The Data NanoAnalytics Group’s mission is “To accelerate the development of autonomous research tools and workflows capable of scientific discovery in nanoscale synthesis and characterization by combining simulations, physics-driven machine learning methods and instrument automation.”
Data NanoAnalytics
Our group is known for developing and implementing physics-driven machine learning experimental workflows to increase the capabilities of instruments for both nanoscale characterization and synthesis. To this end, we develop and apply machine learning tools to analyze multi-modal data streams, including imaging data from microscopy, hyperspectral data (e.g., STEM-EELS) and other forms of spectroscopy.
We utilize these algorithms, alongside human-in-the-loop knowledge injection, automation, simulations, and edge computing, to steer autonomous instruments and assist in materials modification, optimization, and ultimately, in discovering new physics. We are well-recognized for our physics-driven approach to autonomous science, through awards (e.g., R&D100 Award 2023), granted patents, and a history of publications and invitations to lecture on this topic.
We use these new capabilities to explore the frontiers of nanoscience: to study and perturb interfaces in ferroelectric and ferroelastic materials, explore structural phase transitions as a function of electrical stress and temperature, manipulate topological structures, rearrange individual atoms deterministically to form structures with designer properties with a microscope, and optimize growth of chalcogenides to enhance opto-electronic properties, amongst others.
