Jtam-A4.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 50, 3, pp. 855-869, Warsaw 2012 50th Anniversary of JTAM PERFORMANCE AND DESIGN OF PRECISION-CAST SHAPE MEMORY ALLOY BRAIN SPATULA Hisaaki Tobushi, Kento Mitsui, Kohei Takeda Aichi Institute of Technology, Department of Mechanical Engineering, Toyota, Japan e-mail: tobushi@aitech.ac.jp Kazuhiro Kitamura Aichi University of Education, Department of Technology Education, Kariya, Japan Yukiharu Yoshimi Yoshimi Inc., Obu, Japan Inorder todevelopabrainspatulamadeofa shapememoryalloy (SMA), thispaperdiscusses the bending characteristics of a newbrain spatula precision-cast in aTiNi SMA.The results obtained can be summed up as follows. (1) With respect to an SMA-brain spatula having the same length andwidth as the existing typemade of copper, if the new cast SMA spatula is 1.2 times as great in thickness as the existing copper spatula, or if a conventionally rolled SMA spatula is 1.3 times as great, the SMA instrument will offer the same bending rigidity andwithstand almost as great a bending force as the copper one. (2)Expressing the bending fatigue life of the copper or SMA in the region of low-cycle fatigue as a power function of the maximum bending strain, the fatigue life of the SMA is longer than that of the copper. For both the cast and the rolled SMAs, the fatigue life is longer under pulsating-plane bending than under alternating plane bending. Key words: shape memory alloy, brain spatula, bending 1. Introduction The shapememory effect (SME) and superelasticity (SE) are characteristic behaviors of a shape memory alloy (SMA). A strain of several percent can be recovered by heating in the case of the SMEorunloading in the case of SE.Thesebehaviors occurdue to themartensitic transformation (MT) and its reverse transformation. The large recovery stress and great amounts of energy dissipation and storage associated with theMT are effectively exploited in an SMA (Funakubo, 1987; Duerig et al., 1990; Saburi, 2000; Chu and Zhao, 2002; Otsuka and Wayman, 1998). The development of applications for SMAs as intelligent materials has therefore attracted worldwide attention. SMAs are now in use across wide fields of industry, electrical manufacturing, and medical and leisure technologies, to name only a few. A brain spatula or brain retractor is an instrument used in surgery to hold a brain incision open while a deep cerebral tumor is operated on. A schematic image of how this is done is shown in Fig. 1. As shown, the spatula is used in a bent form, fitting the shape and depth of the individual brain. After the operation, the spatula is struck with a miniature hammer to return it to its original flat form, after which it is sterilized by heating and is then ready for reuse. The usual material for brain spatulas hitherto has been copper, but owing to the irrecoverable loss of evenness that develops on the copper surface after each use, in practice the instrument has to be disposed of after being used only a few times. If an SMA material is used instead, the original flat shape can be restored automatically and precisely through the working of the SME during the sterilization heating in the autoclave. This not only saves time and dispenses with 856 H. Tobushi et al. the need for hammering, but also means that the unevenness left from plastic deformation is much reduced so that the spatula can be reused over and over. Fig. 1. Illustration of a brain spatula in surgical use The instruments treated in this study have been recently developed and are precision-cast in a lost-wax process of the self-combustion high-temperature synthesis method (Yoshimi et al., 2008). The brain spatula needs to form itself into unique shapes according to the brain configurations of individual patients and, as canbe seen fromthe examples inFig. 2, the spatulas themselves also come in intricate forms that are a challenge to create in theTiNi SMAmaterial. The precision casting technique makes it much easier for these tools to bemanufactured. Fig. 2. Examples of precision-cast TiNi SMA brain spatulas; (a) whole rake-type brain spatula against a centimeter scale, (b) tip of the rake-type brain spatula, (c) batch of brain spatulas just after precision casting The conditions required of a brain spatula are, first, that it can be bent to any desired shape to fit the brain being operated on and, second, that it should have sufficient rigidity to be able to keep the brain incision open during the operation. These requirements can be specified in terms of the bendingdeformation properties of the instrument, including its bending rigidity. To Performance and design of precision-cast shape memory alloy... 857 evaluate the reliability of abrain spatula fromthepoint of viewof safety, the fatigue properties of thematerial are crucial, and to investigate these, the spatula has to be subjected to cyclic plane bending. This paper is the first reported study of the plane bending properties of a TiNi SMA considered as the material for a brain spatula. For the present study, in pursuit of the development of the SMAbrain spatula, tension tests were conducted to examine stress-strain relations in three materials: the newly introduced cast TiNi SMA, a TiNi SMA of the conventional rolled type, and copper of the type used in brain spatulas up until now. The shape and dimension requirements of an SMA brain spatula were investigated on the basis of the bending deformation properties of a strip cantilever. Further, an investigation was also conducted into the fatigue properties of thematerials, which is a question of extreme importance for the practicalities of their cyclic use. 2. Experimental method 2.1. Materials and specimens The materials used in the experiment were a new-type cast Ti-49.7at%Ni SMA, a conven- tional rolled Ti-50.0at%Ni SMA, and copper as used in brain spatulas hitherto. The new cast SMA was produced by precision casting using a lost-wax process from a self-combustion high- temperature synthesismethod (Yoshimi et al., 2008). Samples of the cast and rolled SMAswere shape-memorized by fixing them in a flat plane for 40min at a furnace temperature of 753Kand then quenching them inwater. The starting andfinishing temperatures for theMTof the SMAs, Ms and Mf, and those for the reverse transformation, As and Af, were obtained from theDSC (differential scanning calorimetric) tests. The values obtained were Ms = 326K, Mf = 312K, As = 342K, Af = 365K for the rolled SMA and Ms = 358K, Mf = 283K, As = 314K, Af = 386K for the cast SMA. The specimens used in the tension tests were the uniform rec- tangular bars with a thickness t=1.0mm, width w =1.2mm and length l= 160mm for the rolled and cast SMAs, with corresponding values of t= 1.0mm, w = 8.5mm and l =140mm for the copper. In the fatigue tests, all specimens were bars with dimensions of t = 1.0mm, w=1.2mm and l=80mm. 2.2. Experimental apparatus An SMA characteristic-testing machine was used for the tension test (Tobushi et al., 1992). The testingmachinewas composed of a tension control for loading andunloading and a heating- cooling device to control temperature. Displacements of the specimen were measured using an extensometer with a gauge length of 50mm. Fig. 3. Experimental apparatus for alternating-plane bending fatigue test; 1O specimen, 2O grip, 3 O speed controller, 4O cycle counter, 5Omotor, 6O power source, 7O crank 1, 8O crank 2, 9O bearing 858 H. Tobushi et al. Amachine for the testing of alternating-plane bending fatigue (Furuichi et al., 2003), for use in the fatigue tests, is shown inFig. 3. In the fatigue testing, themaximumbending strain occurs in the central part of the specimen 1O, and the tensile and compressive bending strains occur alternately. Usewas alsomade in the fatigue tests of amachine for the testing of pulsating-plane bending fatigue (Tobushi et al., 2001). In the alternating- andpulsating-plane bending tests, the maximum bending strain on the surface of the specimen was selected and the number of cycles to failure was obtained under a constant frequency. A scanning electron microscope (SEM)was used to observe the fracture surface of the specimen. 2.3. Experimental procedure The tension tests were carried out under a constant strain rate in air at room temperature below the Mf point of the SMA bars. Since the yielding of the SMA occurs under low stress in the M-phase, the SMA-brain spatula can be bent with a very small force. In the case of SMA bars in the M-phase, a residual strain appears after unloading. SMA bars showing the residual strainwere heated up to temperatures above the Af point under no load. In this heating process of the SMA bars, the residual strain was found to diminish due to the reverse transformation between the As and Af points. The tests for alternating- andpulsating-plane bending fatiguewere carried out in air at room temperature. Every specimen fractured in the central part of its length between the grips, which is where the bending strain reaches its maximum value. The maximum bending strain εm on the surface of the specimen was obtained from the radius of curvature at the point of fracture. The frequency f was found to be 8.33Hz (500cpm) and 3.33Hz (200cpm), respectively. 3. Deformation properties of materials used for brain spatula 3.1. Tensile deformation properties The stress-strain curves of the copper, the rolled and the cast SMAs obtained from the tension test under a strain rate of dε/dt = 2 · 10−4 s−1 are shown in Fig. 4. As can be seen in Fig. 4, a linear elastic deformation occurs in the initial loading stage and yielding occurs thereafter. The modulus of elasticity E determined from the slope of the initial stress-strain Fig. 4. Tensile stress-strain curves for copper, rolled SMA and cast SMA asmaterials for brain spatulas curve is 40GPa for the rolled SMA, 54GPa for the cast SMA and 95GPa for the copper. Approximating the elastic and yield regions of the stress-strain curves to two straight lines, the yield stress σM was determined from the intersection of these lines. This gave a value of 68MPa for the rolled SMA, 168MPa for the cast SMA and 240MPa for the copper. In the case of the copper, the deformation seen at strain levels of above 0.2% occurs due to plastic Performance and design of precision-cast shape memory alloy... 859 deformation with dislocations. For example, in the unloading process from a strain of 4%, there is a strain recovery of 0.25% corresponding to the elastic deformation, leaving the residual strain as permanent strain. In the cases of the SMAs, however, since thematerial is in theM-phase at room temperature below the Mf point, yielding occurs due to rearrangements in theM-phase. In the unloading process from the same strain of 4%, there is a strain recovery of 0.6% for the rolled, and 0.8% for the cast SMA, leaving the residual strains of 3.4% and 3.2% respectively after unloading. Fig. 5. Strain-temperature curves for the rolled and cast SMA in tension, with unloading followed by heating without loading The relationships between strain and temperature for the cast and rolled SMAs, obtained from tension testswith unloading followed by heating in the absence of load, are shown inFig. 5. The symbols As:C, As:R, Af:C and Af:R in Fig. 5 represent the starting and finishing tempe- ratures for the reverse-transformation in the cases of the cast and rolled SMAs, respectively. In the heating process under no load, strain begins to recover gradually around As and disappears altogether around Af. The SME behind this strain recovery occurs as a result of the reverse transformation from theM-phase to the parent (austenite) phase. 3.2. Comparison of characteristic values for deformation Table 1 sets out the values of the modulus of elasticity E, the yield stress σM, the yield strain εM and the hardeningmodulus k for the copper, the rolled and the cast SMAsmaterials, as obtained from the tension tests. As can be seen, both themodulus of elasticity and the yield stress are lower for the rolled and the cast SMAs than for the copper. These differences account for the residence to bendingdeformation in the SMAmaterials, and also for the bending rigidity which allows the brain incision to be held open in a particular shape during the operation. In the next section, the bending deformation properties of brain spatulas made of these materials will be discussed. 4. Bending characteristics of copper and SMA brain spatulas In order to design an SMA brain spatula, it is important to evaluate the force required for bending of the spatula in operational use and the bending rigidity required for brain incision to be held open during the operation.With a focus on these two bendingdeformation properties in spatulasmade of copper and SMA, the required specifications for an SMAbrain spatula will be clarified.Conceiving of the existing type of brain spatula as a bendable cantilevermade of a strip of material with a uniform rectangular cross-section, the length of the strip can be expressed by l, the width of the cross-section by w and the thickness by t. 860 H. Tobushi et al. Table 1. Values of modulus of elasticity, yield stress, yield strain and hardening modulus for copper, rolled SMA and cast SMAs Materials Copper Rolled SMA Cast SMA (M-phase) (M-phase) Modulus of elasticity 95 40 54 E [GPa] Yield stress 240 68 168 σM [MPa] Yield strain 0.25 0.17 0.34 εM [%] Hardening modulus 0 2.6 2.5 k [GPa] 4.1. Bending rigidity required to hold the brain spatula in its bent form In order to maintain the chosen displacement in the part of the brain that is opened during the operation, it is required that the brain spatula should remain stable in its bent form (see Fig. 1). This requirement can be quantified in terms of the maximum permitted deflection in a cantilever made of the specifiedmaterials. Let us consider, in particular, the conditions required in each case for obtaining the samemaximum deflection ymax in response to the same force F applied at the top of the cantilever. The maximum deflection of the cantilever ymax can be expressed using the secondmoment of area Iz =wt 3/12 from the theory of elasticity as follows ymax = Fl3 3EIz = 4Fl3 Ewt3 (4.1) Assuming that this maximumdeflection ymax in strips subjected to the same force F coincides for the copper and SMAmaterials, the following equation is obtained ymax = 4Fl3Cu ECuwCut 3 Cu = 4Fl3SMA ESMAwSMAt 3 SMA (4.2) From the practical point of view of a brain operation, the width w and the length l of an SMA brain spatula are expected to offer the same values as for a conventional spatula made of copper. Therefore, it is appropriate to assume that the length and width dimensions of both kinds of spatula will coincide, leaving only the thickness t to differ. The thickness for the SMA spatula tSMA can be found from Eq. (4.2) as follows tSMA = tCu 3 √ ECu ESMA (4.3) If the values for the modulus of elasticity shown for the copper and SMA materials in Table 1 are now substituted in Eq. (4.3), the bending rigidities of the three types of spatulas will be found to coincide when the rolled SMA spatula has a thickness of 1.3 times, and the cast SMA spatula a thickness of 1.2 times that of the copper spatula. These are the conditions, in other words, in which the same deflection and bending rigidity can be obtained from the SMA brain spatulas as from the copper one. 4.2. Force required to bend brain spatula In a brain operation, the surgeon has to bend the spatula to fit the exact shape and depth of the part of the brain being opened. The force required to bend the brain spatula is evaluated as Performance and design of precision-cast shape memory alloy... 861 the force applied at the top of the cantilever to obtain the requiredmaximumbending strain.As shown in Fig. 6a, the yield (transformed) regions of the strip during bending occur at the inner and outer surfaces of the cantilever. A schematic stress-strain diagram and the distributions of bending strains and stresses in the strip are shown in Figs. 6b, and 7, respectively. In Fig. 6b, it is assumed that the yielding region can be expressed by the linear hardening with a hardening modulus k while the yield stress and yield strain are expressed by σM and εM, respectively. In these figures, the maximum bending stress and strain are denoted by σm and εm, respectively. It is assumed that the yield stress σM is the same in tension as in compression. Fig. 6. (a) Elastic and transformed regions of the specimen in bending; (b) stress-strain diagram Fig. 7. Bending strain (a) and stress (b) distribution in the specimen Let us first consider the bending moment Me required to produce the elastic region of the strip in bending. The thickness of the elastic region tM is tM =(εM/εm)t. Since the maximum bending stress in the elastic region σmax is the sameas the yield stress σM, the bendingmoment required for the elastic region Me is Me = σMZ, where Z denotes the section modulus and Z =wt2M/6. Accordingly, the bendingmoment Me for the elastic region is Me = σMwt 2 6 (εM εm )2 (4.4) Next, let us turn our attention to the bendingmoment My required to create the yield (transfor- med) region of the strip. The center of the yield region yc from the neutral axis z is calculated as yc = tc 2 = t 4 ( 1+ εM εm ) (4.5) The area Ay of the yield region in the tension side of the cross section is Ay = wt 2 ( 1− εM εm ) (4.6) Considering the stress distribution in the yield region and the symmetric match between the tension and compression sides, the bending moment My required to create the yield region is obtained as follows 862 H. Tobushi et al. My = 1 2 (σM +σm)Aytc = 1 8 [2σM +k(εm−εM)]wt 2 [ 1− (εM εm )2] (4.7) The total bending moment M required to bend the brain spatula is the sum of the bending moment Me for the elastic region and the bendingmoment My for the yield region M =Me+My = wt2 24 { 2σM [ 3− (εM εm )2] +3k(εm−εM) [ 1− (εM εm )2]} (4.8) Since the bendingmoment M =Fl, the force required is F = wt2 24l { 2σM [ 3− (εM εm )2] +3k(εm−εM) [ 1− (εM εm )2]} (4.9) The condition that the forces required to benda copper spatulawith k=0and an SMAspatula coincide can be expressed as follows F = σMCuwCut 2 Cu 12lCu [ 3− (εMCu εmCu )2] (4.10) = wSMAt 2 SMA 24lSMA { 2σMSMA [ 3− (εMSMA εmSMA )2] +3kSMA(εmSMA−εMSMA) [ 1− (εMSMA εmSMA )2]} If the length l, thewidth w and themaximumbending strain εm of both spatulas coincide only the thickness to differ, the thickness of the SMA brain spatula is tSMA = tCu √ √ √ √ √ √ √ 2σMCu [ 3− ( εMCu εm )2] 2σMSMA [ 3− ( εMSMA εm )2] +3kSMA(εm−εMSMA) [ 1− ( εMSMA εm )2] (4.11) The yield stress σM, the yield strain εM and the hardeningmodulus k for the three types of materials are shown inTable 1.Using these values, the thickness of the SMAbrain spatula tSMA can be obtained fromEq. (4.11). The calculated relations between the thickness ratio tSMA/tCu and themaximum bending strain εm are shown in Fig. 8. If thematerial of the thickness tSMA Fig. 8. Relationship between the thickness ratio tSMA/tCu and the maximum bending strain εm calculated fromEq. (4.11) shown in this figure is used, an SMA spatula can be bent by the same force as that required for a conventional copper spatula. As can also be seen fromFig. 8, themaximumbending strain εm exerts very little effect on the ratio of tSMA to tCu except in an area of small εm values close to the yield strain εM. When the thicknesses of the rolled and cast SMA spatulas are 1.85 times and1.20 times that of the copper spatula, respectively, the SMAspatula can bebentbyapplying the same force as with an existing copper spatula. Performance and design of precision-cast shape memory alloy... 863 4.3. Shape of SMA brain spatula Thediscussion in Sections 4.1 and4.2will havemade clear that if the length and thewidthof a rolled SMAbrain spatula are the same as those of the existing copper typewhile the thickness is 1.3 times as large, the rolled SMA spatula is capable of the same bending rigidity and can be easily bent using a smaller force than required for the copper one. Similarly, if the thickness of the cast SMA spatula is 1.2 times that of the copper one, the cast SMA spatula is capable of the same bending rigidity and can be bent by the same force as an existing copper one. 5. Bending fatigue properties 5.1. Fatigue life in alternating-plane bending Figure 9 shows the relationship between the maximum bending strain εm and the number of cycles to failure Nf for the rolled SMA, the cast SMA and the copper one, as obtained from the alternating-plane bending fatigue test. For all threematerials, the larger the εm, the smaller the Nf. It can also be seen that the fatigue life of the rolled SMA is 100 times longer than that of the copper, and that the fatigue life of the cast SMA is also 100 times longer that of the copper in the region where εm is small and 10 times longer in the region where it is large. The fatigue life of the copper is several hundredcycles at εm =1%,which is almost the sameas the low-cycle fatigue life of normalmetals (Shigley andMischke, 1989). In the case of copper, the yield strain is causedby slip in the crystals andplastic deformation occurs repeatedly, resulting in short fatigue life. In the SMAs, the yield strain is not caused by permanent slip but by a recoverable rearran- gement of theM-phase, which accounts for the long fatigue life. In the case of the cast SMA, as already observed in Fig. 4, stress increases with an increase in strain, and this in turn results in the shortening of the fatigue life with an increase in themaximumbending strain. The influence of the frequency on the fatigue life of the threematerials is not clear for f =3.33Hz and8.33Hz. For the case of a TiNi SMA wire, it has been found that the lower the frequency, the longer the fatigue life (Tobushi et al., 2000). This being so, the frequency of an SMA brain spatula in practical use should be lower than this range between 3.33Hz and 8.33Hz. In practice, therefore, the actual fatigue life of the material in use will be longer than that obtained in the present study. Fig. 9. Fatigue life curves for copper, rolled SMA and cast SMA under alternating-plane bending In the low-cycle fatigue region of Fig. 9, the fatigue life curves of all three materials tested can be approximated by straight lines. Since the relationship between εm and Nf can also be expressed by a straight line on the logarithmic graph, the relationship can be expressed by a 864 H. Tobushi et al. power function similar to the fatigue life for TiNi SMAwires and tubes (Matsui et al., 2004) as follows εmN β f =α (5.1) where αdenotes εm at Nf =1and β denotes the slopeof the logεm−logNf curve, respectively. The results calculated fromEq. (5.1) with values of β=0.58 and α=0.25 for copper, β=0.41 and α= 0.67 for the rolled SMA and β = 0.14 and α= 0.05 for the cast SMA are shown by solid lines in Fig. 9. As can be seen, these calculated results generally match up well with the inclinations obtained by testing. 5.2. Fatigue life in pulsating-plane bending Figure 10 shows the relationship between maximum bending strain εm and the number of cycles to failure Nf for the rolled SMA, the cast SMA and the copper one, as obtained from the pulsating-plane bending fatigue test. Here too, the larger the εm, the smaller the Nf for all three materials. The fatigue life of the rolled SMA is 100 times longer than that of the copper, and that of the cast SMA is 40 times longer. The fatigue life of the copper is two thousand cycles at εm = 2%. Again as observed in Fig. 4, the yield stress is higher for the cast SMA than for the rolled SMA, resulting in shorter fatigue life. Fig. 10. Fatigue-life curves for copper, rolled SMA and cast SMA used for brain spatula under pulsating-plane bending As in Fig. 9, the fatigue life curves of all three materials in the low-cycle fatigue region can be approximated by straight lines, meaning that the relationship can be expressed by a power function of Eq. (5.1). The calculated results from Eq. (5.1) with β = 0.35 and α = 0.26 for copper, β = 0.26 and α = 0.43 for the rolled SMA and β = 0.29 and α = 0.49 for the cast SMAare shown by the solid lines in Fig. 10. As can be seen, these calculated results again show a good overall match with the test inclinations. 5.3. Comparison of fatigue life The relationships between the maximum bending strain εm and the number of cycles to failure Nf obtained from the alternating- and pulsating-plane bending fatigue tests are brought together in Fig. 11 for the rolled, and in Fig. 12 for the cast SMA. In both cases, the fatigue life is shorter in alternating-plane bending than in pulsating-plane bending. Since the amount of work dissipated in each cycle is greater in the case of alternating-plane bending, a longer fatigue life can be obtained for an SMA brain spatula in use by fixing the bending direction to result in pulsating bending. The influence of the dissipatedwork Wd on the difference in fatigue life between the cases of alternating- and pulsating-plane bending can be explained as follows. The stress-strain relations and the amount of work dissipated Wd at each cycle of pulsating- Performance and design of precision-cast shape memory alloy... 865 or alternating-plane bending are shown, respectively, in diagrams (a) and (b) in Fig. 13, where it is assumed that the yielding region is expressed by the linear hardening with a hardening modulus k and that the yield stress σM is constant under cyclic deformation and has the same value under tension and compression. The area enclosed by the hysteresis loop gives a measure of the work dissipated Wd per unit volume in each cycle. Fig. 11. Relationship between the maximum bending strain and number of cycles to failure for the rolled SMA under alternating-plane and pulsating-plane bending Fig. 12. Relationship between the maximum bending strain and number of cycles to failure for the cast SMA under alternating-plane and pulsating-plane bending Fig. 13. Stress-strain diagram and dissipated work Wd in the initial cycle under pulsating-plane bending (a) and alternating-plane bending (b) Thework dissipated in N =2 is expressed by the following equation for the pulsating-plane bending Wd = ( 1− k E )[ σM ( 2− k E ) +k ( 1− k E )( εm− σM E )][( 1− k E )( εm− σM E ) − σM E ] (5.2) 866 H. Tobushi et al. while the work dissipated for the same cycle of alternating-plane bending is Wd =4 E−k E+k ( εm− σM +kεm E+k ) (σM +kεm) (5.3) where E and εm represent the elastic modulus and themaximum bending strain in each cycle, respectively. It can be seen from Eqs. (5.2) and (5.3) that Wd increases in proportion to both σM and εm. The quantitative relationship between the fatigue life Nf and dissipated work Wd obtained fromEqs. (5.2) and (5.3) is shown in Fig. 14 for the rolled, and in Fig. 15 for the cast SMA. In both cases, the relationship between Nf and Wd can be matched approximately by a single straight line taking in both the alternating- and the pulsating-plane bending. In other words, the unified relationship between the fatigue life Nf and dissipated work Wd under both alternating- and pulsating-plane bending can be expressed by the same power function WdN λ f =µ (5.4) where µ and λ denote Wd in Nf =1 and the slope of the logWd− logNf curves, respectively. Results calculated from this equation are shown by the solid lines in Fig. 14, with values of λ= 0.45 and µ= 408MJ/m3, for the rolled SMA and in Fig. 15, with values of λ= 0.29 and µ=88MJ/m3, for the cast SMA.As can be seen, the solid lines show a good overall matchwith the test results. In the case of the rolled SMA in Fig. 14, the range 2 ·103 ¬Nf ¬ 5 ·10 4, in the number of cycles to failure represents the fatigue life under alternating-plane bending, while the range 2 ·104 ¬Nf ¬ 1 ·10 6 represents that under pulsating-plane bending. The corresponding ranges for the cast SMA in Fig. 15 are 1 · 101 ¬ Nf ¬ 3 · 10 4 cycles under alternating-plane bending and 1 ·104 ¬Nf ¬ 1 ·10 6 cycles under pulsating-plane bending. For the cast SMA, the value of Wd ismuch larger under alternating-plane bending thanunder pulsating-plane bending. That is why the fatigue life of a cast SMA is shorter under alternating-plane bending for a large maximum bending strain εm. To sum up, as already mentioned above, the unified relationship between Wd and Nf is of great usefulness for evaluating the fatigue life of a rolled SMA or a cast SMA under conditions of either alternating- or pulsating-plane bending. Fig. 14. Relationship between the dissipated work Wd and number of cycles to failure for the rolled SMA under alternating-plane and pulsating-plane bending 5.4. Temperature rise under cyclic bending The surface portion of the specimen is subjected to the effects of dissipated work in each cycle as discussed in Section 5.3. Therefore, the temperature of the specimen can be expected to increase under cyclic bending. The surface temperature in the central part of a rolled SMA strip was measured using a thermocouple under conditions of alternating-plane bending at a frequency of f =3.33Hz. The relationship between the incremental temperature rise ∆T and Performance and design of precision-cast shape memory alloy... 867 Fig. 15. Relationship between the dissipated work Wd and number of cycles to failure for the cast SMA under alternating-plane and pulsating-plane bending Fig. 16. Relationships between the incremental temperature rise and time for the rolled SMA at two different maximum bending strains, under alternating-plane bending at f =3.33Hz Fig. 17. Relationships between the temperature rise saturation points andmaximum bending strains, for the rolled and cast SMAs under alternating-plane bending at at f =3.33Hz time is shown in Fig. 16. In the first 20 seconds or so, the temperature increases rapidly, but once the amount of heat generated comes to match the amount released a saturation level is reached for ∆T . The relationship between the saturation point for the temperature rise ∆Ts and the maxi- mumbending strain εm is shown inFig. 17 for the rolled and the cast SMAsat f =3.33Hz.The saturation point increases in proportion to the rise in εm. The value of ∆Ts is a little higher for the cast than for the rolled SMA, rising to around 30K at εm =3%. The yield stress σM also increases with temperature, at a rate of 5MPa/K in the case of a TiNi SMA. For ∆T =30K, this gives a yield stress increase of 150MPa.This explainswhy the fatigue damage is large under alternating-plane bending for a large εm. In practical use in a brain spatula, the frequency will be low and the temperature rise only small, resulting in a much longer fatigue life than that obtained in the present study. 868 H. Tobushi et al. 6. Conclusions In order to develop the SMA-brain spatula, themechanical characteristics of theTiNi cast- and rolled-SMAs and the copper one used for the brain spatula were compared based on the tensile deformation properties, and the characteristics of the SMA-brain spatula were discussed. The fatigue properties of these materials were also investigated by pulsating- and alternating-plane bending fatigue tests. The results obtained can be summarized as follows. (1) Based on the yield stress and themodulus of elasticity of the copper and the TiNi SMAs, the bending deformation properties of the SMA-brain spatulawere estimated by assuming the condition to use the brain spatula as the bending of the strip cantilever.With respect to the SMA-brain spatula for the same length and width as the existing copper one, if the thickness of the conventional rolled-SMA spatula is 1.3 times as large as that of the existing copper-brain spatula, the SMA spatula can hold the same bending rigidity and can be bent by a smaller force than the existing copper one. If the thickness of the new cast-SMA spatula is 1.2 times as large as that of the existing-copper spatula, the SMA spatula can hold the same bending rigidity and can be bent by the same force as the existing copper one. (2) With respect to the alternating- and pulsating-plane bending fatigue, the fatigue life of both the copper and the SMAs in the region of low-cycle fatigue is expressed by a power function of the maximum bending strain. The fatigue life of the conventional rolled SMA and the new cast SMA is longer than that of the existing copper. The fatigue life of the rolled SMA is longer than that of the cast SMA. The fatigue life of the new cast and rolled SMAs in the pulsating-plane bending is longer than that in the alternating-plane bending. The fatigue life of the rolled SMA and cast SMA for alternating- and pulsating- plane bendings can be expressed by the unified relationship with a power function of the dissipated work. (3) The abovementioned characteristics of the SMA-brain spatula obtained in this studywill be substantially applied to the development not only for the brain spatula but also for other retractors and instruments used in other surgery operations. Acknowledgment The experimentalwork of this studywas carried outwith the assistance of students inAichi Institute of Technology, to whom the authors wish to express their gratitude. The samples of the existing copper- brain spatula were supplied byMizuhoCo., Ltd., to whom the authorswish to express their gratefulness. The authors also wish to extend their thanks to the Basic Research (C) of Grant-in-Aid of Scientific Research supportedbyJapanSociety forPromotionof Sciences and theProgramofSupport forAdvanced Corporate Networking supported by Chubu Bureau, Ministry of Economy, Trade and Industry, METI for their financial support. References 1. Chu Y.Y., Zhao L.C., Eds., 2002, Shape Memory Materials and Its Applications, Trans. Tech. Pub., Switzerland 2. Duerig T.W., Melton K.N., Stockel D., Wayman C.M., eds., 1990, Engineering Aspects of Shape Memory Alloys, Butterworth-Heinemann, 1-35 3. Funakubo H., Ed., 1987, Shape Memory Alloys, Gordon and Breach Science Pub., 1-60 4. Furuichi Y., Tobushi H., Ikawa T., Matsui R., 2003, Fatigue properties of a TiNi shape- memory alloywire subjected to bendingwith various strain ratios,Proc. Instn.Mech. Engrs., 217, Part L: Materials: Design and Applications, 93-99 Performance and design of precision-cast shape memory alloy... 869 5. Matsui R., Tobushi H., Furuichi Y., Horikawa H., 2004, Tensile deformation and rotating- bending fatigue properties of a highelastic thin wire, a superelastic thin wire, and a superelastic thin tube of NiTi alloys,Trans. ASME, J. Eng. Mater. Tech., 126, 384-391 6. Otsuka K., Wayman C.M., Eds., 1998, Shape Memory Materials, CambridgeUniversity Press, Cambridge 7. Saburi T., Ed., 2000, Shape Memory Materials, Trans. Tech. Pub., Switzerland 8. Shigley J.E., Mischke C.R., 1989,Mechanical Engineering Design, 5th ed.,McGraw-Hill, New York, 270-274 9. TobushiH.,NakaharaT., ShimenoY.,HashimotoT., 2000,Low-cycle fatigue ofTiNi shape memory alloy and formulation of fatigue life, Trans. ASME, J. Eng. Mater. Tech., 122, 186-191 10. Tobushi H., OkumuraK., NakagawaK., Takata K., 2001, Fatigue properties of TiNi shape memory alloy,Trans. Mater. Res. Soc. Jap., 26, 1, 347-350 11. Tobushi H., Tanaka K., Kimura K., Hori T., Sawada T., 1992, Stress-strain-temperature relationship associatedwith the R-phase transformation in TiNi shapememory alloy, JSME Inter. J., Ser. I, 35, 3, 278-284 12. Yoshimi Y.,KitamuraK., TokudaM., InabaT., Asai J.,WatanabeY., 2008,Ti-Ni shape memory alloys precision casting products and its process, Proc. Int. Conf. Shape Memory and Superelastic Tech., 387-396 Konstrukcja i właściwości użytkowe odlewanej szpatułki chirurgicznej wykonanej ze stopu z pamięcią kształtu Streszczenie W pracy opisano problem projektowania szpatułki używanej w chirurgii mózgu, wykonanej ze stopu z pamięcią kształtu (SMA). Przedyskutowano charakterystyki zginania nowej szpatułki wykonanej z pre- cyzyjnie odlewanego stopu TiNi.Wyniki badań podsumowano w następujący sposób: (1) w porównaniu do konwencjonalnych, miedzianych szpatułek o tej samej długości i szerokości, jeśli odlewana szpatułka SMA jest 1.2 raza grubsza od miedzianej, lub 1.3 raza, gdy wykonana została przez walcowanie, instru- ment ten będzie posiadał taką samą sztywność giętną i wytrzyma niemalże to samo obciążenie zginające, jak w przypadku narzędzia miedzianego; (2) wyrażając wytrzymałość zmęczeniową szpatułki miedzianej w rejonie niskocyklicznych obciążeń jako funkcję potęgową maksymalnego odkształcenia przy zginaniu, stwierdzono, że szpatułki SMA wykazują dłuższą żywotność od konwencjonalnych.W obydwu przypad- kach – odlewanej i walcowanej szpatułki SMA – wytrzymałość zmęczeniowa jest większa dla płaskiego jednostronnego zginania od wytrzymałości przy zginaniu przemiennym. Manuscript received January 30, 2012; accepted for print February 12, 2012