Official websites use .gov
A .gov website belongs to an official government organization in the United States.
Secure .gov websites use HTTPS
A lock (
) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.
Summary
The Autonomous Formulation Lab (AFL) at NIST links AI-driven experimental planning with robotic experimental platforms and in-situ small-angle scattering (SANS/SAXS/USAXS) and other flow-cell compatible measurements (e.g., rheology, specstrocopy). The AFL prepares and delivers complex fluid mixtures to flow cells, runs measurements, and learns from results in a closed loop to map phase behavior, understand physical properties, and optimize formulations faster. Built with open-source software and modular hardware developed with NCNR and the nMat consortium, AFL supports partners with reproducible methods, data standards, and shareable datasets to accelerate materials innovation.
Description
The Autonomous Formulation Laboratory is a joint program between the nMat consortium, NCNR, and NIST's Material Measurement Lab focused on accelerating materials discovery and formulation optimization through artificial intelligence (AI) and machine learning (ML) directed, multimodal scattering experiments. The core of the AFL platform is an open source platform to prepare liquid mixtures via pipetting, transfer those mixtures to a measurement cell, perform a SANS (or SAXS or other method) experiment, and provide the data to an AI guidance server.
Formulation engineering is at the core of soft materials-based products that cut across industry and everyday life: paints, vaccines, fuels, and shampoos are just a few examples of products that are formulated with tens to hundreds of individual components. Companies frequently need to (re)formulate products to address environmental regulations, reduce cost, improve performance, etc. Frequently, changes in one component (for example, replacing a fragrance) can cause complex structural changes that affect the stability bounds of a given product and ultimately impact performance in unexpected ways. The AFL seeks to map changes to underlying structure to enable rational formulation design.
The AFL is being developed by NIST NCNR and MML staff with beamtime from other facilities such as the Cornell CHESS synchrotron, the Advanced Photon Source at Argonne, the ISIS Neutron source in the UK, and the Swiss spallation neutron source at the Paul Scherrer Institute in Switzerland. The AFL platform provides manual/programmed grid and autonomous, AI-driven measurements across multiple techniques and user facilities. Key neutron capabilities will be to automatically determine compositions for contrast matched samples, and determine multi-component phase maps for surfactant formulations.
The AFL is a radically open platform; our code and instrument designs are public on GitHub in the AFL-automation, AFL-agent, AFL-hardware packages. The design of the platform is further documented in our 2023 Chemistry of Materials publication, which includes extensive details in the Supporting Information, and in forthcoming papers.