`
`[etal and Ceramic
`
`-iomaterials
`
`
`VolumeI
`Strength and Surface
`
`Editors
`
`Paul Ducheyne, Ph.D.
`Associate Professor of
`Biomedica! Engineering
`Associate Professor of
`Orthopaedic Surgery Research
`University of Pennsylvania
`Philadelphia, Pennsylvania
`
`,
`
`Garth W. Hastings, Ph.D., D.Sc., C.Chem., F.R.S.C.,
`
`Acting Head
`Biomedical Engineering Unit
`North Staffordshire Polytechnic
`Honorary Scientific Officer
`North Staffordshire Health Authority
`Medica! Institute
`Hartshill, Stoke-on-Trent
`England
`
`CRC Series in Structure-Property Relationship of Biomaterials
`Series Editors-in-Chief
`Garth W. Hastings and Paul Ducheyne
`
`CRC Press, Inc.
`Boca Raton, Florida
`
`
`
`IRONBURG EX2020, Page1
`
`IRONBURG EX2020, Page 1
`
`

`

`
`
`Ms
`ees
`aL
`
`ii:
`
`Library of Congress Cataloging in Publication
`Main entry undertitle:
`
`
`
`
`
`
`
`
`
`(CRC series in structure-propery relationship of biomaterials)
`Bibliography: p.
`
`Includes index.
`
`Contents: v. 1. Structure — ¥. 2. Strength and
`
`surface celated behavior,
`
`|. Ceramics in medicing, 2. Metals in surgery.
`
`I. Ducheyne, Paul.
`JI. Hastings, Gari W.
`TIL. Series.
`
`
`R&S7.CAMA?=1984 610".28 83-}5018
`
`
`ISBN 0-8493-6261-X (v.
`LD
`
`ISBN 9-8493-6262-8 (v. 2}
`
`
`Metal and ceramié biomaterials.
`
`
`
`
`
`
`
`
`
`
`
`‘This book represents information obtained from authentic and highly regarded sources. Reprinted material is
`quoted with permission, and sources are indicated. A wide varicly of references are listed. Every reasonable ¢ffort
`has been madeto pive reliable data and information, but the author and the publisher cannot, assume responsibility
`for the validity of al? materials or for the consequences of their use,
`
`All rigtts reserved. This book, or any parts thereof, may not be reproduced in any form without written consent
`from the publisher,
`
`Direct afl inquiries to CRC Press, [nc., 2000 Corporate Bivd., N.W., Boca Raton, Florida, 33431.
`
`® 1984 by CRC Press,Inc.
`
`International Standard Book Number 0-8493-6261-X (v. 1}
`International Standard Book Number 0-8493-6262-8 (v. 2)
`
`Library of Congress Card Nomber 83-5018
`Printed in the United States
`
`
`IRONBURG EX2020, Page 2
`
`
`
`
`
`
`
`
`
`IRONBURG EX2020, Page 2
`
`

`

`Volume LH: Strength and Surface
`
`63
`
`Chapter 3
`
`SHAPE MEMORY ALLOYS
`
`R, Kousbroek
`
`TABLE OF CONTENTS
`
`Introduction.......0.0.5vce c cede edad tere Neb b rea cece bent g ener epee eterna es 64
`
`
`
`on
`dD.
`
`{V.
`
`Vv.
`
`
`
`
`
`
`
`
`Historical Background 2.2... 0.00: cee serene cere ere ees cect nese ee ee teen eeeeee 64
`
`
`
`Nature and Mechanism of Shape Memory Alloys. .......---+s--s¢rerrrereeeee es 65
`A.
`Martensitic Transformation .......--00cseee eee e ee ce erent etter ess 65
`B.
`Martensitic Morphology .....-...0.2:0eceeese crc ee eet eer en ne sree etese ss 66
`1.
`Thermally Induced Martensite ....- 016-4000 seer cr ecee reser renee 66
`
`
`2.
`Stress-Induced Martensite....-... 0000.2 eect eerste erent 67
`3,
`Reoriented Martensite....... 2.062 eee eee center n err ies 68
`
`
`Crystallographic Requirements For Shape Memory Alloys ......-...---+5 69 |
`
`
`Shape Memory Effect 0.2... 0002. ..sesreeeeenerreeeste eres snes e ss ses ee 70
`:
`1.
`One-Way Shape Memory Effect ...........:sereeeces eee cterereres 70
`aI
`
`
`2.
`Two-Way Shape Memory Effect........-.--0+--.:serreerscteeess 72
`tf
`
`
`FB.——-Pseuddo-Elasticity ....22....0scceceneeeeee sree etn ee ee nee ea neers arenes sess B i
`
`
`i
`1.
`Pscudo-Elasticity by Transformation .......-..0--..s-r see eeeeenes 73
`
`
`|
`2.
`Pseudo-Elasticity by Reorientation ........)....reeseeree teeters .. 74
`
`
`4
`3.
`Pseudo-Elasticity by Transformation and Reorientation .........+- 74
`
`
`Overview of the Coherence between Structure and Memory Eftects...... 75
`F,
`
`
`Structure Related Méchanical Properties .....-......---er seers tees terres entices 15
`i
`A. Young's Modulus and VieldeStress.....20200ce cece cae eee nents 73
`
`
`4
`B.
`Recovery SUCSS ... 6. cseiee eee eee ee ener re renee ere ees 77
`
`
`Cc.
`Shape Memory Fatigue.....-..26se rece eeerer esse et ere eee rene nese ee ees 78
`
`
`:|
`Shape Memory Alloys for Medicat Application... ....... 0. .-ee rene sere s ee eere ines 79
`44
`A.
`Biocompatibitity.....0..--0c002ce ee ceeeee seen ters ceeer etre seers tnen aes 80
`
`
`B.
`Design Considerations......20.ccccsseeceree cece ceterteeretree seer ct: BO
`4
`
`
`
`
`Cc.
`Potential Applications ..... 0.60. ..ceeretr renner eee renee reese ences e ess $1
`1.
`Dentistry and Orthodontics. .....--.-----6ss see ere seen ee ener 81
`2.
`Orthopedics 0.0.2... 50: cee ee erence renters n terete ees 83
`3.
`Rehabilitation .....c.0.. erect ee eee e eee eet nearness 84
`4.
`Heart and Vascular Surgery.......2.00ceeeeeeree cere t eee eee 85
`Concluding Remarks.........00c0cscrsveeeeeerenegecsen eager ener ste 87
`WL.
`Acknowledgrients .... 0.00... 000see cet e reece entree tere geneEE 87
`References. .......6+ eee en yee E EE eR EE REE EEE EEE SEE E ETT ETE TEESE 87
`
`
`
`rig
`
`:|
`
`|
`|
`
`
`
`ul
`
`;
`
`
`
`IRONBURG EXx2020, Page 3
`
`IRONBURG EX2020, Page 3
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`64
`
`Metal and Ceramic Biomaterials
`
`1. INTRODUCTION
`
`In the past few years the shape memory alloys have attracted the attention of metailurgists
`and design engineers because of a number of remarkable properties which open a revolu-
`tionary way ofdesigning on the basis of entirely new principles compared with conventional
`alloys. The moststriking featutes of this family of alloys are the shape memory effect,
`the
`pseudo-elasticity, and the very high damping capacity. A short definition of these effects
`will be given here for convenience, but a more extensive description of the shape memory
`effect and the pseudo-elasticity, which are important in the context of this chapter, will
`foilow later.
`Shape memory effect —- This is the phenomenon by which, after an apparent plastic
`deformation, a metal alloy tipon heating starts to remember its original shape at a certain
`temperature and returns to its deformed shape upen costing.
`Pseudo-elasticity — The effect by which a material recovers the induced “‘plastic’’ strain
`upon unloading is known as the pseudo-elasticity. Thé amount of this reversible strain is
`much greatcr than the classical, elastic strain. In constrast to the shape memory effect, the
`temperature remains constant.
`the damping
`High damping capacity — As a result of a change in internal structure,
`capacity of these alloys can be varied considerably as a function of temperature, resulting
`in a very high damping capacity in certain temperature ranges.
`These three features are associated with a martensitic transformation and this association
`limits the family of alloys exhibiting these special effects. In principle, afl the alloys which
`exhibit a martensitic transformation are potential shape memoryalloys, bud experience shows
`that the effects only appear significantly in alloys having a reversible martensitic transfor-
`mation, e.g., nickel- and copper-based alloys (¢.g., Ti-Ni and Cu-Zn-Al).
`Although the first bazic information concerning the shape memory alloys was already
`observed about 40 years ago, it was in the |96Qs that the usefulness was recognized. Since
`that time many potential applications of the shape memoryalloys have been suggested.
`In this review the potential uses of the shape memory alloys in medical applications will
`be highlighted (Section V). However,first, in order to understand fully the working principles
`of the biomedical devices, the basic metallurgical mechanisms of the shape memory alloys
`and the related effects will be discussed (Sections II] and EV).
`
`iH. HISTORICAL BACKGROUND
`
`The first observed shape memory phenomenonis the pseudo-elasticity. In 1932 Olander
`observed this ina Au-Cd alloy and called it “‘rubber-like’’ behavior,'
`In the 1950s this
`phenomenon was also recognized in other alloys, e.g., In-Tl, Cu-Zn, and Cu-Al-Ni. Because
`of the great amount of reversible strain, this effect is also called *‘superelasticity’’. The
`Maximum amount of reversible strain has been observed in a Cu-Al-Ni single crystal with
`a recoverable elastic strain of 24%.?
`The first steps on the discovery of the shape memory effect were made in 1938 by Greninger
`and Mooradian,* observing the formation and disappearance of martensite with falling and
`tising temperature ina Cu-Zn alloy. However, this basic phenomenon of the memory effects,
`the thermoelastic behavior of the martensite phase, was first extensively studied 10 years
`later by Kurdjumov and Khandros.7
`Since the first observation of the shape memory effect in Au-Cd in 1951,> the effect has
`also been reported in other alloys, e.g., Cu-Zn, In-T], Cu-Al-Ni, Ag-Cd, Ag-Zn, Cu-Al,
`Fe-Pt, Nb-Ti, and Ni-Ad, but the great breakthrough came in the early 1960s, when Buehler
`et al.” of the U.S. Naval Ordnance Laboratory (now called the U.S. Naval Surface Weapons
`Center) discovered the shape memory effect in an equiatomic alloy of nickel and titanium,
`
`
`
`IRONBURG EX2020, Page 4
`
`

`

` Volnme Hy: Strength ana Surface
`
`
`
`65
`
`since then popularized under the name Nitino! (Nickel-Titanium Naval Ordnance Labora-
`tory). With this alley complete recovery of a maximum strain of 8% can be achicved by
`the shape memory effect, associated with a considerable foree, which can perform wark.?*
`As Nitinol is difficult and expensive to mauulacture and Fabricaie, the attention of metal-
`lurgists reverted lo one ofthe first shape memoryalloys Cu-Zn (brass). With the discovery
`that addition ef small amounts of aluminium to brass raised its transition temperature con-
`siderably ,” the shape memory effect of this new Cu-Zn-Al alley could now be used in many
`practical applications at or near room temperature as was already possible with Nitinal, The
`great advantage of this Cu-Zn-Al alloy in comparison with Nitinolis thatit is much cheaper
`and much easier to machine and fabricate. Since £969, a major part of the fundamental
`research on Cu-2n-Ai shape memory alloys was done by Delaey et al!"
`Several applications-of the Ti-Ni and Cu-Zn-Al shape memory alloys have been developed,
`e.g., tube fitting systems, self-erectable structures, clamps, greenhouse window openers,
`thermostatic devices, different thermomechanical applications for automobiles, heat-engines.
`and biomedical applications. Several international symposia have been devoted exclusively
`to these alloys. In 1968a first syrnposium concerning the shape memory alloy Nitinal was
`held in the U.S." followed in 1975 by the first international symposium on shape memory
`affects in alloys and applications at Toronto, Canada. Since 1976 shape memory alloys
`and the mechanisms are an ¢ver recurring topic at such conferences as the International
`- Conference on Martensitic Transfermations GCOMAT).
`I. NATURE AND MECHANISM OF SHAPE MEMORY ALLOYS
`The striking features of shape memory alloys are all closely related to the martensitic
`transformation,
`It
`is thus of value to describe the martensitic transformation and related
`
`phenomenafirst. Ii should also be noted that the exact nature and mechanism governing the
`
`
`behavior of shape memory atloys is not identical for all shape memory alloys.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`A. Martensitic Transformation
`Cohen et al.'4 formulated the next definition of martensitic transformation: “A martensitic
`transformation is a lattice-distarlive, virtually diffusionless stryctural change having a dom-
`inant deviatoric component and associated shape change such that strain energy dominates
`the kinetics and morphology during the transformation.” [fdiffusion rues the transformation,
`atoms are changing places in the lattice in an uncoordinated way over long ranges. The
`resulting process can be fully described by the diffusion Jaws of Fick or by the method of
`Matano. However, if as is the case in the martensitic transformation, a diffustonless trans-
`formation occurs, 2 Coordinated movement of large blocks of atoms is decisive for the
`resulting structure. The most well-known martensitic transformation is the phase transition
`responsible for the hardening of carbon steel caused by quenching after annealing at high
`temperature.
`[n this case the austenilic, high-temperature fee-structure changes imto the
`martensitic, bet-structure en cooling.
`From now on, regardless of the crystal structure, we will denote the high-temperature
`phase as the austenitic phase, while the product of the transformation will be called mar-
`iensitic. In the present context, these terms are certainly not limited to ferrous materials.
`Concomitant with the homogeneous lattice deformation, caused by the movement of the
`large blecks of atoms, dominant deviatoric shear displacements can cause an extemal meas-
`urable shape change. The associated strain energy will exert a dominant influence on the
`kinetics and morphology of the transformation. it is in this dominant influence ofthe strain
`energy on the growth characteristics that many nonferrous martensitic systems show the
`shape memory effect, whereas most ferrous martensites do not, because of differences in
`
`IRONBURG EX2020, Page 5
`
` 2tseompRGRREMn
`
`RATT
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`IRONBURG EX2020, Page 5
`
`

`

`
`
`
`
`‘
`
`66
`
`Metal and Ceramic Biomateriais
`
`growth behavior. The effect of these differences in growth behavior on the potential of an
`alloy tu exhibit the shape memory effect by transformation is rélated to the driving force at
`the onset of growth.'? It is generally agreed that the reason for the transformation is the
`difference between the free energies of the austenitic and martensitic structure, i.e., one
`kind of interatomic bonding is energetically favored at lower temperatures, below some
`equilibrium temperature, T.,.
`If the formation and disappearance of the martensitic phase is directly responsive to
`alternation in temperature, i.e., a thermoelastic behavior, the driving force is small (as-in
`ihe nonferrous systems}. while the stress necessary to induce growth is accordingly smail.
`If the driving force at the onset of growth is large due to the absence of a thermoelastic
`equilibrium (the ferrous systems), growth of the martensite proceeds at high rates (of the
`order of the velocity of sound} which leads toa strongly disturbed, defective grain boundary
`whichis subsequentiy essentially immobile. In this last case the shape memory:effect appears
`for efficient application purposes between too broad and therefore unacceptable temperature
`boundaries (260 to 300°C}.
`
`B. Martensitic Merphology
`Martensitic transformations can be induced by changes in temperature as well as by ihe
`application of stress. This can be explained by the following effects: (1) the free enthalpy
`of the austenitic and martensitic phase and so their equilibria depend not only on changes
`in temperature and composition, but also on stress; and (2) the nucleation and growth process
`are associaled with shear strains and these will interact with stresses acting within, or applied
`io, the specimen. These thermodynamic and kinetic effects are strongly dependent on the
`direction of stresses with respect to the lattice orientations. Thus,
`two main groups of
`martensite can be recognized: (t) thermally induced martensite and (2) stress-induced
`martensite.
`
`!. Thermally Induced Martensite
`Thermally induced martensite is mainly characterized byits temperature dependence. It
`forms and grows continuously as the temperature is lowered and shrinks ard vanishes
`continuously as the temperature is raised. In this case the transformation proceeds essentially
`in equilibrium between the chemical driving energy of the transformation and a resistive
`energy whose dominating component is the stored elastic energy. The transformation as-
`sociated with this kind of thermally induced martensite is called a thermoelastic martensitic
`transformation and characterized by its transition lemperature A,, A;, M, and M; (Figure
`1A).
`The growth rate appears to be governed solely by the rate of change in temperature.'5-'6
`However,if the transformation occurs spontancously wheneverthe chemical driving energy
`largely execeds the resistive energy (i.e.,
`the growth rate is independent of the rate of
`temperature change), the resulting martensite is not called thermoelastic anymore, but burst
`Martensite.
`The thermoclastic, as well as the burst martensitic transformation, are frequently either
`partially of fully self-accommodating. The martensite forms either in zig-zag arrays (burst
`martensite), in packets (massive martensite}, in groups, or in bands.
`If the martensitic transformation is sclf-accommodating, the orientation of the growing
`martensite plate with respect to the orientation of its neighboris the one that is energetically
`most stabie in that particular strain field. The strain associated with one variant compensates
`the sitain in the other variants. This requires the accommodation ofalt the local distortions
`involved in the formation of the individual
`lamellae with minimum macroscopic strain
`regardless of crystal direction. Because of the absence of sufficient time to permit any
`relaxation of the resulting local stresses, high densities of dislocations are observed in, self
`accommodating martensite. These dislocations are formed as the result of Jocal strain energy
`
`IRONBURG EX2020, Page 6
`
`IRONBURG EX2020, Page 6
`
`

`

`
`
`
`
`
`TEMPERATURE
`
`
`;
`Fhermoetastic martensitic transformation in function of
`FIGURE ! (A).
`
`
`lemperature, with A, = start transformation on heating; Ay * end trans-
`H
`
`
`formation on heating; M, = slart transformation on cooling; M,; = end
`
`transformation on cooling,all at zero stress; and V/V = fractional volume
`of the martensite product.
`
`—
`
`
`
`My
`
` Volume H: Strength and Surface
`67
`
`al
`he
`ne
`ne
`
`to
`yn
`il.
`Lic
`ne
`
`we
`
`; i i; i
`
`pe
`
`py
`e3
`ss
`ed
`he
`of
`ed
`
`
`TEMPERATURE
`
`
`
`
`Schematica! transformation of bursl martensite on cool-
`FIGURE | (8).
`ing. ‘The initial bursts are expected to be large, decreasing rapidly as the
`
`
`martensitic fractional volume increases.”
`
`
`
`as an alternative of the local breakdown of the lattice and not in response to any specific
`shear stress as is usuaily the case.
`
`
`Therefore these dislocations are immobile and cailed accommodation disiocations, nec-
`essary to accommodate the microscopic transformation strain.'? The maxinium number of
`
`
`possible martensite orientations within one grain depends on the crystal symmetry of the
`
`
`austenitic phase. For a cubic austenitic phase, 24 martensite plate variants can occur.”
`In the absence of external applied stresses and when the volume change ts negligible, the
`
`
`
`
`thermally induced rnartensitic transformation is characterized by random martensite plate
`variants, resulting in a minimum or zero macroscopic shape deformation. However, ifa
`
`
`j
`constant external applied uniaxial stress assists the thermally induced martensitic transfor-
`
`
`|
`are expected togrow or, in the case of self-accommodating formations, certain martensite
`i
`mation thermodynamically, only a limited nurtnber of thermoelastic martensite plate variants
`
`
`
`
`plate variants will become dominant in the different groups. This will lead to an external
`macroscopi¢ shape change.
`
`
`
`if
`rst
`
`ne
`iy
`tes
`ys
`vn
`ny
`lf
`‘BY
`
`2. Stress-lnduced Martensite
`The stress-induced martensite is a mechanical analogue to the thermally induced marten-
`
`i;
`
`j
`|
`
`
`
`site. In this case the transformation proceeds continuously with increasing applied uniaxial
`
`
`IRONBURG EX2020, Page 7
`
`IRONBURG EX2020, Page 7
`
`

`

`-_
`
`68
`
`Metal and Ceramic Biomateriats
`a |
`
`¥
`
`STRESS
`
`
`
`!
`:
`
`which is followed by the immediate and also stress-assisted transformation of this austenite
`
` Stress-induced, thermoelastic martensitic transformation in
`FIGURE 2.
`function ofthe applied stress. with o* *™ = stress at which transformation
`austenite to martensite starts and o “~ 4 = stress at which the reverse
`transformation rnariensite to auslenite starts.
`
`siress and is reversed continuously when the stress is decreaséd (Figure 2), while the tem-
`perature remains constant. Stress-induced martensite can be thermoelastic as well as burst-
`type martensite,
`Again, those martensite plates will preferentiallygrow, which are most favorably oriented
`with regard to the extemally applied uniaxial stress. Therefore, stress-induced martensite
`will have a strongly textured microstructure. The influence of éxternal stresses on the
`martensitic transformation can be expressed by using the temperature My, defined as the
`temperature at which the transformation to martensite can take place under an externally
`applied stress. The maximum temperature M, = (MP)quc depends on the stress conditions,
`specimen orientation or texture, the austenitic yield strength, and, possibly, other factors.
`it cannot be considered as an inherent characteristic of the material itself. '*
`Tf the deformation teniperature, T,, is approached on cooling and if A, < T, < M,,
`martensite formed from the. austenitic phase will disappear on the removal of the external
`stress. However, if M, < T; < A,, the stress-induced martensite variants remain predom-
`inantly thermodynamically stable upon unloading. This includes that under stressed condi-
`tions the M,-temperature is higher than in unstréssed conditions.
`
`3. Reoriented Martensite
`If a fully thermally induced martensite structure is stressed at T < M, the martensite
`structure will be reoriented. This means that with regard to the applied stress certain pret-
`erential martensite variants will grow at the expensé of the less suitably oriented martensite
`variants. The stress necessary to initiate the reorientation decreases with inereasing number
`of transformation cycles, probably due to sweeping out existing defects along the martensite
`plate boundaries. Several theories are in circulation about the mechanism related to the
`reorientation. One proposed mechanism is the elastic twinning and untwinning in crystals,
`ic., a reversible motion of existing twin boundaries resulting in a motion of the martensite
`plate bourdary.'* This implies that the thermaily induced martensite has to be intemally
`twinned,
`Another mechanism for the reorientation of the martensite plates has been proposed by
`Wasilewski. '*#9-2 If the martensite is stressed at T < M, a stress-induced austenite variant,
`transformed out of the martensite, should be a transient intermediate transformation step,
`
`IRONBURG EX2020, Page 8
`
`IRONBURG EX2020, Page 8
`
`

`

`
`
`Volume iH: Strength and Surface
`
`69
`
`M3 63M!
`
`to another martensite variant with another orientation with regard to the original martensite,
`Le,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`This transformation is Jimited between the upper temperature M, and the lower temperature
`A,, corresponding with the minimum temperature at which the transformation M *, B can
`occur, The reverse transformation will occur in the opposite direction upon unleading. The
`existence of the intermediate step-is a source of discussion. Other investigators describe a
`same kind of transformation by reorientation M — M’ withoutthis intermediate step.?'**
`Although the exact theory is not yet clear,
`in crystallagraphic terms the two models are
`equivalent. Only the atomic position before and after the application of the stress can be
`considered, but not the path followed by the atoms.
`Whether the reortented martensite variant remains irreversible upon unloading or not, is
`dependent on the kind of shape memoryalloy. If for a certain amountofstrain the reoriented
`martensite is thermodynamically stable upon unloading, one has the situation required for
`one-way shape memory, induced by reorientation (Section ITI.D.1), while in the reverse the
`requirements are present for pseudo-elasticity by reorientation (Section IL.E.2).
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Cc. Crystallographic Requirements for Shape Memory Alloys
`Originally the shape memory effect was ascribed to stress-assisted compositional changes ,°~4
`but this theory did not stand firm for long. In 1966 the shape memory effect in Ti-Ni was
`for the first time related to a thermoelastic martensite transformation by Zijderveld et a}.*7-**
`This association seemed to be too limited. Because in Ti-Ni, one has observed that 50 to
`80% of the specimen apparently transforms instantaneously. This implies that the growth
`rate is not governed by the rate of change in temperature and therefore does not meet the
`definition of thermoelastic transformation, although Ti-Ni exhibits a very striking memory
`behavior.
`A more general relation should be that eny material exhibiting a forward and a reverse
`martensitic transformation, thermally or stress induced, is a potential shape memory alloy.
`It is only necessary to determine the exact conditions, associated with a complete and useful
`shape recovery, under which the martensitic transformation occurs. *7
`Another selection rule for shape memory alloys is the existence of an ordered structure. *
`It has been shown that the ordered Fe,Pt-alloy exhibits a thermoelastic martensitic trans-
`formation and a related shape memory effect, while a disordered Fe,Pt-alloy of the same
`composition displays neither,’**? This can be explained on the basis that the shape memory
`effect, as well as the pseudo-elasticity, originates from a complete crystallographic revers-
`ibility of the martensitic transformation. Otsuka and Shimizu* have shown that a path in
`the reverse transformation in ordered alloys is unique,
`in contrast to multiple paths in
`disordered alloys. They propose that the complete crystallographic reversibility ofthe mar-
`tensitic transfermation is characteristic of ordered alloys and rationalize the fact that the
`shape memory effect has usually been observed in ordered alloys. Besides, in the case of
`an ordering, treversible plastic accommodation requires the creation of superdislocations,
`possessing energies which are multiples of those of the dislocations in the disordered lattice,
`while the matrixyield stress increases considerably as the result of the ordering (Figure 3).
`Therefore, ordered alloys with their reversible martensitic transformation and their absence
`of lattice invariant plastic accommodation at the habit plane are much more favorable for
`shape memory purposes than the disordered ailoys. There exisis only one exception to this
`among shape memory alloys: the disordered In-TI alloy.’ However, here it involves a fee
`+ fet transformation, which is reversible even in disordered alloys. This stems from the
`fact that the lattice correspondence is unique in the reverse transformation because of the
`very simple lattice changé and lower symmetry of the fet phase.”?
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`IRONBURG EX2020, Page 9
`
`IRONBURG EX2020, Page 9
`
`

`

`70
`
`Metal and Ceramic Biomaterials
`
`STRESS
`
`
`
`ORDERED
`
`
`
`DISORDERED
`
`LEVEL OF SHEAR AND YOLUME
`STRAINS DURING PLATE GROWTH
`
`——
`
`STRAIN
`
`Hypothetical stress-strain curves for ordered and disordered
`FIGURE3.
`martensite, In thé latler case transformation strains exceed (he matrix for
`miurtensite) elastic limit and interface coherenceis iost.'*
`.
`
`Another requirementfor obtaining shape memory properties is that the martensite should
`be internally twinned.'® If the deformed alloy has to revert ta its initial State, ¢.g., when
`heated, the deformation process must be reversible, i.¢., the shape memory alloy should
`hot contain mobile dislocations. A possible mechanism for plastic deférmation without
`mobile dislocations is internal detwinning of themartensitic substructure, The deformation
`can result from a selective detwinning process according to whieh one of the two avin
`orientations grows at
`the expense of the other, However,
`this theory seems subject to
`reservations. For example, Cu-Zn alloys are not internally twinned, yet they exhibit the
`shape memory effect.
`Althcugh not all
`the crystallagraphic requirements are conclusive, the shape memory
`effects are a reality. Fortunately, a detailed understanding of the physics of a process is not
`always necessary for successful exploitation.
`
`D. Shape Memory Effect
`When a conventional material is plastically deformed, a permanent deformation remains
`after unloading. The material can only regainits original shape by a new plastic deformation.
`The shape memory alloys are distinguished from the “usual” alloys By the fact that,
`when these alloys are apparently plastically deformed, they reinembertheir preceding shape
`on changing the temperature. Depending on ‘the initial amount of deformation,
`they can
`regain their original shape (totally or partiatly) during heating. One can recognize two kinds
`of shape memory effects: (1) one-way shape memory effect and (2) two-way stiape memory
`effect.
`
`1, One-Way Shape Memory Effect
`Figure 4 depicts schematically the one-way shape memory effect. The macroscopic de-
`formation, which may not exceed a critical strain limit, is accompanied by 2 martensitic
`transformation, not reversed by removing the applied stress. If the specimen is fully mar-
`tensitic at the onset of the deformation (T, <M), the existing thertnally induced mattensite,
`M, is tcansformed by reorientation, according to one of the mechanisms of Section ULB.3,
`to the martensite variant M’. The reorientation takes place at the regidn BC of Figure 4.
`On unloading, the material recovers elastically, but a permanent deformation AD remains
`with a thermodynamically stable maztensite structure. If the material is subsequenily heated,
`the recovery of this deformation starts at the temperature A,. The martensite. variant obtained
`by reorientation reverts to the austenitic phase.
`In the ideal case, strain recovery is complete if A, is reached, For Nitiiol, strains of the
`order of 6 to 8% maycompletely recover. If subsequent changes in temperature,
`in the
`
`
`
`IRONBURG EX2020, Page 10
`
`
`
`i
`
`'
`
`IRONBURG EX2020, Page 10
`
`

`

` Volume ff: Strength and Surface
`
`STRESS
`
`
`
`71 STRAIN
`TEMPERATURE FIGURE 4. Qne-way shape memary effect.
` sal
`~~e
`
`.
`
`eaEe
`
` TEMPERATURE
`
`
`Stress-assisted transformation in Cu-ALNi.””
`FIGURE 5.
`
`
`
`absence of external forces, do not change the macroscopic shape anymore, the phenomenon
`is called the oné-way shape memory effect.
`Another possibility is that the specimen is partially both martensitic and austenitic, If, for
`
`example, M, < T, < M, (Fa reached on cooling}, the austenitic phase is transformed upon
`
`unloading to the irrevérsible stress-induced martensite variant, resulting in a predominantly
`
`martensitic stfucture, concomitant with a remaining deformation,
`
`The last possiblity is that T, > A,, i-¢., the specimen is fully austenitic. For Cu-Al-Ni
`
`the transformation path does not simply consist of a forward and a reverse.transformation
`
`as usual, but consists of four steps”? (Figure 5). By deformation of the austenitic structure
`
`B, stress-induced martensite @’ will grow, resulting in a new +/-martensite variant, con-
`
`comitant with a permanentstrain upon cooling below M,underconstant stress (i.e., stress-
`
`assisted transformation) and after subsequent unloading. This strain can be annihilated by
`
`subsequent heating above the A,-A, range, wherethe y ‘-martensite variant regainsits original
`
`B-austenite structure.
`
`i
`
`
`
`
`
`
`|
`
`Pik
`
`
`
`
`
`
`
`IRONBURG EX2020, Page 11
`
`IRONBURG EX2020, Page 11
`
`

`

`32
`
`Metal and Ceramic Biomaterials STRESS
`
`TEMPERATURE
`
`
`
`:
`
`the number of martensite variants, formed when an alloy is repeatediy heated and cooled
`
`FIGURE 6.
`Two-way shape memory effect,
`A remarkable effect on heating is the appearance of an external measurable force, which
`can perform work. Because of the higher symmetry of the austenitic structure compared
`withthat of the martensitic siructure, the specimen strongly prefers to regain the austenitic
`state on heating. A part of the transformation energy is released as this recovery stréss.
`2. Two-Way Shape Memary Effect
`In contrast with the one-way shape memory effect, subsequent cooling does indeed in-
`fluence the macroscopic shape, while no external forces are applied. As well as the material
`remembering its undeformed shape on heating,
`it also remembers the deformed shape on
`cooling. Figure 6 depicts this phenomenon schematically. After deformation and unloading
`the material, a permanent deformation, AB, remains. Due to an initial nonuniform defor
`mation (¢.g., bending) above a critical strain €,
`(for Nitinol, €, = 8%), a plastic deformation
`of the martensitic stfucture, stress-induced or reoriented, is introduced, associated with
`internal stresses and préferential nucleationsites. These stresses and nucleation sites control
`the growth of a very select number of strongly textured martensite variants.
`Due to