Encyclopedia of Smart Materials (Vols 1 and 2) - M. Schwartz (2002) Episode 10 pps

Smart Surfaces—Controlled Porosity, “Chemical Valve” Environmentally controlled change in macromolecular sizefrom a compact hydrophobic globule to an expanded hy-drophilic coil is exploi

to hydrophilic below it has been successfully used for

detaching mammalian cells Mammalian cells are

nor-mally cultivated on a hydrophobic solid substrate and are

detached from the substrate by protease treatment, which

often damages the cells by hydrolyzing various

membrane-associated protein molecules The poly(NIPAAM)-grafted

surface is hydrophobic at 37◦C because this temperature

is above the critical temperature for the grafted polymer

and that cells that are growing well on it A decrease in

temperature results in transition of the surface to the

hy-drophilic state, where the cells can be easily detached from

the solid substrate without any damage Poly(NIPAAM)

was grafted to polystyrene culture dishes using an electron

beam Bovine hepatocytes, cells that are highly sensitive

to enzymatic treatment, were cultivated for 2 days at 37◦C

and detached by incubation at 4◦C for 1 h Nearly 100%

of the hepatocytes was detached and recovered from the

poly(NIPAAM)-grafted dishes by low-temperature

treat-ment, whereas only about 8% of the cells was detached from

the control dish (57) The technique has been extended

to different cell types (58,59) It is noteworthy that

hep-atocytes recovered by cooling retained their native form

had numerous bulges and dips, and attach well to the

hy-drophobic surface again, for example, when the

tempera-ture was increased above the conformational transition of

poly(NIPAAM) On the contrary, enzyme-treated cells had

a smooth outer surface and had lost their ability to attach

to the surface Thus, cells recovered by a temperature shift

from poly(NIPAAM)-grafted surfaces have an intact

struc-ture and maintain normal cell functions (58)

The molecular machinery involved in cell-surface

de-tachment was investigated using temperature-responsive

surfaces (60) Poly(NIPAAM)-grafted and nongrafted

sur-faces showed no difference in attachment, spreading,

growth, confluent cell density, or morphology of bovine

aortic endothelial cells at 37◦C Stress fibers, peripheral

bands, and focal contacts were established in similar ways

When the temperature was decreased to 20◦C, the cells

grown on poly(NIPAAM)-grafted support lost their

flat-tened morphology and acquired a rounded appearance

sim-ilar to that of cells immediately after plating Mild

agi-tation makes the cells float free from the surface without a

trypsin treatment Neither changes in cell morphology nor

cell detachment occurred on ungrafted surfaces Sodium

azide, an ATP synthesis inhibitor, and genistein, a tyrosine

kinase inhibitor, suppressed changes in cell morphology

and cell detachment, whereas cycloheximide, a protein

syn-thesis inhibitor, slightly enhanced cell detachment

Phal-loidin, an actin filament stabilizer, and its depolymerizer,

cytochalasin D, also inhibited cell detachment These

find-ings suggest that cell detachment from grafted surfaces

is mediated by intracellular signal transduction and

re-organization of the cytoskeleton, rather than by a simple

changes in the “stickiness” of the cells to the surface when

the hydrophobicity of the surface is changed

One could imagine producing artificial organs using

temperature-induced detachment of cells Artificial skin

could be produced as the cells are detached from the

support not as a suspension (the usual result of

protease-induced detachment) but preserving their intercellular

contacts Fibroblasts were cultured on the

poly(NIPAAM)-collagen support until the cells completely covered the

surface at 37◦C, followed by a decrease in temperature toabout 15◦C The sheets of fibroblasts detached from thedish and within about 15 min floated in the culture medium(57) The detached cells could be transplanted to anotherculture surface without functional and structural changes(34) Grafting of poly(NIPPAM) onto a polystyrene sur-face by photolitographic technique creates a special pat-tern on the surface, and by decreasing temperature, cul-tured mouse fibroblast STO cells are detached only fromthe surface area on which poly(NIPAAM) was grafted (61).Lithographed films of smart polymer present supports forcontrolled interactions of cells with surfaces and can di-rect the attachment and spreading of cells (62) One couldenvisage producing artificial cell assemblies of complex ar-chitecture using this technique

Smart Surfaces—Temperature Controlled Chromatography

Surfaces that have thermoresponsive hydrophobic philic properties have been used in chromatography HPLCcolumns with grafted poly(NIPAAM) have been used forseparating steroids (63) and drugs (64) The chromato-graphic retention and resolution of the solutes was stronglydependent on temperature and increased as temperatureincreased from 5 to 50◦C, whereas the reference columnpacked with nonmodified silica displayed much shorter re-tention times that decreased as temperature decreased.Hydrophobic interactions dominate in retaining solutes

/hydro-at higher temper/hydro-ature, and the preferential retention ofhydrogen-bond acceptors was observed at low tempera-tures The effect of temperature increase on the reten-tion behavior of solutes separated on the poly(NIPAAM)-grafted silica chromatographic matrix was similar to theaddition of methanol to the mobile phase at constant tem-perature (65)

The temperature response of the poly(NIPAAM)-silicamatrices depends drastically on the architecture of thegrafted polymer molecules Surface wettability changesdramatically as temperature changes across the range32–35◦C (corresponding to the phase-transition tempera-ture for NIPAAM in aqueous media) for surfaces wherepoly(NIPAAM) is terminally grafted either directly to thesurface or to the looped chain copolymer of NIPAAM and

N-acryloylhydroxysuccinimide which was initially coupled

to the surface The wettability changes for the loop-graftedsurface itself were relatively large but had a slightly lowertransition temperature (∼27◦C) The restricted conforma-tional transitions for multipoint grafted macromoleculesare probably the reason for the reduced transition tem-perature The largest surface free energy changes amongthree surfaces was observed for the combination of bothloops and terminally grafted chains (30)

Introduction of a hydrophobic comonomer, methacrylate, in the polymer resulted in a decreasedtransition temperature of about 20◦C Retention of

buthyl-steroids in poly(NIPAAM-co-buthylmethacrylate)-grafted

columns increases as column temperature increases Thecapacity factors for steroids on the copolymer-modifiedsilica beads was much larger than that on poly(NIPAAM)-grafted columns The effect of temperature on steroidretention on poly(NIPAAM-co-buthylmethacrylate)-

grafted stationary phases was more pronounced compared

to supports modified with poly(NIPAAM) Furthermore,

retention times for steroids increased remarkably as the

buthylmethacrylate content increased in the copolymer

The temperature-responsive elution of steroids was

strongly affected by the hydrophobicity of the grafted

polymer chains on silica surfaces (63)

The mixture of polypeptides, consisting of 21–30 amino

acid residues (insulin chain A,β-endorphin fragment 1–27

and insulin chain B) could not be separated at 5◦C(below

the transition temperature) on copolymer-grafted matrix

At this temperature, the copolymer is in an extended

hydrophilic conformation that results in decreased

inter-actions with peptides and hence short retention times

in-sufficient to resolve them The mixture has been easily

separated at 30◦C, when the copolymer is collapsed,

hy-drophobic interactions are more pronounced, and

reten-tion times sufficiently long for resolving polypeptides (66)

Large protein molecules such as immunoglobulin G

demon-strate less pronounced changes in adsorption above and

below the transition temperature Only about 20% of the

protein adsorbed on poly(NIPAAM)-grafted silica at 37◦C

(above the LCST) were eluted after decreasing

tempera-ture to 24◦C (below the transition temperature) (67)

Quan-titative elution of proteins adsorbed on the matrix via

hydrophobic interactions has not yet been demonstrated,

although protein adsorption on poly(NIPAAM)-grafted

ma-trices could be somewhat controlled by a temperature

shift A successful strategy for temperature-controlled

protein chromatography proved to be a combination of

temperature-responsive polymeric grafts and

biorecogni-tion element, for example, affinity ligands

The access of the protein molecules to the ligands

on the surface of the matrix is affected by the

transi-tion of the polymer macromolecule grafted or attached to

the chromatographic matrix Triazine dyes, for example,

Cibacron Blue, are often used as ligands for dye-affinity

chromatography of various nucleotide-dependent enzymes

(68) Poly(N-vinyl caprolactam), a thermoresponsive

poly-mer whose critical temperature is about 35◦C interacts

effi-ciently with triazine dyes Polymer molecules of 40000 MW

are capable of binding up to seven to eight dye molecules

hence, the polymer binds via multipoint interaction to the

dye ligands available on the chromatographic matrix At

elevated temperature, polymer molecules are in a

com-pact globule conformation that can bind only to a few

lig-ands on the matrix Lactate dehydrogenase, an enzyme

from porcine muscle has good access to the ligands that

are not occupied by the polymer and binds to the column

Poly(N-vinyl caprolactam) macromolecules undergo

tran-sition to a more expanded coil conformation as temperature

decreases Now, the polymer molecules interact with more

ligands and begin to compete with the bound enzyme for

the ligands Finally, the bound enzyme is displaced by the

expanded polymer chains The temperature-induced

elu-tion was quantitative, and the first reported in the

litera-ture when temperalitera-ture change was used as the only

elut-ing factor without any changes in buffer composition (69)

Small changes in temperature, as the only eluting factor,

are quite promising because there is no need in this case to

separate the target protein from an eluent, usually a

com-peting nucleotide or high salt concentration in dye-affinity

chromatography

Smart Surfaces—Controlled Porosity, “Chemical Valve”

Environmentally controlled change in macromolecular sizefrom a compact hydrophobic globule to an expanded hy-drophilic coil is exploited when smart polymers are used

in systems of environmentally controlled porosity, so called

“chemical valves.” When a smart polymer is grafted to thesurface of the pores in a porous membrane or chromato-graphic matrix, the transition in the macromolecule affectsthe total free volume of the pores available for the solventand hence presents a means to regulate the porosity of thesystem

Membranes of pH-sensitive permeability were ted by grafting smart polymers such as poly(methacrylicacid) (70), poly(benzyl glutamate), poly(2-ethylacrylicacid) (71), poly(4-vinylpyridine) (72), which changetheir conformation in response to pH Thermosensitive

construc-chemical valves have been developed by grafting

poly(N-acryloylpyrrolidine), poly(N-n-propylacrylamide), orpoly(acryloylpiperidine) (73), poly(NIPAAM) alone (33,74)

or in copolymers with poly(methacrylic acid) (74) insidethe pores For example, grafted molecules of poly(benzylglutamate) at high pH are charged and are in extendedconformation The efficient pore size is reduced, andthe flow through the membrane is low (“off-state” of themembrane) As pH decreases, the macromolecules areprotonated, lose their charge, and adopt a compact confor-mation The efficient pore size and hence the flow throughthe membrane increases (“on-state” of the membrane)(71) The fluxes of bigger molecules (dextrans of molecularweights 4400–50600) across a temperature-sensitive,poly(NIPAAM)-grafted membrane were effectively con-trolled by temperature, environmental ionic strength,and degree of grafting of the membrane, while the flux ofsmaller molecules such as mannitol was not affected bytemperature even at high degree of membrane grafting(75) The on-off permeability ratio for different molecules(water, Cl− ion, choline, insulin, and albumin) rangedbetween 3 and 10 and increased as molecular weight in-creased (76) An even more abrupt change of the on-off per-meability ratio was observed for a membrane that had nar-row pores formed by heavy ion beams when poly(NIPAAM)

or poly(acryloyl-L-proline methyl ester) were grafted (77).Different stimuli could trigger the transition of thesmart polymer making it possible to produce membraneswhose permeabilities respond to these stimuli When

a copolymer of NIPAAM with triphenylmethane cocianide was grafted to the membrane, it acquiresphotosensitivity—UV irradiation increases permeationthrough the membrane (78) Fully reversible, pH-switchable permselectivity for both cationic and anionicredox-active probe molecules was achieved by deposit-ing composite films formed from multilayers of amine-

leu-terminated dendrimers and poly(maleic

anhydride-co-methylvinyl ether) on gold-coated silicon (79)

When the smart polymer is grafted inside thepores of the chromatographic matrix for gel permeationchromatography, the transition of grafted macromoleculesregulates the pore size and as a result, the elution profile

of substances of different molecular weights As the perature is raised, the substances are eluted progressivelyearlier indicating shrinking of the pores of the hydrogel

Figure 6 Schematic of a “chemical valve.” Glucose oxidase is

immobilized on a pH-responsive polyacrylic acid grafted onto a

porous polycarbonate membrane: (a) poly(acrylic acid) is in an

ex-panded conformation that blocks insulin transport; (b) the

oxida-tion of glucose is accompanied by a decrease in pH and the

transi-tion of poly(acrylic cid) into a compact conformatransi-tion that results in

opening of the pores and transport of insulin [redrawn from (82)].

beads composed of cross-linked

poly(acrylamide-co-N-isopropylacrylamide) (80) or porous polymer beads with

grafted poly(NIPAAM) (81)

When using a specific biorecognition element, which

recognizes specific substances and translates the signal

into a change of physicochemical properties, for

exam-ple, pH, a smart membrane that changes its permeability

in response to particular substances can be constructed

Specific insulin release in response to increasing glucose

concentration, that is, an artificial pancreas, presents an

everlasting challenge to bioengineers One of the potential

solutions is a “chemical valve” (Fig 6) The enzyme,

glu-cose oxidase, was used as a biorecognition element, capable

of specific oxidation of glucose accompanied by a decrease

in pH The enzyme was immobilized on pH-responsive

poly(acrylic acid) graft on a porous polycarbonate

mem-brane In neutral conditions, polymer chains are densely

charged and have extended conformation that prevents

insulin transport through the membrane by blocking the

pores Under exposure to glucose, the pH drops as the

re-sult of glucose oxidation by the immobilized enzyme, the

polymer chains adopt a more compact conformation that

diminishes the blockage of the pores, and insulin is

trans-ported through the membrane (82) Systems such as this

could be used for efficient drug delivery that responds to the

needs of the organism A membrane that consists of

poly(2-hydrohyethyl

acrylate-co-N,N-diethylaminomethacrylate-co-4-trimethylsilylstyrene) undergoes a sharp transition

from a shrunken state at pH 6.3 to a swollen state at

pH 6.15 The transition between the two states changesthe membrane permeability to insulin 42-fold Copolymercapsules that contain glucose oxidase and insulin increaseinsulin release five fold in response to 0.2 M glucose Afterglucose removal, the rate of insulin release falls back to theinitial value (83)

Alternatively, reversible cross-linking of polymermacromolecules could be used to control the porosity in

a system Two polymers, poly(m-acrylamidophenylboronic acid-co-vinylpyrrolidone) and poly(vinyl alcohol) form a gel

because of strong interactions between boronate groupsand the hydroxy groups of poly(vinyl alcohol) When

a low molecular weight polyalcohol such as glucose isadded to the gel, it competes with poly(vinyl alcohol) forboronate groups The boronate–poly(vinyl alcohol) com-plex changes to a boronate–glucose complex that results

in eventual dissolution of the gel (84) In addition to aglucose oxidase-based artificial pancreas, the boronate–poly(vinyl alcohol) system has been used for constructingglucose-sensitive systems for insulin delivery (29,85–87).The glucose-induced transition from a gel to a sol statedrastically increases the release of insulin from the gel.The reversible response to glucose has also been designedusing another glucose-sensitive biorecognition element,Concanavalin A, a protein that contains four sites that canbind glucose Polymers that have glucose groups in the

side chain such as poly(vinylpyrrolidone-co-allylglucose)

(26) or poly(glucosyloxyethyl methacrylate) (88), are versibly cross-linked by Concanavalin A and form a gel.The addition of glucose results in displacing the glucose-bearing polymer from the complex with Concanavalin Aand dissolving the gel

re-Reversible gel-formation of thermosensitive blockcopolymers in response to temperature could be utilized

in different applications Poly(NIPAAM) block copolymerswith poly(ethylene oxide) which undergo a temperature-induced reversible gel–sol transition were patented asthe basis for cosmetics such as depilatories and bleach-ing agents (89) The copolymer solution is liquid atroom temperature and easily applied to the skin where

it forms a gel within 1 min Commercially availableethyl(hydroxyethyl)celluloses that have cloud points of65–70◦C have been used as redeposition agents in wash-ing powders Adsorption of the precipitated polymer on thelaundry during the initial rinsing period counteracts read-sorption of dirt when the detergent is diluted (90)

Liposomes That Trigger Release of the Contents

When a smart polymer is attached somehow to a lipidmembrane, the transition in the macromolecule affects theproperties of the membrane and renders the system sensi-tive to environmental changes To attach a smart polymer

to a lipid membrane, a suitable “anchor” which could beincorporated in the membrane, should be introduced intothe macromolecule This could be achieved by copolymer-izing poly(NIPAAM) with comonomers that have large hy-

drophobic tails such as N, N-didodecylacrylamide (91),

us-ing a lipophilic radical initiator (92) modifyus-ing copolymers(93), or polymers that have terminally active groups (94)

with a phospholipid Alternatively, smart polymers have

been covalently coupled to the active groups in the

hy-drophilic heads of the lipid-forming membrane (95)

Interesting and practically relevant materials for

study-ing the behavior of smart polymers attached to lipid

mem-branes, are liposomes, self assembled 50–200 nm vesicles

that have one or more (phospho)lipid bilayers which

en-capsulate a fraction of the solvent Liposomes are stable

in aqueous suspension due to the repulsive forces that

ap-pear when two liposomes approach each other Liposomes

are widely used for drug delivery and in cosmetics (96)

The results of a temperature-induced conformational

transition of a smart polymer on the liposomal

sur-face depend significantly on the fluidity of the liposomal

membrane When the membrane is in a fluid state at

temperatures both above and below the polymer transition

temperature, the collapse of the polymer molecule forces

anchor groups to move closer together by lateral diffusion

within the membrane The compact globules of collapsed

polymer cover only a small part of the liposomal surface

Such liposomes have a low tendency to aggregate because

the most of their surface is not covered by the polymer

Naked surfaces contribute to the repulsion between

lipo-somes On the other hand, when the liposomal membrane

is in a solid state at temperatures both above and below

the polymer transition temperature, the lateral diffusion

of anchor groups is impossible, and the collapsed polymer

cannot adopt a compact globule conformation but spreads

over the most of the liposomal surface (97) Liposomes

whose surfaces are covered to a large degree by a collapsed

polymer repel each other less efficiently than intact

lipo-somes The stability of a liposomal suspension is thereby

decreased, and aggregation and fusion of liposomes takes

place, which is often accompanied by the release of the

liposomal content into the surrounding medium (98)

When the liposomal membrane is perturbed by the

con-formational transition of the polymer, both the aggregation

tendency and liposomal permeability for incorporated

substances are affected Poly(ethacrylic acid) undergoes a

transition from an expanded to a compact conformation in

the physiological pH range of 7.4–6.5 (99) The pH-induced

transition of poly(ethacrylic acid) covalently coupled to the

surface of liposomes formed from phosphatidylcholine

results in liposomal reorganization into more compact

micelles and concomitant release of the liposomal content

into the external medium The temperature-induced

tran-sition of poly(NIPAAM-co-N,N-didocecylacrylamide) (100)

or poly(NIPAAM-co-octadecylacrylate) (101), incorporated

into the liposomal membrane, enhanced the release of the

fluorescent marker, calcein, encapsulated in

copolymer-coated liposomes Liposomes hardly release any marker

at temperatures below 32◦C (the polymer transition

temperature), whereas the liposomal content is released

completely within less than a minute at 40◦C To increase

the speed of liposomal response to temperature change, the

smart polymer was attached to the outer and inner sides of

the lipid membrane The polymer bound only to the outer

surface if the liposomes were treated with the polymer

af-ter liposomal formation When the liposomes were formed

directly from the lipid–polymer mixture, the polymer was

present on both sides of the liposomal membrane (91)

Changes of liposomal surface properties caused bypolymer collapse affect liposomal interaction with cells.Liposomes modified by a pH-sensitive polymer, partiallysuccinilated poly(glycidol), deliver calcein into culturedkidney cells of the African green monkey more effi-ciently compared to liposomes not treated with the poly-mer (102) Polymeric micelles formed by smart polymersand liposomes modified by smart polymers could be usedfor targeted drug delivery Polymeric micelles have beenprepared from amphiphilic block copolymers of styrene(forming a hydrophobic core) and NIPAAM (forming athermosensitive outer shell) The polymeric micelles werevery stable in aqueous media and had long blood circu-lation because of small diameter, unimodal size distribu-tion (24± 4 nm), and, a low critical micellar concentration

of around 10µg/mL At temperatures above the polymer

transition temperature (32◦C), the polymer chains thatform an outer shell collapse, become more hydrophobic, andallow aggregation between micelles and favoring bindinginteractions with the surface of cell membranes Thus, hy-drophobic molecules incorporated into the micelles are de-livered into the cell membranes These micelles are capable

of site-specific delivery of drugs to the sites as temperaturechanges, for example, to inflammation sites of increasedtemperature (103)

Smart Polymers in Bioanalytical Systems

Because smart polymers can recognize small changes inenvironmental properties and respond to them in a pro-nounced way, they could be used directly as sensors ofthese changes, for example, a series of polymer solutionsthat have different LCSTs could be used as a simple ther-mometer As salts promote hydrophobic interactions anddecrease the LCST, the polymer system could “sense” thesalt concentration needed to decrease the LCST belowroom temperature A poly(NIPAAM)-based system thatcan sense NaCl concentrations above 1.5% was patented(104) The response of the polymer is controlled by a bal-ance of hydrophilic and hydrophobic interactions in themacromolecule Using a recognition element that can senseexternal stimuli and translate the signal into the changes

of the hydrophilic/hydrophobic balance of the smart mer, the resulting system presents a sensor for the stimu-lus If the conjugate of a smart polymer and a recognition el-

poly-ement has a transition temperature T1in the absence and

T2in the presence of stimuli, fixing the temperature T in the range T1 < T < T2allows achieving the transition of asmart polymer isothermally by the external stimulus (105)

An example of such a sensor was constructed using trans–cis isomerization of the azobenzene chromophore when ir-radiated by UV light The transition is accomplished by anincrease in the dipole moment of azobenzene from 0.5 D

(for the trans-form) to 3.1 D (for the cis-form) and hence a

significant decrease of hydrophobicity Irradiation with UVlight results in increasing the LCST from 19.4 to 26.0◦C forthe conjugate of the chromophore with poly(NIPAAM) Thesolution of the conjugate is turbid at 19.4◦C< T < 26.0◦C,but when irradiated, the conjugate dissolves because thecis-form is below the LCST at this temperature The sys-tem responds to UV light by transition from a turbid to

transparent solution The termination of UV irradiation

results in a slow return of the system to its initial

tur-bid state (105) A few other light-sensitive systems were

proposed that use different chromophores:

triarylmethyl-cyanide (106) and leuconitriles (107)

The hydrophobicity of the recognition molecule was also

changed by chemical signals Poly(NIPAAM) containing

11.6 mol% of crown ether 9 has a LCST of 31.5◦C in the

absence of Na+ or K+ ions, 32◦C in the presence of Na+,

and 38.9◦C in the presence of K+ Thus, the introduction

of both Na+and K+ions leads to the dissolution of the

in-soluble polymer at that temperature At 37◦C, this effect is

achieved only by K+ions (108)

From better understanding of ligand–host interactions

and development of new highly selective binding pairs (e.g.,

by using combinatorial libraries to find ligands of high

affinity for particular biomolecules), one could expect that

smart polymer systems will be used as “signal amplifiers”

to visualize a physicochemical event, which takes place

in a recognition element, by a pronounced change in the

system—conversion of a transparent solution into a turbid

one or vice versa

Antibody–antigen interactions present nearly ideal

analytical selectivity and sensitivity developed by nature

Not surprisingly, they are increasingly used for a broad

variety of bioanalytical applications Different analytical

formats have been developed The common feature of

the most of them is the requirement for separating an

antibody–antigen complex from a nonbound antibody or

antigen Traditionally, the separation is achieved by

cou-pling one of the components of antibody–antigen pair to a

solid support The binding step is followed by washing

non-bound material Interactions of the soluble partner of the

binding pair with the partner coupled to the support are

often accompanied by undesired diffusional limitations,

and hence, incubation times of several hours are required

for analysis Because smart polymers can undergo

tran-sition from the soluble to the insoluble state, they allow

combining the advantages of homogeneous binding and,

after the phase transition of the smart polymer has taken

place, easy separation of the polymer precipitate from the

supernatant The essential features of an immunoassay

that uses smart polymers (named PRECIPIA) are as

follows The covalent conjugate of poly(NIPAAM) with

monoclonal antibodies to the κ-chain of human

im-munoglobulin G (IgG) are incubated for 1 h at room

tem-perature (below the LCST of the conjugate), and the IgG

solution is analyzed Then plain poly(NIPAAM) (to

facili-tate thermoprecipitation of polymer–antibody conjugates)

and fluorescently labeled monoclonal antibodies to theγ

-chain of human IgG are added The temperature is raised

to 45◦C, the precipitated polymer is separated by

centrifu-gation, and fluorescence is measured in the supernatant

(109) Immunoassay systems that use

temperature-induced precipitation of poly(NIPAAM) conjugates with

monoclonal antibodies are not inferior in sensitivity to the

traditional heterogenous immunoassay methods, but

be-cause the antigen–antibody interaction takes place in

solu-tion, the incubation can be shortened to about 1 h (110,111)

The limitations of PRECIPIA as an immunoassay

tech-nique are essentially the same as those of affinity

pre-cipitation, namely, nonspecific coprecipitation of analyzed

protein when poly(NIPAAM) precipitates Polyelectrolytecomplexes that have a low degree of nonspecific protein co-precipitation have also been successfully used as reversiblysoluble carriers for PRECIPIA-type immunoassays (112).The conjugate of antibody and polyanion poly(methacrylicacid) binds to the antigen within a few minutes, and thepolymer hardly exerts any effect on the rate of antigen–antibody binding Subsequent addition of a polycation,

poly(N-ethyl-4-vinyl-pyridinium bromide) in conditions

where the polyelectrolyte is insoluble, results in tive precipitation of the antibody–polymer conjugatewithin 1 min The total assay time is less than 15 min (10)

quantita-In principle, PRECIPIA-type immunoassays could beused for simultaneous assay of different analytes in onesample, provided that conjugates specific toward theseanalytes are coupled covalently to different smart poly-mers that have different precipitating conditions, for ex-ample, precipitation of one conjugate by adding a polymericcounterion followed by thermoprecipitation of the secondconjugate by increasing temperature

Reversibly Soluble Biocatalysts

The transition between the soluble and insoluble state ofstimuli-responsive polymers has been used to develop re-versibly soluble biocatalysts A reversibly soluble biocat-alyst catalyzes an enzymatic reaction in a soluble stateand hence could be used in reactions of insoluble or poorlysoluble substrates/products As soon as the reaction is com-pleted and the products are separated, the conditions (pH,temperature) are changed to promote precipitation of thebiocatalyst The precipitated biocatalyst is separated andcan be used in the next cycle after dissolution The re-versibly soluble biocatalyst acquires the advantages of im-mobilized enzymes (ease of separation from the reactionmixture after the reaction is completed and the possibilityfor biocatalyst recovery and repeated use in many reactioncycles) but at the same time overcomes the disadvantages

of enzymes immobilized onto solid matrices such as sional limitations and the impossibility of using them inreactions of insoluble substrates or products

diffu-Biocatalysts that are reversibly soluble as a function of

pH have been obtained by the covalent coupling of

lysozyme to alginate (113); of trypsin to

poly(acrolein-co-acrylic acid) (114); and of cellulase (115); amylase (115);

α-chymotrypsin, and papain (116) to crylate-co-methacrylic acid) A reversibly soluble cofactor

poly(methylmetha-has been produced by the covalent binding of NAD toalginate (117) Reversibly soluble α-chymotrypsin, peni-

cillin acylase, and alcohol dehydrogenase were produced

by coupling to the polycation component of polyelectrolyte

complexes formed by poly(methacrylic acid) and

poly(N-ethyl-4-vinyl-pyridinium bromide) (118)

Biocatalysts that are reversibly soluble as a function

of temperature have been obtained by the covalent pling of α-chymotrypsin and penicillin acylase to a par- tially hydrolysed poly(N-vinylcaprolactam) (119); and of

cou-trypsin (120); alkaline phosphatase (121),α-chymotrypsin

(122), and thermolysin (123,124) to NIPAAM copolymersthat contain active groups suitable for covalent coupling ofbiomolecules Lipase was coupled to a graft copolymer com-

posed of NIPAAM grafts on a poly(acrylamide-co-acrylic

acid) copolymer (125) No significant differences in

bio-catalytic properties were found forα-amylase coupled to

poly(NIPAAM) via single-point or multipoint mode Both

enzyme preparations demonstrated increased

thermosta-bility and the absence of diffusional limitation when

hy-drolyzing starch, a high molecular weight substrate (126)

The temperature of a protein–ligand interaction was

con-trolled by site-directed coupling of terminally modified

poly(NIPAAM) to a specifically constructed site (close to

a biotin binding site) on a genetically modified

strepta-vidin (127)

Biocatalysts which are reversibly soluble as a function of

Ca2 +concentration were produced by covalent coupling of

phosphoglyceromutase, enolase, peroxidase, and pyruvate

kinase toα s1-casein The enzyme casein conjugates are

sol-uble at a Ca2 + concentration below 20 mM but

precipi-tate completely at a Ca2 +concentration above 50 mM The

precipitate redissolves when EDTA, a strong Ca2 +-binding

agent is added (128)

The reversible flocculation of latices has been used to

produce thermosensitive reversibly soluble (more precisely

reversibly dispersible) biocatalysts using trypsin (129),

papain (130), and α-amylase (131) Latices sensitive to

a magnetic field have been used to immobilize trypsin

andβ-galactosidase (132) Liposomes that have a

polymer-ized membrane, that reversibly aggregates on

chang-ing salt concentration have been used to immobilize

α-chymotrypsin (133).

The most attractive application of reversibly soluble

biocatalysts is repeated use in a reaction which is

diffi-cult or even impossible to carry out using enzymes

im-mobilized onto insoluble matrices, for example,

hydroly-sis of water-insoluble phlorizidin (134); hydrolyhydroly-sis of high

molecular weight substrates such as casein (123,130) and

starch (115); hydrolysis of insoluble substrates such as

cellulose (135) and raw starch (corn flour) (7,134,136–

138); production of insoluble products such as peptide,

benzyloxycarbonyl-L-tyrosyl-N ω-nitro-L-arginine (116) and

phenylglycine (139)

The hydrolytic cleavage of corn flour to glucose is an

example of successfully using a reversibly soluble

bio-catalyst, amylase coupled to

poly(methylmethacrylate-co-methacrylic acid), in an industrially interesting process

(136) The reaction product of the process, glucose, inhibits

the hydrolysis The use of a reversibly soluble biocatalyst

improves the efficiency of the hydrolysis which is carried

out at pH 5, at which the amylase–polymer conjugate is

soluble After each 24 h, the pH is reduced to 3.5, the

un-hydrolyzed solid residue and the precipitated conjugate

are separated by centrifugation, the conjugate is

resus-pended in a fresh portion of the substrate at pH 5, and the

hydrolysis is continued The conversion achieved after 5

cy-cles is 67%, and the activity of the amylase after the fifth

cycle was 96% of the initial value (136)

CONCLUSION

In the future, one looks forward to further developments

and the commercial introduction of new smart polymers

whose transition temperatures and pH are compatible with

physiological conditions or conditions for maximal stability

of target biomolecules, such as temperatures of 4–15◦C and

pH values of 5–8 Additional prospects will stem from abetter understanding of the mechanism of cooperativeinteractions in polymers and increasing knowledge ofstructure–property correlations to enable rational synthe-sis of smart polymers that have predefined properties Due

to the possibility of combining a variety of biorecognition

or biocatalytic systems and the unique features of smartpolymers, expectations are running high in this area Onlytime and more experimentation will determine whethersmart polymers will live up to their generous promises

BIBLIOGRAPHY

1 R Dagani, Chem Eng News, pp 30–33 (1995).

2 H Brønsted and J Kopecek, in Polyelectrolyte Gels ties, Preparation, and Applications, R.S Harland and R.K.

Proper-Prud’homme, eds., American Chemical Society, Washington

DC, 1992, pp 285–304.

3 V.A Kabanov, Polym Sci 36: 143–156 (1994).

4 H.G Schild, Prog Polym Sci 17: 163–249 (1992).

5 E.Y Kramarenko and A.R Khokhlov, Polym Gels Networks

6: 46–56 (1998).

6 Eudragit, R¨ohm Pharma GMBH Information Materials, 1993.

7 K Hoshino, H Yamasaki, C Chida, S Morohashi,

M Taniguchi, and M Fujii, J Chem Eng Jpn 30: 30–37

11 M.B Dainiak, V.A Izumrudov, V.I Muronetz, I.Yu Galaev,

and B Mattiasson, Bioseparation 7: 231–240 (1999).

12 U Dissing and B Mattiasson, J Biotechnol 52: 1–10 (1996).

13 A.A Tager, A.P Safronov, S.V Sharina, and I.Yu Galaev,

Col-loid Polym Sci 271 (1993).

14 M Heskins and J.E Guillet, J Macromol Sci Chem A2:

1441–1455 (1968).

15 K Suwa, K Morishita, A Kishida, and M Akashi, J Polym.

Sci Part A: Polym Chem 35: 3087–3094 (1997).

16 I.Yu Galaev and B Mattiasson, Enzyme Microb Technol 15:

354–366 (1993).

17 F Liu, F.H Liu, R.X Zhuo, Y Peng, Y.Z Deng, and Y Zeng,

Biotechnol Appl Biochem 21: 257–264 (1995).

18 S Mori, Y Nakata, and H Endo, Protein Expression Purif.

5: 151–156 (1994).

19 F.M Winnik, A.D Davidson, G.H Hamer, and H Kitano,

Macromolecules 25: 1876–1880 (1992).

20 G Chen and A.S Hoffman, 373: 49–52 (1995).

21 M Maeda, C Nishimura, A Inenaga, and M Takagi,

Reac-tive Functional Polym 21: 27–35 (1993).

22 D Umeno, T Mori, and M Maeda, Chem Commun 1433–

26 S.J Lee and K Park, J Mol Recognition 9: 549–557 (1997).

27 K.-Y.A Wu and K.D Wisecarver, Biotechnol Bioeng 39:

447–449 (1992).

28 E Kokufuta and S Matsukawa, Macromolecules 28: 3474–

3475 (1995).

29 S Kitano, I Hisamitsu, Y Koyama, K Kataoka, T Okano,

M Yokoyama, and Y Sakurai, Proc 1st Int Conf Intelligent Mater., 1993, pp 383–388.

30 T Yakushiji, K Sakai, A Kikuchi, T Aoyagi, Y Sakurai, and

T Okano, Langmuir 14: 4657–4662 (1998).

31 H Kawaguchi, in Biomedical Functions and Biotechnology of

Natural and Artificial Polymers, M Yalpani, ed., ATL Press,

1996, pp 157–168.

32 H Kubota and N Shiobara, Reactive Functional Polym 37:

219–224 (1998).

33 Y.M Lee and J.K Shim, Polymer 38: 1227–1232 (1997).

34 H.A VonRecum, S.W Kim, A Kikuchi, M Okuhara,

Y Sakurai, and T Okano, J Biomed Mater Res 40: 631–639

42 M.B Dainiak, V.A Izumrudov, V.I Muronetz, I.Yu Galaev,

and B Mattiason, Anal Biochem (submitted).

43 M.B Dainiak, V.A Izumrudov, V.I Muronetz, I.Yu Galaev,

and B Mattiason, Biochim Biophys Acta 1381: 279–285

46 K Hoshino, M Taniguchi, T Kitao, S Morohashi, and

T Sasakura, Biotechnol Bioeng 60: 568–579 (1998).

47 P.- ˚A Albertson, Partition of Cell Particles and

Macro-molecules Wiley, NY, 1986.

48 P Hughes and C.R Lowe, Enzyme Microb Technol 10: 115–

122 (1988).

49 G Johansson and F Tjerneld, in Highly Selective Separations

in Biotechnology, G Street, ed., Blackie Academic &

Profes-sional, London, 1994, pp 55–85.

50 M Lu, P.- ˚ A Albertsson, G Johansson, and F Tjerneld,

J Chromatogr B 680: 65–70 (1996).

51 T.T Franko, I.Yu Galaev, R Hatti-Kaul, N Holmberg,

L B ¨ulow, and B Mattiasson, Biotechnol Tech 11: 231–235

57 N Yamada, T Okano, H Sakai, F Karikusa, Y Sawasaki,

and Y Sakurai, Makromol Chem Rapid Commun 11: 571–

576 (1990).

58 T Okano, N Yamada, M Okuhara, H Sakai, and Y Sakurai,

Biomaterials 16: 297–303 (1995).

59 T Okano, A Kikuchi, Y Sakurai, Y.G Takei, T Aoki, and

N Ogata, J Controlled Release 36: 125–133 (1995).

60 M Yamato, M Okuhara, F Karikusa, A Kikuchi,

Y Sakurai, and T Okano, J Biomed Mater Res 44: 44–52

(1999).

61 Y Ito, G Chen, Y Guan, and Y Imanishi, Langmuir 13:

2756–2759 (1997).

62 T Bohanon, G Elender, W Knoll, P Koeberle, J Lee,

A Offenhaeusser, H Ringsdorf, E Sackmann, and J Simon,

J Biomater Sci Polym ed., 8: 19–39 (1996).

63 H Kanazawa, Y Kashiwase, K Yamamoto, Y Matsushima,

A Kikuchi, Y Sakurai, and T Okano, Anal Chem 69: 823–

75 S Akerman, P Viinikka, B Svarfvar, K Putkonen,

K Jarvinen, K Kontturi, J Nasman, A Urtti, and

P Paronen, Int J Pharm 164: 29–36 (1998).

76 R Spohr, N Reber, A Wolf, G.M Alder, V Ang, C.L.

Bashford, C.A Pasternak, O Hideki, and M Yoshida, J

Con-trolled Release 50: 1–11 (1998).

77 H Omichi, M Yoshida, M Asano, N Nagaoka, H Kubota,

R Katakai, R Spohr, N Reber, A Wolf, G.M Alder, V Ang,

C.L Bashford, and C.A Pasternak, Nucl Instrum Methods

Phys Res Sect B 131: 350–356 (1997).

78 S Kurihara, Y Ueno, and T Nonaka, J Appl Polym Sci 67:

82 Y Imanishi and Y Ito, Pure Appl Chem 67: 2015–2021

(1995).

83 K Ishihara and K Matsui, J Polym Sci Polym Lett ed 24:

413–417 (1986).

84 S Kitano, K Kataoka, Y Koyama, T Okano, and

Y Sakurai, Makromol Chem., Rapid Commun 12: 227–233

(1991).

85 K Kataoka, H Miyazaki, B Miyazaki, T Okano, and

Y Sakurai, J Am Chem Soc 120: 12694–12695 (1998).

86 S Kitano, Y Koyama, K Kataoka, T Okano, and Y Sakurai,

J Controlled Release 19: 162–170 (1992).

87 D Shiino, K Kataoka, Y Koyama, M Yokoyama, T.

Okano, and Y Sakurai, J Controlled Release 28: 317–319

(1994).

88 K Nakamae, T Miyata, A Jikihara, and A.S Hoffman,

J Biomater Sci Polymer Ed 6: 79–90 (1994).

89 Cosmetic Bases and Cosmetics Containing Polymers with

Sol-Gel Transition Temperature, Jpn Kokai Tokkyo Koho

JP 09227329 A2, Sep 2, 1997, Y Mori, M Yoshioka, and

T Mukoyama.

90 K Lindell, Doctoral Thesis, Lund University, Sweden, 1996.

91 H Hayashi, K Kono, and T Takagishi, Bioconjugate Chem.

10: 412–418 (1999).

92 H Ringsdorf, J Simon, and F.M Winnik, in Colloid-Polymer

Intercations Particulate, Amphiphilic, and Biological faces, P.L Dubin and P Tong, eds., American Chemical Soci-

95 D.A Tirrell, in Pulsed and Self-Regulated Drug Delivery,

J Kost, ed., CRC, Boca Raton, 1990, pp 110–116.

96 D.D Lasic, Trends Biotechnol 16: 307–321 (1998).

97 H Ringsdorf, J Venzmer, and F.M Winnik, Angew Chem.

Int Ed Engl 30: 315–318 (1991).

98 H Hayashi, K Kono, and T Takagishi, Bioconjugate Chem.

9: 382–389 (1998).

99 S Sugaj, K Nitta, N Ohno, and H Nakano, Colloid Polym.

Chem 261: 159–165 (1983).

100 K Kono, A Henmi, H Yamashita, H Hayashi, and

T Takagishi, J Controlled Release 59: 63–75 (1999).

101 H Hayashi, K Kono, and T Takagishi, Biochim Biophys.

Acta 1280: 127–134 (1996).

102 K Kono, T Igawa, and T Takagishi, Biochim Biophys Acta.

1325: 143–154 (1997).

103 S Cammas, K Suzuki, C Sone, Y Sakurai, K Kataoka, and

T Okano, J Controlled Release 48: 157–164 (1997).

104 Sensors for Concentrations of Salts, Jpn Kokai Tokkyo Koho

JP 63206653 A2 Aug 25, 1988, Y Tomita.

105 M Irie, in Responsive Gels: Volume Transitions II., K Dusek,

ed., Springer-Verlag, Berlin, 1993, pp 50–65.

106 R Kr¨oger, H Menzel, and M.L Hallensleben, Macromol.

111 Y.G Takei, M Matsukata, T Aoki, K Sanui, N Ogata,

A Kikuchi, Y Sakurai, and T Okano, Bioconjugate Chem.

5: 577–5582 (1994).

112 B.B Dzantiev, A.N Blintsov, A.F Bobkova, V.A Izumrudov,

and A.B Zezin, Dokl Biochem (in Russian) 342: 549–552

119 Y.E Kirsh, I Yu Galaev, T.M Karaputadze, A.L Margolin,

and V.K Svedas, Biotechnology (Russian), 184–189 (1987).

120 H Lee and T.G Park, Biotechnol Prog 14(3): 508–516

(1999).

121 A.L Nguyen and J.H.T Luong, Biotechnol Bioeng 34: 1186–

1190 (1989).

122 J.-P Chen, J Chem Technol Biotechnol 73: 137–143 (1995).

123 F Liu, G Tao, and R Zhuo, Polymer J 25: 561–567 (1993).

124 K Hoshino, M Taniguchi, H Kawaberi, Y Takeda,

S Morohashi, and T Sasakura, J Ferment Bioeng 83: 246–

252 (1997).

125 S Takeuchi, I Omodaka, K Hasegawa, Y Maeda, and

H Kitano, Makromol Chem 194: 1991–1999 (1993).

126 J.-P Chen, D.-H Chu, and Y.-M Sun, J Chem Technol.

Biotechnol 69: 421–428 (1997).

127 P.S Stayton, T Shimoboji, C Long, A Chilkoti, G Chen,

J.M Harris, and A.S Hoffman, Nature 378: 472–474

(1995).

128 K Okumura, K Ikura, M Yoshikawa, R Sasaki, and

H Chiba, Agric Biol Chem 48: 2435–2440 (1984).

129 H Kitano, C Yan, and K Nakamura, Makromol Chem 192:

2915–2923 (1991).

130 A Kondo, K Imura, K Nakama, and K Higashitani,

J Ferment Bioeng 78: 241–245 (1994).

131 T Kondo, T Urabe, and K Yoshinaga, Colloids Surf A:

Physicochem Eng Aspects 109: 129–136 (1996).

132 A Kondo and H Fukuda, J Ferment Bioeng 84: 337–341

(1997).

133 Y Sun, X.H Jin, X.-Y Dong, K Yu, and X.Z Zhou, Appl.

Biochem Biotechnol 56: 331–339 (1996).

134 K Hoshino, M Taniguchi, H Ueoka, M Ohkuwa, C Chida,

S Morohashi, and T Sasakura, J Ferment Bioeng 82: 253–

139 Yu.V Galaev, Prikl Bioxim Mikrobiol 30: 167–170 (1994).

140 I.Yu Galaev and B Mattiasson, Trends Biotechnol 17: 335–

340 (1999).

POLYMERS, FERROELECTRIC LIQUID

Ferroelectric materials are a subclass of pyro- and

piezo-electric materials (Fig 1) They are very rarely found in

crystalline organic or polymeric materials because

ferro-electric hysteresis requires enough molecular mobility to

reorient molecular dipoles in space So semicrystalline

polyvinylidene fluoride (PVDF) is nearly the only known

compound (1) On the contrary, ferroelectric behavior is

very often observed in chiral liquid crystalline materials,

both low molar mass and polymeric For an overview of

fer-roelectric liquid crystals, see (2) Tilted smectic liquid

crys-tals that are made from chiral molecules lack the symmetry

plane perpendicular to the smectic layer structure (Fig 2)

Therefore, they develop a spontaneous electric

polariza-tion, which is oriented perpendicular to the layer normal

and perpendicular to the tilt direction Due to the

liquid-like structure inside the smectic layers, the direction of

the tilt and thus the polar axis can be easily switched in

external electric fields (see Figs 2 and 4)

Here, we discuss materials (LC-elastomers) that

com-bine a liquid crystalline phase and ferroelectric properties

(preferable the chiral smectic C∗phase) in a polymer

net-work structure (see Fig 3) The coupling of the liquid

crys-talline director to the network or the softness of the

net-work is chosen so that reorientation of the polar axis is still

possible Thus densely cross-linked systems, that possess

a polar axis but cannot be switched (3) will be excluded

FerroelectricPyroelectricPiezoelectric

PS

E

Figure 1 Ferroelectric hysteresis that shows the spontaneous

polarization PS of a ferroelectric material reversed by an applied

electric field E.

It is the role of the network (1) to form a rubbery matrixfor the liquid crystalline phase and (2) to stabilize a direc-tor configuration LC-materials that have these propertiescan be made either (see Fig 3) by covalently linking themesogenic groups to a slightly cross-linked rubbery poly-mer network structure (see Fig 3a) (4–6) or by dispersing

a phase-separated polymer network structure within a lowmolar mass liquid crystal (see Fig 3b) (8,9) Both systemspossess locally a very different structure They may show,however, macroscopically similar properties

LC-elastomers (see Fig 3a) have been investigated indetail (4–7) Although the liquid crystalline phase transi-tions are nearly unaffected by the network, the networkretains the memory of the phase and director pattern dur-ing cross-linking (7) In addition, it freezes fluctuations ofthe smectic layers and leads to a real long range order

in one dimension (11) An attempt to change the tor pattern by electric or magnetic fields in LC-elastomersleads to a deformation of the network and to an elasticresponse (see Fig 4) As a consequence of this, nematicLC-elastomers could never be switched in electric fields, ifthe shape of the elastomer was kept fixed For freely sus-pended pieces of nematic LC-elastomers, shape variations

direc-in electric fields have been observed sometimes (12,13) Inferroelectric liquid crystals, the interaction with the elec-tric field is, however, much larger Thus, it has been possi-ble to prepare real ferroeletric LC-elastomers (see Fig 4)(14,15) In these systems, the polymer network stabilizesone switching state like a soft spring It is, however, softenough to allow ferroelectric switching Therefore the fer-roelectric hysteresis can therefore be measured in thesesystems It is, however, shifted away from zero voltage (seeFig 4)

SYNTHESIS OF FERROELECTRIC LC-ELASTOMERS

The ferroelectric LC-elastomers described so far (14–17,44–46) are mostly prepared from cross-linkable ferroelec-tric polysiloxanes (see Fig 5), which are prepared by hy-drosilylation of precursor polysiloxanes (18) The cross-linking is finally initiated by irradiating a photoradicalgenerator, which leads to oligomerization of acrylamide oracrylate groups (see Fig 5) The functionality of the netpoints is thus high (equal to the degree of polymerization)and varies with the cross-linking conditions

The advantage of this photochemical-initiated linking is that the crosslinking can be started—at will afterthe liquid crystalline polymer is oriented and sufficientlycharacterized in the uncross-linked state (see Fig 6) Theadvantage of using polymerizable groups (acrylates) forcross-linking is that small amounts of these groups are suf-ficient to transform a soluble polymer into a polymer geland that the chemical reactions happens far away from themesogen Cinnamoyl moieties, on the other hand (19), re-quire a high concentration of these groups for cross-linking.The dimers thus formed are, in addition, nonmesogenic.Figure 7 summarizes the ferroelectric LC-elastomers dis-cussed in this article Two different positions of cross-linkable groups are used In polymer P1, the cross-linkinggroup is close to the siloxane chains, which are known to

Elektroden

Figure 2 Schematic of the bistable

switch-ing of a ferroelectric liquid crystal in the

“surface stabilized FLC” configuration.

(a)

(b)

Figure 3 Network: soft, can be transformed like rubber band, but retains its shape and couples to

director orientation because (a) director is preferably parallel (or perpendicular) to polymer chains (LC-elastomer) (4–8) (b) Director aligns (parallel) to chains in oriented phase-separated polymer network structure (low molar mass LC in LC-thermoset) (8,37).

microphase separate from the mesogenic groups (18,19)

Therefore, the crosslinking should proceed mostly within

the siloxane sublayers In polymers P2 and P3, the

cross-linking group is located at the end of mesogens

There-fore, the cross-linking should proceed mostly between

dif-ferent siloxane layers (see Fig 7) A comparison of these

elastomers allows evaluating structure–property

Figure 4 Schematic of the ferroelectric LC-elastomer and its two switching states (14): (a) A

polymer chain acts as cross-linking point by connecting different mesogenic groups attached to the main polymer chains A ferroelectric switching in this elastomer extends polymer chains (b) The entropy elasticity arising from this acts like a spring that stabilizes one state (c) For the uncross- linked system (left) the hysteresis is symmetrical to zero voltage and both states are equal After cross-linking in one polar state (right), only that state is stable with no electric field, and the hysteresis is no longer symmetrical to zero voltage.

were determined in a careful study by Kocot et al (22)

It seems that the electroclinic effect is especially strong

in these polysiloxanes (15) This has implications for the

freezing of a memory of the tilt angle present during

cross-linking Therefore, ferroelectric elastomers, which have

been crosslinked in the smectic A phase while applying

an electric field, produce a stable macroscopic polarization

(tilt) after cooling into the smectic C* phase (17)

Properties of Ferroelectric LC-Elastomers The

crosslink-ing reactions of a series of copolymers analogs to polymer

P2, but differing in the amount of cross-linkable groups

were studied by FTIR spectroscopy (16) These

measure-ments show a decrease of the acrylamide double bond on

irradiation Conversions between 60 to 84% were observed

The uncertainity of the conversion, however, is high

be-cause only very few double bonds are present in polymer

P2 and they are visible in the infrared spectrum at ratherlow intensity

Mechanical measurements, which show how this chemical crosslinking (conversion of double bonds) leads to

photo-an elastic response of the network are, however, still at thebeginning because photo-cross-linking can be performedonly in thin layers of some microns It is best performedbetween two glass slides to exclude oxygen

AFM measurements of photo-cross-linked free ing films show changes in topology during stretching (seeProperties of Ferroelectric LC-Elastomers—AFM Imaging

stand-of Thin Films) (23) They do, however, not allow measuringelastic moduli

The most promising approach to obtaining elastic datafor these ferroelectric elastomers is investigation of LC-elastomer balloons (25,26) For this purpose, an experi-mental setup was developed on the basis of an apparatus

LC network

OCH3

OCH3O

H2N NH2 THF

O Cl CH3

OSi

O(CH2)5 N

HCOO

DCC, DMAP, THF, CH2Cl2

0.9 n

2.7 n 0.1 n

λ = 365 nm in smectic C*

Cl

Figure 5 Synthetic route to the

cross-linkable polysiloxane P2 and the ing preparation of the oriented smectic C* network using UV light in the presence

follow-of a photoinitiator acetophenone) (14).

Monodomain

Electricalfield

Figure 6 Preparation of polar smectic C* monodomains (14,15)

(ITO: indium tin oxide).

designed to study smectic bubbles (25) Freely suspended

films of the uncross-linked material behave like ordinary

smectic films They can be inflated to spherical bubbles

sev-eral mm in diameter (the thickness of a smectic-layer skin

is about 50 nm) These bubbles are stabilized by the

smectic-layer structure and their inner pressure p is

re-lated to the surface tension and the bubble radius R by

the Laplace–Young equation, p ∝ 1/R After exposure to

UV light, the material is cross-linked, and an anisotropic

elastomer is formed When the cross-linked bubbles are

inflated / deflated, the radius–pressure curve reverses its

slope and gives direct access to the elastic moduli of the

ma-terial (26) Because the deformation during inflation of the

balloon is isotropic in the layer plane, the material should

contract in the direction of the layer normal

Mechanical measurements of chemically crosslinked

LC-elastomers have been made extensively (4,5,27,28,

36,41–43) For these systems, it can be shown that

stretch-ing allows orientation of the liquid crystalline phase In

ideal situations, it is thus possible to prepare a ferroelectric

monodomain by stretching (28,30,36) This result can be

rationalized as a two-stage deformation process (see Fig 9)

(36) This possibility of orienting or reorienting the polar

axis mechanically is the basis for the piezoelectric

proper-ties to be discussed later Ferroelectric switching could not

be observed for any of the chemically crosslinked systems

This may occur because chemically cross-linked films aretoo thick (several 100 µm compared to about 10 µm for

photochemically cross-linked systems) and the electricfield applied is therefore too small In addition, the cross-linking density in chemically cross-linked systems is pre-sumably higher

Ferroelectric Properties LC-Elastomers The ferroelectric

properties of the photochemically crosslinked elastomersE1 to E3 differ significantly and depend on the topology

of the network formed For the systems that have layer cross-linking (see Fig 7, E2 and E3), the switch-ing time is increased greatly Therefore, spontaneouspolarization can no longer be determined by the triangu-lar wave method Slow switching is, however, still possi-ble and therefore ferroelectric hysteresis can be measuredoptically (see Figs 4c and 10) (14,15) After photochem-ical cross-linking in a ferroelectric monodomain, the fer-roelectric hysteresis shows stabilization of the orientationpresent during cross-linking At zero external voltage, onlythis state is stable The second switching state can, how-ever, be reached Therefore, the network acts like a springthat stabilized one state because switching to the otherstate leads to a deviation from the most probable confor-mation of the polymer chain (32) (see Fig 10) Then, theshift of the center of the hysteretic loop away from zerovoltage gives the magnitude of the electric field necessary

inter-to balance the mechanical field of the network The metry of the hysteresis increases with the cross-linkingdensity (17) For high cross-linking densities, switchingremains possible only if the spontaneous polarization israther high (17) Otherwise, the network prohibits switch-ing The asymmetry of ferroelectric switching could also

asym-be proven by polarized FT-IR spectroscopy (33) Increasingthe temperature of this ferroelectric elastomer leads to nar-rowing of the hysteretic loop, which is lost at the transition

to the smectic A phase (see Fig 10)

This behavior is best interpreted by plotting the liquidcrystalline potential, the elastic potential of the network,and their superposition in one graph (15) (see Fig 11) Asthe network is formed in the smectic C* phase, an internalelastic field is created, which has its minimum value forthe tilt angle and tilt direction during cross-linking Othertilt angles are destabilized

The elastomer that has preferable intralayer linking (E1, see Fig 7) shows completely different behav-ior (see Fig 12) (17,34) In this case, the switching timeincreases by less than a factor of 2, the polarization canstill be determined, and measurement of the ferroelectrichysteresis shows no stabilization of the switching statepresent during cross-linking Then, the coupling betweenthe orientation of the mesogens and the network confor-mation is obviously very weak The network stabilizes thesmectic layer structure (see Properties of Ferroelectric LC-Elastomers—AFM Imaging of Thin Films), but it does notstabilize the tilt direction Therefore, the polar axis can beswitched easily This is the result of the network topology(see Fig 7) in which interlayer cross-linking is rare

cross-Properties of Ferroelectric LC-Elastomers—AFM Imaging

of Thin Films Freestanding films can be prepared from

O O

O

Cl (CH 2 ) 10

CH 2

Si O

H 3 C

O Si

H 3 C CH 2 CH 2 CH 2 O

O O

CH 2

Si O

H 3 C

O Si

H 3 C CH 2

O Si

H P2

O (CH 2 ) 10

CH 2

Si O

H 3 C

O Si

O C O

Figure 7 Chemical structure and phase transition temperatures of polymers P1–3 (17) (a) P1 is

designed to favor intralayer cross-linking (b) P2–3 forming an interlayer network.

50

PS

P1P2P3

Figure 8 Temperature dependence of the spontaneous

polariza-tion PS for the polymers P1–3 measured by the triangular wave

method (17).

First deformation Second deformation

Figure 9 Two-step deformation process of a chiral smectic C*

elastomer that displays macroscopic polarization at the end (36).

855

Figure 10 Temperature dependence of the optical hysteresis of

elastomer E2 (S C * 49 ◦C SA) (15) (a) Ferroelectric behavior of the

S C * phase (42, 44, 46, and 48 ◦C, respectively) (b) Electroclinic

behavior of the S A phase (50, 54, and 56 ◦C, respectively).

−20000

Figure 11 Effect of network force on the free energy density (15)

2 K below the phase transition, S C * phase: (◦) calculated potential

of the SC*phase, ( ♦) force due to the network (r) superposition of

both.

010203040

50

P1 at 200 V / 10 µmP1 + 1 wt % photoinitiatorE1

T-Tc [°C]

Figure 12 Temperature dependence of the switching timeτ

(de-fined as 0–100% change in transmission) for P1 and E1 [see (17) for comparison].

uncross-linked polymers (see also the smectic balloons inthis context) They can be photo-cross-linked and trans-ferred to a solid substrate Thereafter, the topology of thefilms can be imaged by AFM, which gives direct visualiza-tion of the smectic layer structure at low temperatures Theuncross-linked polymers can only be imaged at low temper-atures, deep inside the smectic phase and in the tappingmode, which does not induce strong lateral forces At highertemperatures, the sample is too soft and mobile to allowimaging Cross-linked elastomers, on the other hand, aremechanically stable, and films sustain the tapping modeand also the contact mode of the atomic force microscope(35) This holds both for intra- and interlayer cross-linkedsystems Because measurements can be done in all phases,

it is also possible to determine the change of the tic layer thickness at the phase transitions directly Forelastomer E1, for example, the smectic layer thickness is4.2 nm in the smectic C* phase (36◦C, tilt angle about 30◦)

smec-It increases to 4.4 nm at 50◦C in the smectic A phase (35).This corresponds to X-ray measurements

To analyze the impact of the molecular structure onnetwork properties, elastomers are compared, which areidentical except for the molecular position of the cross-linkable group: (1) elastomer E1 that has cross-linkablegroups attached to the backbone via a short spacer(intralayer cross-linking) and (2) elastomer E2 where thecross-linkable group is in the terminal position of a meso-genic side group (intralayer cross-linking) (23,24) Whenmechanical stress (stretching) is imposed on thin films inhomeotropic orientation, the two elastomers react differ-ently to the deformation (23,24), as seen by AFM imag-ing of the surface topology (see Fig 13) For elastomer E1,

“intralayer” cross-linking results in two-dimensional works in the backbone layers, separated by liquid-like FLCside-group layers Because there are practically no verticalconnections in this intralayer network, no vertical distor-tions occur Therefore, this elastomer can be stretched up

net-to 100%, the surface remains smoth, and the layers deform

(a) (b)

Figure 13 Surface topography of as prepared and stretched transferred films of elastomer

Scale bars 1µm, height scale 25 nm valid for all images The surface of all polymers show plateau

patterns In the E2 samples, the lateral strain leads to surface deformation (24).

affinely In elastomer E2, a three-dimensional, “interlayer”

network is formed; the system reacts by distorting the

smectic layering Therefore, only smaller stretching ratios

are accessible and the surface roughens and buckles during

stretching The distortion strength increases with a higher

cross-linking density

Piezoelectric Properties of Ferroelectric LC-Elastomers.

Because a ferroelectric material has to be piezoelectric (see

Fig 1), observation of a piezoresponse is natural It has

been observed for ferroelectric LC-elastomers (14,15,44–

46) and also for more densely cross-linked systems (36–

39) for which no ferroelectric switching could be observed

For the elastomers described here it is, however, possible

to change the piezoresponse (14) by reorienting the

po-lar axis in an external field (see E2 in Fig 14) For this

experiment, the polar axis was kept in one orientationduring cross-linking This resulted in a positive piezore-sponse (see Fig 14) Thereafter, the direction of the polaraxis was inverted by applying an external field of oppositedirection Then, the external field was removed and thepiezocoefficient was measured At first, a piezoresponse ofopposite sign (negative) but identical value is determined(see Fig 14) In the field-free state, this piezoresponse con-tinuously decreases, it goes through zero, increases again,and finally reaches the original positive value This ex-periment is comparable to the hysteresis measurements

of Figs 4 and 10 because it shows that two polar statesare accessible, but the one present during cross-linking isstabilized

The shape variation under application of an externalelectric field was most intensively studied for microtomized

Figure 14 Relaxation of the piezocoefficient d33of elastomer E2

at room temperature after reversal poling at 65 ◦C (SAphase) (14).

piezes of ferroelectric elastomers, which had been oriented

by drawing (44–46) These experiments show only a small

shape variation if the field is applied parallel to the polar

axis of the monodomain The effects become, on the other

hand, rather large if the smectic layer structure (chevron

h0

E = 0

for E = 0 for E >0 for E <0

E ≠ 0

z smectic layers z smectic layers

Single smectic layer

Electricfield E

∆h/z

z

h0−∆h

Figure 15 In the SA* phase, the mesogenic parts (depicted as

ellipsoids) of the elastomeric macromolecule stand upright (θ = 0◦)

inside the single smectic layers By applying a lateral electric field

(perpendicular to the plane of the paper), a tilt angleθ that is

proportional to the electric field E can be induced (electroclinic

effect) The sign of the tilt depends on the sign of the electric field E.

Hence, each smectic layer shrinks byh/z twice during one period

of the electric field The shrinkagehof the whole film is measured

by the interferometer as an optical phase shift between the sample

beam and the reference beam.

Table 1.

Electrostrictive strainε

Material (at electric field strength E) Ref Free-standing FLCE film 4% lateral strain (1.5 MV/m) 40 Spin-coated FLCE ∼1% lateral layer thickness 40

shrinkage (2.0 MV/m)

Lead magnesium niobate—

Lead titanate ceramics

electron irradiated poly(vinylidene fluoride trifluoro- ethylene) copolymer PBLG monomolecular, 0.005% (300 MV/m) 49 grafted layer of

poly-γ

-benzyl-L -glutamate

texture) rearranges (44,46) To get a large electrostrictiveresponse, which can be understood on a molecular level, thegeometry presented in Fig 15 was chosen (40) The appli-cation of an electric field parallel to an smectic layer leads

to a tilt of the mesogens (electroclinic effect) Thereby, thethickness of the layer decreases For a stack of layers, theeffect sums up over all layers As a result, the thicknessperpendicular to the smectic layers decreases if a field isapplied parallel to the layers (see Fig 15) Because the elec-troclinic effect is relatively large in these polymers (15,33),

a large variation in thickness is expected

X-ray diffraction measurements prove the electricallyinduced shrinkage of single smectic layers A freestand-ing ultrathin (75 nm) film of a ferroelectric liquid crys-talline elastomer (similiar to E2) was used to measure theshape variation (electrostrictive response) associated withthis (40) It was measured by a high precision (± 3 pm

at 133 Hz) Michelson interferometer The measurements(see Fig 16) exhibit extremely high electrostrictive strains

of 4% in an electric field of only 1.5 MV/m, which is, toour knowledge, a new world record for the correspond-

ing electrostrictive coefficient a The effect exhibits

typi-cal electroclinic behavior, which means that it is caused

by an electrically induced tilt of the chiral LC molecules

As a consequence of chirality, the primary strain is pendicular to the applied field Hence, a new material thathas a giant electrostriction effect is introduced, where theeffect can be fully understood on a molecular level Thecharacterization of these materials is summarized inTable 1

per-CONCLUSION

Ferroelectric LC-elastomers represent an interesting class

of material because they combine the ordering of liquidcrystalline ferroelectric phases and the rubber elasticity

of polymer networks Switching of the electric polarizationleads to deformation of the polymer network, equivalent to

1234567

Figure 16 Electrostriction of a ferroelectric

LC-elastomer (40) Big diagram: Thickness ationh as a function of the applied ac volt- age Uac Interferometric data were obtained

vari-at the fundamental frequency of the electric

field (piezoelectricity, first harmonic: +) and

at twice the frequency (electrostriction, second harmonic: o) Sample temperature: 60 ◦C.

Inset: Electrostrictive coefficient a (+) versus temperature At the temperature where the non-cross-linked polymer would have its phase transition S C *–S A * (about 62.5 ◦C), the tilt an-gle of 0 ◦is unstable That is why the electrocliniceffect is most effective at this temperature An electric field of only 1.5 MV/m is sufficient to in- duce lateral strains of more than 4%.

stretching a spring, and creates a stress in the network of

polymer chains

The interaction of mesogens and the network can be

var-ied by using different topologies of net points: Crosslinking

is carried out either within the siloxane sublayers

(lead-ing to fast switch(lead-ing elastomers) or between the siloxane

sublayers (resulting in an elastomer that favors the

ferro-electric switching state in which the cross-linking reaction

took place)

Because the orientation of the smectic phase couples

to the polymer network, electromechanical measurements

show a piezoelectric effect Mechanical deformation leads

to polarization or an external electric field to deformation of

the sample Applying the electric field parallel to the

smec-tic layer structure leads to an extremly high electrostrictive

strain of 4% in an electric field of only 1.5 MV/m

ACKNOWLEDGMENT

The work in this summary was possible only by close

co-operation with several groups from physics and physical

chemistry We give special thanks to the groups of Profs

F Kremer and R Stannarius (Leipzig) A Helfer is

than-ked for setting up the manuscript

BIBLIOGRAPHY

1 M.G Broadhurst and G.T Davis in Medical Applications

of Piezoelectric Polymers P.M Galetti, D.E De Rossi, and

A.S De Reggi, eds., Gordon and Breach, Amsterdam, 1988,

pp 3–14.

2 J.W Goodby, R Blinc, N.A Clark, S.T Lagerwall, M.A Osipov,

S.A Pikin, T Sakurai, K Yoshino, and B Zeks Ferroelectric

Liquid Crystals, Principles, Properties and Applications,

Ferroelectricity and Related Phenomena Gordon Breach,

Philadelphia, 1991.

3 A Hult, F Sahl´en, Trolls ˚as, S T Lagerwall, D Hermann,

L Komitov, P Rudquist, and B Stebler Liq Cryst 20: 23–28

(1996).

4 R Zentel Angew Chem 101: 1437–1445 (1989).

5 M Brehmer and R Zentel Mol Cryst Liq Cryst 243:

353–376 (1994).

6 F.J Davis J Mater Chem 3: 551–562 (1993).

7 W Gleim, H Finkelmann in Side Chain Liquid Crystalline Polymers C.B McArdle, ed., Blackie and Son, Glasgow, 1989;

H Finkelmann in Liquid Crystallinity in Polymers, A.Ciferri,

10 G.R Mitchell, P.M.S Roberts, K Ahn, F.J Davis, C Hasson,

H Hischmann, and J.A Pople, Macromol Symp 117: 21–31

(1997).

11 G.C.L Wong, W.H de Jeu, H Shao, K.S Liang, and R Zentel.

Nature 389: 576–579 (1997).

12 R Zentel, Field Liq Cryst 1: 589 (1986).

13 N.R Barnes, F.J Davis, and G.R Mitchell, Mol Cryst Liq.

Cryst 13: 168 (1989).

14 M Brehmer, R Zentel, G Wagenblast, and K Siemensmeyer,

Macromol Chem Phys 195: 1891–1904 (1994).

15 M Brehmer, R Zentel, F Giesselmann, R Germer, and

P Zugenmaier, Liq Cryst 21: 589–596 (1996).

16 E Gebhard, M Brehmer, R Zentel, J Reibel, G Decher,

H.M Brodowsky, and F Kremer The Wiley Polymer Networks Group Review Series, Vol 1, Chemical and Physical Networks,

K te Nijenhuis and W.J Mijs, eds., Wiley, Chichester, 1998,

pp 387–397.

17 E Gebhard and R Zentel, Macromol Chem Phys 201:

902–910; 911–922 (2000).

18 H Poths and R Zentel, Liq Cryst 16: 749–767 (1994).

19 L.-C Chien and L.G Cada, Macromolecules 27: 3721 (1994).

20 S Diele, S Oelsner, F Kuschel, B Hisgen, H Ringsdorf, and

R Zentel, Makromol Chem 188: 1993 (1987).

21 H Poths, E Wischerhoff, R Zentel, A Sch¨onfeld, G Henn,

and F Kremer, Liq Cryst 18: 811–818 (1995).

22 A Kocot, R Wrzalik, J.K Vij, M Brehmer, and R Zentel,

25 R Stannarius and C Cramer, Europhys Lett 42: 43 (1998).

26 H Sch ¨ uring, R Stannarius, C Tolksdorf, and R Zentel,

Macromolecules 34: 3962–3972 (2001).

27 J K ¨upfer and H Finkelmann, Macromol Chem Phys 195:

1353–1367 (1994).

28 K.-H Hanus, H Kapitza, F Kremer, G Reckert, S Vallerien,

and R Zentel, Polym Prepr 30: 493 (1989).

29 R Zentel, Liq Cryst 3: 531 (1988).

30 R Zentel, G Reckert, S Bualek, and H Kapitza, Makromol.

38 M Mauzac, H.T Nguygen, F.G Tournilhac, and S.V.

Yablonsky, Chem Phys Lett 240: 461 (1995).

39 S.U Vallerien, F Kremer, E.W Fischer, H Kapitza, R Zentel,

and H Poths, Makromol Chem., Rapid Commun 11: 593

(1990).

40 W Lehmann, H Skupin, C Tolksdorf, E Gebhard R Zentel,

P Kr ¨uger, M L¨osche, and F Kremer, Nature 410: 447–450

44 F Kremer, W Lehmann, H Skupin, L Hartmann, P.

Stein, and H Finkelmann, Polym Adv Technol 9: 672–676

(1998).

45 W Lehmann, P Gattinger, M Keck, F Kremer, P Stein,

T Eckert, and H Finkelmann, Ferroelectrics 208(1–4):

373–383 (1998).

46 W Lehmann, L Hartmann, F Kremer, P Stein,

H Finkelmann, H Kruth, and S Diele, J Appl Phys.

Z OUNAIES ICASE/NASA Langley Research Center Hampton, VA

an external stimulus

The most popular smart materials are piezoelectric terials, magnetostrictive materials, shape-memory alloys,electrorheological fluids, electrostrictive materials, andoptical fibers Magnetostrictives, electrostrictives, shape-memory alloys, and electrorheological fluids are used asactuators; optical fibers are used primarily as sensors.Among these active materials, piezoelectric materialsare most widely used because of their wide bandwidth,fast electromechanical response, relatively low power re-quirements, and high generative forces A classical defini-tion of piezoelectricity, a Greek term for “pressure elec-tricity,” is the generation of electrical polarization in amaterial in response to mechanical stress This phe-nomenon is known as the direct effect Piezoelectric ma-terials also display the converse effect: mechanical defor-mation upon application of an electrical charge or signal.Piezoelectricity is a property of many noncentrosymmet-ric ceramics, polymers, and other biological systems Pyro-electricity is a subset of piezoelectricity, whereby the po-larization is a function of temperature Some pyroelectricmaterials are ferroelectric, although not all ferroelectricsare pyroelectric Ferroelectricity is a property of certaindielectrics that exhibit spontaneous electric polarization(separation of the center of positive and negative electriccharge that makes one side of the crystal positive and theopposite side negative) that can be reversed in direction byapplying an appropriate electric field Ferroelectricity isnamed by analogy with ferromagnetism, which occurs inmaterials such as iron Traditionally, ferroelectricity is de-fined for crystalline materials, or at least in the crystallineregion of semicrystalline materials In the last couple ofyears, however, a number of researchers have explored thepossibility of ferroelectricity in amorphous polymers, that

ma-is, ferroelectricity without a crystal lattice structure (1)

Table 1 Comparison of Properties of Standard Piezoelectric Polymer and Ceramic Materials

31 (pm/V) (mV-m/N) k31 Salient Features Polyvinylidene fluoride 28 240 0.12 Flexible, lightweight, low

impedance Lead zirconium titanate 175 11 0.34 Brittle, heavy, toxic (PZT)

aValues shown are absolute values of constants.

Characteristics of Piezoelectric Polymers

The properties of polymers are very different from those

of inorganics (Table 1), and they are uniquely qualified to

fill niche areas where single crystals and ceramics

can-not perform as effectively As can-noted in Table 1, the

piezo-electric strain constant (d31) for the polymer is lower than

that of the ceramic However, piezoelectric polymers have

much higher piezoelectric stress constants (g31) which

in-dicates that they are much better sensors than ceramics

Piezoelectric polymeric sensors and actuators offer the

advantage of processing flexibility because they are

lightweight, tough, readily manufactured in large areas,

and can be cut and formed into complex shapes

Poly-mers also exhibit high strength and high impact

resis-tance (2) Other notable features of polymers are low

di-electric constant, low elastic stiffness, and low density,

which result in high voltage sensitivity (excellent sensor

characteristic) and low acoustic and mechanical impedance

(crucial for medical and underwater applications)

Poly-mers also typically possess high dielectric breakdown and

high operating field strength, which means that they can

withstand much higher driving fields than ceramics

Poly-mers offer the ability to pattern electrodes on the film

surface and pole only selected regions Based on these

features, piezoelectric polymers possess their own

estab-lished area for technical applications and useful device

configurations

Structural Requirements for Piezoelectric Polymers

The piezoelectric mechanisms for semicrystalline and

amorphous polymers differ Although the differences are

distinct, particularly with respect to polarization

stabil-ity, in the simplest terms, four critical elements exist for

all piezoelectric polymers, regardless of morphology These

essential elements are: (1) the presence of permanent

mole-cular dipoles, (2) the ability to orient or align the molemole-cular

dipoles, (3) the ability to sustain this dipole alignment once

it is achieved, and (4) the ability of the material to undergo

large strains when mechanically stressed (3)

SEMICRYSTALLINE POLYMERS

Mechanism of Piezoelectricity in Semicrystalline Polymers

Semicrystalline polymers must have a polar crystalline

phase to render them piezoelectric The morphology of such

polymers consists of crystallites dispersed within phous regions, as shown in Fig 1a The amorphous regionhas a glass transition temperature that dictates the me-chanical properties of the polymer, and the melting temper-ature of the crystallites dictates the upper limit of the use

amor-Electric Field Direction

1c Electrically Poled

Crystalline Region

1a Melt Cast

Amorphous Region

1b Mechanically Oriented

Stretch Direction

Figure 1 Schematic illustration of random stacks of amorphous

and crystal lamellae in the PVDF polymer: (a) the morphology after the film is melt cast; (b) after orientation of the film by me- chanically stretching several times its original length; and (c) after depositing metal electrodes and poling through the film thickness.

temperature The degree of crystallinity in such polymers

depends on the method of preparation and thermal

his-tory Most semicrystalline polymers have several

polymor-phic phases, some of which may be polar Mechanical

ori-entation, thermal annealing, and high-voltage treatment

are all it has been shown, effective in inducing crystalline

phase transformations Stretching the polymer aligns the

amorphous strands in the film plane, as shown in Fig 1b,

and facilitates uniform rotation of the crystallites by an

electric field Depending on whether stretching is

uniax-ial or biaxuniax-ial, the electrical and mechanical properties

(and therefore the transduction response) are either highly

anisotropic or isotropic in the plane of the polymer sheet

Electrical poling is accomplished by applying an electric

field across the thickness of the polymer, as depicted in

Fig 1c An electric field of the order of 50 MV/m is

typi-cally sufficient to effect crystalline orientation Polymers

can be poled by using a direct contact method or corona

discharge The latter is advantageous because contacting

electrodes is not required and samples of large area can

be poled continuously This method is used to manufacture

commercial poly(vinylidene fluoride) (PVDF) film Some

re-searchers have also successfully poled large areas of

poly-mer films by sandwiching them between polished metal

plates under a vacuum This method eliminates

electri-cal arcing of samples and the need for depositing metal

electrodes on the film surface The amorphous phase of

semicrystalline polymers supports the crystal orientation,

and polarization is stable up to the Curie temperature

This polarization can remain constant for many years if it

is not degraded by moisture uptake or elevated

tempera-tures

Piezoelectric Constitutive Relationships

The constitutive relationships that describe piezoelectric

behavior in materials can be derived from thermodynamic

2

3

15

Piezoelectricity is a cross coupling among the elastic

variables, stress X , and strain S, and the dielectric ables, electric charge density D and electric field E Note that D is named in analogy to the B field in ferromag-

vari-netism, although some authors also refer to it as tric or electric displacement There does not seem to be astandard nomenclature; however, it is the opinion of theauthors of this article that electric charge density is a bet-ter description of this property The combinations of these

dielec-variables define the piezoelectric strain constant d, the terial compliance s, and the permittivity ε Other piezo- electric properties are the piezoelectric voltage constant g, stress constant e , and strain constant h given by the equa-

ma-tions in Table 2 For a given constant, the first definition inthe table refers to the direct effect, and the second refers tothe converse effect The piezoelectric constants are inter-related through the electrical and mechanical properties

Table 2 Definitions of Piezoelectric Constants

of the material Electric field strength and displacement

charge density are related through the dielectric constant,

εε0(whereε0is the permittivity of free space), and stress

and strain are related through the compliance according

to

d ij = ε0 ε i g i j , (3)

e ij = s ij d ij (4)

The polarization P is a measure of the degree of

piezo-electricity in a given material In a piezoelectric material, a

change in polarizationP results from an applied stress X

or strain S under conditions of constant temperature and

zero electric field A linear relationship exists betweenP

and the piezoelectric constants Due to material anisotropy,

P is a vector that has three orthogonal components in the

1, 2, and 3 directions Alternatively, the piezoelectric

con-stants can be defined as

The electrical response of a piezoelectric material is a

function of the electrode configuration relative to the

di-rection of the applied mechanical stress For a coefficient

d i j, the first subscript is the direction of the electric field

or charge displacement, and the second subscript gives the

direction of the mechanical deformation or stress The C2 ν

crystallographic symmetry typical of synthetic oriented,

poled polymer film leads to cancellation of all but five of

the d i j components (d31 , d32, d33, d15, and d24) If the film is

poled and biaxially oriented or unoriented, d31 = d32, and

d15= d24 Most natural biopolymers possess D∞

symme-try which yields a matrix that possesses only the shear

piezoelectricity components d13 and d25 Because the d33

constant is difficult to measure without constraining the

lateral dimension of the sample, it is typically determined

from Eq (7) which relates the constants to the hydrostatic

piezoelectric constant, d3h.

d3h= d31 + d32 + d33 (7)

The electromechanical coupling coefficient k i j

repre-sents the conversion of electrical energy into mechanical

energy and vice versa The electromechanical coupling can

be considered a measure of transduction efficiency and is

always less than unity as shown here:

k2= electrical energy converted to mechanical energy

input electrical energy ,

(8a)

k2= mechanical energy converted to electrical energy

input mechanical energy .

Figure 3 Typical ferroelectric hysteretic behavior for PVDF.

Some k coefficients can be obtained from a measured d

Ferroelectricity in Semicrystalline Polymers

At high electric fields, the polarization in semicrystallinepolymers such as PVDF is nonlinear with the applied elec-tric field This nonlinearity in polarization is defined ashysteresis The existence of spontaneous polarization toge-ther with polarization reversal (as illustrated by a hystere-sis loop) is generally accepted as proof of ferroelectricity.Figure 3 is an example of the typical hysteretic behavior

of PVDF Two other key properties typically reported forferroelectric materials are the coercive field and the rema-

nent polarization The coercive field Ec, which marks the

point where the hysteresis intersects the horizontal axis,

is about 50 MV/m at room temperature for many

ferroelec-tric polymers The remanent polarization Pr corresponds

to the point where the loop intersects the vertical axis The

values of Ec and Prdepend on temperature and frequency

The Curie temperature Tc, is generally lower than but close

to the melting temperature of the polymer Below Tc, the

polymer is ferroelectric and above Tc, the polymer loses itsnoncentrosymmetric nature

Although ferroelectric phenomenon has been well umented in ceramic crystals, the question of whetherpolymer crystallites could exhibit dipole switching wasdebated for about a decade after the discovery of piezo-electricity in PVDF Inhomogeneous polarization throughthe film thickness that yielded higher polarization on thepositive electrode side of the polymer led to speculationsthat PVDF was simply a trapped charge electret Thesespeculations were dispelled when X-ray studies (6) demon-strated that polarization anisotropy vanishes due to highpoling field strengths and that true ferroelectric dipole re-orientation occurs in PVDF Luongo used infrared to at-tribute the polarization reversal in PVDF to 180◦ dipolerotation (7) Scheinbeim documented the same via X-ray

doc-pole analysis and infrared techniques for odd-numbered

nylons (8)

State of the Art

Pioneering work in the area of piezoelectric polymers (9)

led to the development of strong piezoelectric activity in

polyvinylidene fluoride (PVDF) and its copolymers with

trifluoroethylene (TrFE) and tetrafluoroethylene (TFE)

These semicrystalline fluoropolymers represent the state

of the art in piezoelectric polymers and are currently the

only commercial piezoelectric polymers Odd-numbered

nylons, the next most widely investigated semicrystalline

piezoelectric polymers, have excellent piezoelectric

pro-perties at elevated temperatures but have not yet been

embraced in practical application Other semicrystalline

polymers including polyureas, liquid crystalline polymers,

biopolymers and an array of blends have been studied for

their piezoelectric potential and are summarized in the

fol-lowing section The chemical repeat unit and piezoelectric

constants of several semicrystalline polymers are listed in

Table 3

Polyvinylidene Fluoride (PVDF) Interest in the electrical

properties of PVDF began in 1969 when it was shown (9)

that poled thin films exhibit a very large piezoelectric

co-efficient, 6–7 pCN−1, a value about ten times larger than

had been observed in any other polymer As seen in Table 3,

PVDF is inherently polar The spatially symmetrical

dis-position of the hydrogen and fluorine atoms along the

poly-mer chain gives rise to unique polar effects that influence

the electromechnical response, solubility, dielectric

proper-ties, and crystal morphology and yield an unusally high

di-electric constant The didi-electric constant of PVDF is about

Table 3 Comparison of Piezoelectric Properties of Some Semicrystalline Polymeric Materials

polar The most stable, nonpolarα phase results upon

cast-ing PVDF from a melt and can be transformed into the larβ phase by mechanical stretching at elevated temper-

po-atures or into the polarδ phase by rotating the molecular

chain axis in a high electric field (∼130 MV/m) (10) The

β phase is most important for piezoelectricity and has a

dipole moment perpendicular to the chain axis of 2.1 Dthat corresponds to a dipole concentration of 7× 10−30

Cm After poling PVDF, the room temperature ization stability is excellent; however, polarization andpiezoelectricity degrade as temperature increases and areerased at its Curie temperature Previously, it was believedthat polarization stability was defined only by the melt-ing temperature of the PVDF crystals Recently, however,some researchers suggest that the polarization stability

polar-of PVDF and its copolymers is associated with bic interactions between injected, trapped charges andoriented dipoles in the crystals (11) They hypothesizethat the thermal decay of the polarization is caused bythe thermally activated removal of the trapped chargesfrom the traps at the surface of the crystals The role oftrapped charges in stabilizing orientation in both semicrys-talline and amorphous polymers is still a subject thatneeds further study The electromechanical properties ofPVDF have been widely investigated For more details, the

coulom-reader is referred to the wealth of literature that exists on

the subjects of piezoelectric, pyroelectric, and ferroelectric

properties (2,6,12,13), and the morphology (14–16) of this

polymer

Poly(Vinylidene Fluoride–Trifluoroethylene and

Tetrafluoroethylene) Copolymers Copolymers of

polyvinyli-dene fluoride with trifluoroethylene (TrFE) and

tetraflu-oroethylene (TFE) also exhibit strong piezoelectric,

pyroelectric, and ferroelectric effects These polymers

are discussed together here because they behave

sim-ilarly when copolymerized with PVDF An attractive

morphological feature of the comonomers is that they

force the polymer into an all-trans conformation that has

a polar crystalline phase, which eliminates the need for

mechanical stretching to yield a polar phase P(VDF–

TrFE) crystallizes to a much greater extent than PVDF

(up to 90% crystalline) and yields a higher remanent

polarization, a lower coercive field, and much sharper

hysteretic loops TrFE also extends the use temperature

by about 20◦C to close to 100◦C Conversely, copolymers

with TFE exhibit a lower degree of crystallinity and a

suppressed melting temperature, compared to the PVDF

homopolymer Although the piezoelectric constants for

the copolymers are not as large as those of the

homopoly-mer, the advantages of P(VDF–TrFE) in processibility,

enhanced crystallinity, and higher use temperature make

it favorable for applications

Researchers have recently reported that highly ordered,

lamellar crystals of P(VDF–TrFE) can be made by

anneal-ing the material at temperatures between the Curie

tem-perature and the melting point They refer to this

mate-rial as a “single crystalline film.” A relatively large single

crystal of P(VDF–TrFE) 75/25 mol% copolymer was grown

that exhibits a room temperature d33= −38 pm/V and a

coupling factor k33 = 0.33 (17).

The result of introducing defects into the crystalline

structure of P(VDF–TrFE) copolymer on electroactive

ac-tuation has been studied using high electron irradiation

(18) Extensive structural investigations indicate that

elec-tron irradiation disrupts the coherence of polarization

do-mains (all-trans chains) and forms localized polar regions

(nanometer-sized, all-trans chains interrupted by trans

and gauche bonds) After irradiation, the material

ex-hibits behavior analogous to that of relaxor ferroelectric

systems in inorganic materials The resulting material is

no longer piezoelectric but rather exhibits a large electric

field-induced strain (5% strain) due to electrostriction The

basis for such large electrostriction is the large change in

the lattice strain as the polymer traverses the

ferroelec-tric to paraelecferroelec-tric phase transistion and the expansion

and contraction of the polar regions Piezoelectricity can

be measured in these and other electrostrictives when a dc

bias field is applied Irradiation is typically accomplished

in a nitrogen atmosphere at elevated temperatures using

irradiation dosages up to 120 Mrad

Polyamides A low level of piezoelectricity was first

re-ported in polyamides (also known as nylons) in 1970 (19)

NHC

O

CH2

H2C

CH2HNO

H2C

CH2

H2C

OHN

CH2NH

H2C

CH2

H2C

CH2CNHONylon 5

(b)

Figure 4 Schematic depiction of hydrogen-bonded sheets

show-ing dipole directions in the crystal lattices of (a) even (nylon 4) and (b) odd polyamides (nylon 5).

A systematic study of odd-numbered nylons, however, tiated in 1980 (20), served as the impetus for more than

ini-20 years of subsequent investigations of piezoelectric andferroelectric activity in these polymers The monomerunit of odd nylons consists of even numbers of methy-lene groups and one amide group whose dipole moment is3.7 D Polyamides crystallize in all-trans conformationsand are packed to maximize hydrogen bonding betweenadjacent amine and carbonyl groups, as seen in Fig 4 for

an even-numbered and an odd-numbered polyamide Theamide dipoles align synergistically in the odd-numberedmonomer, resulting in a net dipole moment The amidedipole cancels in an even-numbered nylon, although re-manent polarizations have been measured for some even-numbered nylons, as discussed later in this article Theunit dipole density depends on the number of methylenegroups present, and polarization increases from 58 mC/m2for nylon-11 to 125 mC/m2 for nylon-5 as the number ofmethylene groups decreases (8)

Polyamides are known hydrophilics Because water sorption is associated with hydrogen bonding to the polaramide groups, the hydrophilicity increases as the density

ab-of amide groups increases Water absorption in nylon-11and nylon-7 has been shown to be as high as 4.5% (byweight) and more than 12% for nylon-5 (21), whereas it

is less than 0.02% for PVDF and its copolymers Studieshave shown that water absorption can have a dramatic ef-fect on the dielectric and piezoelectric properties of nylons;however, water does not affect the crystallinity or orienta-tion in thermally annealed films (21) Thus, films can bedried to restore their original properties

At room temperature, odd-numbered nylons havelower piezoelectric constants than PVDF; however, when

examined above their glass transition temperatures,

they exhibit comparable ferroelectric and piezoelectric

properties and much higher thermal stability The

piezo-electric d and e constants increase rapidly as temperature

increases Maximum stable d31 values of 17 pC/N and

14 pC/N are reported for nylon-7 and nylon-11,

respec-tively Corresponding values of the electromechanical

coupling constant k31 are 0.054 and 0.049 Studies have

also shown that annealing nylon films enhances their

polarization stability because it promotes denser packing

of the hydrogen-bonded sheet structure in the crystalline

regions and hinders dipole switching due to lowered free

volume for rotation (22)

Though widely studied, piezoelectric polyamides have

not been widely used in applications due in part to their low

room temperature piezoelectric response and the problem

of moisture uptake

Liquid Crystalline Polymers Liquid crystals consist of

highly ordered rodlike or disklike molecules At their

melt-ing points they partially lose crystalline order and generate

a fluid but ordered state They can form layered structures

called smectic phases or nematic phases that have an

ap-proximately parallel orientation of the molecular long axis

It was first predicted in 1975 that spontaneous polarization

could be achieved in liquid crystals based on symmetry

ar-guments (23) Subsequently, it has been shown that

liq-uid crystalline molecules whose chiral carbon atoms link

a mesogenic group and end alkyl chains may exhibit

fer-roelectric behavior in the smectic C phase (SmC∗) (24) In

this phase, the molecular axis tilts from the normal to the

layer plane, and the molecular dipoles align in the same

direction, yielding a net polarization If such liquid

crys-talline molecules are introduced into the backbone or as a

side group on a polymer, a ferroelectric liquid crystalline

polymer can be obtained There are three requirements for

spontaneous polarization in a liquid crystal: a center of

chirality, a dipole moment positioned at the chiral center

that acts transverse to the molecular long axis, and a tilted

smectic phase (25)

Polyureas Polyureas are thermosets, long used as

in-sulators in a number of applications Until a few years

ago, ureas were available mostly as insoluble powders or

highly cross-linked resins In 1987, a vapor deposition

poly-merization method was successfully developed that was

later applied to synthesizing polyureas (26a,27) Typically,

a vapor deposition technique is used by evaporating OCN–

R1–NH2and H2N–R2–NH2monomers simultaneously on

a substrate (where R1and R2are various aliphatic or

aro-matic groups) This prevents cross-linking and allows

pro-cessing thicknesses in the hundreds of nanometers to tens

of micrometers

An exploration (27) of the dielectric and pyroelectric

properties of polyureas films led to the discovery of their

piezoelectricity From the early 1990s to the present,

various aromatic and aliphatic polyureas were

synthe-sized, and it was shown that they are piezoelectric (28,29)

Aromatic polyureas were the first polyurea structures

identified as piezoeletric They exhibit a piezoelectric

e constant of 15 mC/m2 and have high temperature bility which remains independent of temperature up to

sta-200◦C Their pyroelectric coefficient is high due to their

low dielectric loss compared to other polymers The d

con-stant is about 5 pC/N at room temperature and increases

as temperature increases (28)

Owing to their structures, aliphatic polyureas possesshigher flexibility in their molecular chains Similarly topolyamides, hydrogen bonds play a large role in stabiliz-ing the orientation polarization that is imparted Polyureasthat have an odd number of methyl groups exhibit overallnonzero polarization Polyurea-9 was synthesized and pro-

cessed first, and an e constant of 5 mC/m2 was reported(28) Then, polyureas that have a smaller number of car-bons were attempted because it was surmised that theyshould lead to a higher density of urea bond dipoles Towardthat end, polyurea-5 was synthesized, and it was found

that the e and d constants are twice those of polyurea-9.

Aliphatic polyureas exhibit ferroelectric hysteresis and inaddition, are piezoelectric when they have odd numbers

of methyl groups Their thermal stability and tric coefficients depend highly on the poling temperature(typically 70–150◦C) but are lower than those of aromaticpolyureas

piezoelec-Biopolymers Piezoelectricity of biopolymers was first

reported in keratin in 1941 (30) When a bundle of hairwas immersed in liquid air, an electric voltage of a fewvolts was generated between the tip and the root Whenpressure was applied on the cross section of the bundle,

an electric voltage was generated Subsequently tricity has been observed in a wide range of other biopoly-mers including collagen (31,32), polypeptides like poly-γ -

piezoelec-methylglutamate and poly-γ -benzyl-L-glutamate (33,34),oriented films of DNA (35), poly-lactic acid (36), and chitin

(37) Most natural biopolymers possess D∞ symmetry, sothey exhibit shear piezoelectricity A shear stress in theplane of polarization produces an electric displacementperpendicular to the plane of the applied stress and re-sults in a−d14 = d25piezoelectric constant The piezoelec-tric constants of biopolymers are small relative to syntheticpolymers; they range in value from 0.01 pC/N for DNA

to 2.5 pC/N for collagen The electromechanical effect insuch polymers is attributed to the internal rotation of po-lar atomic groups linked to asymmetrical carbon atoms.Keratin and some polypeptide molecules assume an

α-helical or a β-tcrystalline structure in which the CONH

dipoles align synergistically in the axial direction.Currently, the physiological significance of piezoelec-tricity in many biopolymers is not well understood, but

it is believed that such electromechanical phenomena mayhave a distinct role in biochemical processes For example,

it is known that electric polarization in bone influencesbone growth (38) In one study, a piezoelectric PVDF filmwas wrapped around the femur of a monkey Within weeks,

a remarkable formation of new bone was observed Themotion of the animal caused deformation of the film thatproduced a neutralizing ionic current in the surroundingtissue This minute fluctuating current appears to stimu-late the metabolic activity of bone cells and leads to theproliferation of bone

AMORPHOUS POLYMERS

The purpose of this section is to explain the mechanism

and key components required for developing

piezoelectric-ity in amorphous polymers and to present a summary of

the polarization and electromechanical properties of the

amorphous polymers currently under investigation

Mechanism of Piezoelectricity

Dielectric Theory The piezoelectricity in amorphous

polymers differs from that in semicrystalline polymers and

inorganic crystals in that the polarization is not in a state

of thermal equilibrium, but rather a quasi-stable state

due to the freezing-in of molecular dipoles The result is

a piezoelectric-like effect A theoretical model for polymers

that have frozen-in dipolar orientation was presented to

explain piezoelectricity and pyroelectricity in amorphous

polymers such as polyvinyl chloride (39)

One of the most important properties of an amorphous

piezoelectric polymer is its glass transition temperature

(the temperature below which the material exhibits

glass-like characteristics, and above which it has rubber-glass-like

properties) because it dictates use temperature and

de-fines the poling process conditions Orientation

polariza-tion of molecular dipoles is responsible for piezoelectricity

in amorphous polymers It is induced, as shown in Fig 5,

by applying an electric field Epat an elevated temperature

(Tp≥ Tg) where the molecular chains are sufficiently

mo-bile and allow dipole alignment with the electric field

Par-tial retention of this orientation is achieved by lowering the

temperature below Tg) in the presence of Ep, resulting in

a piezoelectric-like effect The remanent polarization Pris

Figure 5 Poling profile of an amorphous polymer.

directly proportional to Epand the piezoelectric response.The procedure used to prepare a piezoelectric amorphouspolymer clearly results in both oriented dipoles and space

or real charge injection The real charges are usually centrated near the surface of the polymer, and they areintroduced due to the presence of the electrodes Interest-ingly, some researchers (40,41) have shown that the pres-ence of space charges does not significantly affect piezo-electric behavior The reason is twofold The magnitude ofspace charges is usually not significant with respect to po-larization charges Secondly, space charges are essentiallysymmetrical with respect to the thickness of the polymer;therefore, when the material is strained uniformly, the con-tribution to the piezoelectric effect is negligible

con-In what follows, the origins of the dielectric tion to the piezoelectric response of amorphous polymers

contribu-are addressed The potential energy U of a dipole µ at an

angleθ with the applied electric field is U = µ E cos θ

Us-ing statistical mechanics and assumUs-ing Boltzman’s bution of dipole energies, the mean projection of the dipolemomentµ Ein the direction of the applied electric field isobtained:

This is the Langevin equation that describes the

de-gree of polarization in a sample when an electric field E is applied at temperature T Experimentally, a poling tem- perature in the vicinity of Tg) is used to maximize dipolemotion The maximum electric field that may be applied,typically 100 MV/m, is determined by the dielectric break-down strength of the polymer For amorphous polymers,

µE/kT is much less than one; this places these systems

well within the linear region of the Langevin function The

remanent polarization Pris simply the polarization duringpoling minus the electronic and atomic polarizations that

relax at room temperature, once the field Ep is removed.The following linear equation for remanent polarizationresults when the Clausius–Mossotti equation is used torelate the dielectric constant to the dipole moment (42):

It can be concluded that remanent polarization andhence the piezoelectric response of a material are deter-mined byε; this makes it a practical criterion to use when

designing piezoelectric amorphous polymers The dielectricrelaxation strengthε may be the result of either free or

cooperative dipole motion Dielectric theory yields a matical approach for examining the dielectric relaxation

mathe-ε due to free rotation of the dipoles The equation

incor-porates Debye’s work based on statistical mechanics, theClausius–Mossotti equation, and the Onsager local fieldand neglects short range interactions (43):

εcalculated= Nµ2

3kT ε0



n2+ 23

where N is the number of dipoles per unit volume, k is the

Boltzman constant, ε(0) is the static dielectric constant,

and n is the refractive index One way to measure Pr in

amorphous polymers is the thermally stimulated current

(TSC) method (refer to section on characterization) Prcan

be calculated from the liberated charge during TSC, and by

reconciling that with the Onsager relationship, the dipole

density can be calculated:

Pr= Nµ2Ep

3kTp



ε∞+ 23

2

3ε(0)

2ε(0) + ε∞ . (11)The piezoelectric constants are related to the polarization

From basic thermodynamics,

A molecular theory of the direct piezoelectric effect in poled

amorphous piezoelectric polymers has been developed An

expression for the hydrostatic coefficient appears in the

original paper (41) Later, this theory was extended, and

an equation for d31was obtained (44,45) By differentiating

Eq (11) and modifying it to account for dimensional effects

such as stretching (44,46),

d31= Pr(1 − γ )S11+ Pr(1− γ )

3 (ε∞− 1)S11 , (13)whereγ is Poisson’s ratio, ε∞ is the permittivity at high

frequencies, and S11is the compliance of the polymer The

first term accounts for dimensional effects, and the second

term gives the contribution of the local field effect

Polarizability and Poling Conditions Designing an

amor-phous polymer that has a large dielectric relaxation

strength and hence piezoelectric response requires the

ability to incorporate highly polar groups at high

concen-trations and cooperative dipole motion A study of the

re-lationship between relaxation times, poling temperatures,

and poling fields is crucial to achieving optimal dipole

alignment Theoretically, the higher the electric field, the

better the dipole alignment The value of the electric field

is limited, however, by the dielectric breakdown of the

poly-mer In practice, 100 MV/m is the maximum field that can

be applied to these materials Poling times need to be of

the order of the relaxation time of the polymer at the

pol-ing temperature

During poling, the temperature is lowered to room

tem-perature, while the field is still on, to freeze in the dipole

alignment In a semicrystalline material, however,

locking-in the polarization is supported by the crystalllocking-ine

struc-ture of the polymer, and it is therefore stable above the

glass transition temperature of the polymer Because the

remanent polarization in amorphous polymers is lost in

the vicinity of Tg, their use is limited to temperatures well

below Tg This means that the polymers are used in their

glassy state, where they are quite stiff and thus limit the

ability of the polymer to strain as stress is applied A

piezo-electric amorphous polymer may be used at temperatures

near its Tgto optimize its mechanical properties, but not

too close so as to maintain the remanent polarization

Although there are few data that address the ity of piezoelectric activity in amorphous polymers, it isclear that time, pressure, and temperature can contribute

stabil-to dipole relaxation in these polymers For a given cation and use temperature, the effect of these parameters

appli-on the stability of the frozen-in dipole alignment should bedetermined

Examples of Amorphous Piezoelectric Polymers

The literature on amorphous piezoelectric polymers is muchmore limited than that for semicrystalline systems this is

in part because no amorphous piezoelectric polymers; haveresponses high enough to attract commercial interest.Much of the previous work on amorphous piezoelectricpolymers was in nitrile-substituted polymers, includingpolyacrylonitrile (PAN) (47–49), poly(vinylidene cyanidevinyl acetate) (PVDCN/VAc) (50–54), polyphenyletherni-trile (PPEN) (55,56), and poly(1-bicyclobutanecarbonitrile)(57) Weak piezoelectric activity in polyvinyl chloride(PVC) and polyvinyl acetate (PVAc) has also been found(11,41,58,59) The most promising of these materials arevinylidene cyanide copolymers that exhibit large dielectricrelaxation strengths and strong piezoelectricity Table 4shows the molecular structures of the most commonlyencountered amorphous piezoelectric polymers

Polyvinylidene Chloride The carbon–chlorine dipole in

polyvinylidene chloride (PVC) has been oriented to duce a low level of piezoelectricity The piezoelectric andpyroelectric activities generated in PVC are stable and re-producible PVC has been used as a basis for understandingand studying piezoelectricity in amorphous polymers (39)

pro-The piezoelectric coefficients d31of PVC are reportedly inthe range of 0.5–1.3 pC/N This response was improved

by simultaneous stretching and corona poling of film (44)

The enhanced piezoelectric coefficient d31ranged from 1.5–5.0 pC/N

PVDCN Copolymers In 1980, exceptionally strong

piezoelectric activity was found (50) in the amorphouscopolymer of VDCN and VAc The copolymer was poled

at 150◦C (20◦C below its Tg) and cooled to room

tempera-ture in the electric field A Pr= 55 mC/m2was obtained in

a poling field of 50 MV/m That is comparable to the ProfPVDF When local ordering, or paracrystallinity, is inher-ent in the polymer or is induced by mechanical stretching,

an increase in the value of the remanent polarization is served For example, some researchers (51) assert that thelarge discrepancy between the measured and calculated

ob-ε for PVDCN-VAc may be attributed to locally ordered

regions in the polymer.εcalculated= 30 for the copolymerPVDCN/VAc, andεmeasured= 125 (51) This large discrep-ancy in the values ofε is indicative of cooperative mo-

tion of several nitrile dipoles within the locally orderedregions of the polymers Cooperativity means that multi-ple nitrile dipoles respond to the applied electric field in

a unified manner, instead of each dipole acting dently Although the existence of cooperative dipole motionclearly increases the piezoelectric response of amorphouspolymers, the mechanisms by which cooperativity can be

indepen-Table 4 Structure, Polarization, and Tgof Piezoelectric Amorphous Polymers

systematically incorporated into the polymer structure

re-main unclear at this time (59)

The large relaxation strength exhibited by PVDCN/VAc

gives it the largest value of Prand hence d31of all of the

amorphous polymers A number of authors have suggested

that PVDCN-VAc also exhibits ferroelectric-like behavior

(51–53) due to switching of the nitrile dipoles in an ac field

The switching time is long compared to that of a normal

ferroelectric polymer

Other VDCN Polymers The homopolymer of vinylidene

cyanide is thermally unstable (60) and highly sensitive

to moisture, but VDCN can be polymerized with a

vari-ety of monomers in addition to VAc, such as vinyl

ben-zoate (VBz), methyl methacrylate (MMA) and others that

form highly alternating chains All of these copolymers

show some degree of piezoelectricity although lower than

PVDCN-VAc, which is explained by different activation

en-ergies for dipole orientation in the glassy state and

differ-ent chain mobility that depends on the side group

Polyacrylonitrile Polyacrylonitrile (PAN) is one of the

most widely used polymers Shortly after it was shown

that the PVDCN-VAc system is piezoelectric, researchers

turned their attention to PAN due to its similarity to the

aforementioned polymers The presence of the large nitrile

dipole in PAN indicated that it can be oriented by an

ap-plied electric field PAN presented some challenges not

en-countered in other nitrile-substituted polymers, however

Although theoretical calculations predicted strong

piezo-electric behavior, it was difficult to pole Several

investiga-tors (47–49) proposed that the difficulty of poling PAN in

the unstretched state is related to the strong dipole–dipoleinteraction of nitrile groups of the same molecule that re-pel each other, and thus prevent normal polarization Uponstretching, the intermolecular dipole interactions facilitatepacking of the individual chains and give rise to orderedzones The remanent polarization of both unstretched andstretched PAN has been measured using the thermallystimulated current method (TSC), and a twofold increase

in remanent polarization (TSC peak at 90◦C) was observedfor PAN that was stretched to four times its original length(47) Another approach is the copolymerization of PAN withanother monomer Researchers have reported reduction ofthe hindering effect of the dipole–dipole interactions andenhancement of the internal mobility of the polymer seg-ments when PAN is copolymerized with polystyrene ormethyl methacrylate Ferroelectric behavior has been ob-served in P(AN-MMA), where, for given temperature andfield conditions, a characteristic hysteretic loop is obtained(49) It was concluded that it may be one rare examplewhere both ferroelectric and frozen-in dipole orientationsare superimposed

Nitrile-Substituted Polyimide Amorphous polyimides

that contain polar functional groups were synthesized(61–63) and investigated for use as high temperature piezo-electric sensors (β-CN) APB/ODPA polyimide is one such

system The (β-CN) APB/ODPA polyimide possesses the

three dipole functionalities shown in Table 5 Typically,the functional groups in amorphous polymers are pendent

to the main chain The dipoles, however, may also residewithin the main chain of the polymer, such as the anhy-dride units in the (β-CN) APB/ODPA polyimide The nitrile

Table 5 Values of Dipole Moments Within

Nitrile-Substituted Polyimide

Dipole Moment

Pendent nitrile group 4.18

Main chain dianhydride group 2.34 Main chain diphenylether group 1.30

to dipole is pendent a phenyl ring (µ = 4.2 D), and the two

anhydride dipoles (µ = 2.34 D) are within the chain,

re-sulting in a total dipole moment of 8.8 D per repeat unit

The remanent polarization Prof the (β-CN)APB/ODPA

polymer found by the thermally stimulated current

method (TSC) was approximately 20 mC/m2 when poled

at 80 MV/m above the Tgfor 1 hour (64) Excellent

ther-mal stability was observed up to 100◦C, and no loss of the

piezoelectric response was seen after aging at 50◦C and

100◦C for as long as 500 hrs

Partially cured films of the (β-CN)APB/ODPA system

were simultaneously corona poled and cured to enhance

dipolar orientation and minimize localized arcing during

poling The aligned polar groups should be immobilized

by additional imidization and subsequent cooling in an

electric field Both the Tg and the degree of imidization

increase almost linearly as the final cure temperature is

increased (64) The value of Prwas higher for films cured at

lower temperatures The mobility of the molecules in a

par-tially imidized state should be higher than that in the fully

cured state and therefore produce a higher degree of dipole

orientation

The importance of dipole concentration in ultimate

po-larization is evident from a comparison of polyacrilonitrile

(PAN) and the polyimide (β-CN) APB/ODPA PAN has a

single nitrile dipole per repeat unit (µ = 3.5 D) resulting

in a dipole concentration of 1.34 × 1028 m−3 This

trans-lates into an ultimate polarization of 152 mC/m2(20) The

(β-CN) APB/ODPA polyimide, on the other hand, has a

to-tal dipole moment of 8.8 D per monomer The dipole

con-centration of (β-CN) APB/ODPA, however, is only 0.136 ×

1028m−3, resulting in an ultimate polarization of

40 mC/m2, which is less than a fourth of that of PAN As a

result, similar polyimides that have increased nitrile

con-centrations were synthesized and characterized Studies

of these polymers show that polarization is significantly

increased by increasing dipole concentration Structure–

property investigations designed to assess the effects of

these dipoles on Tg, thermal stability, and overall

polariza-tion behavior are currently being pursued

Even-Numbered Nylons Nylon 6I and 6I/6T exhibit a

D–E hysteretic loop across a temperature range of 30–65◦C

at a fixed maximum field of 168 MV/m (65) The remanent

polarization increases as the temperature increases Note

that nylon 6I and 6I/6T are completely amorphous The Pr

is about 30 mC/m2

Aliphatic Polyurethane Some researchers (1) have

sug-gested that aliphatic polyurethane systems exhibit electricity that stems from the amorphous part at temper-atures above the glass transition temperature This “liquidstate” ferroelectricity is very peculiar, seems to exist, and

ferro-is supported by the hydrogen bonds present

CHARACTERIZATION AND MODELING

Most piezoelectric characterization methods were oped for crystalline ceramics and had to be adapted forpiezoelectric polymers Methods based on resonance anal-ysis and equivalent circuits that can be used to characterizesemicrystalline PVDF and its copolymers are outlined

devel-in IEEE standards (66) Details for applydevel-ing resonanceanalysis to piezoelectric polymers have recently been ex-plored (67) Due to the lossy nature of some polymers,the IEEE standards are not adequate, and other tech-niques are needed to describe piezoelectric properties moreaccurately

Quasi-static direct methods are both versatile and wellsuited to investigating fully the piezoelectric response ofpolymers Direct methods of this type are especially ap-propriate for amorphous polymers Thermally stimulatedcurrent measurements (TSC) (68) are used to measure theremanent polarization imparted to a polymer, and directstrain or charge measurements are used to investigate thepiezoelectric coefficients with respect to the electric field,frequency, and stress

TSC is a valuable tool for characterizing piezoelectricpolymers After poling a polymer, a measure of the cur-rent dissipation and the remanent polarization as a func-tion temperature can be obtained by TSC As the sample isheated through its glass transition temperature (or Curietemperature for a semicrystalline polymer) at a slow rate(typically 1–4◦C/min), the depolarization current is mea-sured by an electrometer The remanent polarization isequal to the charge per unit area and is obtained from thedata by integrating the current with respect to time andplotting it as a function of temperature:

Figure 6 illustrates a typical TSC result Because

perma-nent dipoles are immobile at temperatures well below Tg,the current discharge remains low in this temperature

range As temperature increases to and beyond the Tg, ever, the onset of dipole mobility contributes to a significantincrease in the current peak The peak in the current andthe subsequent polarization maximum usually occurs in

how-the vicinity of how-the Tg.

Direct methods for measuring the strain that resultsfrom applying a field or vice versa, applying a strain,and measuring the accumulated charge, are abundant.Interferometers, dilatometers, fiber-optic sensors, opti-cal levers, linear variable displacement transducers, andoptical methods are employed to evaluate the piezoelec-tric strain (converse effect) (69–72) The “out-of-plane” or

thickness piezoelectric coefficient d33can be ascertained as

a function of the driving field and frequency The coefficient

−5051015

Figure 6 Plot of thermally stimulated current for a typical

amor-phous poled polymer.

is measured based on the equation,

where S33 is the strain and E3is the applied electric field

A modified Rheovibron, or similar techniques, have

been used to measure the direct piezoelectric effect, where

charges accumulated on the surfaces of the polymer are

measured (59) The piezoelectric coefficient d31can be

ob-tained by straining the polymer in the direction of

ap-plied stress using a force F A charge Q is generated on

the surface of the electrodes A geometric factor is used to

produce a geometrically independent parameter, surface

charge density per unit applied stress:

d31= Q/(WL)

which has units of pC/N W , L, and t are the width, length,

and thickness of the sample, respectively

Modeling

The methodology for modeling piezoelectric behavior in

polymers varies, depending on the targeted properties

Ap-proaches cover the range from macroscale to micro and

atomistic scales A detailed review of computational

meth-ods applied to electroactive polymers has been published

(73)

PolymerTop electrode

Bottom electrode

Stress

Stress

+ + + +

Figure 7 Direct effect in polymers.

In some cases, modeling can predict behavior whereexperiments cannot Using molecular dynamics, the ori-entation polarization of the (β-CN) APB/ODPA polymer

was assessed by monitoring the angleθ that the dipoles make with an applied electric field (74) The bulk Prwascalculated, and the results agreed extremely well withexperimental results (61) Computational modeling, how-ever, gave insight into the contributions of the variousdipoles present in a way experimental results could not.The model predicted that 40% of the orientation polariza-tion was due to the dianhydride within the backbone of theODPA monomer and demonstrated the importance of theflexible ether linkage (oxygen atom) in facilitating dipolealignment Modeling insight of this kind is invaluable inguiding the synthesis of new materials

Modeling of PVDCN-VAc can also play a role in standing the cooperative motion responsible for the highdielectric relaxation strength of this class of polymers,not possible experimentally (75) Recently, mesoscalesimulation was used to describe polarization reversal inPVDF films (76)

under-Applications and Future Considerations

The potential for applying piezoelectric and other troactive polymers is immense To date, ferroelectric poly-mers have been incorporated into numerous sensing andactuating devices for a wide array of applications Typicalapplications include devices in medical instrumentation,robotics, optics, computers, and ultrasonic, underwater,and electroacoustic transducers One important emergingapplication area for electroactive polymers is the biomedi-cal field where polymers are being explored as artificialmuscle actuators, as invasive medical robots for diagnos-tics and microsurgery, as actuator implants to stimulatetissue and bone growth, and as sensors to monitor vas-cular grafts and to prevent blockages (77,78) Such appli-cations are ideal for polymers because they can be madebiocompatible, and they have excellent conformability andimpedance that match body fluids and human tissue Theintent of this article is not to detail specific applications; theinterested reader may consult excellent sources on appli-cations of piezoelectric and ferroelectric polymers (79–81)

elec-In the future, we believe that fertile research areas forpiezoelectric polymers will include work to enhance theirproperties, to improve their processibility for incorporation

in devices, and to develop materials that have a broaderuse temperature range Fundamental structure–propertyunderstanding has enabled the development of numeroussemicrystalline and amorphous polymers Based on thisknowledge, future research that focuses on property en-hancement via new chemistries that have higher dipoleconcentrations and incorporate dipole cooperativity mayyield improved materials Property enhancements mayalso be gained from processing studies to alter polymermorphology such as those used to make “single crystalline”fluoropolymers Development of materials that can operate

in extreme environments (high temperature and ent temperature) is also important for expanding the use ofpiezoelectric polymers Piezoelectric and pyroelectric con-stants of polymers are considerably lower than those of

subambi-ferroelectric inorganic ceramics Improvements in

proper-ties by incorporating polymers in composites with

inor-ganics to obtain higher electromechanical properties and

better mechanical properties are also valuable To date,

piezoelectric polymer–ceramic composites have been made

wherein the polymer serves only as an inactive matrix

for the active ceramic phase This is due to the mismatch

in permittivity between the polymer and ceramic which

makes it difficult to pole both phases Research that results

in active polymer and ceramic phases could yield

interest-ing electromechanical properties

ACKNOWLEDGMENTS

The authors express sincere appreciation to Dr J.A

Young (Lawrence Livermore) for her technical insight in

the area of amorphous piezoelectric polymers The

au-thors thank Suzanne Waltz (NASA Langley) for graphics

assistance

BIBLIOGRAPHY

1 T Yuki, K Shida, T Koda, and S Ikeda, 10th Int Symp

Elec-trets, 1999, pp 675– 677.

2 G.T Davis, in Polymers for Electronic and Photonic

Appli-cations, C.P Wong, ed., Academic Press, Boston, MA, 1993,

p 435.

3 M.G Broadhurst, and G.T Davis, in Electrets, G.M Sessler,

ed., Springer-Verlag, NY, 1980, Vol 33, p 283.

4 W.P Mason, Physical Acoustics and the Properties of Solids.

Van Nostrand, Princeton, NJ, 1958.

5 J.F Nye, Physical Properties of Crystals Clarendon, Oxford,

1957.

6 R.G Kepler and R.A Anderson, J Appl Phys 49 (3): 1232–

1235 (1978).

7 J.P Luongo, J Polym Sci A-2 10: 1119–1123 (1972).

8 B.Z Mei, J.I Schienbeim, and B.A Newman, Ferroelectrics

144: 51–60 (1993).

9 H Kawai, Jpn J Appl Phys 8: 975 (1969).

10 K Tashiro, H Tadokoro, and M Kobayashi, Ferroelectrics

32: 167 (1981).

11 T Furukawa, IEEE Trans Electr Insul 24: 375–393 (1989).

12 G.M Sessler, J Acoust Soc Am 70 (6): 1596–1608 (1981).

13 A.J Lovinger, Science 220 (4602): 1115–1121 (1983).

14 J.B Lando and W.W Doll, J Macromol Sci.-Phys B2, 205

(1968).

15 A.J Lovinger, in Developments in Crystalline Polymers, D.C.

Basset, ed., Applied Science, London, UK, 1982, Vol 1,

19 H Kawai and I Henji, Oyo Buturi 39: 413 (1970).

20 B.A Newman, P Chen, K.D Pae, and J.I Scheinbeim, J Appl.

30 A.J.P Martin, Proc Phys Soc 53: 183 (1941).

31 E Fukada, in Ferroelectric Polymers, Chemistry, Physics and Application, Marcel Dekker, H S Nalwa, ed., NY, 1995,

p 393.

32 H Maeda and E Fukada, Biopolymers 21: 2055 (1982).

33 M Date, S Takashita, and E Fukada, J Polym Sci A-2

36 E Fukada, Rep Prog Polym Phys Jpn 34: 269 (1991).

37 Y Ando, E Fukada, and M.J Glimcher, Biorheology 14: 175

(1977).

38 E Fukada, in Mechanisms of Growth Control, R.O Becker,

ed., Charles C Thomas, Springfield, IL, 1981, p 192.

39 M.G Broadhurst, C.G Malmberg, F.I Mopsik, and W.P.

Harris, in Electrets: Charge Storage and Transport in electrics, M.M Perlman, ed., The Electrochemical Society,

Di-Princeton, NJ, 1973, pp 492–504.

40 M.G Broadhurst, W.P Harris, F.I Mopsik, and C.G.

Malmberg, Polym Prepr 14: 820 (1973).

41 F.I Mopsik and M.G Broadhurst, J Appl Phys 46: 4204–

45 H Stefanou, J Appl Phys 50: 1486–1490 (1979).

46 R Meyrueix and O Lemonnier, Phys D Appl Phys 27: 379–

386 (1994).

47 R.J Comstock, S.I Stupp, and S.H Carr, J Macromol Sci.

Phys B13 (1): 101–115 (1977).

48 H Ueda and S.H Carr, Polym J 16: 661–667 (1984).

49 H.von Berlepsch, W Kunstler, A Wedel, R Danz, and D Geib,

IEEE Trans Electr Insul 24: 357–362 (1989).

50 S Miyata, M Yoshikawa, S Tasaka, and M Ko, Polym.

J 12: 857–860 (1980).

51 T Furukawa, M Date, K Nakajima, T Kosaka, and I Seo,

Jpn J Appl Phys 25: 1178–1182 (1986).

52 I Seo, Ferroelectrics 171: 45–55 (1995).

53 P.A Mirau and S.A Heffner, Polymer 33: 1156–1161 (1992).

54 M Sakurai, Y Ohta, Y Inoue, and R Chujo, Polym Commun.

57 H.K Hall, R.J.H Chan, J Oku, O.R Hughes, J Scheinbeim,

and B Newman, Polym Bull 17: 135–136 (1987).

58 V Bharti and R Nath, J Appl Phys 82: 3488–3492 (1997).

59 Z Ounaies, J.A Young, and J.S Harrison, in An Overview of

the Piezoelectric Phenomenon in Amorphous Polymers, I.M.

Khan and J.S Harrison, eds., American Chemical Society,

Washington, DC, 1999.

60 S Tasaka, N Inagaki, T Okutani, and S Miyata, Polymer 30:

1639–1642 (1989).

61 Z Ounaies, J.A Young, J.O Simpson, and B.L Farmer, Mater.

Res Soc Proc.: Mater Smart Syst II, Materials Research

So-ciety, Pittsburgh, PA, 1997, Vol 459, p 59.

62 J.O Simpson, Z Ounaies, and C Fay, Mater Res Soc Proc.:

Mater Smart Syst II, Materials Research Society, Pittsburgh,

PA, 1997, Vol 459, p 53.

63 Z Ounaies, C Park, J.S Harrison, J.G Smith, and J Hinkley,

SPIE Proc., Electroactive Polym Actuators Devices, Newport

Beach, CA, 1999, Vol 3669, p 171.

64 C Park, Z Ounaies, J Su, J.G Smith, Jr., and J.S.

Harrison, Mater Res Soc Proc.: Electroactive Polym., 1999,

Vol 600.

65 Y Murata, K Tsunashima, and N Koizumi, 8th Int Symp.

Electrets (ISE 8), 1994, pp 709–714.

66 IEEE Standard on Piezoelectricity, (IEEE Standard

176-1987); Institute of Electrical and Electronic Engineers, NY,

1989.

67 S Sherritt and Y Bar-Cohen, in Electroactive Polymer (EAP)

Actuators as Artificial Muscles: Reality, Potential and

Chal-lenges, Y Bar-Cohen, ed., SPIE Press, Bellingham, WA, 2001,

p 405.

68 J van Turnhout, Polym J 2: 173–191 (1971).

69 J.T Dawley, G Teowee, B.J.J Zelinski, and D.R Uhlmann,

Piezoelectric Characterization of Bulk and Thin Film

Ferro-electric Materials Using Fiber Optics MTI Instruments Inc.,

73 K.E Wise, in Electroactive Polymer (EAP) Actuators as

Artifi-cial Muscles: Reality, Potential and Challenges, Y Bar-Cohen,

ed., SPIE Press, Bellingham, WA, 2001, pp 267–284.

74 J.A Young, B.L Farmer, J.A Hinkley, Polymer 40: 2787 (1999).

75 K.E Wise, 3rd SIAM Conf Math Aspects Mater Sci.,

Philadelphia, PA, May 21–24, 2000.

76 T Koda, K Shibasaki, and S Ikeda, Comp Theor Polym Sci.

10: 335–343 (2000).

77 Y Bar-Cohen, in Electroactive Polymer (EAP) Actuators

as Artificial Muscles: Reality, Potential and Challenges,

Y Bar-Cohen, ed., SPIE Press, Bellingham, WA, 2001,

p 615.

78 D DeRossi and P Dario, in Medical Applications of tric Polymers, P.M Galletti, D.E DeRossi, and A.S DeReggi,

Piezoelec-eds., Gordon and Breach, 1988, p 83.

79 G.M Garner, in The Applications of Ferroelectric Polymers,

T Wang, J Herbert, and A Glass, eds., Blackie, London, UK,

1988, p 190.

80 T.R Meeker, in The Applications of Ferroelectric Polymers,

T Wang, J Herbert, and A Glass, eds., Blackie, London, UK,

1988, p 305.

81 E Yamaka, in The Applications of Ferroelectric Polymers,

T Wang, J Herbert, and A Glass, eds., Blackie, London, UK,

1988, p 329.

POWER INDUSTRY APPLICATIONS

J STRINGER EPRI Palo Alto, CA

R H RICHMAN Daedalus Associates Mountain View, CA

INTRODUCTION

Many industrial and commercial sectors are made up

of separate, relatively small components: aerospace andground transport, for example In such applications, trulysmart materials systems (SMS) are defined as those thatrespond autonomously to changes in their operating con-ditions That is, they detect the onset of “illness” and takesteps to effect a cure In contrast, the electric power indus-try is a vast, interconnected enterprise, not a collection ofindividual entities It is a far-flung grid, fed and drainedcontinuously at variable rates SMS must be consistentwith this character: owing to the extent of the grid and

the way it operates, just detecting illness and determining its location constitutes smartness Moreover, users of elec-

tricity expect reliability everywhere the grid reaches, so

avoiding failures (outages) is paramount Thus, distributed

SMS are often required to achieve reliability Some othersectors (e.g., highways) also will rely on distributed SMS,but in electricity systems the SMS must function undermore hostile conditions of temperature, pressure, aggres-sive chemicals, and especially, intense electric and/or mag-netic fields Early detection of deviations or trouble is vital

In a sense, successful performances of SMS in the powerindustry equates to buying time for a rational response.Whatever else they might do by way of autonomous re-sponses (derating a unit, eliminating an ailing componentfrom a redundant set, etc.), the crucial action for SMS inelectricity systems is to notify a central authority about an(impending) illness and where that illness is

Although investment in SMS research by the power dustry has been substantial, utilization of SMS within theindustry is still in the early stages There is little doubt,however, that the influence of SMS will be profound in thefuture

in-Table 1 Electric Power Industry Segments

Median Number of

% of U.S Customers per

888 (distribution only)

59 (generation and transmission)

(supplier to other segments)

OVERVIEW OF THE ELECTRIC POWER INDUSTRY

Size, Composition, and Economic Pressures

There are four sectors in the electric power industry:

investor-owned utilities, rural electric cooperatives, public

power systems, and federal power agencies The number

of organizations in each category, percentage of U.S

cus-tomers served by each segment, and the median number of

customers served by each are summarized in Table 1 Total

installed capacity is on the order of 750 GW (1), with

ap-proximately 627,400 miles of transmission lines of 11 kV or

more (2), and 3,600,000 miles of distribution lines (3)

An-nual expenditures for generation and for transmission and

distribution are approximately equal at about $12 billion

each

In the past, utilities have been highly regulated, with

specified rates of return, protected service territories, and

the obligation to provide power to any customer within a

defined service territory Over the last few years, protected

markets and assured customers have come to an end, and

nonutility generators (NUGs) have appeared on the scene

It is estimated that there are now over 3000 NUGs in the

United States and that 75% of all new capacity since 1980

has been supplied by NUGs Competitive pressure is

per-vasive; lowering costs, improving and expanding customer

services, diversifying into other businesses, and striving

for better utilization of corporate assets have become the

primary drivers

Problem Costs and Solution Benefits

A principal role for smart materials in the power industry

would be to lower costs and improve system reliability An

indication of the scale of the challenge is provided by the

three rules of thumb that follow

1 A typical coal-fired power plant is approximately

35% efficient For a 500 MW generating unit, loaded

at 70% plant factor, a 0.1% efficiency improvementwould reduce annual costs by about $150,000 at atypical busbar electricity cost of $0.017/kWh For asix-unit plant, annual savings would therefore beabout $900,000

2 Most electricity is generated by large units and there

is a high premium for keeping those units on line

When high-merit units are out of service, ment power must often be provided by less-efficient(more costly) sources Unscheduled outages caused

replace-by equipment failure or significant unit tion are particularly costly Outage lengths depend

degrada-on availability of spares, difficulty of access, degree

of damage, and so on, and can range from hours tomonths Replacement power cost for a one-day out-age of a 500 MW unit is on the order of $240,000(at $0.02/kWh) Emergency (unscheduled) power pur-chases can cost even more; $0.03–0.04/kWh is not un-usual

3 Most generation and power delivery equipment isoperated well below theoretical limits so as to pro-vide reasonable factors of safety Furthermore, whendeterioration of structural or functional integrity issuspected, the afflicted equipment is derated, that

is, operated at a lower level of performance ing a 500 MW unit by 5% (25 MW) is equivalent to

Derat-$12,000 per day or $4,380,000 per year of lost enue (at $0.02/kWh) A transmission line that must

rev-be operated rev-below its normal thermal limit can duce revenues by $1,750,000 per year

re-The Power Industry Environment

Although combustion turbines, often combined with recovery boilers, are increasingly popular for new instal-lations, the generation of power from thermal sources isstill preponderantly by steam turbines Hence, a primaryactivity for many utilities is the raising of steam Fossil-fueled boilers in large central stations produce as much

heat-as 1260 kg of steam per second (10,000,000 lb of steamper hour) at pressures ranging from 13.8 to 24.8 MPa(2000 to 3600 psig) and temperatures up to 580◦C(1100◦F) Nuclear-fueled, pressurized-water reactors oper-ate at coolant pressures of 15.2 to 15.9 MPa (2200–2300psig) and a temperature of about 316◦C (600◦F) to producesteam at 6.2 to 7.3 MPa (900–1050 psig) In other words,steam plants are characterized by hostile environments:high temperatures, high pressures, and aggressive fluids.The damage mechanisms that ensue when plant compo-nents are exposed to these conditions are summarized inTable 2

Steam plant environments are demanding, but most ofthem are not unique; high temperatures and pressures,

Table 2 Damage Mechanisms in Large Components of Fossil-Fueled Generating Plants

or System Fatigue (Note 1) Corrosion Creep Fatigue Erosion (Note 2) (Note 3)

steam, and lower Drums

reheat piping

steam chests, valve casings

intermediate-pressure (IP) turbine rotors

turbine rotors

rotating and stationary

exchangers

Notes: 1 Environmentally assisted fatigue includes mechanisms such as corrosion fatigue and stress corrosion cracking.

2 Wear includes abrasion, galling, rolling contact, fretting, etc.

3 Embrittlement includes temper embrittlement, graphitization, and hydrogen embrittlement.

and reactive process streams, are also found, for instance,

in the chemical processing and metal production

indus-tries In contrast, the environments associated with

elec-tricity generation and delivery are special in that they

involve strong electromagnetic fields and large electric

cur-rents As a consequence, the health of insulation materials,

with their extreme sensitivity to relatively low

temper-atures (100–180◦C, or 212–356◦F) and mild mechanical

forces, dominates concerns about reliability and failure

avoidance Of course, there are also the usual threats to

integrity, such as cyclic stresses, corrosion, and wear

Dam-age mechanisms operative in this sector of the power

in-dustry are given in Table 3

Table 3 Damage Mechanisms in Electrical and Hydroelectric Apparatus

Component

or System

Degradation of Insulation Thermal Mechanical Chemical Electrical Aging

Mechanical

mentally Mechanical Trans- Assisted Corro- Over- fer by tation Fatigue sion load Arcing Erosion Wear Generators

so that early warning of impending dysfunction becomeseven more critical In addition, there are implicationsabout longevity and redundancy requirements for SMS,

since SMS embedded in the insulation of overhead

con-ductors or in concrete structures, for example, cannot be

replaced

Constraints on SMS Development for the Power Industry

All of the foregoing environmental factors, especially the

extended nature of the electricity grid and the

perva-sive presence of strong electromagnetic fields, impose

limitations on SMS solutions to power industry

chal-lenges In the first place, SMS from other industries will

not, in general, be directly transferable in their

origi-nal form either because they have not been designed

for the specific tasks of the contemplated application

or because electrically based SMS cannot function

prop-erly in intense electromagnetic fields Second, the spatial

extent of electricity systems dictates that the most

ur-gent job for SMS is to detect damage or deviant

condi-tions and report the location to a central authority

Au-tonomous actions by SMS are less important The SMS

constituent that matches this requirement is smart

sen-sors In fact, smart sensors are the aspect of the SMS that

has been most thoroughly investigated for utility

applica-tions Here, transference from another industry was vital:

without the R&D base built by the communications

indus-try, the technology of fiber-optic sensors might not have

matured

Undoubtedly, a wide variety of power industry problems

will someday be addressed by truly smart, autonomous,

fully integrated SMS At this juncture, however, the main

thrust for SMS in utility systems has been aimed at

de-veloping and applying smart sensors and, to less extent,

smart sensor-actuators

SMART SENSORS

Requirements and Scope

Need for Better Measurements In recent years, advances

in electronics (signal processing and conditioning) have

re-sulted in improved instrumentation and controls for power

plants Many utilities have replaced analog and pneumatic

controls with distributed digital control systems (DCSs) in

order to reduce operating costs and to increase the

capa-bility to respond to system dispatch A DCS can regulate

processes with less than 0.25% uncertainty, compared with

2% to 3% uncertainty for analog controls However, the

conventional sensors used in power plants cannot, without

frequent calibration, provide the levels of accuracy

appro-priate for sophisticated DCSs Thus, smart sensors offer

the means to increase measurement accuracy and thereby

realize the benefits of DCSs

Opportunities for better measurements are

numer-ous in every sector of the power industry For example,

Sachdeva (4) lists 1393 essential sensors in a typical

500 MW coal-fired power plant Of these, approximately

1150 are devoted to sensing temperature and pressure In

forward-looking utilities that are developing

methodolo-gies for early diagnosis of problems, monitoring of critical

components has resulted in 4000 sensors in some

gener-ating units (5) Likewise, an NSF workshop (6) identified

26 utility problems associated with transmission and tribution that could be alleviated with improved sensors;affected components included lattice steel towers, woodenpoles, overhead conductors and insulators, undergroundtransmission and distribution facilities, and substations

dis-Of course, not all of these sensors need to be smart; formany of them the smartness can be in central control Still,

a significant portion must be smart in order to overcomethe effects of difficult environments

Performance in the Environment The utility industry can

be thought of as the agency for three sequential tions: (1) converting thermal energy to mechanical energy,(2) converting mechanical energy to electrical energy, and(3) electricity delivery The environments characteristic ofeach sector are summarized in Table 4 For the purposes

opera-of this article the second and third functions are similarenough in their environments that they are discussed to-gether as electricity systems in what follows

The first prerequisite for SMS in the power industry

is that they be able to perform reliably in the ment That is, they must produce accurate signals undernormal as well as perturbed conditions for prolonged pe-riods of time Again, their priority is to inform a centralauthority that there is a problem or that damage has oc-curred, and where Autonomous action can also be part

environ-of the SMS mandate; electricity systems are quite dant, for example, so a smart sensor could initiate close-down of a redundant component that is about to fail But

redun-a report of illness is the primredun-ary responsibility As will beseen, most smart sensors intended for power-industry ap-plications are based on fiber optics because of their immu-nity to electromagnetic interference (including lightningstrikes and ground faults), as well as their small size andlight weight

Thermal Plant Environments Pressure Sensing Many pressure transmitters now in-

stalled in power plants rely on fill fluids to separate cess fluids from the gauge mechanisms These devices aresubject to failures when fill fluids leak, failures that aredifficult to detect In addition, all conventional pressuresensors drift with time, a condition that necessitates dis-proportionately large efforts to restore accuracy and ver-ify operability These concerns have prompted an assess-ment of modern sensor technologies to select candidates foradaptation to use in power plants (7) Among the findings

pro-is the conclusion that fiber-optic pressure sensors could

provide more extensive sampling of process pressures, animproved mode of signal transmission and processing, andfreedom from electromagnetic and radio-frequency inter-ference in nuclear and fossil-fueled steam plants

As a step toward implementing improvements in sure sensing, a fiber-optic transducer based on the micro-bend attenuation of light transmitted through an opticalfiber (Fig 1) has been developed (8) Since the diaphragmdeflects linearly with pressure, process pressure is mea-sured by the diminution of light transmission through analuminum-coated fiber with core/clad/coating dimensions

pres-of 150/180/210 µm Performance of the transducer was

Table 4 Classification of Environments in the Power Industry

Convert thermal to rSteam plant rHigh temperaturesmechanical energy rCombustion turbines rHigh pressures

rAggressive fluids

rRelatively close tocentral control Convert mechanical to rGenerators rNear-ambient

rHigh electricaland magnetic fields

rHigh mechanical forces

rRelatively close

to central control Electricity delivery rTransformers rNear-ambient

rOverhead transmission pressures

rUnderground cable and magnetic fields

rHigh electrical currentsthrough contacts

rOften far from centralcontrol

characterized in laboratory and field trials at pressures to

22.75 MPa (3300 psig) and temperatures to 438◦C (820◦F),

with measurement error calculated to be 1.2% of full

scale This smart sensor, developed for coal liquefaction

service by the U.S Department of Energy, would be

Diaphragm

Input

Opticalfiber

Pressure

Sensor detail

Output

DiaphragmWeldSpacerMicrobend sensor

CylinderBolt

WeldPackingseal

Steeltubing

Opticalfibers

Flangepressureseal

30 mm

42 mm

Figure 1 Diagram of the fiber-optic pressure transducer (8).

suitable for pressure measurements in nuclear steamgenerators

Strain Measurement Strain sensing as an indicator

of structural health in high-temperature components is

Reflectedlight signal

Opticalfiber

Incidentlight

Lamp

Lens

Fizeauanalyzer

2 × 2Coupler

correlationfunction

Cross-Linearphotodiodearray

Figure 2 Schematic of interferometric strain sensor (9).

becoming more important as power-generating equipment

ages Conventional foil and wire strain gauges are not

re-liable at temperatures above 250◦C (482◦F) for long times,

largely owing to the unavailability of adhesives that can

withstand those temperatures Therefore, a fiber-optic

sen-sor to monitor strain in boilers, headers, steam pipes, and

other high-temperature structures has been developed for

operation at temperatures up to 650◦C (1202◦F)

Based on the Fabry-Perot interferometric technique,

the system is shown schematically in Fig 2 A

broad-band optical beam is conducted into a quartz tube

contain-ing two fibers, each with a partially reflectcontain-ing mirror at

the end A small air gap or resonance cavity between the

two mirrors forms an extrinsic Fabry-Perot

interferome-ter (Fig 3) Beams reflected from the two mirrors ininterferome-ter-

inter-fere, travel back toward the detector, and enter an optical

wedge (Fizeau analyzer) Reflected light from the

Fabry-Perot sensor is transmitted maximally where the

opti-cal path length matches the dimension of the Fabry-Perot

cavity Thus, when strain changes the cavity length, there

is a corresponding shift in the intensity maximum

trans-mitted through the optical wedge A linear photodiode

ar-ray at the back of the optical wedge detects the

transmit-ted beam Because this is a frequency-modulatransmit-ted sensor, it

is insensitive to light attenuation; the signal reaching the

photodiode array need only contain enough information for

decoding (10)

Coating

Epoxy

reflectingmirrors Air gap

Partially-(fabry-perotcavity)

Fusedjoints

Quartztube

Transmitter/receiver fiber Target fiber

Figure 3 Schematic of Fabry-Perot interferometer (9).

This small, lightweight device has been field tested intwo applications at a power station: a spot welded on amain steam line operating at 565◦C (1049◦F), and in athermowell inside a reheat steam line at 538◦C (1000◦F).Feasibility of on-line monitoring of strain in structures athigh temperatures was clearly demonstrated

Measurements with Fiber Bragg Gratings A potentially

more versatile sensor technology is embodied bywavelength-modulated fiber Bragg gratings (FBG), whichare created in low-cost, commercially available fibers EachFBG reflects a characteristic wavelength that changes asthe grating periodicity changes with temperature and/orstrain Research sponsored by the power industry hasconcluded that, in principle, FBGs can serve as generictransducer elements to measure temperature, pressure,strain, vibration, acoustical disturbances, electrical andmagnetic field strengths, and the concentrations of certainchemical species (11)

A primary issue in developing FBGs for utility use inhigh-temperature environments is their stability Expo-sures to high temperatures and temperature cycling hasdemonstrated that FBGs are quite robust: FBGs enduredlong-time cycling between ambient and 427◦C (800◦F)without degradation Use temperature can be extended

to 650◦C (1202◦F) only if very low loads (strains) are posed on the fiber (11) FBGs for measuring temperature,pressure, and strain are now in prototype development Al-though several approaches to chemical sensing with FBGshave been explored in laboratory studies, they are not yetready for prototyping

im-Sensors for Combustion Processes Accurate and reliable

sensors for very high temperature environments wouldhave multiple benefits for thermal plant: avoidance of dam-age to heat-transfer surfaces, combustion controls for in-dividual burners (not possible at present), reduction ofnoxious emissions, and structural health monitoring ofcritical components, to name a few The chief barrier todeveloping smart systems based on fiber optics for combus-tion environments is the temperature constraint on silicafibers Fiber function is limited by the temperatures at

which (1) the dopants in conventional silica begin to

diffuse rapidly enough to affect adversely both signal

atten-uation and wave-guiding properties, and (2) the silica

soft-ens These temperatures are approximately 800 to 900◦C

(1472–1652◦F)

Since crystalline sapphire has reasonable optical

prop-agation properties, melts above 2000◦C (3632◦F), and is

corrosion resistant, it is the basis for research aimed at

developing fiber-optic sensors capable of operating above

1500◦C (2732◦F) Schemes under investigation include

sapphire-fiber-based Fabry-Perot interferometers, FBGs,

and intensity-based sensors, as well as ways to modify fiber

coatings and claddings that make sensors able to

func-tion in combusfunc-tion gases and superheated water

environ-ments (9)

One especially important challenge for

high-temperature sensing is management of emissions from

combustion In control systems for nitrogen oxides (NOx),

ammonia (NH3) or urea (NH2–CO–NH2) is injected into

combustion gas to react with NOx and produce molecular

nitrogen and water Postcombustion NOx reduction must

avoid significant NH3in the exhaust, both to comply with

emission regulations and to keep from fouling downstream

components A feedback control system is needed, but

no reliable, real-time NOx or NH3 sensors have been

available

An investigation is exploring measurement of nitric

oxide (NO) levels on the basis of radiative emissions from

single molecular transitions that are well separated from

emission features associated with other constituents in

the flow The system consists of feedback-stabilized,

scan-ning Fabry-Perot interferometers linked with

thermoelec-trically cooled wavelength detectors A digital system

con-trols cavity lengths for wavelength scanning One detector

monitors NO upstream of the injection plane and a second

monitors NH3downstream Signals from both are fed into

the injection system controller, which then determines in

real time the quantity of NH3 or urea to be injected for

optimal NO removal Tests of the prototype system are

un-derway at a utility power station (9)

pH Measurement Corrosion behavior of steam-plant

materials is determined to a large extent by the pH and

electrochemical potential of the circulating water Since the

physicochemical properties of water are highly sensitive to

temperature, there is strong incentive to develop on-line

sensors for pH that can be used at system temperatures,

rather than relying on analytical extrapolations from grab

samples cooled to ambient, as is the current practice

How-ever, attempts to develop on-line sensors for pH at

ele-vated temperatures must confront two problems:

degra-dation of the sensor materials by hot water; and, in the

high-purity water characteristic of cooling loops,

interac-tions of the sensor with the water can affect the pH being

measured

A sensor was developed by incorporating a pH-sensitive

organic dye (8-hydroxypyrene-1,3,6-trisulphonic acid) in

a polyacrylamide polymer at the end of an optical fiber

(Fig 4) By choosing a dye with two absorption peaks, the

sensor indicates the pH as the ratio of the two peaks;

long-term leaching of the dye does not compromise the

Optical fiber

HPTS polymer

Glass capillaryO-ring

Nylon mesh

Figure 4 Schematic of the fiber-optic pH sensor (12).

measurements pH determinations were made successfullyover the course of a one-year immersion at 38◦C (100◦F),and pH changes associated with addition of 1 ppm morpho-line were measured consistently (12)

Water at higher temperatures exacerbates stabilityproblems for fiber-optic reflectors A 250 nm thick, mul-tilayer (titanium–platinum–gold) coating for sapphire wasfound to be stable to 180◦C (356◦F) In conjunction with

an azo chromophore (dinitrophenyl-azo-napthol) as anindicator, pH was measured reproducibly in the labora-tory for 160 hours at 50◦C (122◦F), which was the stabil-ity limit of the cellulose acetate waveguide (13) In anycase, azo-based indicators are stable only to about 100◦C(212◦F) While none of the pH-sensing systems devised sofar are acceptable for use in high-temperature water, pre-liminary experiments have shown that a copper phthalo-cyanine chromophore is stable to at least 200◦C (392◦F)

It is expected that the next step, an incremental ment, will be a fiber-optic pH probe that is usable to 150◦C(302◦F)

improve-Electricity Systems Environments Distributed Fiber-Optic Temperature Sensor (DFOTS) In-

sulator temperature is a key factor in the safe and reliableoperation of motor and generator windings, transformers,circuit breakers, underground cables, and overhead trans-mission lines Although temporary overload conditions donot normally cause thermal damage to conductors, higherthan normal temperatures do have a cumulative effect ofshortening insulation life On-line methods to locate andmeasure “hot spots” have traditionally not been available,largely owing to the fact that most transducers are elec-trically based In other words, conventional sensing de-vices are usually incompatible with the environment of

an electrical system Furthermore, local temperature surements (e.g., 0.1 m [4 in] long or less) need to be madethroughout an electrical system that may be hundreds ofmeters (feet) long

Distance

Input

Hot spot

Sensing fiber(attached or embedded)Terminal

Motors, Generators, and Circuit Breakers To address this

challenge, the DFOTS system has been developed for

de-tecting hot spots in low-temperature (≤150◦C [302◦F]),

high-voltage environments It is based on optical time

do-main reflectometry (OTDR), which was devised by the

telecommunications industry for fault location in

fiber-optic telephone lines A light pulse transmitted by an

optical fiber is gradually attenuated by absorption and

Rayleigh scattering The scattered light returns in a

direction opposite to that of the injected light pulse To

make a DFOTS, the fiber must be modified along its

length such that the local backscattering changes as a

function of the local temperature change This was

accom-plished by coating the fiber with a UV-curable polymer that

changes refractive index reversibly with temperature (14)

A change in the local intensity of backscattered light serves

as a sensitive indicator of temperature, and the elapsed

time of the returned pulse indicates the location of the

tem-perature change, as shown in Fig 5

The DFOTS system has monitored winding

tempera-tures in rotor and stator windings of motors and generators

in trials at power plants (15) and in switchgear (14) It is

accurate to within 5◦C (9◦F) over the range 0 to 150◦C (32 to

302◦F) A 100 ps laser pulse is capable of resolving hot spots

only 2 cm (0.8 in) long over a fiber length of 40 m (131 ft)

In retrofit applications, the DFOTS fiber can be strung

be-tween windings bars and slot wedges; in new windings, the

sensor can be incorporated in the high-voltage groundwall

insulation in direct contact with the conductors In either

application, impending generator problems could be

diag-nosed rapidly and major winding failures prevented

Im-proved remaining-life assessments would also ensue from

knowing real thermal histories of insulation

Underground Cables A similar DFOTS has been devised

for temperature monitoring of underground power cables

The power transfer capability of a buried cable is strongly

affected by thermal conditions along the length of the

cir-cuit: burial depth, ambient earth temperature, soil thermal

resistivity, and the like Power transfer could be optimized

if actual temperatures along the cable are known in real

time

When a sharply pulsed laser beam (1050 nm) is

in-jected into a standard multimode fiber, very weak,

ther-mally dependent, molecular vibrations produce reflections

along the fiber length These reflections, known as Ramanbackscattering (in contrast to Rayleigh backscattering inthe previous example), can be detected by OTDR andthe signals processed to determine temperature along thelength of the optical fiber A commercial system (York Sen-sors, Ltd., DTS-80) based on Raman scattering was in-stalled in a duct containing 66 kV power cables Ampacityanalysis of the measurements enabled the utility to upratethe cable circuit by 8%, without exceeding the 90◦C (194◦F)maximum design temperature for the cables (16)

Strain in End Windings The stator winding of an AC

gen-erator is comprised of conductors (coil sides or bars) thatare housed in slots in the stator core At the end regionsadjoining the active length of the machine, pairs of conduc-tors are linked by end connections to form coils The com-plete set of end connections at each end constitutes an endwinding End windings cannot be supported as securely asthe conductors, which are in slots; if blocks and lashingsbecome loosened by the forces of starting and stopping orsystem faults, the end windings can vibrate and progres-sively damage insulation Undetected end-winding vibra-tion can lead to a forced outage in a relatively short time.Standard strain gauges (metal foil or wire) cannot function

in the strong electromagnetic environment of an operatinggenerator

A fiber-optic sensor based on microbending has been veloped to measure strain in end windings during opera-tion (9) The sensor attaches directly to an end turn andconverts deflection to strain Output of the device is linearover the range ±1000 µm/m (1000 microstrain) and 5 to

de-100 Hz, with a resolution of±5 µm/m

Monitoring Transformers Substation transformers are

large, oil-filled devices and are among the most expensivecomponents in an electric-power network The cost of a fail-ure, or an outage to repair a unit, can exceed the originalcost of the transformer within five days if the cost of re-placement power from a less-efficient station is accounted.Deterioration of transformer oil results from excessivetemperature, aging, and electrical discharges through theoil Oil has an effective lifetime of about one millionhours at 90◦C (194◦F) but only about 100 hours at 180◦C(356◦F)

Winding Temperatures In principle, temperatures in

transformer windings could be measured by DFOTS based

on Rayleigh or Raman backscattering However, it has beenfound that transformer oil penetrates the fiber jacket and

is absorbed by the polymer cladding of Rayleigh-basedsensors, and all-silica Raman-based sensors have inad-equate spatial resolution (about 5 m [16.4 ft]) at theirpresent stage of development Point monitoring of wind-ing temperatures in real time is the best that can bedone right now by measuring the temperature-dependentfluorescent decay time of a photoluminescent sensor mate-rial (manganese-activated magnesium fluorogermanate).Pulses of blue light power the phosphors; fluorescencereturning in the all-silica fiber is detected and inter-preted in terms of sensor temperature (17) Decay time

is a well-characterized, intensity-dependent property of