P011-07
Using Graph-Based Reaction Enumeration Techniques to Find Hidden Atmospheric Chemical Reactions
Monday, 7 December 2020: 16:24
Virtual
Romulo Cruz1,2, Jessica Ray1, Siddhant Sharma1, Eduardo A Lozano Garcia1, Aayush Arya1, Jakob Lykke Andersen3, Henderson James Cleaves II4,5, Rana Doğan1 and Paul Rimmer6, (1)Blue Marble Space Institute of Science, Young Scientist Program, Seattle, WA, United States, (2)National University of Engineering, Laboratory of Physical Chemistry Research, Faculty of Sciences, Av. Tupac Amaru 210, Rimac, Lima., Peru, (3)University of Southern Denmark, Department of Mathematics and Computer Science, Odense M, Denmark, (4)Tokyo Institute of Technology, Earth-Life Science Institute, Tokyo, Japan, (5)Institute for Advanced Study, School of Natural Sciences, Princeton, NJ, United States, (6)University of Cambridge, Department of Earth Sciences, Cambridge, United Kingdom
Abstract:
When studying the atmospheric composition of other planets using mass spectrometric or IR spectroscopic techniques, there are likely to be signals that could be difficult to assign to a specific molecule type, or whose identification could be questioned by the fact that it does not have a reasonable synthesis route [1]. This suggests molecules studied in atmospheric chemistry can be divided into four sets: 1. measured and correctly identified species, 2. measured and incorrectly identified species, 3. unmeasured and probable species, and 4. unmeasured and improbable species. Species in set 3 can be identified in two ways: First, by observing a phenomenon that needs an explanation, then writing out possible mechanisms for reactions those species could be combinatorially involved in, or second, by making a map of all possible mechanisms, which allows one to have on hand a library of possible mechanisms, which can be prioritized in terms of probability according to abundance of species, network connectedness, presence of autocatalytic loops, etc.
We have attempted to build such a map, in part using the MedØlDatschgerl (MØD) reaction enumeration software [2], previously developed for automatically generating reaction networks from graph grammar formulations of chemistries. MØD allowed construction of a network that contains reactions for various common and otherwise plausible atmospheric chemical species containing H, C, N, O and S, restricted to molecules with no more than 6 heavy atoms (e.g. those besides H) to manage computation time. The network is initiated with a set of seed molecules and in each iteration all the reactions made possible by a constructed reaction library are applied. We also employ a list of “forbidden molecules”, based on chemical stability considerations or steric impediments related to the aforementioned set 4 species. We explain here the study of reaction networks generated in various atmosphere types, detailing their evolution during each iteration. Highly connected reactions in these networks are identified and the modeled results compared with experimental and observational measurements.
References
[1]. Cleaves, H. J., et al. (2019). J Chem Inf Model, 59(10), 4266-4277.
[2]. Andersen, J. L., et al. (2017). Philos. Trans. R. Soc. A, 375(2109), 20160354.
