Automatic Determination of the Non-Covalent Stable Conformations of the NO 2 -Pyrene Cluster in Full Dimensionality (81D) Using the vdW-TSSCDS Approach.
Ramón L Panadés-BarruetaDenis DuflotJuan SotoEmilio Martínez-NúñezDaniel PeláezPublished in: Chemphyschem : a European journal of chemical physics and physical chemistry (2024)
We present the detailed topographical characterisation (stationary points and minimum energy paths connecting them) of the full dimensional (81D) intermolecular potential energy surface associated with the non-covalent interactions between the NO 2 radical and the pyrene (C 16 H 10 ) molecule. The whole procedure is (quasi) fully automated. We have used our recent algorithm vdW-TSSCDS as implemented on the freely-available AutoMekin software package. To this end, a series of inexpensive classical trajectories using forces from a low-level (semi-empirical) theory are used to sample the configuration space of the system in the search for candidates to first order saddle points. These guess structures are determined by means of a graph-theory based algorithm using the concept of adjacency matrix. Low-level optimizations are followed by re-optimizations at a final high-level of theory (DFT and CCSD(T)-F12 in our case.). The resulting set of stationary points and paths connecting them constitutes the so-called reaction network. In the case of NO 2 -pyrene, this network exhibits four major basins which can be characterized by their point-group symmetry. A central one, of global C 2 symmetry, comprises the global minimum (as well as all other permutationally related conformers) together with the corresponding C 2v saddle points connecting them. This central basin is connected to three others of lower C 1 symmetry. The latter can be distinguished by the projection of the position of the NO 2 nitrogen atom on the pyrene plane in combination with the relative orientation of the oxygen pair pointing either inwards, outwards, upwards or downwards.
Keyphrases
- deep learning
- machine learning
- neural network
- convolutional neural network
- liquid chromatography
- climate change
- depressive symptoms
- high resolution
- magnetic resonance imaging
- magnetic resonance
- minimally invasive
- density functional theory
- computed tomography
- molecular docking
- risk assessment
- image quality
- quantum dots
- crystal structure