Reconstruction of particle size distribution from cross-sections

Controlling the microstructure enables higher energy density and lower energy consumption of a battery. Although particle size distribution is an important property of microstructures, its study is hindered by limited analytical tools. In this study, we precisely estimate the 3-dimensional (3D) spherical size distribution from a 2-dimensional circular size distribution. Here, we introduce the least absolute shrinkage and selection operator (LASSO) regularization method to handle the existing issues in 3D reconstruction efficiently. Using a virtual structure from various predefined distributions, we demonstrate that the LASSO regression outperforms other regularization methods in predicting the original distribution. Finally, we suggest an effective number of cross sections, that is, the minimum required number of cross sections, for 3D reconstruction consisting of spherical particles.

Automated summary

Controlling microstructure improves battery energy density and reduces energy consumption. This study introduces the LASSO regularization method to accurately estimate 3D spherical size distribution, and demonstrates its superiority in predicting the distribution compared to other methods.