DeepStruc: Towards structure solution from pair distribution function data using deep generative models

Abstract

Structure solution of nanostructured materials that have limited long-range remains a bottleneck in materials development. We present a deep learning algorithm, DeepStruc, that can solve a simple nanoparticle structure directly from a Pair Distribution Function obtained from total scattering data by using a conditional variational autoencoder (CVAE). We first apply DeepStruc to PDFs from seven different structure types of monometallic nanoparticles, and show that structures can be solved from both simulated and experimental PDFs, including PDFs from nanoparticles that are not present in the training distribution. We also apply DeepStruc to a system of hcp, fcc and stacking faulted nanoparticles, where DeepStruc recognizes stacking faulted nanoparticles as an interpolation between hcp and fcc nanoparticles and is able to solve stacking faulted structures from PDFs. Our findings suggests that DeepStruc is a step towards a general approach for structure solution of nanomaterials.

Extracting structural motifs from pair distribution function data of nanostructures using explainable machine learning

“…Compared to crystallographic methods relying on long-range order, PDF analysis can be applied for nanomaterials [3][4][5] , disordered 1,6,7 , or amorphous materials 3,5,8 . However, structure solution from the PDF is not possible except in a very few simple cases 9 , using either the Reverse Monte Carlo method 10 or the LIGA algorithm 11,12 . In the absence of broadly applicable ab initio nanostructure determination methods, it is, therefore, necessary to propose reasonable starting models and to then 'refine' the model parameters against the data using local minimization methods.…”