Apr 24, 2024
5:00pm - 7:00pm
Flex Hall C, Level 2, Summit
Brenden Pelkie1,Abdul Moeez1,Lilo Pozzo1
University of Washington1
Brenden Pelkie1,Abdul Moeez1,Lilo Pozzo1
University of Washington1
Many nanomaterial synthesis procedures require the selection of several experimental parameters to control the properties of the material that is produced. While this can provide detailed control over material properties, navigating this high dimensional parameter space to select values that enable desired outcomes can be an intractable process. While the applications of laboratory automation makes the systematic testing of parameter combinations more accessible, complete combinatorial explorations can still be inefficient or impossible for large systems. Closed loop optimization methods like Bayesian optimization provide a method to autonomously select experimental parameter values to test, on-the-fly during an ongoing experiment. These approaches use machine learning methods to select parameters that are likely to improve a material’s performance, as defined by a researcher-defined metric. Implementation of fully closed loop experiments requires capabilities for automated sample synthesis, characterization, and data processing. We are developing infrastructure to enable closed loop optimization for nanoparticle systems. To enable automated synthesis, we have developed an open-hardware robotic motion platform to enable flexible lab automation for advanced synthetic capabilities and automated flow synthesis tools for nanoparticle synthesis. Online characterization is enabled by an autosampler and automated data processing workflow for lab-scale x-ray scattering. These capabilities are applied to gain morphological control of various nanoparticle synthesis processes.