Computation and Sensitivity Analysis of the Deformation-Gradient Tensor Reconstruction in Dark-Field X-ray Microscopy

TL;DR

This paper develops an inverse modeling framework for dark-field X-ray microscopy (DFXM) that enables the reconstruction of the full deformation gradient tensor in bulk materials, enhancing strain measurement capabilities at the nanoscale.

Contribution

It introduces a novel formalism and computational framework for reconstructing the deformation gradient tensor from DFXM data, including sensitivity analysis for error estimation.

Findings

Reconstruction of the full deformation gradient tensor is feasible using oblique diffraction geometry.

The framework allows forward calculation of angular shifts and inverse reconstruction of strain and rotation tensors.

Sensitivity analysis quantifies errors in tensor component reconstructions.

Abstract

Spatially resolved strain measurements are crucial to understanding the properties of engineering materials. Although strain measurements utilizing techniques such as transmission electron microscopy and electron backscatter diffraction offer high spatial resolution, they are limited to surface or thin samples. X-ray diffraction methods, including Bragg Coherent Diffraction Imaging and X-ray topography, enable strain measurements deep inside bulk materials but face challenges in simultaneously achieving both high spatial resolution and large field-of-view. Dark-field X-ray Microscopy (DFXM) offers a promising solution with its ability to image bulk crystals at the nanoscale while offering a field-of-view approaching a few hundred $\mu$m. However, an inverse modeling framework to explicitly relate the angular shifts in DFXM to the strain and lattice rotation tensors is lacking. In this paper, we develop such an inverse modeling formalism. Using the oblique diffraction geometry, enabling access to noncoplanar symmetry-equivalent reflections, we demonstrate that the reconstruction of the full deformation gradient tensor ($\mathbf{F^{(g)}}$) is possible. We also develop the computational framework to both forward calculate the anticipated angular shifts and reconstruct the average $\mathbf{F^{(g)}}$ for an individual pixel from DFXM experiments. Finally, utilizing the established formalism and computational framework, we present methods for sensitivity analysis to relate individual components of the rotation or strain tensor to specific angles of DFXM. The developed sensitivity analysis also enables explicit computation of the errors associated with the reconstruction of each component. The formalism, the computational framework, and the sensitivity analysis established in this paper should assist both the interpretation of past DFXM experiments and the design of future DFXM experiments.