Automated Functional Group Analysis of Mass Spectrometry via Machine Learning
Abstract
The icy plume of Enceladus has been sampled both by the cosmic dust analyzer (CDA) and the ion neutral mass spectrometer (INMS) aboard the Cassini mission to Saturn. Due to difficulties that arise in implementing mass spectrometry aboard planetary missions, enhanced computing methods will be beneficial to improving the data return of the Cassini mission. In an attempt to counter limitations of the mass spectrometers in terms of mass resolution and mass range, machine learning methods are utilized to explore options for mass spectrometry-based functional group analysis. Before this technique can be utilized on hyper velocity impact ionization spectra like those from the Cassini probe, it needs to be tested and validated on mass spectral data for known compounds. Multiple machine learning methods were tested using labeled electron ionization data to determine which methods are most promising for future development. This approach may be useful for interpretation of mass spectral data in planetary science missions. In which issues such as low instrument mass resolution and/or mass range, limited sample replicates and/or lack of chromatography, or tandem mass spectral analysis tools, can make unambiguous identification of specific chemical species challenging.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFM.P32A..03N