THERMO-MECHANICAL CHARACTERIZATION OF NiTiCu AND ...
Transcript of THERMO-MECHANICAL CHARACTERIZATION OF NiTiCu AND ...
THERMO-MECHANICAL CHARACTERIZATION OF NiTiCu AND NiTi
SMA ACTUATORS: INFLUENCE OF PLASTIC STRAINS
David A. Miller* and Dimitris C. Lagoudas§
Aerospace Engineering Department,
Texas A&M University
College Station, TX 77843-3141
ABSTRACT
The focus of this study is the thermomechanical characterization and comparison
between two different shape memory alloys (SMAs) quantifying the effect of plastic strain on the
transformation characteristics of SMA actuators. In this study, the thermomechanical response
and transformation characteristics of a NiTiCu and a NiTi SMA are studied as a function of the
induced plastic strain for four different loading paths: 1) an elastic-plastic loading of the
austenitic phase, 2) a stress-induced martensitic phase transformation, 3) an elastic-detwinning-
plastic loading of the martensitic phase and 4) thermally-induced phase transformation under a
constant applied stress. Each loading path is repeated multiple times, with an incremental change
of the total applied strain, to determine the effect of accumulated plastic strain on the phase
transformation characteristics of the two SMA material systems. The effects of plastic strain are
quantified by measurements of recoverable strain during a thermal cycle under zero applied
stress, measurements of heat of transformation from a differential scanning calorimeter and
microstructural evaluations.
* Graduate Research Assistant, E-mail: [email protected]
§ Professor, E-mail: [email protected]
2
1. INTRODUCTION
Over the last decade shape memory alloys have seen growing use in the mechanical,
medical, and aerospace industries (Birman 1997). Most of the applications have been 1-D in
nature where wires, strips, and rods were employed, i.e. as actuators in active wings (Garner et
al, 1999), robotic systems and self-extracting microstructures (Sachdeva and Miyazaki 1990).
Unfortunately, the inability to identify material properties and the inherent complex
thermomechanical behavior of SMAs has stifled widespread use. Material properties of SMAs
can undergo significant changes with differences in the chemical composition, cold work, heat
treatment and thermomechanical cycling. The addition of copper as a ternary element, for
example, has demonstrated favorable results for the development of SMA actuators due to a
reduction in the temperature hysteresis, while the yield stress is significantly lower than binary
NiTi (Funakubo 1987).
The austenite to martensite phase transformation characteristics of NiTi SMAs has been
shown in previous studies to be related to the presence of lattice defects introduced by cold
working (Hebda and White 1995, Liu and McCormick 1990, McNeese 1998, Matsumoto 1992,
Moraweic et al. 1995). The process of drawing NiTi wires imparts large plastic deformations, i.e.
cold work, on the material in the martensitic condition (Lin et al 1994, Jackson 1972). To reduce
the bulk material to typical wire diameters of 0.3 mm to 2 mm, multiple drawings are performed,
with each followed by an anneal at temperatures in the region of 800°C to 900°C for roughly 15-
30 minutes. Qualitatively, the effects of cold work on the transformation properties have been
shown to be independent of material composition for small percentage changes in binary
material. In general, the effects of cold work are a reduction in the transformation temperatures,
smaller reversible strains and increases in the yield stress for slip in martensite. Also, a high
3
dislocation density restricts phase boundary movement and the development of stress-induced
and reoriented martensite, which results in increased hardening in the stress-strain curve (Filip
and Mazanec 1995).
During the lifespan of the SMA actuator, loss of actuation can occur through repeated
cycling due to plastic strain development, and unlike common ductile metals, plastic deformation
in SMAs can be induced by the martensitic phase transformation and can occur at relatively low
stress levels. There are many experimental results on stress-induced martensite at temperatures
above the austenite finish temperature, Af, (pseudoelastic response) showing the effects of strain
level, stress level, cycle number, pre-straining and strain-rate on the transformation
characteristics (Lim and McDowell 1994, Shaw and Kyriakides 1994, McCormick et al. 1993,
Tobushi, et al. 1993, Eucken and Duerig 1989, Miyzaki et al. 1986, and Otsuka and Shimizu
1986). In these works, plastic deformation is developed during the loading as a result of the
phase transformation process and the amount of plastic strain remains small compared to the
overall applied strain. Miyazaki, et al. (1981) investigates large plastic deformations for
isothermal mechanical loading, however, no quantitative results are given for changes in the
transformation characteristics. Bo and Lagoudas (1999) and Liu and McCormick (1990) have
studied the development of transformation induced plastic strain and two-way strain during
thermally induced cyclic phase transformations under a constant applied stress. Two-way strain
is the term used to describe the strain that develops during the austenitic to martensitic phase
transformation under zero load. This strain is the result of dislocation arrangements that guide
the formation of martensite variants to a preferred orientation, thus resulting in an overall change
in length. These works have shown that two-way strain and plastic deformation are developed
through multiple cycles of these loading paths. While in Bo and Lagoudas (1999) large plastic
4
strains have been observed for fully annealed NiTi SMA, much smaller strains have been
measured in Liu and McCormick (1990) for NiTi with a cold worked microstructure.
Even though the transformation induced plastic strain development has been studied
extensively, the influence of plastic strains on transformation characteristics has not been fully
addressed. The main focus of this study is to quantify the effects of plastic strains developed
during a thermomechanical loading on the transformation characteristics of binary NiTi and
NiTiCu specimens. In addition to plastic strain induced during the phase transformation, the
influence of plastic strain induced in the austenitic phase and the detwinned martensitic phase on
the transformation characteristics will be studied in this work.
The paper is divided in the following manner. The experimental procedures employed in
this study are detailed in section two describing the material selection, equipment and
mechanical tests performed. Section three presents the experimental results and is divided into
four sub-sections detailing the four loading paths performed. The fourth section discusses the
results and the conclusions are given in the fifth and final section of the paper.
5
2. EXPERIMENTAL PROCEDURE
To investigate the influence of plastic strains on the transformation characteristics of
SMAs, a thorough experimental investigation was performed. Motivated by the multiple phases
that occur during the application of SMAs, tests were performed for each material phase
independently, i.e., pure austenitic and pure martensitic loading, and for loading paths which
involve a phase change, i.e. stress-induced martensite and thermally-induced martensite. These
loading paths are depicted in Figure 1. In this section the specimens, experimental procedures
and equipment used to perform this investigation are fully described.
The geometry of SMA specimens has largely been driven by the desired applications into
which the SMAs are envisioned. Numerous SMA test specimens utilizing different cross sections
have been experimentally investigated including rectangular specimens (Gall et al 1999), bar
specimens (Howard 1995), wire specimens (Shaw and Kyriakides 1995), square specimens (Liu
and McCormick 1990), thin film specimens (Miyazaki and Ishida 1992) and tubular specimens
(Lim and McDowell 1999). However, over the last decade, SMAs have been primarily employed
in 1-D applications where small strips and wires are required. Due to this fact and the knowledge
that SMA response is strongly a function of the prior processing, it becomes necessary to test the
response of SMAs in the geometry in which it will be applied. Therefore, the specimens chosen
for this experimental study are in the form of wires.
The specimens utilized in this study were provided by Memry Corporation and are a Ni-
45at%Ti - 10at%Cu wire with a diameter of 0.6 mm and a Ni- 50at%Ti wire with a diameter of
0.91 mm, both in the as-drawn condition. Before testing, a 600°C/ 30-minute heat treatment was
performed on the NiTiCu specimen. A Perkin-Elmer Pyris 1 Differential Scanning Calorimeter
(DSC) was utilized to determine the phase transformation temperatures and heat of
6
transformation, ∆H, (Jardine 1989) measured as the total area under the curve during the heating
and cooling cycle, for each specimen prior to testing and after the testing is completed. DSC
results for the specimens prior to testing are shown in Figure 2, with the hatched potion of the
NiTi curve delineating the area used for the calculation of ∆H.
SMA specimens were loaded under three different isothermal mechanical loading paths,
labeled as loading path 1, 2 and 3 in Figure 1. As seen in the figure, these paths each represent a
significantly different test; loading path one represents an elastic-plastic loading of the austenitic
phase, loading path two represents a stress-induced martensitic phase transformation, and
loading path three represents an elastic-detwinning-plastic loading of the martensitic phase. Each
of these loading paths are repeated multiple times, incrementing the total applied strain by 1.0%
to 2.0% for each loading, to determine the effect of plastic strain on the phase transformation
characteristics of the SMA. To quantify the effect of the plastic strain, each mechanical loading
cycle is followed by a thermal cycle under a constant stress of 5 MPa from which the
transformation temperatures and two-way strain are measured. The final thermo-mechanical
loading performed were multiple thermal cycles under a constant applied stress, shown as
loading path 4 in Figure 1, a loading path commonly used to induce a two-way shape memory
effect in the SMA.
All mechanical tests were performed on an MTS servo-hydraulic load frame equipped
with a custom built environmental chamber, see Figure 3. The environmental chamber uses a
resistive heating coil to heat the specimen and liquid CO2 for cooling. The temperature of the
specimen is measured in three locations on the specimen using K-type thermocouples held in
contact with the specimen and mounted with a thermally conductive paste to ensure good heat
transfer qualities between the specimen and thermocouple. The thermocouples are .005” in
7
diameter to ensure a fast response and small disruption of the temperature field on the specimen.
A control program utilizing LabWindows was used to control the crosshead displacement,
heater and CO2 operations and collect the temperature of the specimen, force and displacement.
The mechanical loading was performed in displacement control with a constant strain rate of
1x10-4 in/in/sec and the thermal cycles were performed at approximately 2 °C/min. The reported
strain levels were calculated using ε=∆L/L0 where the length, L0, represents the original
specimen length and ∆L is measured from the crosshead displacement. Due to the small loads
that were applied to the wire as compared to the overall stiffness of the test frame, a machine
compliance correction was not computed into the results. The thermal strain induced in the grips
due to the heating and cooling cycles was measured and determined to be a negligible quantity in
the strain calculations.
It has been shown that phase fronts occur in annealed SMAs and that diffused phase
fronts occur in cold worked SMAs due to multiple nucleation sites (Shaw and Kyriakides 1995,
and Howard, 1995). Since hardening and multiple phase fronts exist during the mechanical
loading, and extensometers and strain gauges are not suitable for measuring the strain in the
specimen unless specially designed for this application. Since plastic strain development cannot
be confined to a specific gauge section, a global strain measurement is the obvious choice for
strain measurement. Thus, the crosshead displacement is used for the strain calculation in the
SMA specimen. The effect of the grips on the overall stress field is minimized by the length of
the specimen contrasted to the diameter of the wire. Slipping of the specimen in the grip was
closely monitored and did not occur in the any of the tests.
8
3. EXPERIMENTAL RESULTS
3.1. Loading Path 1: Austenitic Plastic Loading
Loading path 1 entails a mechanical loading of the austenitic microstructure into the
plastic region of deformation, see Figure 1. The temperature for loading path 1 was chosen such
that the specimen would plastically deform before the formation of stress induced martensite. It
was calculated using estimated values of the martensitic stress-temperature slope, Cm in Figure
1, the initial DSC result for the martensitic start temperature and an estimate of the yield stress of
austenite. Using this criterion, tests were performed at 155°C for the NiTiCu specimen, and
120°C for the NiTi specimen. The mechanical response of the austenitic phase is similar to that
of typical engineering materials for both material systems, showing initial yielding followed by
linear hardening, as seen in the stress-strain curves in Figure 4a for the NiTiCu specimen and
Figure 5a for the NiTi specimen. Measurements from the initial loading cycle result in an
austenitic elastic Young’s Modulus, EA, of 67.9 GPa and 64.5 GPa, and a 0.2% offset yield stress
of 290 MPa and 370 MPa for the NiTiCu specimen the NiTi specimen, respectively. Following
each mechanical unloading, a thermal cycle under a constant stress of 5 MPa is applied to
evaluate the effect of the plastic strain on the transformation characteristics. A summary graph of
the strain-temperature results is seen Figure 4b and Figure 5b for each specimen, where each
thermal cycle shown starts at the strain level from which the previous mechanical loading ended.
From each thermal cycle, the transformation temperatures and two-way strain are measured and
plotted as a function of the plastic strain and included in Figure 6. For both material systems, the
strain-temperature response shows the growth of a negative two-way strain, and an enlarging of
the hysteresis loop. The data shows that similar levels of two-way strain are created for each
9
material system, however, the NiTi specimen peaks before attaining this value and begins to lose
the two-way shape memory effect at plastic strains larger than 8.5%. The negative two-way
strain is contrary to the positive two-way strain typically seen in SMAs, however, it does provide
insight into the training and transformation process. From these results, it is seen that
dislocations developed in the austenitic microstructure are inherited into the martensitic
microstructure. As temperature-induced martensite is formed, the internal stresses created by the
plastic deformation influence the martensitic variant formation and a global negative strain
results. Fremond and Miyazaki (1996) have experimentally observed that plastic deformation in
the austenitic state results in zero two-way strain, however, it is seen in both material systems
presented in this work. Post-test DSC results are shown in Figure 7 and Figure 8, along with a
comparison of the original DSC result and the final strain-temperature curve for the NiTiCu and
NiTi specimens respectively. As seen in the figures, the latent heat of transformation is reduced
and shifted down in temperature for both material systems. Also, the austenite to martensite peak
for the NiTi specimen is broadened such that a discernable peak is not obvious, a fact not seen
for the NiTiCu specimen and attributed to initial cold-drawn condition of the NiTi wire. The
values for the latent heat of transformation, ∆H, are given in Table 1. Using the average ∆H
value of A⇒M and M⇒A, a reduction in latent heat to 49.6% and 56.7% of the pre-tested value
is shown for the NiTiCu and NiTi specimen. The large drop in the ∆H values is attributed to the
dislocations inhibiting the amount of material allowed to undergo the phase transformation (Lim
and McDowell, 1994).
10
3.2. Loading Path 2: Stress-Induced Martensitic Loading
The second mechanical loading path, see Figure 1, was conducted at temperatures such
that stress-induced martensite was created and the pseudoelastic response would be seen upon
unloading. Using initial DSC results, isothermal mechanical loadings were performed at 85°C
for the NiTiCu and 40°C for the NiTi specimen, temperatures above austenitic finish. Figure 9a
and Figure 10a show the stress-strain results for the stress-induced martensitic loading path for
the NiTiCu and NiTi specimens respectively. It is of interest to note that the pseudoelastic
response is not seen in the stress-strain results, a fact explained by Otsuka and Shimizu (1986)
and Miyazaki et al (1982) as a result of having a low critical stress for slip. To fully recover the
detwinned strain imparted into the specimen during the mechanical loading, the specimens are
heated upon unloading under a constant load of 5 MPa to a temperature 50°C above the
austenitic finish temperature. The oscillations seen in the stress-strain response in Figure 9 are
attributed to ± 1.5°C oscillations of the specimen temperature induced by to the inability of the
temperature controller to adequately respond to the release of latent heat. The increase of the
specimen temperature requires a higher stress for the creation of stress-induced martensite, as
seen in Figure 1 and described in Shaw and Kyriakides (1995), and the material response is an
increase of the slope in the stress-strain response. As the specimen temperature decreased, the
stress required for transformation also decreased and the slope of the stress-strain response was
also decreased. After the specimens were heated to recover the remaining detwinned martensite,
a thermal cycle under 5 MPa was performed to evaluate the effect of the plastic strain on the
transformation characteristics. A summary graph of the strain-temperature results is seen in
Figure 9b and Figure 10b for each specimen, where each thermal cycle shown starts at the strain
level from which the previous mechanical loading ended As seen in the figures, plastic strain
11
was applied into the specimen in each cycle, which in turn, altered the stress-strain response for
the following cycle and imparted a two-way strain as evidenced in the strain-temperature
response. Figure 11 shows the development of two-way strain for both specimens as a function
of the induced plastic strain. The data points elucidate the trend of two-way strain development
as the plastic strain level increases, and shows that similar two-way strain levels are achieved for
both specimens. As the level of plastic strain increased, Figure 9a and Figure 10a show that the
stress level at the initiation of stress-induced martensite was lowered. Also, increased hardening
was exhibited during the phase transformation in the stress-strain response as the level of plastic
strain increased. These two observations are both related to the broadening of the transformation
temperature region as seen in the DSC and two-way strain graphs shown in Figure 12 and Figure
13. The final DSC for both specimens shows essentially no phase transformation peak in the
austenitic to martensitic transformation due to the extreme broadening of the transformation over
the entire temperature range, a fact which is echoed in the strain-temperature curves in Figure 9
and Figure 10. Using the average ∆H value of A⇒M and M⇒A, a reduction in latent heat to
29.1% and 33.0% of the pre-tested value is shown for the NiTiCu and NiTi specimen,
respectively.
3.3. Loading Path 3: Martensitic Plastic Loading
Mechanical loading path 3, see Figure 1, was performed at temperatures such that
martensitic detwinning and plastic deformation were the primary deformation mechanisms. The
isothermal mechanical loading was performed at 22°C for the NiTiCu specimen and -30°C for
the NiTi specimen and the stress-strain and strain temperature results are included in Figure 14
and Figure 15. Upon unloading to zero stress in each cycle, the specimen was held at a constant
12
stress of 5 MPa and heated to a temperature 50 °C above Af to recover the detwinned strain and
measure the permanent strain in the specimen. The strain-temperature results show the
development of two-way strain, which is quantified in Figure 16. Measurements from the initial
loading cycle result in an martensitic elastic Young’s Modulus, EM, of 16.0 GPa and 23.5 GPa,
for the NiTiCu specimen the NiTi specimen, respectively. Both specimens show similar stress-
strain behavior with respect to the development of plastic strain.
The initial cycles, where the plastic strain level is small, show a Lüders type of
deformation (Miyazaki et al 1981), i.e. the strain level increases without an increase in the stress
level. However, as the plastic strain level is increased, the Lüders type of deformation begins to
disappear, the stress level for the initiation of detwinning decreases and a dramatic hardening
effect is seen resulting in large stress levels. This deformation mode is utilized for the cold
working and drawing procedures (Lin, Wu, and Lin 1994), and these results are evidence to the
lack of Lüders deformation in heavily cold worked materials. However, is has not been shown
that cold worked materials display the two-way strain observed in the plastically deformed
specimens of this study, but it could be concluded from these results that for small amounts of
cold work a two-way strain could be developed. Figure 16 shows the history of two-way strain
development as a function of the plastic strain. As seen in the stress-induced martensitic case, the
final two-way strain values are similar, however, the plastic strain level for the NiTi specimen is
only half of the NiTiCu specimen, and might in fact decrease at higher levels of plastic strain.
The final DSC curves are shown in Figure 17 for the NiTiCu and Figure 18 for the NiTi
specimens. The figures show the extreme broadening of the phase transformation temperature
range and the reduction of the latent heat of transformation. Using the average ∆H value of
13
A⇒M and M⇒A, a reduction in latent heat to 25.7% and 17.5% of the pre-tested value is shown
for the NiTiCu and NiTi.
3.4. Loading Path 4: Thermal Cycles under Constant Applied Stress
The fourth loading path applied to the specimens consisted of multiple thermal cycles
through the transformation temperatures, i.e. from T< MF to T>AF, under a constant applied
stress of 200 MPa, see Figure 1. The strain-temperature results of the first fifty thermal cycles are
included in Figure 19 and Figure 20 for the NiTiCu and NiTi specimens, respectively. These
results show the development of plastic strain with each thermal cycle and the saturation of the
plastic strain as the number of cycles increased. The transformation strain in the final cycle was
similar for each material system, 5.25% for the NiTiCu and 5.30% for the NiTi specimen,
however, the NiTiCu specimen had nearly twice the plastic strain accumulation at 8.65%
compared to 4.02% for the NiTi specimen. Figure 21 and Figure 22 show the initial and final
DSC results along with the final two-way strain as a function of temperature for the NiTiCu and
NiTi specimens. In these graphs, the shifting and broadening effect on the DSC curves induced
by the loading is shown, as well as a final two-way strain of 2.1% for the NiTiCu specimen and
3.2% for the NiTi specimen. The latent heat of transformation for each specimen is shown in
Table 1. Using the average ∆H value of A⇒M and M⇒A, a reduction in latent heat to 44.6%
and 42.0% of the pre-tested value is shown for the NiTiCu and NiTi specimens respectively.
14
4. DISCUSSION
The method in which plastic strain is induced into the specimen is shown in the
experimental results to strongly influence the amount of two-way strain developed. Plastic
deformation of the austenitic microstructure showed the development of a negative two-way
strain. This unique result was seen in two separate material systems and shows the effect
dislocations have on the deformation characteristics of SMAs. A negative two-way strain
indicates that dislocations in the austenitic microstructure influence the preferred orientation of
the variants formed during thermally induced martensite.
Of the three loading paths that introduced positive two-way strains, the single-phase
martensitic loading path developed the highest two-way strain for both material systems.
Additionally, the rate of two-way strain development with plastic strain accumulation was
highest for the martensitic loading path, resulting in higher two-way strains at lower plastic
strains. This observation implies that dislocation development through the detwinning process
and plastic slip of the martensitic microstructure generates a higher preferential ordering of the
martensitic variants, and thus a higher two-way strain, than does dislocation development
through stress-induced martensite or thermally induced martensite. This result may be explained
by examining the deformation mechanisms involved with the stress-induced martensitic loading
path and the temperature-induced martensitic loading path. Lim and McDowell (1994) describe
that plastic strain in SMAs cycled in the pseudoelastic regime is a combination of the existence
of martensite which is “locked in” by the dislocations generated at the austenitic-martensitic
phase boundaries during the transformation and plastic deformation of the induced martensitic
phase itself.
15
McCormick and Lui (1994) make a similar argument that for thermal cycles under a
constant load an increasing fraction of martensite is retained at the end of each cycle, and
comprises a portion of the permanent strain. Plastic slip of the thermally induced martensitic
phase will not be present unless the cycles are applied at a stress level above the yield limit of the
martensitic phase. Therefore, the existence of plastic strain induced by these loading paths,
loading paths 2 and 4 in this study, may be a combination of the retention of the martensitic
phase and plastic deformation of the martensitic phase resulting in higher plastic strains with
lower two-way strains. Experimental verification of this fact is seen in post-test microstructural
evaluation of the specimens, and shows that retained martensite is present in the austenitic phase
for loading path 2 and 4, as well as loading path 3.
A representative microstructure is shown in Figure 23 for the martensitic plastic
deformation NiTi specimen after being heated above the austenitic finish temperature and
photographed at room temperature, above the martensitic start temperature. This micrograph
shows the existence of martensitic plates, marked by the arrows, throughout the grains
confirming the existence of retained martensite. Additionally, Figure 24 shows the evolution of
the martensitic strain recovered upon heating as a function of the plastic strain in loading path 2
and 3. The reduction of detwinned strain for both specimens is further evidence that martensite is
retained in the austenitic microstructure for both material systems. Micrographs of the NiTiCu
austenitic microstructure are not currently available due to the elevated temperature of the
austenitic finish temperature. However, using the results of the recoverable strain, it can be
concluded that retained martensite is present for the NiTiCu specimen as well.
For each loading path a broadening and downward shift in temperature of the
transformation peaks is seen in the DSC results, showing that the influence on the transformation
16
temperatures is independent of the method in which plastic strain is induced. The shift in the
transformation temperatures is created by the existence of dislocations, which introduces
obstacles for the moving interfaces (Moraweic et al 1995). The method of applied plastic strain
does show a consistent influence on the heat of transformation, ∆H, for both material systems
studied. For all loading paths a significant reduction in the heat of transformation was observed,
which can be attributed to retention of the martensitic phase at temperatures above austenite
finish (Moraweic et al 1995). In both material systems, the martensitic detwinning loading path
had the lowest heat of transformation, followed by the stress-induced martensitic loading path
and thermally-induced loading path. The austenitic loaded material retained the highest heat of
transformation for both material systems. From these results is apparent that the martensitic
detwinning loading path creates the highest amount of retained martensite. Qualitative
microstructural evaluation of the specimens supports this result.
17
5. CONCLUSIONS
In this study dislocations have been introduced, quantified by a measurable permanent
strain, and the effect of the plastic deformation on the transformation characteristics of two-way
strain, transformation temperature and heat of transformation has been reported.
Pure austenitic plastic deformation for both NiTiCu and NiTi has been shown to induce a
negative two-way strain, a contrasting result to published literature. Positive two-way strains
have been produced for loading paths involving stress-induced martensite, temperature-induced
martensite and martensitic detwinning. For both the NiTiCu specimen and the binary NiTi
specimen the martensitic detwinning loading path resulted in the highest two-way strain. Also,
this two-way strain was developed at the lowest level of plastic strain. Therefore, the martensitic
detwinning loading path is established as the best method of two-way strain development.
Post-test DSC results show a significant reduction in the heat of transformation, ∆H. The
influence of the plastic strain on ∆H is due to the retention of martensite in the austenitic
microstructure. The generation of dislocations associated with the plastic strain “locks in” the
martensitic microstructure, thus removing a portion of the microstructure from the
transformation and reducing the heat of transformation. The reduction of available
microstructure for transformation is also observable through the reduction of recoverable strain
and microstructural evaluation.
18
ACKNOWLEDGEMENTS
The authors acknowledge the financial support of Air Force Office of Scientific Research under
grant No. F49620-98-1-0041.
REFERENCES
Birman, V. 1997 Review of Mechanics of Shape Memory Alloy Structures Applied Mechanics
Reviews 50 (11) pp 629-645.
Bo, Z. and Lagoudas, D.C. 1999 Thermomechanical Modeling of Polycrystalline SMAs Under
Cyclic Loading, Part III: Evolution of Plastic Strains and Two-Way Memory Effect Int J
Eng Sci., at press.
Bo Z., Lagoudas, D.C., Miller, D.A. 1999 Material Characterization of SMA Actuators under
Non-Proportional Thermomechanical Loading J. Engineering Materials and Technology
121 pp 75-85.
Eucken, S., and Duerig, T.W. 1989 The Effects of Pseudoelastic Prestraining on the Tensile
Behavior and Two-way Shape Memory Effect in aged NiTi Acta metall 37 (8) pp 2245-
2252.
Filip, P. and Mazanek, K. 1995 Influence of Work Hardening and Heat Treatment on the
Substructure and Deformation Behavior of NiTi Shape Memory Alloys Scr. Metall, Mater.
32 (9) pp. 1375-138.
Fremond, M and Miyazaki, S. 1996 Shape Memory Alloys CISM Courses and Lectures No. 351
(New York: SpringerWien).
Funakubo, H (ed) 1987 Shape Memory Alloys (London: Gordon and Breach).
19
Gall, K., Sehitoglu, H., Chumlyakov, Y.I., Kireeva, I.V. and Maier H.J. 1999 The Influence of
Aging on Critical Transformation Stress Levels and Martensitic Start Temperatures in NiTi
Part II: Discussion of Experimental Results J. Engineering Materials and Technology 121
pp 28-37.
Garner,L.J, Wilson, L.N., Lagoudas, D.C. and Rediniotis, O.K. 1999 Development of a Shape
Memory Alloy Actuated Biomimetic Vehicle Smart Materials and Structures published in
this issue.
Hebda, D. and White, S.R.1995 Effect of Training Conditions and Extended Thermal Cycling on
Nitinol Two-way Shape Memory Behavior Smart Materials and Structures 4 pp 298-304.
Howard, S.D. 1995 The Thermomechanical Constitutive Experimentation of NiTi Shape
Memory Alloy Strips and Rods Master’s Thesis (Texas A&M University, College Station,
TX).
Jackson, C.M., Wagner, H.J., Wasilewski, R.J. 1972 55-Nitinol- The Alloy with a Memory: Its
Physical Metallurgy, Properties, and Applications NASA-SP5110.
Jardine, A.P. 1989 Calorimetric techniques for the evaluation of thermal efficiencies of shape
memory alloys J. Material Science 24 pp2587-2593.
Lim, T.J., and McDowell, D.L. 1994 Degradation of an Ni-Ti alloy During Cyclic Loading,
Proceedings of the 1994 North American Conference on Smart Structures and Materials,
SPIE, Orlando, Florida pp 153-165.
Lim, T.J. and McDowell, D.L. 1999 Mechanical Behavior of a NiTi-Shape Memory Alloy under
Axial-torsional Proportional and Non-Proportional Loading J. Engineering Materials and
Technology 121 pp 9-18.
20
Lin, H.C., Wu, S.K. and Lin, J.C. 1994 The Martensitic transformation in Ti-rich Shape Memory
Alloys Mater. Chem. Phys. 37 pp 184-190.
Liu, Y., and McCormick, P.G. 1990 Factors Influencing the Development of Two-Way Shape
Memory effect in NiTi Acta Metall. Mater 38 (7) pp 1321-1326.
Matsumoto, H., 1993 Transformation behavior of NiTi in relation to thermal cycling and
deformation Physica B 190 pp 115-120.
McCormick, P.G., Liu, Y. and Miyazaki, S. 1993 Intrinsic thermal-Mechanical Behavior
Associated with the stress-induced martensitic transformation in NiTi Material Science and
Engineering A167 pp51-56.
McCormick, P.G. and Liu, Y.1994 Thermodynamic Analysis of the Martensitic Transformation
in NiTi-II. Effect of Transformation Cycling Acta Metallurgica el Materialia 42 (7) pp
2407-2413.
Miyazaki, S., Imai, T., Igo, Y. and Otsuka, K. 1986 Effect of Cyclic Deformation on the
Pseudoelasticity Characteristics of Ni-Ti Alloys Metal. Trans.A 17A pp 115-120.
Miyazaki, S. and Ishida, A. 1992 Shape Memory Characteristics of Sputter-depositedTi-Ni Thin
Films Material Transactions JIM 34 pp 14-19.
Miyazaki, S., Ohmi, Y. Otsuka, K. and Suzuki, Y. 1982 Characteristics of Deformation and
Transformation Pseudoelasticity in NiTi Alloys J. de Physique 15 pp 287-292.
Miyazaki, S., Otsuka, K. and Suzuki, Y. 1981 Transformation Pseudoelasticity and Deformation
Behavior in a Ti-50.6at%Ni Alloy Scr. Metall. 15 pp 287-292.
Morawiec, H., Stroz, D. and Chrobak, D. 1995 Effect of Deformation and Thermal Treatment of
NiTi Alloy on Transition Sequence J. de Physique, Suppl 5 pp C2-205-209.
21
Otsuka, K., Shimizu, K., 1986, Pseudoelasticity and Shape Memory Effects Int. Metals Review
31 (3) pp 93-114.
Sechdeva, R.C. and Miyazaki, S. 1990 Superelastic Ni-Ti Alloys in Orthodontics, in T.W.Duerig
et al (eds.) Engineering Aspects of Shape Memory Alloys pp 452-469.
Shaw, J.A. and Kyriakides, S. 1995 On the thermomechanical behavior of NiTi J. Mechanics
and Physics of Solids 43 (8) pp 1243-1282.
Tobushi, H., Onashi, Y., Saidi, H., Hori, T. and Shirai, S. 1993 Pseudoelasticity of NiTi Shape
Memory Alloy (Dependence on Maximum Strain and Temperature) JSME Int.J. 36 (3) pp
314-318.
22
TM os AosM of Aof
MartensiteAustenite
σPlastic Deformation
LoadingPath 1
LoadingPath 4
LoadingPath 3
LoadingPath 2
Cm
Figure 1 Loading Paths In Stress/Temperature Space
0.45
0.55
0.65
0.75
0.85
0.95
1.05
-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110
Temperature, °C
Hea
t F
low
, W/g
(en
do u
p)
NiTiCuNiTi
∆H
Figure 2 Initial DSC results for NiTiCu and NiTi specimens showing heat of
transformation, ∆∆H
23
LoadCell
Grip
ThermocouplesSMASpecimen
LVDT Servo-HydraulicActuator
Furnace
Figure 3 Experimental Set-up
-0.01
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
0.11
0.12
0.13
0.14
0.15
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180
Temperature, °C
Austenitic Plastic Deformation
0
100
200
300
400
500
600
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16
Strain
Figure 4 Stress-Strain and Strain-Temperature Results for Austenitic deformation on aNiTiCu SMA
24
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120
Temperature, °C
memaust :120°C
0
100
200
300
400
500
600
700
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Strain
Figure 5 Stress-Strain and Strain-Temperature Results for Austenitic
deformation on a NiTi SMA
-0.008
-0.007
-0.006
-0.005
-0.004
-0.003
-0.002
-0.001
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18
Plastic Strain
Tw
o-W
ay S
trai
n
NiTiCu Specimen
NiTi Specimen-
Figure 6 Two-way Strain vs. Plastic Strain for Austenitic
Deformation on NiTiCu and NiTi specimens
25
0.13
0.131
0.132
0.133
0.134
0.135
0.136
0 10 20 30 40 50 60 70 80 90 100 110 120 130
Temperature, °C
Stra
in
0.5
0.6
0.7
0.8
0.9
1
1.1
Hea
t F
low
, W/g
(en
do u
p)
strain
DSC-Austenitic Series
DSC- Pre-Test
Figure 7 Initial and Final DSC and Two-Way Strain for NiTiCu Specimen deformed in
Austenitic Condition
0.15
0.152
0.154
0.156
0.158
0.16
-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100
Temperature, °C
Stra
in
0.35
0.45
0.55
0.65
0.75
0.85
0.95
Hea
t F
low
, W/m
g (e
ndo
up)
strainPost Test DSCPre-Test DSC
Figure 8 Two-Way strain and Initial and Final DSC for NiTi Specimen Deformed in
Austenitic Condition
26
Table 1 Heat of Transformation values for each loading history, ∆∆H (J/g)
Pre-Test
Condition
Austenitic
Plastic
Loading
Thermal
Cycling
under stress
Stress-Induced
Martensitic
Loading
Martensitic
Plastic Loading
M⇒⇒A A⇒⇒M M⇒⇒A A⇒⇒M M⇒⇒A A⇒⇒M M⇒⇒A A⇒⇒M M⇒⇒A A⇒⇒M
NiTiCu 12.41 -12.60 5.94 -6.45 5.79 -5.36 3.69 -3.61 3.23 -3.20
NiTi 15.56 -15.38 9.18 -8.36 8.62 -4.39 6.29 -2.07 3.63 -1.77
ActTrans- SeriesT=85 °C
0
100
200
300
400
500
600
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Strain
Summary-Acta Tranformation
00.010.020.030.040.050.060.070.080.090.1
0.110.120.130.140.150.160.17
-30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Temperature, °C
Figure 9 Stress-Strain and Strain-Temperature Results for Stress-Induced MartensiticDeformation on a NiTiCu SMA
27
memtrans- :40°C
0
50
100
150
200
250
300
350
400
450
500
550
600
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24
Strain
Stre
ss, M
Pa
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
-70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110Temperature, °C
Stra
inFigure 10 Stress-Strain and Strain-Temperature Results for Stress-Induced Martensitic
Deformation on a NiTi SMA
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15
Plastic Strain
2-w
ay S
trai
n
NiTiCu Specimen
NiTi Specimen
Figure 11 Two-way strain development for Stress-Induced Martensitic Loading as a
function of Plastic Strain
28
0.12
0.125
0.13
0.135
0.14
0.145
0.15
0.155
0.16
-30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Temperature, °C
Stra
in
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
Hea
t F
low
, W/m
g (e
ndo
up)
strainPost-Test DSCPre-test DSC
Figure 12 Initial and Final DSC and Two-Way Strain for NiTiCu Specimen through the
Stress-Induced Martenitic Loading
0.13
0.135
0.14
0.145
0.15
0.155
0.16
0.165
0.17
0.175
-60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110
Temperature, °C
Stra
in
0.5
0.6
0.7
0.8
0.9
1
1.1
Hea
t F
low
, W/g
(en
do u
p)
Strain
Post-Test DSC
Pre-Test DSC
Figure 13 Initial and Final DSC and Two-Way Strain for NiTi Specimen through the
Stress-Induced Martenitic Loading
29
Martensitic-plastic strainT=22°C
-100
0
100
200
300
400
500
600
700
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 0.22 0.24 0.26 0.28
Strain
Stre
ss, M
Pa
Acta-Martensite
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Temperature, °C
Stra
in
Figure 14 Stress-Strain and Strain-Temperature Results for Martensitic Deformation on aNiTiCu SMA
-0.002
0.018
0.038
0.058
0.078
0.098
0.118
0.138
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120
Temperature, °C
Stra
in
0
100
200
300
400
500
600
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Strain
Stre
ss, M
Pa
Figure 15 Stress-Strain and Strain-Temperature Results for Martensitic Deformation on aNiTi SMA
30
Martensitic 2way strain
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
0.045
0.05
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16
Plastic Strain
Tw
o-w
ay s
trai
n NiTiCu Specimen
NiTi Specimen-
Figure 16 Two-way strain development for the Martensitic Loading as a function of Plastic
Strain
31
0.14
0.145
0.15
0.155
0.16
0.165
0.17
0.175
0.18
0.185
-40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Temperature, °C
Stra
in
0.5
0.6
0.7
0.8
0.9
1
1.1
Hea
t F
low
, W/m
g (e
ndo
up)
Strain
Post-Test DSC
Pre-Test DSC
Figure 17 Initial and Final DSC and Two-Way Strain for NiTiCu Specimen through a
Martensitic Loading
32
-0.005
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100
Temperature, °C
Stra
in
0.5
0.6
0.7
0.8
0.9
1
1.1
Hea
t F
low
, W/m
g (e
ndo
up)
Strain
Post-Test DSC
Pre-Test DSC
Figure 18 Initial and Final DSC and Two-Way Strain for NiTi Specimen through a
Martensitic Loading
33
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170
Temperature, °C
Stra
in
Figure 19 Thermal Cycles under Constant 200 MPa for NiTiCu
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.1
-30 -20 -10 0 10 20 30 40 50 60 70 80
Temperature, °C
Stra
in
Figure 20 Thermal Cycles under Constant Stress for NiTi
34
0
0.005
0.01
0.015
0.02
0.025
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160
Temperature, °C
Stra
in
0.5
0.6
0.7
0.8
0.9
1
1.1
Hea
t F
low
, W/m
g (e
ndo
up)
6 Mpa
Post-Train DSC
Pre-Test DSC
Figure 21 Two-way strain and Initial and Final DSCs for thermally cycled NiTiCu
Specimen
-0.005
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
-80 -70 -60 -50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110
Temperature, °C
Stra
in
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
Hea
t F
low
, W/m
g (e
ndo
up)
Strain
Pre-Test DSCPost-Test DSC
Figure 22 Two-way strain and initial and final DSC results for thermally cycled NiTi
Specimen
35
Figure 23 Microstructure showing residual martensite after heating above austenite finish
Figure 24 Recoverable strain as a function of the plastic strain for the NiTi Specimen