My Contributions:
-
Design and prototyping of multi-material printing hardware (SolidWorks)
-
Hyperspectral characterization of semiconductors
-
Wrote code for automated characterization (Python, C++)
-
Wrote code for non-convex machine learning optimization (Python)
Functional materials have vast and high-dimensional composition spaces which makes discovering optimized compositions intractable with conventional synthesis tools. Conventional experimental methods of exploring material composition spaces are slow and resource intensive due to being manual processes and requiring trial-and-error experimentation. Thus, the question is posed: How can we design an optimized functional material from this highly dimensional and vast composition space such that it has a high performance for a given application?
In this project, machine learning algorithms are integrated into novel, high-throughput synthesis hardware to accelerate the rate of material composition exploration by 1000x relative to conventional methods. First, inverse design algorithms predict a set of materials to synthesize based on user-defined target properties. Second, a multi-material printer, capable of printing thousands of unique experiments as droplets, synthesizes a large set of these predicted materials. Third, automatic characterization methods quantify the target properties of the experimentally synthesized materials. Fourth, computer vision and machine learning algorithms are integrated into the synthesis loop to autonomously discover synthesis conditions that generate optimized compositions without any intervention of a domain expert. Lastly, the inverse design algorithm is updated with the machine learning output to suggest a new set of materials to print until the target property is achieved.
The culmination of these processes enables the high-throughput and low-cost discovery of high-performance and novel functional materials for a wide array of applications.