Rapid determination of local composition in quasi-binary, inhomogeneous material systems from backscattered electron image contrast

Quantification of different phases in multiphase materials and the determination of compositions in material systems with extended solubility of the components are crucial for the engineering of material properties. In the field of thermoelectrics, multiphase materials are of considerable interest due to the additional scattering caused by different phases, reducing the lattice thermal conductivity and consequently improving the thermoelectric performance of the material. Size, shape, amount, and composition of secondary phases need to be assessed and optimized to be able to tune the material properties. SEM-EDX is the state-of-the-art method for such phase analysis. It suffers, however, from a low throughput and a limited spatial resolution. In this work, we demonstrate a simple relation between the grey value in backscattered electron images and the chemical composition allowing for a rapid method for phase quantification (with respect to composition as well as compositional fraction) in quasi-binary multiphase or wide solubility material systems. Applying this method to Mg2SixSn1−x, the Si and Sn content of each individual point on the BSE images can be calculated. The introduced SEM image analysis is much faster compared to EDX mapping as it requires only two EDX point measurements for calibration. Moreover, it also provides a superior spatial resolution