|
|
|
|
Damping in H-free and H-doped Shape Memory Alloys
Usually, four temperatures characterize the MT; these are: the start (Ms ,) and finish (Mf ) temperatures for the direct MT, on cooling, the start (As) and finish (Af) temperatures for the reverse MT, on heating. The MT is a displacive first-order structural transition associated with short-distance collective atomic motions. In b-phase alloys the MT is usually regarded as being due to the combination of shear and shuffling strains and gives rise to a number of long-period martensitic structures of orthorhombic symmetry, such as B19, 3R, 6R, 9R etc.
The main aim of this paper is to outline some of the most recent progresses made in the investigation of the damping properties of NiTi and NiTiCu SMA containing various amount nH of hydrogen, at kHz frequencies. In this frequency range the transient contribution IFTr to the overall damping is expected to be negligible. Thus, the study of the stationary contributions IFPT (associated to the martensitic transition) and IFint (intrinsic to the austenite or martensite phase), turns out to be easier since elaborated procedures to separate the transient from the stationary contributions are not required, contrary to the case of data at 1 Hz.
DAMPING OF H-FREE SHAPE MEMORY ALLOYS
Damping measured in the austenite is small, typically of the order of 10-4 , regardless of the alloy composition, the vibration frequency and, to some extent, the thermo-mechanical history of the material. Also plastic deformations seem not to increase the background damping beyond the 10-4 range. On approaching the MT from high temperatures the damping starts increasing at 20-40 degrees above the start temperatures of the B2-B19', B2-R and B2-B19 transitions. This damping increase probably arises from the interaction of the applied acoustical or ultrasonic strains with the order parameter fluctuations.
The IF int is higher in the martensite than in the austenite and depends on the thermo-mechanical history of the material. As seen in fig. 1, a well developed broad peak PAM (FWHH @ 50 K) associated with the B2-B19' transition occurs in a Ni50.8Ti49.8 alloy, prior solubilised (1273 K) then water quenched. The IFint of the B19' martensite at intermediate temperatures (150 K) is small and a minor peak (Pd) appears below 100 K. PAM is steeper on its high temperature side and wider than the transition region (Ms - Mf @ 15 K). The low-temperature tail of the peak extends far below Mf , indicating that at least part of the damping spectrum is not directly associated with the transition. Peak Pd is thermally activated (W=0.11-0.13 eV, t0 =10-10s) and most likely it is a Bordoni-type relaxation due to dislocation arrays introduced by quenching.
With changing the frequency from 1.2 to 6.8 kHz no shift occurs on the high temperature side of PAM , while a shift of only few degrees takes place on the low temperature side. This behavior suggests that PAM may result from the superposition of two processes, one frequency dependent the other not. Likely, the frequency dependent relaxation might be the counterpart of peak P2 occurring in ferroelectrics which is known to be due to viscous motions of domain walls. If this assumption is correct the frequency dependent component of PAM is to be ascribed to stress-induced motions of phase interfaces and/or of twin boundaries. Thus, the actual contribution IFPT of the transition would only be the higher-temperature component of the PAM spectrum. This component arises from the coupling of the applied stress-field to the order parameter of the MT.
The main damping features of the binary alloys Ni50.8Ti 49.2 and Ni49Ti51 , prior solubilised, then deformed and aged (673 K) can be summarised as follows (fig. 2): a) The damping IFint of both the R and the B19' martensites is high; b) a broad non-thermally activated peak PTWM appears at around 125 K; c) a stepwise increase in damping is found on cooling and from time to time a narrow peak occurs in the proximity of the B2-R transition.
Fig. 1. IF spectrum and related Young's modulus in a solubilised and quenched Ni50.8Ti49.2 alloy.
 Fig. 2. Q-1 and E in a solubilised then deformed and aged Ni50.8Ti49.2 alloy.
EFFECT OF HYDROGEN DOPING
Hydrogen has a dramatic effect on the Young's modulus and on the damping spectra of NiTi based alloys. As can be seen in fig. 3a, where the plotted data refer to the Ni50.8Ti 49.2 alloy, the height of the overall PAM increases by a factor of six with increasing the H content nH (nH =H/Me at.) from 0 to 0.008. The peak temperature in the H-doped is slightly lower than in the undoped material, suggesting that H enhances the relaxational component P2 of the overall peak PAM . Assuming that this component is due to stress-induced viscous motion of phase interfaces and/or of twin boundaries, the enhancement of the peak will be caused by an increase in their viscous drag coefficient. It is conceivable that the relaxational stationary component sometime observed in non-intentionally loaded materials may be partly due to H present in the starting material or captured during solubilisation in argon and/or water quenching. The idea that H might have been responsible, to some extent, of the overall peak PAM in the solubilised material (fig.1) is supported by the fact that the same specimens outgassed and furnace cooled under vacuum exhibited a much smaller peak PAM .
With further increasing the H content the overall peak PAM decreases and a second peak, PH , grows above Ms in the B2 austenite (fig. 3b). PH is thermally activated (W=0.5 eV; t 0 =(1 ± 4)x10 -12 s) and its height linearly increases with nH . Thus, this peak can be interpreted as being due to stress-induced changes in the short-range order parameter of H dipoles in solid solution (Snoek-type effect) or within the metal hydrides (Zener-type effect), which are known to form at higher H contents. The concomitant decrease of PAM (fig. 3b) is most likely due to pinning of phase interfaces and/or of twin boundaries by isolate H atoms or by hydride particles. The presence of hydrogen reduces the IFint and increases the Young's modulus of the B19' martensite.
|
|
|
|
|
|
|
|
|
|
|
These two related effects can also be easily accounted for in terms of pinning processes of twin boundaries by H. Thus, the experiments strongly support the idea that the wide stationary peaks observed in non-intentionally H-doped material, when associated with low dissipative martensites, are partly due to viscous motions of interfaces and/or of twin boundaries interacting with residual H impurities.
Fig. 3a. Effect on Q -1 and E of H loading.
Fig. 3b. As in 3a but for heavier H loading.
Effects of hydrogen similar to the ones found with the binary alloys were also observed with the ternary alloys Ni40Ti50 Cu10 , prior solubilised and quenched, then deformed and aged (673K) (see fig. 4), as well as with Cu30Ti50Cu20 , prior solubilised (1173 K) then simply water quenched. Big partially overlapping peaks PH and PTWH are introduced at around 250 K and 300 K, respectively, the Young's modulus is increased while the IFint of the B19 martensite at low temperatures is decreased. All these observations confirm the idea that peak PH is due to H dipoles, PTWH to viscous motion of phase interfaces and/or of twin boundaries and that isolated H atoms or hydride particles act as pinning points for these two-dimensional defects, similar to what happens with the B19' martensite. Finally, it can be concluded that the big decrease in the Young's modulus between Ms and Mf represents a "modulus defect" associated with mobile interfaces and twins, similar to the case of dislocations in plastically deformed metals.
Fig. 4. Effect of H on Q -1 and E of a Ni40Ti 50Cu10 alloy.
 Fig. 5 Comparison of data on H doped alloys.
It has been known for some time that NiTi based alloys are high damping materials. Actually, the stationary damping contributions IFint and IFPT are not very high (typically of the order of 0.01). Thus, it is of interest to call attention to the fact that H doping indeed improves the damping properties of NiTi alloys. In fig. 5 a comparison is made of the damping peaks observed in the various H doped NiTi(Cu) alloys. It is clear from this figure that PTWH , major component of PAM follows the transition along the temperature scale, while PH does not. This different behavior is consistent with the assumption that the temperature dependence of the relaxation time of peak PTWH is the same as that for domain wall relaxation P2 occurring in dielectrics, while that for PH is exponential in 1/T. The observed relaxation peaks are stable during ageing at temperatures at least as high as 373 K and are isothermal effects. Ti-rich binary alloys and NiTiCu ternary alloys, especially for compositions up to 10%, are well suited for applications as their transition temperatures are close to room temperature and their workability is reasonably good. Hydrogen can easily be introduced into NiTi(Cu) alloys either by electrolysis or by gas-phase solid-surface reaction and does not come out after doping unless heated up to temperatures higher than about 500 K.
In conclusion, H doping appears to be a viable technique to improve the performance of NiTi alloys when employed as high damping materials. Moreover, it is to be emphasised that H is also a powerful tool to probe the elastic and anelastic properties of NiTi(Cu) alloys.
|
|
|
|
|
|
|
|
|
References
- Coluzzi B., Biscarini A., Campanella R., Trotta L., Mazzolai G., Tuissi A. and Mazzolai F.M., 1999, Acta Mater., 47 , 1965.
- Liang, X. L., Ren, X., Shen, H. M., Wang, Y. N., Otsuka, K. and Suzuki, T., 2001, Scripta Mater., 45 , 591.
- Huang, Y. N., Wang, Y. N. and Shen, H. M., 1992, Phys. Rev. B, 46 , 3290.
- Nishida, M., Ohgi, H., Itai, I., Chiba, A. and Yamauchi, K., 1995 Acta Mater., 43 , 1219.
- Biscarini A., Campanella R., Coluzzi B., Mazzolai G., Trotta L., Tuissi A. and Mazzolai F. M, 1999, Acta Mater., 47 , 4525.
- Biscarini, A., Coluzzi B., Mazzolai G., Tuissi A. and Mazzolai F. M., 2003, J. Alloys and Compds., 355 , 53
- Mazzolai F. M., Biscarini A., Campanella R., Coluzzi B., and Mazzolai G., Rotini A., Tuissi A., 2003 Acta Mater., 51 , 573.
(top) |
|
|
|
|
|
