ML-enhanced Fast and Interpretable Simulation of IR Spectra

Theoretical IR (infrared) spectroscopy is a powerful tool for assisting chemical structure identification. However, approaches based on quantum chemical calculations suffer from either high computational cost (e.g., density functional theory, DFT) or insufficient accuracy (semi-empirical methods). 

Hence, we introduce a new approach based on the universal machine learning models of AIQM series targeting the gold-standard coupled-cluster level, going beyond the typical DFT accuracy (see our preprint). 

AIQM methods, particularly, newly introduced AIQM2, can provide IR spectra with accuracy close to DFT and the speed close to a semi-empirical GFN2-xTB method. To ensure high speed and interpretability, our implementation is based on the harmonic approximation with the frequenciesscaled by factors that we found empirically.

You can easily do such calculations yourself using MLatom, which comes with tutorials and example scripts. We also offer easy-to-use tool for interpretation of the IR spectra by visualizing vibrational normal modes and assigning them to corresponding IR bands.

Preprint:

• Yi-Fan Hou, Cheng Wang, Pavlo O. Dral*.Accurate and Affordable Simulation of Molecular Infrared Spectra with AIQM Models. 2024.
  Preprint on ChemRxiv: https://doi.org/10.26434/chemrxiv-2024-r5prt (2024-11-08).