THE INFLUENCE OF STRAIN-RATE, AMPLITUDE AND TEMPERATURE ON THE HYSTERESIS OF A PSEUDOELASTIC Cu-Zn-Al SINGLE CRYSTAL

The hysteresis loop, described during the formation of stressinduced pseudoelastic martensite in a Cu-Zn-A1 betaphase single crystal, was studied as a function of strain-rate, deformation amplitude and temperature. The aPtm, the stress at which the transformation starts at a given constant temperature, is strain-rate independent but the hysteresis described by the stress-strain curve shows a maximum at intermediate strain-rates (3.3 x sec-l) For very low strain rates (3.3 x sec-l) the relative energy-loss ( A W / W ) was independent of amplitude. The amplitude-dependence was the strongest at intermediate strain-rates. The absolute hysteresis, i.e. energy loss, measured at constant strain-rate seems to be independent of temperature in the region (Ms+60, Ms+llO). The change in hysteresis as a function of strain-rate can be attributed to the heating and cooling of the sample due to the exothermic character of the beta-martensite transformation and the reverse endothermic transformation. At very low strain-rates the transformatio? occurs isothermally so that nearly no hysteresis is found. Only at ver:: high strain-rates the sample is deformed adiabatically. Introduction. The beta-to-martensite transfornation in Cu-Zn-A1 alloys is associated with a damping peak and the internal fric-tion in martensite is much higher than in the beta-phase (1). However, this behaviour was only measured when the temperature rate was different from zero. However, at dT/dt equal to zero, this is at constant temperature, a relaxation occurs so t3at the internal friction in the martensitic phase clecreases wit!? time and even the internal friction peak nearly disappears with time (2,3). This effect was also already noticed in CuAl-Lii (4) and Ti-Ni alloys (5). It is generally accepted that in the martensitic state tie movement of interfaces greatly contribute to the enerpj loss (I), however the influence of these movements on the crystal structure is not clear yet. In the transformation re~ion, the internal friction is due to the gi-owing or disappearing of one uhase into the other and, due to the measuring technique, combined with an alternating movement of the interfaces between the two phases. Above the temperature Af the martensitic phase can also be induced by stress or strain. At this temperature, the beta-phase is stable in a stress free sample, upon removing the stress, the stress-induced martensite will disappear. This SIX gives rise to large el-astic cleformations. This ~seu6oeiastic effec2 is more pronounced in single crystals and is much lower in polycrystals ( 5 ) . The study of t!?e hysteretic loop, described by a pseudoelastic curve of a L3 single crystal may reveal Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:19815155 C5-1008 JOURNAL DE PHYSIQUE important information concerning the mechanism and energy dissipation during the stress-induced growth of martensite. Experimental procedure. A Cu-Zn-A1 alloy with composition 57.24 at % Cu, 21.S at % Zn and 11.10 at % A1 was made by induction furnace melting an eigth kilogram ingot to get a good homogenisation. The ingot was hot extruded to a bar of 12 mm diameter, a sample of 250 mm length was taken for inclusion in an evacuated quartztube to make a single crystal with the modified Bridgmen-technique. A heat-treatment of 10 min at 800°C followed by water-quenching was applied to the single crystal. The transformation temperatures, K =-42"C, W --5Z0C, A =-50°C, A --22OC, were ff determined by DSC and resistivity measurements. C, tensile sample was made from the single crystal with the following dimensions : length 150 mm, diameter of the grips 12 mm, gauge length 40 mm, gauge diameter 6.25 mm. THe sample was tested in an Instron-25 Ton tensile machine with a heating/cooling chamber. The strain rates applied to the sample were between 3.3 x sec-I (0.05 mm/min) and 6.7 x 10-~sec-~(100 mm/min). The strain was measured by a 25 mm/50 76 extensometer. A chromel-alumel thermocouple was pressed against the sample. The orientation of the tmsile axis was determined by the Iaue back reflection method and is shown in the insert of fig. 1. Results. Figure 1 gives the stress-strain curve at room temperature for complete stress-induced transformation, with a pseudoelastic elongation of 8.65 %. The stress necessary to induce martensite, is temperature-dependent (6). This relationship is shown in figure 2 and is determined at a strain rate of 1.3 x10-~ sec-I . The line is calculated by the linear regression method. oP =1.93T-442 with o in IPa and T in K. At three different temperatures, 293 K (RT), 323 K and 353 K, pseudoelastic loops were described at different amplitudes (1, 2, 4 and 8 % deformation) and different strain rates. Figure 3 shows the curves taken at room temperature for four different strain rates and .8 % deformation. From an analysis of all these curves it can be concluded that : 1. up)m is strain-rate independent. The transformation starts always at the same stress and strain. 2. The hysteresis described by the loops shows a maximum at intermediate strain rates. Figure 4 shows the change of the relative hysteresis, which is the area enclosed by the hysteresis loop devided by the area under the "horizontal" part of the loading curve, as a function of the natural logarithm of the strain rate at 8 % deformation and at two temperatures. 3. The loading curve shows a strain hardening effect increasing with increasing deformation rate. 4. The hysteresis H, as defined by Otsuka et al. (7), being the difference between the stress at eP'm of the loading and unloading curve, is first increasing but decreases again at higher strain rates. 5. The hysteresis, proportional to the area of a closed loop, seems, within the experimental error, and at the same strain rates to be temperature independent. Figure 1 The total pseudoelastic loop at room temperature and at 6 = 1.3 x sec-I. The orientation of the tensile axis is shown in the insert.