Autonomous x-ray scattering

Abstract

Autonomous experimentation (AE) is an emerging paradigm that seeks to automate the entire workflow of an experiment, including---crucially---the decision-making step. Beyond mere automation and efficiency, AE aims to liberate scientists to tackle more challenging and complex problems. We describe our recent progress in the application of this concept at synchrotron x-ray scattering beamlines. We automate the measurement instrument, data analysis, and decision-making, and couple them into an autonomous loop. We exploit Gaussian process modeling to compute a surrogate model and associated uncertainty for the experimental problem, and define an objective function exploiting these. We provide example applications of AE to x-ray scattering, including imaging of samples, exploration of physical spaces through combinatorial methods, and coupling to in-situ processing platforms. These uses demonstrate how autonomous x-ray scattering can enhance efficiency, and discover new materials.

Roadmap on data-centric materials science

“…Beaucage and Martin report on the development of an open liquid handling platform for autonomous formulation and x-ray scattering [205]. Yager and coworkers review this new paradigm and show how autonomous x-ray scattering can enhance efficiency and help discover new materials [206].…”

Domain-specific chatbots for science using embeddings

“…An emerging trend in experimental sciences is autonomous experimentation (AE), wherein the measurement loop is closed using a decision-making algorithm that selects high-quality experiments to perform. 75–77 For instance, researchers have demonstrated that a synchrotron X-ray scattering beamline can autonomously explore physical parameter spaces, 78,79 reconstructing a high-quality model of the space and even discovering new materials or structures. 80 Existing approaches have typically used grounded ML modeling approaches (such as Gaussian process regression).…”

Deep Learning for Determination of Properties of Semidilute Polymer Solutions

“…The application of artificial intelligence (AI) in establishing structure–property correlations opened a new direction in polymer research. – The majority of research in this direction is focused on predicting polymer properties from a monomer chemical structure or their distribution along the polymer backbone. , In particular, data-driven approaches have utilized unsupervised learning and reverse engineering analysis methods , to expand the analysis of simulation and experimental data. – However, leveraging artificial intelligence to accelerate data analysis remains largely unexplored. , To address these challenges and the big data requirements needed for neural network training, we developed an approach that utilizes a scaling theory of semidilute polymer solutions – in conjunction with a convolutional neural network (CNN). – By generating large theoretical datasets using the confirmed scaling relationships, we eliminated the reliance on extensive experimental datasets. The CNN viability was tested on experimental datasets for the concentration dependence of specific viscosity in solutions of polymers with different molecular weights.…”