Prakt. Met. Sonderband 52 (2018) 209 
angm ie 
ul i From the tension-torsion specimens longitudinal cuts are made in the area of the gauge 
Sra i aly length. The samples are then polished and etched by standard methods. Before the EBSD 
art a recordings some of the polished surface was removed using a soft field ion beam (FIB) 
Sima method (ion slicing). The typical resolution of the EBSD measurements is above the 
N die te expected dimensions of inter-lath retained austenite films. EBSD scans confirmed the 
‚sten N assumption that there is no noteworthy amount of retained austenite in the samples at 
be dig Ya room temperature. 
har 7 In the current recordings a step size of 50 nm has been used and an area comprising 100 
ak A fe prior austenite grains (PAGs) has been investigated for each specimen as a compromise 
Ws between resolution and statistical significance. 
Tansfomatiy 
nant select 3. Results 
| Ses fhe i 
a The average prior austenite grain size (PAGS) of all three specimens was found to be 
4 0s f 15 ym estimated from the fully transformed microstructure. The martensite block size is 
after te Wh about one quarter of that, but blocks are much more elongated in shape than PAGs. (Fig. 
in on 1) gives an overview of the microstructural state of the three specimens in terms of 
modation Js orientation imaging microscopy (OIM) maps. To obtain an optimal contrast, the inverse 
Sh Mv pole figure (IPF) color-code has been overlaid with the image quality (IQ) parameter as 
hai y Case grayscale. The latter is also useful for gaining some insight into the distribution of strain in 
vior Can be a microstructure. 
3 higher initial 
0 form certain TAO3 (higher annealed) TR32 {torsion 120 MPa) TR25 (tensile 120 MPa) 
Ure, boundary 
tain load after 
nens are, after 
ooled to room 
cannot fully be 50 um 
re heated fo a = 
to ensure ful Fig. 1: Overview images of the microstructures of the three samples. Color-code: inverse 
2 at a cooling pole figure (IPF), grayscale: Image quality (IQ). Left: weakest sub-block heterogeneity due 
3 lower cooling to higher annealing. A shear load (center) leads to a much more pronounced sub-block 
ing before the heterogeneity compared to a tensile load of the same magnitude. 
een thermally The heavy plastic deformation due to the repeated TRIP effect involves block-substructure 
size remains sliding and rotation. For the higher annealed specimen the lowest degree of sub-block 
°C and one a orientation heterogeneity has been found as expected. However, although (Fig. 1) 
at the lower generally suggests a higher dislocation density for the torsion sample (TR32), the 
ı and the other maximum value of geometrically necessary dislocations as evaluated using an OIM 
equivalent {0 software with the according calculation method is almost constant at 7.3 x 10% [m1] in all 
yr temperature three specimens. The measurement method can be found in [12] and considers both edge 
and screw dislocations. Maximum misorientations of 15° and first nearest neighbor pixels 
have been considered for the gradient calculation. (Fig. 2) demonstrates that even in the