**3. Snake-like SMA actuators**

40 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

method does not change so much (from 0.91% to 0.75%).

procedure leads to improve the actuation performance.

been designed to work.

right).

As previously stated, the heating process of a SMA actuator can be easily obtained by Joule effect. In this case, it has to be considered that the shape of the electrical pulse used for the actuation strongly affects the functional properties of SMA [27]. As an example, Figure 6 shows the time required by the wires to recover the 3% of deformation (actuation time) when it is heated by two different current pulses. At the 1st cycle ramp and step electrical pulse employed to heat two 80μm wires were designed to have the same electrical efficiency but the actuation time is 400ms by step current pulse while it is 623ms by ramp. After 5·103 cycles the wire heated by step pulse employs 618ms, so 218ms more than the time employed at the first cycle. As opposite, the ramp pulse leads to an increase of just 22ms. After 5·104 cycles the actuation time related to the wire heated by step is even higher than the one related to wire heated by ramp. This behavior leads to a drastic decrease of the step heating method efficiency (from 0.91% to 0.55%), vice versa the efficiency of the ramp heating

Then, in order to achieve fatigue performance acceptable for the specific device, the right electrical pulse has to be chosen considering the number of cycles that the SMA actuator has

Recently, the effect of drawing procedure on functional fatigue of thin NiTi wires has also been investigated [21]. Basically, 80μm NiTi wires were produced through two different drawing procedures reaching the same final cold working level before shape setting. These two processes differ for the number of drawing steps carried out to reach a certain cold working level before each heat treatment. After the last thermal treatment the specimen that underwent to a less number of drawing steps shows a narrow thermal hysteresis, even after thermo-mechanical cycling (see Figure 7). It means that adopting a severe drawing

**Figure 6.** Actuation time tests performed by electrical heating (on the left) and by efficiencies (on the

The main focus of miniaturization is to develop devices with high mechanical performance in very limited spaces. In the SMA field, the mini-actuator may be composed by only the SMA element or by a mini-modular device in which the SMA element is embedded and acts as the core part. In the latter case, some issues typical of SMA miniaturization should be taken into a serious consideration: first of all, the fixing points of the SMA element and the electrical contacts should be adequately designed to prevent abrupt rupture, and then the use of a wire geometry generally implies the employment of additional mechanical parts which may affect the long run performance of the SMA element.

The mechanical work resulting from the SMA shape recovery is used to produce linear or rotational motion in smart actuators. To attain high strokes in limited space, designers developed several miniature devices which include SMA elements with a straight or a helical spring shape [31]. These two shapes present two opposite mechanical behaviors: the straight conformation allows high forces but smart wire arrangements should be employed in order to achieve also high strokes. As an example, Jansen et al. used two sets of seven pulleys to arrange a 1000mm long SMA wire with a diameter of 0.200mm, and attain 3% of stroke under a constant load of about 130MPa [32]. The use of additional mechanical components may affect the overall performance of the SMA element, leading to friction which could influence the final functional parameters of the device. As opposite, the helical spring shape guarantees high strokes without the use of any further mechanical part but fairly small forces can be attained especially when embedded in miniature devices.

Hence, straight and helical spring conformations represent two opposite limits in the force versus stroke plane. In order to answer both to the need of stroke and force between these two limits, new SMA geometries should be studied.

In a recent work, a new SMA conformation has been proposed [33]; it consists in a planar wavy formed NiTi wire, named snake-like arrangement. To understand the mechanical performance of this unusual conformation as respect to other shapes, three NiTi samples

with different geometry, but with identical elemental composition, thermal history and power consumption were tested by means of a standard tensile test. In this investigation the NiTi wire was formed to have the straight conformation, the helical spring and the snakelike arrangements, and austenite and martensite mechanical responses were assessed separately. As can be seen in Figure 8 the snake-like shape has a mechanical behavior between the wire and the helical spring ones.

**Figure 8.** Comparison between austenite and martensite mechanical behavior of different NiTi wire geometries for actuators.

The snake-like conformation is characterized by four main parameters: the number of curvatures, N, the curvature radius, R, the distance between two consecutive curvatures, D, and the height, H. A snake-like specimen can be prepared by constraining a SMA wire in the snake-like arrangement by using a drilled aluminum bar and a series of nails; heating at high temperature is finally used to fix the shape. Figure 9 reports a snake-like sample prepared from a commercial NiTi wire having the diameter of 0.2mm, formed at 500°C during ten minutes and quenched in water at room temperature, with N=4, H=5.27mm±0.07mm, R=0.54mm±0.03mm and D=0.82mm±0.06mm.

**Figure 9.** Snake-like NiTi wire with N=4, H=5.27mm±0.07mm, R=0.54mm±0.03mm and D=0.82mm±0.06mm.

As previously stated, during the ten minutes formation process the curved parts of the snake element are in contact with metal devices. This fact does not modify the calorimetric properties of the SMA element in all its length, as can be seen in Figure 10 where the overlapping of the DSC scans relative to the straight and the curved segments is clearly visible.

42 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

between the wire and the helical spring ones.

geometries for actuators.

D=0.82mm±0.06mm.

with different geometry, but with identical elemental composition, thermal history and power consumption were tested by means of a standard tensile test. In this investigation the NiTi wire was formed to have the straight conformation, the helical spring and the snakelike arrangements, and austenite and martensite mechanical responses were assessed separately. As can be seen in Figure 8 the snake-like shape has a mechanical behavior

**Figure 8.** Comparison between austenite and martensite mechanical behavior of different NiTi wire

H=5.27mm±0.07mm, R=0.54mm±0.03mm and D=0.82mm±0.06mm.

**Figure 9.** Snake-like NiTi wire with N=4, H=5.27mm±0.07mm, R=0.54mm±0.03mm and

The snake-like conformation is characterized by four main parameters: the number of curvatures, N, the curvature radius, R, the distance between two consecutive curvatures, D, and the height, H. A snake-like specimen can be prepared by constraining a SMA wire in the snake-like arrangement by using a drilled aluminum bar and a series of nails; heating at high temperature is finally used to fix the shape. Figure 9 reports a snake-like sample prepared from a commercial NiTi wire having the diameter of 0.2mm, formed at 500°C during ten minutes and quenched in water at room temperature, with N=4,

**Figure 10.** DSC scans of the straight and the curved part of the snake-like NiTi wire.

For small deformation values (see Figure 11), this snake-like shape shows an almost complete recovery of the deformation both for austenite and for martensite.

The relative higher output stroke performed by the snake-like NiTi wire as respect to the straight wire conformation is substantially due to the bend/unbend movement of the curvatures. In order to avoid plastic deformation and to achieve simultaneously a significant output stroke, the snake sample, formed as previously explained, should be stressed with a load near 0.1N, as forces lower and higher than 0.1N do not guarantee a proper functioning of this snake element: in fact, lower forces do not induce martensite formation, as opposite higher forces cause plastic deformations. As an example, in Figure 12 the curves derived from five thermal cycles of the snake-like NiTi sample under two constant loads, 0.1N and 0.2N is depicted. By comparing the two trends, it can be seen that in both cases the snake sample lose part of its shape recovery during cycling; this is much more visible for the specimen stressed with 0.2N.

Fatigue life was studied under the constant load of 0.1N; the snake-like sample was electrically activated with 0.6A with a maximum power consumption of 1W and a work frequency of 0.1Hz. Fracture takes place at the apex of that curvature most far from the applied load and from Figure 13 it can be observed that it occurs near the 7,5x104 cycle. The fractured section was analyzed through SEM observation, see the inset of Figure 13; it can be seen a region with the striation of the crack propagation departing from a nucleation site placed on the inner surface of the snake and a region characterized by dimples, typical of ductile materials.

After the cycling test, the DSC scan shows a feeble shifting of the transformation temperatures towards higher values; a better definition of the rhombohedral phase during heating is also visible (see Figure 14).

**Figure 11.** Force versus displacement of austenite and martensite of the snake-like NiTi wire.

**Figure 12.** Hysteresis curves of the snake-like NiTi sample treated at 500°C for 10 minutes under 0.1N and 0.2N constant load.

When embedded in mini-modular device, the snake-like wire does not lose its mechanical performance. As an example, in Figure 15 a miniature rotational actuator composed by two mutual antagonist NiTi snake-like wire is reported [34]. The device exerts a stroke of about 120° under a constant torque of 0.1Nmm, as shown in Figure 16; it is activated by 0.6A with a maximum power consumption of 1W and a work frequency of 0.05Hz. Due to the particular construction, plastic deformation of the SMA element is prevented as the SMA elements never exceed 6mm of deformation.

**Figure 13.** Snake-like NiTi sample fatigue test result under 0.1N load.

heating is also visible (see Figure 14).

and 0.2N constant load.

elements never exceed 6mm of deformation.

After the cycling test, the DSC scan shows a feeble shifting of the transformation temperatures towards higher values; a better definition of the rhombohedral phase during

**Figure 11.** Force versus displacement of austenite and martensite of the snake-like NiTi wire.

**Figure 12.** Hysteresis curves of the snake-like NiTi sample treated at 500°C for 10 minutes under 0.1N

When embedded in mini-modular device, the snake-like wire does not lose its mechanical performance. As an example, in Figure 15 a miniature rotational actuator composed by two mutual antagonist NiTi snake-like wire is reported [34]. The device exerts a stroke of about 120° under a constant torque of 0.1Nmm, as shown in Figure 16; it is activated by 0.6A with a maximum power consumption of 1W and a work frequency of 0.05Hz. Due to the particular construction, plastic deformation of the SMA element is prevented as the SMA

**Figure 14.** Snake-like NiTi sample DSC scan before and after the fatigue test

**Figure 15.** Mini rotational actuator activated by two antagonists snake-like NiTi wires.

**Figure 16.** Fatigue test results of the mini rotational actuator under 0.1Nmm constant torque.

#### **3.1. Micro snake-like shape memory alloy elements**

In the last years the trend of miniaturization of components and products has been evidently introduced in several industrial fields, such as biomedical, electronics, aerospace and mechanics [35-36]. SMAs, as it already happens in biomedical applications [14,15,37-40], can be machined using laser machining with appreciated and high qualitative results [12]. The decision to investigate this method for the production of SMA actuators is considered an interesting topic because it shows some relevant advantages in comparison with other technologies, such as high productivity, high quality and repeatability, when applied in the industrial world.

Hence, the snake geometry can be easily scale down to the micro scale by the employment of laser technology with appreciated and high qualitative results. Because laser material processing is a thermal process, in which an heat source is used to remove a certain volume of material from the work piece, some thermal damages are evidently obtained, such as melted material, heat affected zone and oxides. To remove these defects, chemical etching and electro-chemical polishing have to be performed on the snake SMA samples [37,41]. Figure 17 depicts SEM images of the snake NiTi sample after laser, chemical etching and electro-chemical processes; top and bottom surfaces as well as snake sample section are shown. It can be seen that laser damages are mainly visible on the bottom surface and that the chemical etching remove the great part of defects. Moreover, a significant loss of mass and consequent reduction of the dimensions of the micro-snake can be evidently observed from SEM pictures due to the material removal of the post-processing. The snake NiTi sample geometry and calorimetric properties change after each process, as can be seen in Figure 18 and in Figure 19 respectively. In particular, the geometrical dimension are referred to a micro-snake obtained starting from a 120μm NiTi sheet. As concern the DSC data, it can be observed that the laser machining produces a refinement of the rhombohedral phase peak both on cooling and heating while the martensite peak barely change as respect to the unworked 120mm thick sheet, being almost spread and invisible. The laser process also yields a conspicuous diminution of the rhombohedral phase transformation temperatures and of the martensite starting temperature while the other characteristic transformation temperatures do not have an analogous variation being almost unchanged. Chemical etching and electro-polishing steps do not visibly alter any phase transformation temperature both on cooling and on heating, even if some little change due probably to the hydrogen introduction in the chemical process, is visible [45,46].

46 Smart Actuation and Sensing Systems – Recent Advances and Future Challenges

**3.1. Micro snake-like shape memory alloy elements** 

industrial world.

**Figure 16.** Fatigue test results of the mini rotational actuator under 0.1Nmm constant torque.

In the last years the trend of miniaturization of components and products has been evidently introduced in several industrial fields, such as biomedical, electronics, aerospace and mechanics [35-36]. SMAs, as it already happens in biomedical applications [14,15,37-40], can be machined using laser machining with appreciated and high qualitative results [12]. The decision to investigate this method for the production of SMA actuators is considered an interesting topic because it shows some relevant advantages in comparison with other technologies, such as high productivity, high quality and repeatability, when applied in the

Hence, the snake geometry can be easily scale down to the micro scale by the employment of laser technology with appreciated and high qualitative results. Because laser material processing is a thermal process, in which an heat source is used to remove a certain volume of material from the work piece, some thermal damages are evidently obtained, such as melted material, heat affected zone and oxides. To remove these defects, chemical etching and electro-chemical polishing have to be performed on the snake SMA samples [37,41]. Figure 17 depicts SEM images of the snake NiTi sample after laser, chemical etching and electro-chemical processes; top and bottom surfaces as well as snake sample section are shown. It can be seen that laser damages are mainly visible on the bottom surface and that the chemical etching remove the great part of defects. Moreover, a significant loss of mass and consequent reduction of the dimensions of the micro-snake can be evidently observed from SEM pictures due to the material removal of the post-processing. The snake NiTi sample geometry and calorimetric properties change after each process, as can be seen in Figure 18 and in Figure 19 respectively. In particular, the geometrical dimension are referred to a micro-snake obtained starting from a 120μm NiTi sheet. As concern the DSC data, it can be observed that the laser machining produces a refinement of the rhombohedral phase peak both on cooling and heating while the martensite peak barely change as respect to the unworked 120mm thick sheet, being almost spread and invisible. The laser process also yields a conspicuous diminution of the rhombohedral phase transformation temperatures

**Figure 17.** SEM images of the snake NiTi sample surfaces and section after laser, chemical etching and electro-polishing processes.

**Figure 18.** Variation of the geometrical parameters of the micro-snake during the fabrication process (h: high, R: radius of curvatures, w: width, t: thickness).

**Figure 19.** DSC scans before laser machining and in correspondence of each fabrication process.

The fabrication process previously described can be used to produce snake SMA sample of different length. As an example, Figure 20 depicts the SEM image of a micro-snake NiTi sample with 19 curvatures (i.e. 3.8mm of length) obtained starting from a 120μm NiTi sheet treated at 400°C for 15 minutes and quenched in water.

**Figure 20.** SEM image of a micro-snake NiTi sample obtained starting from a 120μm NiTi sheet.

Strain-recovery and thermo-mechanical cycling tests demonstrate the high performance of this micro-snake sample. As an example, under the constant load of 16mN, it reaches about 40% of recovery with a cycling stabilization within few hundreds of cycles, when heated by electrical current (130mA) with a maximum power consumption of 0.3W and a frequency of 0.15Hz.
