The project NAMDIA was divided into different subtasks: (i) Theoretical study of a typical molecule, HPALD, and its photochemical reactions upon light absorption; (ii) Develop and benchmark a protocol to determine if a given VOC molecule is likely to absorb light under sun irradiation; (iii) Develop new theoretical methodologies for the simulation of the light-absorption process and the generation of simple kinetic models. Finally, a central goal of the NAMDIA project was to bridge theory and experiment around the question of VOC photochemistry.
The VOC molecule HPALD-C6 is a close parent of an important VOC molecule and is stable enough to be studied experimentally. HPALD molecules have been proposed to release OH radicals upon light irradiation, and could explain the larger concentration of OH radicals often measured in specific regions such as on top of tropical forests as compared to predicted theoretical value. We performed an in-depth characterization of the photochemistry of HPALD-C6 by simulating its different electronic states, how they are coupled together, and simulate all-atom dynamics of the molecule after light absorption. A clear limit of the methodology employed is the level of theory to describe the electronic structure of HPALD-C6, a constant issue with theoretical photochemical studies. To alleviate this problem, we developed a model for HPALD (based on Zhu-Nakamura theory and implemented in the software Mesmer) in a reduced number of nuclear dimensions, selected based on our all-atom simulations, and we performed high-level electronic structure calculations to feed the model. Such model indicates that specific excited-state effects (`nonadiabatic effects') increases the lifetime of the excited molecule by a factor of at least two, and highlighted an exciting mechanism known as "diabatic trapping". Additional calculations are currently ongoing to validate our results.
An important question that arose from project (i) is: how to simulate sunlight absorption explicitly? We developed a new theoretical method, coined XFAIMS, that includes explicitly the effect of light during the dynamics of a molecule, naturally promoting it in the excited states (J. Chem. Phys., 2016). A second project emanated during discussions with the spectroscopy group at the University of Bristol (Prof. Andrew Orr-Ewing) on the mechanism of 'intersystem crossings processes'. These discussions led to a successful experimental/theoretical collaboration, during which we studied the process of intersystem crossings for a cobalt complex (Ang. Chem. Int. Ed., 2017).
We benchmarked most of the available strategies to determine the absorption spectra of VOC molecules, and compared their result to experimental ones. We developed different protocols based on the nature of the studied molecule and applied those protocol to predict the absorption spectra HPALD-C6 (which is not available experimentally). In collaboration with two experimental groups (Prof. Andrew Orr-Ewing and Dr. Max Mcgillen), we employed these strategies to predict the absorption spectra of molecules produced by the reaction of a Criegee's intermediate with simple alcohols (ACS Earth Space Chem., 2017).
Computing accurately and efficiently the electronic configuration of molecules of the size of atmospheric VOCs constitutes a critical bottleneck for nowadays theoretical methods. We therefore developed a new strategy to push the limits of the gold-standard method called 'EOM-CCSD'. The technique was tested on a typical atmospheric molecule, acrolein, and was efficient enough to study it in isolation as well as in the aqueous phase (J. Phys. Chem. Lett, 2017).
