Investigation of Reversible Transformation Strain Variation in Shape Memory Alloys

ABSTRACT

Planar anisotropy of shape memory strain in polycrystalline Ti-55.4wt%Ni materials was studied by investigating the temperature induced martensite transformation in TiNi and the mechanism of SME degradation during the thermo-mechanical cycling. In the experiment, constant-load tests were done on ten specimens which were spark-cut from the TiNi sheet in various angles to the rolling direction (RD).


It was observed that the reversible transformation strain was the largest when the angle to RD is 20°, hence having the maximum SME.  The inverse relationship between the transformation strain and the angle to RD was also studied.

INTRODUCTION

Smart structures are characterized by the ability to sense and respond to external stimuli in an appropriate manner and on an appropriate time scale.  The shape memory alloys (SMAs) are candidate materials for the actuators of these smart structures since they have the ability to change properties and are able to function in a controlled response to a change in environment or operating conditions.


Shape memory alloys are the metal compounds which have the capability to sustain and recover relatively large strains without undergoing plastic deformation [1].  This ability to fully recover large strains is a result of a reversible thermo-elastic martensite transformation which means that the martensite fraction in the SMA will change with rising and falling temperature.  The austenite-to-martensite transformation starts at M_s (martensite start) and on further cooling to a temperature M_f (martensite finish), the transformation is complete and the specimen is said to be in the martensitic state.  This austenite-to-martensite transformation temperature M_s depends on the composition of the SMA and the previous thermo-mechanical treatments.  In order to totally recover the large strain which occurs during the martensite transformation, the SMA is heated above the martensite-to-austenite transformation temperature A_s (austenite starts) and the martensite-to-austenite transformation will end at A_f (austenite finish).  This strain recovery phenomenon is usually referred as the shape memory effect (SME) [2].  The reversible thermo-elastic martensite transformation can be induced by variation of either temperature or stress.  In this  study, we use the temperature induced martensitic transformation such that the specimens were thermally cycled under a constant load, hence keeping the stress constant throughout the experiment.


SMAs exhibit good shape memory and mechanical properties in cryogenic temperature and high temperature.  Therefore, Titanium-Nickel (TiNi) which is also referred as Nitinol (Nickel Titanium Naval Ordnance Laboratory, named after the laboratory where it was discovered and first extensively studied in the early 60s) was used in our experiment as it has satisfactory performance in extreme temperatures.  Another reason for the selection of TiNi for the experiment is that TiNi is vastly used for actuators compared to other shape memory alloys, such as CuZnAl and CuAlNi.  The active control is facilitated by electrical heating and since TiNi is easier as compared to Cu based alloys to be processed into fine wires and also possesses higher electrical resistivity, it is a very good choice for actuators.


This project investigated the temperature induced martensite transformation in TiNi and aimed to study the mechanism of SME degradation during the thermal mechanical cycling.  The definite orientation relationship between the crystal lattices of the austenite and martensite leads to the fact that the SME  depends on the crystallographic direction.  Therefore, the SME of a polycrystalline TiNi alloy sheet is expected to be texture-related.  Hence, the focus of this paper is to understand the planar anisotropy of shape memory strain in the polycrystalline TiNi alloys and thus the crystallographic direction in which SME is maximum is studied.

EXPERIMENTS

The materials used in the present experiment are Ti-55.4wt%Ni specimens with a thickness of 1mm.  The TiNi sheet was made by vacuum arc melting, followed by hot forging and hot rolling.  In the shape memory processing of the TiNi sheet, the sheet was held at 800°C for 60min and cooled in the furnace.


Ten specimens (figure 1) for the thermo-mechanical test were spark-cut from the TiNi sheet in direction 'Angle to RD'= 0°, 10°, 20°, 30°, 40°, 50°, 60°, 70°, 80° and 90° to the rolling direction (RD).   The specimens were submitted to traction under a constant load of 3kN by using a computer-controlled tensile machine.  In  the experiment, the specimens were initially subjected to a cooling rate of 2°C/min and were then heated back at a heating rate of  2°C/min (figure 2).  The heating and cooling rates were achieved using a temperature chamber.  The strain was measured by an extensometer clamped on the specimen.

RESULTS AND DISCUSSION

Figures 3a to 3d show the experimental results for the specimen with 'Angle to RD'= 40°.  Similar results for the other nine specimens were also obtained.
 
As shown in figures 3a and 3b, the force was kept at a constant value and the temperature was varied throughout the experiment.  This shows that the experiment was correctly done since our present study is on the temperature induced martensite transformation in TiNi under constant loading.  The times and temperatures for the starting and ending stages of the martensite transformation can be seen from the strain-time and strain-temperature graphs in figures 3c and 3d.  Similarly, the times and temperatures for the starting and ending stages of the austenite transformation can also be seen from the two figures. It can be seen from figure 3d that the specimen is elongating on cooling, whereas on heating strain recovers due to the reverse martensite-to-austenite transformation, hence showing SME.


During the experiment, sometimes a permanent elongation was built up during the thermal cycle.  This permanent strain, E_perm (figure 4) is defined using the length difference between the austenite after and before the transformation.  There is still doubt about the origin of the permanent strain during transformation under a constant load.  The experimentally observed permanent strain may be due to the formation of dislocations necessary to accommodate the shape changes.  Since no mechanical system is ideal, the possible formation of dislocations is also an indication of dissipation of energy in the transformation system.  In addition to the defect production, the dissipation of energy is also due to interfacial friction and acoustic emission caused by nucleation and growth of martensite plates and interactions between them during the transformation.
 
From figure 4, it can be seen that two types of transformation strains are important in the study of  SME.  One is the transformation strain for the austenite-to-martensite transformation, E_AMand the other is the reversible transformation strain for the martensite-to-austenite transformation, E_MA.


In order to understand the planar anisotropy of shape memory strain in the polycrystalline TiNi alloys, the orientation dependence of the reversible transformation strain was plotted in figure 5.
 
Figure 5 shows that the largest reversible transformation strain occurs at 'Angle to RD'=20°.  A general inverse relationship between the transformation strain and the 'Angle to RD' can be seen, that is the transformation strain decreases as 'Angle to RD' increases.

CONCLUSION AND RECOMMENDATION

In order to investigate the planar anisotropy of shape memory strain in the polycrystalline TiNi alloys, constant-load tests were done on ten specimens which were spark-cut from the TiNi sheet in various  angles to the rolling direction (RD).


It was found that the specimen in which 'Angle to RD'=20° had the largest reversible transformation strain, hence having the maximum SME. In addition, a general inverse relationship between the transformation strain and the 'Angle to RD' was found.  This means that SME decreases with increasing 'Angle to RD'.


The strain gauge must be properly clamped to the specimen throughout the experiment.  Hence if the clamping became loose at any one time in the experiment, a loss of accuracy would occur.  In addition, the deformation constraint during the transformation may also affect the accuracy.


Further experimental work is needed to clarify the cause of permanent strains so as to have a deeper understanding of the mechanism of SME degradation during the thermo-mechanical cycling.

REFERENCES

1) H.T. Banks, R,C. Smith and Y.Wang  (1996).  Smart material structures.  Paris:  John Wiley and Sons Ltd.
2) P.F. Gobin and J. Tatibouet (Ed.).  (1996).  Third international conference on intelligent materials.  United States of America:  The International Society for Optical Engineering.
3) G.K. Haritos and A.V. Srinivasan (Ed.).  (1991).  Smart structures and materials.  United States of America: The American Society of Mechanical Engineers.
4) C.A. Rogers, and R.C. Rogers  (1992).  Recent advances in adaptive & sensory materials & their application.  United States of America: Technomic Publishing Co. inc.
5) L.C. Brinson and B. Moran (Ed.).  (1994).  Mechanics of phase transformations and shape memory alloys.  United States of America:  The American Society of Mechanical Engineers.