Evaluating the predictive character of the method of Constrained Geometries Simulate External Force with Density Functional Theory.

## Abstract

from [1]:

Mechanochemistry is a fast-developing field of interdisciplinary research with a growing number of applications. Therefore, many theoretical methods have been developed to quickly predict the outcome of mechanically induced reactions. Constrained geometries simulate External Force (CoGEF) is one of the earlier methods in this field. It is easily implemented and can be conducted with most DFT codes. However, recently, we observed totally different predictions for model systems of epoxy resins in different conformations and with different density functionals. To better understand the conformational and functional dependence in typical CoGEF calculations we present a systematic evaluation of the CoGEF method for different model systems covering homolytic and heterolytic bond cleavage reactions, electrocyclic ring opening reactions and scission of non-covalent interactions in hydrogen-bond complexes. From our calculations we observe that many mechanochemical descriptors strongly depend on the functional used, however, a systematic trend exists for the relative maximum Force. In general, we observe that the CoGEF procedure is forcing the system to high energetic regions on the molecular potential energy profiles, which can lead to unexpected and uncorrelated predictions of mechanochemical reactions. This is questioning the true predictive character of the method.

## Contact

Christian R. Wick

Friedrich-Alexander-University Erlangen-Nürnberg (FAU), Faculty of Science, Department of Physics, PULS Group, Interdisciplinary Center for Nanostructured Films (IZNF), Cauerstrasse 3, 91058, Germany

## License

Creative Commons Attribution 4.0 International

## Context

Dataset to paper [1]

## Contents

• All COGEF trajectories in xyz format.
• All CoGEF distances and DFT Energies in csv format.
• compounds.json: JSON file, containing SMILES, InChI and InChIKey descriptors for all compounds investigated as dictionary

The following DFT levels of theory were investigated:

• B3LYP/6-31G(d)
• B3LYP-D3BJ/def2-SVP
• BP86-D3/def2-SVP
• PBE1PBE/def2-SVP
• M06-D3/def2-SVP

## Folder structure

• - compound_X : data set for compound number X (numbering corresponds to the numbering scheme in [1])
  • the xyz trajectories follow the following naming convention: "DFT_method"_"unrestricted/restricted".xyz
  • the csv files follow the naming convention: "DFT_method"_"unrestricted/restricted".xyz.csv

## Software

### COGEFF calculations: COGEF.py v1.8.0

Zenodo release:

https://doi.org/10.5281/zenodo.7079733

### DFT calculations:

Gaussian 16 Rev B [2]

## Funding

This research was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - 377472739/GRK 2423/1-2019 FRASCAL.

## References

[1] C. R. Wick, E. Topraksal, D. M. Smith, A.-S. Smith, "Evaluating the predictive character of the method of Constrained Geometries Simulate External Force with Density Functional Theory.", Forces in Mechanics, 9, 100143; doi:10.1016/j.finmec.2022.100143

[2] Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H.; et al. Gaussian 16 Rev. B.01, 2016.