自驱动实验系统会自己“种”材料

Core Insights - A self-driven experimental system developed by the University of Chicago's Pritzker School of Molecular Engineering can autonomously synthesize and optimize materials without continuous human intervention [1][2] - This system integrates robotic automation and machine learning algorithms to create a closed-loop operation from experiment execution to performance measurement and result analysis [1][2] Group 1: System Overview - The system focuses on Physical Vapor Deposition (PVD) technology, which is sensitive to temperature, time, material purity, and environmental conditions, making accurate predictions challenging [2] - Traditional methods require manual adjustments and typically take over a day per experiment, leading to inefficiencies [2] - The new robotic system automates all PVD steps, including sample handling, film preparation, and performance testing [2] Group 2: Machine Learning Integration - Collaboration with computer scientists led to the development of specialized machine learning algorithms that guide the system in synthesis and analysis while dynamically adjusting experimental conditions [2] - Users only need to input desired film performance metrics, and the machine learning model autonomously plans the experimental path [2] Group 3: Performance Validation - The system was tested by aiming to produce silver films with specific optical properties, achieving the target in an average of just 2.3 experiments [3] - The self-driven system was able to explore various process conditions comprehensively, accomplishing what would take human teams weeks in just a few runs [3] Group 4: Cost Efficiency - The entire setup is significantly cheaper than previously developed commercial automated systems, costing an order of magnitude less [4]