Development of the Tensor CT Algorithm for Strain Tomography Using Bragg-edge Neutron Transmission

The tensor CT algorithm for strain tomography using the Bragg-edge neutron transmission spectroscopy is presented. Crystal lattice strain is not scalar but is a tensorwhich changesdepending on the observation angle. Therefore, since traditional“scalar” CT algorithms cannot be applied to tomography of strain, the development of a “tensor” CT algorithm is needed. Aiming at further developments in the future, we first developed a ML-EM based versatile tensor tomography using ofa simple algorithm withsmall restriction. The basic concept is to simultaneously reconstruct multiple strain-tensor components (scalar quantities of normal strain and shear strain) existing at a certain position. In the actual CT image reconstruction, it is important to consider the angular dependence of each tensor component. Through the simulation studies on axially-symmetric and axially-asymmetric distributionscomposed of two strain components and experimental demonstration using the axially-symmetric VAMAS standard sample, we found some important points for strain-tensor tomography. The angle-dependent back-projection procedure of ML-EM is indispensable fortomography of each tensor component,butsuch function also causes animage distortion which can average each strain value along each strain direction. Also, we found that the optimization of the angle-dependent back-projection procedure is important for further improvements of the tensor CT algorithm.