New machine learning model advances atmospheric science
Large fraction of atmospheric aerosol particles are formed in the air by condensing organic vapors. Aerosol particles can further grow to form cloud condensation nuclei. Aerosol particles containing different types of organic molecules have varying chances of forming cloud droplets.
The formation of cloud droplets can be modelled using thermodynamic properties, such as saturation vapor pressure and phase equilibria.
Thermodynamic properties can be estimated with computational methods. However, the most accurate thermodynamic models require a lot of computer time.
A machine learning model developed at the Ä¢¹½Ö±²¥ will speed up the calculations of large organic compounds. Thermodynamic properties can therefore be calculated for a larger set of molecules. These properties can be further utilized in global climate models to predict the effects of aerosol particles on climate change.
The research results were published 19th of October 2022 in The Journal of Physical Chemistry Letters: .
The research was funded by Academy of Finland. The research was carried out at the Ä¢¹½Ö±²¥ in the Computational Nanoscience research group of the Nanoscience Center.
For further information:
Postdoctoral Researcher Noora Hyttinen, noora.x.hyttinen@jyu.fi