Analyzing Small Angle Scattering Profiles of Soft Materials with Polydisperse Sizes and Shapes using CREASE

Advances in particle synthesis techniques have led to the emergence of nanoscale and colloidal particles that have anisotropic shapes. The anisotropy in the shapes of these building blocks can lead to the assembly of complex structures with unique electronic and optical properties. Small angle X-ray scattering (SAXS) has achieved prominence as a characterization technique for soft materials due to its ability to provide structural information about the system across multiple length scales. The structural information obtained through SAXS is in the reciprocal space and each SAXS profile can correspond to multiple three-dimensional (3D) real space structural interpretations. Therefore, there is a need for computational methods that convert the SAXS profile to all possible 3D structures representative of that profile. Our lab's CREASE method has been successfully utilized to identify families of structural features and corresponding representative structures from scattering profiles for a variety of soft material shapes and dimensions. In this work, we make use of machine learning enhanced CREASE method to identify ellipsoidal and spherical particle systems with distributions of shapes and sizes that act as possible 3D structural representations for the same experimental scattering profile. Experimental groups can then choose between the CREASE identified degenerate 3D structural representations based on complementary information from imaging techniques such as TEM.