Machine Learning Enabled Discovery of Complex Barocaloric Materials

Current refrigiration and air-coonditining technology is primarily based on vapor compression cycles of volatile hydrofluorocarbons, which have large global warming potential and whose unaviodable leakage into the atmosphere contribute significantly to global greenhouse gas emissision.

With the rapididly increasing deployment of air-conditioning and cooling worldwide, there is a great need for more efficient and clean cooling technology. Solid-state cooling based on barocaloric materials is a promising early-stage emerging such technology.

Here, an external mechanical pressure is used to reversibly drive a solid-solid phase transformation, which is used to design caloric cooling cycles.In this proposal, I will use and develop tools of modern computational and data-driven materials sciene to study, barocaloric rotationally disordered crystals, a particularily promising group of barocaloric materials.

These materials have phase transformations from a state with ordered to one with rotationally disorderd molecules.

This disordering process carries a large entropy change, and can be sensitively tuned by pressure, which allows for colossal barocaloric effects.I will utilize state-of-the-art machine learning interatomic potentials and develop a workflow to computationally screen, discover and understand this class of promising materials.

The objective is to provide a route towards the discovery of materials for future efficient solid-state cooling technology.