Predicting binding energies of astrochemically relevant molecules via machine learning

Abstract

The behaviour of molecules in space is to a large extent governed by where they freeze out or sublimate. The molecular binding energy is thus an important parameter for many astrochemical studies. This parameter is usually determined with time-consuming experiments, computationally expensive quantum chemical calculations, or the inexpensive, but inaccurate, linear addition method. In this work we propose a new method based on machine learning for predicting binding energies that is accurate, yet computationally inexpensive. A machine learning model based on Gaussian Process Regression is created and trained on a database of binding energies of molecules collected from laboratory experiments presented in the literature. The molecules in the database are categorized by their features, such as mono- or multilayer coverage, binding surface, functional groups, valence electrons, and H-bond acceptors and donors. The performance of the model is assessed with five-fold and leave-one-molecule-out cross validation. Predictions are generally accurate, with differences between predicted and literature binding energies values of less than $\pm$20\%. The validated model is used to predict the binding energies of twenty one molecules that have recently been detected in the interstellar medium, but for which binding energy values are not known. A simplified model is used to visualize where the snowlines of these molecules would be located in a protoplanetary disk. This work demonstrates that machine learning can be employed to accurately and rapidly predict binding energies of molecules. Machine learning complements current laboratory experiments and quantum chemical computational studies. The predicted binding energies will find use in the modelling of astrochemical and planet-forming environments.

Accepted in astronomy and astrophysics

References in corpus (15)

• Formation of methyl formate and other organic species in the warm-up phase of hot molecular cores
• Astrochemistry and compositions of planetary systems
• Desorption of CO and O2 interstellar ice analogs
• Discovery in space of ethanolamine, the simplest phospholipid head group
• Binding energies of interstellar molecules on crystalline and amorphous models of water ice by ab-initio calculations
• Early Science from GOTHAM: Project Overview, Methods, and the Detection of Interstellar Propargyl Cyanide (HCCCH${}_2$CN) in TMC-1
• Propargylimine in the laboratory and in space: millimetre-wave spectroscopy and first detection in the ISM
• Isomers in Interstellar Environments (I): The Case of Z- and E-Cyanomethanimine
• Ammonia snow-lines and ammonium salts desorption
• Machine Learning of Interstellar Chemical Inventories
• Desorption Kinetics and Binding Energies of Small Hydrocarbons
• Nitrile versus isonitrile adsorption at interstellar grains surfaces: I - Hydroxylated surfaces
• Thermal desorption of astrophysically relevant molecules from forsterite(010)
• Diffusion and Clustering of Carbon Dioxide on non-porous Amorphous Solid Water
• Thermal desorption of astrophysical relevant ice mixtures of acetaldehyde and acetonitrile from olivine dust

Cited by in corpus (10)

• Carbon-Chain Chemistry in the Interstellar Medium
• Laboratory and Computational Studies of Interstellar Ices
• An overview of desorption parameters of Volatile and Complex Organic Molecules: A systematic dig on experimental literature
• Machine learning-accelerated chemistry modeling of protoplanetary disks
• Astrochemical Modeling of Propargyl Radical Chemistry in TMC-1
• Molecular Mobility of Extraterrestrial Ices: Surface Diffusion in Astrochemistry and Planetary Science
• Simulation of CH${}_3$OH ice UV photolysis under laboratory conditions
• Infrared Spectra of Solid-State Ethanolamine: Laboratory Data in Support of JWST Observations
• Identifying the most constraining ice observations to infer molecular binding energies
• Identification of molecular line emission using Convolutional Neural Networks