First World Congress on Artificial Intelligence in Materials and Manufacturing (AIM 2022): Autonomous Materials Research
Program Organizers: Taylor Sparks, University of Utah; Michael Dawson-Haggerty, Kerfed, Inc.; Elizabeth Holm, University of Michigan; Jin Kocsis, Purdue University; Adam Kopper, Mercury Marine; Benji Maruyama; James Warren, National Institute of Standards and Technology
Tuesday 9:30 AM
April 5, 2022
Room: William Penn Ballroom
Location: Omni William Penn Hotel
Session Chair: Howard Joress, National Institute of Standards and Technology
9:30 AM Break
10:00 AM Invited
Employing Artificial Intelligence to Accelerate Development and Implementation of Materials and Manufacturing Innovations: Elizabeth Holm1; George Spanos; 1Carnegie Mellon University
Although artificial intelligence (AI) has great potential for enabling and accelerating materials and manufacturing (M&M) innovations, there is currently not a consistent plan for how to implement AI techniques into materials and materials-related manufacturing processes. Thus, on behalf of the Office of Naval Research and the National Institute of Science and Technology, The Minerals, Metals & Materials Society (TMS) has engaged in a science and technology accelerator study on Employing Artificial Intelligence to Accelerate Development and Implementation of Materials and Manufacturing Innovations. This recently completed study and report are the result of the volunteer efforts of a team of internationally known subject matter experts, and this presentation will provide some key highlights, concerning in particular (1) the value proposition and key application areas of AI in M&M, (2) key gaps, barriers, and enablers, and (3) recommendations and detailed action plans to help realize the great potential of AI in M&M.
10:30 AM
Autonomous Corrosion Resistant Coatings Development Using a Scanning Droplet Cell: Howard Joress1; Jason Hattrick-Simpers2; Najlaa Hassan3; Trevor Braun1; Justin Gorham1; Brian DeCost1; 1NIST; 2University of Toronto; 3University of Wisconsin
Electrochemical deposition of alloy coatings is a difficult process to study and optimize, having a large parameter space including electrolyte chemistry, fluid flow, and electronic parameters.This makes autonomous materials science ideal for development of new alloys and increased understanding of electrochemical deposition. To this end, we have developed a fully automated, millifluidic scanning droplet cell (SDC) system capable of both synthesis via electrodeposition and characterization via electrochemical corrosion, macroscopic imaging, and roughness measurements. Because the tool is flexible and highly automated it is readily compatible with active machine learning, creating an autonomous materials design platform. We demonstrated use of this tool as an autonomous corrosion analysis tool on prefabricated libraries as well as for close-loop electrodeposition and characterization campaigns. We will discuss the various components of the autonomous research cycle using the SDC including synthesis, automating corrosion measurements and analysis, and utilizing machine learning to control the instrument.
10:50 AM
Automating the Discovery of New Halide Perovskites with RAPID and ESCALATE: Joshua Schrier1; 1Fordham University
In this talk, I will describe our Robotic-Accelerated Perovskite Investigation and Discovery (RAPID) project. The first generation of RAPID uses inverse temperature crystallization (ITC) to grow halide perovskite single crystals for x-ray structure determination and bulk characterization using commercial liquid handling robots. The second generation of RAPID uses antisolvent vapor diffusion, expanding the types of chemical processes we can study. In parallel, we’ve developed a reusable data management software, ESCALATE (Experiment Specification, Capture and Laboratory Automation Technology) to facilitate data and metadata capture. Using the RAPID+ESCALATE technology stack, we’ve been able to discover dozens of new compounds, increased scientist productivity, used statistical “automated serendipity” to use ambient condition fluctuations to identify (and subsequently test) new hypotheses regarding crystal formation, and benchmarked a variety active- and meta-learning algorithms for experiment selection in a laboratory setting.
11:10 AM
The nSoft Autonomous Formulation Laboratory: X-Ray and Neutron Scattering for Industrial Formulation Discovery: Peter Beaucage1; Tyler Martin2; 1National Institute of Standards and Technology; 2NIST
While scattering methods (SAXS, SANS, WAXS) are workhorse techniques for characterizing model macromolecular formulations, they have not been widely used to characterize real products, largely because the large number of components (10-100) often precludes rational mapping between component fractions, structure, and product stability. Multimodal characterization and machine learning (ML) tools promise to greatly reduce the expense of exploring the stability boundaries of a particular, desirable phase in highly multicomponent products. Here we describe the development of the Autonomous Formulation Laboratory, a highly adaptable platform capable of autonomously synthesizing and characterizing liquid mixtures with varying composition and chemistry using x-ray and neutron scattering in addition to a suite of secondary measurements such as optical imaging, UV-vis-NIR and capillary rheometry.
11:30 AM
Unsupervised Topological Learning Approach for Crystal Nucleation in Pure Metals and Alloys: Sebastien Becker1; Emilie Devijver1; Rémi Molinier1; Philippe Jarry1; Noel Jakse1; 1Université Grenoble-Alpes
Theoretical understanding of crystal nucleation is still a challenging issue as experimental confirmation remains out of reach for bulk materials. Large-scale atomic-level simulations are therefore a promising substitute for such experiments, and molecular dynamics (MD) of million to billion atoms may indeed lead to meaningful results. Machine Learning (ML) tools propose powerful methods to analyse such a large amount of MD-generated big data. An unsupervised ML approach based on topological descriptors using persistent homology concepts is proposed to reveal the structural features of atomic arrangements without a priori knowledge on the studied system. This approach is applied to monatomic metals and extended to aluminium-based alloys. Both translational and orientational orderings are thus evidenced together with nucleation pathways, whose revealed features are beyond the hypotheses of the Classical Nucleation Theory. This promising methodology more generally opens the route to an autonomous and in-depth investigation of atomic level mechanisms in material science.
11:50 AM
Machine Learning-Guided Materials Discovery Enabled by Automated Experimentation: Olexandr Isayev1; 1Carnegie Mellon University
Modern polymer science is plagued by the curse of multidimensionality; the large chemical space imposed by including combinations of monomers into a statistical copolymer overwhelms polymer synthesis and characterization technology and limits the ability to systematically study structure–property relationships. To tackle this challenge in the context of 19F MRI agents, we pursued a computer-guided materials discovery approach that combines synergistic innovations in automated flow synthesis and machine learning (ML) method development. A software controlled, continuous polymer synthesis platform was developed to enable iterative experimental–computational cycles that resulted in the synthesis of 397 unique copolymer compositions within a six-variable compositional space. The non-intuitive design criteria identified by ML, which was accomplished by exploring less than 0.9% of overall compositional space, upended conventional wisdom in the design of 19F MRI agents and lead to the identification of >10 copolymer compositions that outperformed state-of-the-art materials.