Strain estimation by electron back scattered diffraction

The quality of the Kikuchi patterns obtained for example by electron backscattered diffraction is altered when the dislocation density increases in strained materials, and consequently if can theoretically be correlated to the strain for monotonic deformation pathes. For that purpose, in a first step, it is necessary to define a qualify factor in the Hough space or with the Burns method currently used when the Kikuchi patterns are automatically analysed. Then, and it is the most important point, the quality factor must be linked to the strain through a "calibration" curve independent of crystallographic orientation. The main objective of this study is to test the reliability of this method in the "simple", case of the hardness test. This test is simulated using the finite element method and the obtained results are compared to those measured by electron backscattered diffraction. Several "calibration" methods are defined. The first one consists in defining a calibration curve from the simulation and then verifying it for some other tests with different loads inducing different strain scales. Two other experimental methods have been also applied. In the first case, it is assumed that the behaviour of an anisotropic polycrystalline material can be represented by a I calibration curve defined from single crystals with the same orientation and deformed in channel die with several strains. In the second case, the "calibration" curve is determined from a torsion test that is very interesting since the relationship between the plastic strain and the sample radius is theoretically linear, Because of some difficulties in the numerical analysis such as the choice of parameters in the rheological and the friction laws useful for the calculation or the choice of the analysed subsurface (the strain gradient is very important in the sample thickness), it appears difficult to use the finite element method to determine a "calibration" curve for the hardness test. On an other hand, the results obtained from the two experimental "calibration" methods give similar results. They are also close to those obtained with the simulation. However, and taking into account in particular the remark concerning the choice of the subsurface, it becomes difficult to evaluate the precision of strain measurements. To estimate exactly this precision, it is necessary to modify the comparative procedure by choosing for example other laboratory tests where the strain is more evenly distributed in the sample and so to prevent any ambiguity for the comparison with the strains measured by electron backscattered diffraction.