negative This means that the effect of transverse

0.0008

0.00120.001

0.00160.0014

Figure 5 Variation of the fractional change in transverse

electri-cal resistivity (solid curve) and of strain with time (dashed curve)

during dynamic uniaxial tensile loading at increasing stress

am-plitudes within the elastic regime for a silica-fume cement paste.

resistivity increase overshadows the effect of transverseshrinkage The increase in resistivity is a consequence

of uniaxial tension In contrast, under uniaxial sion, the resistance in the stress direction decreases after

compres-28 days of curing Hence, the effects of uniaxial tension

on transverse resistivity and of uniaxial compression onlongitudinal resistivity are different; the gauge factors arenegative and positive for these cases, respectively.The similarity of the resistivity change in the longitudi-nal and transverse directions under uniaxial tension sug-gests similarity for other directions as well This meansthat the resistance can be measured in any direction tosense the occurrence of tensile loading Although the gaugefactor is comparable in both longitudinal and transverse di-rections, the fractional change in resistance under uniaxialtension is much higher in the longitudinal direction than inthe transverse direction Thus, the use of the longitudinalresistance for practical self-sensing is preferred

Cement-Matrix Composites for Damage Sensing

Concrete, with or without admixtures, can sense major andminor damage—even damage during elastic deformation—due to the increase in electrical resistivity that accom-panies damage (44,48,63–65) That both strain and dam-age can be sensed simultaneously through resistance mea-surement means that the strain/stress condition (duringdynamic loading) under which damage occurs can beobtained, thus facilitating damage origin identification.Damage is indicated by an increase in resistance, which

is larger and less reversible when the stress amplitude

is higher The increase in resistance can be sudden ing loading It can also be a gradual shift of the baselineresistance

dur-Figure 6 (64) shows the fractional change in resistivityalong the stress axis, as well as the strain during repeatedcompressive loading at increasing stress amplitude forsilica-fume cement paste after 28 days of curing The strainvaries linearly with stress up to the highest stress ampli-tude The strain returns to zero at the end of each loading

0.10.050.15

0.250.2

0.30.35

Figure 6 Variation of the fractional change in electrical

resis-tivity and of strain (negative for compressive strain) with time during dynamic compressive loading at increasing stress ampli- tudes within the elastic regime for a silica-fume cement paste after

28 days of curing.

Figure 7 Fractional change in resistance and strain during

re-peated compressive loading at increasing and decreasing stress

amplitudes, the highest of which was 60% of the compressive

strength, for carbon-fiber concrete after 28 days of curing.

cycle During the first loading, the resistivity increases

due to damage initiation During the subsequent

unload-ing, the resistivity continues to increase, probably due

to opening of microcracks generated during loading

Dur-ing the second loadDur-ing, the resistivity decreases slightly

as the stress increases up to the maximum stress of the

first cycle (probably due to the closing of microcracks) and

then increases as the stress increases beyond this value

(probably due to the generation of additional microcracks)

During unloading in the second cycle, the resistivity

increases significantly (probably due to the opening of

microcracks) During the third loading, the resistivity

es-sentially does not change (or decreases very slightly) as the

stress increases to the maximum stress of the third cycle

(probably due to the balance between microcrack

genera-tion and microcrack closing) Subsequent unloading causes

the resistivity to increase very significantly (probably due

to the opening of microcracks)

Figure 7 shows the fractional change in resistance

and strain during repeated compressive loading at

in-creasing and dein-creasing stress amplitudes for carbon-fiber

(0.18 vol.%) concrete (with fine and coarse aggregates)

af-ter 28 days of curing The highest stress amplitude is

60% of the compressive strength A group of cycles in

which the stress amplitude increases cycle by cycle and

then decreases cycle by cycle back to the initial low stress

amplitude is here referred to as a group Figure 7 shows the

results for three groups The strain returns to zero at the

end of each cycle for any of the stress amplitudes,

indicat-ing elastic behavior The resistance decreases upon loadindicat-ing

in each cycle, as in Fig 1 An extra peak at the maximum

stress of a cycle grows as the stress amplitude increases,

resulting in two peaks per cycle The original peak (strain

induced) occurs at zero stress, and the extra peak (damage

induced) occurs at maximum stress Hence, during loading

from zero stress within a cycle, the resistance drops, then

increases sharply, and reaches the maximum resistance of

the extra peak at the maximum stress of the cycle Upon

subsequent unloading, the resistance decreases, then

in-creases as unloading continues, and reaches the maximum

resistance of the original peak at zero stress In the part of

this group where the stress amplitude decreases cycle bycycle, the extra peak diminishes and disappears, leavingthe original peak as the sole peak In the part of the secondgroup where the stress amplitude increases cycle by cycle,the original peak (peak at zero stress) is the sole peak,except that the extra peak (peak at the maximum stress)returns in a minor way (more minor than in the first group)

as the stress amplitude increases The extra peak grows asthe stress amplitude increases, but, in the part of the sec-ond group in which the stress amplitude decreases cycle

by cycle, it quickly diminishes and vanishes, as in the firstgroup Within each group, the amplitude of the variation

in resistance increases as the stress amplitude increasesand decreases as the stress amplitude subsequentlydecreases

The greater the stress amplitude, the larger and theless reversible the damage-induced increase in resistance(the extra peak) If the stress amplitude has been experi-enced before, the damage-induced increase in resistance(the extra peak) is small, as shown by comparing the re-sult of the second group with that of the first group (Fig 7),unless the extent of damage is large (Fig 8 for the high-est stress amplitude of>90% of the compressive strength).

When the damage is extensive (as shown by a decrease

in modulus), a damage-induced increase in resistanceoccurs in every cycle, even at a decreasing stress amplitude,and it can overshadow the strain-induced decrease inresistance (Fig 8) Hence, the damage-induced increase inresistance occurs mainly during loading (even within theelastic regime), particularly at a stress above that in priorcycles, unless the stress amplitude is high and/or damage

is extensive

At a high stress amplitude, the damage-induced crease in resistance, cycle by cycle, as the stress amplitudeincreases, causes the baseline resistance to increase irre-versibly (Fig 8) The baseline resistance in the regime ofmajor damage (that decreases in modulus) provides a mea-sure of the extent of damage (i.e., condition monitoring).This measure works in the loaded or unloaded state In

Time (sec)

− 0.001 0.000 0.001 0.002

0.003 80%

Figure 8 Fractional change in resistance and strain during

re-peated compressive loading at increasing and decreasing stress amplitudes, the highest of which was>90% of the compressive

strength, for a carbon-fiber concrete after 28 days of curing.

−20 −10 10

10002000

−1000

−2000

20Voltage (V)

Figure 9 Current-voltage characteristic of a carbon-fiber

silica-fume cement paste at 38 ◦C during stepped heating.

contrast, the measure using the damage-induced increase

in resistance (Fig 7) works only during an increase in

stress and indicates the occurrence of damage (whether

minor or major), as well as the extent of damage

Cement-Matrix Composites for Temperature Sensing

A thermistor is a thermometric device that consists of a

material (typically a semiconductor, but in this case a

ce-ment paste) whose electrical resistivity decreases with a

rise in temperature The carbon-fiber concrete described

previously for strain sensing is a thermistor because its

resistivity decreases reversibly as temperature increases

(59); the sensitivity is comparable to that of

semicon-ductor thermistors (The effect of temperature will need

to be compensated for in using the concrete as a strain

sensor.)

Figure 9 (59) shows the current–voltage characteristic

of carbon-fiber (0.5% by weight of cement) silica-fume (15%

by weight of cement) cement paste at 38◦C during stepped

heating The characteristic is linear below 5 V and

devi-ates positively from linearity beyond 5 V The resistivity is

obtained from the slope of the linear portion The voltage

at which the characteristic starts to deviate from linearity

is called the critical voltage

Figure 10 shows a plot of resistivity versus

tempera-ture for carbon-fiber silica-fume cement paste during

heat-ing and coolheat-ing The resistivity decreases upon heatheat-ing,

and the effect is quite reversible upon cooling That the

resistivity increases slightly after a heating/cooling cycle

is probably due to thermal degradation of the material

Figure 11 shows the Arrhenius plot of log conductivity

(conductivity= 1/resistivity) versus the reciprocal of the

absolute temperature The slope of the plot gives the

acti-vation energy, which is 0.390 ± 0.014 and 0.412 ± 0.017 eV

during heating and cooling, respectively

Results similar to those of carbon-fiber silica-fume

cement paste were obtained with carbon-fiber (0.5% by

weight of cement) latex (20% by weight of cement) cement

paste, silica-fume cement paste, latex cement paste, and

plain cement paste However, for all of these four types of

cement paste, (1) the resistivity is higher by about an

or-der of magnitude, and (2) the activation energy is lower

0

5 10 15 20 25 30 35 40 45 501

23

Cooling

Heating6

Figure 10 Plot of volume electrical resistivity vs temperature

during heating and cooling for a carbon-fiber silica-fume cement paste.

by about an order of magnitude, as shown in Table 1 Thecritical voltage is higher when fibers are absent (Table 1).The Seebeck (60–62,97) effect is a thermoelectric effectwhich is the basis for using thermocouples for temper-ature measurement This effect involves charge carriersthat move from a hot point to a cold point within a material,thereby resulting in a voltage difference between the twopoints The Seebeck coefficient is the negative of the voltagedifference (hot minus cold) per unit temperature differencebetween the two points (hot minus cold) Negative carriers(electrons) make it more negative, and positive carriers(holes) make it more positive

The Seebeck effect in carbon-fiber-reinforced cementpaste involves electrons from the cement matrix (62) andholes from the fibers (60,61), such that the two contribu-tions are equal at the percolation threshold, a fiber contentfrom 0.5–1.0% by weight of cement (62) The hole contri-bution increases monotonically as fiber content changesabove and below the percolation threshold (62)

Due to the free electrons in a metal, cement that tains metal fibers such as steel fibers is even more negative

con-in thermoelectric power than cement without fiber (97)

Figure 11 Arrhenius plot of log electrical conductivity vs

recip-rocal absolute temperature for a carbon-fiber silica-fume cement paste.

Table 1 Resistivity, Critical Voltage, and Activation Energy of Five Types of Cement Pastes

Activation Energy (eV) Formulation

Resistivity at 20 ◦C( cm)

Critical Voltage

at 20 ◦C (V) Heating CoolingPlain (4.87 ± 0.37) × 105 10.80 ± 0.45 0.040 ± 0.006 0.122 ± 0.006

Silica fume (6.12 ± 0.15) × 105 11.60 ± 0.37 0.035 ± 0.003 0.084 ± 0.004

Carbon fibers + (1.73 ± 0.08) × 104 8.15 ± 0.34 0.390 ± 0.014 0.412 ± 0.017

silica fume Latex (6.99 ± 0.12) × 105 11.80 ± 0.31 0.017 ± 0.001 0.025 ± 0.002

Carbon fibers + (9.64 ± 0.08) × 104 8.76 ± 0.35 0.018 ± 0.001 0.027 ± 0.002

latex

The attainment of a very negative thermoelectric power is

attractive because a material whose thermoelectric power

is positive and a material whose thermoelectric power is

negative are two very dissimilar materials; their junction

is a thermocouple junction The greater the dissimilarity,

the more sensitive the thermocouple

Table 2 and Fig 12 show the thermoelectric power

re-sults The absolute thermoelectric power is much more

neg-ative for all of the steel-fiber cement pastes compared to all

of the carbon-fiber cement pastes An increase in the steel

fiber content from 0.5% to 1.0% by weight of cement makes

the absolute thermoelectric power more negative, whether

or not silica fume (or latex) is present An increase in the

steel fiber content also increases the reversibility and

lin-earity of the change in Seebeck voltage with the

tempera-ture difference between the hot and cold ends, as shown by

comparing the values of the Seebeck coefficient obtained

during heating and cooling in Table 2 The values obtained

during heating and cooling are close for the pastes that

have the higher steel-fiber content but are not so close for

the pastes that have the lower steel-fiber content In

con-trast, for pastes that have carbon fibers in place of steel

fibers, the change in Seebeck voltage with temperature

dif-ference is highly reversible for both carbon-fiber contents

of 0.5 and 1.0% by weight of cement, as shown in Table 2

by comparing the values of the Seebeck coefficient obtained

during heating and cooling

Table 2 Volume Electical Resistivity, Seebeck Coefficient (µV/◦C) with Copper as the Reference, and the Absolute

Thermoelectric Power (µV/◦C) of Various Cement Pastes with Steel Fibers (S

f ) or Carbon Fibers (C f )

Cement Paste Fibers(%) ( cm) Coefficient Thermoelectric Power Coefficient Thermoelectric Power

Table 2 shows that the volume electrical resistivity

is much higher for the steel-fiber cement pastes thanthe corresponding carbon-fiber cement pastes This is at-tributed to the much lower volume fraction of fibers in theformer (Table 2) An increase in the steel-or carbon-fibercontent from 0.5 to 1.0% by weight of cement decreases theresistivity, though the decrease is more significant for thecarbon-fiber cement than for the steel-fiber cement Thatthe resistivity decrease is not large when the steel fibercontent is increased from 0.5 to 1.0% by weight of cementand that the resistivity is still high at a steel-fiber content

of 1.0% by weight of cement suggest that a steel-fiber tent of 1.0% by weight of cement is below the percolationthreshold

con-With or without silica fume (or latex), the change of theSeebeck voltage with temperature is more reversible andlinear at a steel-fiber content of 1.0% by weight of cementthan at a steel fiber content of 0.5% by weight of cement.This is attributed to the larger role of the cement matrix

at the lower steel-fiber content and the contribution of thecement matrix to the irreversibility and nonlinearity Ir-reversibility and nonlinearity are particularly significantwhen the cement paste contains no fiber

From the practical point of view, the steel-fiber fume cement paste that contains 1.0% steel fibers byweight of cement is particularly attractive for use intemperature sensing because the absolute thermoelectric

0

Figure 12 Variation of the Seebeck voltage (with copper as the

reference) vs the temperature difference during heating and

cool-ing for a steel-fiber silica-fume cement paste containcool-ing 1.0% steel

fibers by weight of cement.

power is the highest (−68 µV/◦C) and the variation of the

Seebeck voltage with the temperature difference between

the hot and cold ends is reversible and linear The

abso-lute thermoelectric power is as high as that of commercial

thermocouple materials

Joints between concretes that have different values of

thermoelectric power, made by multiple pouring, provide

concrete thermocouples (98)

Cement-Matrix Composites for Thermal Control

Concretes that can inherently provide heating through

Joule heating, provide temperature sensing, or provide

temperature stability through a high specific heat (high

thermal mass) are highly desirable for thermal control of

structures and energy saving in buildings Concretes of low

electrical resistivity (69–78) are useful for Joule heating,

concrete thermistors and thermocouples are useful for

tem-perature sensing, and concretes of high specific heat (66–

68,99) are useful for heat retention These concretes

in-volve the use of admixtures such as fibers and silica fume

For example, silica fume introduces interfaces that

pro-mote the specific heat (66); short carbon fibers enhance

electrical conductivity (74) and produce p-type concrete

(62) Plain concrete is n-type (62)

Figure 13 (74) gives the volume electrical resistivity of

composites after 7 days of curing The resistivity decreases

a lot as the fiber volume fraction increases, whether or not

a second filler (silica fume or sand) is present When sand

is absent, the addition of silica fume decreases the

resistiv-ity at all carbon-fiber volume fractions except the highest

volume fraction of 4.24%; the decrease is most significant

at the lowest fiber volume fraction of 0.53% When sand is

present, the addition of silica fume similarly decreases the

resistivity; the decrease is most significant at fiber volume

fractions below 1% When silica fume is absent, the

addi-tion of sand decreases the resistivity only when the fiber

volume fraction is below about 0.5%; at high fiber volume

fractions, the addition of sand even increases the resistivity

due to the porosity induced by the sand Thus, the addition

of a second filler (silica fume or sand) that is essentially

(a)

(b)

(c)(d)

Figure 13 Variation of the volume electrical resistivity of

cement-matrix composites with carbon-fiber volume fraction (a) without sand, with methylcellulose, without silica fume; (b) without sand, with methylcellulose, with silica fume; (c) with sand, with methylcellulose, without silica fume; (d) with sand, with methylcellulose, with silica fume.

nonconducting decreases the resistivity of the compositeonly at low volume fractions of the carbon fibers, and themaximum fiber volume fraction for decreased resistivity islarger when the particle size of the filler is smaller The de-crease in resistivity is attributed to the improved fiber dis-persion that results from the presence of the second filler.Consistent with improved fiber dispersion is increased flex-ural toughness and strength due to the presence of the sec-ond filler

Table 3 (67,100) shows the specific heats of cementpastes The specific heat is significantly increased byadding of silica fume It is further increased by the fur-ther addition of methylcellulose and defoamer It is stillfurther increased by the still further addition of carbonfibers The effectiveness of fibers in increasing the specificheat increases in the following order: as-received fibers,

O3-treated fibers, dichromate-treated fibers, and treated fibers This trend applies whether the silica fume isas-received or silane-treated For any of the formulations,silane-treated silica fume gives higher specific heat thanas-received silica fume The highest specific heat is exhib-ited by the cement paste that contains silane-treated sil-ica fume and silane-treated fibers The specific heat is 12%higher than that of plain cement paste, 5% higher than that

silane-of the cement paste containing as-received silica fume andas-received fibers, and 0.5% higher than that of the cement

Table 3 Specific Heats (J/g K,± 0.001) of Cement Pastes.

The Value for Plain Cement Paste (with Cement and

Water only) Is 0.736 J/g K

As-Received Silane-Treated Formulationa Silica Fume Silica Fume

paste containing as-received silica fume and silane-treated

fibers Hence, silane treatment of fibers is more valuable

than treatment of silica fume to increase the specific heat

Table 4 (67,100) shows the thermal diffusivities of

ce-ment pastes The thermal diffusivity is significantly

de-creased by adding silica fume The further addition of

methylcellulose and defoamer or the further addition of

fibers has relatively little effect on thermal diffusivity

Sur-face treatment of the fibers by ozone or dichromate slightly

increases the thermal diffusivity, whereas surface

treat-ment of the fibers by silane slightly decreases the thermal

diffusivity These trends apply whether the silica fume is

as-received or silane-treated For any of the formulations,

silane-treated silica fume gives slightly lower (or

essen-tially the same) thermal diffusivity than as-received silica

fume Silane treatments of silica fume and of fibers are

about equally effective in lowering the thermal diffusivity

Table 5 (67,100) shows the densities of cement pastes

The density is significantly decreased by adding silica

fume It is further decreased slightly by the addition

of methylcellulose and defoamer It is still further

de-creased by the addition of fibers The effectiveness of

Table 4 Thermal Diffusivities (mm 2 /s,± 0.03) of Cement

Pastes The Value for Plain Cement Paste (with Cement

and Water Only) Is 0.36 mm 2 /s

As-Received Silane-Treated Formulationa Silica Fume Silica Fume

Table 5 Density (g/cm 3± 0.02) of Cement Pastes The

Value for Plain Cement Paste (with Cement and Water Only) Is 2.01 g/cm 3

As-Received Silane-Treated Formulationa Silica-Fume Silica-Fume

or dichromate slightly increases the thermal ity, whereas surface treatment of the fibers by silanehas a negligible effect These trends apply whether thesilica fume is as-received or silane-treated For any ofthe formulations, silane-treated silica fume gives slightlylower (or essentially the same) thermal conductivity thanas-received silica fume Silane treatments of silica fumeand of fibers contribute comparably to reducing the ther-mal conductivity

conductiv-Table 6 Thermal Conductivities (W/mK,± 0.03) of Cement

Pastes The Value for Plain Cement Paste (with Cement and Water Only) Is 0.53 W/m K

As-Received Silane-Treated Formulationa Silica-Fume Silica-Fume

Table 7 Thermal Behavior of Cement Pastes and Mortars

Without Silica With Silica Without Silica With Silica

bProduct of density, specific heat, and thermal diffusivity.

Sand is a much more common component in concrete

than silica fume It differs from silica fume in its relatively

large particle size and negligible reactivity with cement

Sand gives effects that are opposite from those of silica

fume; sand addition decreases the specific heat and

in-creases the thermal conductivity (99)

Table 7 (99) shows the thermal behavior of cement

pastes and mortars Comparison of the results on cement

paste without silica fume and those on mortar without

sil-ica fume shows that sand addition decreases the specific

heat by 13% and increases the thermal conductivity by

9% Comparison of the results on cement paste with silica

fume and those on mortar with silica fume shows that sand

addition decreases the specific heat by 11% and increases

the thermal conductivity by 64% That sand addition has

more effect on the thermal conductivity when silica fume

is present than when silica fume is absent is due to the

low value of the thermal conductivity of cement paste that

contains silica fume (Table 7)

Comparison of the results on cement paste without and

with silica fume shows that silica fume addition increases

the specific heat by 7% and decreases the thermal

conduc-tivity by 38% Comparison of the results on mortar without

and with silica fume shows that silica fume addition

in-creases the specific heat by 10% and dein-creases the thermal

conductivity by 6% Hence, the effects of silica fume

addi-tion on mortar and cement paste are in the same direcaddi-tion

That the effect of silica fume on the thermal conductivity

is much less for mortar than for cement paste is mainly

due to the fact that silica fume addition increases the

den-sity of mortar but decreases the denden-sity of cement paste

(Table 7) That the fractional increase in specific heat due

to silica fume addition is higher for mortar than cement

paste is attributed to the low value of the specific heat of

mortar without silica fume (Table 7)

Comparison of the results on cement paste with

sil-ica fume and those on mortar without silsil-ica fume shows

that sand addition gives a lower specific heat than silica

fume addition and a higher thermal conductivity than

sil-ica fume addition Because sand has a much larger particle

size than silica fume, sand has much less interfacial area

than silica fume, though the interface may be more diffuse

for silica fume than for sand The low interfacial area for

sand is believed responsible for the low specific heat and

the higher thermal conductivity because slippage at theinterface contributes to the specific heat and the interfaceacts as a thermal barrier

Silica fume addition increases the specific heat of ment paste by 7%, whereas sand addition decreases it by13% Silica fume addition decreases the thermal conduc-tivity of cement paste by 38%, whereas sand addition in-creases it by 22% Hence, silica fume addition and sandaddition have opposite effects The cause is believed to beassociated mainly with the low interfacial area for sandand the high interfacial area for silica fume, as explained inthe last paragraph The high reactivity of silica fume com-pared to sand may contribute to the observed difference be-tween silica fume addition and sand addition, though thiscontribution is believed to be minor because the reactivityshould have tightened up the interface, thus decreasingthe specific heat (in contrast to the observed effects) Thedecrease in specific heat and the increase in thermal con-ductivity upon sand addition are believed to be due to thehigher level of homogeneity within a sand particle thanwithin cement paste

ce-Cement-Matrix Composites for Vibration Reduction

Vibration reduction requires high damping capacity andhigh stiffness Viscoelastic materials such as rubber havehigh damping capacity but low stiffness Concretes thathave both high damping capacity (two or more ordershigher than conventional concrete) (Table 8) (101) and highstiffness (Table 9) (101) can be obtained by adding surface-treated silica fume to concrete Steel-reinforced concretesthat have improved damping capacity and stiffness can beobtained by surface treating the steel (say, by sand blast-ing) before incorporating it in concrete (Table 10) (102) or

by using silica fume in concrete (101) Due to its small

Table 8 Loss Tangent (Tan δ, ±0.01)

Plaina 0.016 <10−4 <10−4 <10−4 Sand <10−4 <10−4 <10−4 <10−4 Sand + silica fume 0.021 0.14 0.01 <10−4

aNo sand, no silica fume.

Table 9 Storage Modulus (GPa,±0.2)

Sand + silica fume 13.11 14.34 13.17 13.11

aNo sand, no silica fume.

particle size, silica fume in concrete introduces interfaces

that enhance damping Sand blasting of a steel rebar

in-creases the interfacial area between steel and concrete,

thereby enhancing damping Carbon-fiber addition has

rel-atively small effects on damping capacity and stiffness

(103)

POLYMER-MATRIX COMPOSITES

FOR SMART STRUCTURES

Polymer-matrix composites for structural applications

typ-ically contain continuous fibers such as carbon, polymer,

and glass fibers because continuous fibers tend to be more

effective as a reinforcement than short fibers

Polymer-matrix composites that contain continuous carbon fibers

are used for aerospace, automobile, and civil structures In

contrast, continuous fibers are too expensive for reinforcing

concrete Because carbon fibers are electrically

conduct-ing, whereas polymer and glass fibers are not, carbon-fiber

composites are predominant among polymer-matrix

com-posites that are intrinsically smart

Background on Polymer-Matrix Composites

Polymer-matrix composites that contain continuous carbon

fibers are important structural materials due to their high

tensile strength, high tensile modulus, and low density

They are used for lightweight structures such as satellites,

aircraft, automobiles, bicycles, ships, submarines, sporting

goods, wheel chairs, armor, and rotating machinery (such

as turbine blades and helicoptor rotors) Due to the

re-cent emphasis on repair of civil infrastructural systems,

Table 10 Loss Tangent, Storage Modulus, and Loss Modulus of Mortars with and Without Steel Reinforcement

aA: No rebar B: As-received steel rebar C: Ozone treated steel rebar D: Sand-blasted steel rebar.

composites are beginning to be used to repair concretestructures and bridges, even though they are much moreexpensive than concrete Because the price of carbon fibershas been dropping steadily during the last two decades, thespectrum of applications has been widening tremendously.The continuous carbon fibers used are primarily based

on either polyacrylonitrile (PAN) or mesophase pitch.Mesophase-pitch-based carbon fibers, if heat treated tohigh temperatures exceeding 2500◦C, can be graphitizedand attain very high values of tensile modulus and thermalconductivity (in-plane), in addition to improved oxidationresistance The high thermal conductivity is attractivefor thermal management, which is particularly importantfor electronics (i.e., heat sinks, etc.) However, graphitizedfibers tend to have relatively low strength due to the ease

of shear between the graphite layers, and they are veryexpensive On the other hand, PAN-based fibers cannot begraphitized, though they compete well with mesophase-pitch-based fibers which have not been graphitized, bothmaterials exhibit reasonably high values of both strengthand modulus and are not very expensive These fibers arethe most widely used among carbon fibers The fabrication

of both pitch-based and PAN-based carbon fibers involvesstabilization (infusibilization) and then carbonization (con-version from hydrocarbon molecules to a carbon network).Graphitization optionally follows carbonization

Due to the importance of carbon-fiber polymer-matrixcomposites for structural applications, the mechanical be-havior of these materials has been much investigated.Much less work has been done to study the electricalbehavior (104–111) On the other hand, due to the fact thatcarbon fibers are much more conductive than the polymermatrix, the electrical behavior gives much information onthe microstructure, such as the degree of fiber alignment,the number of fiber–fiber contacts, the amount of delami-nation, and the extent of fiber breakage Such information

is useful for scientific understanding of the properties ofthe composite and is also valuable for giving the compositethe ability to sense its strain, damage, and temperature

in real time via electrical measurement In other words,the strain, damage, and temperature affect the electrical

behavior, such as electrical resistance, which thus serves

to indicate strain, damage, and temperature In this way,

the composite is self-sensing, that is, intrinsically smart,

without the need for attached or embedded sensors (such as

optical fibers, acoustic sensors and piezoelectric sensors),

which raise the cost, reduce the durability and, in the case

of embedded sensors, weaken the structure

Carbon fibers are electrically conductive, whereas

the polymer matrix is electrically insulating [except for

the rare situation in which the polymer is electrically

conductive (112)] The continuous fibers in a composite

laminate are in the form of layers called laminae Each

lam-ina comprises many bundles (called tows) of fibers in a

poly-mer matrix Each tow consists of thousands of fibers There

may or may not be twist in a tow Each fiber has a diameter

that typically ranges from 7–12µm The tows within a

lam-ina are usually oriented in the same direction, but tows in

different laminae may or may not be in the same direction

A laminate whose tows in all of the laminae are oriented

in the same direction is said to unidirectional A laminate

whose tows in adjacent laminae are oriented at a 90◦

an-gle is said to be crossply In general, an anan-gle of 45◦ and

other angles may also be involved in the various laminae,

as desired to attain the mechanical properties required

for the laminate in various directions in the plane of the

laminate

Within a lamina whose tows are in the same direction,

the electrical conductivity is highest in the fiber direction

In the transverse direction in the plane of the lamina, the

conductivity is not zero, even though the polymer matrix is

insulating because there are contacts between fibers of

ad-jacent tows (113) In other words, a fraction of the fibers of

one tow touch a fraction of the fiber of an adjacent tow here

and there along the length of the fibers These contacts

re-sult from the fact that fibers are not perfectly straight or

parallel (even though the lamina is said to be

unidirec-tional) and the flow of the polymer matrix (or resin) during

composite fabrication can prevent a fiber from being

pletely covered by the polymer (even though, prior to

com-posite fabrication, each fiber may be completely covered

by the polymer, as in the case of a prepreg, a fiber sheet

impregnated with polymer) Fiber waviness is known as

marcelling Thus, the transverse conductivity gives

infor-mation on the number of fiber–fiber contacts in the plane

of the lamina

For similar reasons, the contacts between fibers of

ad-jacent laminae make the conductivity in the

through-thickness direction (direction perpendicular to the plane

of the laminate) nonzero Thus, the through-thickness

con-ductivity gives information on the number of fiber–fiber

contacts between adjacent laminae

Matrix cracking between the tows of a lamina decreases

the number of fiber–fiber contacts in the plane of the

lam-ina, thus decreasing the transverse conductivity Similarly,

matrix cracking between adjacent laminae [as in

delami-nation (114)] decreases the number of fiber–fiber contacts

between adjacent laminae, thus decreasing the

through-thickness conductivity This means that the transverse and

through-thickness conductivities can indicate damage as

matrix cracking

Fiber damage (as distinct from fiber fracture)

de-creases the conductivity of a fiber, thereby decreasing the

longitudinal conductivity (conductivity in the fiber tion) However, due to the brittleness of carbon fibers, thedecrease in conductivity due to fiber damage before fiberfracture is rather small (115)

direc-Fiber fracture causes a much larger decrease in the gitudinal conductivity of a lamina than fiber damage Ifthere is only one fiber, a broken fiber results in an open cir-cuit, i.e., zero conductivity However, a lamina has a largenumber of fibers and adjacent fibers can make contact hereand there Therefore, the portions of a broken fiber still con-tribute to the longitudinal conductivity of the lamina As aresult, the decrease in conductivity due to fiber fracture isless than it would be if a broken fiber did not contribute

lon-to conductivity Nevertheless, the effect of fiber fracture

on longitudinal conductivity is significant, so that tudinal conductivity can indicate damage as fiber fracture(116)

longi-The through-thickness volume resistance of a laminate

is the sum of the volume resistance of each of the laminae

in the through-thickness direction and the contact tance of each of the interfaces between adjacent laminae(i.e., the interlaminar interface) For example, a laminatethat has eight laminae has eight volume resistances andseven contact resistances, all in the through-thickness di-rection Thus, to study the interlaminar interface, it is bet-ter to measure the contact resistance between two laminaerather than the through-thickness volume resistance of theentire laminate

resis-The contact resistance between laminae can be sured by allowing two laminae (strips) to contact at a junc-tion and using the two ends of each strip to make fourelectrical contacts (117) An end of the top strip and anend of the bottom strip serve as contacts for passing cur-rent The other end of the top strip and the other end ofthe bottom strip serve as contacts for voltage measure-ment The fibers in the two strips can be in the same di-rection or in different directions This method is a form ofthe four-probe method of measuring electrical resistance.The configuration is illustrated in Fig 14 for crossply andunidirectional laminates To make sure that the volumeresistance within a lamina in the through-thickness direc-tion does not contribute to the measured resistance, thefibers at each end of a lamina strip should be electricallyshorted together by using silver paint or other conductingmedia The measured resistance is the contact resistance

mea-of the junction This resistance, multiplied by the area mea-ofthe junction, gives the contact resistivity, which is indepen-dent of the area of the junction and depends only on thenature of the interlaminar interface The unit of contactresistivity is m2, whereas that of the volume resistivity

is m.

The structure of the interlaminar interface tends to bemore prone to change than the structure within a lamina.For example, damage as delamination is much more com-mon than damage as fiber fracture Moreover, the struc-ture of the interlaminar interface is affected by the inter-laminar stress (whether thermal stress or curing stress),which is particularly significant when the laminae are notunidirectional (because the anisotropy within each lam-ina enhances the interlaminar stress) The structure ofthe interlaminar interface also depends on the extent ofconsolidation of the laminae during composite fabrication

AD

B+

−C

CurrentInsulating paper

Bottom lamina

(b)

Figure 14 Specimen configuration for measuring of the

con-tact electrical resistivity between laminae (a) Crossply laminae.

(b) Unidirectional laminae.

The contact resistance provides a sensitive probe of the

structure of the interlaminar interface

The volume resistivity in the through-thickness

direc-tion can be measured by using the four-probe method, in

which each of the two current contacts is a conductor loop

(made by silver paint, for example) on each of the two outer

surfaces of the laminate in the plane of the laminate, and

each of the two voltage contacts is a conductor dot within

the loop (114) An alternate method is to have four extra

long laminae in the laminate, extend out to serve as

electri-cal leads (118) The two outer leads are for current contacts,

and the two inner leads are for voltage contacts The use of

a thin metal wire inserted at an end into the interlaminar

space during composite fabrication to serve as an electrical

contact is not recommended because the quality of the

elec-trical contact between the metal wire and carbon fibers is

hard to control and the wire is intrusive in the composite

The alternate method is less convenient than the method

involving loops and dots, but it approaches the ideal

four-probe method more closely

To attain zero conductivity in the through-thickness

di-rection of a laminate, it is necessary to use an insulating

layer between two adjacent laminae (119) The insulating

layer can be a piece of writing paper Tissue paper is

in-effective in preventing contacts between fibers of adjacent

laminae due to its porosity The attainment of zero

conduc-tivity in the through-thickness direction allows the

lami-nate to serve as a capacitor This means that the structural

composite stores energy by serving as a capacitor

Polymer-Matrix Composites for Strain Sensing

Smart structures that can monitor their own strain are

valuable for structural vibration control Self-monitoring

of strain (reversible) has been achieved in carbon-fiber

epoxy-matrix composites without using embedded or tached sensors (118,120–123) because the electrical resis-tance of the composite in the through-thickness or longitu-dinal direction changes reversibly with longitudinal strain(gauge factor up to 40) due to a change in the degree of fiberalignment Tension in the fiber direction of the compositeincreases the degree of fiber alignment, thereby decreasingthe chance for fibers of adjacent laminae to touch one an-other As a consequence, the through-thickness resistanceincreases, and the longitudinal resistance decreases.Figure 14 (121) shows the change in longitudinal re-sistance during cyclic longitudinal tension in the elasticregime for a unidirectional continuous-carbon-fiber epoxy-matrix composite that has eight fiber layers (laminae) Thestress amplitude is equal to 14% of the breaking stress Thestrain returns to zero at the end of each cycle Because ofthe small strains involved, the fractional resistance change

at- R/R0is essentially equal to the fractional change in tivity The longitudinal R/R0decreases upon loading and

resis-increases upon unloading in every cycle, such that R

irre-versibly decreases slightly after the first cycle (i.e., R/R0does not return to zero at the end of the first cycle) Athigher stress amplitudes, the effect is similar, except thatboth of the reversible and irreversible parts of R/R0arelarger

Figure 15 (121) shows the change in the thickness resistance during cyclic longitudinal tension inthe elastic regime for the same composite The stress am-plitude is equal to 14% of the breaking stress The through-thickness R/R0 increases upon loading and decreases

through-020406080100

140160180

00.020.040.060.080.10.120.140.16

20 40 60 80 100

Time (s)

120 140 160 180 200

Figure 15 Longitudinal stress and strain and fractional

resis-tance increase ( R/R0 ) obtained simultaneously during cyclic sion at a stress amplitude equal to 14% of the breaking stress for

ten-a continuous-fiber epoxy-mten-atrix composite.

140160180

00.020.040.060.080.10.120.140.16

−0.020

−0.0100.010.020.030.04

Figure 16 Longitudinal stress and strain and the

through-thickness R/R0 obtained simultaneously during cyclic tension

at a stress amplitude equal to 14% of the breaking stress for a

continuous-fiber epoxy-matrix composite.

upon unloading in every cycle, such that R irreversibly

de-creases slightly after the first cycle (i.e., R/R0 does not

return to zero at the end of the first cycle) Upon

increas-ing the stress amplitude, the effect is similar, except that

the reversible part of R/R0is larger

The strain sensitivity (gauge factor) is defined as the

reversible part of ...

14 016 018 0

00.020. 040 .060.080 .10 .12 0 . 14 0 .16

20 40 60 80 10 0

Time (s)

12 0 14 0 16 0 18 0 200

Figure 15 Longitudinal stress and strain and fractional ... thermoelectricpowers) of T-300 and P-25, whether the junction isunidirectional or crossply Pristine P -1 0 0 and pristineP -1 2 0 are both slightly n-type Intercalation with sodiummakes P -1 0 0 and P -1 2 0 strongly n-type... makes P -1 0 0 and P -1 2 0 strongly p-type

pris-A junction comprising bromine-intercalated P -1 0 0 andsodium-intercalated P -1 0 0 has a positive thermocouplesensitivity that is close to the sum of