[Case Study] Machine Learning Enabling Accurate Prediction of Precursor Volatility

Predict the evaporation or sublimation temperature with an accuracy of ±9°C on average, calculating hundreds of complexes per second.

A New Path to Precursor Development: Schrödinger's Machine Learning This predictive model opens a new avenue for designing new precursors with improved performance, optimizing not only the deposition and chemistry but also the temperature at which they can evaporate or sublime to be supplied as vapor. This advancement allows for a much broader range of structural changes to be screened computationally than before, enabling the generation of candidate precursors for experimental synthesis and testing that are less risky and more innovative. With this volatility model and the computational workflow for reactivity and decomposition based on Schrödinger's quantum mechanics, a complete design kit for vapor phase deposition and etching is provided, accelerating research on materials and processes for new technologies. *For 50 common metal and metalloid complexes, the evaporation or sublimation temperature at a given vapor pressure is predicted with an accuracy of ±9°C (about 3% of absolute temperature). *It can compute hundreds of complexes per second, resulting in a fast turnaround time. *For more details, please refer to the PDF document or feel free to contact us.

Software (middle, driver, security, etc.)