Closing the loop: autonomous experiments enabled by machine-learning-based online data analysis in synchrotron beamline environments

Abstract

Recently, there has been significant interest in applying machine-learning (ML) techniques to the automated analysis of X-ray scattering experiments, due to the increasing speed and size at which datasets are generated. ML-based analysis presents an important opportunity to establish a closed-loop feedback system, enabling monitoring and real-time decision-making based on online data analysis. In this study, the incorporation of a combined one-dimensional convolutional neural network (CNN) and multilayer perceptron that is trained to extract physical thin-film parameters (thickness, density, roughness) and capable of taking into account prior knowledge is described. ML-based online analysis results are processed in a closed-loop workflow for X-ray reflectometry (XRR), using the growth of organic thin films as an example. Our focus lies on the beamline integration of ML-based online data analysis and closed-loop feedback. Our data demonstrate the accuracy and robustness of ML methods for analyzing XRR curves and Bragg reflections and its autonomous control over a vacuum deposition setup.

Neural network analysis of neutron and X-ray reflectivity data incorporating prior knowledge

“…In recent years, machine learning has emerged as an alternative to classical methods of analysing surface scattering data (Hinderhofer et al, 2023), being attractive due to its very fast prediction times and its ability to be incorporated into the operating pipelines of large-scale measurement facilities. In particular, fast machine-learning-based solutions are ideal for enabling an experimental feedback loop during reflectometry measurements to be performed in real time (Pithan et al, 2023;Ritley et al, 2001). While many machine learning approaches dedicated to reflectivity exist (Greco et al, 2019(Greco et al, , 2021(Greco et al, , 2022Mironov et al, 2021;Doucet et al, 2021;Aoki et al, 2021;Kim & Lee, 2021;Andrejevic et al, 2022), most of them do not directly address the inherent ambiguity of the reflectivity data, instead training neural networks over specific parameter domains where the ambiguity is not prominent enough to prevent convergence.…”