360 Prakt. Met. Sonderband 30 (1999)
assumption that the generation of dislocations is linked with the “true plasticity”, a residual plastic
strain is connected with applied training.
The second mechanism is based on the stabilization of preferentially oriented martensitic variants,
which are retained after the heating above the original A¢ temperature. Only the most favorably
oriented variants of martensite grow during applying an external load (1). However, if the applied
load is high enough or repeated, the deformation process is also accompanied by generation of
dislocations. According to the assumption of stabilized martensite formation, the generated
dislocations do not allow the martensite to shrink and disappear completely after heating to Ar
temperature. During cooling, the stabilized martensite grows preferentially. Moreover, the internal
stress related to the growth of preferentially oriented martensite modifies the arrangement of further
variants, which also form in preferential orientation during subsequent cooling. At heating, the less
stable variants disappear, however, those originally stabilized by presence of dislocations remain in
the matrix. In other words, this mechanism requires an incomplete reverse transformation and the “
residual deformation observed after training is also considered as a prerequisite of the TWSME. :
Authors (4, 10) tried to distinguish between so called intrinsic and extrinsic TWSME. They
disproved two previous explanations. Following their assumptions, the most important influence of
training procedure is not the development of internal oriented stresses, but rather a microstructural
anisotropy inducing a thermodynamic anisotropy. In other words, the crystallographically os
equivalent variants are not thermodynamically equivalent after training.
The presence of “ghost martensite” was also discussed as a possible reason for TWSME (11).
Different opinions were published concerning the influence of R-phase formation in TiNi based
alloys on the stability of TWSME. According to Stachowiak and McCormick (5), the formation of
R-phase diminishes the extent of obtained reversible strain as well as the stability of TWSME. In
contrast to this conclusion, the authors (12, 13) observed enhanced stability of TWSME in cases
where the R phase formation precedes the martensitic transformation B2 — B19".
The stability of TWSME is an even more important parameter than the extent of reversible
deformation and the generated stress. The complex degradation mechanisms occur in TiN alloys
during repeated heating and/or loading cycles (14). The accompanied generation of dislocations
(work hardening) strongly influences the maximum level of generated stresses and the extent of |
reversible strain (15 to 17). The presented paper is devoted to the study of two different training N
procedures and to the influence of different work hardening stage on the substructure. mechanical
properties and stabilitv of TiNi elements with TWSME. &
Experimental ;
The experimental material used in this study was a commercial Ti - 50.4 at. % Ni alloy supplied by
FIBRA Ltd. (Czech Republic) in the form of wire with diameter d = 5 mm produced by several cold
drawing steps with intermediate annealing in vacuum at 800°C/30min/water (18). The wires were 7
deformed with the constant bending strain of 4%. The transformation temperatures Tr (the wo
temperature at which the rhombohedral R-phase starts to form), Ms (the temperature of the B19' =
martensite start) and As (the temperature at which austenitic B2 phase finishes) were established by wn
use of the measurements of electric resistance versus temperature dependencies (19). The “
microstructure of specimens was studied by a TEM Jeol 200CX and Hitachi H7100. Thin foils were i
finished by two different methods: i) spark cutting (thickness = 400um), grinding (= 100um) and om
polishing using a twin jet polisher in an electrolyte of HCIO4 and CH3COOH at ~15 V and
temperature T=0°C. ii) diamond saw (thickness = 200um), prethinning in dimpler D500 (= 50um),
