Decoding defect statistics from diffractograms via machine learning

Cody Kunka,Rémi Dingreville,Surya R Kalidindi,Apaar Shanker,Elton Y Chen

doi:10.1038/s41524-021-00539-z

Abstract

Diffraction techniques can powerfully and nondestructively probe materials while maintaining high resolution in both space and time. Unfortunately, these characterizations have been limited and sometimes even erroneous due to the difficulty of decoding the desired material information from features of the diffractograms. Currently, these features are identified non-comprehensively via human intuition, so the resulting models can only predict a subset of the available structural information. In the present work we show (i) how to compute machine-identified features that fully summarize a diffractogram and (ii) how to employ machine learning to reliably connect these features to an expanded set of structural statistics. To exemplify this framework, we assessed virtual electron diffractograms generated from atomistic simulations of irradiated copper. When based on machine-identified features rather than human-identified features, our machine-learning model not only predicted one-point statistics (i.e. density) but also a two-point statistic (i.e. spatial distribution) of the defect population. Hence, this work demonstrates that machine-learning models that input machine-identified features significantly advance the state of the art for accurately and robustly decoding diffractograms.

Highlights

Diffraction techniques can probe large volumes of material while maintaining high resolution in both space and time[1,2,3,4]
We examined virtual selected area electron diffraction (SAED) patterns produced from atomistic simulations of irradiated copper
Choosing defect statistics To understand the defect statistics generated from our atomistic simulations, we first examined the defect mechanisms as a function of the number of irradiation events, which we iteratively introduced over time

Summary

Introduction

Diffraction techniques can probe large volumes of material while maintaining high resolution in both space and time[1,2,3,4] These techniques are widely used to provide structural characterizations across a variety of scientific fields, including biology[5,6,7], materials science[1,8,9], and polymer physics[10,11]. The difficulty of decoding diffractograms has greatly limited their utility[12] Challenges concern both steps of the decoding process: (i) identifying the key features in the diffractogram and (ii) modeling their relationships to the desired structural characterizations. Assessments of 1D diffraction line profiles, such as from conventional X-ray diffraction (XRD), have focused on the positions and widths of the peaks. We recently demonstrated that fitting two popular width models with the same XRD data yielded opposite trends in characteristic size[13]

Methods

Results

Conclusion