122 Prakt. Met. Sonderband 41 (2009) 
variants that are essentially guided by the internal stress field [5]. During the training process, the with 
resulting stress field assists the formation and growth of preferentially oriented martensite variants melt: 
when the martensitic transformation proceeds in a specimen on cooling [6]. desis 
It is well known that in NiTi alloys a change of the transformation temperatures and appearance of 
an intermediate phase can take place due to thermomechanical training, thermal cycling, aging at an 
appropriate temperature, and other processing techniques [7]. The precipitation of an intermediate 3 
phase in NiTi brings about an excellent shape memory property with a small hysteresis. Therefore, 
the presence of the intermediate phase is useful in engineering fields [8]. 
For practical applications, control of transformation characteristics i ite important. Therefore, th Som: 
practical app , sfo onc csisqu p ore, the 
main object of this paper was to study the influence of training parameters on the efficiency, be p 
transformation temperatures, and intermediate phase of Ni-50.3 at. % Ti melt-spun ribbons. N A 
devi 
fully 
2 Experimental Procedure obse 
thick 
The NiTi alloy ingots were prepared by vacuum arc-melting on a water-cooled copper hearth in a thick 
reduced Ar atmosphere. To ensure homogeneity, arc-melting was repeated three times for each show 
alloy. Spec 
The melt-spun ribbons were produced under a 200 mbar He atmosphere using quartz-glass crucibles cont: 
with a nozzle diameter of 0.9 mm, coated internally with Y,O3. By applying an Ar overpressure of struc 
90 mbar within the crucible, the melt was ejected onto the surface of a polished Cu wheel (200 wate 
mm diameter) having a circumferential wheel speed of 5 to 30 ms. The distance between the obtai 
nozzle and the wheel surface was 2 mm. 
For inducing a two-way shape memory effect there are several methods; one is based on 
temperature cycling under constant stress which is an effective method for introducing a TWSME 
and allows a good control of the training parameters [9, 10], however is limited to linear shape 
changes. 
In this study, a bending training method was applied to NiTi ribbons which consists of several 
thermal cycles under constant bending strain followed by load-free thermal cycling. The value of 
spontaneous shape change during heating and cooling was recorded by photographs. During the 
constrained training and subsequent free thermal cycling, temperature was changed from room 
temperature (RT) to 250 °C. While the TWSME was executed several hundred times, the changes 
in the deformation behaviour and the stability of the effect were continuously observed. 
Specimens for metallographic investigation were cut from the longitudinal cross sections of melt- 
spun ribbons. They were ground on SiC paper to a final mesh size of 2400 and polished. The 
samples were etched in a solution containing HF:HNO;:CH3;COOH = 2:4:5 with etching times Fig. 1 
between 2 and 25 seconds and then studied by optical microscopy. For TEM experiments, thin discs 24:5 
(3 mm in diameter) were mechanically ground to a thickness of 60 pm and polished in an 
electrolyte consisting of 20% sulphuric acid and 80% methanol at 273-278 °K using a twin jet 3.1 
electro-polisher according to [11]. Ton-milling was then performed on the samples for 30 minutes. In th 
The energy of the Ar+ ions was 3.6 keV, the angle to the surface of the specimen was +4°, and the (abo! 
specimen was kept at a rotation rate of 3 rpm during ion-milling. Conventional TEM studies were time: 
performed on a Philips CM12 at an acceleration voltage of 120 kV. incre 
The phase transformation temperatures of trained ribbons were measured using a Mettler DSC 821e am 
device (Mettler Toledo GmbH, Schwerzenbach, CH). Thermograms were recorded under static air defo 
from —40 to 180 °C at a heating/cooling rate of 10 K min’. A circular sample disc was cut with a that 
punch (5 mg), put in a 20 ul aluminium pan, and sealed with a perforated lid. Tensile specimens