smart shape memory alloys

SHAPE MEMORY ALLOYS: FUNCTIONAL AND SMART

R. STALMANS and J. VAN HUMBEECK

** This text was presented in the seminar "smart materials and technologies - sensors, control systems and regulators, October 1995, Prague, Czech Republik" **

Abstract: Shape memory alloys offer attractive potentials such as: reversible strains of several percent, generation of high recovery stresses, and high power/weight ratios. This text gives an overview of the shape memory functions. The biomedical possibilities are illustrated by recent developments and products. The prospects for use in actuators and smart materials are also highlighted. A table with property values for different classes of shape memory alloys is also included.

Introduction

Shape memory alloys are basically functional materials. They are more important for what they do (= an action) than for what they are (= a material) [Hu92]. The specific shape memory properties are illustrated in Fig. 1 to Fig. 5 by the example of a shape memory spring [St93]. The functions that those alloys can perform, are divided into five main categories:

Free recovery refers to applications in which the single function of the memory element is to cause motions or strains up to 8% (see fig.1: one way memory effect and fig.2: two way memory effect).
Fig.1. The one way memory effect. The sample is deformed (A to B) and unloaded (B to C) at a temperature below Mf. The residual deformation is restored during heating to a temperature above Af.

Fig.2. The two way memory effect. A spontaneous shape change occurs during cooling to a temperature below Mf (A to B). This shape change is recovered during subsequent heating to a temperature above Af (B to C).
Constrained recovery includes applications in which the memory element is prevented from changing shape and thereby generates stresses up to 800 MPa (see fig.3).
Fig.3. The generation of shape recovery stresses. The sample is deformed (A to B) and unloaded (B to C) at a temperature below Mf. Recovery stresses are generated during heating (D to E) starting from the contact temperature Tc (D), situaed between As and Af.
Actuator or work production applications are those in which there is a motion against a bias force and thus work, up to 5 J/g, is done by the shape memory element.
Fig.4. The work output. The sample is deformed at a temperature below Mf (A to B), followed by unloading (B to C) and again loading with a weight W (C to D). Shape recovery occurs at an opposing force W during heating to a temperature above Af (D to E), so work is done.
Superelastic or pseudoelastic applications are isothermal in nature and involve the storage of potential energy (see fig.5).
Fig.5. The superelastic effect. The sample is strongly deformed at relatively low stresses (A to B) at a temperature above Af. During subsequent unloading a complete shape recovery occurs (B to C).
High damping capacity [Hu85]: these alloys show in the martensitic state a strong amplitude dependent internal friction. For impact loads, the specific damping capacity can be as high as 90 %.

These exclusive functional properties of shape memory alloys are closely linked to a temperature or stress induced martensitic transformation occurring in a metastable state of the alloy [De91, Ah86]. This transformation from the high temperature phase (austenite) to the low temperature phase (martensite) shows hysteresis; e.g. considerable hysteresis can be found in the strain-temperature (e.g. fig.1-2), stress-strain (e.g. fig.5) and stress-temperature relations. It follows that the transformation cannot be characterised by a single value; i.e. the temperature induced transformation has to be characterised by four temperatures; Ms and Mf (resp. As and Af) to indicate the temperatures at which the martensitic (resp. the reverse martensitic) transformation starts and finishes. The overall transformation describes an hysteresis of the order of 10 to 40 K while the difference between Af and Mf is of the order of 20 to 60 K . A more detailed description can be found in [St92]. Important to notice is also that many characteristics, such as the Youngs' modulus and the electrical resistivity, change drastically during transformation.
Many alloy systems [De91] show shape memory behaviour but only a few of them have been developed on a commercial scale (Ni-Ti, Ni-Ti-Cu, Cu-Zn-Al). Other shape memory alloys are close to market introduction (Cu-Al-Ni, Fe-Mn-Si) while still others have interesting potentials but are difficult to produce or suffer from brittleness (Ni-Al, Ni-Ti-Zr).
Since a few years, shape memory alloys have found their specific niches in many domains of industrial activities. As a consequence a steadily growing amount of different applications are now produced at large volumes. Many successful or potentially successful applications were recently presented on the "First International Conference on Shape Memory and Superelastic Technologies (SMST)" [N95], that was attended by almost 200 people.

In this paper we would like to highlight some recent developments in shape memory applications.

Shape Memory Actuators using the one or two way memory effect

Most industrial applications of Shape Memory Alloys (SMA) have been used for on/off applications such as cooling circuit valves, fire detection systems, clamping devices and many others. Commercial on/off applications are available in very small sizes such as the miniature actuator of A.M.T. [Mo92] for loads up to 1N and with an activation time of 0.1 sec. On the other hand one can also find truss actuators and Shape Memory Actuated Cylinders (SMACs) for loads up to 400 N [Ma91]. These actuators are usually electrically actuated. Although Ni-Ti SMAs are more expensive and more difficult to machine than Cu-based SMAs, there are several reasons why nearly only Ni-Ti SMAs are used for actuation purposes; Ni-Ti SMAs have a larger electrical resistivity, and allow much higher working stresses and strains. Recently, a lot of research efforts have been directed towards continuous position and force control of shape memory actuators [Re95]. SMAs offer important advantages in actuation mechanisms, as summarised below.

1. Simplicity, compactness, and safety of the mechanism:
The actuator can be reduced to a single SMA element, i.e. an electrically activated SMA wire. The stroke and force can be easily modified by the selection of the SMA element, e.g. SMA wire vs. SMA spring. Additional parts such as reduction gears are not required. Hence, the use of SMAs can result in a simplified, more compact and more reliable device.

2. Creation of clean, silent, spark-free and zero gravity working conditions:
Since friction is absent in activated SMA elements, the production of dust particles can be avoided. Conversely, a dusty environment has no influence on the action of SMA elements. Since there are also no additional vibrating parts, the activation is nearly noiseless. The acoustic emission created by the martensitic transformation can only be detected by very sensitive detectors. While no high-voltage or electrical switches are required, SMA actuators can work completely spark-free allowing them to operate in highly inflammable environments. SMA actuators can be controlled in such a way that accelerations of the order of only a few microg are generated. These very smooth movements are therefore extremely suitable for space applications where even small accelerations can influence the global movement of spacelabs or satellites.

3. High power/weight (or power/volume) ratio's:
K. Ikuta [Ik90] compared all types of actuating technologies (from small DC motors to gas turbines). He concluded that SMA actuators offer the highest power to weight ratio at low levels of weight (below 100 grams), as illustrated in Fig. 6.

Fig.6. Schematic representation of the locus of SMA-actuators in a power density vs. weight diagram. The area enclosed between the two border lines indicate where other types of actuators are situated (AC and DC motors, rotary and piston engines, turbines, ...). After K. Ikuta [Ik90].

This means that shape memory alloys are extremely attractive in microactuator technology. Therefore, it is expected that SMA-actuators will become a very important design tool in the important and rapidly growing field of micro-actuation. Examples of prototypes have been already described by Ikuta [Ik91], Walker [Wa90] and Johnson [Jo92]. Recently a Brite project has been approved for the development of remote controlled microactuators for medical applications [Br93]. One of the challenges in this field is the production of high quality and reliable thin foils of Ni-Ti by magnetron sputtering [Jo91,Mi92] or melt spinning.

Additional advantages of Ni-Ti SMAs are the excellent corrosion resistance and biocompatibility. Some drawbacks on the use of SMA-actuators should however also be considered:

1. Low energy efficiency:
It can be easily calculated that the maximum theoretical efficiency of a Carnot cycle between Af and Mf is of the order of 10 %. In reality, the conversion of heat into mechanical work is much less efficient with the result that the real efficiency is at least one order smaller than the theoretical Carnot value. This efficiency is also to a large extent determined by the design and shape of the SMA-actuator. For example, the stress and strain distribution over the cross section of a helical spring is not constant. This implies that more material is needed for generating the same force. This has a negative effect on the efficiency and the bandwidth of the spring based actuator because, for the same output, more material has to be heated and cooled. Therefore, SMA wires are mostly used in SMA actuators since this offers the advantage of optimum use of the material.

2. Limited bandwidth due to heating and cooling restrictions:
Shape memory actuators can be heated in different ways, radiation or conduction (thermal actuators) and by inductive or resistive heating (electrical actuators). For a fast and homogeneous response, resistive heating offers the most attractive solution and is therefore also widely used. The response speed is mainly limited by the cooling capacities. In addition, the martensitic transformation is exothermic (15-22 J/g) which means that extra heat has to be removed during cooling. Ways to increase the limited cooling speed are (i) the use of rectangular flat strips rather than round wires, (ii) forced cooling by a moving liquid and (iii) miniaturisation of the SMA-elements.

3. Degradation and fatigue [Hu91]:
The reliability of shape memory devices depends on its global lifetime performance. Time, temperature, stress, strain, strain mode and the amount of cycles are in this respect important external parameters. Internal parameters that can have a strong influence on the lifetime are: the alloy system, the alloy composition, the heat treatment, and the processing. For general purposes, the maximum memory effect, strain and/or stress, will be selected depending on the required amount of cycles. The following table, presented by D. Stöckel in 1992 [St92] can be used as a guideline for standard binary Ni-Ti alloys. It should however be remarked that special treatments and ternary alloys such as Ni-Ti-Cu can yield much higher values of maximum strains and stresses.

Cycles	Max. strain	Max. stress
1	8 %	500 MPa
100	4 %	275 MPa
10000	2 %	140 MPa
100000+	1 %	70 MPa

4. Complex control
Shape memory alloys show a complex three dimensional thermomechanical behaviour with hysteresis. Moreover, this behaviour is influenced by a large number of parameters. It follows that there are in general no direct and simple relations between the temperature and the position or force. Therefore, accurate position or force control by SMA actuators requires the use of powerful controllers and the experimental determination of complex data. Many mathematical models are being developed nowadays by different research groups to overcome this important limitation.

Smart materials with embedded shape memory elements

The actuators discussed above can be considered as parts of smart structures; the actuation function is performed by discrete SMA elements. SMA elements can be also easily integrated in matrix materials, yielding smart or adaptive materials. In comparison with alternative 'actuating' or 'sensing' materials, SMAs offer several important advantages and extra possibilities: (i) much larger reversible strains (up to 8%), (ii) ability to generate extremely high stresses (up to 800 MPa), (iii) large reversible changes of mechanical and physical characteristics, (iv) high damping capacity, and (v) ability to generate gradually stresses and strains. Therefore, many experts believe that SMAs offer very promising prospects in this new, rapidly evolving field of materials research. The rapidly increasing interest in these materials is also caused by the fact that thin SMA wires can be easily embedded into advanced structural materials, such as polymer matrix composites, without losing the structural integrity of the matrix material. By embedding shape memory elements into a polymer matrix or polymer matrix composite novel material characteristics can be generated: (i) upgraded shape memory characteristics, e.g. larger shape memory effects which are also less sensitive to degradation, (ii) upgraded structural characteristics, e.g. self-repairing properties resulting in an increased resistance against fatigue and buckling of the polymer matrix composite, (iii) combination characteristics, e.g. structural polymer matrix composite with adjustable shape, and (iv) completely new product characteristics, e.g. polymer matrix composites with adaptive stiffness and frequency modes. A first prerequisite to the 'design' of these smart materials is that the materials behaviour of the composing elements is known and predictable. As mentioned above however, the thermomechanical behaviour of SMAs is non-linear with hysteresis. In spite of many attempts, the models presented in literature have not yet been very successful in the quantitative description of shape memory behaviour. Another prerequisite is a thorough knowledge of the SMA-matrix interface and its stability during cyclic applications. The curing of the matrix material can influence the behaviour of the embedded and thus restrained shape memory elements. In summary, although smarts material with embedded shape memory elements offer very promising prospects, several topics have to be further investigated before commercial applications can be developed. A lot of progress can be expected in the next five years.

Biomedical applications

It cannot be denied that the largest commercial successes of shape memory alloys are nowadays situated in the field of bioengineering and biomedical applications [N95].
The shape memory property generally used here is superelasticity. Most successful is the use of orthodontic archwires. When deflected, these superelastic archwires (in contrast to similar stainless steel wires) will gradually return to their original shape exerting a small and nearly constant force on the misaligned teeth. The result is much less patient discomfort and more efficient and faster tooth movement. Those archwires also save also valuable chair time because fewer archwire changes and adjustments are required.
A second fast growing field is the use of instruments for non-invasive surgery, minimal access surgery or laparoscopy based on the superelastic properties of Ni-Ti. Complex but very flexible instruments even capable of going around corners are now used and can perform complex and delicate tasks through a trochar (a tube with a diameter of 5-10 mm).
Ni-Ti elements are also used in needles, stylets, guidewires, catheters, stents, filters, tissue anchoring and connection, flow control devices, rhinosurgical instruments, orthodontic implants,... The ability of very thin tubes (outer diameter smaller than 100 microm) allows the use in arteries or the application of small stents for angioplasty. The latter opens the possibility of temporary stenting in coronary angioplasty. The possibility of removal after healing avoids the problems associated with permanent implants viz. long term anticoagulation medication and excessive cell growth around a foreign body.
Increasing interest appears in the field of orthopedic applications, prostheses and implants. In this field one faces however the problem of biocompatibility although several studies show attractive results.

Some remarks on the design of SMA devices

Shape memory applications cannot be designed in the usual way. One reason is that the mechanical and physical properties change to a large extent during transformation, so that single property values cannot be used. Another reason is that shape memory alloys, different from structural materials, are used because of their five functional properties and, as indicated before, these functional properties can be substantially modified by alloy composition, etc... Therefore, the design of shape memory applications requires either (i) a thorough knowledge of shape memory behaviour, (ii) a close co-operation with an SMA supplier, or (iii) the use of specific computer programs [St89] which might evolve to expert systems.
This table presents a summary of physical, mechanical, economical and functional data of the two main classes of shape memory alloys; Ni-Ti and Cu-Zn-Al . Those data, collected from literature, only give a rough idea and should be handled with caution.

Conclusion

The functional properties of shape memory alloys offer unique opportunities in many fields of industrial activities. At present, most commercial successes are related to the use of superelasticity in biomedical applications. In near future many new commercial successes can be expected in various domains and especially in microactuator technology and in smart materials developments. Important to notice is that the design of shape memory applications always require a specific approach, completely different from the design with structural materials.

Acknowledgements

R. Stalmans and J. Van Humbeeck acknowledge the N.F.W.O. (National Fund for Scientific Research, Belgium) for a grant as Postdoctoral Researcher, resp. Research Leader.

References

[Ah86]: M. Ahlers, Progress in Mat. Science 30, 3, 1986, 135-186.
[Br93]: Brite-Euram project 7596, Shape memory alloy micro-actuators for medical applications "SAMA", 1993.
[De91]: L. Delaey, Mat. Sc. and Technol. 5, in "Phase Transformations in Materials", ed. P. Haesen, VCH, 1991, 339-404.
[Du90]: T.W. Duerig, Proc. Icomat 1989, Mat. Sc. Forum, Vol. 56-58, 1990, 679-692.
[Hu85]: J. Van Humbeeck, Proc. Int. Symp. 13-17, Toronto, Ed. B.B. Rath and M.S. Misra, ASM, 1985, 5-24.
[Hu91]: J. Van Humbeeck, Journal de Physique IV, C4, 1991, 189-197.
[Hu92]: J. Van Humbeeck, Mat. Res. Soc. Symp. Proc., Vol. 246, 1992, 377-378.
[Hu96]: J. Van Humbeeck and R. Stalmans, Characteristics of shape memory alloys, in: Shape Memory Behaviour, ed. K. Otsuka, to be published.
[Ik90]: K. Ikuta, Proc. IEEE Workshop, 1990, 2156-2161.
[Ik91]: K. Ikuta, M. Tsukamoto, S. Hirose, Proc. IEEE MEMS Workshop 1991, 103-108.
[Jo91]: A.D. Johnson, J. Micromech. Microeng. 1 ,1991, 34-41.
[Jo92]: A.D. Johnson, J.D. Bush, A.R. Curtis, C. Sloan, Mat. Res. Soc. Symp. Proc., Vol. 276, 1992, 151-159.
[Ma91]: J.B. Maclean, J.L. Draper, M.S. Misra, J. of Intell. Mater. Systems and Structures, Vol. 2, July 1991, 261-280.
[Mi92]: S. Miyazaki, A. Ishida, A. Takei, Proc. Int. Symp. on Measurements and Control of Robotics (SMCR- 92), Tsukuba 1992.
[Mo92]: W. Van Moorleghem, D. Reynaerts, H. Van Brussel, J. Van Humbeeck, Int. Conf. on New Actuators, Actuator 92, Bremen, 1992, 225-227.
[N95]: Proceedings of SMST-94 (Shape Memory and Superelastic Technologies), ed. A. Pelton, MIAS, Monterey, 1995.
[Re95]: D. Reynaerts, H. Van Brussel, Proceedings of SMST-94 (Shape Memory and Superelastic Technologies), ed. A. Pelton, MIAS, Monterey, 1995, 271- 275.
[Sc92]: L. McD Schetky, MRS-Proceeding, Vol. 246, Ed. C.T. Liu et al., 1992, 299-307.
[St89]: R. Stalmans, J. Van Humbeeck and L. Delaey, in: The martensitic transformation in science and technology, ed. E. Hornbogen, DGM, Oberursel, Germany, 1989, 207-213.
[St92]: D. Stöckel, Int. on New Actuators, Actuator 1992, Bremen, 79-84.
[St93]: R. Stalmans, Ph.D. thesis, dep.MTM- K.U.Leuven, April 1993.
[Wa90]: J.A. Walkers, K.J. Gabriel, M. Mehregang, Sens. and Act., Vol. A21, 1990, 243-246.

For suggestions, comments and more information, contact Rudy Stalmans