Beyond the Ground State: Predicting Electron Ionization Mass Spectra Using Excited-State Molecular Dynamics

Area of Science:

• Computational Chemistry
• Spectroscopy
• Metabolomics

Background:

• Predicting electron ionization mass spectra is crucial for small molecule identification in metabolomics.
• Current methods often rely on ground-state approximations, limiting accuracy.
• Incorporating excited states can provide a more comprehensive understanding of ionization processes.

Purpose of the Study:

• To develop and validate an algorithm that includes excited states in molecular dynamics for predicting 70 eV electron ionization mass spectra.
• To improve the accuracy of mass spectral predictions for small molecules, particularly in metabolomics.
• To identify fragmentation reactions missed by ground-state methods.

Main Methods:

• Utilized the binary-encounter-Bethe (BEB) model to calculate ionization cross sections for electronic states.
• Employed a fast orthogonalization model/single and double configuration interaction (OM2/CISD) method combined with the GFN1-xTB semiempirical model for excited state calculations.
• Applied excited-state molecular dynamics to predict mass spectra and compared results with ground-state methods.

Main Results:

• Excited-state molecular dynamics demonstrated improved accuracy compared to ground-state calculations for urocanic acid mass spectra.
• Excited-state corrections identified significantly more true positive ions for histidine pathway intermediates than OM2 or GFN methods.
• The excited-state models successfully predicted fragmentation reactions absent in ground-state predictions.
• The mixed excited-state method achieved the best average mass spectral similarity scores across 48 molecules.

Conclusions:

• Adding excited-state calculations is recommended for predicting electron ionization mass spectra of small molecules in metabolomics.
• The developed algorithm enhances the predictive power for spectral analysis and fragmentation patterns.
• This approach offers a more accurate and comprehensive tool for molecular identification.