You are on page 1of 6

Effect of aging and solution annealing on transformation and

deformation mechanism of super-elastic Ni50.9%-Ti alloy in nano- scale
A. Ghabchi
a,1,2,3
, J. Khalil-Allafi
1
, X.W. Liu
4
, O. 6|GHUEHUJ
4
, E. Turunen
2
, S-P Hannula
4
1
Sahand University of Technology, Department of Materials Engineering, Tabriz, Iran
2
Technical Research Center of Finland (VTT), Espoo, Finland
3
Center for Thermal Spray Research, Material Science and Engineering Department, SUNY, Stony Brook, USA
4
Department of Materials Science and Engineering, Helsinki University of Technology, Espoo, Finland
Abstract. Effects of aging and solution annealing treatment on the super elastic respond of the cold-worked Ni50.9%-Ti thin
wire was studied in nano-scale utilizing instrumented nanoindentation. The nanoindentation results were coupled with the
tensile test results to provide better insight to the governing deformation mechanisms. The first applied heat treatment
consisted oI solution annealing treatment (850´C Ior 60 minutes) Iollowed by aging at 450´C Ior 30 minutes and as second
heat treatment the cold-worked wire was aged at 450´C Ior 30 minutes with no prior solution annealing treatment. Both
tensile test and nanoindentation results showed better super elastic behavior for the wire subjected to second treatment.
Indentation displacement bursts were observed in loading course after both treatments. However, slope changes in unloading
course of indents were recorded after second heat treatment which might be due to reverse phase transformation. Surface
asperities on the post-indented surface after first treatment were observed far away from indentation stress field by atomic
force microscopy.
1. Introduction
Micro- and nanoelectromechanical (MEMS and NEMS) systems are gaining great interest in wide variety of industries
as actuators and sensors. Nevertheless, for any materials developed for this purpose there is a question of reliability in
operation and manufacturing while the size of the material is decreasing to sub-micron or nano-scale. Another
deficiency which rises from decreasing the size scale is the amount of output work. Application of martensitic materials
can be a solution to overcome such intrinsic issues. Krulevitch et al. demonstrated that a NiTi actuator has the highest
work output among other actuator materials [1]. Thus, the martensitic transformation can be exploited as stress-induced
in super-elasticity and as thermally induced in one way or two way shape memory effect, provides more flexibility in
designing smart MEMS [2]. Characterization of stress induced martensitic (SIM) transformation in nanoscale is a
challenging task to accomplish. One potential technique to understand the SIM in nanoscale is nanoindentation.
Nanoindentation returns the continuous indentation load-displacement response of a material with resolution of micro-
newton in load and nano-meter in depth. This technique provides a unique tool for studying the local mechanical
properties as well as insight into the associated deformation mechanism in that scale [3]. The nanoindentation technique
has been used for characterization of martensitic materials [4-6], but information of studies with bulk material is
limited. However, even for a simple material structure, interpretation of load-displacement response is not a trivial task
due to the complexity of stress field in the volume of material and the possible different deformation mechanisms, not
to mention the complexity of the crystal structure or the effect of the possible special surface layers, such as oxides.
In this study we address the deformation mechanisms in Ni-rich NiTi wires under tensile tests and
nanoindentation. As the tensile test samples the whole bulk volume, nanoindentation returns the response of material
only from few hundreds nanometers volume of material. Different phases in the studied material were obtained by
applying different heat treatments in order to study their influence on deformation mechanisms. The surfaces, before
and after nanoindentation were studied by atomic force microscopy (AFM).
a
e-mail: Arash.Ghabchi@vtt.fi
ESOMAT 2009, 02009 (2009)
DOI:10.1051/esomat/200902009
© Owned by the authors, published by EDP Sciences, 2009
This is an Open Access article distributed under the terms of the Creative Commons Attribution-Noncommercial License
(http://creativecommons.org/licenses/by-nc/3.0/), which permits unrestricted use, distribution, and reproduction in any
noncommercial medium, provided the original work is properly cited.
Article available at http://www.esomat.org or http://dx.doi.org/10.1051/esomat/200902009
2. Experimental
2.1 Materials and heat treatments
The thin wires (diam. 0.35 mm) of the nickel-rich nickel-titanium containing 50.9 at% Ni (Memry Corp. USA) were
used in this study. During the manufacturing (drawing) the wire was subjected to a ~40% cold-work with 5%
increments with no heat treatment in final step. On as-reseived material two different heat treatments were applied in
order to obtain certain crystallographic structures in the samples to be studied. The first set of wires were solution
annealed at 850´C Ior 60 minutes Iollowed by aging treatment at 450´C Ior 30 minutes (these samples are later
addressed with SA). For the second set only aging treatment at 450´C Ior 30 minutes was carried out without any prior
solution annealing treatment (samples marked with A). In both cases, after each heat treatment, samples were quenched
in 25´C water. Solution annealing treatment was conducted in a vacuum quartz tube in the presence of titanium powders
in order to minimize the effect of oxidation. Aging treatment was carried out in argon atmosphere.
2.2 Differential scanning calorimetry (DSC)
The transformation behavior during heating and cooling was investigated using a differential scanning calorimeter
Netzsch DSC 200F3. The DSC samples with masses between 20 and 40 mg were heated up to 100´C where they were
held for three minutes to establish thermal equilibrium. Then the DSC measurement started by cooling down to -100´C
in 10´C/min. At -100´C the samples were again held Ior three minutes and then heated up to 100´C with a heating rate
oI 10´C/min.
2.3 Tensile tests
Tensile test samples were cut to 90 mm length and the extensometer gage length was set to 70 mm. To evaluate the
super-elasticity in loading/unloading and the ability of the wires to recover deformation in macroscale, tensile tests were
carried out at room temperature (27´C) up to 7 % strain (Zwick/Roell FR010). To prevent effect of heating the samples
due to phase transformations during the test, the strain rate was set to 0.2 mm/min (e.g. static condition).
2.4 Nanoindentation
The samples for the nanoindentation were made by mounting and mechanical polishing of cross section of wires with
0.5µm diamond paste. The mechanical polishing was done with applying very low load to minimize the surIace plastic
deformation and keeping for longer duration for removing the deformed layer as much as possible. Hysitron
TriboindenterII, a nanoindentation instrument (USA) with Berkovich diamond tip (E
i
= 1140 GPa, ȣ
i
= 0.07, tip radius
R = 230 nm) was used for the nanoindentation tests in which the load-displacement response of the sample was
recorded. All indentation tests were carried out in load-controlled condition. Four indents with maximum load of 40, 80,
100 and 300 µN were made with 15 seconds elapsing time at the maximum load. Nanoindentation instrument was
equipped with atomic force microscope (AFM) that enables us to study the surface before and after indentation with
high resolution.
3. Results and Discussions
3.1 Phase transformations
Fig. 1 depicts the DSC graphs of both (A) and (SA) samples. The A sample (Fig. 1b) that was aged exhibits apparently
only one peak during cooling and two peaks during heating. M
s
= 37
o
C, M
f
= 15
o
C, A
s
= 22
o
C and A
f
= 40
o
C. The
solution annealed and aged, SA, sample (Fig. 1a) shows three-step transformation in the cooling regime (M
s
= -22.5
o
C,
M
f
= -43
o
C, R
s
= 37.5
o
C and R
f
= 23
o
C). In a binary NiTi alloy, two martensitic phases can be formed on cooling: the
first transformation from the cubic high temperature phase (B2) to the rhombohedral R phase, while further cooling
results in transformation B2+R B19´ [7,8]. The first R-phase transition is associated with 1 % transformation strain,
while formation oI the B19´ is associated with 10 % transformation strain. The possible precipitates and dislocations in
the structure promote the R-phase as they act as barriers to Iormation oI B19´ |9]. In sample SA solution annealing has
led to the recrystallization of the material and releasing the dislocation structure formed during drawing of the wire, and
with applying secondary heat treatment (aging) the precipitates will be introduced into the microstructure and
consequently R-phase formation will be favored. However, in Sample A the aging temperature has not been high
enough to cause the recrystallization and releasing the dislocation structure, and the observed one transformation peak is
ESOMAT 2009
02009-p.2
related to the B2 R and B2 B19´ transIormations that the peaks are overlapped. This explanation is plausible due
to presence of two peaks in heating of the A Sample.
a) b)
Fig.1. DSC curves for the wire with a) the SA treatment, and b) the A treatment.
Fig. 2 shows the optical microstructure of the solution annealed SA Sample. The grains are not any more oriented
according to the deformation direction of drawing as in the A Sample and the grain size is significantly larger than as-
received wire (the microstructure is not presented here due to limitation of optical image that can not show the fine
grains in that range).
Fig.2. Microstructures of the SA Sample.
3.2 Tensile behavior
Figure 3 shows the tensile test results of SA and A samples at room temperature. The SA sample at testing temperature
has a microstructure mixture of austenite and R-phase. Due to solution annealing which decreases the dislocation
density the critical shear stress for slip is low. However, two deformation mechanisms might be considered; i)
deformation due to slip and ii) deformation due to twinning. In the loading there is a linear elastic region followed by a
nonlinear region right before the starting of the plateau. The nonlinearity might be due to austenite to R-phase
transformation. In the transition from nonlinear to plateau there was no stress-peak observed which can be an indication
of different stress-induced transformations such as B2Æ B19´and RÆ B19´. Also, increasing the slope oI plateau can
be an indication of deformation by slip mechanism and there is no distinct region indicating martensite plasticity. The
scenario seems more plausible while in unloading we have partial recovery. The mechanism responsible for the
recovered portion of deformation is reverse transformation of twinned martensite. The residual strain can be an
indication of deformation mechanism associated with slip or can be related to the presence of R-phase in the testing
temperature. The aged sample at testing temperature had the microstructure mixture of austenite and martensite. The
stress-strain curve of aged sample can be divided into three regions. In the first region there is elastic deformation of
austenite phase (and some elastic-plastic deformation of existing martensite) followed by a maximum stress peak which
is starting point of austenite to martensite transformation. The last portion of loading part consists of the elastic and/or
plastic deformation of martensite.
ESOMAT 2009
02009-p.3
Fig. 3. The tensile test results of the A (red line) and SA (black line) samples at room temperature.
3.3 Nanomechanical behavior
Figure 4 shows the load-displacement response of both A and SA Samples under different maximum loads.
Disregarding of the sample and maximum applied loads, in all cases discontinuities (sudden increase in depth) during
the loading were observed. Nevertheless, there were no significant discontinuities observed in unloading (except for
100µN A sample). Discontinuities (or 'pop-ins¨) in loading portion are shown by arrows in figure 4. For some of the
conventional metals it has been shown that discontinuities in the loading portion of nanoindentation is an indication of
dislocation nucleation and/or movement [10]. Nevertheless, in this experiment, for the aged sample, A Sample, such
explanation is not valid as the final depth of indent for the indents with maximum load, equal or smaller than 100µN is
close to zero and there is no trace of permanent deformation on the surface of material. In this case we can consider that
observed pop-ins (particularly Ior 40, 80 and 100µN indents Ior 'A¨ sample) are related to stress induced martensitic
transformations (SIM).
a) b)
c) d)
Fig. 4. The load-displacement response in the nanoindentation study of both the A (red) and SA (black) Samples under a) 300µN, b)
100µN, c) 80µN and d) 40µN load.
The full recovery of A Sample was also observed in tensile test. Due to full recovery of indents in A Sample (40, 80
and 100µN indents) reverse transformation must be happened. However, we did not observe any significant
ESOMAT 2009
02009-p.4
discontinuity in the unloading portion as Frick et al. [11] reported. But for the A Sample sudden slope change of
unloading curve was observed in 100 and 80 µN maxmimum loads. We hypothesize the sudden slope change of
unloading portion is an indication of reverse martensitic transformation. This explanation is plausible while there is not
any known phenomena related to reverse dislocation movement upon unloading. By comparing the load-displacement
response for both samples, it is clear that the number of pop-ins in the A Sample is more than that of SA Sample. This is
in agreement with the tensile test behavior of these materials. As the SA Sample did not show clear superelasticity
because of the presence of R-phase in the material at the beginning of the tests which restricts partially the stress-
induced martensitic transIormation. Also due to the large grain size (~40µm) one grain was indented in this material.
The A sample had signiIicantly Iiner grains (··1µm) and with increasing the indent load more grains undergo stress
induced martensitic transformation. Figure 5 shows the atomic force microscope (AFM) image of the A sample before
and after nanoindentation. No indent mark was observed Ior indents with 40, 80 and 100µN maximum load. The only
visible indent mark is made by indent with 300µN load.
In figure 5b, for the SA Sample in addition to the indent marks also a surface phenomena was observed as some
surface asperities after nanoindentation (Fig. 5b). This might be related to permanent martensites which were formed
during the nanoindentation and due to very smooth surface of sample (less than 10nm) became visible under AFM.
However, such phenomenon was not observed in all indents. We assume the formation of asperities strongly depends on
the location of indent. Another surface related observation was reported for shape memory CuAlNi which was very
limited to the vicinity of nanoindentation mark [12]. To deeply understand the origin of this phenomenon, further
studies are needed with having the knowledge of indenting spot and crystallographic orientation. Additionally, a
nanoindentation stage with capability of heating can be used to verify the martensitic transformation which can be
reverted upon heating.
Fig. 5. AFM image of the A Sample a) before and b) after nanoindentation test showing nearly full recovery of the indentation marks
except Ior 300µN indent: and the SA Sample c) beIore and d) aIter nanoindentation presenting surface asperities after the indentation
test.
Fig.6a shows the schematic diagram for the indentation of NiTi shape memory alloys which illustrates different
information given by the indentation curve, such as hardness, energy dissipation and energy recovery (adopted from
Gall et al. [13].) The recoverable energy, E
r
, will be higher if a significant reverse transformation or lattice relaxation
takes place upon unloading. Fig. 7b shows the calculated E
r
values (percentile) for the A and SA Samples at different
maximum loads. With the loads 300, 100, and 80 µN the maximum recoverable energy is obtained for the A Sample,
while recoverable energy is clearly lower for the SA sample. Observed higher recoverable energy in the A Sample is
suggested to be related to the larger amount of the recovered stress-induced martensite in this material during
unloading. Additionally, immobilized dislocations by the small grain size, possible residual deformation structure from
the wire drawing, and perhaps also by the precipitates in the structure, result in large elastic strain which is recoverable.
In the SA sample the dislocation structure is minimized and the grain size is larger, due to solution annealing, both
minimizing the barrier for dislocation movement and resulting in permanent plastic strain. Furhermore, as the amount of
austenite for SIM transformation in this material was less than in the A Sample, it may be that also the amount of
recovering SIM is less. In addition, the high dislocation activity stabilizes the martensite plates and reducing further the
E
r
in SA Sample. The recoverable energy for both samples increased with decreasing the maximum indentation load
ESOMAT 2009
02009-p.5
Irom 300 to 80 µN. As the applied stress decreases, the possible permanent slip deIormation oI the structure should
decrease. Therefore, this behaviour may indicate that the main part of the observed recoverable energy originates from
the reverse SIM transIormation. However, by Iurther decreasing the maximum load to 40 µN, the E
r
values for both the
A and SA Samples decrease. In the A Sample, recoverable energy decreases from 72.6 % to 60 %, while the decrease in
the SA Sample is only couple of percentages. This may be related to the higher stress level needed for the SIM
formation in the A sample, e.g. with the lower applied stress, the amount of the SIM to be formed is less, and
consequently, the recoverable energy related to the reverse transformation is also smaller.
a) b)
Fig.6. a) Information obtained from the indentation curve according to Gall et al. [13]. b) Recoverably energy (E
r
) for the A (dashed)
and SA (black) Samples with applied maximum indentation loads oI 300, 100, 80 and 40 µN.
4. Conclusions
The difference in tensile behavior of SA and A Samples was addressed by critical slip stress and presence of different
dominant microstructures such as austenite and R-phase which leads to improved and poor superelasticty respectively.
Discontinuities in loading and sudden slope change in unloading portion of P-h curves were attributed to the SIM and
reverse SIM transformations respectively. Recoverable energy was measured as a mean to ability of material to undergo
the reverse SIM transformation. A new surface related asperities phenomenon was reported which was visualized by
AFM imaging before and after nanoindentation.
Acknowledgements
AG is grateful for constructive discussions with Professor Gouldstone from Northeastern University, Department of
Mechanical, Industrial and Manufacturing Engineering, and Professor Frick from Wyoming University, Department of
Mechanical Engineering.
References
[1] P. Krulevitch et al., J. MEMS 5, 271 (1996).
[2] K. Bhattacharya, R.D. James. Science 307, 53 (2005).
[3] A. Gouldstone, H.-J. Koh, K.-Y. Zeng, A.E. Giannakopoulos, S. Suresh, Acta Mater. 48, 2277 (2000).
[4] G.A. Shaw, W.C. Crone, Mat. Res. Soc. Symp. Proc. 791, 215, (2003).
[5] K.S.S.E. Raju, S, Bysakh, M.A. Sumesh, S.V. Kamat, S. Mohan, Mat. Sci. Eng. A, 476, 267 (2008).
[6] A. Kumar, D. Singh, D. Kaur, Surf. Coat. Techn. 203, 1596 (2009).
[7] J. Khalil Allafi, X. Ren, G. Eggeler, Acta Mater. 50, 793 (2002).
[8] M.C. Carroll, Ch. Somsen, G. Eggeler, Scripta Mater. 50, 187 (2004).
[9] J. Khalil-allafi, A. Dlouhy, G. Eggeler, Acta Mater. 50. 4255 (2002)
[10] T.F. Page, W.C. Oliver, C.J. McHargue, J Mater Res 7, 450 (1992).
[11] C.P. Frick, T.W. Lang, K. Spark, K. Gall, Acta Mater. 54, 2223 (2006).
[12] W.C. Crone, H. Brock, A. Creuziger, Exp. Mech. 47, 133 (2007).
[13] K. Gall, K. Juntunen, H.J. Maier, H. Sehitoglu, Y.I. Chumlyakov, Acta Mater. 49, 3205 (2001).
ESOMAT 2009
02009-p.6

ESOMAT 2009 2. Results and Discussions 3. 100 equipped with atomic force microscope (AFM) that enables us to study the surface before and after indentation with high resolution. the strain rate was set to 0.g. 1a) shows three-step transformation in the cooling regime (Ms = -22. During the manufacturing (drawing) the wire was subjected to a ~40% cold-work with 5% increments with no heat treatment in final step. a nanoindentation instrument (USA) with Berkovich diamond tip (Ei = 1140 GPa. Mf = 15oC. and with applying secondary heat treatment (aging) the precipitates will be introduced into the microstructure and consequently R-phase formation will be favored. On as-reseived material two different heat treatments were applied in order to obtain certain crystallographic structures in the samples to be studied. 1 depicts the DSC graphs of both (A) and (SA) samples.5oC. static condition). The possible precipitates and dislocations in the structure promote the R-phase as they 9]. sample (Fig. The solution annealed and aged. after each heat treatment. tensile tests were % strain (Zwick/Roell FR010). results in transformation B2+R while f ociated with 10 % transformation strain.2 Differential scanning calorimetry (DSC) The transformation behavior during heating and cooling was investigated using a differential scanning calorimeter Netzsch DSC 200F3. In a binary NiTi alloy. Experimental 2. Then the DSC measurement started by cooling down to s were eating rate 2. in Sample A the aging temperature has not been high enough to cause the recrystallization and releasing the dislocation structure. 2. USA) were used in this study.8]. To prevent effect of heating the samples due to phase transformations during the test. The first set of wires were solution annealed these samples are later addressed with SA). and the observed one transformation peak is 02009-p. M s = 37oC.1 Materials and heat treatments The thin wires (diam. In both cases.2 mm/min (e. samples were quenched vacuum quartz tube in the presence of titanium powders in order to minimize the effect of oxidation. To evaluate the super-elasticity in loading/unloading and the ability of the wires to recover deformation in macroscale. Hysitron TriboindenterII.3 Tensile tests Tensile test samples were cut to 90 mm length and the extensometer gage length was set to 70 mm. 1b) that was aged exhibits apparently only one peak during cooling and two peaks during heating.1 Phase transformations Fig. Four indents with maximum load of 40. All indentation tests were carried out in load-controlled condition. Rs = 37.9 at% Ni (Memry Corp. two martensitic phases can be formed on cooling: the first transformation from the cubic high temperature phase (B2) to the rhombohedral R phase. SA. However. while further cooling [7.2 . 80.5oC and Rf = 23oC).07. 2.35 mm) of the nickel-rich nickel-titanium containing 50. Aging treatment was carried out in argon atmosphere. Mf = -43oC. The DSC samples with masses between 20 and 40 mg were held for three minutes to establish thermal equilibrium. For the second set only was carried out without any prior solution annealing treatment (samples marked with A). In sample SA solution annealing has led to the recrystallization of the material and releasing the dislocation structure formed during drawing of the wire.4 Nanoindentation The samples for the nanoindentation were made by mounting and mechanical polishing of cross section of wires with deformation and keeping for longer duration for removing the deformed layer as much as possible. The A sample (Fig. tip radius R = 230 nm) was used for the nanoindentation tests in which the load-displacement response of the sample was recorded. i = 0. As = 22oC and Af = 40oC. The first R-phase transition is associated with 1 % transformation strain. 0. 3.

ESOMAT 2009 related to the B2 R and B2 to presence of two peaks in heating of the A Sample. The scenario seems more plausible while in unloading we have partial recovery. The mechanism responsible for the recovered portion of deformation is reverse transformation of twinned martensite. 3. Fig. Due to solution annealing which decreases the dislocation density the critical shear stress for slip is low. two deformation mechanisms might be considered. The stress-strain curve of aged sample can be divided into three regions. However. The aged sample at testing temperature had the microstructure mixture of austenite and martensite. i) deformation due to slip and ii) deformation due to twinning.1. The last portion of loading part consists of the elastic and/or plastic deformation of martensite. The SA sample at testing temperature has a microstructure mixture of austenite and R-phase.2 Tensile behavior Figure 3 shows the tensile test results of SA and A samples at room temperature. In the transition from nonlinear to plateau there was no stress-peak observed which can be an indication of different stress-induced transformations such as B2 be an indication of deformation by slip mechanism and there is no distinct region indicating martensite plasticity. In the loading there is a linear elastic region followed by a nonlinear region right before the starting of the plateau. The residual strain can be an indication of deformation mechanism associated with slip or can be related to the presence of R-phase in the testing temperature. This explanation is plausible due a) b) Fig.3 . Fig. and b) the A treatment.2. The grains are not any more oriented according to the deformation direction of drawing as in the A Sample and the grain size is significantly larger than asreceived wire (the microstructure is not presented here due to limitation of optical image that can not show the fine grains in that range). 02009-p. DSC curves for the wire with a) the SA treatment. Microstructures of the SA Sample. The nonlinearity might be due to austenite to R-phase transformation. 2 shows the optical microstructure of the solution annealed SA Sample. In the first region there is elastic deformation of austenite phase (and some elastic-plastic deformation of existing martensite) followed by a maximum stress peak which is starting point of austenite to martensite transformation.

in this experiment. 3. we did not observe any significant 02009-p. The full recovery of A Sample was also observed in tensile test. However. in all cases discontinuities (sudden increase in depth) during the loading were observed. Due to full recovery of indents in A Sample (40.4 .3 Nanomechanical behavior Figure 4 shows the load-displacement response of both A and SA Samples under different maximum loads. there were no significant discontinuities observed in unloading (except for pop-ins ) in loading portion are shown by arrows in figure 4. 4. such explanation is not valid as the final depth of indent for the indents with maximum load close to zero and there is no trace of permanent deformation on the surface of material. Nevertheless. The tensile test results of the A (red line) and SA (black line) samples at room temperature. Nevertheless. The load-displacement response in the nanoindentation study of both the A (red) and SA (black) Samples under load. Disregarding of the sample and maximum applied loads. A Sample. For some of the conventional metals it has been shown that discontinuities in the loading portion of nanoindentation is an indication of dislocation nucleation and/or movement [10]. 80 s) reverse transformation must be happened. 3. for the aged sample.ESOMAT 2009 Fig. a) b) c) d) Fig. In this case we can consider that observed popduced martensitic transformations (SIM).

for the SA Sample in addition to the indent marks also a surface phenomena was observed as some surface asperities after nanoindentation (Fig. Furhermore. will be higher if a significant reverse transformation or lattice relaxation takes place upon unloading. and perhaps also by the precipitates in the structure. [11] reported. as the amount of austenite for SIM transformation in this material was less than in the A Sample. Additionally. while recoverable energy is clearly lower for the SA sample. The recoverable energy for both samples increased with decreasing the maximum indentation load 02009-p.6a shows the schematic diagram for the indentation of NiTi shape memory alloys which illustrates different information given by the indentation curve. result in large elastic strain which is recoverable. In addition. Fig. the high dislocation activity stabilizes the martensite plates and reducing further the Er in SA Sample. due to solution annealing. it is clear that the number of pop-ins in the A Sample is more than that of SA Sample. This might be related to permanent martensites which were formed during the nanoindentation and due to very smooth surface of sample (less than 10nm) became visible under AFM. Fig. 5. both minimizing the barrier for dislocation movement and resulting in permanent plastic strain. Figure 5 shows the atomic force microscope (AFM) image of the A sample before and after nanoindentation. Er. As the SA Sample did not show clear superelasticity because of the presence of R-phase in the material at the beginning of the tests which restricts partially the stress. In the SA sample the dislocation structure is minimized and the grain size is larger. immobilized dislocations by the small grain size. By comparing the load-displacement response for both samples. We assume the formation of asperities strongly depends on the location of indent. AFM image of the A Sample a) before and b) after nanoindentation test showing nearly full recovery of the indentation marks nanoindentation presenting surface asperities after the indentation test. [13]. This is in agreement with the tensile test behavior of these materials. This explanation is plausible while there is not any known phenomena related to reverse dislocation movement upon unloading.5 . We hypothesize the sudden slope change of unloading portion is an indication of reverse martensitic transformation. Another surface related observation was reported for shape memory CuAlNi which was very limited to the vicinity of nanoindentation mark [12]. such phenomenon was not observed in all indents. However. 7b shows the calculated E r values (percentile) for the A and SA Samples at different maximum loads. Fig. Wi the maximum recoverable energy is obtained for the A Sample. such as hardness. further studies are needed with having the knowledge of indenting spot and crystallographic orientation. Observed higher recoverable energy in the A Sample is suggested to be related to the larger amount of the recovered stress-induced martensite in this material during unloading.ESOMAT 2009 discontinuity in the unloading portion as Frick et al. The A sample had induced martensitic transformation. it may be that also the amount of recovering SIM is less. But for the A Sample sudden slope change of unloading curve was observed in 100 and 80 maxmimum loads. a nanoindentation stage with capability of heating can be used to verify the martensitic transformation which can be reverted upon heating. possible residual deformation structure from the wire drawing. To deeply understand the origin of this phenomenon. Additionally.) The recoverable energy. energy dissipation and energy recovery (adopted from Gall et al. No indent The only In figure 5b. 5b).

Chumlyakov. References [1] P. 133 (2007). b) Recoverably energy (Er) for the A (dashed) and SA 4.V. 187 (2004).S. Mat. Symp. 02009-p. recoverable energy decreases from 72. 791. A. Brock. [9] J. 1596 (2009). In the A Sample.. A. Department of Mechanical Engineering. Proc. H. 2223 (2006). A new surface related asperities phenomenon was reported which was visualized by AFM imaging before and after nanoindentation. Eggeler. Maier. Gall. G. 271 (1996). Ch. Science 307.P. K.-Y. C. [13]. H. Kumar. the amount of the SIM to be formed is less. a) Information obtained from the indentation curve according to Gall et al. Mohan.ESOMAT 2009 decrease. Gall. and Professor Frick from Wyoming University. Singh. 50. Gouldstone. H.C. 215. Industrial and Manufacturing Engineering. Acta Mater. S. [12] W. J Mater Res 7. Eggeler.A. A. Ren. R. [7] J. Sehitoglu. Surf. Koh.C. and consequently. Khalil-allafi. [11] C. 53 (2005). S.S. Res. M. Exp. the recoverable energy related to the reverse transformation is also smaller.J. W. Khalil Allafi.6 . 203. 267 (2008). [13] K. Acta Mater. Carroll. Crone.W. S. MEMS 5. A. Acknowledgements AG is grateful for constructive discussions with Professor Gouldstone from Northeastern University. Shaw.g. McHargue. Acta Mater. [6] A. 50. [4] G.A. S. Crone. 3205 (2001). Mat. Soc. Dlouhy. 48. Raju. Creuziger. X. Frick. Acta Mater. e. (2003). Sumesh.E.C. K. [5] K. Acta Mater. 4255 (2002) [10] T.I. 54. H. Kaur.D.J. Oliver. Techn.6. Therefore.E. T. Eggeler. 793 (2002). Discontinuities in loading and sudden slope change in unloading portion of P-h curves were attributed to the SIM and reverse SIM transformations respectively. [2] K. Department of Mechanical. Scripta Mater. Recoverable energy was measured as a mean to ability of material to undergo the reverse SIM transformation. Bhattacharya. Juntunen. Zeng. a) b) Fig. J. G. 49. K. this behaviour may indicate that the main part of the observed recoverable energy originates from r values for both the A and SA Samples decrease. Page. 476. [3] A. while the decrease in the SA Sample is only couple of percentages. 47. James. K. Suresh. Y. This may be related to the higher stress level needed for the SIM formation in the A sample. Mech. 50. Conclusions The difference in tensile behavior of SA and A Samples was addressed by critical slip stress and presence of different dominant microstructures such as austenite and R-phase which leads to improved and poor superelasticty respectively. 2277 (2000). with the lower applied stress. Lang. G. Spark. [8] M. W. Somsen.6 % to 60 %. Bysakh. D.F. Krulevitch et al.C. Kamat. Sci. Eng. Coat.-J. 450 (1992). Giannakopoulos. D.