Leveraging Automation and Machine Learning To Bridge the Sustainability Data Gap

Abstract Materials with minimal environmental impact are urgently needed if we want to meet climate goals. The chemical industry, responsible for approximately 5% of global greenhouse gas emissions, is a critical sector for reducing or eliminating human-caused carbon dioxide. Designing inherently sustainable products and processes, emphasizing renewable feedstocks, greener solvents, and energy-efficient methods are at the core of “sustainable chemistry”. However, robust metrics and comprehensive chemical data are needed to measure the progress toward sustainability. Life cycle sustainability assessment is a key framework for fully quantifying potential sustainability impact but has several limitations that need to be overcome including a disconnect between the data demands of AI for predictive sustainability metrics and the data that exist. Advanced data-driven methods, coupled with standardized data workflows, can help bridge these gaps. Integrating ML with automated experimentation enables closed-loop optimization, embedding sustainability directly into chemical process design while acquiring high fidelity data, which can be later used.