Logo image
A Bayesian desmearing algorithm for Bonse-Hart USANS with anisotropic scattering
 

A Bayesian desmearing algorithm for Bonse-Hart USANS with anisotropic scattering

Chi-Huan Tung, Guan-Rong Huang, Yangyang Wang, Jan-Michael Carrillo, Yuya Shinohara, Changwoo Do, Gernot Rother, Yingrui Shang Wei-Ren Chen
The Journal of chemical physics, Vol.164(16), 164108
28/04/2026
: 42023729
Chemistry Chemistry, Physical Physical Sciences Physics Physics, Atomic, Molecular & Chemical Science & Technology
Ultra-small-angle neutron scattering (USANS) using Bonse-Hart optics provides micrometer-scale structural insights but suffers from severe slit-geometry smearing. While well-established for isotropic systems, quantitative desmearing of anisotropic data remains a challenge because conventional corrections break down for non-radial scattering. We address this by developing a resolution-aware Bayesian framework that explicitly incorporates anisotropy via an affine deformation to the scattering pattern, guided by the principle of parsimony. This results in orientation-resolved point-spread functions that enable a self-consistent determination of both the resolution and deformation parameters. Using Gaussian process regression with uncertainty quantification and a probabilistic correction for multiple scattering, we demonstrate the framework's effectiveness through numerical benchmarks and experimental studies of a stretched polymer melt. Our approach enables the seamless integration of SANS and USANS data, facilitating quantitative structural analysis of deformed materials at nanometer to micrometer scales.
1
Logo image