Prakt. Met. Sonderband 30 (1999) 361 
« Dlastıc and atomic milling (Maxmill 360B) in order to avoid the influence of liquid environment with H*. 
Mechanical properties were tested using an Instron 1196 machine equipped with a heating element. 
Vii Temperature of the TiNi samples was measured by Ni-NiCr thermocouples, which were spark- 
I welded to the central part of specimens. Stability of trained specimens was tested using a specially 
Cale designed testing device in which the specimens were working against biasing spring with stiffness 
ne of 15 Nmm'. This device enabled us to measure the extent of reversible bending strain in 
genen dependence on the number of performed working cycles. The critical extent of reversible strain €= 
no oA 2% was defined as a measure of functional element. Mechanical properties, transformation 
¢ en | temperatures and microstructure, respectively, were inspected after performing 10,000 working 
fuer cycles. 
J 
ph, The thermomechanical training consisted of bending of the samples (€ = 4 %) at room temperature 
vd he and following 20 thermal sequences each of them composed of heating the material to 135°C and 
SE subsequent cooling to room temperature (20 to 22°C). Two training procedures were applied: i) so- 
or called “soft training” in which TiNi elements were constrained by a biasing spring with a stiffness 
E The of 100N mm? (specimens are further referred as A type specimens). ii) “Hard training” in which 
ene ! samples were totally constrained by a rigid frame without a possibility to change their shape during 
SL heating (specimens B). 
ical Results and discussion 
The transformation temperatures Tr, Mj, and A¢ of investigated materials are summarized in Table 
1. The determined changes of transformation temperatures well correspond to previously stated 
a principles (14 to 16). 
NI ase 
pa | Specimen ' Te [°C]_ | M, [°C] | A{°C] 
Rang As-received (not trained or tested) 43 | 39 0 
+s After “soft training”, samples A | 50 | 31 185 
After “hard training”, samples B | 60 130 190 
After 10,000 working cycles, samples A | 53 34 5 
ers After 10,000 working ei Samples B [70 36 (3 
N Table 1: Transformation temperatures of investigated alloys. 
il 
extent The M; temperature decreased after the training. Both soft and hard training processes are linked 
of an with work hardening (increasing of dislocation density). The decrease of M; is attributed to the 
nechanica lowering of the interface mobility (1, 15, 16) and to the increase of transformation enthalpy as well 
as entropy, whereby the entropy chances are more pronounced and the equilibrium temperature 
between austenitic and martensitic phases decreases (20). The increase of Af temperature after 
training is dominantly connected with lowering of the interface mobility during recovery at heating. 
applied by In accordance with earlier presented results, the sweeping effect occurs during repeating of the 
vera old transformation cycles (14). Contrary to the training procedures, the movement of interfaces 
wies WEI becomes easier with increasing number of working cycles and this is connected with the subsequent 
To (he increase of M; and lowering of Af temperature. Beside this influence, the effect of the elastic energy 
f te BY stored upon thermoelastic transformation plays an important role (21). The effect of elastic energy 
lished 7 was summarized by Ortin et al. (22) and it can be expressed for thermoelastic transformation 
(19). The according to Kurdjumov (23) and Olson (24) in the following form: Agcn+2Age=0. Age 
fols Were corresponds to the change of chemical free enthalpy related to parent phase — martensite 
Num) ar’ transformation and 2Ag,, is the elastic energy stored in matrix after formation of martensite. This 
15 V and equation means that half of the chemical free enthalpy change is stored as the elastic energy in the 
r= Sum) matrix. This elastic energy resists the forward transformation and assists the reverse transformation.